AAPS PharmSciTech Volume 10, Number 2 6.09a PDF

AAPS PharmSciTech, Vol. 10, No.
2, June 2009 (# 2009)

DOI: 10.1208/s12249 009 9209 2
Research Article
Glycosylation of Aromatic Amines I: Characterization of Reaction Products
and Kinetic Scheme
Madhushree Y. Gokhale,1,3,4 William R. Kearney,2 and Lee E. Kirsch1
Received 21 August 2008; accepted 19 February 2009; published online 21 March 2009
Abstract. The reactions of aliphatic and aromatic amines with reducing sugars are important in both drug
stability and synthesis. The formation of glycosylamines in solution, the rst step in the Maillard reaction,
does not typically cause browning but results in decreased potency and is hence signicant from the
aspect of drug instability. The purpose of this research was to present (1) unreported ionic equilibria of
model reactant (kynurenine), (2) the analytical methods used to characterize and measure reaction
products, (3) the kinetic scheme used to measure reaction rates and (4) relevant properties of various
reducing sugars that impact the reaction rate in solution. The methods used to identify the reversible
formation of two products from the reaction of kynurenine and monosaccharides included LC mass
spectrometry, UV spectroscopy, and 1 D and 2 D 1H 1H COSY NMR spectroscopy. Kinetics was studied
using a stability indicating HPLC method. The results indicated the formation of and glycosylamines
by a pseudo rst order reversible reaction scheme in the pH range of 1 6. The forward reaction was a
function of initial glucose concentration but not the reverse reaction. It was concluded that the reaction
kinetics and equilibrium concentrations of the glycosylamines were pH dependent and also a function of
the acyclic content of the reacting glucose isomer.
KEY WORDS: glucose; glycosylamines; glycosylation; kinetics; Maillard reaction; reducing sugar
aromatic amine.
INTRODUCTION
The reactions of aliphatic and aromatic amines with
aldehydes to form imines are important in both drug stability
and synthesis. The reactions of glycosidic aldehydes with
amines involve the added complexity of monosaccharide
equilibria in solution whereby the reversible formation of
an imine (acyclic sugar moiety) is in equilibrium with the
glycosylamine isomers (the reduced cyclic sugars) (1,2).
Amadori products may subsequently form and give rise to
additional degradation products. The complex so called
Maillard reaction pathways result in potency loss and
product discoloration. All reducing sugars have been
implicated in these browning reactions including glucose,
lactose, galactose, mannose, and dextrates (3). Glucose is
commonly used as a solute in solutions for parenteral
administration; lactose is used as an excipient in both
parenteral and enteric formulations.
Division of Pharmaceutics, College of Pharmacy, The University of

Iowa, Iowa City, Iowa 52242, USA.
2
The Nuclear Magnetic Resonance Facility, College of Medicine, The
University of Iowa, Iowa City, Iowa 52242, USA.
3
Bristol Myers Squibb Company, 1 Squibb Drive, New Brunswick,
New Jersey 08903, USA.
4
To whom correspondence should be addressed.
(e mail: madhush reeg@gmail.com)
Examples of drug instability due to the Maillard

reaction include both primary and secondary amines as in
the case of amphetamines, insulin, uoxetine, sulfa drugs,
procainamide, metoclopropamide, and daptomycin to name
a few (4 9). For example, Duval et al. studied the browning
of dextroamphetamine sulfate solutions containing lactose
and found that the solution darkens on storage at 50C with
the formation of brownish precipitate (4). Using infrared
data and thin layer chromatography, they concluded that the
product was a Schiff base formed from the reaction of the
primary amine in dextroamphetamine and the carbonyl
group of lactose. Secondary amines in generic formulations
of uoxetine HCl containing lactose as a common ingredient
were shown to be less stable than the formulations contain
ing starch due to the Maillard reaction between the drug and
lactose (5). Dextrates obtained by the controlled hydrolysis
of starch used as excipients are also known to exhibit the
browning reaction due to the presence of varying amounts
of dextrose (4,10,11).
The majority of previous studies primarily focused on
reactions of glucose with model aliphatic amines such as
amino acids (12 16) and the browning process due to the
formation of Amadori products and advanced Maillard
reaction products (15,17 21). However, the formation of
glycosylamines in solution, the rst step in the Maillard
reaction, does not typically cause browning but results in
decreased potency and is hence signicant from the aspect of
drug instability. The glycosylamines of cyclic sugars can exist
in equilibria in solution predominantly as the and forms
317
1530-9932/09/0200-0317/0 # 2009 American Association of Pharmaceutical Scientists
318
similar to the parent sugar though the equilibrium composi
tions may differ. The , glycosylamines mutarotate and
hydrolyze through the acyclic imine (22,23).
In the area of drug stability, only a few cases have been
reported for reactions of weakly basic aromatic amine
containing drugs in the presence of reducing sugars like
glucose forming glycosylamines. For example, the reversible
formation of glycosylamines for procainamide in the presence
of 5% dextrose solution was reported in water (24,25). In
another study, a 20 30% procainamide potency loss (aromat
ic amine pKa =2.75) in 24 h was reported following admixing
with glucose solutions (26). The formation of glucosylamines
has also been reported for the oral suspension of sulfame
thoxazole (aromatic amine pKa =1.69) in the presence of
glucose and a limit for the amount of glycosylamines
allowed in the suspension has been reported for such
formulations (27). In another study involving dissolution
testing of Bactrim DS tablets (sulphamethoxazole, 800 mg
and trimethoprim, 160 mg) in the presence of 5% glucose,
glucosylamines readily formed with the sulphamethoxazole
and suggested the potential of reduced bioavailability of
sulphamethoxazole in the presence of dietary glucose (7).
Lucida et al. studied the reaction of glucose and sulphame
thoxazole and showed a pH and temperature dependent
rate of reversible formation of glycosylamines in the acidic
to neutral pH range of 1 6 (6). Thus, the kinetics of
formation of glycosylamines with respect to solution con
ditions such as pH, buffers and temperature, effect of
aldehydic content of sugar, and an understanding of the
mechanism of their formation is a key factor in predicting
reactivity and degradation kinetics of these sugar amine
reactions. A systematic study involving formation and
characterization of the glycosylamines and a detailed kinetic
and mechanistic understanding of their formation for weakly
basic aromatic amines and sugar carbonyls has not been
reported.
Daptomycin, a cyclic lipopeptide antibiotic, has been
reported to react with 5% dextrose and other reducing sugars
to form reversible products that decrease the potency of the
drug in acidic solutions (28). The peptide portion of
daptomycin consists of 13 amino acid residues connected at
the N terminal tryptophan to a decanoyl aliphatic group. It
has six ionizable groups: four carboxylic acid side chains and
two primary amines from the side chains of ornithine and a
weakly basic amine kynurenine. The pKa of kynurenine in
daptomycin was determined to be 0.8 using UV spectroscopy
(29). Reactions of daptomycin with glyceraldehyde resulting
in the formation of Schiffs base are also reported (9).
Reaction products formed with dextrose were proposed to
be the glycosylamines formed at the kynurenine aromatic
amine by reaction with the aldehydic glucose.
This paper represents the rst in a series of papers
describing the kinetics and mechanisms of the reactions of
reducing sugars and weakly basic amines in aqueous solution,
solid state, and pharmaceutical formulations. The main
objectives of this rst paper is to present (1) unreported ionic
equilibria of model reactant (kynurenine), (2) the analytical
methods used to characterize and measure reaction products,
(3) the kinetic scheme used to measure reaction rates, and (4)
relevant properties of various reducing sugars that impact the
reaction rate in solution.
Gokhale, Kearney, and Kirsch

Kynurenine is a weakly basic aromatic amino acid with
three ionizable groups viz. an aromatic amine, an alpha
carboxylic acid, and an alpha amine (Fig. 1) with a molecular
weight of 208.1. Glucose is one of the most common reducing
sugars (MW = 180) that exists in solution as cyclic six
membered rings as or glucose which equilibrates through
an acyclic aldehyde by mutarotation. The and glucose
differ in the position of the anomeric proton (proton at the C1
position in the ring) being in the equatorial or the axial
positions, respectively. The acyclic form of the sugar is the
reactive aldehyde in carbonyl amine reactions. All reducing
sugars can undergo reactions with amines owing to the
aldehydic form. Thus, isomers of glucose like galactose,
mannose, gulose, and allose with varying proportion of
the acyclic form can all undergo reactions with amines (30).
Although the six membered ring forms are the more stable
thermodynamically, glucose also exists as ve membered
ring forms in very low concentrations (30). The reaction
between kynurenine and glucose (Fig. 2) gives rise to
the imine that exists in equilibrium with the cyclic ,
glycosylamines.
Reactions were initiated with kynurenine and these
monosaccharides to isolate the reaction products and identify
them. A reaction scheme was proposed based on observed
kinetics for these sugar amine reactions. The aromatic amine
pKa for kynurenine was determined using UV spectroscopy;
the amine acid pKa was determined using potentiometric
titration and the carboxylic acid pKa was determined using
2D HMQC nuclear magnetic resonance (NMR) spectroscopy.
Reaction products were generated by heat stressing mixtures
of kynurenine in the presence of excess sugars under acidic
conditions at various pH values. Separation of the reaction
products was achieved by reversed phase high performance
liquid chromatography (RP HPLC) and product character
ization was done by liquid chromatography mass spectrome
try (LC MS). For reactions of kynurenine with glucose,
additional characterization studies were conducted by collect
ing fractions of products which were lyophilized and evaluat
ed by UV spectroscopy, 1 D 1H NMR, 2 D 1H 1H COSY, and
decoupling NMR experiments.
After product identication, a reaction scheme was
proposed and conrmed by determining the effect of varying
concentrations of glucose on the apparent pseudo rst order
loss of kynurenine. The effect of pH on the extent and half
life of the reaction was determined in acidic aqueous
solutions (pH values between 1 and 6). The effect of varying
acyclic content of various sugars on reactivity was also
studied.
NH3+
NH3+
C
CH2
C
H
Fig. 1. Structure of kynurenine
COOH
Glycosylation of Aromatic Amines I: Characterization of Reaction Products and Kinetic Scheme
319
H OH
OH
NH2
HO
H
NH3+
HO
OH
CH2
C
H
H
O
COO-
kynurenine
D-glucose
H OH
H OH
HO
HO
H
NH3+
OH
C
CH2
C
H
COO-
imine
H OH
H OH
H O
H O
HO
HO
H
H
H
NH
H2
C CH
NH3+
OH
H
COO-
-glycosylamine
HO
HO
H
H
H
OH
NH
C
H2
C CH
NH3+
COO-
-glycosylamine
Fig. 2. Reaction of kynurenine and glucose forming imine and and glycosylamines
METHODS AND MATERIALS

Kynurenine was obtained from ICN Biomedicals. Anhy
drous glucose was obtained from Fisher (Springeld, NJ,
USA) as was hydrochloric acid (1 N), sodium hydroxide
(0.1 N), and sodium chloride. D allose was obtained from
TCI America; D gulose was obtained from O Micron Bio
medicals, Inc.; and D Mannose and D galactose were
obtained from Sigma Aldrich. All chemicals were reagent
grade. Solvents used for chromatography were HPLC
grade.
pKa Determinations
The pKa value of the aromatic amine was determined
using a UV spectrophotometric method (Hewlett Packard HP
8453) at 401C (31). A 5.07 mM of L kynurenine stock
solution was used to prepare a series of 0.507 mM solutions of
kynurenine in the pH range of 0.46 3.58 with buffers
(hydrochloric acid, acetate buffer, and phosphate buffer)
and sodium chloride (to adjust to a constant ionic strength
of =0.5). The absorbance at 281 nm was the wavelength of
maximum difference between the ionized and unionized
species and was used to estimate the apparent pKa of the
aromatic amine by non linear regression (JMP) using an
equation relating the absorbance to the fraction and absorp

tivity of each species (31).
The pKa value for the alpha amine group on the amino
acid side chain was determined using potentiometric titration
(DL25, Mettler Inc.). All measurements were carried out at
401C for a 0.01 M L kynurenine solution. The pKa was
determined using the Gran plot method.
The alpha carboxylic acid pKa was determined using
NMR spectroscopy (heteronuclear multiple quantum corre
lation spectroscopy) by following changes in chemical shifts of
the alpha carbon proton and beta carbon protons on the side
chain of kynurenine as a function of pH. The HMQC
experiment monitors protons attached to only the carbon
skeleton of the molecule, specically by one bond in the
version used.
Kynurenine stock solution (0.020 M) was prepared in
deuterated water. The pH was measured at 40C. The pH
electrode was equilibrated and calibrated using pH 7 and
pH 4 standard buffer solutions also equilibrated at 401C.
Kynurenine solution was pipetted into a water jacketed
vessel, equilibrated to 401C and titrated using standardized
1.0 N HCl added in aliquots of 0.010 mL using an automatic
titrator (DL25, Mettler Inc.). One millimolar sodium 2, 2
dimethyl 2 silapentane 5 sulfonate (DSS) was also added to
the solution for referencing the NMR spectra. During the
titration, 0.65 mL of kynurenine solution was removed at the
320
desired pH and subjected to a 1D proton and a 2D HMQC

NMR experiment using a 500 MHz NMR, Varian Inc.,
spectrometer at 40 1C. The parameters for the 1 D
experiment were: averaged number of transients (nt)=8,
spectral width (sw)=6,000 Hz, number of points (np)=6,000
points processed with 0.2 Hz line broadening (lb) and zero
lled to 128 k points, and a recycle delay of 5 s (d1) in D2O.
For the 2 D HMQC experiment the parameters were as
follows: sw1=30,000, ni=128 s phase 1, 30,000/125.659=
12 ppm in proton. The spectra for the HMQC experiments
were directly referenced to DSS at high pH and indirectly
referenced at low pH.
After obtaining the NMR spectra, the sample was
reintroduced into the titration vessel and the titration was
continued until the next desired pH value was obtained. The
proton on the chiral carbon and the protons on the
carbon in the side chain were monitored for changes in
chemical shifts as a function of the deprotonation of the alpha
carboxylic acid group. Plots of chemical shifts versus pD for
each of the two groups of protons were sigmoidal in nature.
pKa was estimated using non linear regression (JMP V5.0.1,
SAS Institute) using the following equation: (32,33)

obs acid
10 pH
pH
10
10 pKa

base
10 pKa
pH
10
10 pKa

1
where obs is the observed chemical shift, acid is the chemical

shift of the most acidic species (pH* 1.02) and base is the
chemical shift of the most basic species (pH* 3.75) (pH*
denotes the uncorrected pH in deuterium oxide).
For separation and LC MS analysis of reaction products

of kynurenine with galactose, mannose, gulose, and allose the
column temperature and run time for the HPLC method was
further modied to range from 25 40C and 35 to 50 min,
respectively
Reaction Product Formation and Isolation
For reactions of kynurenine with glucose, a reaction
mixture of 1.2 mM kynurenine and 0.5 M glucose in hydro
chloric acid (pH 3.45) was prepared and stored at 401C in a
Teon coated rubber stoppered glass vial. Aliquots of reac
tion mixture were removed and diluted in acetate buffer
(0.5 M, pH 5.8) to quench the reaction. Reactions of 1.2 mM
kynurenine with 0.5 M glucose solution were conducted in
5.0010 4 N HCl at 40C. A semi preparative Shimadzu RP
HPLC system consisting of a SCL 10A system controller, two
LC 10ATVP pumps, and SPD 10A UV VIS detector was
used. Separation of reactants and products were obtained
using a Phenomenex Synergi semi prep C18 column, 250
21.2 mm using a 1% methanol water mobile phase, at a ow
rate of 7 mL/min, detection wavelength of 257 nm and a run
time of 80 min. Two products (I and II) were formed.
Products were collected over dry ice, transferred to a bench
lyophilizer (Virtis lyocentre) and lyophilized under vacuum
until the product appeared to be completely dry upon which
they were removed from the lyophilizer, sealed in plastic
tubes and stored at 20C until use. The integrity of the
lyophilized products was reconrmed by HPLC and the
products were used for identication by UV and NMR
spectroscopy.
Reaction Product Identification and Characterization
HPLC Analysis
HPLC analyses were performed using a Shimadzu RP
HPLC system consisting of an SCL 10AVP system controller,
LC 10ATVP pumps, SIL 10ADVP auto injector, SPD 10AVP
UV VIS detector, CTO 10ASVP column oven, and a FRC 10A
fraction collector. Chromatograms were integrated and data
stored using Class VP Chromatography Data System software
(Version 4.2). The column used was a Phenomenex HydroRP 18,
4.6250 mm, 4 column.
For reactions of kynurenine and glucose, an isocratic
method with a mobile phase composition of 1% methanol
and 99% water, a ow rate of 1 mL/min, injection volume of
10 lt, run time of 35 min, sample temperature of 4C, column
temperature of 25C, and detection wavelength of 257 nm
was used. The HPLC method was validated using standard
compendial methods. Calibration curves for kynurenine in
the range of 0.2 mM to 1 mM were used for calculating
concentrations. The precision of the analytical method was
determined by multiple analyses of 0.50 mM solutions of
kynurenine. The coefcient of variation was determined to be
3.6%. The accuracy of the method was determined by
estimating the percent recovery for 44 kynurenine test
samples of known concentrations in the range 0.2 to
1.0 mM. The mean, range, and CV for the test set were
98.9%, 91 110%, and 4.8%, respectively. This method was
used for identication of products for kynurenine glucose
mixtures using LC MS and all further kinetic analyses.
Reaction products were identied using UV, NMR, and

mass spectrometry. Reaction mixtures of 1.2 mM kynurenine
and 0.5 M monosaccharides (glucose, mannose, galactose,
gulose, and allose) were subjected to a quadrupole ion trap
type mass spectrometer (Thermo Finnigan, San Jose, CA,
USA), with a LCQ Deca Ion trap MS, a Surveyor LC, and
PDA detector. A positive electrospray type ESI ionization
method was used. The sheath gas and auxiliary gas ow rates
were 40 and 10, respectively. Nitrogen was used as the sheath
and the auxiliary gas. The capillary temperature was 350C
and the capillary voltage was 3 V.
The isolated, lyophilized products from reactions of
kynurenine with glucose were dissolved in water and scanned
by UV spectroscopy from 190 450 nm using a UV spectro
photometer (Hewlett Packard HP 8453 Diode Array) to
detect any observable spectral changes between the reactant
(kynurenine) and the products.
NMR spectroscopy was conducted on kynurenine and
the lyophilized products of kynurenine glucose reactions to
determine their structure. Kynurenine and each isolated
lyophilized product (I and II) were dissolved in 0.8 mL of
deuterated solvent dimethylsulfoxide (DMSO). A 0.65 mL
aliquot of this solution was introduced into the NMR tube
and subjected to 1 D 1H NMR experiments using a 500 MHz
NMR (Varian, Inc.) spectrometer. Kynurenine and product I
was also subjected to 2 D 1H 1H COSY NMR experiments,
but not product II due to its low concentration. Decoupling
321
Table I. Experimental Conditions to Determine the Scheme for the Reactions of 1.2 mM Kynurenine and 0.5 M Glucose at 40C and Ionic
Strength of 0.1 0.5 in Hydrochloric Acid or Phosphate Buffer
Buffer
Total concentration M103
NaCl (M)
Glucose (M)
Kynurenine (mM)
pH
HCl
HCl
HCl
NaH2PO4/Na2HPO4
30.0
10.0
0.100
0.121
0.0902
0.0891
0.100
0.032
0.501
0.505
0.514
0.508
1.20
1.25
1.25
1.19
1.66
2.11
4.39
5.94
experiments were also performed on products I and II setting

the decoupler frequency at the frequency of the aromatic
amine proton of kynurenine. This simplied the splitting
pattern of the protons on the anomeric carbon for products I
and II. The coupling constants for the simplied decoupled
doublets were calculated for the two products and compared
to literature values for the / form of glucose to conrm the
identity of the / isomers. The NMR parameters for the 1 D
1
H NMR experiments for determination of kynurenine
structure were as follows; np=60,000, number of transients=
4, spectral width=6,000, d1=5 s, temperature=25.4, and lb=
0.2. Similarly parameters for 2 D 1H 1H COSY experiment
for kynurenine were as follows: nt=4, sw=60,000, np=2,048,
lb=not used, fn=2,048, temperature=25.4, and d1=4. For the
2 D acquisition sw1=6,000 and ni=512 in seconds. Similarly,
parameters for 1 D 1H NMR determination for product I were,
np=60,000, nt=16, sw=6,000, d1=10 s, temperature=25.4, lb=
0.2, and fn=131,072. The parameters for 2 D 1H 1H COSY
experiment were nt=4, sw=6,000, np=4,096 (direct dimension)
lb=not used, fn=2,048, temperature=25.4, and d1=7. For the
2 D acquisition sw1=6,000 and ni=1,024 in seconds (indirect
dimension). For the decoupling experiment for product I the
peak at 9 ppm was decoupled and the following parameters
were used: nt=8, sw=6,000, lb=0.2, fn=131,072, temperature=
25C, and d1=10 s. The parameters for 1 D 1H spectral
determination for product II were nt=256, sw=6,000, d1=10 s,
temperature=25.4, lb=0.25, and fn=131,072. For the decoupling
experiment for product II the peak at 9.01 ppm was decoupled
and the following parameters were used: nt=256, sw=6,000,
lb=0.25, fn=131,072, temperature=25C, and d1=10 s.
reactions were carried out in the pH range of 1 6. The

solution conditions are given in Table I. Aliquots were
removed from reaction mixtures at appropriate times
quenched with acetic acid/sodium acetate quench buffer
(0.5 M at pH 5.8) and stored at 4C in auto injector and
analyzed by HPLC. Standard calibration curves were used for
calculating the concentration of kynurenine and the reaction
products in the reaction mixtures. The concentration time
proles for the loss of kynurenine and the appearance of
glycosylamines were constructed and reaction extent and
half lives were measured as function of pH.
Reactions were also conducted with 1 mM kynurenine
and various initial concentrations of glucose ranging from
0.19 0.70 M in sodium acetate/acetic acid buffer (total buffer
concentration 0.279 M) at pH 3.4 and a temperature of 40C.
The reaction conditions are described in Table II.
To study the kinetics with other glucose isomers, reaction
mixtures of 1.2 mM kynurenine and 0.5 M of monosaccharides
(galactose, mannose, allose, and gulose) in hydrochloric acid
(pH 2.7) at an ionic strength of 0.1 were stored at 401C in
Teon coated rubber stoppered glass vials. Aliquots of the
reaction mixture were removed, diluted with acetate quench
buffer, and HPLC analyses were performed. The extent of
reaction and reaction half life as a function of the acyclic
content were determined.
Kinetics of Kynurenine and Reducing Monosaccharides
The apparent Ka value of 0.01445 (pKa =1.84) at 40C for

the aromatic amine was calculated from the UV absorption
data at the analytical wavelength of 281 nm and the Ka value
of 8.99510 10 (pKa =9.05) at 40C of the alpha amino group
was determined using potentiometric titration.
The pK a of the alpha carboxylic acid group was
determined using 1 D 1H NMR spectroscopy. The proton
on the chiral carbon and the protons on the carbon in the
side chain were monitored in the 2 D HMQC experiment for
changes in chemical shifts as a function of the deprotonation
of the terminal carboxylic group.
The two monitored protons showed a characteristic pH
dependent change in chemical shift. The chemical shifts versus
pD (pH+0.4) for each of the two groups of protons resulted in a
sigmoidal curve. The proton on the chiral carbon gave a pKa
value of 2.57 (Fig. 3a), while the protons on the carbon in the
side chain gave a pKa value of 2.50 (Fig. 3b). The average
carboxylic acid pKa was estimated to be 2.53 in deuterium oxide
or 2.11 in water (as described in the DISCUSSION section).
Reaction mixtures of 1.2 mM kynurenine and 0.50 M

glucose in hydrochloric acid or phosphate buffer at the
desired pH and ionic strength of 0.10 or 0.50 were stored at
401C in Teon coated rubber stoppered glass vials. The
Table II. Experimental Conditions of Reactions of 1 mM Kynurenine

with Varying Concentration of Glucose in the Range of 0.19 0.70 M
in Acetic Acid/Sodium Acetate Buffer at pH 3.4 at 40C
No.
Glucose (M)
Kynurenine (mM)
pH
1
2
3
4
0.193
0.323
0.504
0.701
1.05
1.06
1.05
1.05
3.49
3.50
3.40
3.49
RESULTS
pKa Determinations for Kynurenine
322
4.65
4.05
4.6
chemical shift (ppm)
chemical shift (ppm)
4.55
4.5
4.45
4.4
3.95
3.9
4.35
3.85
4.3
4.25
1.5
2.5
3.5
4.5
pD
3.8
1.5
.
2.5
3..5
4.5
pD
Fig. 3. Chemical shift in the proton frequency for the a chiral carbon proton and b carbon protons as a function of pD in the pH* range of
1.02 3.76. The solid circles are the experimental data and the solid line was estimated based on non linear regression using JMP
Isolation and Characterization of Reaction Products

The HPLC chromatograms for reactions of kynurenine
and glucose showed the presence of three peaks; kynur
enine and two reaction products; product I and product II
in the order of their elution at retention times of 17.3 min,
23 min, and 27.4 min, respectively (Fig. 4). The two
products were formed reversibly and were isolated for
further characterization.
Kynurenine and reaction mixture samples were analyzed
using a LC mass spectrometer. The kynurenine sample showed
a peak with m/z ratio of 209.1, which corresponded to the
molecular ion of kynurenine and one proton (MH)+. The rst
peak in the reaction mixture chromatogram corresponded to
the kynurenine molecular ion with m/z of 209.1. The two
reaction product peaks that eluted from the reaction mixture of
kynurenine and glucose corresponded to m/z ratio of 371
(MH)+.
Fig. 4. Representative chromatogram of reaction mixture of kynur

enine and glucose peak at 17.3 min is kynurenine, at 23 min is product
I and at 27.4 min is product II
Reactions of kynurenine with galactose, mannose, allose,

and gulose also gave two major products in addition to the
kynurenine peak. In reactions of galactose, the two peaks
were not completely separated and co eluted as a single peak.
The LC MS analysis indicated that the products formed with
the other monosaccharides (galactose, mannose, gulose, and
allose) also corresponded to a m/z ratio of 371 (MH)+
indicating that the same products were formed for reactions
of weakly basic amine with all reducing monosaccharides.
The UV scans of the isolated products looked similar to
the UV scan of kynurenine exhibiting the same three
wavelength maxima of 228 230 nm, 257 260 nm, and 360
362 nm. These results are consistent with the formation of the
glycosylamines.
The 1 D proton spectrum of kynurenine, showed two
distinct regions of peaks: one at 6 8 ppm with ve peaks that
corresponded to the aromatic protons and the other region at
3 4 ppm with three peaks that corresponded to the side chain
protons. The assignments of the protons for kynurenine were
conrmed by performing a 2 D 1H 1H COSY experiment
(Table III). The COSY spectrum also showed two distinct
regions: the aromatic protons in the region of 6 8 ppm and
the side chain protons in the region of 3 4 ppm. The 1 D
spectra and the 2 D spectrum for product I both showed two
regions associated with the aromatic protons at 6 9 ppm and
the side chain protons at 3 4 ppm similar to the kynurenine
spectra. In addition to these protons, the 1 D spectra for
products I and II showed the sugar protons in the region of 3
4 ppm overlapping with the side chain protons of the amino
acid. Also, the 1 D spectra showed protons at 4.5 and 5.1 ppm
representing the anomeric protons for glucose and protons at
9 and 9.1 ppm representing the aromatic amine protons for
products I and II, respectively. The COSY spectrum for
product I showed a cross peak between the anomeric proton
(4.5 ppm) and the aromatic amine (9 ppm) thereby conrming
the identity of the reaction product to be a glycosylamine and
not an imine (Fig. 5). The identity of the and glycosylamines
was conrmed based on the coupling constants of 7.98 and 3.67
conditions by maintaining constant pH, temperature, and

ionic strength and glucose concentrations in high molar
excess. The chromatograms for the reaction mixtures showed
the same three peaks at all pH values; kynurenine and the
two products indicating that the reaction proceeded to the
and glycosylamines at all pH values. Typical calibration
curves of peak area versus concentration for kynurenine and
the glycosylamines (concentration range of 0.2 1 mM) were
linear (R2 >0.99). The concentration time proles (Fig. 6)
indicated that kynurenine reversibly formed the two
glycosylamines and that equilibrium between all three
species was reached over the duration of the reaction. The
and glycosylamines equilibrated rapidly with each other
and existed as a constant ratio of 6:1( glycosylamine
glycosylamine) at all pH values. The simplest scheme
consistent with the observed kinetics was:
Table III. Chemical Shift Assignments in ppm for Kynurenine

Protons for the 2 D 1H 1H COSY Spectrum in Deuterated DMSO
Region
Position
Chemical shift (ppm)
Aromatic region
H1
H2
H3
H4
NH
CH
CH2
6.75
7.25
6.5
7.7
7.2
3.6
3.5, 3.25
Side chain region
323
determined for the products I and II, respectively, by decoupling

experiments conducted for products I and II. The assignments
for the glycosylamines are listed in Table IV.
Determination of Reaction Scheme
kynurenine
kfobs
! glycosylamines
krobs
Reactions of kynurenine in the presence of excess

glucose were carried out in the pH range of 1 to 6.5. The
kinetic studies were performed under pseudo rst order
where kfobs is the rate constant for the forward reaction and
krobs is the rate constant for the backward reaction. The
sugar region
anomeric proton
H OH
H O
HO
HO
H
H
H
NH
OH
H
C
1
2
NH 3
H2
C CH
COO -
aromatic region
H3
aromatic amine proton
H1
H2
H4
Fig. 5. COSY spectrum of product I showing the region of 3 4 ppm corresponding to overlapped peaks and cross peaks of the side chain and
sugar protons and 6 9 ppm corresponding to peaks and cross peaks for the aromatic protons. (Bottom right) Expanded region showing the
region of 6 7 ppm corresponding to overlapped peaks and cross peaks of the aromatic protons. The peaks at 6.7, 6.9, 7.4, and 7.8 ppm
correspond to the aromatic ring protons H3, H1, H2, and H4, respectively. (Top left) Expanded region showing the region of 3 3.7 ppm. Two
distinct spin systems corresponding to the aromatic side chain and sugar protons
324
Region
Aromatic region
form
form
Side chain region

Sugar region
form
form
Position
NH
NH
H1
H2
H3
H4
CH
CH2
H1
H1
H2
H3
H4
H5
H6
H6
Chemical shift
9.1
9.0
6.9
7.4
6.7
7.8
3.6
3.5, 3.25
5.1
4.5
3.1
3.16
3.3
3.3
3.65
3.44
observed rate constant (kobs) for the overall loss of kynurenine

was determined based on the pseudo rst order reversible
kinetics of the reaction. Table V lists the reaction half lives and
equilibrium concentrations of kynurenine for the reactions of
kynurenine and glucose in the acidic pH range of 1 to 6.
Reactions were conducted with 1.2 mM kynurenine with
various initial concentrations of glucose from 0.19 0.70 M.
The data was t using reversible reaction kinetics and the
forward and reverse rate constants were determined using the
Eq. 2 (Table VI). Typical CV values for rate constant
estimates were 10%.
Reactions of kynurenine with other reducing sugars viz.
galactose, mannose, gulose, and allose also resulted in the
formation of the glycosylamines. The reaction kinetics for
these reducing monosaccharides was also described by the
reversible reaction scheme shown in Eq. 2. The total acyclic
content for the various isomers of glucose were obtained from
literature, based on measurements in aqueous solutions in
deuterium oxide at 30C obtained by 13C NMR for these
sugars (34). The acyclic content was reported to be lowest for
glucose and allose at 0.009% while gulose had the highest at
0.082%. Equilibrium concentrations of kynurenine for
reactions with allose, gulose, galactose, and mannose
indicated that the equilibrium concentration was constant
through all the sugars averaging at 64.3% (3.4%) of the
initial amine concentration. Reaction half lives indicated the
longest half life for glucose and allose at 0.6 h and 0.5 h,
respectively, and the shortest for gulose at 0.06 h.
estimated to be 1.84 at 40C using UV spectroscopy.

Kynurenine is a weakly basic aromatic amine similar in
basicity with primary aromatic amine containing drugs such
as sulfamethoxazole (pKa=1.69) (6), procainamide (pKa=
2.75) (26), benzocaine (pKa=2.49), or procaine (pKa=2.28)
(35) to name a few (31). The alpha amine pKa value was
determined to be 9.05 at 40C by potentiometric titration.
Protonation of the reactive amine dramatically reduces its
nucleophilicity thereby substantially minimizing its reactivity.
Hence, the alpha amine is largely unreactive in the acidic pH
range of 1 6 and the reactions with glucose were limited to
those involving the weakly basic aromatic amine. The alpha
carboxylic acid pKa for amino acids typically lies in the range
of 2 3. This is very close to the aromatic amine pKa and
hence to avoid interference by the aromatic amine pKa, an
HMQC type NMR experiment was used to determine the
pKa for the carboxylic acid. In the HMQC experiment,
protons attached to the carbon skeleton of the molecule were
monitored. Since the carboxylic acid proton is exchangeable,
the protons on the carbon atoms ( and carbons of the side
chain) neighboring to the carboxylic acid group were
monitored as a function of pH. The protonation or deproto
nation of the carboxylic acid changed the shielding of the
neighboring protons and caused a change in chemical shift of
those protons as a function of pH. This chemical shift was
plotted as a function of pD to yield a sigmoidal curve that was
used to estimate the pKa value. The standard correction for
estimation in deuterium was applied to the observed meter
reading (36,37). pKa value of 2.57 for the carbon proton
and 2.50 for the carbon protons gave an average pKa value
of 2.53 for the carboxylic acid pKa at 40C in D2O. However,
the values for ionization constants of carboxylic acids in
deuterium oxide and water differ by a factor of 0.5 0.6 due
primary isotope effects (38,39). The difference in the ioniza
tion constant (pKa) of carboxylic acids of amino acids in
1.5
Conc (mM)
Table IV. Chemical Shift Assignments for the Glycosylamines Based

on 1 D 1H NMR, 2 D NMR in Deuterated DMSO
0.5
DISCUSSION
10
time (hrs)
Determination of pKa Values

Kynurenine is an aromatic amino acid with three
ionizable groups; the aromatic amine, the alpha amine, and
the alpha carboxylic acid. The aromatic amine pKa was
Fig. 6. Typical concentration area time proles of reaction of 1.2 mM

kynurenine (solid circle) and 0.5 M glucose in HCl/NaCl at 0.1 ionic
strength and 40C. Appearance of glycosylamine (solid square),
glycosylamine (solid triangle), and mass balance (multiplication
symbol) are depicted
325
Table V. Overall Observed Rate Constant (kobs), Half Lives (h), and Percent of Initial Concentration of Kynurenine at Equilibrium for the
Reactions of 1.2 mM Kynurenine with 0.5 M Glucose in Hydrochloric Acid and Phosphate Buffer at an Ionic Strength of 0.1 0.5 and 40C in
the pH Range of 1 to 6.5
Buffer
pH
kobs (h 1)
Half life (h)
Percent of initial concentration at equilibrium
HCl
HCl
HCl
NaH2PO4/Na2HPO4
1.66
2.11
4.39
5.94
3.87
2.43
0.118
0.026
0.179
0.285
5.87
26.86
78.58
72.20
65.24
63.27
The estimated rate constant values for the forward and reverse reactions are contained in paper Glycosylation of aromatic amines II: Kinetics
and Mechanisms of the Hydrolytic Reaction between Kynurenine and Glucose (in press)s
water and deuterium oxide can be described by a linear

relation of the form (y=mx+h) as:
pKD
pKH $pK a b pKH
where a=0.332 and b=0.040 (38). Substituting the pKa value

in D2O (pKD) of 2.53 for kynurenine in the Eq. 3, the nal
pKa value for kynurenine in water (pKH) was 2.11. This is
consistent with the typical pKa values in the range of 2 3
reported for the alpha carboxylic acids of amino acids (31).
Identification of Reaction Products
In the pH range of 1 6, the aromatic amine of kynurenine
is largely unprotonated and therefore present in a reactive
form, whereas the aliphatic alpha amine is largely protonated
and unreactive. The reaction of kynurenine and glucose gave
rise to two products I and II. LC MS, UV spectroscopy, and
NMR were used to determine the structure of the two
products. Both the products, I and II, showed a m/z of 371 or
identical molecular mass of 370 indicating that the products
were either the glycosylamines or the imine. The m/z of 393 in
the mass spectra for the two products corresponds to molecular
ion plus sodium. Though the expected products were the
glycosylamines, as has been reported for reactions of amines
with sugars (6,8,26) this was further conrmed using UV and
NMR spectroscopy. The two products showed similar UV
spectral properties to that of kynurenine. The presence of an
additional double bond in an imine would result in a shift of the
UV wavelength maxima to a longer wavelength. Since no such
shifts were observed, the two products were most likely the two
glycosylamines. This supposition was further conrmed using
NMR spectroscopy. 1 D 1H NMR, 2 D 1 H 1 H COSY
experiments and decoupling experiments were conducted to
conrm the identity of the products.
Table VI. Estimated Values for the Forward (kfobs) and Reverse
Rate (krobs) Constants for Reactions of 1 mM Kynurenine with 0.19
0.70 M Glucose in Acetic Acid/Sodium Acetate Buffer at pH 3.4 and
40C
Concentration of glucose (M)
kfobs (h 1)
krobs (h 1)
0.193
0.323
0.504
0.701
0.25
0.50
0.75
1.19
1.84
2.56
2.29
2.77
The chemical shift assignments for the protons on the

amino acid kynurenine are listed in Table III. The proton at
3.6 ppm was typical of the chiral carbon proton and showed
a doublet of doublets due to the two nonequivalent protons
on the side chain that appeared at a chemical shift of 3.5 and
3.2 ppm. The proton at 7.2 ppm showed a typical broad
singlet characteristic of an amine proton. A triplet at 6.5 ppm,
a doublet at 6.75 ppm, another triplet at 7.3 ppm, and a
doublet at 7.7 ppm each corresponded to one proton on
the aromatic ring of kynurenine. In the COSY spectrum, the
peaks across the diagonal represent the 1 D NMR peaks. The
cross peaks perpendicular to the peaks along the diagonal
represented connectivity between corresponding peaks along
the diagonal. Protons up to two to three bond distances away
typically show cross peaks. Occasionally longer range cou
plings also show in a COSY. In the aromatic region of 6
8 ppm, cross peaks were observed for four peaks along the
diagonal except one peak which was the broad amine peak.
The assignments of the aromatic region protons in kynur
enine were based on the splitting pattern of peaks in the 1 D
spectrum for kynurenine and product I, and the cross peaks in
the COSY spectrum. Substituent group rules and resonance
effects were also used to verify the assignments in the
aromatic region.
The identity of the glycosylamine was determined using
the following:
1. The presence of the anomeric proton and aromatic
amine proton indicates the presence of an intact sugar
ring.
2. A cross peak in the COSY spectrum for product I
showed that the anomeric proton and the amine
proton are connected in the structure through two to
three bonds.
3. Decoupling experiments identied the and forms
of the glycosylamines.
4. Chemical shifts were assigned to complete the structural
identication.
1 D NMR spectrum of product I and II in deuterated
DMSO showed a doublet at 9.0 and 9.1 ppm that were
assigned to the aromatic amine protons for products I and II,
respectively. The triplet at 4.5 ppm for product I and at
5.1 ppm for product II was typical of the anomeric proton on
the glucose and was assigned to be the C1 proton (anomeric
proton) on the glucose ring. A 2 D 1H 1H COSY experiment
was conducted for product I (Fig. 5). It showed a distinct
cross peak between the proton at 9.0 ppm and the proton at
326

0.7
4.5 ppm, which indicated that they were two to three bonds
away, and hence the two peaks were assigned to be the amine
and the anomeric proton, respectively. In order to further
conrm these assignments and identify the and forms, the
decoupling experiments were used that provided two pieces
of information:
The decoupling experiment was performed by setting the

decoupler frequency at the frequency of the aromatic amine
peak for each of the products I and II. Thus, the effect of the
aromatic amine proton on the anomeric proton of glucose
was removed and simplifying the splitting pattern. The
anomeric proton splitting pattern changed from a triplet to a
doublet. This conrmed that the protons at 9.0 and 9.1 ppm
and the anomeric protons at 4.5 and 5.1 ppm were connected
in the structure for products I and II, respectively. The
coupling constant (J) for the decoupled doublets was
determined to be 7.98 for product I and 3.67 for product II.
These coupling constants were typical to that of the glucose
(J~8) and the glucose (J~4), respectively (6,30,40). The
result supported the assumption that product I was the
glycosylamine and the product II was the glycosylamine.
The more stable form, the glycosylamine, appeared in a
higher ratio in the reaction mixture, in agreement with the
observations of Sunderland and coworkers for the reaction of
glucose and sulfamethoxazole (6).
Figure 5 shows the expanded COSY spectrum in the
region of the aromatic protons for product I. The chemical
shift assignments were done using the kynurenine assign
ments as a reference and also considering the splitting
patterns of the peaks in the 1 D NMR spectrum. The region
of 3 3.7 ppm in the COSY spectrum showed two spin systems
k (h-1)
allose
0.5
half life (hr)
1. Decoupling of either of those protons affected the

splitting pattern of the other proton, indicating the
two protons were connected.
2. It could be used to distinguish between the isomeric
forms i.e. / forms.
glucose
0.6
0.4
0.3
mannose
0.2
galactose
gulose
0.1
0
0
0.02
0.04
0.06
0.08
0.1
total acyclic content (%)

Fig. 8. Half life (h) versus total acyclic content (%) for reactions of
1.2 mM kynurenine with 0.5 M monosaccharides for reactions at
pH 2.7, 0.1 ionic strength, and 40C
overlapping each other, one from the protons in the side

chain of the amino acid and the other from the sugar protons
(Fig. 5). One of the spin system overlayed with the
kynurenine side chain region and was assigned accordingly.
The other spin system was due to the sugar protons. Specic
proton assignments for the sugar part of the glycosylamine
(Product I) was done by initially assigning the anomeric
proton (C1) at 4.5 ppm followed by assigning the proton at
3.1 ppm as C2 based on the cross peak observed in the COSY
spectrum. Similarly the subsequent protons were assigned
using cross peaks, peak integrations, and comparing typical
assignments reported in literature for a D glucose (30).
Proton signals for the amino acid side chain and sugar ring
which lie around 3.2 3.35 ppm are also overlapped by the
water peak due to interference of water signals in this region
of the spectrum. The chemical shift assignments for the
glycosylamines are reported in Table IV. The chemical shifts
for the glycosylamines for the anomeric protons correlated
well with those determined by Lucida and co workers (6).
The product II was identied based on the mass, the coupling
constant (J), and similar chemical linkages identied for the
1 D spectra of products I and II. However, individual
chemical shift assignments for product II were not done
because the low concentrations and instability of the sample
resulted in a noisy spectrum. The characterization studies
demonstrated that for the reactions of kynurenine and
glucose, product I formed was the glycosylamine and the
product II was the glycosylamine.
Reaction Scheme and Kinetics
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
sugar Conc (M)

Fig. 7. The forward (solid circle) and reverse (solid square) rate
constants as a function of initial glucose concentration in sodium
acetate/acetic acid buffer at pH 3.49 and 40C
The reaction of kynurenine and glucose reversibly

formed glycosylamines in the pH range of 1 to 6. Based on
the concentration time proles, only three species existed in
solution in detectable quantities, kynurenine and the two
glycosylamines. The glycosylamines equilibrated in solution
with each other through the acyclic imine. This is analogous
to the , glucose that exists in equilibrium in solution with

the aldehydic form (23). However, no detectable concen
trations of the imine were obtained under the experimental
conditions. No loss in mass balance was observed. A simple
reversible scheme (Eq. 2) was used to describe the data. The
pseudo rst order rate constant (kobs) for the loss of
kynurenine was determined based on reversible kinetics.
The rate of formation of glycosylamines was pH dependent.
For example, the reaction half life at pH 1.66 was 0.179 h
while the half life at pH 5.94 was 26.86 h (Table V). pH
dependent reaction rates for the reactions of daptomycin with
some monosaccharides were also reported earlier by Inman
and Kirsch (28). The reaction extent was also a function of
pH as shown in Table V.
Reaction scheme conrmation was obtained by conduct
ing a series of reactions with various concentrations of excess
glucose. The forward rate constant (kfobs) was determined to
be a function of sugar concentration; however, the reverse
reaction rate constant (kfobs) was independent of the glucose
concentration (Table VI, Fig. 7).
Reactions of kynurenine with the other monosacchar
ides indicated that the extent of reaction was not a function
of acyclic content; however, the reaction half life was a
function of the total acyclic content (Fig. 8). This indicated
that the reaction kinetics for these sugar weakly basic amine
reactions were a function of the inherent sugar properties
and a higher equilibrium acyclic concentration resulted in
increased reactivity.
CONCLUSIONS
Kynurenine contains a weakly basic aromatic amine that
reacts with the aldehydic moiety of reducing sugars under
mildly acidic conditions to form glycosylamines. The reaction
occurs reversibly via a steady state imine intermediate, and
the reaction rate is dependent on the concentration of the
acyclic form of the sugar. The pH dependent mechanisms of
this reaction are the subject of the second paper in this series.
REFERENCES
1. G. P. Ellis, and J. Honeyman. Advances in carbohydrate
chemistry. Glycosylamines. 10:95 125 (1955).
2. W. Pigman and D. Horton. The carbohydrates: Chemistry and
Biochemistry 2nd edition, IB ed., Academic Press, New York,
1980, pp. 881 921.
3. V. Kumar, and G. S. Banker. Maillard reactions in chemistry, food
and health; Maillard reaction and drug stability, The Royal
Society of Chemistry, Cambridge, 1994, pp. 20 27.
4. R. N. Duvall, K. T. Koshy, and J. W. Pyles. Comparison of
reactivity of amphetamine, metamphetamine, and dimethylam
phetamine with lactose and related compounds. J. Pharm. Sci.
54:607 611 (1965b).
5. D. D. Wirth, S. W. Baertschi, R. A. Johnson, S. R. Maple, M. S.
Miller, D. K. Hallenbeck, and S. M. Greg. Maillard reaction of
lactose and uoxetine hydrochloride, a secondary amine. J.
Pharm. Sci. 871:31 39 (1998).
6. H. Lucida, J. E. Parkin, and V. B. Sunderland. Kinetic study of
the reaction of sulfamethoxazole and glucose under acidic
conditions I. Effect of pH and temperature. Int. J. Pharm.
202:47 61 (2000).
7. J. E. Parkin, K. F. Ilett, and D. A. Joyce. Reaction of
sulphamethoxazole in Bactrim DS with carbohydrates to form
glucosylamines in vitro and in vivo. Eur. J. Pharm. Sci. 5:347 353
(1997).
327
8. A. Sianipar, J. E. Parkin, and V. B. Sunderland. The reaction of

procainamide with glucose following admixture to glucose
infusion. Int. J. Pharm. 176:55 61 (1998).
9. W. Muangsiri, W. R. Kearney, L. M. Teesch, and L. E. Kirsch.
Studies on the reactions between daptomycin and glyceralde
hyde. Int. J. Pharm. 289:133 150 (2005).
10. S. M. Blaug, and W. T. Huang. Interaction of dextroamphet
amine sulfate with dextrates in solution. J. Pharm. Sci. 62:653
655 (1973).
11. S. M. Blaug, and W. T. Huang. Browning of dextrates in solid
solid mixtures containing dextroamphetamine sulfate. J. Pharm.
Sci. 63:1415 1418 (1974).
12. C. Petriella, J. Chirife, S. L. Resnik, and R. D. Lozano.
Correlation between induction time and rate of browning in
heated model solutions of glucose and lysine. Int. J. Food Sci.
Tech. 23:415 418 (1988).
13. H. C. Warmbier, R. A. Schnickels, and T. P. Lazuba. Nonenzymic
browning kinetics in an intermediate moisture model system:
effect of glucose to lysine ratio. J. Food Sci. 48:981 983 (1976).
14. C. M. Lee, B. Sherr, and Y. N. Koh. Evaluation of kinetic
parameters for a glucose lysine Maillard reaction. J. Agric. Food
Chem. 32:379 382 (1984).
15. T. P. Lazuba, and S. A. Massaro. Browning and amino acid loss
in model total parental nutrition solutions. J. Food Sci. 55:821
826 (1990).
16. L. N. Bell, K. L. White, and Y. H. Chem. Glycine loss and
Maillard browning as related to the glass transition in a model
food system. J. Food Sci. 63:625 628 (1998a).
17. T. P. Lazuba. Interpreting the complexity of the kinetics of the
Maillard reaction. In T. P. Lazuba, G. A. Reineccious, V. M.
Monnier, J. OBrien, and J. Baynes (eds.), Maillard reactions in
chemistry food and health, 5th ed., The Royal Society of
Chemistry, Cambridge, 1994, pp. 176 181.
18. C. G. A. Davies, B. L. Wedzich, and C. Gillard. Kinetics model
of the glucose glycine reaction. Food Chem. 60:323 329 (1997).
19. S. E. Fayle, and J. A. Gerrard. The Maillard reaction, The Royal
Society of Chemistry, Cambridge, UK, 2002.
20. W. M. Baisier, and T. P. Lazuba. Maillard browning kinetics in
a liquid model system. J. Agric. Food Chem. 40:703 713
(1992).
21. J. A. Stamp, and T. P. Lazuba. Kinetics of the Maillard reaction
between aspartame and glucose in solution at high temperature.
J. Food Sci. 48:543 545 (1983).
22. H. S. Isbell, and H. L. Frush. Mutarotation, hydrolysis and
rearrangement reactions of glycosylamines. J. Org. Chem.
23:1309 1319 (1958).
23. B. Capon and B. E. Connett. The mechanism of hydrolysis of N
aryl D glucosylamines. J. Chem. Soc. 4497 4502 (1965).
24. V. Das Gupta. Complexation of procainamide with dextrose. J.
Pharm. Sci. 719:994 996 (1982).
25. V. Das Gupta. Complexation of procainamide with hydroxide
containing compounds. J. Pharm. Sci. 722:205 207 (1983).
26. A. Sianipar, J. E. Parkin, and V. B. Sunderland. Chemical
incompatibility between procainamide hydrochloride and glu
cose following intravenous admixture. J. Pharm. Pharmacol.
46:951 955 (1994).
27. The United States Pharmacopeia. 23:1463 (1995).
28. E. L. Inman and L. E. Kirsch. An improved diluent formulation
for daptomycin., in PCT Int. 1990: European.
29. W. Muangsiri, and L. E. Kirsch. The kinetics of the alkaline
degradation of daptomycin. J. Pharm. Sci. 908:1066 1075 (2001).
30. H. S. Khadem. Carbohydrate chemistry, monosaccharides and
their oligomers, Academic Press Inc, San Diego, California, 1988,
pp. 11 56.
31. A. Albert, and E. P. Searjeant. The determination of ionization
constants. A laboratory manual, 3rd ed., Chapman and Hall, New
York, 1984, pp. 70 73.
32. W. Schaller, and A. D. Robertson. pH, ionic strength, and
temperature dependences of ionization equilibria for the carbox
yl groups in Turkey ovomucoid third domain. Biochemistry.
34:4714 4723 (1995).
33. W. R. Forsyth, and A. D. Robertson. Insensitivity of perturbed
carboxylic pKa values in the ovomucoid third domain charge
replacement at a neighboring residue. Biochemistry. 39:8067
8072 (2000).
328
34. Y. Zhu, J. Zajicek, and A. S. Serianni. Acyclic forms of [1 13C]
aldohexoses in aqueous solution: Quantication 13C NMR and
deuterium isotope effects on tautomeric equilibria. J. Org. Chem.
66:6244 6251 (2001).
35. G. Volgyi, R. Ruiz, K. Box, J. Comer, E. Box, and K. Takacs
Novak. Potentiometric and spectrophotometric pKa determination
of water insoluble compounds: Validation study in a new cosolvent
system. Anal. Chim. Acta. 583:418 428 (2007).
36. P. K. Glasoe, and F. A. Long. Use of glass electrodes to
measure acidities in deuterium oxide. J. Phys. Chem. 641:188
190 (1960).
37. D. D. Perrin, and B. Demsey. Buffers for pH and metal ion
control, Chapman and Hall, London, 1974, pp. 81 82.

38. I. N. Gordon and B. M. Lowe. The dissociation constants of
amino acids in deuterium oxide. J. Chem. Soc. Chem. Commun.
13(803 804) (1970).
39. G. C. K. Roberts. NMR of macromolecules. A practical
approach. In D. Rickwood, and B. D. Hames (eds.), The
Practical Approach Series, Oxford University Press, New York,
1993, pp. 22 23.
40. P. A. Harmon, W. Yin, W. E. Bowen, R. J. Tyrell, and R. A.
Reed. Liquid Chromatography mass spectrometry and proton
nuclear magnetic resonance characterization of trace level
condensation products formed between lactose and the amine
containing diuretic hydrochlorothiazide. J. Pharm. Sci. 897:920
929 (2000).
AAPS PharmSciTech, Vol. 10, No. 2, June 2009 ( # 2009)

DOI: 10.1208/s12249 009 9212 7
Brief/Technical Note
Physicochemical Properties of Solid Dispersions of Gliclazide
in Polyvinylpyrrolidone K90
S. Biswal,1,2 J. Sahoo,1 and P. N. Murthy1
Received 27 June 2008; accepted 13 February 2009; published online 25 March 2009
KEY WORDS: dissolution; gliclazide; polyvinylpyrrolidone K90; solid dispersion.
INTRODUCTION
Gliclazide is a second generation hypoglycemic sulfonyl
urea which is useful in the treatment of type 2 diabetes
mellitus (1). Following oral administration, however, glicla
zide exhibits slow rate of absorption and interindividual
variation in bioavailability. Stated problems of gliclazide might
be due to its poor water solubility and slow dissolution rate (2
4). But gliclazide exhibits good tolerability, low incidence of
hypoglycemic effect, low rate of secondary failure, and low rate
of progression of diabetic retinopathy (2,5). Hence, gliclazide
appears to be a drug of choice in long term sulfonylurea therapy
for treatment of type 2 diabetes mellitus. Few attempts have
been made for improvement of solubility and bioavailability of
gliclazide including complexation with cyclodextrin (6,7) or
cyclodextrin hydroxypropylmethylcellulose (8) and using PEG
400 (9) as per present literature. The authors investigated the
physicochemical characteristics and dissolution behaviors of
gliclazide in physical mixtures as well as solid dispersions with
polyethylene glycol 6000 in a previous study (10).
The main objective of this work was to investigate the
physicochemical characteristics of gliclazide in physical mix
tures (PMs) and solid dispersions (SDs) prepared by using
polyvinylpyrrolidone K90 (PVP K90). The possible interac
tions between gliclazide and PVP K90 in both solid and liquid
states were investigated. Interaction in solid state was investi
gated by Fourier transform infrared (FT IR) spectroscopy, X
ray diffraction (XRD) analysis, and differential scanning
calorimetry (DSC). Interaction in solution was studied by
phase solubility analysis and dissolution experiments.
MATERIALS AND METHODS
Materials
A gift sample of gliclazide was received from Aristo
Pharmaceuticals Ltd. (Mumbai, India). PVP K90 was re
1
Royal College of Pharmacy and Health Sciences, Andhapasara

Road, Brahmapur, 760002, Orissa, India.
2
To whom correspondence should be addressed. (e mail: sudarsan
mpharm@yahoo.co.in)
ceived from BASF (Germany). Double distilled water was

used throughout the study and all the other chemicals used
were of analytical grade.
Preparation of SDs
The SDs of gliclazide with PVP K90 containing three
different weight ratios (1:1, 1:2, 1:5; gliclazide/PVP K90) and
denoted as SD 1/1, SD 1/2, and SD 1/5, respectively, were
prepared by solvent evaporation method. In the solvent
evaporation method, to a solution of gliclazide in chloroform,
an appropriate amount of PVP K90 in solution was added.
The solvent was evaporated under reduced pressure at 40C
by using a rotary evaporator and the resulting residue was
dried under vacuum for 3 h. The mixture was stored
overnight in a desiccator. The hardened mixture was pow
dered in a mortar, sieved through a 100 mesh screen, and
stored in a screw cap vial at room temperature until further
use (11,12).
The PMs having the same weight ratio as SDs were
prepared by thoroughly mixing the required amount of
gliclazide and PVP K90 for 10 min in a mortar. The resulting
mixtures were sieved through a 100 mesh sieve and denoted
as PM 1/1, PM 1/2, and PM 1/5, respectively. The mixtures
were stored in a screw cap vial at room temperature until
use.
Phase Solubility of Gliclazide

Solubility determinations were performed in triplicate
according to the method of Higuchi and Connors (13). In
brief, an excess amount of gliclazide was taken into a screw
capped glass vial to which 20 ml of aqueous solution
containing various concentrations of PVP K90 was added.
Then, the samples were shaken at 370.5C for 72 h in a
water bath (Remi Pvt Ltd, Mumbai). After 72 h, samples
were ltered through a 0.45 m membrane lter. The ltrate
was suitably diluted and analyzed spectrophotometrically at the
wavelength of 227 nm using a UV VIS spectrophotometer
(Shimadzu 1700, Pharm Spec, Japan).
329
330
Biswal, Sahoo and Murthy
Dissolution Studies
Dissolution studies of gliclazide in powder form, SDs,
and PMs in triplicate were performed by using the US
Pharmacopoeia (USP) model digital tablet dissolution test
apparatus 2 (Lab India Ltd, Mumbai) at the paddle rotation
speed of 50 rpm in 900 ml 0.1 N HCl containing 0.25% (w/v)
of sodium lauryl sulfate as a dissolution medium at 370.5C
(14). The SDs or PMs equivalent to 30 mg of gliclazide were
weighed using a digital balance and added into the dissolution
medium. At the specied times (every 10 min for 1 h), 10 ml
samples were withdrawn by using syringe lter (0.45 m;
Sepyrane, Mumbai) and then assayed for gliclazide content
by measuring the absorbance at 227 nm using the UV Visible
spectrophotometer (Shimadzu UV 1700, Pharm Spec). Fresh
medium (10 ml), which was prewarmed at 37C, was added to
the dissolution medium after each sampling to maintain a
constant volume throughout the test.
Differential Scanning Calorimetry
The DSC measurements were performed on a DSC 6100
(Seiko Instruments, Japan) differential scanning calorimeter
with a thermal analyzer. All accurately weighed samples
(about 1.541 mg of gliclazide or its equivalent) were placed in
sealed aluminum pans, before heating under nitrogen ow
(20 ml/min) at a scanning rate of 10C min 1 from 25C to
250C. An empty aluminum pan was used as a reference.
The Gibbs free energy of transfer (Gtr) of gliclazide

from pure water to the aqueous solutions of carrier was
calculated as
$Gtr
2:303 RT log
In Vitro Dissolution Data

In the present investigation, model independent and
model dependent approaches are used for comparison of
dissolution proles. Model independent approaches are
based on the ratio of area under the dissolution curve
(dissolution efciency) or on mean dissolution time (15,16).
Percent dissolution efciency (%DE) and mean dissolution
time were also computed to compare the relative perfor
mance of various concentrations of carrier in SDs and PMs.
The magnitude of %DE at 10 min (%DE10 min) and 30 min
(%DE30 min) for each formulation was computed as the
percent ratio of area under the dissolution curve up to time t
(10 and 30 min), to that of the area of the rectangle described
by 100% dissolution at the same time. The magnitude of
mean dissolution time for each formulation was calculated
using PCP Disso v3 software (Pune, India).
Rt
Fourier-Transform Infrared Spectroscopy
%DE
Ydt
Y100 t
In the model dependent approaches, release data

were tted to ve kinetic models including the zero order
(Eq. 4), rst order (Eq. 5), Higuchi matrix (Eq. 6), Peppas
Korsmeyer (Eq. 7), and Hixson Crowell (Eq. 8) release
equations. PCP Disso v3 software (Pune, India) was used to
nd best t model.
R k1 t
FT IR spectra were obtained by using an FT IR spec

trometer 430 (Jasco, Japan). The samples (gliclazide or SDs
or PMs) were previously ground and mixed thoroughly with
potassium bromide, an infrared transparent matrix, at 1:5
(sample/KBr) ratio, respectively. The KBr disks were pre
pared by compressing the powders at a pressure of 5 tons for
5 min in a hydraulic press. Forty scans were obtained at a
resolution of 4 cm 1, from 4,600 to 300 cm 1.
where SSos is the ratio of molar solubility of gliclazide in

aqueous solution of PVP K90 to that of the same medium
without PVP K90.
X-ray Diffraction
The X ray powder diffraction patterns were obtained at
room temperature using a PW1710 X ray diffractometer
(Philips, Holland) with Cu as anode material and graphite
monochromator, operated at a voltage of 35 kV, current
20 mA. The samples were analyzed in the 2 angle range of
5 70, and the process parameters were set as: scan step size
of 0.02 (2) and scan step time of 0.5 s.
So
Ss
log UR
k2 t
2:303
p
R k3 t
log R log k4 n log t
UR1=3 k5 t
Dissolution Data Analysis

Phase Solubility Studies
The value of apparent stability constant, Ks, between
drug carrier combinations was computed from the phase
solubility proles, as described below
Ks
Slope
:
Intercept1 Slope
where R and UR are the released and unreleased percen

tages, respectively, at time t; k1, k2, k3, k4, and k5 are the rate
constants of zero order, rst order, Higuchi matrix, Peppas
Korsmeyer, and Hixon Crowell models, respectively.
Properties of Gliclazide in Polyvinylpyrrolidone K90
331
The phase solubility diagram investigated in 0.1 N HCl

(pH 1.2) was linear in a wide range of PVP K90 concen
trations and correspond to AL type proles (13). The stability
constant was found to be 1.20 ml 1 mg. At 18% (w/v)
concentration of PVP K90, the solubility of gliclazide
increased by 4.6 fold (Table I). Table I presents the values
of Gibbs free energy associated with the aqueous solubility of
gliclazide in the presence of PVP K90.
ratio, respectively. From Table II, Q20 as well as Q30 values

indicate, in general, PVP K90 based formulations (SDs and
PMs) at high carrier levels exhibited higher dissolution rates
than those at low polymer levels (PM 1:1 and SD 1:1 with PM
1:5 and SD 1:5, respectively).
The %DE values were computed at two times, showing
the early and late phase of dissolution study for comparative
analysis of all the formulations. The %DE values in the initial
time period of dissolution study, i.e., %DE10 min, provide
comparative information for very fast releasing formulations,
whereas %DE30 min provides relative information about both
fast and slow releasing formulations. The value of %DE10 min
for pure gliclazide (9.16%) was enhanced in PMs (21.94%) as
well as in SDs (49.87%). The value of %DE30 min for the pure
drug increased to 45.11% in PMs and up to 73.20% in SDs
(Table II).
The obtained values of mean dissolution time (MDT) for
pure gliclazide, PMs, and SDs are presented in Table II. The
MDT of gliclazide was 12.5 min and it decreased to 7.09 min
in SD with PVP K90 at 1:5 (gliclazide/PVP K90) ratio,
whereas in case of PMs, MDT decreased to 9.43 min.
Table III shows the regression parameters obtained after
tting various release kinetic models to the in vitro dissolution
data. In vitro release data of the drug are best tted to
Korsemeyer Peppas model with n value of 0.7249, hence
exhibits non Fickian diffusion. For all PMs (PM 1:2 and PM
1:1), the Higuchi matrix is best tted except PM 1:5 which
exhibits Korsemeyer Peppas model. All the SDs best tted to
Korsemeyer Peppas model with n values less than 0.4500
tended to exhibit Fickian diffusion characteristics except SD
1:5 which exhibits rst order release kinetics (17,18).
Dissolution Studies
Q10, Q20, and Q30 values (percent drug dissolved within

30 min) are reported in Table II (coefcient of variation is
less than 10% in each case). The onset of dissolution of pure
gliclazide is very low, about 18.46% of drug being dissolved
within 10 min (Q10). SDs of gliclazide with PVP K90
considerably enhanced dissolution rates within 10 min com
pared to pure gliclazide such as SD at 1:1 (gliclazide/PVP
K90) ratio up to 76.29%, SD at 1:2 ratio up to 77.12%, and
SD at 1:5 ratio up to 77.73%, respectively. In case of PMs of
gliclazide, Q10 values increased up to 25.36% at 1:1, up to
33.1% at 1:2, and up to 43.88% at 1:5 (gliclazide/PVP K90)
The DSC curve of pure gliclazide exhibited a single

endothermic response corresponding to the melting of the
drug. Onset of melting was observed at 170.8C, the
corresponding heat of fusion (HF) was 171.2 J/g (Fig. 1 A)
(10). During scanning of PVP K90, a broad endotherm
ranging from 80C to 120C was observed, due to the
presence of water. DSC thermograms of PMs and SDs of
gliclazide and PVP K90 always exhibited complete absence of
melting peak of the drug at 170.8C and the broad endotherm
due to the presence of water ranging from 60C to 120C
(Fig. 1 D). Repeated scanning of PVP K90 based formula
Table I. Effect of PVP K90 Concentration and Gibbs Free Energy on

Solubility of Gliclazide in 0.1 N HCl
Concentration
of PVP K90
(% w/v)
Concentration
of gliclazide
(mg ml 1)
at 37C
Gtr (J/mol)
0
2
4
6
8
10
12
14
16
18
0.81
0.81
1.21
1.53
1.87
2.28
2.61
2.96
3.33
3.7
0
1,143.55
1,749.30
2,266.5
2,672.73
3,020.83
3,345.22
3,651.21
3,920.44
1
2
3
4
5
6
7
8
9
10
PVP K90 polyvinylpyrrolidone K90
RESULTS
Table II. In Vitro Dissolution Prole of drug, Physical Mixtures, and Solid Dispersions of Gliclazide in pH 1.2 (0.1 N HCl)
Dissolution parameters
Sr. no.
Formulation
1
2
3
4
5
6
7
Drug
PM 1:1
PM 1:2
PM 1:5
SD 1:1
SD 1:2
SD 1:5
Q10
min
18.46
25.36
33.1
43.88
76.29
77.12
77.73
Q20
min
32.67
38.93
48.8
58.54
80.36
87.12
93.95
Q30
min
40.82
56.68
59.5
65.77
88.24
92.45
95.84
%DE10
9.16
12.68
16.55
21.94
38.15
45.56
49.87
min
%DE30
23.67
30.88
37.21
45.11
66.92
70.16
73.20
min
MDT (min)
12.5
13.66
11.24
9.43
7.25
7.24
7.09
PM physical mixture, SD solid dispersion of gliclazide prepared by the solvent evaporation method, %DE percent dissolution efciency, MDT
mead dissolution time
332

Table III. Statistical Parameters of Various Formulations After Fitting Drug Release Data to Various Release Kinetic Models
Zero order model
First order model
H M model
P K model
Formulations
k1
k2
k3
Drug
PM 1:1
PM 1:2
PM 1:5
SD 1:1
SD 1:2
SD 1:5
0.9813
0.8256
0.7319
0.5379
0.1647
0.2008
0.1099
1.4736
1.2789
1.3776
1.5087
2.0615
2.1641
2.2337
0.9939
0.8975
0.8582
0.7312
0.7982
0.8956
0.9362
0.0184
0.0190
0.0213
0.0247
0.0533
0.0685
0.0903
0.9943
0.9764
0.9695
0.9265
0.8589
0.8646
0.8560
7.1358
8.6575
9.4130
10.4269
14.3741
15.0883
15.5967
0.9952
0.9560
0.9608
0.9421
0.9812
0.9809
0.9166
H C model
k4, n
3.4734,
8.5742,
15.1667,
26.6687,
57.9627,
58.4002,
60.4486,
0.7249
0.5041
0.3712
0.2439
0.1171
0.1287
0.1284
K5
0.9905
0.8774
0.8214
0.6742
0.6450
0.7345
0.7612
0.0057
0.0055
0.0061
0.0069
0.0123
0.0143
0.0165
H M Higuchi matrix, P K Peppas Korsmeyer, H C Hixon Crowell, R correlation coefcient, k1 k5 constants of release kinetics, PM physical
mixture, SD solid dispersion of gliclazide prepared by the solvent evaporation method
Figure 2 shows the X ray diffraction pattern of pure

gliclazide, PVP K90, its physical mixture, and solid dispersion.
In the X ray diffraction pattern of gliclazide, sharp peaks are
present at 2 of 10.59, 14.98, 17.2, 17.85, 18.15, 22.07, 25.42,
26.25, 26.75, and 29.5 (Fig. 2 A), and it suggests that gliclazide
is a crystalline material. Pure PVP K90 shows absence of
peaks in diffraction spectrum (Fig. 2 B). Peaks characteristic
of the drug were observed in the X ray diffractogram of
physical mixture and solid dispersion.
of gliclazide is characterized by the absorption of carbonyl

(C=O) sulfonyl urea group at 1,706 cm 1 (6). In the spectra of
SDs and PMs, this band was shifted towards higher wave
number at 1,711 and 1,709 cm 1, respectively (6). Also, the
NH group which is located at 3,265 cm 1 from the IR
spectrum of gliclazide was shifted to 3,630 cm 1 in SDs. The
sulfonyl group bands are located at 1,349 and 1,162 cm 1 in
pure gliclazide. In SDs, the asymmetrical vibration peak of
S = 0 band was shifted from 1,349 to 1,342 cm 1 with
decreased frequencies. In SDs, the symmetrical stretching
vibration band of S=0 was shifted from 1,162 to 1,113 cm 1
with decreased frequencies. The spectrum of PVP K90
exhibited important bands at 2,953 cm 1 (C H stretch) and
1,652 cm 1 (C=O). A very broad band was also visible at
3,446 cm 1 which was attributed to the presence of water,
conrming the appearance of broad endotherm in DSC run
due to the presence of water (20).
FT-IR Spectroscopy
DISCUSSION
The IR spectra of SDs and PMs were compared with the

standard spectrum of gliclazide (Fig. 3 A). The IR spectrum
The results of phase solubility are in accordance with the

well established formation of soluble complexes between
Fig. 1. DSC thermograms of pure gliclazide a, pure PVP K90 b,

gliclazide PVP K90 PMs at 1:2 ratio c, and gliclazide PVP K90 SDs
at 1:2 ratio d
Fig. 2. X ray diffractograms of pure gliclazide a, pure PVP K90 b,

at 1:2 ratio d
tions led to disappearance of the endotherm, which is due

to evaporation of water during the rst run. The presence
of broad endothermic peak of PVP K90 and formulations
on DSC thermogram was reported by several researchers
(19 21).
X-ray Diffractions
Properties of Gliclazide in Polyvinylpyrrolidone K90
Fig. 3. FT IR spectrograms of pure gliclazide a, pure PVP K90 b,

at 1:2 ratio d
water soluble polymeric carriers and poorly water soluble

drugs (17). Gtr values were all negative for PVP K90 at
various concentrations indicating the spontaneous nature of
the drug solubilization. The values decreased by increasing
PVP K90 concentration, demonstrating that the reaction
became more favorable as the concentration of PVP K90
increased.
The results of the dissolution study indicate an improve
ment of dissolution rate of gliclazide in solid dispersion. The
rate of dissolution increases as the concentration of PVP K90
increases in SDs. The improvement of dissolution rate is
possibly caused by several factors. Such factors are: (a) the
strong hydrophilic character of PVP K90, which improves the
water penetration and the wettability of the hydrophobic
gliclazide; (b) the optimal dispersion of gliclazide to PVP K90;
(c) the absence of crystals (amorphous dispersions) corre
sponds to lower energy required for dissolution; and (d) the
intermolecular hydrogen bonds and the molecular dispersion
of gliclazide on PVP leads to partial miscibility, improving the
hydrophilic characteristics of the drug substance via inter
actions within the polymer (22). The improvement of
dissolution rate of gliclazide in PMs is due to increased
wettability of the drug powder (23).
Thermograms of SD (Fig. 1 D) and PM showed the
absence of a gliclazide melting endothermic peak. The
absence of drug melting endothermic peak in PM and SD
indicate that gliclazide is present in amorphous form within
PMs and SDs. The inhibition of crystallization of drug in
dispersions results in the amorphous form of gliclazide.
Crystallization inhibition is attributed to two effects: inter
actions such as hydrogen bonding between the drug and the
polymer and the entrapment of the drug molecules in the
polymer matrix during solvent evaporation or a combination
of both (24). Numerous studies have shown that polymers
like povidone used in solid dispersions can inhibit the
crystallization of drugs resulting in an amorphous form of
the drug in the solid dispersions (25,26). The formation of an
amorphous form of gliclazide in SDs is due to the combina
tion of the two stated effects (22,24). In order to verify the
333
DSC results and to exclude the possibility of existence of
crystalline material in solid dispersion, these systems were
again evaluated by using X ray diffraction analysis.
The prominent peaks from pure gliclazide such as at 2
of 10.59, 14.98, 17.2, etc. were observed clearly at the same
position in the PMs and in SDs, and their intensities were
decreased by 55 60% and by more than 80%, respectively.
The X ray diffraction ndings suggested that some portion of
gliclazide still existed in the same crystal structures of the
pure drug, but the relative reduction of diffraction intensity of
gliclazide in SDs at these angles suggests that the crystal size
was reduced to microcrystal form (27). The nding of the
XRD such as existence of some portion of gliclazide in the
same crystal structures of the pure drug is not in agreement
with the DSC nding (complete conversion of drug to
amorphous form) (28). The XRD ndings again suggest that
the melting peak of gliclazide in SDs and PMs containing
some portion of crystalline gliclazide was also absent. This is
due to the interaction between gliclazide and polymer in solid
state (28). This further conrms that DSC is not useful for
examining the solid state of drug within the PMs and SDs
(28). The existence of some portion of gliclazide in the same
crystal structures of pure drug was conrmed by the XRD
study but not by the DSC study. The absence of an
endothermic peak of gliclazide in SDs and PMs suggests an
amorphous form of the drug, but again the presence of a
sharp peak of the drug with lesser intensity indicates some
portion of crystalline drug. DSC and XRD results are in
agreement when the crystalline form of the drug is converted
into amorphous form completely. Absence of endothermic
and diffraction peak of the drug in dispersion indicate an
amorphous form, where DSC and XRD results are in
agreement with each other. Absence of an endothermic peak
of the drug in dispersion and presence of a diffraction peak in
XRD with less intensity indicate that the drug is partially
converted into amorphous form, and the DSC and XRD
results, which are not in agreement with each other, could be
due to the presence of some portion of the crystalline drug
and the presence of polymer. Microcrystals are formed as a
consequence of evaporation of solvent during the preparation
of solid dispersions by the solvent evaporation technique.
Evaporation of solvent increases the viscosity very rapidly
leading to a decrease in drug mobility preventing re
crystallization. When the solvent is evaporated completely,
drug molecules are frozen in the polymer. A crystal lattice is
not formed, but the drug molecules are of randomly
dispersed order over only a few molecular dimensions
(20). The existence of interaction between the drug and
polymer is again suggested by FT IR study.
The shift in the peaks associated with C=O, S=O, and
NH groups of gliclazide indicates some sort of solid state
interactions between the drug and the polymer in SD and
PM. The interactions are due to intermolecular hydrogen
bonding between the drug and polymer. An intermolecular
hydrogen bond was expected to occur between the hydrogen
atom of the NH group of gliclazide and one of the lone pair
electrons of C=O group of polymer and/or C=O group of
gliclazide and one of the hydrogen atoms of PVP K90 (20).
The shift in the peaks associated with the S=O group of
gliclazide could be due to involvement of complexation with
PVP K90 (29,30). The FT IR data suggest the formation of
334
intermolecular hydrogen bonding between gliclazide and

PVP K90. The absence of the endothermic peak of the drug
in SDs or PMs is due to intermolecular hydrogen bonding and
partial conversion drug into amorphous form.
10.
Conclusions
11.
The solubility and dissolution rate of gliclazide can be

enhanced by formulating SDs of gliclazide with PVP K90.
The solubilization effect of PVP K90, reduction of particle
aggregation of the drug, formation of microcrystalline or
amorphous drug, increased wettability and dispersibility, and
development of intermolecular hydrogen bonding are respon
sible for the enhanced solubility and dissolution rate of
gliclazide from its SDs and to some extent in PMs. The
results of infrared spectroscopy, X ray diffractometry, and
DSC indicate that some sort of interactions such as intermo
lecular hydrogen bonding or complexation between the
functional groups of gliclazide and PVP K90 have occurred
in the molecular level and show microcrystallinity of glicla
zide in the solid dispersion, which increased the solubility and
the dissolution of gliclazide from the solid dispersion.
12.
13.
14.
15.
16.
17.
ACKNOWLEDGMENTS
18.
The authors are grateful to Aristo Pharmaceuticals Pvt.

Ltd, Mumbai, India. Mr. S. Biswal acknowledges the Principal
of the Royal College of Pharmacy and Health Sciences,
Brahmapur, for providing the necessary facilities.
19.
REFERENCES
1. Reynolds JEF, editor. Martindale. The Extra Pharmacopoiea
XXX. 30th edn. London: The Pharmaceutical Press; 1993. p.
279 80.
2. Palmer KJ, Brogden RN. Gliclazide, an update of its pharmaco
logical properties and therapeutic efcacy in NIDDM. Drugs
1993;46:92 125.
3. Lobenberg R, Amidon GL. Modern bioavailability, bioequiva
lence and biopharmaceutics classication system: new scientic
approaches to international regulatory standards. Eur J Pharm
Biopharm. 2000;50:3 12.
4. Desai KGH, Kulkarni AR, Aminabhavi TM. Solubility of
rofecoxib in presence of methanol, ethanol and sodium lauryl
sulfate at (298.15, 303.15 and 308.15) K. J Chem Eng Data.
2003;48:942 5.
5. Harrower AD. Comparison of efcacy, secondary failure rate
and complications of sulfonylurea. J Diabetes Complicat.
1994;8:201 3.
6. zkan Y, Atay T, Dkman N, Isimer A, Aboul Enein YH.
Improvement of water solubility and in vitro dissolution rate of
gliclazide by complexation with cyclodextrin. Pharm Acta
Helv. 2000;74:365 70.
7. Winters S, York P, Timmins P. Solid state examination of a
gliclazide: beta cyclodextrin complex. Eur J Pharm Sci.
1997;5:209 14.
8. Aggarwal S, Singh PN, Mishra B. Studies on solubility and
hypoglycemic activity of gliclazide beta cyclodextrin hydrox
ypropylmethylcellulose complexes. Die Pharmazie. 2002;57:
191 3.
9. Hong SS, Lee SH, Lee YJ, Chung SJ, Lee MH, Shim CK.
Accelerated oral absorption of gliclazide in human subjects from
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
a soft gelatin capsule containing a PEG 400 suspension of

gliclazide. J Control Release. 1998;51:185 92.
Biswal S, Sahoo J, Murthy PN, Giradkar PR, Avari JG.
Enhancement of dissolution rate of gliclazide using solid
dispersions with polyethylene glycol 6000. AAPS PharmaSciTech.
2008;9(2):563 70.
Trapani G, Franco M, Latrofa A, Pantaleo MR, Provenzano
MR, Sanna E, et al. Physicochemical characterization and in vivo
properties of Zolpidem in solid dispersions with polyethylene
glycol 4000 and 6000. Int J Pharm. 1999;184:121 30.
Trapani G, Franco M, Latrofa A, Tullio C, Provenzano MR,
Serra M, et al. Dissolution properties and anticonvulsant activity
of phenytoin polyethylene glycol 6000 and polyvinylpyrrolidone
K 30 solid dispersions. Int J Pharm. 2001;225:63 73.
Higuchi T, Connors K. Phase solubility techniques. Adv Anal
Chem Instrum. 1965;4:17 123.
Damian F, Blaton N, Naesens L, Balzarini J, Kinget R,
Augustinjns P, et al. Physicochemical characterization of solid
dispersions of the antiviral agent UC 781 with polyethylene
glycol 6000 and Gelucire 44/14. Eur J Pharm Sci. 2000;10:311 22.
Khan CA, Rhodes CT. The concept of dissolution efciency. J
Pharm Pharmacol. 1975;27:48 9.
Arias MJ, Gines JM, Moyano JR, Rabasco AM. Dissolution
properties and in vivo behaviour of triamterene in solid
dispersions with polyethylene glycols. Pharm Acta Helv.
1996;71:229 35.
Costa P, Lobo JMS. Modeling and comparison of dissolution
proles. Eur J Pharm Sci. 2001;13:123 33.
Korsemeyer RW, Gurney R, Doelker E, Buri P, Peppas NA.
Mechanisms of solute release from porous hydrophilic polymers.
Int J Pharm. 1983;15:25 35.
Albertini B, Cavallari C, Passerini N, Gonzlez Rodrguez ML,
Rodriguez L. Evaluation of b lactose, PVP K12 and PVP K90 as
excipients to prepare piroxicam granules using two wet granula
tion techniques. Eur J Pharm Biopharm. 2003;56:479 87.
Van den Mooter G, Augustijns P, Blaton N, Kinget R. Physico
chemical characterization of solid dispersions of temazepam with
polyethylene glycol 6000 and PVP K30. Int J Pharm.
1998;164:67 80.
Khougaz K, Clas SD. Crystallization inhibition in solid dispersion
of MK 0591 and poly (vinylpyrrolidone) Polymers. J Pharm Sci.
2000;89(10):1325 34.
Karavas E, Ktistis G, Xenakis A, Georgarakis E. Miscibility
behavior and formation mechanism of stabilized felodipine
polyvinylpyrrolidone amorphous solid dispersions. Drug Dev Ind
Pharm. 2005;31:473 89.
Ford JL. The current status of solid dispersions. Pharm Acta
Helv. 1986;61:69 88.
Bettinetti GP, Mura P, Giordano F, Setti M. Thermal behaviour
and physico chemical properties of naproxen in mixtures with
polyvinylpyrrolidone. Thermochim Acta. 1991;199:165 71.
Yamashita K, Nakate T, Okimoto K, Ohike A, Tokunaga Y,
Ibuki R, et al. Establishment of new preparation method for
solid dispersion formulation of tacrolimus. Int J Pharm.
2003;267:79 91.
Tantishaiyakul V, Kaewnopparat N, Ingkatawornwong S. Prop
erties of solid dispersions of piroxicam in polyvinylpyrrolidone.
Int J Pharm. 1999;18:143 51.
Valizadeh H, Nokhodchi A, Qarakhani N, Zakeri Milani P,
Azarmi S, Hassanzadeh D, et al. Physicochemical characteriza
tion of solid dispersions of indomethacin with PEG 6000, Myrj
52, lactose, sorbitol, dextrin, and Eudragit E100. Drug Dev Ind
Pharm. 2004;30(3):303 17.
Nair R, Gonen S, Hoag SW. Inuence of polyethylene glycol and
povidone on the polymorphic transformation and solubility of
carbamazepine. Int J Pharm. 2002;240:11 22.
Hosono T, Tsuchiya S, Matsumura H. Model of interaction of
ajmaline with polyvinylpyrrolidon. J Pharm Sci. 1980;69:824 6.
Najib NM, Suleiman M, Malakh A. Characteristics of the in vitro
release of ibuprofen from polyvinylpyrrolidone solid dispersions.
Int J Pharm. 1986;32:229 36.
AAPS PharmSciTech, Vol. 10, No. 2, June 2009 (# 2009)

DOI: 10.1208/s12249 009 9214 5
Research Article
Agglomerates Containing Pantoprazole Microparticles: Modulating
the Drug Release
Renata P. Raffin,1,4 Paolo Colombo,2 Fabio Sonvico,2 Alessandra Rossi,2 Denise S. Jornada,1
Adriana R. Pohlmann,1,3 and Silvia S. Guterres1
Received 1 October 2008; accepted 21 February 2009; published online 25 March 2009
Abstract. Pantoprazole loaded microparticles were prepared using a blend of Eudragit S100 and
Methocel F4M. The accelerated stability was carried out during 6 months at 40C and 75% relative
humidity. In order to improve technological characteristics of the pantoprazole loaded microparticles,
soft agglomerates were prepared viewing an oral delayed release and gastro resistant solid dosage form.
The agglomeration was performed by mixing the pantoprazole microparticles with spray dried mannitol/
lecithin powders. The effects of factors such as the amount of lecithin in the spray dried mannitol/lecithin
powders and the ratio between pantoprazole microparticles and spray dried mannitol/lecithin powders
were evaluated. The pantoprazole loaded microparticles present no signicant degradation in 6 months.
The agglomerates presented spherical shape, with smooth surface and very small quantity of non
agglomerated particles. The agglomerates presented different yields (35.5 79.0%), drug loading (58
101%), and mechanical properties (tensile strength varied from 44 to 69 mN mm 2), when the spray
dried mannitol/lecithin powders with different lecithin amounts were used. The biopharmaceutical
characteristics of pantoprazole microparticles, i.e., their delayed release properties, were not affected by
the agglomeration process. The gastro resistance of the agglomerates was affected by the amount of
spray dried mannitol/lecithin powders. The ratio of lecithin in the spray dried mannitol/lecithin powders
was the key factor in the agglomerate formation and in the drug release proles. The agglomerates
presenting better mechanical and biopharmaceutical characteristics were prepared with 1:2 (w/w) ratio of
pantoprazole loaded microparticles and mannitol/lecithin (80:20) powder.
KEY WORDS: agglomerates; delayed release; gastro resistance; microparticles.
INTRODUCTION
Polymeric drug delivery systems have potential therapeu
tic advantages in comparison with conventional dosage forms
such as the reduction of drug side effects (1), the improvement
of therapeutic effect (2), and the control of drug release (3),
decreasing the administration frequency (4). Among the
different techniques used to prepare drug loaded micropar
ticles, spray drying is a one stage continuous process of easy
scaling up, which is barely dependent on the solubility of drugs
and polymers (5,6). This technique provides microparticles,
diameters of which range from few to several tens of micro
meters with relatively narrow size distribution (5). The
production of a spray dried powder involves the droplet
Programa de Ps Graduao em Cincias Farmacuticas, Faculdade

de Farmcia, Universidade Federal do Rio Grande do Sul, Av.
Ipiranga 2752/405, 90610 000, Porto Alegre, Brazil.
2
Dipartimento Farmaceutico, Universit degli Studi di Parma, Parma,
Italy.
3
Departamento de Qumica Orgnica, Instituto de Qumica,
Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.
4
To whom correspondence should be addressed. (e mail: renata.
rafn@ufrgs.br)
formation from the atomized suspension, solution or emulsion

followed by their solidication driven by the solvent evapora
tion (7). The desired microstructure can be created from the
complex mixture of polymers, surfactants and the appropriate
drug that gives the in use properties of the product (8). The
particle size of spray dried powders depends on the viscosity of
the feed solution, design of the atomizer, air pressure, and air/
spray contact (9,10). The boiling point of the solvent, as well as
the spray drier scale, can affect particle sizes (10,11). Micro
particulated dried powders, granules, or agglomerates can be
used as oral dosage forms (4,9,12) and nasal powders (13,14).
Among the various drug delivery devices used to sustain
the drug release, the hydrophilic matrix systems are generally
preferred because of their ability to swell, coalesce, and form a
viscous gel. The viscous layer acts as a diffusional barrier to the
drug (15). Cellulose derivatives are frequently chosen to
develop such systems due to their low toxicity and cost. Hydro
xypropylmethylcellulose (HPMC) is a nonionic cellulose deriv
ative that presents different grades of substitution. Methocel
F4M is a HPMC with viscosity of 3800 mPa s (2% solution in
water, 20C). Methacrylate copolymers (Eudragit) are inter
esting candidates for the production of microparticles by spray
drying since they are inert and freely soluble in organic solvents
(16,17). Eudragit S100 is a pH dependent enteric copolymer
composed of methacrylic acid and methylmethacrylate mono
335
Raffin et al.
336
mers, rendering it soluble in aqueous solutions presenting pH
higher than 7 (18). In this way, it is insoluble in the mouth and in
the stomach but starts to dissolve in the duodenum (pH around
6). Eudragit S100 has in its structure a hydrophobic monomer
unit (methyl methacrylate) and another monomeric unit
(methacrylic acid), the behavior of which is dependent on its
protonation state. At pH over 7, the carboxylic groups became
ionized and the polymer soluble; at lower pH, the carboxylates
groups become not ionized and the polymer precipitates (19). In
previous works, Methocel F4M was blended with Eudragit
S100, and its aqueous solution was spray dried to produce
pantoprazole loaded microparticles (10,12).
Pantoprazole (5 (diuoromethoxy) 2 [[(3,4 dimethoxy 2
pyridinyl)methyl] sulnyl] benzimidazole) is a prodrug used
in the treatment of digestive ulcers, gastroesophageal reux
disease, as well as an adjuvant in the eradication of the
Helicobacter pylori (20). In acid medium, pantoprazole turns
into a cationic sulfenamide, which is its active form. It accu
mulates in the highly acidic environment of the parietal cell
canalicular lumen and it is activated (20). The active form, a
tetracyclic cationic sulfenamide, reacts with thiol group of
cysteines 813 and 822 of the transmembranal H+/K+ ATPase.
This conversion must occur inside the gastric parietal cells, so
pantoprazole must be absorbed intact by the gastrointestinal
tract (20).
Pantoprazole loaded microparticles showed both charac
teristics of gastro resistance and controlled release (10). Eudra
git S100 was responsible for the gastro resistance, whereas the
presence of hydroxypropylmethylcellulose (HPMC) delayed
the release in comparison with formulations prepared without
this polymer (9). These microparticles were formed by a solid
solution of the polymers, and the pantoprazole was molecularly
dispersed in the blend (12). Those microparticles also demon
strated an anti ulcer activity in rats in an ethanol induced ulcer
in vivo model (12). The scaling up of the spray drying process
was validated according to the following parameters: total solid
concentration in the solution feed, type of atomizer, air pressure,
and air/spray contact (10). The more adequate conditions were
6.6% of solids, two uid nozzle atomizer, co current ow, and
air pressure of 196 kPa (10).
Unfortunately, the attainment of biopharmaceutical attrib
utes by microparticles is opposed by the small size of the
particles that leads to powders with bulk volume and problem
atic ow for dosage forms manufacturing (14,21). In several
pharmaceutical applications, particles might be ne for drug
delivery but coarse enough to facilitate the solid dosage form
preparation.
Often, the transformation of microparticles in solid dosage
forms involves granulation and compaction, provoking irrevers
ible modications of the microparticle range size and structure
(22). In particular, this technological problem could be solved
using soft agglomeration, a process in which the powder size is
enlarged by constructing weak clusters of primary micropar
ticles (14). Soft agglomerates are easily broken down by air
turbulence or water uptake, reconstituting the original size of
the microparticles. Weak cohesion bonds due to capillary, van
der Waals, or electrostatic forces hold together the primary
particles in soft structures (23). The quantity and the nature of
these interactions, as well as the method of production,
determine the agglomerate structures (24). Recently, a new
procedure for agglomerating microparticles has been described
(25). Morphine crystals have been agglomerated in soft clusters

by processing the physical mixture of the drug with spray dried
mannitol/lecithin powders (25). The lecithin was used as a
binder to improve the interparticle cohesion reinforcing the
internal structure of agglomerates (26). Taking those ndings
into account, we hypothesized that the soft agglomeration pro
cedure could be applied in preparing stable soft agglomerates of
pantoprazole loaded microparticles formulated using a blend of
Eudragit S100 and Methocel F4M. Additionally, the pur
pose of this research was to study the accelerated stability of
Eudragit S100 and Methocel F4M blended microparticles,
as well as to study the release control of pantoprazole by
changing the composition of the spray dried mannitol/lecithin
powders in the agglomerates.
Materials
Sesquihydrate sodium pantoprazole (Mw =432.4, pKa1 =
3.92, pKa2 =8.19, freely soluble in water, very slightly soluble
in phosphate buffer at pH 7.4, and practically insoluble in n
hexane) was purchased from Henrifarma (So Paulo, Brazil).
Eudragit S100 [(poly(methacrylic acid co methyl methacry
late) 1:2, average Mw =135,000, pKa approximately 6, soluble
in methanol, ethanol, in aqueous isopropyl alcohol, and acetone,
as well as in 1 N sodium hydroxide] was a kind gift from
Almapal (So Paulo, Brazil, produced by Rohm, Germany).
Methocel F4M [HPMC, methoxyl content of 27 30%,
hydroxypropyl content of 4 7.5%; soluble in cold water, prac
tically insoluble in chloroform, ethanol (95%), and ether] was
provided by Colorcon (So Paulo, Brazil, produced by Dow
Chemical, USA). Mannitol (Ph. Eur.) was a gift of Lisapharma
(Como, Italy), and lecithin (Lipoid S45) was supplied by Lipoid
AG (Ludwigshafen, Germany). Acetonitrile was of high
performance liquid chromatography (HPLC) grade and sup
plied by Tedia (Faireld, USA). Sodium hydroxide was
obtained from Merck (Darmstadt, Germany). Ethanol, hydro
chloric acid, monohydrate sodium phosphate, and hexane were
of analytical grade.
Preparation of Pantoprazole-Loaded Microparticles
Pantoprazole loaded microparticles were prepared in pilot
scale as previously described (10). Briey, the feed solution
consisted of 36 g of Eudragit S100 in 2,000 mL NaOH solution
(3 g L 1). After its complete dissolution, Methocel F4M (18 g)
was added, and the solution was kept at 10C for 24 h. Sodium
pantoprazole (9 g) was added just before spray drying. The
solution was then spray dried in pilot scale equipment (Model
PSD 52 APV1Anhydro, Denmark) presenting the cylindrical
dryer chamber of 1.0 m diameter and 2.3 m of total height. A
two uid pneumatic atomizer with external mixing was used. In
this nozzle, the liquid to be atomized is discharged through a
central hole diameter of do =1.5 mm, whereas the atomizing air
is injected through a ring area around the liquid hole. The
atomizing air pressure was 196 kPa, the inlet temperature was
1701C, and the ow rate was 2 L h 1. Room temperature and
humidity were controlled (241C and 542% of relative
humidity). One batch was prepared and used for the
Agglomerates Containing Pantoprazole Microparticles

accelerated stability. Another batch was produced and used in
the preparation of all agglomerates.
Accelerated Stability Tests
of Pantoprazole-Loaded Microparticles
Transparent glass vials containing 0.5 g of microparticles
were stored for 6 months in a stability chamber at 40C and
75% relative humidity (RH) (27,28). During 180 days, sealed
and non sealed vials were evaluated every 30 days for their
drug content. Humidity was gravimetrically determined.
Drug Loading
The drug loading was assayed by a validated HPLC
method (29). Briey, an amount of agglomerates, equivalent
to 10 mg of pantoprazole, was weighed and magnetically stirred
with 40 of 0.05 molL-1 NaOH solution for 1 h in a volumetric
ask. The volume was completed to 50 mL, and drug
concentration was determined after ltration (0.45 m) by
HPLC (Perkin Elmer serie 200, detector UV Vis) using a
LiChrospher RP18 (Merck) column. The mobile phase
consisted of acetonitrile/phosphate buffer pH 7.4 (35:65 v/v).
The ow rate was 1 mLmin 1, and the drug was detected at
290 nm. Retention time was 6.8 min. Linearity was achieved
between 0.5 and 20 g mL 1 and presented a correlation
coefcient of 0.9994. The relative standard deviation for its
precision was 0.47% and for its reproducibility 1.13%. Accuracy
was between 95.4% and 101.4% for different concentrations.
Preparation and Characterization
of Spray-Dried Mannitol/Lecithin Powders
The feed solution of mannitol and lecithin was prepared
as follows: lecithin (15%, 17.5%, and 20%) was dissolved in
60 mL ethanol. In another beaker, mannitol (85%, 82.5%,
and 80%) was dissolved in water (340 mL). Both solutions
were then mixed, and the nal ratios between them were
85:15, 82.5:17.5, and 80:20 (w/w). The resulting solutions had
15% of ethanol and 4.5% of solids. The solutions were spray
dried in a laboratorial spray drier Buchi Mini Spray Dryer B
191 (Buchi Laboratoriums Tecnik, Flawil, Switzerland) using
ow rate of 6.5 mL min 1, inlet temperature of 902C,
aspiration set in 100%, and air ow of 500 NL h 1. The three
spray dried mannitol/lecithin powders were used as excipients
for the agglomeration with pantoprazole loaded microparticles.
They were characterized in terms of yield, humidity content,
morphology, and specic surface area.
The yield, expressed in percent, was calculated by the
ratio between the mass obtained and the mass of mannitol
and lecithin added to the solution. The particle size distribu
tions of spray dried mannitol/lecithin powders were measured
using a laser light diffraction apparatus (series 2600 Malvern
Instruments Ltd, Spring Lane South Malvern, Worcestershire,
UK). Particles were suspended in ethyl acetate, a non solvent
for these materials. Particle size was expressed as median
volume diameter. After gold sputtering, the morphology of
the spray dried mannitol/lecithin powders was assessed by
scanning electron microscopy (SEM) using accelerating volt
age of 15 kV (JSM 6400, Jeol Ltd, Tokyo, Japan).
337
The specic surface areas of spray dried mannitol/
lecithin powders were determined by the Brunauer, Emmet,
and Teller multipoint technique (BET) (30). The nitrogen
adsorption desorption isotherms of previous degassed organ
ic solids, under vacuum at 40C, were determined at liquid
nitrogen boiling point in a homemade volumetric apparatus,
using nitrogen as probe. The pressure was measured using
capilar mercury barometer and the results were compared to
an alumina pattern.
Water content was assayed by Karl Fisher titration (Titro
Matric 1S, Crison, Alella, Spain). Different test methods to
determine owability, which vary in nature and degree of force
and energy transmission, quantify the macroscopic properties of
a powder in different ways. Therefore, different test procedures
provide different answers and powder classications (31).
Flowability, as well as bulk and tapped densities, were measured
according to the European Pharmacopoeia (32). The compress
ibility index was calculated according to USP (33). The
differential scanning calorimetry (DSC) was performed in a
DSC 60 (Shimadzu, Kyoto, Japan) at 10C min 1 from 140C
to 200C.
Preparation of the Agglomerates
The agglomerate preparation is summarized in Fig. 1.
Pantoprazole loaded microparticles and spray dried mannitol/
lecithin powders were mixed in a Turbula apparatus (Wab, Basel,
Switzerland) for 3 h. Five grams of the mixture of pantoprazole
loaded microparticles and spray dried mannitol/lecithin powders
were put on the top of two sieves stack with nominal apertures of
106 and 850 m, respectively (10 cm sieves, Endecotts Ltd,
London, UK). The mixture was vibrated for 10 min on a
laboratory sieve shaker (amplitude 2 3; Analysette 3 Fritz
model, Fritsch GMBH, Idar Oberstein, Germany). Agglomer
ates between 106 and 850 m were collected. The reprocess of
the non agglomerated powder and the crushing of larger
agglomerates was repeated eight times. The ratios tested were
1:1, 1:2, and 1:3 (w/w) (Table I). The drug content measured by
HPLC was used to determine powder homogeneity.
Characterization of the Agglomerates
The yield of the agglomeration process was calculated by
dividing the weight of the agglomerates (106 850 m) by the
total weight of powder before agglomeration, multiplied by 100.
The agglomerates were examined by SEM, as described
before. Mean size distribution was veried by sieving (106, 250,
425, 500, 600, 710, 850 m, 10 cm, Endecotts Ltd, London,
UK). The average diameter was calculated by determining the
mass retained in each sieve. The specic surface area was
calculated by BET method (30). The nitrogen adsorption
desorption isotherms of previous degassed organic solids were
obtained, under vacuum at 40C, at liquid nitrogen boiling point
in a homemade volumetric apparatus, using nitrogen as probe.
The pressure measurements were made using a capillary Hg
barometer and also an active Pirani gauge. The results were
systematically compared with an alumina standard reference.
Flowability and compressibility index were determined
using the same procedure described above for the spray dried
mannitol/lecithin powders.
Raffin et al.
338
Fig. 1. Preparation of the pantoprazole microparticles, spray dried mannitol/lecithin and agglom
erates. Asterisk indicates variable concentrations
Agglomerates (2 g) were tested for resistance using a

friabilometer (Model 300, Nova tica, Brazil) operating at
25 rpm for 4 min (32). The agglomerates were separated from
the powder during the test using a 106 m sieve. The recovered
agglomerates were weighed, and the percentage of powder loss
was calculated.
In order to determine the tensile strength, a single agglo
merate (n=8) was placed on a mobile platform under the mea
suring head of a calibrated load cell (514 QD, DS Europe, Milan,
Italy) (14). The very slow movement of the platform caused the
compression of the agglomerate against the measuring head. The
force time curve was recorded using the Scope v 3.5 software
(AdInstruments Ltd, Oxfordshire, UK). From the crushing force
(F), the tensile strength () was calculated (Eq. 1) (14).
2:8F
d2
where d is the agglomerate diameter.
Table I. Composition of the agglomerates, as well as the nal amount

of lecithin present in the agglomerates
Agglomerates
Lecithin in
excipient
microparticles
(%)
Pantoprazole/
excipient
microparticles
ratio
Percentage of
lecithin
in the agglome
rates (%)
A
B
C
D
E
15.0
17.5
17.5
20.0
20.0
1:3
1:2
1:3
1:1
1:2
11.25
11.67
13.12
10.00
13.33
The agglomerate disintegrations in aqueous media (phos

phate pH 7.4 or 0.1 M HCl) were recorded under an optical
stereomicroscope (magnication of 20; Citoval 2, Jena,
Germany) connected to a video camera (JVC, Tokyo, Japan).
The disintegration tests were performed by placing the agglom
erates (425 500 m) over a microscope glass and wetting them
with 50 L of each medium: phosphate buffer pH 7.4 or 0.1 M
HCl at 37C. The disintegration time was measured on 25 30
agglomerates as referred to the time for deagglomeration of the
globular structure.
To determine the drug release prole, size 00 hard gelatin
capsules without coloring agent were lled with a mass of
agglomerates corresponding to 15 mg of pantoprazole. In vitro
drug release tests were conducted in USP Dissolution Appa
ratus II at 150 rpm at 37C. To evaluate the gastro resistance,
the hard gelatin capsules containing the agglomerates (A to E)
were exposed to 300 mL of 0.1 M HCl. After 1 h, an aqueous
solution (600 mL) composed of NaOH (2.6 g) and KH2PO4
(6.12 g) was added into the dissolution vial in order to reach
pH 7.4 (13,34,35). At this step, sampling began from 0 up to
600 min.
In order to determine if the agglomerates were able to
release 100% of the encapsulated drug, the dissolution was
evaluated in phosphate buffer pH 7.4 for 480 min in the same
apparatus and conditions as referred for the gastro resistance
test.
Pantoprazole concentrations were determined by UV at
295 nm (Vankel UV/Vis spectrometer). The analytical method
was previously validated (36). Gastro resistance proles were t
to monoexponential and biexponential models, using the
MicroMath Scientist software (Salt Lake City, UT, USA).
The best t was chosen considering the highest model selection
criteria (MSC, given by the software), the highest determination
coefcient, and the best graphic adjustment.
339
The Korsmeyer Peppas model was used (Eq. 2) to

determine the drug release mechanism.
ft atn
where ft (dimensionless) is the fraction of pantoprazole

released at time t (min), a (min 1) is a constant that
incorporates structural and geometric characteristics of the
carrier, and n (dimensionless) is the release exponent that
indicates the release mechanism (37).
The adequacy of the models was evaluated considering
the best correlation coefcient and the best model selection
criteria (MSC), both provided by the software, as well as the
best graphic adjustment.
The contact angle was assessed using a Dynamic Contact
Angle Meter and Tensiometer (DCAT 11, Dataphysics,
Filderstadt, Germany). The capillary constant was measured
using n hexano, and the contact angle of the pantoprazole
loaded microparticles and the agglomerates was measured
with water (38).
Statistical Analysis
A one way analysis of variance was employed in the
comparison of the experimental data for signicance at p values
lower than 0.05.
RESULTS AND DISCUSSION
Accelerated Stability Tests
for Pantoprazole-Loaded Microparticles
One batch of pantoprazole loaded microparticles pre
senting mean particle size of 22.01.2 m were obtained as
off white powders (62%) with drug content of 128 mg of
pantoprazole per gram of microparticles and the residual
Fig. 3. The spray dried mannitol/lecithin powders prepared with

15.0% a, 17.5% b, and 20% c of lecithin (magnication, 1,000)
Fig. 2. Drug content of the microparticles during 180 days of

accelerated conditions storage. Sealed and non sealed vials were
evaluated (n 3). Error bars indicate standard deviations
moisture of 2%. Within 180 days of storage, no variation in

the weight was veried either to sealed or non sealed vials,
indicating that pantoprazole microparticles are not hygro
scopic in spite of the use of NaOH in the preparation.
Regarding the drug content, both samples (sealed and non
sealed vials) were stable (Fig. 2). The decay in pantoprazole
content was less than 5% after 6 months of storage.
Raffin et al.
340
Fig. 4. Spray dried mannitol/lecithin containing 17.5% of lecithin a and agglomerates C b (magnication, 8,000)
Characterization of the spray-dried mannitol/lecithin powders

The spray dried mannitol/lecithin powders presented
lecithin contents of 15.0%, 17.5% and 20.0% (w/w). The yield
of the spray drying process was not affected by the lecithin
concentration (approximately 55% for all formulations). In
addition, the particle sizes were not inuenced by the lecithin
concentration, and all the three spray dried mannitol/lecithin
powders had mean diameters of 3.7 m, evidently smaller than
the size of the pantoprazole loaded microparticles (mean
particle size of 22.01.2 m). The spray dried mannitol/
lecithin powders had mean diameters compatible to those
described in the literature for powders prepared using the
same type of equipment (11,39), in which a model plasmid was
encapsulated in PLGA/stearylamine microparticles by spray
drying (39). The microparticles became more uniform in size
and less aggregated and more owable with the increase of
stearylamine. Particle size ranged from 2.1 to 8.3 m, depend
ing on the stearylamine concentration (39). In another
research, DNA microparticles were prepared with polyethyle
nimine as a complexation agent for DNA (11). The size of the
microparticles obtained by spray drying were similar, ranging
from 2.5 to 8.1 m. SEM revealed microparticles with regular
shapes and smooth surfaces (11).
All the powders were composed of spherical particles
(Fig. 3). The powders showed a tendency to form aggregates
as the content of lecithin increased. Their moisture content
were similar for all samples (lower than 2.0%). Particles were
spherical shaped (Fig. 4a), and the mannitol was crystalline
after the spray drying process (veried by DSC, data not
shown). Those spray dried powders presented bulk density
between 0.2 and 0.3 g.cm 3, poor packing (compressibility
index between 22 and 31) and did not ow according to the

European Pharmacopoeia test (32).
Characterization of the Agglomerates

Agglomerates were obtained with yields from 35% to 79%
(Table II), presenting drug loading from 58% to 100%. The
lecithin content used in the present work was higher than the
one described in our previous work (40) due to the lower density
and cohesiveness of the pantoprazole loaded Methocel/Eudra
git microparticles. In the present work, the incomplete drug
loading occurred because of the segregation of the two powders
(spray dried mannitol/lecithin powders and pantoprazole load
ed microparticles). In these cases, the lecithin content (used as a
binder) was not sufcient to aggregate all the amount of
pantoprazole loaded microparticles in larger clusters (the
agglomerates). Using the spray dried mannitol/lecithin powder
containing 15% of lecithin, agglomerates were prepared with
1:3 (w/w) ratio; otherwise, a complete segregation of the two
powders was observed. Agglomerates A presented satisfactory
yield and drug loading of 85.3%.
Mannitol/lecithin powders containing 17.5% of lecithin
were also prepared. However, the 1:1 (w/w) ratio of mannitol/
lecithin powder and pantoprazole loaded microparticles also
presented segregation of powders. Agglomerates B prepared
with a 1:2 (w/w) ratio presented lower yield and lower drug
loading than agglomerates A, and less than 60% of the drug was
incorporated in the clusters. The 1:3 (w/w) ratio was also tested
(agglomerates C). In this case, the agglomerates presented the
highest agglomeration yield (79%) and complete drug loading
(100%).
Table II. Characteristics of the agglomerates (n 3)

Agglomerates
Yield (%)
Drug loading (%)
Bulk density (g cm 3)
Tapped density (g cm 3)
Compressibility (%)
Flowability (s)
A
B
C
D
E
59.41.6
35.34.8
79.00.9
62.42.5
76.93.7
85.34.6
57.90.8
101.02.3
100.33.2
95.51.3
0.240.01
0.210.01
0.220.02
0.150.01
0.190.01
0.280.01
0.260.01
0.230.03
0.170.01
0.240.01
11.81.0
18.10.9
11.32.8
11.80.5
19.50.1
122.222.0
135.633.8
131.614.8
237.929.8
276.830.2
341
Fig. 5. SEM images of agglomerates a, b, c, d, and e (magnication, 100)
In order to maintain the ethanol proportion in the spray

drying solution, the maximum concentration of lecithin tested
was 20% in the spray dried mannitol/lecithin powders due to
the lecithin solubility in this solvent. The agglomeration of
this powder with the pantoprazole loaded microparticles at
the ratios of 1:1 (w/w) and 1:2 (w/w) showed yields of 62 %
and 77%, respectively (Table II). Furthermore, agglomerates
D and E presented complete drug loading. Using the spray
dried mannitol/lecithin powder with 15% of lecithin, it was
not possible to agglomerate all the amount of pantoprazole
loaded microparticles. In the same way, agglomerates B
showed the highest segregation, as well as the lowest yield.
The agglomerates presented spherical shape (Fig. 5). The
globule surface was smooth with a very small quantity of non
agglomerated particles on the surface. The agglomerate sur

face was characterized by small spray dried mannitol/lecithin
powders embedding larger pantoprazole microparticles with
out evident bridges among them (agglomerate C was chosen
as an example in Fig. 6). In detail, some material, likely
lecithin, was spread out over the particles leading to particles
more closely connected (Fig. 4b), which was particularly
evident for the agglomerates containing the mannitol/lecithin
powder prepared using higher content of lecithin (agglomer
ates E). This result suggested that, in the mannitol/lecithin
powder, lecithin could be located at the surface acting as a
binder. The specic surface area of the agglomerates was
measured (Table III). The surface areas of the microparticles
(close to 98 m2 g 1) and of the mannitol/lecithin powders (close
Raffin et al.
342
Russo and co workers (20). The samples prepared with 1:2

(w/w) ratio of pantoprazole microparticles and spray dried
mannitol/lecithin powders had higher tensile strength values.
The agglomerates prepared with 1:1 and 1:3 (w/w) ratios
presented lower values of tensile strength. The 1:2 (w/w) ratio
seemed the optimal composition to improve resistance.
Therefore, the agglomerates presented good resistance while
owing and poor resistance when compressed. Based on these
features, they are suitable for lling hard gelatin capsules
viewing oral administration of these agglomerates containing
pantoprazole microparticles.
Disintegration and Dissolution of the Agglomerates
Fig. 6. Photomicrograph of the surface of agglomerate C (magnica

tion, 1,000)
to 60 m2 g 1) were used to calculate the expected surface area of

the agglomerates, considering the ratio between the components.
In all cases, the measured and the expected areas were very close,
indicating that no changes in the particle structure occurred
during the agglomeration process.
The agglomerates showed values of bulk densities
around 0.20 g cm 3, whose values were higher than those
observed for the pantoprazole microparticles (0.06 g cm 3)
but still corresponding to a loose packing arrangement of
particles (Table II). The density values were determined by
the ratio between pantoprazole loaded microparticles and
mannitol/lecithin powders. The bulk densities were higher for
the agglomerates prepared with higher amounts of spray
dried mannitol/lecithin powder. The compressibility indexes
were around 11 for agglomerates A, C, and D. On the other
hand, agglomerates B and E presented slightly higher values
of compressibility index (18 and 19, respectively).
The agglomeration process, determining the organization
of particles in the globular structure, favored the packed
arrangement of powder bed over primary microparticle pow
ders. The compressibility index that is related to the powder
owability was improved by agglomeration. Pantoprazole
microparticles, as well as the three different spray dried
mannitol/lecithin powders, showed ow in innite time (the
entire samples failed to ow under conditions prescribed for the
owability test). In contrast, the agglomerates owed well
(Table II). Agglomerates A, B, and C presented higher
owability than agglomerates D and E (p<0.01). However, all
products were classied as free owing powders. In summary,
the agglomerates showed characteristics of packing arrange
ment and owability more adequate for handling and dose
metering than the pantoprazole loaded microparticles.
In order to evaluate the resistance of the agglomerates
during transportation, the friability test was performed for the
agglomerates. Values were statistically similar (p=0.32) from
1.06% to 2.48% (Table III). Increasing the amount of mannitol/
lecithin powders,l the friability increased, demonstrating that
there were some particles not embedded in the globular
structure. Pantoprazole agglomerates had a very low resistance
to crushing, and the tensile strength values (Table III) were
between 44 and 69 mN mm 2, similar to those reported by
In phosphate buffer pH 7.4, the agglomerates were

slowly penetrated and slightly swollen by the solvent, main
taining the globular structure. Only agglomerates prepared
with 1:3 (w/w) ratio (A and C) disintegrated after 2 min. The
soluble spray dried mannitol/lecithin powders in large quan
tity inuenced the gel layer formation of HPMC in the
pantoprazole loaded microparticles, facilitating the water
penetration and the drug diffusion. The mannitol acted as a
disintegrant, and increasing mannitol amount resulted to
faster disintegration. Agglomerates B and E prepared with
1:2 (w/w) ratio disintegrated more slowly within 5 10 min.
Agglomerates D did not disintegrate within 20 min. In order
to understand the inuence of lecithin and mannitol on the
disintegration behavior, agglomerates constituting exclusively
of spray dried mannitol/lecithin powder were tested at
pH 7.4. These agglomerates disintegrated within 2 min,
indicating that the slow disintegration is dependent of HPMC
and that the spray dried mannitol/lecithin powder acted as a
disintegrant in the agglomerates. Varying the amount of the
spray dried mannitol/lecithin powders in the agglomerates,
the modulation of the disintegration time was achieved.
The stability of all samples (pantoprazole loaded micro
particles and agglomerates) in phosphate buffer pH 7.4 was
evaluated, showing that the pantoprazole loaded micropar
ticles and the agglomerates reached 100% of pantoprazole
dissolution after 500 min. The drug did not degrade in this
experimental condition. These results indicated that the spray
drying and the agglomeration techniques, as well as the
medium used in the release experiments, did not affect the
stability of pantoprazole.
Concerning the gastro resistance evaluation, the agglom
erates showed different results from the pantoprazole micro
particles in terms of dissolution prole and total amount after
the acid step (Fig. 7). Pantoprazole microparticles showed
Table III. Specic surface area and mechanical properties of the

agglomerates
Agglomerates
Specic surface
area (m2 g 1)
Friability
(%) (n 3)
Tensile strength
(mN mm 2) (n 8)
A
B
C
D
E
76
85
70
73
78
2.441.37
1.060.67
2.470.36
1.670.44
1.140.60
44.08.6
61.64.2
54.17.7
52.36.7
69.35.5
Fig. 7. Gastro resistance of pantoprazole microparticles and agglom

erates (A to E) (n 3) Error bars indicate standard deviations. Drug
release is reported in phosphate buffer pH 7.4 after exposure of 1 h in
0.1 M HCl
92% of pantoprazole content after exposure to acid medium.

Agglomerates D prepared with 1:1 (w/w) ratio and 20.0% of
lecithin, presented 70% of pantoprazole after the acid step.
This percentage is similar to that reported for the correspon
dent pantoprazole loaded microparticle formulation prepared
in laboratory scale (12). Agglomerates B prepared with 1:2
(w/w) ratio, and 17.5% of lecithin presented the lowest
gastro resistance (51%). This formulation also showed low
agglomeration yield and incomplete drug loading. Agglomer
ates E prepared with 1:2 (w/w) ratio and 20.0% of lecithin
showed 91% of gastro resistance. This formulation showed a
gastro resistance similar to that observed for the pantopra
zole loaded microparticles before agglomerating. Agglomer
ates A and C presented very similar proles, and the amount
of drug stabilized was approximately 87%.
The gastro resistance proles were mathematically mod
eled to t mono or biexponential equations. The prole of
the pantoprazole loaded microparticles t the monoexponen
tial model (MSC=4.38; r2 =0.996), and the half life of drug
release was 155.8 min. In general, drug delivery systems
containing water soluble drugs follow the monoexponential
model (41). On the other hand, all agglomerate proles t the
biexponential model (Table IV).
The biexponential release proles have two different
release rates (burst and sustained release phases). The initial
burst was higher for agglomerates A and C, which presented
Ab parameters close to 0.4. The burst phase half lives were
16.1 and 19.2 min for agglomerates A and C, respectively.
These agglomerates contained a higher amount of mannitol
(in the form of spray dried mannitol/lecithin powders) that
acted as a disintegrant, facilitating the inux of water inside
the agglomerates and reducing the gel layer formation by the
swelling of HPMC. Those agglomerates presented sustained
phase half lives of 330.1 and 301.4 min, respectively. Other
wise, agglomerates B, D, and E showed Ab values of 0.21,
0.12, and 0.25, respectively. The Bs values were between 0.5
and 0.9 (Table IV). The amount of spray dried mannitol/
343
lecithin powders inuenced the amount of pantoprazole
released in the burst phase, as well as the release rate,
modulating the drug release.
In order to access the release mechanism, the proles
were modeled to t the Korsmeyer Peppas model. The
pantoprazole loaded microparticle prole showed n value of
0.68 (Eq. 2), demonstrating that the release mechanism is the
anomalous transport. The anomalous transport has interme
diate characteristics between the Fickian diffusion and the
swelling/controlled release (37).
Agglomerates A, B, C, and E presented n values of 0.27,
0.15, 0.25, and 0.28, respectively, indicating the Fickian diffusion
release mechanism for polydisperse systems. On the other hand,
agglomerates D presented the same release mechanism of the
pantoprazole loaded microparticles (n=0.54). In this way, the
drug release mechanism was also changed with the concentra
tion of spray dried mannitol/lecithin powders.
In order to evaluate the wettability, the contact angle
between the microparticles and water and the agglomerates and
water were determined. The contact angle between pantopra
zole loaded microparticles and water was 89.9, while the
contact angles between water and the agglomerates A, B, C,
D, and E were 76.2, 75.1, 65.5, 87.6, and 70.2, respectively.
The presence of mannitol/lecithin powders strongly inuenced
the contact angle and, as a consequence, the release mechanism
of the agglomerates. The hydrophilic powder of mannitol and
lecithin increased the wettability of the agglomerates, facilitat
ing the disintegration of the globular structure. The agglomer
ates presented faster release as the ratio between pantoprazole
loaded microparticles and mannitol/lecithin powders increased.
The high solubility and prompt disintegration of these excipients
altered the gel layer formation around the pantoprazole loaded
microparticles, facilitating the water penetration inside the
agglomerates. It was possible to maintain the release mechanism
and the release rate of the primary pantoprazole microparticles
when 1:1 (w/w) ratio was used. However, those agglomerates
were not capable of stabilizing more than 90% of pantoprazole
as required by the Pharmacopoeia (33). Furthermore, agglom
erates E containing 1:2 (w/w) ratio of mannitol/lecithin powders
(80:20) presented high gastro resistance and an intermediate
release rate (half life of release of 108.8 min).
CONCLUSIONS
The spray drying process was reproducible, and the
pantoprazole loaded microparticles were stable under acceler
ate condition within 6 months. Indeed, pantoprazole micro
particles were not hygroscopic. The agglomeration of
Table IV. Mathematical model of the agglomerate dissolution

proles and t to the biexponential equation (n 3)
Biexponential equation parameters
Agglomerates
Ab
Bs
r2
MSC
A
B
C
D
E
0.42
0.21
0.39
0.12
0.25
0.043
0.075
0.036
0.057
0.034
0.27
0.61
0.29
0.87
0.53
0.0021
0.0007
0.0023
0.0017
0.0026
0.999
0.992
0.999
0.999
0.998
5.9
3.8
5.8
6.7
5.5
Raffin et al.
344
pantoprazole loaded microparticles blended with spray dried
mannitol/lecithin powders was successfully applied to size
enlargement of micronized products that could be damaged by
granulation or compaction. The composition and quantity of the
spray dried mannitol/lecithin powders resulted to be the crucial
factors for the agglomerate quality, in terms of process yield,
drug loading, resistance, friability, and owability. Therefore,
adjusting the content of lecithin used as a binder, it was possible
to agglomerate microparticulated materials that have poor
owability. The presence of mannitol/lecithin powders strongly
inuenced the disintegration and the drug release from the
agglomerates. The agglomerates with more adequate mechan
ical and biopharmaceutical characteristics were prepared with
1:2 (w/w) ratio of pantoprazole loaded microparticles and
mannitol/lecithin powder (80:20).
13.
14.
15.
16.
17.
18.
ACKNOWLEDGEMENT
The authors are grateful for the nancial support of
CNPq/MCT, Universal 2007 CNPq and for the CAPES
fellowship. The nancial support of the Italian Ministry for
University and Research is also gratefully acknowledged. We
thank Prof. Edilson Benvenutti for the BET analysis.
19.
20.
21.
REFERENCES
1. N. S. Barakat, and A. A. E. Ahmad. Diclofenac sodium loaded
cellulose acetate butyrate: Effect of processing variables on
microparticles properties, drug release kinetics and ulcerogenic
activity. J. Microencapsul. 25(1):31 45 (2008).
2. Y. C. Huang, A. Vieira, M. K. Yeh, and C. H. Chiang. Containing
betamethasone pulmonary anti inammatory effects of chitosan
microparticles. J. Bioact. Compat. Polym. 22:30 41 (2007).
3. S. Tamilvanan and B. Sa. In vitro and in vivo evaluation of
single unit commercial conventional tablet and sustained release
capsules compared with multiple unit polystyrene microparticle
dosage forms of ibuprofen. AAPS PharmSciTech. 7(3):Article 72
(2006).
4. K. P. R. Chowdary and Y. Srinivasa Rao. Design and in vitro and
in vivo evaluation of mucoadhesive microcapsules of glipizide for
oral controlled release: A technical note. AAPS PharmSciTech. 4
(3):Article 39 (2003).
5. K. Masters. The spray drying handbook, Longman, New York,
1991.
6. R. C. R. Beck, A. R. Pohlmann, and S. S. Guterres. Nano
particle coated microparticles: preparation and characterization.
J. Microencapsul. 21:499 512 (2004).
7. A. Goula, K. Adamopoulos, and N. Kazakis. Inuence of spray
drying conditions on tomato powder properties. Dry. Technol. 22
(5):1129 1151 (2004).
8. M. I. R. Microencapsulation by spray drying. Dry. Technol.
16:1195 1236 (1998).
9. R. P. Rafn, S. S. Guterres, A. R. Pohlmann, and M. I. R.
Powder characteristics of pantoprazole delivery systems pro
duced in different spray dryer scales. Dry. Technol. 24:339 348
(2006).
10. R. P. Rafn, D. S. Jornada, M. I. R, A. R. Pohlmann, and S. S.
Guterres. Sodium Pantoprazole loaded enteric microparticles
prepared by spray drying: effect of the scale of production and
process validation. Int. J. Pharm. 324:10 18 (2006).
11. C.G. Oster, and T. Kissel. Comparative study of DNA encapsu
lation into PLGA microparticles using modied double emulsion
methods and spray drying techniques. J. Microencapsul. 22:235
244 (2005).
12. R. P. Rafn, L. M. Colom, S. E. Haas, D. S. Jornada, A. R.
Pohlmann, and S. S. Guterres. Development of HPMC and
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
Eudragit S100 blended microparticles containing sodium

pantoprazole. Pharmazie. 62:361 364 (2007).
S. Harikarnpakdee, V. Lipipun, N. Sutanthavibul, and G. C.
Ritthidej. Spray dried mucoadhesive microspheres: Preparation
and transport through nasal cell monolayer. AAPS PharmSci
Tech. 7(1):Article 12 (2006).
P. Russo, F. Buttini, F. Sonvico, R. Bettini, G. Massimo, C.
Sacchetti, P. Colombo, and P. Santi. Chimeral agglomerates of
microparticles for administration of caffeine nasal powders. J.
Drug Deliv. Sci. Technol. 14:449 454 (2004).
T. D. Reynolds, S. A. Mitchell, and K. M. Balwinsky. Investiga
tion of the effect of tablet surface area/volume on drug release
from hydroxypropylmethylcellulose controlled release matrix
tablets. Drug Dev. Ind. Pharm. 28:457 466 (2002).
E. Esposito, F. Cervellati, E. Menegatti, C. Nastruzzi, and R.
Cortesi. Spray dried Eudragit microparticles as encapsulation
devices for vitamin C. Int. J. Pharm. 242:329 334 (2002).
G. F. Palmieri, G. Bonacucina, P. Di Martino, and S. Martelli.
Gastro resistant microspheres containing ketoprofen. J. Micro
encapsul. 19:111 119 (2002).
D. B. Beten, M. Gelbcke, B. Diallo, and A. J. Moes. Interaction
between dipyridamole and Eudragit S. Int. J. Pharm. 88:31 37
(1992).
H. P. de Oliveiraa, J. J. F. Albuquerque Jr., C. Nogueirasa, and J.
Rieumonta. Physical chemistry behavior of enteric polymer in
drug release systems. Int. J. Pharm. 366:185 189 (2009).
S. Cheer, A. Prakash, D. Faulds, and H. Lamb. Pantoprazole
an update of its pharmacological properties and therapeutic use
in the management of acid related disorders. Drugs. 63:101 132
(2003).
B. K. Kim, S. J. Hwang, J. B. Park, and H. J. Park. Character
istics of felodipine located poly( caprolactone) microspheres. J.
Microencapsul. 22:193 203 (2005).
Y. J. Fu, F. L. Mi, T. B. Wong, and S. S. Shyu. Characteristic and
controlled release of anticancer drug loaded poly (D,L lactide)
microparticles prepared by spray drying technique. J. Micro
encapsul. 18:733 747 (2001).
S. Tsantilis, and S. E. Pratsinis. Soft and hard agglomerate
aerosols made at high temperatures. Langmuir. 20:5933 5939
(2004).
R. Boerejn, Z. Ning, and M. Ghadiri. Disintegration of weak
lactose agglomerates for inhalation applications. Int. J. Pharm.
172:199 209 (1998).
P. Russo, C. Sacchetti, I. Pasquali, R. Bettini, G. Massimo, P.
Colombo, and A. Rossi. Primary microparticles and agglomer
ates of morphine for nasal insufation. J. Pharm. Sci. 95:2553
2561 (2006).
R. Moreno Atanasio, and M. Ghadiri. Mechanistic analysis and
computer simulation of impact breakage of agglomerates: Effect
of surface energy. Chem. Eng. Sci. 61:2476 2481 (2006).
Stability testing guideline RE 1, Brazilian National Health
Surveillance Agency, July 29, 2005.
International Conference on Harmonisation of Technical
Requirements for Registration of Pharmaceuticals for Human
Use. Stability Testing of New Drug Substances and Products (Q
1A (R2)), 2003.
R. P. Rafn, L. M. Colom, S. S. Guterres, and A. R. Pohlmann.
Validation of analytical methodology by HPLC for quantication
and stability evaluation of sodium pantoprazole. Quim. Nova.
30:1001 1005 (2007).
S. Brunauer, P. H. Emmet, and E. Teller. Adsorption of gases in
multimolecular layers. J. Am. Chem. Soc. 60:309 319 (1938).
Y. S. L. Lee, R. Poynter, F. Podczeck, and J. M. Newton.
Development of a Dual Approach to Assess Powder Flow from
Avalanching Behavior. AAPS PharmSciTech 1(3):Article 21
(2000).
European Pharmacopoeia. 5th edn., Council of Europe, Stras
bourg, 2005.
US Pharmacopeia XXX, US Pharmacopeial Convention, Rock
ville, MD, 2007.
Farmacopia Brasileira 4a ed., Atheneu, So Paulo, 2005.
H. X. Guo, J. Heinamaki, and J. Yliruusi. Amylopectin as a
subcoating material improves the acidic resistance of enteric
coated pellets containing a freely soluble drug. Int. J. Pharm.
235:79 86 (2002).

36. R. P. Rafn, L. M. Colom, A. R. Pohlmann, and S. S. Guterres.
Preparation, characterization, and in vivo anti ulcer evaluation of
pantoprazole loaded microparticles. Eur. J. Pharm. Biopharm.
63:198 204 (2006).
37. R. W. Korsmeyer, R. Gurny, E. Doelker, P. Buri, and N. A.
Peppas. Mechanism of solute release from porous hydrophilic
polymers. Int. J. Pharm. 15:25 35 (1983).
38. Z. Wong, and E. Schaffer. Experimental study of contact line
dynamics by capillary rise and fall. In K. L. Mittal (ed.),
Contact Angle, Wettability and Adhesion, Brill, Boston, 2003,
pp. 25 37.
345
39. K. N. Atuah, E. Walter, H. P. Merkle, and H. O. Alpar.
Encapsulation of plasmid DNA in PLGA stearylamine micro
spheres: a comparison of solvent evaporation and spray drying
methods. J. Microencapsul. 20:387 399 (2003).
40. R. P. Rafn, P. Colombo, F. Sonvico, F. S. Polleto, G. Colombo,
A. Rossi, A. R. Pohlmann, and S. S. Guterres. Soft Agglomer
ates of pantoprazole gastro resistant microparticles for oral
administration and intestinal release. J. Drug Deliv. Sci. Technol.
17:407 413 (2007).
41. P. Costa, and J. M. S. Lobo. Modeling and comparison of
dissolution proles. Eur. J. Pharm. Sci. 13:123 133 (2001).

DOI: 10.1208/s12249 009 9215 4
Research Article
PolymerMagnesium Aluminum Silicate Composite Dispersions for Improved
Physical Stability of Acetaminophen Suspensions
Thaned Pongjanyakul1,3 and Satit Puttipipatkhachorn2
Received 26 August 2008; accepted 1 March 2009; published online 25 March 2009
Abstract. The aims of this study were to characterize the morphology and size of occulates and the zeta
potential and rheological properties of polymer magnesium aluminum silicate (MAS) composite
dispersions and to investigate the physical properties of acetaminophen (ACT) suspensions prepared
using the composite dispersions as a occulating/suspending agent. The polymers used were sodium
alginate (SA), sodium carboxymethylcellulose (SCMC), and methylcellulose (MC). The results showed
that SA, SCMC, and MC could induce occulation of MAS by a polymer bridging mechanism, leading to
the changes in the zeta potential of MAS and the ow properties of the polymer dispersions. The
microscopic morphology and size of the occulates was dependent on the molecular structure of the
polymer, especially ether groups on the polymer side chain. The residual MAS from the occulation
could create a three dimensional structure in the SA MAS and SCMC MAS dispersions, which brought
about not only an enhancement of viscosity and thixotropic properties but also an improvement in the
ACT occulating efciency of polymers. The use of polymer MAS dispersions provided a higher degree
of occulation and a lower redispersibility value of ACT suspensions compared with the pure polymer
dispersions. This led to a low tendency for caking of the suspensions. The SCMC MAS dispersions
provided the highest ACT occulating efciency, whereas the lowest ACT occulating efciency was
found in the MC MAS dispersions. Moreover, the added MAS did not affect ACT dissolution from the
suspensions in an acidic medium. These ndings suggest that the polymer MAS dispersions show good
potential for use as a occulating/suspending agent for improving the rheological properties and physical
stability of the suspensions.
KEY WORDS: magnesium aluminum silicate; physical stability; sodium alginate; sodium
carboxymethylcellose; suspensions.
INTRODUCTION
In pharmaceutical product development, polymers have
been widely used in coarse dispersion dosage forms, such as
suspensions, for minimizing or controlling sedimentation of
drug particles (1). They can physically stabilize suspensions
by inducing occulation of drug particles. This can prevent
the formation of hard cake sediment which is difcult to
redisperse (2). Furthermore, the rheological characteristics of
polymer dispersions are very important. A shear thinning or
pseudoplastic system is required, that is, the dispersions
should have a high viscosity at rest to retard sedimentation
of drug particles and have a low viscosity when shaking so as
to be easy to pour from a container. Therefore, the concen
tration of polymers used as occulating or suspending agents
needs to be optimized. High concentrations of polymer lead
Faculty of Pharmaceutical Sciences, Khon Kaen University, Khon

Kaen 40002, Thailand.
2
Department of Manufacturing Pharmacy, Faculty of Pharmacy,
Mahidol University, Bangkok 10400, Thailand.
3
To whom correspondence should be addressed. (e mail: thaned@
kku.ac.th)
to a good physical stability of suspensions, but may cause too

high a viscosity to pour easily. For this reason, incorporating a
material, such as clay, which molecularly interacts with
polymers, could enhance viscosity synergism and create a
gel like structure of polymer dispersion (3). This may lead to
an increase in the drug occulating efciency of polymer.
Magnesium aluminum silicate (MAS) is a mixture of
natural smectite clays, in particular montmorillonites and
saponites (4). It is composed of a three lattice layer of
octahedral alumina and two tetrahedral silica sheets (4 6);
the structure of montmorillonite clays is illustrated in Fig. 1a
(5). The surface of the silicate layer contains many silanol
groups (SiOH) which have hydrogen bonding potential with
other substances (7). The layer structures of MAS can be
separated when they are hydrated in water. Once MAS is
hydrated, the weakly positive edges are attracted to the
negatively charged faces. The face to edge attraction of these
colloidal layers creates a three dimensional colloidal structure
throughout the dispersion that exhibits a thixotropic property
(8). Therefore, MAS has been widely used as a pharmaceu
tical suspending and stabilizing agent (4). In addition, the
positively charged edges on the layers of MAS could interact
with anionic polymers such as xanthan gum (9), carbomer
(10), and sodium alginate (SA) (11,12), resulting in viscosity
346
PolymerClay Composite Dispersions for Suspensions
347
Fig. 1. Molecular structure of a montmorillonite clays (5), b mannuronic acid of SA, c SCMC, and d MC
synergism and an increase in the thixotropic properties of

polymeric dispersions. To obtain similar results, MAS was
also used in combination with sodium carboxymethylcellulose
(SCMC) for improving not only physical stability (13) but
also the ow behavior of suspensions (3).
Interaction between a polymer and clay could form a
occulate due to an electrostatic force if the polymer had a
positive charge, such as chitosan (14), and a bridging
mechanism for the polymer chain in the case of non ionic
polymers (15 17). These interactions lead to changes in the
particle size and zeta potential of dispersed phase and also
the ow behavior of composite dispersions. In a previous
study, incorporating MAS into the SA gels provided higher
viscosity and changed rheological behavior from Newtonian
to pseudoplastic with thixotropic properties. This behavior
was due to the formation of an electrostatic force between
the negative charge of the carboxyl groups of SA and the
positively charged sites at the edges of MAS, as well as
intermolecular hydrogen bonding between SA and MAS
(12). However, the occulation and zeta potential of
dispersed phase were not investigated in this report.
SA, SCMC, and methylcellulose (MC) have been used
as suspending agents in suspensions (4). SA and SCMC are
negatively charged polymers that possess carboxyl groups
on their polymer chains (Fig. 1b, c, respectively), but
SCMC has a carboxylmethyl ether group in which the
ether groups may form hydrogen bonds with the silanol
groups on the silicate layer surface of MAS. This may lead
to the different characteristics of the occulates formed.
MC, a non ionic polymer, possesses hydroxyl and methoxyl
ether groups (Fig. 1d) and also has the potential to form
occulates with MAS. However, there are no data available
concerning occulation between these polymers and MAS.
Therefore, the purpose of the present study was to
investigate the size and microscopic morphology of the
occulates, zeta potential, and rheological behavior of
composite dispersions prepared with different ratios of

MAS and these selected polymers. The composite disper
sions prepared were used as occulating/suspending agents
for the suspensions of acetaminophen (ACT) which was
selected as a model drug because an increase in ACT
particle size in suspensions could occur due to the
crystallization of ACT saturated solution (18). This led to
a hard cake after storage. Thus, the use of the composite
dispersion may provide good physical stability to ACT
suspensions. Additionally, the dissolution of ACT from the
suspensions in an acidic medium was investigated.

Materials
MAS (Veegum HV) was obtained from R.T. Vander
bilt Company, Inc. (Norwalk, CT, USA). SCMC (viscosity of
2% dispersion at 25C: 1,550 cps), MC (Methocel, 27.5
32% methoxyl content, viscosity of 2% dispersion at 20C:
3,000 5,500 cps), and SA NF17 (mannuronic acid rich type)
(19) were purchased from Nichirin Chemical Industries, Ltd.
(Japan), Fluka (Buchs, Switzerland), and Srichand United
Dispensary Co., Ltd. (Bangkok, Thailand), respectively. ACT
was obtained from Praporn Darsut Ltd., (Bangkok, Thai
land). ACT used in this study was the stable monoclinic form
in which the melting peak temperature measured using
differential scanning calorimeter was found to be 169C
(20,21). All other reagents used were of analytical grade
and used as received.
Preparation of Dispersions
SCMC (0.5 g) or SA (1 g) was dispersed in distilled water
(30 ml) by agitation for 2 h, whereas MC (0.5 g) was
348
Pongjanyakul and Puttipipatkhachorn
dispersed in hot water (10 ml) and cold water (20 ml) was
then added and stirred for 1 h to get a clear dispersion. MAS
(0.1, 0.5, or 1 g) was prehydrated with hot water (20 ml) for
15 min prior to adding into the polymer dispersions. The
composite dispersions were obtained after adjusting the nal
volume to 100 ml with distilled water and were then stirred
for 30 min and allowed to fully hydrate at room temperature
(25 28C) overnight prior to use.
coefcient, were calculated using exponential formulas as

follows (2):
F N G
LogG N log F
log
where G, F, N, and are the shear rate, the shear stress, the
exponential constant that denes a type of ow, and the
viscosity coefcient, respectively.
Characterization of Composite Dispersions

Microscopic Morphology Studies
Preparation of ACT Suspensions

The microscopic morphology of MAS particles and the
dispersed phase of composite dispersions were investigated
using an inverted microscope (Eclipse TS100, Nikon, Japan)
and viewed using a digital camera (Coolpix 4500, Nikon,
Japan).
Particle Size Determination
The particle size of MAS and the dispersed phase of
composite dispersions were measured using a laser diffraction
particle size analyzer (Mastersizer2000 Model Hydro2000SM,
Malvern Instrument Ltd., UK). The samples were dispersed
in 70 ml of distilled water in a small volume sample dispersion
unit in which the stirrer revolved at a rate of 50 Hz for 30 s
prior to counting. The size frequency distributions were
plotted and particle sizes, measured as volume weighted
mean diameter, were reported. Moreover, D10%, D50%, and
D90%, which were the volume number diameters where the
given percentage of the particles was smaller than that size,
were used to calculate the polydispersity index (PI), indicat
ing the size distribution of the particles. The PI value can be
expressed as (22):
D90% D10%
PI
:
D50%
pH and Zeta Potential Measurement

The pH of all dispersions was measured using a pH
meter (Ion Analyzer 250, Corning, USA). Moreover, the
MAS and polymer MAS dispersions were diluted to obtain
appropriate concentrations, and the zeta potentials of all
samples were measured using a laser Doppler electropho
resis analyzer (Zetasizer model ZEN 2600, Malvern Instru
ment Ltd.). The temperature of the samples was controlled
at 25C.
Rheological Studies
The rheological properties of the dispersions were
studied using a Brookeld digital rheometer (Model DV
III, Brookeld Engineering Laboratories, Inc., Middleboro,
MA, USA). The temperature of all samples was maintained
at 321C. A rheogram of the samples was plotted using
shear rate and shear stress at various revolution rates of
spindle (SC4 34). The rheological parameters of the dis
persions, such as the exponential constant and viscosity
A 2.5 g portion of ACT was placed into a mortar, wetted

using 0.1% (w/v) polysorbate 80 (2 ml), and levigated until a
smooth paste was produced. Next, either polymer or poly
mer MAS composite dispersions were added as a vehicle to
the ACT smooth paste and mixed well. The resultant mixture
was then transferred to a 50 ml graduated cylinder. The
mortar was rinsed several times with a small portion of
vehicle and the mixture obtained was also poured into the
cylinder. Then, the suspension was adjusted to a nal volume
of 50 ml and mixed well prior to testing. The vehicle of the
control suspension in this study was distilled water. In
addition, MAS dispersions in concentrations of 0.1%, 0.5%,
and 1% (w/v) were also used as a vehicle for ACT
suspensions to investigate the occulating efciency of
MAS. The suspensions were prepared as described above.
Characterization of Suspensions
Determination of Sedimentation Volume and Degree
of Flocculation
The ACT suspension was poured into a 10 ml graduated
cylinder and allowed to settle. The volume of the sediment
was recorded at predetermined times over a period of 14 days.
The sedimentation volume and the degree of occulation at
14 days were determined. The sedimentation volume is the
ratio between the volume of sediment at a given time and the
total volume of the suspension, whereas the degree of
occulation is the sedimentation volume of the ACT suspen
sions using polymer or polymer MAS composite dispersion
divided by the sedimentation volume of the control suspen
sion (2,23,24).
Redispersibility Studies
After the suspension in the 10 ml cylinder settled for
14 days, the cylinder was inverted to redisperse the sediment.
The number of inversions required to resuspend the sediment
of the suspension is the redispersibility value (24).
In vitro Dissolution Studies
A USP dissolution apparatus II (Hanson Research,
Northridge, USA) was used to characterize the dissolution
of ACT suspensions. The paddles were rotated at 50 rpm at
37.00.5C. Nine hundred milliliters of 0.1 N HCl was used as
a dissolution medium. Five milliliters of the suspension was

placed into a 10 ml graduated cylinder and weighed. The
suspension was then directly poured into the dissolution
medium and the agitation was started (25). The cylinder was
reweighed and the difference from the starting weight recorded
as the weight of the tested suspension. The corresponding
volume was computed using the density of the suspension,
which was determined using a 25 ml pycnometer. Samples
(5 ml) were collected without replacement of fresh medium at
different time intervals. Then, the collected samples were
ltered with a 0.45 m cellulose acetate membrane. The
amount of ACT dissolved was analyzed spectrophotometrically
at 245 nm (Shimadzu UV1201, Japan).
Microscopic Studies of PolymerMAS Dispersions
MAS dispersions were opaque in nature. Incorporating
MAS into MC, SA, and SCMC dispersions caused a change
from clear viscous solutions to dispersions with different
turbidities. All composite dispersions showed good homoge
neity after fresh preparation. However, different sedimenta
tion behaviors of these polymer MAS dispersions were
observed after settling for 2 days (Fig. 2). The supernatant
of the MAS dispersion was not clear, suggesting a deoccu
lation system (2). Additionally, the MC MAS and SCMC
MAS dispersions led to a clear supernatant after settling for
2 days (Fig. 2b, d, respectively). In contrast, no sedimentation
of the SA MAS dispersion was observed (Fig. 2c). A
photomicrograph of the 0.5% MAS dispersion showed
separated particles of hydrated MAS (Fig. 3a), whereas an
irregular shape of aggregated particles in the MC MAS and
SCMC MAS dispersions were readily observed (Fig. 3b, c,
respectively). Interestingly, a unique arrangement of the
particles was found in the SA MAS dispersions (Fig. 3d).
These results reveal occulate formation between polymers
and MAS. Moreover, the remarkably larger size of the MC
MAS and SCMC MAS occulates relative to the SA MAS
occulates could explain the higher extent in sedimentation
of the MC MAS and SCMC MAS occulates. It is likely that
Fig. 2. Photographs of 0.5% MAS dispersions a without polymer and

b with 0.5% MC, c 1% SA, and d 0.5% SCMC after settling for
2 days
349
the occulation between these polymers and MAS may affect
the characteristics of the composite dispersions, such as zeta
potential and ow behavior.
Characterization of PolymerMAS Dispersions
The addition of polymer into MAS dispersions led to
changes in the characteristics of the size frequency distribu
tion curves, mean particle size, and PI values (Fig. 4 and
Table I). As MAS concentration increased, the size frequen
cy distribution curves of the MC MAS occulates were
shifted to larger particle size (Fig. 4a). The MC dispersion
with 0.1% MAS provided a larger occulate size than that
with 0.5% MAS (Table I). However, the largest occulate
size and the lowest PI value were found when 1% MAS was
used. In the case of the SCMC MAS dispersion, incorpora
tion of MAS provided bimodal size frequency distribution
curves at all concentrations of MAS (Fig. 4b). The peak at the
smaller size was in the same range as that of the MAS
dispersion. With increasing MAS concentrations, the frequen
cy of the peak at the smaller size was increased, whereas that
of the peak at the larger size was reduced. This led to a
decrease in particle size with increasing PI values of the
SCMC MAS occulates. Furthermore, incorporating MAS
did not obviously change the size frequency distribution of
the SA MAS occulates (Fig. 4c), and the SA MAS
occulate size tended to decrease with increasing MAS
(Table I).
The pH of the polymer dispersions was over the range
6.7 7.2 (Table I), resulting in an ionization of carboxyl groups
of SCMC and SA. The zeta potentials of the SCMC and SA
dispersions were approximately 49.3 and 81.8 mV, respec
tively, whereas that of MC dispersion could not be deter
mined due to a very low count rate, indicative of non ionic
polymer of MC. The MAS dispersion possessed a basic pH,
and its zeta potential was found to be 44.7 mV (Table I).
The addition of MAS into the polymer dispersions caused an
obvious increase in the pH because of a mild alkalinity of OH
groups associated with Si, Mg, and Al in MAS (11). The basic
pH of the composite dispersions suggested that SCMC, SA,
and MAS also presented negatively charged molecules. The
incorporation of polymer to the MAS dispersion led to
changes in zeta potential (Table I). The zeta potential of the
MC MAS dispersions was remarkably lower than that of the
MAS dispersion, although MC was a non ionic polymer.
However, the MC MAS occulates also presented a negative
charge with a small value of zeta potential. The SCMC MAS
dispersions showed an obviously higher zeta potential with a
negative charge than both SCMC and MAS dispersions. On
the other hand, incorporating MAS did not obviously affect
the zeta potential of the SA dispersion. The zeta potential
of the SA MAS dispersions was over the range 78.0 to
84.1 mV, values which were close to the zeta potential of the
SA dispersion. Furthermore, it can be seen that the change of
the zeta potential of these polymers was independent upon
the MAS concentration added. However, the change in zeta
potential indicated an interaction between these polymers
and MAS, which was related to the occulation of these
dispersion systems.
The occulation between polymers and MAS is a result
of the molecular interaction between both materials, although
350
Fig. 3. Photomicrographs of MAS particles in 0.5% MAS dispersion a without polymer and b with 0.5% MC, c 0.5% SCMC, d and 1% SA
SA and SCMC have a negative charge that is similar to MAS

and MC is a non ionic polymer. This result suggests that MAS
and these polymers could mainly interact via non ionic
electrostatic interaction, such as intermolecular hydrogen
bonding, which would lead to occulation by a polymer
bridging mechanism (15). In fact, the surface of the silicate
layers of MAS contain many silanol groups (SiOH), which
have a high potential to form hydrogen bonds with the
oxygen atoms of carboxyl, hydroxyl, and ether groups of

these polymers. In the previous study, intermolecular hydro
gen bonding between silanol groups at the surface of MAS
and hydroxyl or carboxyl groups of SA was demonstrated.
Moreover, the negative charges of the carboxyl groups of SA
have an electrostatic interaction with the positively charged
sites at the edges of MAS (12). This introduces numerous
points of contact to create a denser network, resulting in the
Fig. 4. Size frequency distributions of occulates in a 0.5% MC, b 0.5% SCMC, and c 1% SA dispersions containing different concentrations
of MAS
351
Table I. Characteristics of Polymer MAS Composite Dispersions
Component
0.5% (w/v) MAS
0.5% (w/v) MC
+0.1% (w/v) MAS
+0.5% (w/v) MAS
+1% (w/v) MAS
0.5% (w/v) SCMC
+0.1% (w/v) MAS
+0.5% (w/v) MAS
+1% (w/v) MAS
1% (w/v) SA
+0.1% (w/v) MAS
+0.5% (w/v) MAS
+1% (w/v) MAS
Particle sizea (m)
Polydispersity indexa
pHa
Zeta potentialb (mV)
5.50.2
2.330.09
44.72.2
82.40.8
51.91.5
217.415.6
2.200.04
2.370.20
1.750.01
251.517.4
169.715.0
137.214.9
1.550.11
2.470.13
3.530.56
16.31.3
7.11.3
5.90.1
5.540.60
2.470.04
2.280.02
8.360.06
6.850.10
8.580.11
9.310.06
9.230.03
7.200.05
7.570.03
9.230.04
9.230.02
6.710.02
7.370.03
9.130.01
9.120.03
4.340.94
2.830.51
4.361.84
49.314.3
84.39.4
83.15.6
79.02.8
81.84.5
84.12.9
78.92.1
78.11.5
Na
Viscosity coefcienta
[(dyne cm 2)N s]
0.760.09
0.790.12
1.220.04
1.960.01
0.960.08
0.900.07
1.100.05
2.260.01
0.700.04
0.820.07
0.950.02
1.470.11
0.080.01
0.090.01
0.440.03
8.511.52
0.240.05
0.240.02
1.070.16
227.624.3
0.180.01
0.250.04
0.480.01
7.772.96
could not be measured,

count rate is too low
Data are the meanSD of three determinations
b
Data are the meanSD of six determinations
a
unique structure of the occulates observed in the SA MAS

dispersions (Fig. 3d). However, the SA MAS occulates
were not stable and were dissociated by water dilution and
gentle mechanical force during particle size measurement.
Thus, the nal measured size was in the same range as that of
MAS. SCMC has carboxyl groups that are similar to SA, but
it could form larger sized occulates. This can explain why
SCMC or SA could interact with MAS via similar interaction
mechanisms. However, the aliphatic ether groups on the
molecular chain of SCMC could possibly create a hydrogen
bond with MAS rather than the ether groups of SA. This
resulted to the different characteristics of occulates. The
hydrogen bonding mechanism could also be involved in the
formation of the MC MAS occulates. However, increasing
the concentration of MAS led to a decrease in the amount of
the SCMC MAS occulates, but it increased the MAS
particles that were not involved in occulation (residual
MAS), suggesting that the formation of the SCMC MAS
occulates was limited in amount and particle size. On the
other hand, an almost complete occulation between MC and
MAS occurred, leading to a larger occulate size with
increasing concentrations of MAS. Moreover, the MC MAS
occulates gave small zeta potentials with negative charges
that were obviously less than the zeta potential of MAS,
indicating that the occulates were covered by non ionic MC
molecules. This phenomenon was also found in the SCMC
MAS occulates, which presented a negative charge, but the
zeta potentials of these occulates were higher than those of
SCMC and MAS. This suggests that the coverage of SCMC
caused a denser negative charge on the surface of these
occulates.
The effect of MAS addition on the ow curve of the
polymer dispersions is shown in Fig. 5. Nonlinear ow curves
of the composite dispersions were observed when adding
higher concentrations of MAS. In contrast, the ow curves of
pure polymer dispersions showed a linear relationship
between shear rate and shear stress. The rheological param
eters calculated using the shear rate and shear stress of the
up curve are presented in Table I. The exponential constant
(N value) that denes the type of ow of the dispersions was
approximately unity, indicating Newtonian ow. Incorpora

tion of MAS into the polymer dispersions caused an increase
in the N value, suggesting that the ow behavior shifted to
pseudoplastic ow. Furthermore, the viscosity coefcient of
the polymer dispersion increased with an increasing concen
tration of MAS (Table I), suggesting viscosity enhancing
properties of MAS. The up and down curves of the
composite dispersions were coincident when 0.1% and 0.5%
MAS were used. A hysteresis loop was found between up
and down curves of the ow curves when using 1% MAS.
The down curve moved to the left of the up curve when the
MC dispersion contained 1% MAS (Fig. 5a). This phenom
enon is called negative thixotropy or antithixotropy (2). On
the other hand, the down curve of the SCMC and SA
dispersions with 1%MAS fell to the right of the up curve
(Fig. 5b, c, respectively), indicative of a thixotropic property.
The formation of hydrogen bonds between MAS and SA
to create numerous points of contact could produce a loose
three dimensional structure throughout the composite disper
sion that is a gel like structure. The higher the concentration
of MAS, the greater the number of contact points in the
dispersion. These interactions could result in an enhancement
of the viscosity and thixotropic properties of the composite
dispersions. The SCMC MAS dispersions that contained
the occulates and the residual MAS demonstrated that
increasing the amount of MAS led to higher viscosities and
increased thixotropic properties of the composite disper
sions. This suggests that the residual dispersed MAS could
possibly act as a cross linker between SCMC chains to
create numerous points of contact. Thus, the SA MAS and
SCMC MAS dispersions underwent a gel to sol transfor
mation and exhibited shear thinning when shear stress was
applied. The structure started to reform slowly to the
original state after the shear stress was removed. In
contrast, the MC MAS dispersions showed complete oc
culation, the largest occulates were formed when 1%
MAS was used, and the ow curve of this dispersion was
pseudoplastic with antithixotropy. It is possible that antith
ixotropy resulted from an increased collision frequency of
occulates in dispersion, leading to an increase in interpar
352
Fig. 5. Flow curves of a 0.5% MC, b 0.5% SCMC, and c 1% SA dispersions containing different concentrations of MAS. Closed symbols
represent the up curve; open symbols represent the down curve. Each point is the meanSD of three determinations
ticle bonding with time. This bonding ultimately led to the

formation of larger occulates, and a gel like structure was
created. At rest, the larger occulates break up and
gradually return to their original state (2).
Physical Properties of ACT Suspensions
The polymer MAS dispersions were used as occulating/
suspending agents in ACT suspensions. The sedimentation
volumes of ACT suspensions prepared using polymer MAS
dispersions at different times are presented in Fig. 6. The
sedimentation volume of the suspensions using MC and MC
MAS dispersions decreased rapidly in a similar manner to the
control suspension (Fig. 6a). This result was also observed in
the suspensions using SCMC dispersions and SCMC disper
sions with 0.1% and 0.5% MAS (Fig. 6b). This is likely to be
due to the fast sedimentation of ACT particles and polymer
MAS occulates. However, the sedimentation volume at 14
days increased with increasing concentrations of MAS, and
the SCMC MAS dispersions gave higher sedimentation
volumes than the MC MAS dispersions at all concentrations
of MAS. Moreover, it is interesting that the suspensions using
the SCMC 1% MAS dispersion had the highest sedimenta
tion volume, which was close to unity, over the 14 days of the
test. The suspensions using SA and SA MAS dispersions

showed a slow decrease in sedimentation volume with
increasing concentrations of MAS. Moreover, the higher
the concentration of MAS added into the SA dispersion, the
greater the sedimentation volume obtained at 14 days. The
degree of occulation and the redispersibility value at 14 days
of all suspensions are presented in Fig. 7. The degree of
occulation of the suspensions increased with increasing
concentrations of MAS in all polymer and polymer MAS
dispersions used (Fig. 7a). The SCMC MAS dispersions
showed the highest degree of occulation, whereas the lowest
degree of occulation was obtained from the MC MAS
dispersion. When the MAS dispersion was used as a
occulating agent in the ACT suspensions, the degrees of
occulation of the suspensions using 0.1%, 0.5%, and 1%
MAS dispersions were found to be 1, 2.8, and 6.1, respec
tively. The SA MAS and SCMC MAS dispersions yielded a
remarkably higher degree of occulation than the MAS
dispersions, while the results of the MC MAS dispersions
were similar to those of the MAS dispersion when the same
concentration of MAS was compared. The other parameter
used to evaluate the suspensions was the redispersibility
value. The suspensions with the MC MAS dispersions
demonstrated no differences in redispersibility value when
Fig. 6. Sedimentation volume of ACT suspensions prepared using a 0.5% MC, b 0.5% SCMC, and c 1% SA dispersions containing different
concentrations of MAS as suspending agents. Each point is the meanSD, n 3
353
Fig. 7. Degree of a occulation and b redispersibility value of ACT suspensions prepared using different
polymer MAS dispersions as suspending agents. Each value is the meanSD, n 3
compared to that using MC dispersion, even though 1% MAS

was added (Fig. 7b). On the other hand, the redispersibility
value of the suspensions using the SA dispersion decreased
appreciably with increasing concentrations of MAS, although
this suspension had a tendency to cake, which was also seen
in the control suspension (redispersibility value >100).
Moreover, the SCMC 1% MAS dispersion gave the lowest
redispersibility value, which is likely due to the very low
sedimentation volume of this suspension.
The ACT suspension using the SCMC dispersion gave a
higher degree of occulation and a lower redispersibility
value than that using the MC dispersion, and the lowest
degree of occulation and the highest caking tendency of the
suspensions were obtained when using the SA dispersion
(Fig. 7). These results show that the SCMC dispersion
provides the highest efciency in inducing occulation of
ACT particles; the occulation of ACT particles is formed by
a polymer bridging mechanism via hydrogen bonding (26).
Using the MC MAS dispersion, the degree of occula
tion of suspensions increased with increasing concentrations
of MAS, but increasing the degree of occulation of the MC
MAS dispersion had less of an effect than the SA MAS and
SCMC MAS dispersions. Moreover, the redispersibility value
of the suspensions using the MC MAS dispersions was not
noticeably reduced. This result suggests that the large size of
the MC MAS occulates in the dispersions could induce a
fast sedimentation of ACT particles, and a close packed
arrangement of ACT particle sediment and MC MAS
occulates occurred. However, the close packed sediment
did not cause a hard cake because it was redispersible after
shaking. The SA MAS dispersion led to an improvement not
only in the degree of occulation but also in the redispersi
bility values of the suspensions. These data indicate that the
formation of a three dimensional structure of SA and MAS
and the occulating efciency of MAS brought about a
loosely packed arrangement of ACT particles, leading to a
remarkable reduction in caking tendency and an increased
ability to redisperse. The three dimensional structure formed
by the SCMC MAS occulates and the residual MAS could
improve the degree of occulation of the suspensions but
increase the redispersibility value when using MAS over the
range of 0.1 0.5%. This was due to the higher viscosity of the
SCMC MAS dispersions when compared to the other

composite dispersions. In addition, the highest degree of
occulation was obtained when using the SCMC 1% MAS
dispersion because of the higher viscosity and the denser
three dimensional structure which could retard the sedimen
tation of ACT particles. This also led to the lowest
redispersibility value.
In Vitro ACT Dissolution Studies
The dissolution proles of the ACT suspensions are
presented in Fig. 8. The ACT suspensions using SCMC and
MC dispersions led to more than 80% ACT dissolved at
5 min, which was similar to the control suspension. Addition
ally, the suspensions using SCMC 1% MAS and MC 1%
MAS dispersions showed similar results to those using SCMC
and MC dispersions. This nding indicated that SCMC, MC,
and their composite dispersions can be rapidly dispersed in an
acidic medium, which does not affect the dissolution process
of ACT particles. On the other hand, SA and SA 1% MAS
dispersions could retard the ACT dissolution, for which the
Fig. 8. Dissolution proles of ACT suspensions prepared using

different polymer MAS dispersions as suspending agents. Each point
is the meanSD, n 3
354
time to 80% ACT dissolved was approximately 30 min. The
SA dispersion could possibly form an insoluble gel of alginic
acid in an acidic medium (27), resulting in a slower
dissolution of ACT particles embedded in the alginic gel.
The results indicated that the polymer MAS dispersions did
not affect the dissolution of ACT from the suspensions when
compared with the pure polymer dispersions.
CONCLUSION
SA, SCMC, and MC could form occulates with MAS in
dispersions via a polymer bridging mechanism in which the
occulates obtained possessed different characteristics. The
residual MAS from the occulation could create a three
dimensional structure in the SA MAS and SCMC MAS
dispersions, which led to not only an enhancement of viscosity
and thixotropic properties but also an improvement in the
ACT occulating efciencies of polymers. Moreover, the use
of the polymer MAS dispersions provided a higher degree of
occulation and a lower redispersibility value of ACT
suspensions when compared with pure polymer dispersions.
The SCMC MAS dispersions provided the highest ACT
occulating efciency, whereas the lowest ACT occulating
efciency was detected in the MC MAS dispersions. Addi
tionally, the composite dispersions did not affect the dissolu
tion of ACT from the suspensions in an acidic medium. This
study suggests that polymer MAS dispersions demonstrate
good feasibility for use as occulating/suspending agents for
improving the ow behaviors and physical stabilities of
suspensions.
ACKNOWLEDGMENTS
The authors wish to acknowledge the Commission
on Higher Education, Ministry of Education (Bangkok,
Thailand) and the Thailand Research Fund (Bangkok,
Thailand) for research funding (grant no. RMU4980022).
We also thank the Center for Research and Development
of Herbal Health Products and the Faculty of Pharmaceutical
Sciences, Khon Kaen University (Khon Kaen, Thailand) for
technical facilities.
REFERENCES
1. Zatz JL, Berry JJ, Alderman DA. Viscosity imparting agents in
diperse systems. In: Lieberman HA, Rieger MM, Banker GS,
editors. Pharmaceutical dosage forms: disperse systems, vol. 19.
New York: Marcel Dekker; 1989. p. 171 203.
2. Martin A. Physical pharmacy. 4th ed. Philadelphia: Lea &
Febiger; 1993.
3. Viseras C, Aguzzi C, Cerezo P, Lopez Galindo A. Uses of clay
minerals in semisolid health care and therapeutic products. Appl
Clay Sci. 2007;36:37 50.
4. Kibbe HA. Handbook of pharmaceutical excipients. 3rd ed.
Washington: American Pharmaceutical Association; 2000.

5. Grim RE. Clay mineralogy. New York: McGraw Hill Book
Company; 1953. p. 43 83.
6. Alexandre M, Dubois P. Polymer layered silicate nanocompo
sites: preparation, properties and uses of a new class of materials.
Mater Sci Eng. 2000;28:1 63.
7. Gupta MK, Vanwert A, Bogner RH. Formation of physical
stable amorphous drugs by milling with Neusilin. J Pharm Sci.
2003;92:536 51.
8. Zatz JL, Kushla GP. Gels. In: Lieberman HA, Rieger MM,
Banker GS, editors. Pharmaceutical dosage forms: disperse
systems, vol. 19. New York: Marcel Dekker; 1989. p. 495 510.
9. Ciullo PA. Rheological properties of magnesium aluminum
silicate/xanthan gum dispersions. J Soc Cosmet Chem. 1981;32:
275 85.
10. Ciullo PA, Braun DB. Clay/carbomer mixtures enhance emul
sion stability. Cosmet Toilet. 1991;106:89 95.
11. Pongjanyakul T, Priprem A, Puttipipatkhachorn S. Inuence of
magnesium aluminium silicate on rheological, release and
permeation characteristics of diclofenac sodium aqueous gels
in vitro. J Pharm Pharmacol. 2005;57:429 34.
12. Pongjanyakul T, Puttipipatkhachorn S. Sodium alginate magne
sium aluminum silicate composite gels: characterization of ow be
havior, microviscosity, and drug diffusivity. AAPS PharmSciTech.
2007;8:E158 65.
13. Attama AA, Adikwu MU, Esimone CO. Sedimentation studies
on chalk suspensions containing blends of Veegum and detarium
gum as suspending agents. Boll Chim Farm. 1999;138:521 5.
14. Khunawattanakul W, Puttipipatkhachorn S, Rades T, Pongjanyakul T.
Chitosan magnesium aluminum silicate composite dispersions:
characterization of rheology, occulate size and zeta potential. Int J
Pharm. 2008;351:227 35.
15. Yoon SY, Deng Y. Flocculation and reocculation of clay
suspension by different polymer systems under turbulent con
ditions. J Colloid Interface Sci. 2004;278:139 45.
16. Addai Mensak J. Enhanced occulation and dewatering of clay
mineral dispersions. Powder Technol. 2007;179:73 8.
17. Kamibayashi M, Ogura H, Otsubo Y. Shear thickening ow of
nanoparticle suspension occulated by polymer bridging. J
Colloid Interface Sci. 2008;321:294 301.
18. Ziller KH, Rupprecht H. Control of crystal growth in drug sus
pensions. 1. Design of a control unit and application to acetamin
ophen suspensions. Drug Dev Ind Pharm. 1988;14:2341 70.
19. US Pharmacopeia 29, Asian Edition. United States Pharmaco
peia Convention Inc., Rockville, 2006. p. 3420.
20. Di Martino P, Guyot Hermann A M, Conant P, Drache M,
Guyot J C. A new pure paracetamol for direct compression: the
orthorhombic form. Int J Pharm. 1996;128:1 8.
21. Ledru J, Imrie CT, Pulham CR, Colin R, Hutchinson JM. High
pressure differential scanning calorimetry investigations on the
pressure dependence of the melting of paracetamol polymorphs I
and II. J Pharm Sci. 2007;96:2784 94.
22. Torrado JJ, Illum L, Davis SS. Particle size and size distribution
of albumin microspheres produced by heat and chemical
stabilization. Int J Pharm. 1989;51:85 93.
23. Tempio JS, Zatz JL. Flocculation effect of xanthan gum in
pharmaceutical suspensions. J Pharm Sci. 1980;69:1209 14.
24. Junyaprasert VB, Manwiwattanakul G. Release prole compar
ison and stability of diltiazem resin microcapsules in sustained
release suspensions. Int J Pharm. 2008;352:81 91.
25. Ogger KE, Noory C, Gabay J, Shah VP, Skelly JP. Dissolution
proles of resin based oral suspensions. Pharm Technol.
1991;9:84 91.
26. Wen H, Morris KR, Park K. Hydrogen bonding interactions
between adsorbed polymer molecules and crystal surface of
acetaminophen. J Colloid Interface Sci. 2005;290:325 35.
27. Draget KI. Alginates. In: Philips GO, Williams PA, editors.
Handbook of hydrocolloids. Cambridge: Woodhead; 2000.
p. 379 95.

DOI: 10.1208/s12249 009 9211 8
Research Article
Mechanistic Evaluation of the Effect of Sintering on Compritol 888 ATO
Matrices
Monica Rao,1,2 Anuradha Ranpise,1 Sameer Borate,1 and Kaushik Thanki1
Received 2 August 2008; accepted 31 January 2009; published online 31 March 2009
Abstract. The present research studied the effect of sintering technique in the development of a
controlled release formulation for ketorolac tromethamine. The method consisted of mixing drug and
wax powder (Compritol 888 ATO) along with lactose as diluent and talc as lubricant followed by direct
compression at room temperature. The compressed uffy matrices were kept at 80C for 1, 2, and 3 h for
sintering. The sintered tablets were characterized by their physical parameters and in vitro dissolution
prole. The sintering time markedly affected the drug release properties of Compritol 888 ATO
matrices. It is notable that the release rate of ketorolac tromethamine from matrices was inversely related
to the time of sintering. This may be due to the increase in the extent and rmness of sintering which
further compacts the mass so that drug release is affected. Contact angle measurement and scanning
electron microscopy analysis indicated that heat treatment caused the wax to melt and redistribute. This
redistributed wax formed a network like structure in which the drug along with lactose is entrapped. This
particular formed matrix is responsible for retarding the drug release. Fourier transform infrared
spectroscopy results did not show any drug wax interaction due to sintering. Differential scanning
calorimetric and powder X ray diffraction studies ruled out the occurrence of solid solution and
polymorphic changes of the drug. Drug release from the wax tablets with or without sintering was best
described by the Higuchi equation.
KEY WORDS: controlled release; scanning electron microscopy (SEM); sintering; wax.
INTRODUCTION
Controlled release drug delivery is one of the widely
used technologies in dosage form design, and intensive
research has been undertaken in achieving better drug
product effectiveness reliability and safety. Wax has been
used as matrix in pharmaceutical controlled release dosage
forms since decades (1). Waxes provide several advantages
that include good stability at varying pH and moisture levels
and effective retardation of highly water soluble drug from
the matrix (2).
Sintering is dened as the bonding of adjacent particle
surfaces in a mass of powder or in a compact by application of
heat (3). This concept in pharmaceutical science is relatively
recent, but research interests relating to this process have
been growing. According to studies carried out, controlled
release polymeric systems of rifampacin were prepared by
mixing the drug and ethylene vinyl acetate copolymer, which
were then compressed at room temperature. These matrices
were exposed at 60, 70, and 80 for 1.5, 3, and 4.5 h for
sintering. The sintering time markedly affected the drug
release properties of the ethylene vinyl acetate copolymer
Department of Pharmaceutics, AISSMS College of Pharmacy,

Kennedy Road, Near RTO, Pune 01, Maharastra, India.
2
To whom correspondence should be addressed. (e mail: moni
carp 6@hotmail.com)
matrices. The percent release decreased as the sintering

temperature was increased for all formulation (4,5). The
release followed a diffusive mechanism with rst order
release kinetics. A similar effect was seen when rifampicin
was replaced with theophylline (6) and when ethylene vinyl
acetate copolymer was replaced with EudragitRL100 matri
ces (7). In another approach, thermal treatment of tablets
containing Eudragit RS or RL in their structure above the
glass transition temperature of the respective polymer was
found to decrease drug release (8).
Ketorolac tromethamine is a non steroidal anti inammatory
drug usually given in the treatment of rheumatoid arthritis,
osteoarthritis, and ankylosing spondylitis. The dose of conven
tional tablet is 10 mg three times a day. Due to high water
solubility and short plasma half life, development of oral
controlled release formulation of this drug is highly desirable
(9,10).
Compritol 888 ATO composed of glyceryl behnate (an
atomized mixture of mono , di , and tribehenate of glycerol)
has been used as retard material for sustained release dosage
forms. It has low fusion point and hydrophilic lipophilic
balance value of 2. This glyceride mixture is known to exhibit
a complex polymorphism depending on many parameters
such as crystallization rate or temperature of storage (11).
The crystalline lattice of Compritol 888 ATO is composed of
very small amounts of the unstable alpha polymorphic form
characteristic of triacylglycerols, which disappear after ther
mal stress (12).
355
356
Rao, Ranpise, Borate and Thanki
The current study aimed to investigate the effect of

sintering on the release of ketorolac tromethamine from
Compritol 888 ATO matrices and also investigate the changes
in the properties of the drug and/or wax during the sintering
process.
Gift sample of ketorolac tromethamine USP was provid
ed by Ranbaxy Lab Ltd, Goa. India. Compritol was
provided by Gattefosse, France while lactose and talc by
Loba Chemie Ltd. Mumbai. All chemicals were used as
received.
Preparation of Compritol 888 ATO Matrix Tablets
A formulation containing ketorolac tromethamine and
Compritol 888 ATO in a ratio of 1:1 was prepared as per
Table I. The components of each formulation were mixed
together by geometric mixing for a period of 10 min. Tablets
of 150 mg weight were directly compressed using 6 mm
diameter punches on a rotary tablet compression machine
(Mini Press II MT, Rimek). The prepared tablets were
assayed by UV spectrophotometric method at 322 nm.
Sintering of Matrix Tablets
The prepared tablets were then subjected to thermal
treatment by placing on aluminum foil and subjecting to
sintering (3,8,13) at 80C for 1, 2, and 3 h in hot air oven
(Labhosp, Mumbai).
In Vitro Release Studies
Drug release was evaluated by conventional in vitro
dissolution testing. The dissolution test for wax matrix tablet
was performed in triplicate using dissolution test apparatus II
(DA 6D Veego) at 100 rpm in 900 ml distilled water at 37C.
Aliquots of 5 ml were periodically withdrawn and the sample
volume replaced with an equal volume of fresh dissolution
medium. The samples were ltered through Whatman lter
paper and 1 ml of sample was made up to 10 ml with distilled
water and solutions analyzed at 322 nm by UV spectropho
tometer (JASCO, V 530, Japan). Cumulative percentage drug
release was calculated using PCP Disso v2.08 Software
(Poona College of Pharmacy, Pune).
Scanning Electron Microscopy
Scanning electron microscopy (SEM) photomicrographs
were taken by scanning electron microscope (JSM 6360) for
Fig. 1. Effect of sintering on the release of ketorolac tromethamine

from Compritol 888 ATO matrices
studying surface morphology of sintered matrix tablet before

and after sintering. Each sample was mounted to an
aluminum stub using double sided adhesive tape and then
coated with gold palladium alloy using JEOL/EO ne coat
sputter. The surfaces of the tablet were coated with platinum
under an argon atmosphere. The samples were then exam
ined with 200 and 1,000 magnication using a scanning
electron microscope (JEOL JSM 6360A).
Wettability (13)
The contact angle between puried water and tablet
surfaces was determined by placing 10 l of water on the
surface of the tablet using micropipette. Photographs of the
drop in contact with water were taken. Amaranth red was
added to the water to ensure proper visibility of the drop.
Fourier Transform Infrared Spectroscopy
Fourier transform infrared spectroscopy (FTIR) spectra
(ranging 400 4,000 cm 1 ) of ketorolac tromethamine,
Compritol 888 ATO, and a physical mixture of ketorolac
tromethamine and Compritol 888 ATO before and after
heat treatment were investigated using (460plus, Jasco) using
the KBr disk method.
DSC Studies
Table I. Formulation Composition of Wax Matrix Tablet (150 mg)

Sr. no.
Ingredients
Quantity
1
2
3
4
Ketorolac tromethamine USP

Compritol ATO888
Lactose
Talc
40 mg
40 mg
q.s.
7.5 mg
The differential scanning calorimetric (DSC) thermo

grams of ketorolac tromethamine, Compritol 888 ATO, and
a physical mixture of ketorolac tromethamine and Compri
tol 888 ATO before and after heat treatment were recorded
using differential scanning calorimeter (DSC 823 Mettler
Toledo, Japan). Approximately 2 to 5 mg of each sample was
heated in a closed pierced aluminum pan from 30 to 300 at a
Effect of Sintering on Compritol 888 ATO Matrices
357
Fig. 2. Scanning electron microscope images of a unsintered and b sintered tablet surfaces
heating rate of 10C/min under a stream of nitrogen at a ow

rate of 50 ml/min.
PXRD Analysis
Powder X ray diffraction (PXRD) patterns of ketorolac
tromethamine, Compritol 888 ATO, and a physical mixture
of ketorolac tromethamine and Compritol 888 ATO before
and after heat treatment were investigated using powder X
ray diffractometer (PW 1729 X ray Generator, Philips, The
Netherlands). The X rays were Ni ltered CuK1 radiation
with 40 kV and 30 mA over 0 100/2.
Compritol 888 ATO matrices are shown in Fig. 1. Unsin

tered tablets showed higher drug release compared to tablets
subjected to sintering at 80C. Among the tablets subjected to
sintering, desired drug retardation was when tablets were
sintered for 3 h. Drug release from all tablets shows Higuchi
diffusion controlled matrix release. It was also noted that as
the duration of sintering increased, the retardation of drug
release was increased (Fig. 1).

Figure 2 shows the micrographs of surface of the tablet
before and after sintering for period of 3 h at 80C.
Stability Studies
To assess the drug and formulation stability, stability
studies were done according to ICH and WHO guidelines.
Prepared formulations were kept in humidity chamber (Remi
Instrument Ltd., Mumbai) maintained at 40C and 75%
relative humidity (RH) for 3 months. The sample was
analyzed for the physical changes and percent drug content
at interval of 7, 15, 30, 60, and 90 days.
RESULTS
Wettability
The effect of sintering on the wettability of the tablet
surfaces was established by taking the photographs of the
tablet surfaces on which the drop of puried water containing
amaranth red was placed. The contact angle of sintered tablet
was greater than that of unsintered tablet, as shown in Fig. 3.
This clearly indicates that sintering decreases the wettability
of tablet surfaces.
Preparation of Compritol 888 ATO Matrix Tablets

One hundred tablets of 150 mg each were directly
compressed using 6 mm diameter punches on a rotary tablet
compression machine (Mini Press II MT, Rimek). Hardness
of tablets was kept to 5 6 kg/cm2. The assay of the tablets was
within the ofcial limits and was in the range of 97 99%.
The drug release proles of formulations were studied.
The drug release proles of ketorolac tromethamine from

To investigate the interaction during sintering, the FTIR
spectra of ketorolac tromethamine, Compritol 888 ATO, and a
physical mixture of ketorolac tromethamine and compritol 888
ATO before and after sintering were recorded (Fig. 4). In the
spectrum of ketorolac tromethamine, major peaks 3,350 cm 1
[NH stretch]; 1; 725 cm1 C O stretchacid1; 167 cm1 C
O stretchdiaryl ketone; and 3,450 cm 1 [OH (acid)] were seen
in subsequent spectra.
Fig. 3. Photoimages showing contact angle of water with a unsintered tablet and
b sintered tablets
358
Fig. 4. DSC thermograms of ketorolac tromethamine, Compritol

888 ATO, unsintered (control), and sintered tablets
DSC Studies
DSC spectra of ketorolac tromethamine, Compritol
888 ATO, and a physical mixture of ketorolac tromethamine
and Compritol 888 ATO before and after sintering are
shown in Fig. 5.
Fig. 6. X ray diffraction spectra of ketorolac tromethamine, Compri

tol 888 ATO, unsintered (control), and sintered tablets
powder occurs at 2=12.62, 13.99, 16.43, 17.04, 18.41, 20.10,

22.07, 22.89, and 29.39, which shows that ketorolac trometh
amine exists in crystalline form.
PXRD Analysis
Stability Studies
X ray diffractograms of ketorolac tromethamine, Com
pritol 888 ATO, and a physical mixture of ketorolac
tromethamine and Compritol 888 ATO before and after
sintering were recorded (Fig. 6). The gure shows two peaks
in its diffraction pattern at 2=21.33, 23.45. The X ray
diffraction peaks for the ketorolac tromethamine pure
The stability studies of prepared formulations revealed

no signicant changes in the physical parameters when stored
at temperature and humidity conditions of 402C/755%
RH. Samples were withdrawn and retested for drug content
after intervals of 7, 15, 30, 60, and 90 days. Percent drug
content was found in all the prepared formulations ranging
from 95.210.41 to 97.610.37, indicating that no signicant
reduction in the content of the active drug was observed over
a period of 3 months; the percent drug contained is found
within a specied limit of USP. Therefore, there was no
evidence of degradation of drug quantity.
DISCUSSION
Fig. 5. FTIR of ketorolac tromethamine, Compritol 888 ATO,

unsintered (control), and sintered tablets
In this work, we studied the mechanism of the effect of

sintering on the release of ketorolac tromethamine as a
model drug. SEM, PXD, FTIR, and DSC were used to
demonstrate the suggested mechanism. Unsintered tablets
showed higher drug release in 12 h, whereas with sintering
at 80C for 1, 2, and 3 h, percent drug release was found to
decrease, which indicated that increasing the sintering time
decreased the drug release rate from the matrices. Table II
shows the values of coefcients and release rate constants
for the unsintered and sintered (3 h) tablets. The R values
indicate that Higuchi matrix model is the best t model for
both sintered and unsintered tablets. This is further sub
stantiated by Higuchi square root law (Fig. 7). As seen in
Fig. 2, the surface of the tablet is smoother after heat
treatment. SEM micrographs of the surface of the tablets
after sintering show that a thin lm like structure covers the
entire surface, indicating that heat treatment causes the wax
to melt, redistribute, and coat drug and excipient particles,
thus creating new surfaces with lower wettability and which
is responsible for the retardation of drug release from matrix
Effect of Sintering on Compritol 888 ATO Matrices
359
Table II. Coefcients and Release Rate Constants for Different Models
Sr. no.
Formulation
Model
Unsintered
Sintered
(1 h)
Sintered
(2 h)
Sintered
(3 h)
Zero Order
First Order
Korsemeyer Peppas
Hixon Crowell
Higuchi Matrix
Zero Order
First Order
Korsemeyer Peppas
Hixon Crowell
Higuchi matrix
Zero order
First order
Korsemeyer Peppas
Hixon Crowell
Higuchi matrix
Zero order
First order
Korsemeyer Peppas
Hixon Crowell
Higuchi matrix
0.8850
0.8985
0.9985
0.9662
0.9935
0.8718
0.9800
0.9958
0.9561
0.9983
0.8332
0.9104
0.9958
0.9733
0.9979
0.8845
0.9902
0.9969
0.9802
0.9976
0.0472
27.5878
0.2035
8.4338
25.7140
7.0792
0.1114
19.9201
0.0316
20.8353
7.1753
0.1811
26.8916
0.0412
24.3104
4.0471
0.0718
16.3213
0.0194
14.9491
tablets. The duration of sintering also affected the porosity,

with tablets sintered for 1 h showing higher porosity as
compared to those sintered for 3 h.
Contact angle is indicative of the wettability of the tablet
surface. It is an important physical property that will have a
far reaching impact on the release of the drug especially
from a hydrophobic wax matrix. This indicates a decrease in
the wettability of the sintered tablet. The surface of the
tablets made from direct compression consists of a hetero
geneous mixture of the ingredients of the tablet. After
sintering, the hydrophobic wax is more uniformly dispersed
through the compact, and this leads to a decrease in the
wettability of the tablet surface. This is evidenced by the
higher contact angle and further contributes to the retarda
tion of drug release from matrix tablet. When the tablets are
exposed to temperatures over the melting point of the
incorporated wax, the wax present in liquid state will move
through the matrix of the tablet and get lled between the
pores of the matrix without affecting the overall shape of the
tablet. This coating of the drug particles is generally referred
to as in situ micro coating (14).
During the heat treatment, there are possibilities of
interaction between wax and drug which could have an
impact on the retardation of drug release.
Comparison of the spectra of physical mixtures and heat
treated samples of ketorolac tromethamine with Compritol
888 ATO showed no difference in the position of the
absorption bands. The spectra can be simply regarded as the
superposition of those of ketorolac tromethamine and
Compritol 888 ATO. This observation ruled out the
possibility of chemical interaction and complex formation
between these two components by sintering.
DSC was used to examine the thermal behavior of pure
drug and formulation. The thermogram of ketorolac trometh
amine shows a very sharp endothermic peak at 171C and
Compritol 888 ATO shows sharp endothermic at approxi
mately 72C. Comparison of the spectra of physical mixture

(control) and sintered sample shows that no change has
occurred after sintering of sample. Therefore, there is no
evidence of drug interaction or complexation during manu
facturing process and on sintering. Often, DSC is used to
investigate the polymorphic behavior of the compounds, but
X ray diffraction is a much more suitable technique for the
identication of the polymorphs because of their distinctive
diffraction pattern (12).
Polymorphic structure of a drug is an important param
eter which inuences the dissolution rate and bioavailability
of drug (15). To evaluate the effect of sintering, X ray
diffractograms were taken. Compritol 888 ATO shows two
peaks that are due to lipidic polymorphism. This further
reveals that these two peaks are due to the lateral packing of
fatty acid chains specically designated as sub , , , and ,
each of them corresponding to a particular lateral organiza
Fig. 7. Higuchi square root plot
360
tion of hydrocarbon chains (14). The diffraction peaks for
ketorolac tromethamine were located in the same position for
both physical mixture (control) and sintered sample. These
results further indicate the absence of any crystalline change
during the heat treatment.
CONCLUSION
Sintering is dened as bonding of adjacent particle
surfaces in a mass of compact by the application of heat.
Our study showed that drug release prolongation after
sintering can be attributed to the melting and redistribution
of the wax in the tablet matrix structure. Drug release from
all tablets shows Higuchi diffusion controlled matrix release.
Sintering of the tablets results in melting and redistribution of
the wax throughout the matrix and a possible change in the
nature of the pores within the matrix. FTIR results did not
show any drug wax interaction due to heat treatment. DSC
and PXD studies ruled out the occurrence of solid solution
and polymorphic change of the drug.
ACKNOWLEDGMENTS
The authors would like to thank Dr. K.G. Bothara,
Principal, AISSMS College of Pharmacy, Pune for his constant
support and encouragement.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
REFERENCES
1. Y. Zhang, and B. Schwartz. Effect of processing method and heat
treatment on formulation of wax matrix tablet. Pharm. Dev.
Tech. 69:131 144 (2001).
2. S.B. Tiwari, T. Krishnamurthy, M.R. Pai, P. R. Mehta, and P.B.
Choudhary. Controlled release formulation of tramadol hydro
13.
14.
15.
chloride using hydrophilic and hydrophobic matrix system.

AAPS Pharm. Sci. Tech. 4:31 (2003).
C.L. Luk, and H.L. Jane. Sintering in pharmaceutics. In J. Swarbrick,
and J. C. Boylan (eds.), Encyclopedia of Pharmaceutical Technology,
14:2nd ed., Marcel Dekker, New York, 1996, pp. 87 101.
S.B. Rao, P.S. Cheruvu, K.S. Sudha et al. Studies on release of
rifampicin from sintered matrix tablets. Indian J. Pharm. Sci. 63:371
378 (2001).
S.B. Rao, A. Seshsayana, K. Himasankar et al. Design and
evaluation of ethylene vinyl acetate sintered matrix tablets.
Indian J. Pharm. Sci. 65:496 502 (2003).
A. Kondaiah, and K. Prakash. Design of controlled release non
erodible polymeric matrix tablets of theophylline using sintering
technique. Indian J. Pharm. Sci. 64:239 243 (2002).
S.B. Rao, K.H. Sankar, and B.S. Sirisha. Design and evaluation
of Eudragit RL100 sintered matrix tablets. Indian J. Pharm. Sci.
66:202 207 (2004).
S. Azarmi, J. Farid, A. Nokhodechi, and H. Valizadeh. Thermal
treating as a tool for sustained release of indomethacin from
Eudragit RS and RL matrices. Int J. Pharm. 246:171 177
(2002).
D. Jung, E. Mroszczak, and L. Bynum. Pharmacokinetics of
ketorolac tromethamine in humans after intravenous, intramus
cular and oral administration. Eur. J. Clin. Pharmacol. 35:423
425 (1988).
E.J. Mroszczak, F.W. Lee, and D. Combs. Ketorolac trometh
amine absorption, distribution, metabolism, excretion and phar
macokinetics in animals and humans. Drug Metab. Dispos.
15:618 626 (1987).
J.B. Brubach, V. Jannin, B. Mahler, and P. Roy. Structural and
thermal characterization of glyceryl behenate by X ray diffraction
coupled to differential calorimetry and infrared spectroscopy. Int.
J. Pharm. 336:248 256 (2007).
S.B. Souto, W. Mehnert, and R.H. Muller. Polymorphic behav
iour of Compritol 888 ATO as a bulk lipid and as SLN and NLC.
J. Microencapsul. 234:417 433 (2006).
R. Singh, S.S. Poddar, and A. Chivate. Sintering of wax for controlling
release from pellets. AAPS Pharm. Sci. Tech. 8:E1 E9 (2007).
S. Azarmi, F. Ghaffari, R. Lobenberg, and A. Nokhodchi.
Mechanistic evaluation of the effect of thermal treating on
Eudragit RS matrices. II Farmaco. 60:925 930 (2005).
H.M. Abdou. Dissolution, Bioavailability and Bioequivalence,
Mack, Easton, 1989.

DOI: 10.1208/s12249 009 9213 6
Research Article
Intranasal Microemulsion of Sildenafil Citrate: In Vitro Evaluation and In Vivo
Pharmacokinetic Study in Rabbits
Ahmed H. Elshafeey,1,2 Ehab R. Bendas,1 and Osama H. Mohamed1
Received 29 June 2008; accepted 1 March 2009; published online 31 March 2009
Abstract. The purpose of the present study was to prepare intranasal delivery system of sildenal citrate
and estimate its relative bioavailability after nasal administration in rabbits to attain rapid onset of action
with good efcacy at lower doses. Sildenal citrate saturated solubility was determined in different
solvents, cosolvents, and microemulsion systems. For nasal application, sildenal citrate was formulated
in two different systems: the rst was a cosolvent system (S3) of benzyl alcohol/ethanol/water/Transcutol/
taurodeoxy cholate/Tween 20 (0.5:16.8:47.7:15.9:1:18.1% w/w). The second was a microemulsion system
(ME6) containing Oleic acid: Labrasol/Transcutol/water (8.33:33.3:16.66:41.66% w/w). The prepared
systems were characterized in relation to their clarity, particle size, viscosity, pH, and nasal ciliotoxicity. In
vivo pharmacokinetic performance of the selected system ME6 (with no nasal ciliotoxicity) was evaluated
in a group of six rabbits in a randomized crossover study and compared to the marketed oral tablets. The
targeted solubility (>20 mg/ml) of sildenal citrate was achieved with cosolvent systems S1, S3, and S5
and with microemulsion systems ME3 ME6. The saturated solubility of sildenal citrate in cosolvent
system S3 and microemulsion system ME6 were 22.981.26 and 23.791.16 mg/ml, respectively.
Microemulsion formulation ME6 showed shorter tmax (0.75 h) and higher AUC(0 ) (1,412.42 ng h/ml)
compared to the oral tablets which showed tmax equals 1.25 h and AUC(0 ) of 1,251.14 ng h/ml after
administration to rabbits at dose level of 5 mg/kg. The relative bioavailability was 112.89%. In
conclusion, the nasal absorption of sildenal citrate microemulsion was found to be fast, indicating the
potential of nasal delivery instead of the conventional oral administration of such drug.
KEY WORDS: bioavailability; intranasal; microemulsion; nasal; sidenal citrate; solubilization.
INTRODUCTION
Erectile dysfunction (ED) is estimated to affect up to
50% of men between the ages of 40 and 70 years (1,2). ED is
frequently associated with depression, increased anxiety, and
poor self esteem and compromises interpersonal relationships
(3). In diabetes mellitus, ED occurs at an earlier age and has
a higher prevalence. ED is a side effect commonly associated
with use of the selective serotonin uptake inhibitors to treat
depression (4). To induce penile erection, relaxation of the
smooth muscle cells of the corpus cavernosum and associated
arterioles is required. A major component of this relaxation
process is mediated by nitric oxide stimulation of cyclic
guanosine monophosphate (cGMP). In response to sexual
stimulation, the release of nitric oxide by nerve endings and
endothelial cells increases the levels of cGMP to induce the
erection. cGMP is readily hydrolyzed by phosphodiesterase 5
(PDE5) resulting in restoration of quiescent muscle tone and
detumescence (5). Sildenal citrate is selective inhibitor of
PDE5 with IC50 values of 3.9 nM (6). It is rapidly absorbed
Pharmaceutics Department, College of Pharmacy, Cairo University,

Cairo, Egypt.
2
To whom correspondence should be addressed. (e mail:
Ah elshafeey@hotmail.com)
after oral administration with absolute bioavailability 40%. A

high fat meal delays the onset of action for sildenal (7). A
sublingual preparation of sildenal has been developed, and
this will not be affected by food. An initial study with
sublingual sildenal (20 mg) has shown that the mean onset
of action was 15.5 min and lasted for an average of 40 min
with 13/20 of subjects with ED achieving erections (8). Nasal
delivery has been paid attention as an alternative dosage
form. Nasal mucosa offers the possibility for simple and
comfortable drug administration. Nasal mucosa possesses a
total thickness of only about 100 m and consists of various
cell types, which are ciliated and nonciliated columnar cells,
goblet, and basal cells (9). The nasal respiratory region, which
extends backward by approximately 6 8 cm to the nasophar
ynx, is the primary site for drug absorption. The mucosa in
this region consists of an epithelium resting on a basement
membrane connected with two to four layers of submucosa
(10). Epithelial cells in the nasal mucosa are generally in close
apposition, and adjacent cells are extremely adherent.
Cohesion at the apices of epithelial cells occurs in specialized
regions collectively referred to as the junctional complex.
This structure controls the diffusion of ions and molecules
between cells and constitutes a barrier to the intercellular
movement of macromolecules across the epithelium (11). A
lipophilic component in the formulation seems to be advan
tageous, but in case of nasal administration, incompatibilities
361
362
Elshafeey et al.
can be indicated by impairment of ciliary movement. Use

of a microemulsion may minimize these side effects by
administering a lipophilic component in a transparent,
water continuous system (12,13).
Advantages of microemulsion include its ease of prepa
ration due to spontaneous formation, thermodynamic stabil
ity, transparent and elegant appearance, increased drug
loading, enhanced penetration through the biological mem
branes, increased bioavailability (14,15), and less inter and
intra individual variability in drug pharmacokinetics (16).
These advantages make microemulsions attractive drug
delivery systems.
The advantages of nasal route have been suggested as
follows: rapid absorption, higher bioavailability allowing
lower doses, fast onset of therapeutic action, avoidance of
liver or gastrointestinal metabolism, avoidance of irritation of
the gastrointestinal membrane, reduced risk of overdose,
non invasive administration, ease of convenience and self
medication, and improved patient compliance (17). Many
authors concluded that the nasal route has been found to give
improved bioavailability compared to the oral route as in case
of metoclopramide HCl (18), propranolol (19,20), and mid
azolam (21). So, it would be valuable to develop a sildenal
nasal delivery system to increase its bioavailability and hence
decrease the administered dose and also to attain rapid onset
of action.
This study aimed to develop a new sildenal intranasal
formula and to study its relative bioavailability compared to
the conventional oral tablets after intranasal administration to
rabbits.
and Tween 20 were purchased from Sigma Co. (St. Louis, MO,
USA), dimethyl isosorbide (DMI) was obtained from (Aldrich
Chemical Company, Milwaukee, WI, USA).
Preparation of Sildenafil Citrate Cosolvent Systems

Determination of Sildenafil Citrate Saturated Solubility
in Different Solvents
Excess amounts of sildenal citrate were mixed with 2 ml
of various solvents viz. 0.1 N HCl, 10% citric acid solution,
propylene glycol, PEG 400, 10% hydroalcoholic solution,
50% aqueous PEG 400 solution, DMI, and distilled water in
separate vials. The vials were well closed and left to be
shaken at 25C for 24 h in thermostatically controlled shaking
water bath. The suspensions were centrifuged at 5,000 rpm
for 10 min. The contents of each vial were ltered through a
0.45 m lter, and the supernatant was assayed for the drug
content using high performance liquid chromatography
(HPLC) method.
Determination of Sildenafil Citrate Solubility in Different

Cosolvent Systems
Different cosolvent systems were used to determine the
maximum solubility of sildenal citrate by the same method
mentioned in the previous step. Systems S1 to S5 were
prepared according to the composition listed in Table I.

Preparation of Sildenafil Citrate Microemulsion Systems
Materials
Sildenal citrate was a gift from Sigma Co, Egypt; hydro
chloric acid (HCl), citric acid, propylene glycol, polyethylene
glycol 400 (PEG 400), ethanol, isopropyl myristate (IPM),
dioxane, benzyl alcohol, and isobutanol were obtained from
Fisher Scientic (Fair Lawn, NJ, USA); miglyol, Labrasol,
labral M, and Transcutol were kindly obtained from
(Gattefoss, St. Priest, France); oleic acid, taurodeoxycholate,
Determination of Saturated Solubility of Sildenafil Citrate

in Different Oils
This step was performed as previously discussed in order
to select the suitable oil which has a good solubilizing capacity
for sildenal citrate. This oil can be used as the oil phase in
microemulsion system. The oils selected were isopropyl
myristate, miglyol, labral M, and oleic acid.
Table I. Composition of Different Cosolvent Formulations (%, w/w) and Saturated Solubility of Sildenal Citrate in these Solvent Mixtures at
25C
S1
S2
S3
S4
S5
Composition
% w/w
% w/w
% w/w
% w/w
% w/w
Benzyl alcohol
Distilled H2O
Transcutol
Taurodeoxycholate
Tween 20
PEG 400
Ethanol
Isobutanol
DMI
Labrasol
Saturated solubility (mg/ml)SD
17.3
47.7
15.9
1.0
18.1
0.5
47.7
15.9
1.0
18.1
16.8
0.5
47.7
15.9
1.0
18.1
0.5
47.7
15.9
1.0
18.1
0.5
47.7
15.9
16.8
11.8
5
34.902.25
18.881.20
22.981.26
19.351.99
19.1
16.8
10
20.052.13
Intranasal Microemulsion of Sildenafil Citrate
363
Formulation of Different Microemulsion Systems and

Determination of Solubility of Sildenafil Citrate in These
Systems
The pseudoternary phase diagrams of the oil (oleic acid),
distilled H2O and surfactant (Labrasol), and cosurfactant
(Transcutol, plurol or DMI) mixtures were constructed at room
temperature (25C). Appropriate quantities of oleic acid,
Labrasol, and cosurfactant were stirred at high rate until it
formed clear solution, then water was added dropwise to each
mixture under vigorous stirring. The microemulsions were left
to attain equilibrium. The resulting microemulsions were tightly
sealed and stored at 25C for further evaluation. The compo
sition of each microemulsion system is shown in Table II.
The solubilization capacity of each microemulsion for
sildenal citrate was investigated. Increasing amounts of
sildenal citrate were added portion wise to 2 ml of each
microemulsion system with stirring after each addition to
form a clear and transparent liquid till excess solid does not
dissolve. The mixture was stirred at 25C for 24 h. The
solubility of the drug was determined as mentioned before,
but ltration was done without centrifugation.
Measurement of pH, Viscosity, and Droplet Size
The pH of cosolvent systems and microemulsions were
measured using pH meter (Model C G 820, Schott Gerate,
Germany). The viscosities of the cosolvent systems and micro
emulsions were evaluated at 25C using a viscometer (Brook
eld Viscometer, model LVT, USA). The mean diameter of the
droplets of the microemulsions was measured at 25C using
image analysis software. Samples of sildenal microemulsions
were examined microscopically for particle size analysis at a
magnication of 10 and 40 with a Leica image analyzer Model
Q 550IW equipped with Leica DM LB microscope (Cambridge,
England), which consists of binocular microscope equipped with
a computerized digital camera. A drop of microemulsion
preparation was placed on a microscope slide and then was
examined. The mean diameter of the microemulsion droplets
was measured after storage for 3 months at 25C.
Nasal Ciliotoxicity
Nasal ciliotoxicity studies were carried out using in situ
toad palate model with minor modication (22). The upper
palate of the male toads (20 30 g, Animal house of College of
Pharmacy, Cairo University, Egypt) was exposed and treated
with 0.5 ml of tested microemulsion containing sildenal
citrate (20 mg/ml) or with solvent system containing sildenal

citrate (20 mg/ml) for 30 min then rinsed with saline. The
palate was dissected out, and the mucocilia was examined
with an optical microscope (Leica Imaging Systems, Cam
bridge, England). Saline was used as a control.
In Vivo Nasal Absorption
Animal Handling and Drug Administration
Six male New Zealand white rabbits weighing 2.30
0.12 kg were housed individually in stainless steel cages, fed a
commercial laboratory rabbit diet, and allowed free access of
water. The rabbits were fasted for 18 h prior to and during the
pharmacokinetic study. The animals were conscious through
out the duration of the experiments and were held in rabbit
restrainers during blood sampling. In a crossover study with
1 week apart as a wash out period, the selected sildenal
microemulsion was administered intranasally to each rabbit in
a dose of 5 mg/kg divided equally in each of the two nostrils
using a micropipette inserted 1 cm into the nostril. Sildenal
citrate tablet (Viagra 50 mg) was crushed and suspended in
5 ml of saline. The equivalent volume containing 5 mg/kg was
administered orally via gastric intubation for each fasted
rabbit. For all animal studies, the experimental procedures
conformed to the ethical principles of the Center of Applied
Research and Advanced Studies, Faculty of Pharmacy, Cairo
University on the use of the animals. All animals were treated
according to the principles of laboratory animal care
(National Institutes of Health publication #86 32) (23).
Sample Collection and Analysis
After administration of the different formulations, blood
samples (1.5 ml) were collected at time intervals of 0.00, 0.25,
0.50, 0.75, 1.00, 1.25, 1.50, 2.00, 2.50, 3.00, 4.00, 5.00, 6.00, and
7.00 h from the central auricular artery of the rabbits. Blood
samples were allowed to clot and then centrifuged at
3,000 rpm for 10 min. The obtained serum samples were
deep frozen at 20C, pending HPLC analysis. The plasma
analysis of sildenal was performed using a well validated
HPLC UV method after liquid liquid extraction. Aliquots of
0.5 ml plasma were pipetted into 4 ml centrifuge tubes, 100 l
of internal standard (Glimepiride) was added and vortexed
for 10 s. Then, 100 l (1.0 M) hydrochloric acid was added
and then vortexed for 30 s. Ethyl acetate (4.0 ml) was added
and vortexed for 30 s, then centrifuged for 10 min at
4,000 rpm. The organic layer was decanted into clean
Table II. Composition of Different Microemulsion Systems (%, w/w) and Saturated Solubility of Sildenal Citrates in these Systems at 25C
ME1
ME 2
ME 3
ME 4
ME 5
ME 6
ME 7
Composition
% w/w
% w/w
% w/w
% w/w
% w/w
% w/w
% w/w
Oleic acid
LabrasolLabrasol
Transcutol
DMI
Distilled H2O
Saturated solubility (mg/ml)SD
4
40
10
20
26
16.831.78
4
30
10
26
30
18.441.70
8.33
40
10
20
21.67
23.582.34
8.33
33.33
16.67
16.67
25
22.302.48
8.33
40
10
8.33
33.33
16.67
10
20
20
41.67
21.122.11
41.67
23.791.16
50
15.361.10
364
Elshafeey et al.
centrifuge tubes, then evaporated using vacuum concentrator

(Eppendorf Concentrator 5301, Germany) till dryness at
45C. The residue was reconstituted with 100 L of mobile
phase, and then 25 l was injected into the HPLC system
using autosampler (SIL 10A Shimadzu, Japan). The station
ary phase used was Discovery C18 (Supelco; 2504.6 mm,
5 m particle size). The guard column was Thermo C18 (5
4.0 mm, 5 m particle size). The mobile phase consisted of
acetonitrile and (0.05 M) phosphoric acid (50%, 50%, v/v),
adjusted at pH 4.00 with phosphoric acid before ltration.
The mobile phase was delivered into the HPLC apparatus at
ow rate of 0.5 ml/min (isocratic pump, Model LC 20AD,
Shimadzu, Japan). The detection wave length was 220 nm
(ultraviolet variable wavelength detector, Model SPD 20A
Shimadzu, Japan). All assays were performed at ambient
temperature.
Table IV. Saturated Solubility of Sildenal Citrate in Various Oils at

25C
Pharmacokinetic Analysis
Solubilities of sildenal citrate in different oils at 25C

were presented in Table IV. They were 3.180.33, 4.230.39,
and 2.340.21 mg/ml for IPM, miglyol, and labral, respectively.
Oleic acid showed the highest solubility (6.770.54 mg/ml)
compared to other oils so it was chosen as the oil phase for
microemulsion preparation.
The formation of microemulsion systems (ME1 ME7)
was observed at room temperature. During the addition of
water to the selected oily mixtures (mixture of oleic acid,
Labrasol, Transcutol in addition to DMI in all formulations, a
continuous transition from water in oil systems (W/O) to oil
in water (O/W) systems was observed, a transparent, one
phase and low viscous system was obtained. The O/W
microemulsions formed are shown in three component trian
gular diagram (Fig. 1).
The solubilities of sildenal citrate in different micro
emulsion systems at 25C are shown in Table II. The required
solubility of sildenal citrate (>20 mg/ml) was achieved in
microemulsion systems ME3 ME6 which contained 8.33%
oleic acid; there was no statistically signicant difference in
Pharmacokinetic analysis was performed by means of a

model independent method (noncompartmentally) using a
KineticaTM 2000 computer program. The elimination rate
constant (Lz) was obtained as the slope of the linear
regression of the log transformed plasma concentration
values versus time data in the terminal phase. The
elimination half life (t1/2) was calculated as 0.693/Lz. The
area under the curve to the last measurable concentration
(AUC0-t) was calculated by the linear trapezoidal rule. The
area under the curve extrapolated to innity (AUC0-) was
calculated as AUC0-t +Ct/Lz, where Ct is the last measurable
concentration.
RESULTS
Cosolvent Systems
The saturated solubility of sildenal citrate in different
solvents at 25C presented in Table III could be arranged in
an ascending order as follows: 1.400.17, 3.200.11, 3.31
0.22, 3.610.25, 4.050.23, 4.790.40, and 6.890.62 mg/ml
for 10% citric acid, distilled water, propylene glycol, 10%
ethanol in water, PEG 400, 50% PEG 400 in water, and 0.1 N
HCl, respectively. The maximum solubility was achieved in
DMI (9.980.79 mg/ml).
The solubilities of sildenal citrate in different cosolvent
mixtures at 25C (Table I) were 34.902.25, 18.881.20,
Oil
IPM
Miglyol
Labral
Oleic acid
Saturated solubility (mg/ml)

3.180.33
4.230.39
2.340.21
6.770.54
22.981.26, 19.351.99, and 20.052.13 for S1, S2, S3, S4,

and S5, respectively.
Microemulsion Systems
Table III. Saturated Solubility of Sildenal Citrate in Different

Solvents at 25C
Solvent
Distilled water
0.1 N HCl
10% Citric acid
Propylene glycol
PEG 400
10% ethanol in water
50% PEG 400 in water
DMI
Saturated solubility (mg/ml)S.D

3.200.11
6.890.62
1.400.17
3.310.22
4.050.23
3.610.25
4.790.40
9.980.79
Fig. 1. Pseudo ternary phase diagram of a microemulsion system

composed of oil (oleic acid), surfactant (Labrasol), cosurfactant
(Transcutol), and water. Closed triangles microemulsion, closed circles
clear gel, and closed squares emulsion
365
Table V. Droplet Size, Apparent Viscosity, pH, and Clarity of the Selected Microemulsion and Solvent Vehicles (MeanSD, n 3)
Formulation
ME6
ME6 (stored for 3 months at 25C)
S3
Droplet size (nm)
Viscosity (mpas)
pH
Clarity
37.51.9
38.22.2
47233
47545
26523
6.00.5
6.00.5
4.50.5
Clear
Clear
Clear
solubilization for sildenal citrate among these four micro

emulsion formulations (p<0.05).
pH, Viscosity, and Droplet Size
All the prepared systems were transparent and clear
even after storage for 3 months. The pH, viscositi of cosolvent
system S3 and microemulsion system ME6, and the droplet
size of ME6 are shown in Table V. pH of ME6 was 6.00.5
compared to 4.50.5 for the solvent system S3. The droplet
size of ME6 containing sildenal citrate was not signicantly
affected by incorporation of the drug when compared to the
droplet size of microemulsion prepared with no drug. No
signicant droplet size change was found when these prepa
rations were stored at 25C for 3 months.
Nasal Ciliotoxicity
Optical microscope results showed that there were a
great number of cilia with fast rate beating on the edge of
mucosa that was treated with microemulsion formulation
ME6. The same effect was observed when treating the
mucosa with saline. In contrast, the effect of solvent system
S3 on ciliary movement was inhibitory, and no ciliary
movement was observed as shown in Fig. 2.
of the method and the reproducibility of the assay (tables not

shown) revealed that the mean CV% was <3.62.
The pharmacokinetics of sildenal citrate was determined
for the selected microemulsion ME6 composed of Oleic acid/
Labrasol/Transcutol/H2O (8.33:33.33:16.67:41.67%) and was
compared to the peroral tablets. The sildenal citrate content
was consistent in both formulations to achieve a dose of 5 mg/kg.
The mean serum drug concentration time proles after
administration of the oral tablets as well as the nasal sildenal
microemulsion are illustrated in Fig. 3.
From the plasma time prole, it is clear that nearly
twofolds higher serum drug levels were achieved in case of
the nasal microemulsion compared to the oral tablet. The
Cmax values were 713.1622.98 ng/ml for the nasal micro
emulsion, while it was 346.5018.72 ng/ml for the oral tablets
as listed in Table VI. Statistical analysis revealed that the
Cmax was signicantly higher in case of the nasal micro
emulsion system. Concerning the rate of absorption, the
results showed that the nasal microemulsion had signicantly
shorter tmax values of 0.750.00 h compared to 1.250.00 h
for the oral tablets at P<0.05. Moreover, the AUC0 values
were 1,412.4225.87 and 1,251.1430.19 ng h/ml for the nasal
microemulsion and oral tablets, respectively. These values
corresponded to relative bioavailability values of 112.89%.
DISCUSSION
In Vivo Nasal Absorption

The mean percentage recovery of sildenal from quality
control low and quality control high spiked quality control
samples was 95.333.16 and 99.32.65 (CV%=2.96 and 2.12,
respectively), and the mean correlation coefcient of the
standard curves was 0.9969. Within day accuracy and precision
The saturated solubility study in different solvents

showed that 10% citric acid gave the lowest drug solubility,
while DMI gave the highest drug solubility. Although it has
been reported that solubility determinations in various
buffers showed a maximum solubility at acidic pH (24),
sildenal citrate showed the lowest solubility in 10% citric
Fig. 2. Optical microscopic images of cilia on the mucosa half an hour after treatment with a negative control (saline), b
microemulsion ME6, and c solvent system S3
366
Elshafeey et al.
Fig. 3. Plasma concentration time curve of sildenal citrate after

nasal administration of microemulsion ME6 and oral administration
of tablets to rabbits in a dose of 5 mg/kg
acid solution. This may be explained due to the salting out

effect exerted by citric acid. None of the tried solvents was
satisfying in achieving the required solubility (20 mg/ml) to be
used as nasal formulation.
The solubilities of sildenal citrate in different co solvent
mixtures showed that S1 attained the maximum solubility of
sildenal citrate (34.902.25 mg/ml); however, this system
contains high amount of benzyl alcohol (17.3%) which is not
recommended in nasal route of administration as inhalation
of benzyl alcohol may cause headache, vertigo, nausea,
vomiting, and diarrhea. Overexposure may result in central
nervous system depression and respiratory failure (25). Both
S3 and S5 systems are considered suitable systems for
sildenal citrate solubility as the solubilities were 22.981.26
and 20.052.13 mg/ml, respectively. However, S5 contains
high concentration of DMI (16.8%) which may be irritating if
used nasally. Since S3 showed higher solubility with lower risk
of nasal irritancy or toxicity, it was chosen for further studies.
In the preparation of microemulsion systems, oleic acid
was chosen as the oil phase because it gave maximum
solubility of sildenal citrate. It was noted that the solubility
of sildenal citrate was improved by the use of micro
emulsion. Sildenal citrate solubility reached about 23.79
1.16 mg/ml in ME6, an approximately sevenfold increase
compared with intrinsic solubility in water (3.200.11 mg/ml)
and fourfold increase compared to solubility in oleic acid
(6.770.54 mg/ml).
Table VI. Pharmacokinetic Parameters of Sildenal Citrate after

Nasal Administration of Microemulsion and Oral Administration of
Tablets to Rabbits in a Dose of 5 mg/kg
Pharmacokinetic parameters
Cmax (ng/ml)
tmax (h)
AUC0 7 (ng h/ml)
AUC0 (ng h/ml)
Lz (h 1)
t1/2 (h)
MRT (h)
Nasal
Oral
713.1622.98
0.750.00
1369.1123.60
1412.4225.87
0.660.13
1.080.26
2.460.08
346.5018.72
1.250.00
1181.0522.28
1251.1430.19
0.550.10
1.300.25
3.040.16
ME6 was chosen for further investigations as it showed

higher solubility of sildenal citrate compared to other
microemulsion formulations, and it contains the higher
percent of water which is considered as a suitable carrier for
intranasal administration. Although ME3 and ME4 gave high
drug solubility, they were not chosen because they contain
DMI which might be an irritant to nasal mucosa.
pH of ME6 was 6.00.5 compared to 4.50.5 for the
solvent system S3 which indicates that less irritation effect of
the microemulsion system is expected when applied nasally
than the solvent system S3.
The viscosity of ME6 was 47233 mpas, and there was
no signicant change when stored for 3 months at 25C
indicating physical stability.
The droplet size of ME6 containing sildenal citrate was
not affected signicantly by incorporation of the drug. No
signicant droplet size change was found when these prepa
rations were stored at 25C for 3 months, indicating that
sildenal citrate loaded microemulsion was physically stable.
The constituents of preparations intended for nasal
delivery should not adversely affect the mucociliary clearance
system. Therefore, a requirement in formulation develop
ment is absence of nasal mucosal irritation and no inhibition
of ciliary movement. Results showed that there were a great
number of cilia with fast rate beating on the edge of mucosa
when treated with ME6. In contrast, no ciliary movement was
observed in solvent system S3.
Considering the solubilzation capacity, pH, viscosity, droplet
size, and nasal ciliotoxicity, the microemulsion ME6 composed of
oleic acid/Labrasol/Transcutol/H2O (8.33:33.33:16.67:41.67%)
seems to be an optimal formulation for nasal delivery of sildenal
citrate.
In the pharmacokinetic study of intranasal delivery of
sildenal citrate via a microemulsion system compared to the
oral tablets, the plasma time prole showed that nearly
twofolds higher serum drug levels were achieved in case of
the nasal microemulsion. It showed signicantly higher Cmax
and shorter tmax, with relative bioavailability 112.89%.
Knowing that sildenal is well absorbed orally (7), the
signicantly higher bioavailability of its intranasal micro
emulsion compared to its oral tablets could be attributed to
escaping the rst pass metabolism attendant with peroral
drug administration.
Theoretically, it was expected that the relative bioavail
ability of intranasally administered sildenal will reach almost
the double or more, but due to the limited mucosa contact
time in the nose and extensive rst pass metabolism of the
swallowed fraction of the dose, the relative bioavailability was
112.89%. But, on the other hand, focusing on the shorter tmax
attained with the intranasal route of administration compared
to the peroral administration, this can be considered an
important pro.
CONCLUSION
In conclusion, intranasal sildenal citrate formulated as a
microemulsion composed of oleic acid/Labrasol/Transcutol/
H2O (8.33:33.33:16.67:41.67%) represents a safe and viable
approach to achieving rapid onset systemic drug levels and
higher bioavailability through bypass of the liver metabolism
for the management of erectile dysfunction.

REFERENCES
1. Feldman HA, Goldstein I, Hatzichristou DG, Krane RJ, Mckinlay
JB. Impotence and its medical and psychological correlates: results
of the Massachusetts male aging study. J Urol. 1994;151:54 61.
2. Johannes CB, Araujo AB, Felman HA, Derby CA, Kleinman
KP, McKinlay JB. Incidence of erectile dysfunction in men 40
69 years old: longitudinal results from the Massachusetts male
aging study. J Urol. 2000;163:460 3.
3. NIH consensus development panel on impotence. Impotence.
JAMA. 1993;270:83 90.
4. Rosen RC, Lane RM, Menza M. Effects of SSRIs on sexual
function: a critical review. J Clin Psychopharmacol. 1999;19:67 85.
5. Doggrell S. Do vardenal and tadalal have advantages over
sildenal in the treatment of erectile dysfunction? Int J Impot
Res. 2007;19:281 95.
6. Boolell M, Allen MJ, Ballard SA. Sildenal: an orally active type 5
cyclic GMP specic phosphodiesterase inhibitor for the treatment
of penile erection dysfunction. Int J Impot Res. 1996;8:47 52.
7. Rajagopalan P, Mazzu A, Xia C, Dawkins R, Sundaresan P.
Effect of high fat breakfast and moderate fat evening meal on
pharmacokinetics of vardenal, an oral phosphodiesterase 5
inhibitor for the treatment of erectile dysfunction. J Clin
Pharmacol. 2003;43:260 67.
8. Deveci S, Peskircioglu L, Aygun C, Tekin MI, Dirim A,
Ozkardes H. Sublingual sildenal in the treatment of erectile
dysfunction: faster onset of action with less dose. Int J Urol.
2004;11:989 92.
9. Richter T, Keipert S. In vitro permeation studies comparing
bovine nasal mucosa, porcinecornea and articial membrane:
androstenedione in microemulsions and their components. Eur J
Pharm Biopharm. 2004;58:137 43.
10. Su KSE, In Encyclopedia of Pharmaceutical Technology, edited by
J. Swarbrick, and J. C. Baylan, Marcel Dekker, Inc. 175 201 (1993).
11. Noach A, Hurni MA, De Boer AG, Breimer DD, In Drug Absorption
Enhancement. Concept, Possibilities, Limitations and Trends, Ed.:
Harwood Academic Publishers: Chur, Vol. 3, 1994, p. 291 324.
12. Haecker R, Kampfe E, Mattern C. Androgenic substitution for
the ageing male by nasal administration of a precursor of
testosterone. Geneve: Congress: Healthy Ageing of Male, 1998.
367
13. Sullivan DA, Wickham LA, Rocha EM. Androgens and dry eye
in Sogrens syndrome. Ann NY Acad Sci. 1999;876:312 24.
14. Tenjarla SN. Microemulsions: An overview and pharmaceutical
applications. Critical Reviews in Therapeutic Drug Carrier
Systems. 1999;16:461 521.
15. Lieberman HA, Rieger MM, Banker GS. Pharmaceutical dosage
forms: Disperse systems, second ed, Vol 1. New York: Marcel
Decker, 1996, pp. 211 281, 315 70.
16. Kovarik JM, Muller EA, Van Bree JB, Tetzloff W, Kathz K.
Reduced inter and intra individual variability in cyclosporine
pharmacokinetics from a microemulsion formulation. J Pharm
Sci. 1994;83:444.
17. Behl CR, Pimplaskar HK, Sileno AP, deMeireles J, Romeo VD.
Effects of physicochemical properties and other factors on
systemic nasal drug delivery. Adv Drug Del Rev. 1998;29:89 116.
18. Zaki N, Awad G, Mortadaa N, Abd ElHady S. Enhanced
bioavailability of metoclopramide HCl by intranasal administra
tion of a mucoadhesive in situ gel with modulated rheological
and mucociliary transport properties. Eur J Pharm Sci.
2007;32:296 307.
19. Hussain A, Foster T, Hirai S, Kashihara T, Batenhorst R, Jones
M. Nasal absorption of propranolol in humans. J Pharm Sci.
1980;69:1240 43.
20. Chien YW. Nasal drug delivery systems. In: Swacrbrick J, editor.
Novel drug delivery systems. New York: Marcel Dekker, 1992,
pp. 139 96.
21. Knoester PD, Jonker DM, Van der Hoeven RTM, Vermeij TAC,
Edelbroek PM, Brekelmans GJ, de Haan GJ. Pharmacokinetics
and pharmacodynamics of midazolam administered as a concen
trated intranasal spray. A study in healthy volunteers. Br J Clin
Pharmacol. 2002;53:501 7.
22. Jiang XG, Cuii JB, Fang XL, Wei Y, Xi NZ. Toxicity of drugs on
nasal mucocilia and the method of its evaluation. Acta Pharm.
Sin. 1995;30:848 53.
23. Guide for the Care and Use of Laboratory Animals #86 23.
Bethesda, MD: National Institutes of Health. (1985).
24. Hussain AA, Dittert LW. Method of administration of sildenal
to produce instantaneous response for the treatment of erectile
dysfunction. United States Patent. 6,200,591 (2001).
25. Smolinske SC. Handbook of food, drug, and cosmetic excipients.
Boca Raton, FL. CRC Press, 1992, pp. 47 54.

DOI: 10.1208/s12249 009 9218 1
Research Article
Characterization and Stability of Emulsion Gels Based on Acrylamide/Sodium
Acryloyldimethyl Taurate Copolymer
Giulia Bonacucina,1 Marco Cespi,1 and Giovanni F. Palmieri1,2
Received 10 July 2008; accepted 1 March 2009; published online 2 April 2009
Abstract. Sepineo P 600, a concentrated dispersion of acrylamide/sodium acryloyldimethyl taurate
copolymer in isohexadecane, has self gelling and thickening properties and the ability to emulsify oily
phases, which make it easy to use in the formulation of gels and o/w emulsion gels. In this paper, gels
were prepared using a Sepineo P 600 concentration between the 0.5% and 5% (w/w), and then emulsion
gel was also prepared from the 3% Sepineo gel by adding a specic amount of almond oil. All the
prepared systems were analyzed and characterized by oscillation rheology and acoustic spectroscopy. The
particle size of the oil droplets and the microrheological extensional moduli (G and G) of the systems
were determined from acoustic parameters and used together with the classical oscillatory rheological
tests to assess the stability of the systems. Classical oscillatory analysis revealed that the dynamic moduli
were very dependent on polymer concentration; as this parameter increased, there was progressive
improvement in the sample elasticity. In fact, the mechanical spectra of the 0.5% and 1% (w/w) Sepineo
samples were characterized by strong frequency dependence and multiple crossover points, typical of
dilute polymer solution with no organized structure. On the other hand, the 3 5% (w/w) concentration
systems showed typical gel like spectra, marked by the absence of crossover points between the dynamic
moduli and by weak dependence on frequency. Nevertheless, the elastic properties of the gel like
structure even at elevated polymer concentrations were not strongly long lasting, as demonstrated by the
increase of the viscous contribution in the low frequency range during acoustic spectroscopy analysis.
This fact could indicate that the gel structure is characterized by weak polymer polymer interactions, an
advantageous characteristic for topical administration, as the sample is thus easier to rub into the skin.
Finally, both rheology and acoustic spectroscopy indicated that addition of the oily phase caused minimal
changes to the elastic character of the gel. Thus, Sepineo P 600 gel and emulsion gel are very effective
systems for use in topical and other types of applications.
KEY WORDS: acoustic spectroscopy; emulsion gel; Sepineo; viscoelasticity.
INTRODUCTION
In recent years, there has been great interest in the use of
novel polymers with complex functions as emulsiers and
thickeners because the gelling capacity of these compounds
allows the formulation of stable emulsions and creams by
decreasing surface and interfacial tension and at the same
time increasing the viscosity of the aqueous phase. In fact, the
presence of a gelling agent in the water phase converts a
classical emulsion into an emulsion gel.
Recently, new research with emulsion gels has focused
on the characterization of milk protein (used as surfactant
and gellifying agents) o/w emulsions (1 6). Other investiga
tions have examined the thickening and emulsication
properties of water soluble amphiphilic polymers, such as
modied hydroxyethylcelluloses (7 8). Gelled emulsions have
Department of Chemical Sciences, University of Camerino, via S.

Agostino 1, 62032, Camerino, Italy.
2
To whom correspondence should be addressed. (e mail: gianlippo.
palmieri@unicam.it)
been described as composite materials in which oil droplets

behave as ller particles, and free proteins (referring to
protein gel systems) form the gel matrix (9) and also act as
emulsiers. Thus, differences in emulsion gel rheology are
due to the differences in interactions across the interfaces
between the ller particles and the gel matrix (10). Previous
studies demonstrated that the presence of oil droplets could
modify the mechanical characteristics of the protein gel by
modifying its viscoelastic properties (11), substantially in
creasing gel strength (12). In fact, oil droplets increase the
bulk protein concentration so that the latter can strengthen
the gel in the bulk phase of the emulsion (12). Furthermore,
the oil droplets create an interface able to interact with the
gel matrix (11), and the presence of the proteins adsorbed at
the interface assures interactions within the matrix and ller
(13).
Investigations of viscoelastic properties have provided
fundamental characterization of these systems and made
rheology the most widely used technique for such study
(1,2,7,10,12).
In this work, the self gelling properties of the acrylamide/
sodium acryloyldimethyl taurate copolymer (Sepineo P 600),
368
Characterization and Stability of Emulsion Gels

both alone and as dispersing phase for the preparation of
o/w emulsion gels, have been investigated by oscillatory
rheological measurements and acoustic spectroscopy.
Sepineo P 600, based on the concept of droplet hydro
swelling, is a concentrated droplet dispersion of acrylamide/
sodium acryloyldimethyl taurate (a viscous liquid at room
temperature) in isohexadecane as the oily dispersing phase.
The presence in this dispersion of polysorbate 80 is important
for keeping the resultant dispersion stable. When water is
added, these polymer droplets disappear because the polymer
molecules interact with it strongly, instantly forming a stable
semisolid system (14).
The possibility of obtaining stiff and stable gelled phases
with this polymer makes it a good candidate for the
formulation of emulsion gels. Obviously, the presence of the
oily phase can inuence the viscoelastic behavior of Sepineo
P 600 gels, with consequent changes in the characteristics of
the nal emulsion.
Thus, in this study, rheological oscillatory measurements,
in particular, frequency sweep analysis, were used for the
mechanical characterization of gels and emulsions, while
acoustic spectroscopy provided a better understanding of
sample microstructure. This non destructive technique offers
a unique possibility for characterizing concentrated colloidal
dispersions while avoiding dilution, which is a limiting step,
particularly when the tested sample is highly structured. The
parameters dened by acoustic spectroscopy, such as attenu
ation frequency spectra and sound speed, allow the calcula
tion of particle size from 5 nm to 1,000 m (15).
The operating principle of acoustic spectroscopy is based
on the generation of sound pulses that pass through a sample
and are measured by a receiver. During the passage through
the sample, sound is attenuated by the presence of the liquid
medium. The energy changes in intensity and phase are
measured.
There are six mechanisms of sound interaction with a
dispersed system: viscous (related to the shear waves
generated by the particles oscillating in the acoustic pressure
eld due to the difference in the densities between particle
and medium), thermal (related to the temperature gradients
generated near the particles surface), scattering (the same
principle as light scattering), intrinsic (losses of acoustic
energy occur when the sound wave interacts with the particles
and the medium as a homogeneous phase), structural (caused
by the oscillation of a network of particles; this mechanism is
specic to structured systems), and electrokinetic (ultrasound/
double layer interactions). The electrokinetic losses are
negligible in terms of the total attenuation, making it possible
to separate electroacoustic spectroscopy from acoustic spec
troscopy (15).
Through complex processing and modeling of these
kinds of energetic contribution to the total acoustic attenua
tion, it is possible to calculate the particle size and the zeta
potential of the dispersed particles (15).
The acoustic spectrometer can also act as a micro
rheometer, taking into account the fact that, in this case,
longitudinal viscoelastic properties are measured because
the stress is not tangential, as it is in an oscillation experiment
in a rotational rheometer, but normal.
It is possible to demonstrate that ultrasonic absorption
and velocity are related to the real and imaginary part of the
369
complex modulus (16) and that G and G moduli are related
to sound speed (V), sound attenuation (), and frequency
(), using the following equations (16):
G0 V 2
G0 2V 3 =!
Based on these facts, a number of Sepineo P 600 systems

were prepared using a Sepineo concentration ranging be
tween 0.5% and 4%, and after accurate characterization of
Sepineo P 600s gelling ability, almond oil was added to the
3% Sepineo gel. This oil was chosen because it is widely used
in pharmaceutical and cosmetic applications for its practically
inexistent toxicity and its high tolerability. The emulsion
stability over time was monitored by observing the variation
of the rheological parameters and mean diameter of the
droplets.
Sepineo P 600 and Sepicide HB, an anti microbial agent
composed of a mixture of phenoxyethanol, methylparaben,
ethylparaben, propylparaben, and butylparaben, were a kind
gift of Seppic (Paris). Almond oil USP was obtained from
ACEF s.p.a. (Fiorenzuola DArda, Italy), and deionized
water was obtained from an MF3 ion exchange system (San
Salvatore di Cogorno, Genova, Italy).
Gel Preparation
Gels were prepared for simple dispersion by mechanical
stirring (30 min at 300 rpm; Eurostar Digital, IKA Labor
technik) of the Sepineo in the required amount of deionised
water. The Sepineo concentration ranged between the 0.5%
and 5% (w/w).
Systems were then left to rest for 24 h before being
analyzed.
Emulsion Preparation
Emulsions were obtained by dispersing the 3% w/w Sepineo
P 600 in the required amount (20% w/w) of almond oil and then
adding the aqueous phase, containing 0.5% (w/w) of Sepicide
HB, and stirring (Eurostar Digital, IKA Labortechnik,
1,200 rpm).
Samples were left at room temperature for 24 h before
being analyzed and then stored for 3 months at 40C to check
sample stability over time.
Rheological Characterization
Non destructive oscillatory measurements made it possi
ble to obtain the principal rheological parameters, such as the
storage or elastic modulus (G), the loss or viscous modulus
(G), and the loss tangent (tan ). Rheological analyses were
performed in triplicate using a stress control rheometer
(Stress Tech, Reologica) equipped with a cone plate geome
try (4/40) operating in the oscillation mode. The gap was
370
Bonacucina, Cespi and Palmieri
150 m. Several tests were carried out on gels and emulsion

systems:
Oscillation stress sweep. The sample was exposed to
increasing stress (0.05 10 and 0.05 50 Pa) at constant
frequency (1 Hz) and temperature (20C), and the G
values were plotted in logarithmic scale. This test
makes it possible to determine the linear viscoelastic
regime of the sample and therefore to choose the
stress value for the other oscillation tests.
Temperature sweep. This test was performed to charac
terize sample behavior at constant frequency and stress in
a range of temperatures. The experimental parameters
were 1 Hz frequency, 1 Pa stress, a temperature range of
10 60C and a heating rate of 1C/min.
Frequency sweep. The sample was exposed to stepwise
increases of frequency (0.01 50 Hz) at constant stress
(1 Pa) in the eld of linear viscoelasticity, and the
average values of G, G and were calculated at 10,
1, and 0.02 Hz. The frequency range, the G, and the
G values were plotted in logarithmic scale.
Creep/recovery. This test was carried out at 20C at
different values of stress, depending on the system
studied. The selected stress was kept constant for
100 s, then instantly removed, followed by a 200 s
recovery. The creep compliance JC (dened as the
ratio between the measured strain and the applied
stress) was monitored against time. The test was also
used to calculate the viscosity of the sample from the
linear stress/strain region of the retardation curve.
The rheological behavior of emulsions was also studied
by performing the following test:
Shear test + time sweep. This test was performed by
applying a constant stress (200 Pa) for 5 min (shear
test) and then observing the sample recovery for
90 min in the oscillation mode at 1 Pa of stress and
1 Hz of frequency at the temperature of 20C (time
sweep test).
Emulsion stability was checked 1 and 3 months after
preparation, using stress sweep and creep recovery tests and
shear analysis followed by time sweep.
Acoustic Spectroscopy Measurements
The cell of the DT 1200 acoustic and electroacoustic
spectrometer (Dispersion Technology, USA) was lled with
15 ml of Sepineo P 600 gel or the emulsion, and the sound

attenuation and speed were monitored. Analyses were
performed at the gap interval of 0.325 20 mm, in the
frequency range of 3 100 MHz at 20C. Particle size and
rheological G and G moduli were calculated (15). The
possible variation of droplet size and modication of the
microrheological parameters over time were adopted as
stability criteria of the semisolid systems.

Rheological Analysis of the Gels
The rheological behavior of the gels was strictly related
to polymer concentration. Stress sweep results and viscosity
values obtained from the creep recovery test showed
increasing elasticity in the samples, as the Sepineo concen
tration was increased from 0.5% to 3%, as demonstrated by
the values of elastic modulus G and viscosity (Table I). This
behavior was also conrmed by the temperature sweep test,
which showed higher and very similar values of G modulus
for the 3% and 5% samples. This test also showed that there
was no signicant variation in G or G moduli at increasing
temperatures (Fig. 1).
From the frequency sweep analyses, it was possible to
observe that the 3 5% (w/w) concentration samples showed
typical gel like spectra, characterized by the absence of
crossover points between the dynamic moduli and by weak
dependence on frequency (Fig. 2, Table II). In contrast, as
expected, the 0.5% and 1% (w/w) systems presented different
viscoelastic behavior with mechanical spectra characterized
by stronger frequency dependence and multiple crossover
points. These considerations were conrmed by the slope
values of the G G curves versus frequency (Table III). It is
known that the rheological behavior of a polymer in
dispersion can show a modication in the slope values of
the G G versus frequency curves at increasing concentra
tions or molecular weights. In this case, given the strong
frequency dependence, the Zimm or Rouse models seemed
the best tting (17). These models concern the viscoelastic
behavior of an isolated chain that relaxes independently of
the presence of the others and, for this reason, are usually
employed for dilute polymer solutions. However, in this
study, they failed to provide correctly tting results, even for
the 0.5% concentration, perhaps because of the co presence
of an entanglement network at this concentration. Certainly,
at low polymer concentrations, it is possible to assume the
Table I. G and G Obtained from the Stress Sweep Tests and Calculated at the Stress of 1 Pa and Viscosity Values Obtained from the Creep
Recovery Tests of the 0.5 5% Sepineo Gels and of the Sepineo 3%/Almond Oil Emulsion at Different Storage Times
Gels
Emulsion
0.5%
1%
3%
5%
1 day
1 month
3 months
G (Pa)
SD
G (Pa)
SD
Viscosity (Pa s)
SD
0.630
6.670
615.7
665.7
619.6
611.4
608.6
0.03
3.17
16.1
44.4
7.70
4.00
6.50
1.100
4.620
90.67
94.73
68.20
67.80
60.60
0.03
0.84
1.98
11.2
12.5
8.31
5.76
0.4265
47.180
84,414
112,647
92,346
92,123
92,011
0.045
2.540
2,740
6,215
3,421
3,836
4,971
371
Fig. 1. Temperature sweep test of Sepineo P 600 systems (0.5 5% w/w)
existence of topological constraints and thus to imagine that

there are entanglement sites, rather than specic interactions
between polymeric chains. More in depth analysis of the
frequency sweep plot of the 0.5% sample revealed that it
could easily be tted with the Doi Edwards model (17,18)
(Fig. 3), which assumes that the motion of the chain is
conned to a tube like region formed by the surrounding
polymer molecules. Running along the center of the tube is a
primitive chain that represents the shortest path down the
tube, and every deviation from this path is considered a
defect. The motion of these defects allows the chain to move
along the tube with a reptilian motion (reptation). The
plateau modulus, which refers to the plateau observed in
the plot of G versus frequency at intermediate frequencies
and is related to the relaxation modulus, can be calculated
using this type of modeling. In addition, it was possible to
obtain the characteristic system relaxation time (0.5185 s.),
the zero shear viscosity (2.303 Pa s), and the related steady
state compliance from the values of the plateau modulus and
of the system relaxation time. The low value of zero shear

viscosity and plateau modulus (5.338 Pa) obtained from the
tting, and the fact that this model is usually applied to
semidilute and concentrated solutions demonstrated that the
0.5% systems did not possess an organized structure. In fact,
the Doi Edwards theory adequately describes the dynamic
behavior of ideally exible polymers in concentrated non
entangled solution, where the polymer chains do not possess
permanent cross links, and the chain is free to move along the
tube.
Slightly different considerations can be made for the 1%
concentration sample. In fact, the modeling of its frequency
spectra showed better correlation with the tube linear
dilution theory (17,19,20).
This theory considers the shortcomings of reptation
theory applied to linear chains, treating them as two armed
stars (21 23). In this model, relaxation phenomena are
related to arm retraction. This means that the retraction
motion of the arm (for each star) is very similar to the
Fig. 2. Frequency sweep test of Sepineo P 600 systems (0.5 5% w/w)
372
Table II. G and G from the Frequency Sweep of Sepineo P 600

Gels (0.5%, 3% and 5% w/w) Calculated at Frequencies of 10, 1, and
0.02 Hz
% (w/w)
Frequency (Hz)
G (Pa)
SD
G (Pa)
SD
0.5
10
1
0.02
10
1
0.02
10
1
0.02
10
1
0.02
1.830
1.257
0.001
11.26
9.130
2.680
760.0
650.7
530.0
811.3
686.6
545.0
0.93
0.07
0.00
2.84
1.46
1.40
7.21
8.62
4.00
16.8
13.0
13.1
2.320
1.103
0.061
11.06
4.240
2.000
123.7
78.87
62.73
147.0
90.40
73.30
0.26
0.05
0.00
0.92
0.21
0.13
5.51
5.80
6.40
1.73
0.95
1.97
uctuation motion of the contour along the linear chain

(uctuation driven stretching and contraction of the chain
along the tube). It should be taken into account that, in linear
chains, the tube segments, which do not have retracting ends,
will be relaxed by reptation. For these linear chains, the fast
retractions will involve motion near the chain end, and thus,
this model must consider the center of the chains as xed.
Application of this model and calculation of the relative
spectra revealed the increasing elasticity of this system
compared to the 0.5% concentration, as conrmed by the
higher values of the plateau modulus (491.6 Pa), shear
viscosity (870.6 Pa s), and characteristic relaxation time
(170.6 s.). This means that at 1% concentration, Sepineo is
present in dispersion in a more organized structure, which
implies more complex mechanisms of relaxation and higher
polymer polymer or polymer solvent interactions despite the
absence of a three dimensional network (Fig. 4).
Concentrations from 3% to 5% showed very different
mechanical spectra. As conrmed by the slope values very
close to 0, these systems can be considered as gels from a
rheological point of view, and in this case, the spectra were
typical of a solid like sample with a three dimensional
network structure.
Once again, these results conrmed that the rheological
behavior depended strongly on polymer concentration. In
creased concentrations from 0.5% to 5% gave rise to a
progressive change in system mechanical spectra, which could
lead one to think that increasing chain interactions and
formation of entanglements foster the development of
connectivity across the entire system until a network is
formed.
Rheological Analysis of the Emulsions
Based on the rheological characterization of the different
gelled systems, the 3% (w/w) concentration was selected as
the external dispersing phase for the preparation of the
emulsions. In fact, the 0.5% and 1% samples were not stiff
enough, while the considerable elasticity of the 5% system
could compromise correct emulsion formation. As mentioned
in the MATERIALS AND METHODS, almond oil was
added in order to assess the inuence of an oily phase on the

rheology of the dispersing phase and the stability of the
emulsions themselves.
The stress sweep test was used to follow emulsion
stability because the elastic modulus slowly decreases when
the dispersed droplets become bigger but less numerous
(24 27).
The stress sweep analyses performed 1 day after the
preparation of these emulsions, and then at 1 and 3 months,
showed good stability (Table I). It is important to note that
the dynamic moduli values were just slightly higher than
those found for the corresponding Sepineo gel, proving that
the addition of the oily phase did not affect Sepineo
rheological behavior too much (Table I).
The time sweep performed to check sample recovery
after shear showed that emulsions were unable to recover
their structure during the test, as conrmed by the smaller
moduli values (data not shown). This indicated a degree of
sample deformability after shear, which can improve system
spreading and absorption, making it easy to rub the sample
into the skin after topical administration (28,29).
The frequency sweep tests (Fig. 5), performed 1 day after
preparation, conrmed the behavior previously observed
from the stress sweep: The almond oily phase did not
substantially affect the mechanical spectrum of the Sepineo
aqueous phase, showing a very similar dependence on
frequency for both moduli as revealed by the slope values
in Table III.
Furthermore, the emulsion behavior with temperature
(data not shown) conrmed that the systems were not
affected by temperature, as already observed for the Sepineo
gel.
Acoustic Spectroscopy Analysis of the Gels

Acoustic spectroscopy afforded a microrheological char
acterization of the gel samples. It is important to note that the
dynamic moduli obtained from acoustic spectroscopy meas
urements are not comparable with those derived from
rotational rheometers working in the oscillation mode, since
the applied stress is not tangential, as in an oscillatory
experiment, but normal, and the tested frequencies are
much higher.
In any case, analysis of the rheological parameters and,
in particular, of the G modulus (Fig. 6), which is the
frequency dependent modulus, showed some differences
within the various systems, in agreement with the oscillatory
results. Basically, the spectra of the 0.5% and 1% Sepineo
Table III. Slope Values Calculated by the Linear Fitting of the G

and G Average Curves Obtained from the Frequency Sweep of the
0.5 5% Sepineo Gels and of the Sepineo 3%/Almond Oil Emulsion
0.5% gel
1% gel
3% gel
5% gel
Emulsion
Slopes G
Slopes G
1.568
0.243
0.056
0.060
0.057
0.696
0.272
0.115
0.102
0.120
373
progressive loss of the network structure made the viscous
contributions more important, giving rise to the remarkable
increase of the G modulus.
On the other hand, the G modulus did not show
signicant differences among the various systems, even
though the 3% (w/w) and particularly the 5% concentrations
showed slightly higher values for this modulus (Fig. 6). In this
case, the increase in the density and sound speed of the
systems at greater polymer concentration could explain the
results obtained (see Eq. 1).
Acoustic Spectroscopy Analysis of the Emulsions
Fig. 3. Frequency sweep test of Sepineo P 600 systems 0.5% (w/w)

with Doi Edwards spectra tting (continuous line)
samples were identical and characterized by a linear depen

dence on frequency. On the other hand, this linear trend was
followed by a sharp modulus increase at lower frequency
when the Sepineo concentration rose to 3% and 5% (w/w).
The behavior of the latter two concentrations may seem
unusual compared to the classical rheological trend, due to
the fact that during the tests, the polymer chains showed
signicant modication in their deformability. This greater
deformability at low frequency values indicates that the
polymer/polymer interactions responsible for the gel like
behavior were not particularly strong. These interactions
showed short lasting elastic properties that probably failed
when the system was stressed for a longer time. This change
in the system structural characteristics modied the mecha
nisms of interaction between the sound and the samples (see
MATERIALS AND METHOD) and consequently
changed the resultant attenuation parameter. The structural
contribution, which is characteristic of gel like systems, was
dominant at higher frequencies. At low frequencies, the
Fig. 4. Frequency sweep test of Sepineo P 600 systems 1% (w/w) with

tube dilation spectra tting (continuous line)
Ultrasound technique was used to monitor emulsion

microrheological behavior and stability by considering the
variation of both their rheological parameters and particle
size (such as mean diameter).
The frequency spectra used to analyze system stability
over time (Fig. 7) conrmed the stability of the emulsion. In
fact, the G and G plots obtained after 1 day, 1 month, and
3 months were very similar except for the G modulus at low
frequencies. These results are in agreement with the oscilla
tory analysis results that revealed good stability for the
emulsion.
It is interesting to compare the G spectra between the
gel and the emulsion. As can be seen from Figs. 6 and 7, the
emulsion system in the high frequency range exhibited quite
similar behavior, with slightly greater G values compared to
the gel. The presence of the oily phase caused greater
frequency dependence, which can be explained by a more
organized structure and consequently greater initial energy
dissipation during the frequency sweep. As mentioned in the
introduction, the oil in the emulsion gel formation mainly
served as an interface able to interact with the gel matrix,
providing general improvement in the system rheological
characteristics.
Instead, a different behavior was present in the low
frequency region, where the curve related to the system
analyzed after 1 day revealed a trend very similar to that of
the corresponding Sepineo sample (3% w/w concentration),
which showed, as previously mentioned, a certain deform
ability. After 1 and, in particular, after 3 months, this
deformability gradually decreased, indicating a progressive
interfacial stabilization of the emulsion. On the other hand,
the G modulus can be considered practically identical
between gel and emulsion (Figs. 6 and 7).
Fig. 5. Frequency sweep of Sepineo P 600/almond oil emulsion
374
Fig. 6. Mean curves (standard deviation bars are omitted to avoid overlapping) of G and G
moduli from the acoustic spectroscopy of the different Sepineo P 600 concentration samples
(0.5 5% w/w)
The results obtained from the analysis of the particle size

(mean diameter values in micrometer) conrmed the previ
ous statements concerning the stability of the emulsion. Mean
diameter values (micrometer) of the dispersed almond oil
droplets at the storage time of 1 day, 1 month, and 3 months
were, respectively, 1.5670.14, 1.5920.22, and 1.5780.25.
CONCLUSION
Oscillatory rheology and acoustic spectroscopy analyses
were utilized in conjunction in order to characterize the Sepineo
gel systems as well as to study how the addition of oil affects gel
characteristics and the nal stability of the resultant emulsion.
Both techniques revealed that Sepineo P 600 thickens and gels
well, a property that depends strongly on polymer concentra

tion. Concentration increases from 0.5% (w/w) to 5% (w/w)
modied the viscoelastic properties of the Sepineo samples,
changing the typical behavior of a concentrated non entangled
solution to that of a gel like sample. On the other hand, the
microrheological parameters obtained from acoustic spectros
copy showed that the physical interactions forming this gel like
structure were not particularly strong.
Concerning the emulsions, the most important result is
surely the fact that the addition of an oily phase increased
system consistency only minimally. The viscoelastic character
istics depended exclusively on the gel structure.
In conclusion, Sepineo P 600 is a prime candidate for use
in the formulation of gels and emulsion gels with rheological
properties suitable for topical administration.
Fig. 7. Mean curves of G and G moduli from the acoustic spectroscopy of Sepineo/
almond oil emulsion

REFERENCES
1. Reiffers Magnani CK, Cuq JL, Watzke HJ. Composite structure
formation in whey protein stabilized O/W emulsions. I. Inuence
of the dispersed phase on the viscoelastic properties. Food
Hydrocoll 1997;13:303 16.
2. Chen J, Dickinson E, Langton M, Hermansson AM. Mechanical
properties and microstructure of heat set whey protein emulsion
gels: effect of emulsiers. Lebensm Wiss U Technol 2000;33:299
307.
3. Dickinson E, Merino LM. Effect of sugars on the rheological
properties of acid caseinate stabilized emulsion gel. Food Hydro
coll 2002;16:321 33.
4. Chen J, Dickinson E. Viscoelastic properties of protein stabilized
emulsions: effects of protein surfactant interactions. J Agric
Food Chem 1998;46:91 7.
5. Dickinson E. Milk protein interfacial layers and the relationship
to emulsion stability and rheology. Colloids Surf B Biointerfaces
2001;20:197 210.
6. Kerstens S, Murray BS, Dickinson E. Microstructure of b
lactoglobulin stabilized emulsions containing non ionic surfac
tant and excess free protein: Inuence of heating. J Colloid
Interface Sci 2006;296:332 41.
7. Akiyama E, Kashimoto A, Fukuda K, Hotta H, Suzuki T,
Kitsuki T. Thickening properties and emulsication mechanisms
of new derivatives of polysaccharides in aqueous solution. J
Colloid Interface Sci 2005;282:448 57.
8. Landoll LM. Nonionic polymer surfactants. J Polym Sci Polym
Chem Ed 1982;20:443 55.
9. Langley KR. Functional aspects of particle matrix interactions in
composite foods. Food Qual Prefer 1990;2:111 5.
10. Chen J, Dickinson E. Effect of surface character of ller particles
on rheology of heat set protein emulsion gel. Colloids Surf B
Biointerfaces 1999;12:373 81.
11. Van Vliet T. Rheological properties of lled gels. Inuence of
ller matrix interactions. Colloid Polym Sci 1998;266:518 24.
12. Dickinson E, Hong ST. Inuence of water soluble non ionic
emulsier on the rheology of heat set protein stabilized emulsion
gels. J Agric Food Chem 1995;43:2560 6.
13. McClements DJ, Monahan FJ, Kinsella JE. Effect of emulsion
droplets on the rheology of whey protein isolate gels. J Texture
Stud 1993;24:411 22.
14. Seppic brochure. Sepineo P600 an emulsifying/thickening poly
mer for new sensations. http://www.seppic.com
375
15. Dukhin AS, Goetz PJ. Ultrasound for characterizing colloids.
Particle sizing, zeta potential, rheology. The Netherlands:
Elsevier; 2002. p. 75 144.
16. Litovitz TA, Davis CM. Structural and shear relaxation in
liquids. In: Mason WP, editor. Physical acoustics. New York:
Academic; 1964. p. 285 90.
17. Goodwin JW, Hughes RW. Rheology for chemists. Cambridge:
The Royal Society of Chemistry; 2000. p. 146 212.
18. Doy M, Edwards SF. The theory of polymer dynamics. Oxford:
Oxford University Press; 1986.
19. De Gennes PG. Scaling concepts in polymer physics. Ithaca, NY:
Cornell University Press; 1979.
20. Ebert U, Schfer L, Baumgrtner A. Segment motion in the
reptation model of polymer dynamics. I. Analytical investigation.
J Stat Phys 1998;90:1325 73.
21. Milner ST, McLeish TCB. Reptation and contour length
uctuations in melts of linear polymers. Phys Rev Lett
1998;81:725 8.
22. Milner ST, McLeish TCB. Arm length dependence of stress
relaxation in star polymer melts. Macromolecules 1998;31:7479
82.
23. Blottire B, McLeish TCB, Hakiki A, Young RN, Milner ST.
Rheology and tube model theory of bimodal blends of star
polymer melts. Macromolecules 1998;31:9295 304.
24. Tadros T. Application of rheology for assessment and prediction
of long term physical stability of emulsions. Adv Colloid
Interface Sci 2004;108 109:227 58.
25. Korhonen M, Hellen L, Hirvonen J, Yliruusi J. Rheological
properties of creams with four different surfactant combinations
effect of storage time and conditions. Int J Pharm 2001;221:187 96.
26. Chanamai R, McClements DJ. Dependence of creaming and
rheology of monodisperse oil in water emulsions on droplet size
and concentration. Colloids Surf A Physicochem Eng Asp
2000;172:79 86.
27. Masmoudi H, Piccerelle P, Le Drau Y, Kister J. A rheological
method to evaluate the physical stability of highly viscous
pharmaceutical oil in water emulsions. Pharm Res 2006;23
8:1937 47.
28. Jones DB, Woolfson AD, Brown AF. Textural, viscoelastic and
mucoadhesive properties of pharmaceutical gels composed of
cellulose polymers. Int J Pharm 1997;151:223 33.
29. Adeyeye MC, Jain AC, Ghorab MKM, Reilly WJ Jr. Viscoelastic
evaluation of topical creams containing microcrystalline cellu
lose/sodium carboxymethyl cellulose as stabilizer. AAPS
PharmSciTech 2002;3 2:1 10.

DOI: 10.1208/s12249 009 9216 3
Research Article
The Study on the Entrapment Efficiency and In Vitro Release of Puerarin
Submicron Emulsion
Peng-Fei Yue,1,2,5 Xiu-Yun Lu,1 Zeng-Zhu Zhang,3 Hai-Long Yuan,4 Wei-Feng Zhu,1
Qin Zheng,2 and Ming Yang2
Received 17 July 2008; accepted 11 February 2009; published online 21 April 2009
Abstract. The entrapment efciency (EE) and release in vitro are very important physicochemical
characteristics of puerarin submicron emulsion (SME). In this paper, the performance of ultraltration
(UF), ultracentrifugation (UC), and microdialysis (MD) for determining the EE of SME were evaluated,
respectively. The release study in vitro of puerarin from SME was studied by using MD and pressure UF
technology. The EE of SME was 86.5%, 72.8%, and 55.8% as determined by MD, UF, and UC,
respectively. MD was not suitable for EE measurements of puerarin submicron oil droplet, which could
only determine the total EE of submicron oil droplet and liposomes micelles, but it could be applied to
determine the amount of free drug in SMEs. Although UC was the fastest and simplest to use, its results
were the least reliable. UF was still the relatively accurate method for EE determination of puerarin
SME. The release of puerarin SME could be evaluated by using MD and pressure UF, but MD seemed to
be more suitable for the release study of puerarin emulsion. The drug release from puerarin SME at
three drug concentrations was initially rapid, but reached a plateau value within 30 min. Drug release of
puerarin from the SME occurred via burst release.
KEY WORDS: drug release; entrapment efciency; microdialysis; pressure ultraltration technology;
submicron emulsion.
INTRODUCTION
Submicron emulsion (SME), also referred to as lipid
emulsion or lipid microsphere, is a potentially interesting
drug delivery system (1). It can reduce drug hydrolysis and
increase drug bioavailability (2). The SME system has gained
increasing importance for the administration of those drugs
which are poorly soluble in water and oils and simultaneously
toxiferous for intravenous injection. The SME system also
enhances the activity and bioavailability of these drugs (3,4).
Reports on the ability of these systems to enhance the drug
solubility and efcacy and to reduce side effects by incorpo
rating the drug into the lipophilic core (of the oil droplets or
in the core of micelles) or in the interface have appeared in
the literature (5 8).
The entrapment efciency (EE) is one of the most
important physicochemical characteristics of puerarin SME.
Accordingly, it is of crucial importance to accurately quantify
1
Jiangxi University of Traditional Chinese Medicine, Nanchang

330004, China.
2
Key Laboratory of Modern Preparation of TCM, Ministry of
Education, Nanchang 330004, China.
3
94 Hospital of PLA&PLA, Nanchang 330000, China.
4
302 Hospital of PLA&PLA Institute of Chinese Materia Medica,
Beijing 100039, China.
5
To whom correspondence should be addressed. (e mail: ypfpharm@
126.com)
the EE of SME. Many methods have been used to evaluate

the EE of SME, including dialysis bag diffusion, gel ltration,
ultraltration (UF), and ultracentrifugation (UC) (9 12). The
application of UF to the separation of free drug from SME
has been extensively examined in a variety of practical
applications (11,13). Among these methods, UF is the most
widely used technique for determining the EE of nano
particles, including SME.
Recently, the microdialysis (MD) technique has also
been used to determine the EE of nanocapsules, nano
spheres, and nanoemulsions (14). As far as nanoparticle
loading of drugs is concerned, free drug would diffuse into the
probe because there is a concentration gradient of free drug
from the outside to the inside of the MD ber. The molecular
weight cutoff of the MD membrane is such that nano
structures and, therefore, the incorporated drug cannot cross
the membrane. Convection continually renews the solution
around the probe, keeping the concentrations of all compo
nents outside the probe at the bulk solution concentration.
Because little drug is actually removed from the sample, the
overall free drug concentration remains essentially constant
during the experiment. Also, because MD does not change
the uid volume, the total drug concentration remains
constant. Therefore, the equilibrium of entrapped and free
drug is not disturbed by this technique.
In addition, the UC technique has also been used to
determine the EE of SME because it can be used to separate
oil droplets and most of the aqueous phase from the emulsion
376
Entrapment Efficiency and In Vitro Release of Puerarin SME

(15). It is based on the different sedimentation velocities of
mixtures with different densities. After an initial cream
layer formation, where most of the oil droplets are
centrifuged to the top of the tube and compressed, the
water phase of free oil is observed. But considerable
breakdown of the SME is achieved during prolonged
centrifugation, the redistribution of drug molecules will
probably occur, which calls into question the reliability of
the EE evaluation by means of UC.
Puerarin, a naturally occurring isoavone C glycoside
(Fig. 1), was isolated from Pueraria lobata, one of the most
popular traditional Chinese herbal medicines. It was tradi
tionally used to reduce febrile symptoms, decrease myocar
dial consumption of oxygen, increase coronary artery blood
ow, and improve microcirculation (16 18). Due to poor oral
bioavailability, the commercial product puerarin i.v. was
widely used for the treatment of coronary artery disease,
ischemic heart disease, cerebrovascular disease, coronarism,
cerebral angiospasm, cerebral infarction, and cerebral throm
bosis in clinic (19,20). But the clinical efcacy of puerarin i.
v. was limited by severe and acute toxic side effects, such
as intravenous hemolysis (21,22). In the present study, a
new formulation, puerarin SME, has been developed.
Puerarin was incorporated into the SME, which could be
responsible for reducing the hemolysis side effect (23). As
shown in Fig. 2, there could be drug molecules in the
aqueous phase, the oil water phase, and the oil phase,
respectively. Puerarin with aqueous solubility of about
3.46 mg/mL was poorly soluble in oil (24). There was
insufcient loading capability in the aqueous and oil phases
simultaneously, most of the drugs might be present in the
interfacial surface where the molecules of the surfactant
were arranged in order between the aqueous and oil
interface with puerarin molecules located between them.
So, it was very necessary to evaluate the EE and in vitro
release character of puerarin SME.
The main objectives of this study were: (1) to evaluate
MD and UC as techniques for determining the EE of
puerarin SME, UF as reference technique to determine the
EE of SME; (2) to evaluate the release character in vitro
of puerarin SME. In vitro release studies were performed
using by MD technology, and pressure UF technology as
377
Fig. 2. The proposed structure of puerarin SME
reference technique to determine the in vitro release of

puerarin.
Materials
Puerarin was obtained from Xian guochui, purity 99.5%
(Xian, China). Egg phospholipids was purchased from Hua
qing mei hen, purity 82% (Beijing, China). Puried soybean
oil for parenteral use (Tieling BeiYa Pharmaceutical, Tieling,
China), synperonic F68 (BASF AG, Germany). All other
chemicals and reagents were of analytical or chromatographic
grade.
The MD probes were U shaped, made of hollow
cellulose bers (DM 22, 200 m inner diameter and 220 m
outer diameter, with a cutoff at a nominal molecular weight of
5,000 Da; Eicom, Japan), and the active region of the MD
probe was 10 mm in length.
Preparation of Puerarin Submicron Emulsion
The basic formula of the SME was: soybean oil, 12 g; egg
phospholipids, 1.2 g; synperonic F68, 0.1 g; glycerol, 2.5 g;
tocopherol, 300 mg; puerarin, 1 g; double distilled water,
90.0 g.
The preparation of the puerarin SMEs involved six steps,
as follows:
Fig. 1. The chemical structure of puerarin
(a) Preparation of the puerarin phospholipids complexes:

Weighed amount of puerarin and phospholipid at a
weight ratio of 1:1.2 were placed in a 100 mL round
bottom ask and 60 mL of absolute alcohol was added.
The mixture was reuxed at a temperature not exceeding
60C for 3 h. The resultant clear solution was dried at 40
C under vacuum to remove traces of solvents in order to
obtain the puerarin phospholipid complexes.
(b) Preparation of the lipid phase: The puerarin phospholipid
complexes and tocopherol were dissolved in soybean
oil at 55C.
(c) Preparation of the water phase: The water, synperonic
F68, and glycerol were mixed at 55C.
Yue et al.
378
(d) Preparation of the coarse emulsion: The lipid phase was
stirred at 55C at 2,000 rpm by high speed stirrer, and the
water phase was slowly injected into the lipid phase to
obtain coarse emulsion.
(e) Homogenization: A ne emulsion was prepared by
passing the coarse emulsion through a high pressure
homogenizer (GYB40 10S, Donghua Homogenization
Apparatus, China).Homogenization conditions were typ
ically 80 120 MPa and ve to 20 cycles at 25C.
Afterwards, the pH was adjusted to 6 7 with 0.1 N
sodium hydroxide solution.
(f) Sterilization: The emulsion was packed in 15 mL sterile
glass vials under nitrogen. The vials were sealed and the
emulsion was sterilized by autoclaving (autoclaving ster
ilization cabinet, YZM 03BS, Haoersheng Sterilization
Apparatus, China) at 121C for 15 min.
Particle Size of SME

Measurements of the particle size and mean diameter
of the emulsion were performed using a Malvern Zetasizer
(Malvern Instruments, Malvern, UK). The obtained distri
bution was a volume distribution; as characterization
parameters, the LD diameters 50%, 90%, 95%, and 99%
were calculated. For example, a diameter 90% (D90) meant
that 90% of the volume of the particles was below the
given size in nanometers. The width of the particle size
distribution was expressed as the span of dispersity (SD)
(25):
SD
D90 D10
:
D50
The mobile phase consisted of a mixture of methanol

H2O (30:70, v/v) delivered at a ow rate of 1.0 mL/min. The
injection volume was 20 L. UV detection was performed at
250 nm at room temperature.
The Entrapment Efficiency Measurement
Microdialysis Experiment
The relative recovery was determined by immersing the
MD probes in the stirred puerarin standard solution (etha
nol water, 2:8), which contained different concentrations of
puerarin (0.5, 2.0, and 4.0 mg/mL) as the MD medium (CMM).
The probes were perfused with drug free solution at a ow
rate of 4 L/min and the temperature was kept at room
temperature. The concentration of puerarin in the micro
dialysate samples was determined using high performance
liquid chromatography (HPLC) (CMS). The relative recovery
(RR) of puerarin was calculated using the following equation
(26):
RR
CMM
:
CMS
For the EE determination, the MD probe was inserted

into glass vials containing puerarin SME at room tempera
ture. The ow rate was set at 4 L/min 1. After a 30 min
equilibration period, the dialysate from the SME was
collected at 20 min intervals and analyzed by HPLC. The
total puerarin concentration was also measured by HPLC
after dissolution of the puerarin SME. The EE was calculated
according to the following equation:
EE%

1

CMS =RR
r 100%
CT
Light Microscopy Analysis

Light microscopy was performed by a BH 2 microscope
(Chongqing, China). The magnication selected was 1,500
fold, oil immersion applied. Emulsion was undiluted to
analyze. Typically, 20 microscopic elds were analyzed under
polarized light for the detection of remaining drug crystals.
Liposomes Micelles Size Measurement

The mean diameter and size distribution of small
liposomes micelles in the aqueous layer after UC were
also examined by the dynamic light scattering technique
with the NicompTM 380 Particle Sizing System (Santa
Barbara, USA).
Conditions of High-Performance Liquid Chromatography

A HP 1100 chromatographic system consisting of a
quaternary pump (G1100A Quat Pump, Agilent), degasser,
diode array detector (G1100A DAD, Agilent), and HP
Chemstation Data system (Agilent Technologies, Palo Alto,
CA, USA) was used. Chromatographic column Spherisorb
ODS C18 (2504.6 mm, 5 m) was used for chromatographic
separation.
where CT was the total concentration of puerarin and r was

the ratio of the aqueous phase volume of the SME and the
total volume.
Ultrafiltration Test
The free puerarin was determined in the ultraltrate
after separation of the SME by UF and centrifugation
through VIVASPIN 4 lters (molecular weight cutoff
10 kDa). There was nonspecic adsorption of the drug
to the UF membrane. Therefore, it was necessary to
determine the recovery (R) by UF and centrifugation of
the puerarin standard solution (ethanol water, 2:8) con
taining different concentrations of puerarin (6, 8, and
10 mg/mL). The recovery was calculated by the following
equation:
R0
CF
CT
where CT was the concentration of puerarin and CF was

the concentration of puerarin in the ultraltrate.
For EE determination, 4 mL puerarin SME was added
into the VIVASPIN 4 and centrifuged for 1 h at 3,000 rpm.
The concentration of puerarin entrapped in SME was
calculated from the difference between the total and free

drug concentrations, measured in the SME and in the
ultraltrate. The EE was calculated as follows:
EE%

1

CF =R0
r 100%:
CT
Ultracentrifugation Test
About 3 ml of SME was ultracentrifuged in a Hitachi
ultracentrifuge (CS120GXL) at 162,000g at 4C for 1 h. The
EE of the SME was determined by measuring the amount of
puerarin (CW) in the water layer obtained after UC. The EE
was calculated according to the following equation:

EE% 1

CW
r 100%:
CT
The Release Characterization In Vitro of Puerarin

Submicron Emulsion
Pressure Ultrafiltration Technology Test
An Amicon 8050 pressure UF cell tted with a Millipore
YM10 membrane was used (both obtained from Millipore).
That 49 mL of release medium (ethanol pH 7.4 phosphate
buffer, 2:8), was placed into the Amicon cell, and 1 mL of the
SME was injected into the cell while stirring (t=0). At
predetermined time points, pressure was applied such that
the ow rate of the ultraltrate was approximately 1 mL/min
and sample was collected (after discarding the appropriate
amount of solution). An equivalent amount of phosphate
buffer was added to the release medium at each time
point to maintain the total volume at 50 mL. A sample of
the initial dispersion was used to calculate the extent of
drug release. All operations were conducted at ambient
temperature (222C).
Microdialysis Test
For the in vitro release study, the MD probe was inserted
into the receptor compartment containing dialysis medium
(ethanol pH 7.4 phosphate buffer, 2:8; 49 mL) at ambient
temperature (222C). The ow rate was set at 4 L/min.
After a 30 min equilibration period, puerarin SME (1 mL)
was placed into the receptor compartment. The other
dissolution conditions were equivalent to UF. The dialysate
from the SME were collected at predetermined time intervals
and analyzed by HPLC.
379
factor of 1,500. The homogenization was considered to be the
crucial step that affected the particle size of the emulsion.
High pressure emulsication produced a more rapid reduc
tion in particle size. The coarse emulsion was homogenized
for 15 homogenization cycles applying individually 50 and
80 MPa at 25C, samples were collected after two, four, six,
eight, ten 12, and 15 cycles, respectively. Then, the mean
droplet size, droplet size distribution, and presence drug
crystal were determined. There was a slight decrease in mean
particle size with increasing homogenized pressure and cycle
times. The nal particle sizes were lowest (188.14 nm; see
Fig. 3) with no overprocessing on applying the highest
pressure (80 MPa, 15 cycles). However, puerarin emulsion
yielded at 80 MPa had a slightly broad distribution than
others. Figure 3 showed the mean diameters of the emulsion.
The data showed that just 50 MPa for eight to ten cycles
appeared to be sufcient to yield the ne emulsion with small
particle size and narrow distribution range.
To apply high pressure homogenization technology, the
small particle size (approximately 100 200 nm) of the
emulsied oil droplets in the emulsion was obtained, which
resulted in a sizeable interfacial surface area. There would be
sufcient loading capability in the interfacial surface or oil
droplet for puerarin.
Light Microscopy Analysis

PCS covered the size range of about 5 nm 3 m, but it
was limited to the detection of particles undergoing Brownian
motion, so those particles larger than approximately 3 m were
undetectable. In addition, it was difcult to identify non
incorporated drug particles in an ivory white emulsion and it
could not differentiate between similarly sized droplets and
drug crystals (28). Accordingly, light microscopy was applied to
examine puerarin emulsion with concentrations of 10 mg/mL;
it was investigated whether there was the solid drug particles in
the emulsion. In order to increase the probability of detecting
the presence of even only a few solid drug particles, the
puerarin emulsion was not diluted. Figure 4 showed that no
drug crystals were found in the emulsion.

Preparation and Size Distribution of SME
In this study, puerarin was poorly soluble in soybean oil
and slightly soluble in water. Puerarin phospholipid com
plexes were prepared in order to improve the lipophilicity of
puerarin in the oil (27). A concentration of 10 mg/mL
puerarin in the emulsion was chosen for the preparation of
the drug loaded emulsion. No drug crystals were detected in
the puerarin emulsion by microscope with an enlargement
Fig. 3. The average diameter and SD of the homogenized SME as a

function of cycle numbers at 50 MPa. Data was expressed as meanSD
(n 3)
Yue et al.
380
Table I. MD Probe and UF Recoveries
MD
UF
Puerarin (mg/mL)
RR (%)
Puerarin (mg/mL)
R (%)
0.5
2.0
4.0
Mean values
36.12.13
35.71.81
36.62.21
36.10.45
6
8
10
Mean values
96.22.07
97.11.64
97.91.13
97.10.85
MD microdialysis
UF ultraltration
Fig. 4. Light microscopy analysis of emulsion load with 10 mg/mL

drug
means of UC. At the end of the UC, the slightly turbid

aqueous layer was collected for the particle size measure
ment. The mean particle diameter (n=3) was 49.35.46 nm,
which was similar to those obtained for the small liposomes
micelles dispersions. Similar vesicle sizes for liposomes
micelles in aqueous phase of SME were consistent with the
previous report (31,32).
Microdialysis and Ultrafiltration Recoveries
Liposomes Micelles Size

The SMEs consisted of oil droplets in the submicron
range together with small liposomes micelles because of an
excess of lecithin (29). The small liposomes micelles had
closed bilayer structures. This result was an assembly in which
the hydrophobic parts of the molecule were shielded from the
aqueous solvent and the hydrophilic head groups were in
maximum contact with them (30). Therefore, liposomes
micelles could encapsulate drug molecules in the aqueous
phase (see Fig. 5). During the preparation of SME, a portion
of the drug molecules in the oil phase passed into the aqueous
phase, but some were free in the aqueous and the rest were
intercalated into the liposomes micelles.
The liposomes micelles had a very low sedimentation
velocity since their densities were only marginally different
from those of the aqueous medium. It would be very difcult
to separate the liposomes micelles from the aqueous phase by
The recovery of the probe depended on many factors.

The perfusion ow rate was an important factor, which
dened the performance of a MD probe and directly
inuenced the recovery by the probe. The ow rate and the
temperature were set as described earlier. At this perfusion
rate, the relative recoveries of puerarin at concentrations of
0.5, 2.0, and 4.0 mg/mL were shown in Table I. It showed that
the ow rate used in this study was in the upper limit of the
range used for MD technique, but the puerarin concentra
tions (0.5, 2.0, and 4.0 mg/mL) had no effect on the relative
recovery.
To calibrate the free puerarin concentrations for non
specic adsorption by the ltration membrane, the UF
recovery was carried out using the puerarin standard solution
at three different concentrations (6, 8, and 10 mg/mL) as
described above. The UF recovery data were shown in
Table I. The recovery results showed that the concentration
of drug adsorbed by the UF membrane was less than 5%.
The drug concentration with nonspecic adsorption for
MD was less compared with that for UF because of the
smaller surface area of the membrane involved. The surface
area of the MD membranes used in these experiments was
much smaller than that of the UF membranes. If the
membranes were similar with regard to the nonspecic
adsorption sites, the UF membrane would bind more drug
than the MD membrane. In addition, the dialysis samples
were collected until a constant value was reached in order to
eliminate any error due to nonspecic adsorption (33).
The Entrapment Efficiency Study of Puerarin Submicron
Emulsion
Fig. 5. The structure of liposomes micelles
Table II showed the EE of puerarin SME determined by

means of the MD, UF, and UC techniques. The EE of SME
was 86.5%, 72.8%, and 55.8% as determined by MD, UF, and
UC, respectively. The results showed that the EE measured
by MD was signicantly different compared with those
measured by UC and UF (p<0.05).
381
Table II. Trapping Efciencies of Puerarin Lipid Microspheres Determined by MD, UF, and UC (MeanSD, n 3)
Preparation (batch)
EE by MD (%)*
EE by UF (%)*
EE by UC (%)*
1
2
3
Mean values
87.51.17
86.72.06
85.41.33
86.51.06
73.30.94
72.41.12
72.61.42
72.80.47
56.92.35
54.23.17
56.32.74
55.81.42
p<0.05, statistically signicant differences among MD, UF, and UC

EE entrapment efciency; MD microdialysis; UF ultraltration; UC ultracentrifugation
It was thought that the drug concentration in micro

dialysate determined by MD was that of free drug in the
aqueous phase. Although the determination principle of the UF
method also involved a molecular weight cutoff, the liposomes
micelles could pass through the lter membrane because of the
compression of centrifugation force. So the amount of drug in
the ltrate was that of the quantity of free drug and associated
with the liposomes micelles in the aqueous phase. It was for this
reason that the EE determined by UF was lower than that
determined by MD.
Additionally, with respect to UC technology, it was very
difcult to separate liposomes micelles from the aqueous
phase, and breakdown of the SME could occur during the UC
process, which would result in the redistribution of the drug
molecules. Some of drug molecules in the O/W interface
would go into aqueous phase, which was responsible for the
nding that the EE of each formulation evaluated by UC was
lower than those determined by the other two methods.
It was concluded that MD was not suitable for EE
measurements of puerarin submicron oil droplet, which could
only determine the total EE of submicron oil droplet and
liposomes micelles, but it could be applied to determine the
amount of free drug in SMEs. There was one drawback to the UC
approach, the redistribution caused by breakdown of the SMEs.
UF was still the relatively accurate method for the EE
determination of puerarin SME because it could separate the
liposomes micelles from the puerarin SME.
The In Vitro Release Study of Puerarin from the Submicron
Emulsion
To measure drug release from colloidal delivery systems,
it was necessary to dilute the dispersion and monitor
subsequent release of drug from the particles into the
surrounding free solution. This was often not recognized in
studies where methods such as equilibrium dialysis were
employed. Consequently, release was often dictated by
membrane transport effects, making it difcult to reconcile
the results obtained in terms of release of drug from the
delivery system. Pressure UF was utilized as the principal
method in this study because it allowed the colloidal
dispersion to be diluted directly in the release medium and
also provided a snapshot of drug distribution between the
colloidal particles and free solution at the time of ltration
(34). But because the puerarin emulsion contained small
particle size liposomes micelles, the liposomes micelles might
pass through the membrane into the ltrate under high
pressure condition. Additionally, the little release medium
was lost in sampling. These might result in the increase of
drug release in vitro of puerarin emulsion. In this study, we
tried to apply the MD technology to study the release in vitro

of puerarin because the MD technology has some advantages
to overcome the drawback of pressure UF. The liposomes
cannot pass though the dialysis membrane. MD sampling did
not change the net uid balance in the surrounding matrix, so
higher temporal resolution could be achieved than with
pressure UF techniques. Also, because there was no net
release medium loss, samples could be collected continuously
for minutes or hours (35). Due to the advantages of MD
described above, the release in vitro of puerarin emulsion was
studied by means of the MD and the pressure UF technology
was used as a reference.
It was shown (Fig. 6) that the release of puerarin
determined by MD was less (p<0.05) at 10 and 20 min
compared with those determined by pressure UF technology.
After 30 min, the release is not signicantly different (p>
0.05) by MD and UF technologies. It could be because of the
following reasons: The liposomes of puerarin emulsion passed
through the UF membrane into the ltrate under high
pressure at 10 and 20 min. Provided that the same amount
of liposomes passed through the membrane at every UF, then
that the amount of drug in the liposomes was decreasing with
time. Following that more drug in the liposomes would pass
through the membrane at 10 min than at 20 min, so the
difference between UF and MD should be greater at 10 min
than at 20 min. However, the puerarin in emulsion had been
mostly released into the release medium after 30 min. Then,
the liposomes that passed through the ltration membrane
into the ltrate did not have a signicant effect on the release
Fig. 6. Release of puerarin from a SME using the pressure UF

method and MD at different levels of drug loading. Drug load in
dispersions prior to dilution were 2.5, 5, and 10 mg/mL. Data was
expressed as meanSD(n 3)
382
because the puerarin in the liposomes had been released
mostly into the release medium.
Drug release from puerarin SME at the three drug
concentrations was initially rapid, but reached a plateau value
within 30 min (Fig. 6). The extent of release in all cases was
around 80%; full release was not attained because the release
medium was not an innite sink. Drug loading did not
inuence the plateau value for the extent of release, which
indicated that the release in these experiments was under
partition control and the results obtained using this method
were not concentration dependent. A possible reason for this
could be that the water soluble surfactants used to prepare the
SME were providing some solubilization of the drug in the
surrounding aqueous solution. This would lead to partitioning
of the drug in favor of the aqueous phase at equilibrium. The
short time taken to achieve 80% release under nonsink
conditions justied the conclusion that these systems were a
burst release vehicle and that, in sink conditions, the drug
release would be even more rapid. Drug release from the SME
has previously been reported to occur via burst release. For
example, release of diazepam (36) and miconazole (35) from a
SME, determined using pressure UF, was found to be very
rapid when sufcient sink condition was utilized.
CONCLUSION
In this study, puerarin SME was prepared. Puerarin
emulsion could be produced with a drug loading as high as
10 mg/mL. Three different approaches were employed to
estimate the EE of puerarin SMEs. The results showed that
MD was not suitable for EE measurements of puerarin SME,
but it could be applied to determine the amount of free drug
in SMEs. There was one drawback to UC approach, the
redistribution caused by breakdown of the SMEs. To puerarin
SMEs, the redistribution resulted in signicantly lower EE
obtained by UC compared with those obtained by MD and
UF. UF was still the relatively accurate method for EE
determination of SME. The in vitro release results showed
that pressure UF was a useful method which allowed
elucidation of the drug release mechanism from SME, as it
provided an instantaneous snapshot of drug distribution
between the particles and free solution. Compared with
pressure UF technology, MD can be more suitable to
determine the drug release of SME containing some small
nanoparticles phase such as liposomes micelles. The in vitro
release of puerarin from the SME was under partition control
and occurred via burst release assayed by means of MD and
pressure UF technology.
REFERENCES
1. Kawakami S, Yamashita F, Hasida M. Disposition characteristics
of emulsion and incorporated drugs after systemic or local
injection. Adv Drug Deliv Rev. 2000;45:77 88.
2. Nasirideen S. Naproxen incorporated lipid emulsion fomulation
and stability study. J Clin Pharm Ther. 1998;23:57 65.
3. Mizushima Y, Shoji Y, Kato T, Fukushima M, Kurozumi S. Use
of lipid microspheres as a drug carrier for antitumour drugs. J
4. Venkataram S, Awini WM, Jordan K, Rahman YE. Pharmaco
kinetics of two alternative dosage forms for cyclosporine:
liposomes and intralipid. J Pharm Sci. 1990;79:216 9.
Yue et al.
5. Forster D, Washington C, Davis SS. Toxicity of solubilized and
colloidal amphotericin B formulations to human erythrocytes. J
6. Fukui H, Koike T, Saheki A, Sonoke S, Seki J. A novel delivery
system for amphotericin B with lipid nanosphere. Int J Pharm.
2003;265:37 45.
7. Dirk LT, Bradley D, Andersonb WJP. The distribution and
chemical kinetics of prostaglandin E1 in lipid emulsions. Adv
Drug Deliv Rev. 1996;20:155 64.
8. Menashe YL, Levy L, Shoshanna GS, Benita S. Side effect
evaluation of a new diazepam formulation: venous sequela
reduction following intravenous injection of a diazepam emul
sion in rabbits. Pharm Res. 1989;6:510 6.
9. Lpez Pinto JM, Gonzlez Rodrguez ML, Rabasco AM. Effect
of cholesterol and ethanol on dermal delivery from DPPC
liposomes. Int J Pharm. 2005;298:1 12.
10. Patel VB, Misra AN. Encapsulation and stability of clofazimine
liposomes. J Microencapsul. 1999;16:357 67.
11. Patlolla RR, Vobalaboina V. Pharmacokinetics and tissue
distribution of etoposide delivered in parenteral emulsion. J
Pharm Sci. 2005;94:437 45.
12. Cui FD, Shi K, Zhang L, Tao AJ, Kawashima YJ. Biodegradable
nanoparticles loaded with insulin phospholipid complex for oral
delivery: preparation, in vitro characterization and in vivo
evaluation. J Control Release. 2006;114:242 50.
13. Wang T, Deng YJ, Geng YH, Gao ZB, Zou JP, Wang ZX.
Progress of study on niosomes. Biochim Biophys Acta.
2006;1758:222 31.
14. Michalowski CB, Guterres SS, Dalla Costa T. Microdialysis for
evaluating the entrapment and release of a lipophilic drug from
nanoparticles. J Pharm Biomed Anal. 2004;35:1093 100.
15. Han J, Washington C. Partition of antimicrobial additives in an
intravenous emulsion and their effect on emulsion physical
stability. Int J Pharm. 2005;288:263 71.
16. Liu Q, Lu Z, Wang L. Restrictive effect of puerarin on
myocardial infarct area in dogs and its possible mechanism. J
Tongji Med Univ. 2000;20:43 5.
17. Guerra MC, Speroni E, Broccoli M, et al. Comparison between
Chinese medical herb Pueraria lobata crude extract and its main
isoavone puerarin: antioxidant properties and effects on rat
liver CYP catalysed drug metabolism. Life Sci. 2000;67:2997
3006.
18. Lu XR, Gao E, Xu LZ, Li HZ, Kang B, Chen WN, Chen SM,
Chai XS. Puerarin beta adrenergic receptor blocking effect. Chin
Med J. 1987;100:25 8.
19. Liao HB, He ZF, Wang GC, Li HJ, Yue CJ, Chen XF. Research
progress about Radix puerariae. Sci Technol Food Ind.
2003;24:81 2.
20. Fan LL, Sun LH. The protective effect of puerarin against
myocardial reperfusion injury. Chin Med J. 1992;105:7 11.
21. Xuan B, Zhou YH, Yang RL. Improvement of ocular blood ow
and retinal functions with puerarin analogs. J Ocul Pharmacol
Ther. 1999;15:207 16.
22. Zhang LN, Shi HQ. Analysis on 60 cases of pueraria injection
adverse reactions. Chinese Journal of Drug Application and
Monitoring 2004;3:1672 73.
23. Yue PF, Yang HL, Yang M, Zhu WF, Cai PL, Xiao XH. The
study to reduce the hemolysis side effect of puerarin by a
submicron emulsion delivery system. Biol Pharmaceut Bull.
2008;31:45 50.
24. Zhang Y. Studies on physicochemical of crystal avonoids by
polymorphism of puerarin. Hefei University of Technology
Dissertation 2003;72:43 5.
25. Lu B, Wu W. Optimization of preparation of dexamethasone
acetate loaded poly(D,L lactide) microspheres by central com
posite design. Acta Pharmaceutica Sinica 1999;34:387 91.
26. Hansen DK, Davies MI, Lunte SM, Lunte CE. Pharmacokinetic
and metabolism studies using microdialysis sampling. J Pharm
Sci. 1999;88:14 27.
27. Li Y, Pan WS, Chen SL, Yang DJ, et al. Studies on preparation of
puerarin phytosomes and their solid dispersion. Chinese Phar
maceutical Journal 2006;41:1162 7.
28. Chambrier C, Guriraud M, Gibault JP. Medium and long chain
triacglycerols in postoperatice patients: structured lipid versus a
physical mixture. Nutrition 1999;15:274 77.

29. Wang LX, He HB, Tang X, Shao RY, Chen DW. A less irritant
norcantharidin lipid microspheres: formulation and drug distri
bution. Int J Pharm. 2006;323:161 7.
30. Groves MJ, Wineberg M, Brain APR. The presence of liposomal
material in phosphatide stabilized emulsions. J Dispers Sci
Technol. 1985;2:237 43.
31. Uchegbu IF, Vyas SP. Non ionic surfactant based vesicles
(niosomes) in drug delivery. Int J Pharm. 1998;172:33 70.
32. Frzou J, Lai NT, Leray C, Hajri T, Frey A, Cabaret Y,
Courtieu J. Lipid composition and structure of commercial
parenteral emulsions. Biochim Biophys Acta. 1994;1213:149
58.
383
33. Bach A, Frzou J, Frey A. Phospholipid rich particles in commer
cial parenteral fat emulsions. Prog Lipid Res. 1996;35:133 53.
34. Herrera AM, Scott DO, Lunte CE. Microdialysis sampling for
determination of plasma protein binding of drugs. Pharm Res.
1990;7:1077 81.
35. Magenheim B, Levy MY, Benita S. A new in vitro technique for
the evaluation of drug release prole from colloidal carriers by
ultraltration technique at low pressure. Int J Pharm.
1993;94:115 23.
36. Benita S, Levy MY. Submicron emulsions as colloidal drug
carriers for intravenous administration: comprehensive physico
chemical characterization. J Pharm Sci. 1993;82:1069 79.

DOI: 10.1208/s12249 009 9217 2
Research Article
Incorporation in Lipid Microparticles of the UVA Filter, Butyl
Methoxydibenzoylmethane Combined with the UVB Filter, Octocrylene:
Effect on Photostability
Santo Scalia1,2,3 and Matteo Mezzena1
Received 18 August 2008; accepted 1 March 2009; published online 21 April 2009
Abstract. The aim of this study was to reduce the photoinstability of butyl methoxydibenzoylmethane
(BMDBM), the most widely used UVA lter, by incorporating it in lipid microparticles (LMs) alone or
together with the UVB lter octocrylene (OCR), acting also as photostabilizer. Microparticles loaded
with BMDBM or with combined BMDBM and OCR were produced by the hot emulsion technique,
using glyceryl behenate as lipid material and poloxamer 188 as surfactant. The LMs were characterized
by release studies, scanning electron microscopy, and powder X ray diffractometry. The BMDBM and
OCR loading was 15.2% and 10.6%, respectively. In order to reproduce the conditions prevalent in
commercial sunscreen products, the photoprotective efcacy of the LMs was evaluated after their
introduction in a model cream (oil in water emulsion) containing a mixture of UVA and UVB lters. A
small but statistically signicant decrease in BMDBM photodegradation was obtained when the UVA
lter was encapsulated alone into the LMs (the extent of degradation was 28.6% 2.4 for non
encapsulated BMDBM and 26.0% 2.5 for BMDBM loaded microparticles). On the other hand, the co
loading of OCR in the LMs produced a more marked reduction in the light induced decomposition of
microencapsulated BMDBM (the UVA lter loss was 21.5% 2.2). Therefore, incorporation in lipid
microparticles of BMDBM together with the sunscreen OCR is more effective in enhancing the UVA
lter photostability than LMs loaded with BMDBM alone.
KEY WORDS: butyl methoxydibenzoylmethane; lipid microparticles; octocrylene; photodegradation;
sunscreen formulation.
INTRODUCTION
The use of topical products for sun protection is
constantly increasing (1) due to the rising level of public
awareness of the numerous harmful effects (erythema,
cutaneous photoaging, immune suppression, and various
forms of skin cancers) of solar UV radiation (2,3). The
sunscreening ingredients incorporated in these preparations,
referred to as sunscreen agents or UV lters, decrease the
dose of UV rays impacting on the skin by absorbing,
reecting, or scattering the radiation (4).
Although the sunlight induced skin damage has been
attributed mainly to the UVB rays (290 320 nm), more
recently the important contribution of UVA wavelengths
(320 400 nm) has been well documented (5,6). Therefore,
sunscreen products should provide an effective protection
throughout the whole UV range (290 400 nm) of sun
radiation reaching the Earths surface (3,6). In order to
achieve these characteristics, combinations of several UVB
Department of Pharmaceutical Sciences, University of Ferrara,

44100 Ferrara, Italy.
2
Dipartimento di Scienze Farmaceutiche, via Fossato di Mortara, 17,
44100 Ferrara, Italy.
3
To whom correspondence should be addressed. (e mail: sls@unife.it)
and UVA lters are introduced in the formulation of

sunscreening preparations (4,7,8).
Within the class of UVA absorbing substances, butyl
methoxydibenzoylmethane (BMDBM; Fig. 1) is the most
widely used sunscreen compound (7,9,10). It is included in the
list of authorized sunscreen agents in Europe, USA, Aus
tralia, and Japan (11). BMDBM exhibits high absorptive
capacity in the UVA region, but it suffers from marked
decomposition under sunlight irradiation which leads to a
reduction in the protective efcacy of the sunscreen prepara
tion during solar exposure (9,12,13). In addition, its photo
fragmentation results in the formation of free radicals which
may directly or indirectly initiate skin damage (14,15). The
photochemical inactivation of BMDBM is thus a limiting
factor for the formulation of sun care products (9).
The instability of BMDBM under sunlight can be
reduced by the addition of UVB lters, such as octocrylene
or methylbenzylidene camphor, with triplet energy similar to
BMDBM and acting as quenchers of its triplet state (8,9).
Because of the lower effectiveness of methylbenzylidene
camphor and concern with its safety (8), octocrylene (OCR;
Fig. 1), a UVB lter with moderate extinction coefcient, is
generally employed as photostabilizer for BMDBM (9,16).
However, this effect is reduced when in combination with
octyl methoxycinnamate (OMC) (13,15,17). This aspect is
extremely relevant, since OMC is the most widely used UVB
384
Lipid Microparticles Loaded with Sunscreen Combinations
385
Materials
Butyl methoxydibenzoylmethane and octyl methoxycin
namate were supplied by Merck (Darmstadt, Germany).
Octocrylene and poloxamer 188 were from BASF (Ludwig
shafen, Germany). Glyceryl behenate (a mixture of mono ,
di , and tri esters of glycerol and behenic acid) was obtained
from Gattefoss (Cedex, France). Tristearin was purchased
from Fluka Chemie (Bucks, Switzerland). Hydrogenated
soybean phosphatidylcholine was a gift by Cargill (Hamburg,
Germany). The excipients for the cream preparations were
from Sigma Aldrich (Steinheim, Germany) and Henkel (Fino
Mornasco, Italy). Methanol, acetonitrile and water were high
performance liquid chromatography (HPLC) grade from
Merck. All other reagents and solvents were of analytical
grade (Sigma).
Fig. 1. Chemical structures of BMDBM and OCR
Methods
High Performance Liquid Chromatography
absorber worldwide (9,15,16) and its association with

BMDBM dominates the ranking of sunscreen market shares
in most countries (7,9,10). Hence, there is a need for new
systems exhibiting enhanced photostability for BMDBM,
especially when it is combined with the UVB absorber
OMC, as in most commercial formulations (9).
Inclusion complexation with cyclodextrins, encapsulation
in micro and nano particles have been investigated in order
to improve the efcacy and stability of BMDBM under solar
radiation (9,14,18 20).
Recently, attention has been focused on lipid micro
particles (LMs) as a promising carrier system for sunscreen
agents (19,21,22). They consist of a solid lipid core based on
naturally occurring lipids and stabilized by a layer of
surfactant molecules on the surface (23). Consequently, their
components are physiologically compatible and biodegrad
able, providing excellent in vivo tolerability (24). Additional
advantages of LMs include high loading capacity for lipophilic
substances, such as most of the UV lters, and decreased skin
penetration of encapsulated sunscreens (21). Moreover, their
solid matrix protects incorporated actives against decomposi
tion (24,25).
Previous investigations, demonstrating the photostabiliz
ing effect of LMs on the encapsulated sunscreen agents, have
been performed on individual UV lters (19,21,22,26).
However, these systems do not simulate the actual conditions
of use, since commercial sunscreen products always contain a
mixture of UVA and UVB absorbers (7 10). In order to
overcome this drawback and enhance the lipid microparticle
protective effect, the present study reports on the incorpora
tion of BMDBM in LMs together with the UVB lter OCR,
acting also as photostabilizer. The LMs loaded with the
BMDBM/OCR combination were then introduced in a model
sunscreen formulation (emulsion) containing OMC as addi
tional UVB lter, and their inuence on the light induced
degradation of BMDBM was evaluated. For comparison
purposes, LMs containing BMDBM only were also prepared
and examined.
The HPLC apparatus comprised a Model LabFlow 3000

pump (LabService Analytica, Bologna, Italy), a Model 7125
injection valve with a 20 l sample loop (Rheodyne, Cotati,
CA, USA) and a Model 975 UV variable wavelength UV Vis
detector (Jasco, Tokyo, Japan) set at 330 nm, which is the
optimum wavelength to obtain satisfactory UV responses for
the examined sunscreen agents exhibiting different absorp
tion maxima. Data acquisition and processing were accom
plished with a personal computer using Borwin software
(JBMS Developpements, Le Fontanil, France). Sample
injections were performed with a Model 701 syringe (10 l;
Hamilton, Bonaduz, Switzerland). Separations were per
formed according to the method of Simeoni et al. (27), using
a 5 m Zorbax SB CN column (1503.0 mm i.d.) tted with a
guard column (5 m particles, 42 mm i.d.) and eluted
isocratically, at a ow rate of 0.5 ml/min, with methanol
acetonitrile water (35:20:45, v/v/v), containing 0.5% (v/v)
acetic acid. The identities of the UV lter peaks were
assigned by co chromatography with the authentic standards.
Quantication was carried out by integration of the peak
areas using the external standardization method.
Lipid Microparticle Preparation
Lipid microparticles were prepared by adding preheated
(75 85C) water (50 ml) containing 1% (w/w) of previously
dispersed (magnetic stirring) surfactant, to the melted lipid
phase (3.6 g) in which BMDBM (1.0 g) or the BMDBM/OCR
mixture (1.7 g) has been dissolved. The hot aqueous phase
was poured into the melted lipid, rather than the contrary, to
avoid loss of lipid excipients and sunscreen agents during the
manufacturing process. The sample was then mixed
(13,500 rpm for 2 min) with an Ultra Turrax T25 (IKA Werk,
Staufen, Germany) at 75 85C. The resulting oil in water
emulsion was rapidly cooled at room temperature under
magnetic stirring and the formed LMs were recovered by
centrifugation (6,000 rpm for 15 min) and freeze dried.
386
Scalia and Mezzena
In Vitro Release
Photodegradation Studies
The sunscreen dissolution and release from the LMs

were studied by adding BMDBM (5 mg) and OCR (3.5 mg)
or LMs containing an equivalent amount of the sunscreens, to
propylene glycol (50 ml) under mechanical stirring at 50 rpm
and 37C. At appropriate time intervals, 1 ml aliquots of the
release medium were withdrawn and replaced with an equal
volume of fresh medium. The samples were ltered (0.45 m)
and assayed for BMDBM and OCR by HPLC, after dilution
(1:1) with methanol. Each series of experiments was repeated
six times.
Portions (35 45 mg) of the test creams were homoge

neously spread onto a Transpore tape (3M Health Care,
Neuss, Germany) at a level of 2 mg/cm2. The obtained
samples were irradiated for 1 h with a solar simulator (Suntest
CPS+, Atlas, Linsengericht, Germany) equipped with a
Xenon lamp, an optical lter to cut off wavelengths shorter
than 290 nm and an IR block lter to avoid thermal effects.
The solar simulator emission was maintained at 750 W/m2.
The applied UV energy was equivalent to ten minimal
erythemal dose (MED), which is considered representative
of half day solar emission close to the equator (12). After the
exposure interval, the Transpore tape was cut into small
pieces and extracted with ethanol (5 ml) under sonication
(5 min). The sonication was repeated twice with methanol
(5 ml) followed by overnight extraction, under stirring, with
fresh methanol (15 ml). The combined fractions were
adjusted to volume (50 ml) and the obtained sample was
ltered (0.45 m membrane lters) and analyzed by HPLC.
The degree of photodegradation was evaluated by measuring
the percentage of recovered sunscreen agents with respect to
non irradiated samples. The results were the average of at
least ten experiments.
Microparticle Characterization
Microparticle morphological structure was examined by
scanning electron microscopy (SEM; Cambridge Stereoscan
360, Cambridge Instruments, Bar Hill, UK). The particle size
was determined by computerized image analysis (Micro
metrics camera 122CU and software Vision 1.0) of at least
100 particles on photomicrographs obtained with an optical
microscope (Nikon Diaphot inverted microscope, Tokyo,
Japan).
The powder X ray diffraction patterns were recorded on
a D 5000 powder diffractometer (Siemens, Munich, Ger
many) using a voltage of 45 kV and a current of 25 mA for the
generator, with Cu anode material. The wavelength of the
graphite monocromated radiation was 1.5406 . The diffrac
tograms were recorded from 3 (2) to 50 (2) at an angular
speed of 1 (2) per minute using 1 1 1 0.15 slits.
The amount of BMDBM and OCR entrapped in the
LMs was determined by dissolving the microparticles (35
40 mg) in ethanol under sonication (25 min). The obtained
sample was diluted to volume (20 ml), ltered and assayed by
HPLC. The encapsulation efciency was calculated as the
percentage ratio between the quantity of sunscreen agents
entrapped in the microparticles and added to the melted lipid
phase, during preparation. Data were determined from the
average of at least six determinations.
In Vitro Sun Protection Factor Measurement

The in vitro determination of the cream sun protection
factor (SPF) was carried out according to the Diffey and
Robson (28) technique, with minor modications. The
method is based on the measurement of the transmission
spectrum of the UV radiation (290 400 nm) through a
Transpore tape, before and after application (2 mg/cm2)
of the sunscreen preparation. The tape was placed into the
spectrophotometer (Model V 530PC UV VIS; Jasco, Tokyo,
Japan) sample compartment, over the quartz input optics of
the detector. The spectral data were processed with a
personal computer and the SPF calculated according to
Diffey and Robson (28).
Emulsion Formulations
The photolysis experiments were performed in cream
preparations (oil in water emulsions) containing BMDBM
(1%, w/w) and OCR (0.7%, w/w) incorporated in lipid
microparticles. Creams containing equivalent amounts of
plain BMDBM and OCR in conjunction with blank LMs or
BMDBM loaded microparticles with non encapsulated OCR
were also examined. The UVB lter OMC (1%, w/w) was
added to each formulation. The emulsion excipients were:
sorbitan monostearate (2%), polyoxyethylene sorbitan mono
stearate (4.5%), butylated hydroxyanisole (0.05%), octyl
palmitate (6.0%), liquid petrolatum (5.5%), cetearyl alcohol
(5.0%), sodium benzoate (0.1%), glycerin (2.0%), dehydro
acetic acid (0.1%), EDTA (0.1%), and water (67.0%). The
creams were prepared according to the common procedure
used in compounding practice. Blank or loaded lipid micro
particles (6.0 6.5 g per 100 g of cream) were dispersed in
water and added in the cooling phase of the emulsion
preparation at about 40C.
Statistical analysis of data was performed using Students

t test, analysis of variance (ANOVA), and Tukeys post test. P
values<0.05 were considered signicant. All computations
were carried using the statistical software GraphPad Instat
(Graphpad Software, San Diego, CA, USA).
In Vitro Release of UV Filters from Lipid Microparticles
For the preparation of lipid microparticles loaded with
combined BMDBM and OCR, a hot emulsion technique (23)
was employed, utilizing different lipid materials (tristearin,
glyceryl behenate) and surfactants (hydrogenated phosphati
dylcholine and poloxamer 188). To evaluate the inuence of
these excipients on the retention efcacy of the LMs for the
loaded sunscreens, in vitro release studies were performed
using a medium (propylene glycol) in which BMDBM and
OCR were sufciently soluble to ensure sink conditions,

whereas the LMs remained intact. Distinct differences were
observed among microparticles based on tristearin or glyceryl
behenate as lipid matrix and poloxamer 188 or phosphatidyl
choline as stabilizer (Fig. 2). The slowest release rates for
both BMDBM and OCR were achieved by the LMs prepared
with glyceryl behenate and poloxamer 188 (Fig. 2a and b),
which indicated a more efcient incorporation of the UV
lters in this system. The reduction in release was statistically
signicant (ANOVA and Tukeys post test) at 60 and 120 min
(P<0.01). Moreover, the lack of burst effect phenomena
(Fig. 2) suggested that there was no adsorption of the
sunscreens at the microparticle surface. The obtained data
pointed out that the nature of the lipid and surfactant
excipients is an important factor for the LMs release
modulation capacity (23). This effect can be ascribed to lipid
polymorphic transformations (23), different afnity between
the UV lters and the lipid materials (29), and to differences
in the extent of lipid surface coverage by the surfactant (30).
In addition, the different melting points of the examined
Fig. 2. BMDBM (a) and OCR (b) dissolution (empty triangles) and
release from LMs prepared with tristearin and phosphatidylcholine
(filled diamonds), tristearin and poloxamer 188 (empty circles),
glyceryl behenate and phosphatidylcholine (empty squares), or
glyceryl behenate and poloxamer 188 (filled triangles). Values are
meansSD (n 6)
387
lipids (glyceryl behenate, ca 83C; tristearin, ca 65C)
determining the production temperature of the LMs can
affect their release behavior (24). On the basis of the above
results, LMs prepared with glyceryl behenate and poloxamer
188 were used for further experimentation since they
exhibited the greatest retention capacity for the examined
UV lters (Fig. 2).
The highest microparticle yield (percentage ratio be
tween the weight of microparticles and the weight of lipid,
surfactant, and active fed initially) was obtained at a lipid/
emulsier ratio of 7:1. Additional production parameters
including the stirring rate (9,500 17,500 rpm) and time (1
5 min) were evaluated in order to obtain particles with
satisfactory morphological structure and size homogeneity.
The best results in terms of particle size, polydispersity and
surface smoothness were attained using a stirring rate of
13,500 rpm for 2 min.
Lipid Microparticle Characterization
SEM analysis on the optimized LMs, based on glyceryl
behenate and poloxamer 188, showed a spherical shape
although some irregular fragments were present (Fig. 3).
The particle size was between 7 and 25 m (mean diameter,
12.7 m; polydispersity index, 0.69), the majority (73%) of
the population being in the range 7 15 m, which is
appropriate when skin penetration should be minimized, as
for the sunscreen agents (31). In fact, microparticles do not
permeate the skin (31), whereas percutaneous penetration of
nanometer sized particles has been demonstrated (29,32),
though other studies have reported that nanoparticles do not
permeate the stratum corneum, but they diffuse into the hair
follicles (33).
Additional information on the solid state of the LMs was
obtained by powder X ray diffractometry (Fig. 4). The
diffraction pattern of the corresponding physical mixture of
the sunscreen agents with the lipid microparticle excipients
(Fig. 4d) displayed the crystalline peaks of glyceryl behenate
(4.1, 21.1, 23.3; Fig. 4a), poloxamer 188 (18.9, 23.3;
Fig. 4b), and BMDBM (10.7, 13.2, 16.2, 17.3, 19.5, 20.3,
24.9; Fig. 4c); the OCR being an oil did not exhibit any
signal. The characteristic peaks of BMDBM were not
detected in the diffractogram of the lipid particles (Fig. 4e),
suggesting its amorphization in the LMs. On the other hand,
the typical signals of glyceryl behenate were observed in the
LMs pattern (Fig. 4e), indicating that it is, at least partially,
crystalline in the LMs.
The quantity of BMDBM and OCR incorporated into
the LMs was 15.2% 0.4 (w/w) and 10.6% 2.1 (w/w),
respectively. The encapsulation efciency ranged from
76.9% to 80.9%.
Glyceryl behenate and poloxamer 188 based LMs con
taining BMDBM without OCR were also prepared and
characterized. No signicant differences in particle morphol
ogy, dimensional distribution (mean diameter, 10.6 m;
polydispersity index, 0.67), encapsulation efciency (76.6%),
and BMDBM release prole (curve not shown) were
observed compared with the systems containing BMDBM in
combination with OCR. Therefore, the physical character
istics of BMDBM loaded LMs were not affected by the co
loading of OCR.
388
Fig. 3. Scanning electron microscopy (SEM) micrographs of LMs

loaded with BMDBM and OCR
Photodegradation Studies
Previous investigations on the improvement of BMDBM
photostability by its encapsulation in lipid microparticles have
been performed in cream formulations containing BMDBM
as the only UV lter (19,26). Although this option provides
valuable information on the photochemistry aspects (9), the
relevance of these studies to the real conditions of use of sun
protective products is limited, since in a typical sunscreening
preparation not just one but a mixture of UVB and UVA
lters is always employed in order to provide broad spectrum
protection (3,7,9,10). Moreover, as the light induced decom
position of a sunscreen agent is affected by the presence of
other UV absorbers, the resulting photoinstability in UV lter
combinations may be different from that observed for the
individual sunscreen agent (8 10,13).
Scalia and Mezzena

Accordingly, in the present work, the photochemical
behavior of encapsulated BMDBM (UVA lter) was examined
in the presence of the UVB absorbers OCR and OMC, using a
cream (oil in water emulsion) as a vehicle. This system repro
duces the conditions prevailing under the actual application of
sunscreen products, since BMDBM is associated with OMC in
most commercial suncare preparations (7,9,10) and stabilizing
molecules, such as OCR, are often included in order to reduce
the photoinstability of this combination (8,9,16). Moreover, the
oil in water emulsion selected as a model formulation, repre
sents the most common type of sunscreen product (16).
In order to verify the stabilizing effect of OCR,
preliminary photolysis experiments were performed on
creams containing the non encapsulated UV lter combina
tion BMDBM/OMC or BMDBM/OMC/OCR. The emulsions
were applied onto Transpore tapes (a surgical tape
simulating the texture of human skin), irradiated with the
solar simulator and the extent of degradation measured by
HPLC. The percentage losses of BMDBM were 32.62.3 in
the formulation containing the UVA lter in conjunction with
OMC and 28.62.4 in the cream which also included OCR,
the differences between the foregoing values being statisti
cally signicant (P<0.02, unpaired t test). These results
indicated, in accordance with earlier studies (10,13,17) that
the presence of OCR preserved the UVA lter BMDBM
from degradation, though only partially. This stabilizing effect
was also exerted on the UVB absorber OMC (the extent of
OMC decomposition decreased from 62.1% 1.7 to 53.8%
3.1, P<0.01). On the other hand, OCR appeared to be
rather stable under simulated sunlight (the amount lost was
<4.1%), in good agreement with previous reports (8 10).
The following photodegradation studies were carried
out, under the same experimental conditions, on creams
containing BMDBM loaded LMs in conjunction with plain
Fig. 4. Powder X ray diffraction patterns of glyceryl behenate (a), poloxamer 188 (b), BMDBM (c), BMDBM OCR lipoparticle excipient
physical mixture (d), and LMs loaded with BMDBM and OCR (e)

OCR and OMC (Fig. 5). The amount of BMDBM recovered
after irradiation was 74.0%, corresponding to a decomposi
tion of 26.0% 2.5 (Fig. 5). These results indicated a lower
UVA lter degradation compared with the formulation
containing free BMDBM, OMC, and OCR (BMDBM loss,
28.6%; Fig. 5), the difference being statistically signicant
(unpaired t test, P<0.05). However, the observed protective
efcacy of the lipid microparticles was not as marked as that
reported in previous studies based on formulations containing
BMDBM as the only sunscreen agent (19). This discrepancy
may be ascribable to the fact that the photochemical behavior
of BMDBM changes when in mixture with other UV
absorbers (8 10,13). In addition, the photostabilization effect
of OCR, based on triplet state quenching by an energy
transfer mechanism (9), may be hindered when BMDBM is
entrapped in the particle matrix.
Consequently, it seemed interesting to investigate wheth
er the co loading of BMDBM with OCR in LMs could
produce a more distinct improvement in the UVA lter
stability under sunlight. LMs loaded with combined BMDBM
and OCR were incorporated into the cream and compared in
terms of photodegradation with the formulations containing
non encapsulated BMDBM and OCR or lipid microparticles
loaded with BMDBM only in conjunction with free OCR. A
more marked reduction in the light induced decomposition of
the UVA lter to 21.5% 2.1 was achieved by the microen
capsulation of BMDBM together with OCR (Fig. 5). Statis
tical analysis of the data (ANOVA and Tukeys post test)
demonstrated that the differences among the examined
formulations (Fig. 5) were signicant (P<0.01). Moreover,
additional photostability experiments performed on a cream
containing BMDBM/OCR loaded microparticles prepared
with tristearin, instead of glyceryl behenate, and poloxamer
188 indicated that the extent of BMDBM degradation (29.4%
2.2) was not signicantly different compared with the free
UV lter. Therefore, the photostabilization effect of the
examined microparticle systems correlated with their release
modulation capacity (Fig. 2), LMs with higher release rate
exhibiting reduced protective effect. In fact, the release of the
sunscreens from the LMs decreases the UV lter fraction
protected by the lipid particle matrix. These results indicated
that the co loading of OCR signicantly enhanced the
389
photostability of BMDBM encapsulated in LMs, compared
with the systems based on the classical combination of free
BMDBM and OCR or containing the microparticle entrapped
BMDBM with the non encapsulated OCR. This phenomenon
could be traced to a more efcient interaction (triplet state
energy transfer) of the OCR photostabilizer with the BMDBM
molecules in the lipid particle core, without interferences from
emulsion excipients.
The photolysis experiments also pointed out that the
OCR stability under irradiation (Fig. 5) was improved after
its incorporation in the lipid particles (OCR degradation,
<0.6%). In addition, the UVB lter OMC was found to
degrade by 52.9 53.8%, the extent of the light induced
decomposition being similar for all studied creams (Fig. 5).
This was expected, since this UVB lter was introduced in the
tested emulsions in the non encapsulated form. However, the
observed loss in OMC concentration (Fig. 5), which is
consistent with published data (8,13), cannot be considered
as real instability, since it is due to trans cis isomerization, the
obtained cis isomer absorbing at the same wavelength, though
with a lower extinction (9,13).
In Vitro Sun Protection Factor
Since one of the most important criteria for the
assessment of a sunscreen product performance is the sun
protection factor (SPF) (11), this parameter was determined
in vitro in the examined formulations, according to the Diffey
and Robson technique (28). No signicant differences (P>
0.05, ANOVA) were observed among the in vitro SPF values
(6.2 6.7) of the creams containing the various sunscreen
systems described above (i.e., free BMDBM/OCR/OMC;
BMDBM loaded LMs with free OCR/OMC; microencapsu
lated BMDBM/OCR with free OMC). This indicated that
LM incorporation of the sunscreen agents has not modied
their overall UV attenuation characteristics. Another param
eter obtained from the in vitro SPF measurements is the
UVA/UVB ratio (17), an indicator of the UVA absorbing
performance in relation to that in the UVB. For all tested
formulations, the UVA/UVB ratio was nearly the same (1.0
1.1), suggesting that also the UVA protection was not altered
by the microencapsulation process. Moreover, the measured
in vitro SPF and UVA/UVB ratio fullled the general
requirements on sunscreen products (3,17).
CONCLUSIONS
Fig. 5. BMDBM , OCR and OMC photodegradation (%) in their

formulations after 1 h irradiation with the solar simulator. Values are
meansSD of at least ten experiments
The task of photostabilizing BMDBM is a primary aim of

sunscreen formulators (9) because of its role as the UVA
absorber of choice and therefore its potential impact on the
overall sunscreen preparation performance. The results
described in the present study indicate that the co loading
of BMDBM with OCR in lipid microparticles is more
effective in enhancing the UVA lter photostability compared
to LMs loaded with BMDBM alone. In addition, the
microencapsulation process, while limiting the loss of photo
protective capacity due to UV lter decomposition under
sunlight, did not alter the performance of the sunscreen
preparation, as measured by the in vitro SPF and UVA/UVB
ratio. Moreover, at variance with previous studies, the
photostabilizing properties of lipid microparticles have been
390
assessed in a sunscreen mixture typical of commercial sun
protective products and hence simulating realistic conditions
of use. The developed formulations based on BMDBM
loaded LMs could provide a useful alternative to convention
al sun care products containing this UVA lter.
Scalia and Mezzena
15.
16.
ACKNOWLEDGEMENTS
17.
The authors are grateful to MIUR (Ministero dellIstru

zione, dellUniversit e della Ricerca, Rome, Italy) for
nancial support.
18.
REFERENCES
1. Shaath NA, Shaath M. Recent sunscreen market trends. In:
Shaath N, editor. Sunscreens. Boca Raton, FL: Taylor Francis;
2005. p. 929.
2. Matsumura Y, Ananthaswamy HN. Toxic effects of ultraviolet
radiation on the skin. Toxicol Appl Pharmacol 2004;195:298 308.
3. EC Commission Recommendation on the efcacy of sunscreen
products and the claims made relating thereto. Ofcial Journal of
the European Union, 2006;L.265:39 43.
4. Nohynek GJ, Schaefer H. Benet and risk of organic ultraviolet
lters. Regul Toxicol Pharmacol 2001;33:285 99.
5. Agar NS, Halliday GM, Barneston R, Ananthaswamy HN,
Wheeler M, Jones AM. The basal layer in human squamous
tumors harbors more UVA than UVB ngerprint mutations: a
role for UVA in human skin carcinogenesis. Proc Natl Acad Sci
(USA) 2004;101:4954 9.
6. Fourtanier A, Bernerd F, Bouillon C, Marrot L, Moyal D, Seit S.
Protection of skin biological targets by different types of sunscreens.
Photodermatol Photoimmunol Photomed 2006;22:22 32.
7. Dondi D, Albini A, Serpone N. Interactions between different
UVB/UVA lters contained in commercial suncreams and
consequent loss of UV protection. Photochem Photobiol Sci
2006;5:835 43.
8. Gaspar LR, Maia Campos PMBG. Evaluation of the photo
stability of different UV lter combinations in a sunscreen. Int J
Pharm 2006;307:123 8.
9. Bonda CA. The photostability of organic sunscreen actives. In:
Shaath N, editor. Sunscreens. Boca Raton, FL: Taylor Francis;
2005. p. 323 45.
10. Damiani E, Baschong W, Greci L. UV lter combinations under
UV A exposure: concomitant quantication of overall spectral
stability and molecular integrity. J Photochem Photobiol B
2007;87:95 104.
11. Steinberg DC. Regulations of sunscreens worldwide. In: Shaath
N, editor. Sunscreens. Boca Raton, FL: Taylor Francis; 2005. p.
180 3.
12. Tarras Wahlberg N, Stenhagen G, Lark O, Rosn A, Wennberg
AM, Wennerstrm O. Changes in ultraviolet absorption of
sunscreens after ultraviolet irradiation. J Invest Dermatol
1999;113:547 53.
13. Chatelain E, Gabard B. Photostabilization of butyl methoxydi
benzoylmethane (Avobenzone) and ethylhexyl methoxycinna
mate by bis ethylhexyloxyphenol methoxyphenyl triazine
(Tinosorb S), a new UV broadband lter. Photochem Photobiol
2001;74:401 6.
14. Scalia S, Simeoni S, Barbieri A, Sostero S. Inuence of
hydroxypropyl cyclodextrin on photo induced free radical
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
production by the sunscreen agent, butyl methoxydibenzoyl

methane. J Pharm Pharmacol 2002;54:1553 8.
Damiani E, Rosati L, Castagna R, Carloni P, Greci L. Changes in
ultraviolet absorbance and hence in protective efcacy against
lipid peroxidation of organic sunscreens after UV A irradiation. J
Photochem Photobiol B 2006;82:204 13.
Klein K, Palefsky I. Formulating sunscreen products. In: Shaath
N, editor. Sunscreens. Boca Raton, FL: Taylor Francis; 2005. p.
356 70.
Herzog B, Mongiat S, Deshayes C, Neuhaus M, Sommer K,
Mantler A. In vivo in vitro assessment of UVA protection by
sunscreen formulations containing either butyl methoxydiben
zoylmethane, methylene bis benzotriazolyl tetramethylbutylphe
nol or microne ZnO. Int J Cosmet Sci 2002;24:170 85.
Lapidot N, Gans O, Biagini F, Sosonkin L, Rottman C.
Advanced sunscreens: UV absorbers encapsulated in sol gel
glass microcapsules. J Sol Gel Sci Technol 2003;26:67 72.
Iannuccelli V, Sala N, Tursilli R, Coppi G, Scalia S. Inuence of
liposphere preparation on butyl methoxydibenzoylmethane pho
tostability. Eur J Pharm Biopharm 2006;63:140 5.
Xia Q, Saupe A, Mller RH, Souto EB. Nanostructured lipid
carriers as novel carrier for sunscreen formulations. Int J Cosmet
Sci 2007;29:473 82.
Yener G, Incegl T, Yener N. Importance of using solid lipid
microspheres as carriers for UV lters on the example of octyl
methoxy cinnamate. Int J Pharm 2003;258:203 7.
Tursilli R, Piel G, Delattre L, Scalia S. Solid lipid microparticles
containing the sunscreen agent, octyl dimethylaminobenzoate:
effect of the vehicle. Eur J Pharm Biopharm 2007;66:483 7.
Jaspart S, Piel G, Delattre L, Evrard B. Solid lipid micro
particles: formulation, preparation, characterization, drug release
and applications. Expert Opin Drug Deliv 2005;2:75 87.
Mller RH, Radtke M, Wissing SA. Solid lipid nanoparticles
(SLN) and nanostructured lipid carriers (NLC) in cosmetic and
dermatological preparations. Adv Drug Deliv Rev 2002;54:S131
55.
Jee JP, Lim SJ, Park JS, Kim CK. Stabilization of all trans retinol
by loading lipophilic antioxidants in solid lipid nanoparticles. Eur
J Pharm Biopharm 2006;63:134 9.
Scalia S, Tursilli R, Sala N, Iannuccelli V. Encapsulation in
lipospheres of the complex between butyl methoxydibenzoyl
methane and hydroxypropyl cyclodextrin. Int J Pharm
2006;320:79 85.
Simeoni S, Tursilli R, Bianchi A, Scalia S. Assay of common
sunscreen agents in suncare products by high performance liquid
chromatography on a cyanopropyl bonded silica column. J
Pharm Biomed Anal 2005;38:250 5.
Diffey BL, Robson J. A new substrate to measure sunscreen
protection factors throughout the ultraviolet spectrum. J Soc
Cosmet Chem 1989;40:127 33.
Nasr M, Mansour S, Mortada ND, El Shamy AA. Lipospheres as
carrier for topical delivery of aceclofenac: preparation, charac
terization and in vivo evaluation. AAPS PharmSciTech
2008;9:154 162.
Mehnert W, Mder K. Solid lipid nanoparticles. Production,
characterization and applications. Adv Drug Deliv Rev
2001;47:165 96.
Wiechers JW. Avoiding transdermal cosmetic delivery. Cosmet
Toil 2000;115:39 46.
Baroli B, Ennas MG, Loffredo F, Isola M, Pinna R, Lopez
Quintela MA. Penetration of metallic nanoparticles in human
full thickness skin. J Invest Dermatol 2007;127:1701 12.
Lademann J, Richter H, Teichmann A, Otberg N, Blume Peytavi
U, Luengo J. Nanoparticles: an efcient carrier for drug delivery
into the hair follicles. Eur J Pharm Biopharm 2007;66:159 64.

DOI: 10.1208/s12249 009 9219 0
Research Article
Hesperidin Gastroresistant Microparticles by Spray-Drying: Preparation,
Characterization, and Dissolution Profiles
Francesca Sansone,1 Alessandra Rossi,2 Pasquale Del Gaudio,1 Francesco De Simone,1
Rita Patrizia Aquino,1 and Maria Rosaria Lauro1,3
Abstract. Gastroresistant microparticles for oral administration of hesperidin (Hd) were produced by
spray drying using cellulose acetate phthalate (CAP) as enteric polymer in different polymer/Hd weight
ratio (1:1, 3:1, and 5:1), and a series of enhancers of the dissolution rate, such as sodium
carboxymethylcellulose crosslinked (CMC), sodium dodecylbenzene sulfonate (SDBS), or Tween85.
The raw materials and the microparticles were investigated by differential scanning calorimetry, X ray
diffraction, infrared spectroscopy and imaged using scanning electron and uorescence microscopy. In
vitro dissolution tests were conducted using a pH change method to investigate the inuence of
formulative parameters on the dissolution/release properties of the drug. CAP/Hd microparticles showed
a good gastro resistance but incomplete drug dissolution in the simulated intestinal uid (SIF). The
presence of the enhancers in the formulation produced well formed microparticles with different size and
morphology, containing the drug well coated by the polymer. All the enhancers were able to increase the
dissolution rate of Hd in the simulated intestinal environment without altering CAP ability to protect Hd
in the acidic uid. The spray drying technique and process conditions selected were effective in
microencapsulating and stabilizing the avonoid giving satisfactory encapsulation efciency, product
yield, and microparticles morphology, and a complete drug release in the intestine.
KEY WORDS: enhancers of the dissolution rate; gastroresistant spray dried microparticles; hesperidin;
morphological and physicochemical characterization.
INTRODUCTION
Hesperidin (Hd; Fig. 1) is the major avanone glycoside
in sweet orange and lemon obtained as an abundant by
product of Citrus cultivation (1,2). It has been implicated in
the inhibition of enzymes involved in several diseases:
phospholipase A2, lipoxygenase, HMG CoA reductase and
cyclooxygenase (3,4). As a result, hesperidin has antioxidant,
anti inammatory, hypolipidemic, vasoprotective, and choles
terol lowering properties (5). In the last years, encouraging
results on its anticancer activity were observed. A marketed
formulation of hesperidin and diosmin, just used for the
treatment of chronic venous insufciency, was shown to be
active in inhibiting chemical carcinogenesis of the bladder
urinary in mice (6) and neoplastic transformation induced by
3 methylcolanthrene in murine broblasts (7). Hesperidin
was also found to posses an antimutagenic effect on
Salmonella typhimurium against benz () pyrene induced
mutation (8,9), and on rats against methyl N amylnitros
amine induced tumorigenesis (10). Its metabolite hesperetin
1
Dipartimento di Scienze Farmaceutiche, Universit degli Studi di

Salerno, Via Ponte Don Melillo, 84084 Ficiano, SA, Italy.
2
Dipartimento Farmaceutico, Universit degli Studi di Parma, V.le G.
P. Usberti 27/A, 43100 Parma, Italy.
3
To whom correspondence should be addressed. (e mail: lauro@
unisa.it)
as well as naringenin showed in vitro antiproliferative activity

on highly metastatic murine B16 F10 melanoma cells. Also,
hesperitin or naringenin, orally administered in C57BL6/N
mice, induced antimetastatic activity by reduction of cell
proliferation (11,12). About the mechanisms involved in
avonoid mediated antineoplastic activity, the drugs seem to
be activators of enzymes such as transglutaminase (TGAse)
involved in cell differentiation (13,14,15).
The well documented effects of hesperidin are unfortu
nately more evident in vitro than in vivo, due to the instability to
the gastric pH in which the drug undergoes hydrolysis into
aglycone hesperetin and enzymatic degradation. Moreover,
probably due to its crystalline state (16), this avonoid is slightly
soluble in water, a characteristic which leads to a very low
dissolution rate and an irregular absorption of the drug from
oral solid dosage forms in the gastrointestinal tract. In addition,
it is recovered in plasma, urine, and bile as glucoronides and
sulfoglucoronides and the cumulative recovery indicated low
bioavailability (<25%) (2). In contrast to the numerous
publications focused on the pharmacological effect of avo
noids, rather no attention has been addressed yet to their
formulation in order to have appropriate bioavailability.
Thus, the aim of the present work was to produce Hd
gastroresistant microparticles with improved performance for
oral administration by spray drying. Since the design and
development of gastroresistant microparticles containing a
low solubility drug such as Hd requires a compromise between
391
Sansone et al.
392
The selected surfactants were sodium dodecylbenzene sulfo

nate (SDBS) from Sigma Aldrich (Milan, Italy) and Tween
85 from A.C.E.F. s.p.a. (Piacenza, Italy). All other chemicals
used were of reagent grade.
Hesperidin Solubility
Fig. 1. Hesperidin chemical structure
the enhancement of the dissolution rate in the intestinal uid

and protection in the gastric environment, a selected coating of
gastroresistant polymer, cellulose acetate phthalate (CAP), and
a series of enhancers of the dissolution rate were used in the
formulation. CAP has just been employed for microencapsula
tion of other avonoids (17,18) and drugs such as NSAID (19)
and for preventing tablets disintegration and drug release into
the stomach (20). Spray drying is a technique to obtain micro
particulate systems by a one step process transforming liquid
feed into a dried particulate form (18), which process conditions
strongly affect properties of the resultant dried microparticles
such as size, morphology, shape. Moreover, this technique offers
other advantages such as a low amount of residual solvent in the
nal product.
The inuence of formulative parameters on drug disso
lution/release properties as well as on yield, morphology, and
thermal behavior of the produced gastroresistant micro
particles were studied.
Materials
Hesperidin (Hd) was supplied by Sigma Aldrich (Milan,
Italy); Cellulose acetate phthalate (CAP) from Eastman
Kodak (Kingsport, Tennessee, United States); Vivasol
Croscarmellose sodium (carboxymethylcellulose crosslinked,
CMC) from J. Rettenmaier & Shne (Rosenberg, Germany).
The solubility of Hesperidin was previously determined in

water (57.0 mg/l) and in simulated biological uids (prepared
according to USP 31 without enzymes), gastric uid (GF) pH 1.2
(49.1 mg/l) and intestinal uid (IF) pH 7.5 (68.0 mg/l), as follows:
an excess amount of Hd raw was introduced into glass vials
containing 50 ml of solvents; the samples were shaken and then
were stored at room temperature (25C). After 3 days, the liquid
phases were centrifuged for 15 min at 3,000 rpm, the super
natants were ltered with 0.45 m lters and the concentration
of dissolved Hd was determined spectrophotometrically (UV/
Vis spectrometer Lambda 25, Perkin Elmer Instruments, MA,
USA) at a wavelength of 310 nm. The solubility measurements
were made in triplicate. Solubility data were conrmed by
HPLC using the method reported below (Drug Content
Evaluation section).
Microparticles Preparation
Gastroresistant CAP/Hd Microparticles
The composition of the different batches of spray dried
microparticles is reported in Table I. In preliminary experi
ments, the spray drying of organic feed solutions (3:1 v/v
acetone:ethanol) of CAP (1 and 2%) and Hd (3:1, 1:1, and 5:1
polymer/drug ratio) gave no adequately coated micropar
ticles; encapsulation efciency values were below 40%, and
FM images showed a high amount of Hd uncoated by the
polymer (data not shown). For this reason, in the rst set of
experiments, CAP (1% and 2% w/v) was dissolved in
aqueous buffer at pH 7.5 (simulated intestinal uid, IF, pH
7.5, according to USP 31 without enzymes), then variable
Table I. Feed Composition, Drug Concentration, Enhancer Concentration (% w/v), CAP Concentration, % Yield of the Process, Theoretical
Drug Content (TDC %), Actual Drug Content (ADC%) and Encapsulation Efciency (EE %), d50 (mStandard Deviation) of all
Microsystems Prepared
Batches #
Feed composition
#
#
#
#
#
#
#
#
#
#
#
#
#
#
CAP/Hd
CAP/Hd
CAP/Hd
CAP/Hd/CMC
CAP/Hd/SDBS
CAP/Hd/Tween
CAP blank
CAP/CMC
CAP/SDBS
CAP/Tween
Hd sd
Hd/CMC
Hd/SDBS
Hd/Tween
1a
1b
1c
2
3
4
5
6
7
8
9
10
11
12
Drug
concentration
(% w/v)
0.67
2.0
0.4
0.67
0.67
0.67
Enhancer
concentration
(% w/v)
0.4
0.5
0.5
0.4
0.5
0.5
0.67
0.67
0.67
0.67
0.4
0.5
0.5
CAP
concentration
(% w/v)
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
Yield %
94.3
73.0
84.4
75.0
93.0
76.0
100.0
90.0
99.1
94.1
58.4
41.2
88.0
65.2
TDC %
ADC %
EE %
d50 m (SD)
25.0
50.0
16.6
18.2
21.0
21.0
18.5
35.0
12.0
17.5
17.0
18.9
74.0
70.0
72.3
96.2
80.9
90.0
100.0
28.6
57.1
57.1
100.0
25.0
29.7
28.0
87.4
52.0
49.0
12.28
11.50
13.55
80.96
43.18
42.86
3.90
35.86
10.12
6.42
4.37
28.28
19.96
21.76
(0.07)
(0.12)
(0.10)
(2.40)
(1.83)
(1.07)
(0.06)
(1.21)
(0.64)
(0.51)
(0.12)
(0.20)
(1.03)
(2.34)
CAP cellulose acetate phthalate, CMC carboxymethylcellulose crosslinked, SDBS sodium dodecylbenzene sulfonate, Tween Tween 85, Hd
Hesperidin, Hd sd, hesperidin spray dried
Hesperidin Gastroresistant Microparticles by Spray-Drying

amount of Hd (polymer:drug weight ratios 3:1, 1:1, and 5:1)
were suspended in the polymeric solution under magnetic
stirring. CAP/Hd (batches #1a, #1b, and #1c) microparticles
were obtained by spray drying of these feed solutions.
Gastroresistant CAP/Hd/Enhancer Microparticles
In the second set of experiments, Hd was suspended in
2% w/v CAP aqueous buffer feed solution with 3:1 polymer/
drug ratio in the presence of different enhancers of dissolu
tion rate (CMC 0.4% w/v; SDBS or Tween 85 in 0.5% w/v;
Table I) and spray dried to obtain CAP/Hd/CMC (batch #2),
CAP/Hd/SDBS (batch #3), CAP/Hd/Tween (batch #4) gastro
resistant microparticles.
For the preparation of batch #2, Hd and CMC were
previously blended (Hd/CMC weight ratio 1:2.5), using a
Galena Top (Ataena, Tecno Pro srl, Italy) mixer in the
presence of micronizing spheres until uniformity, and the
resultant mixture was suspended into the polymeric CAP
liquid feed.
For the preparation of both CAP/Hd/SDBS (batch #3)
and CAP/Hd/Tween 85 (batch #4) microparticles, Hd was
previously dampened with SDBS (0.5% w/v) or with Tween
85 (0.5% w/v), and then suspended into the polymeric CAP
liquid feed.
Controls
As control, CAP free microparticles made by Hd and
each of the enhancers (Hd/CMC batch #10, Hd/SDBS batch
#11, Hd/Tween batch #12) and drug free blank microparticles
made by CAP and the enhancers (CAP/CMC Batch #6, CAP/
SDBS #7 and CAP/ Tween #8) were also prepared in the
same experimental conditions by spray drying.
The solvent, CAP concentration, and the polymer/drug
weight ratio were kept constant for each formulation. Each
liquid feed was prepared under continuous magnetic stirring
and sonicated for 10 min before to be processed by spray
drying using 6 8 g as total amount of raw materials contained
in 200 ml of total feed volume sprayed.
The spray drying conditions were: a Buchi Mini Spray
Dryer B 191 (Buchi Laboratoriums Tecnik, Flawil, Switzer
land); inlet temperature 125C; outlet temperature 78 84C;
spray ow feed rate 5 ml/min; nozzle diameter 0.5 mm; drying
air ow 500 l/h, air pressure 6 atm, aspirator 100%.
Each preparation was carried out in triplicate. All the
spray dried microparticles were collected and stored under
vacuum for 48 h at room temperature.
The yield of the process, the percentage of theoretic drug
content, the percentage of actual drug content, encapsulation
efciency of gastroresistant and CAP free microsystems and
the mean size (d50) were determined (Table I).
Microparticle Characterization
Particle Size Analyses
Particle size analyses of commercially available Hd raw
material and of all spray dried microparticles prepared were
carried out with a laser light scattering granulometer (Beck
man Counter LS 230, Particle Volume Module Plus, U.K.).
393
The drug and microparticles were suspended in acid water
(pH 3.0); a few drops of each sample were poured into the
small volume cell to obtain an obscuration between 8% and
12%. Particle size distributions were calculated by instrument
software, using the Fraunhofer model. The analysis was made
in triplicate. The results were expressed as the median
diameter of the particles (d50).
Morphology
The morphology of the microsystems was assessed by
scanning electron microscopy (SEM): the microparticles were
coated with Au/Pd and then observed using a scanning
electron microscope (Scanning Microscope JSM 6400 Jeol,
Japan) at different extensions.
For the uorescent microscopy assays (FM), samples
were observed with a Zeiss Axiophot uorescence micro
scope, with 40, 63, and 1001.4 NA plan Apochromat oil
immersion objectives (Carl Zeiss Vision, Mnchen Hallberg
moos, Germany) using standard DAPI (4,6 diamidino 2
phenylindole) optics that adsorb violet radiation (max
372 nm) and emit a blue uorescence (max 456 nm).
Drug Content Evaluation
The drug content in each formulation was assessed by
UV and HPLC.
UV Method. Samples (40 mg) of each batch of micro
particles were dissolved in aqueous buffer solution at pH 7.5,
and the drug content was determined spectrophotometrically
(UV/Vis spectrometer Lambda 25, Perkin Elmer Instru
ments, MA, USA) at 310 nm (1 cm cell; Spectracomp 602,
Advanced Products srl, Milan, Italy). Each analysis was made
in triplicate and the results expressed as average value. Values
were superimposable to those obtained by HPLC analyses.
HPLC Method. HPLC Apparatus (Agilent 1100 series
system) equipped with a Model G pump, a DAD G 1315 A
detector set at max 284 nm loop 20 l, and a 1503.9 mm i.d. C
18 Bondapack column. The eluent was MeOH/H2O (1:1 v/v),
ow rate of 1.0 mL min 1. Linearity: reference standard
solutions were prepared at three concentration levels (0.05
1.0 mg/mL) and were injected (20 l) three times. The standard
curve was analyzed using the linear least squares regression
equation derived from the peak area (regression equation y=
1798.3x 54.938, R2 =0.9996, where y is the peak area and x the
concentration used). Specificity: the peaks associated with Hd
were identied by retention times, UV and MS spectra
compared with standard and conrmed by co injections.
Analysis of the microspheres: samples (15 mg) of three batches
of microparticles were dissolved in 15 mL MeOH, sonicated for
5 min, centrifuged for 10 min at 300 rpm. Hd concentration was
determined in the supernatant solutions using the same
chromatographic conditions of standard Hd. Each analysis was
made in triplicate and the results (ADC), expressed as average
value, are reported in Table I.
The production yields were expressed as the weight
percentage of the nal product compared to the total amount
of polymer and drug sprayed. The encapsulation efciency of
Hd present in all spray dried powders, was calculated from
Sansone et al.
394
the ratio of actual to theoretical drug content (Table I) in dry
microspheres.
Differential Scanning Calorimetry (DSC)
Differential scanning calorimetry was performed on an
indium calibrated Mettler Toledo DSC 822e (Mettler Toledo,
OH, USA) in order to study the thermal behavior on samples
of raw materials, drug free , CAP free , and drug loaded
microparticles. DSC thermograms were recorded by placing
accurately weighed quantities (8 10 mg weighed with a
microbalance MTS Mettler Toledo, OH, USA) of each
sample in a 40 l aluminum pan which was sealed and
pierced. The samples were exposed to two thermal cycles.
For the dehydration cycle, the samples were heated up to
130C with a heating rate of 20C/min and the temperature
was maintained at 130C for 15 min in order to remove the
residual solvent. Afterwards, the samples were cooled at 25C
and heated up to 350C with a heating rate of 10/min. From
this second thermal cycle, the glass transition temperature
(Tg), the melting temperature (Tm), and the heat of fusion
(Hm) were measured for all samples and compared with
each other. The analyses were carried out in triplicate.
test apparatus n.2: paddle, 100 rpm at 37C. The pH change

method (USP 31 drug release test, method A for Enteric
coated Articles) was used, 750 ml of HCl 0.1 N (pH 1.00)
from 0 to 2 h, then the addition of 250 ml of 0.2 M tribasic
sodium phosphate solution to give a nal pH of 6.8 in a total
volume of 1,000 ml. All the dissolution/release tests were
made in triplicate; only the mean values are reported in graph
(standard deviations <5%).
Samples of spray dried microparticles corresponding to
about 15 mg of hesperidin were analyzed spectrophotomet
rically at 310.
RESULTS AND DISCUSSIONS
In a preliminary set of experiments, Hd was spray dried
using CAP (1 and 2% w/v) as coating polymer in different
polymer/drug ratios (1:1, 3:1, 5:1) in order to achieve gastro
resistant microsystems able to avoid the avonoid exposure
to harsh gastric conditions. Only using 2% polymeric
solutions, satisfactory yields of the process and drug encap
sulation efciency were obtained.
Gastroresistant CAP/Hd Microparticles
Fourier Transform Infrared (FTIR) Spectroscopy
Microparticles Production
Fourier transform infrared spectra were obtained using a

Jasco FT 300, (Tokyo, Japan) Fourier transform IR (FTIR)
spectrometer. Samples of each batch of microparticles and the
raw materials were analyzed as KBr discs in the spectral
region 650 4,000 cm 1.
The actual and theoretical drug contents of each batch,

production yields, and encapsulation efciency are reported
in Table I. The results showed the encapsulation efciency
ranging from 70 to 74% for all batches #1a, 1b, and 1c. The
highest actual drug content and encapsulation efciency was
observed for batch #1a (polymer/drug ratio 3:1). As reported
in a previous work (18), very slightly soluble drugs such as
avonoids (2), due to the phase separation, can undergo to
deposition in the chamber during the spray drying process
giving a reduction of the encapsulation efciency. CAP seems
be able to limit this loss process giving acceptable results.
The product yields were higher for batch #1a (94.3%) than
#1b (73.0%) and #1c (84.4%) and satisfactory in view of the low
quantity materials used for the preparation of each batch,
minding the usual loss of the smallest and lightest particles
through the exhaust devices of the spray dryer apparatus.
X Ray Diffraction
X ray diffraction patterns on powder were recorded on a
Rigaku MiniFlex diffractometer (Tokyo, J) using a CuK
radiation (30 kV, 15 mA) at a scanning speed of 0.05/2 s with
a scanning range (2) from 2 to 55.
Long-Term Stability Studies
The stability test was performed according with conven
tional method (long term studies 12 months) reported by
guide lines ICH (International Conference on Harmonization
of Technical Requirements of Pharmaceutical for Human
Use). The formulations were stored at 25C2C /60%RH
5%RH (20) in a climatic chamber (Climatic and Thermostatic
Chamber, Mod.CCP37, AMT srl, Milan, Italy); HPLC and
DSC analysis of each batch were carried out for 12 months,
including four time points (0, 3, 6, and 12 months). Chromato
graphic peaks were identied on the basis of the retention
times and conrmed by co injections. No peaks corresponding
to degradation products of the drug were observed.
In Vitro Dissolution/Release Tests
In vitro dissolution/release tests of Hd from micro
particles were carried out under sink conditions using a
SOTAX AT Smart Apparatus (Basel, CH) on line with a
spectrophotometer (UV/Vis spectrometer Lambda 25, Per
kin Elmer Instruments, MA, USA) and USP 31 dissolution
Morphological Characterization
Laser scattering particle size analysis showed that CAP/
Hd gastroresistant microparticles (batches #1a, #1b, and #1c)
obtained spray drying CAP 2% w/v and Hd in different weight
ratios (1:1, 3:1, 5:1) have a very narrow size distribution with a
mean diameter in the range 11.50 13.55 m (Tab. I).
These gastroresistant microsystems as well as raw
materials (neat Hd, CAP), spray dried Hd, and CAP micro
particles were analyzed by FM (uorescence microscopy) and
SEM (scanning electron microscopy; Figs. 2, 3, and 4). Some
morphological, e.g. diameter, shape of particles and surface
characteristics were observed.
Raw Materials. Hd is commercially available in a crys
talline state with a needle shape (d50 10.211.01) character
ized by a very slight water solubility (Fig. 2a). Hd processed
by spray drying (Fig. 2b) and analyzed by SEM does not
395
Fig. 2. Scanning electron microscopy micrographs of Hd raw material (a, 1,000) and Hd spray dried (batch
#9, b, 1,000)
show changes in its solid crystalline state but only a mean

diameter reduction (d50 4.370.12) with respect to the raw
material was observed (Fig. 2a).
The SEM pictures of CAP raw material show the
presence of akes (Fig. 3a) while spray dried CAP showed
microparticles with spherical trend, some wrinkled and with a
variable particle size however lower than 10 m (Fig. 3b).
Batches #1a, #1b, and 1#c. SEM and FM micrographs of
batch #1a showed particles which are not regular in the shape
but are almost spherical. Most of the particles are well formed
and separated and some are partially collapsed (Fig. 4). Few
uncoated avonoid crystals were observed (Fig. 4). The
particle size is about 12 m, as conrmed by laser scattering
analysis. Similar results were obtained for batches #1b and
#1c (data not shown).
Thermal, FTIR, and X Ray Analyses
Raw Materials. The X ray diffraction and DSC analyses
indicated the same crystallinity grade for both neat commercial
Hd and spray dried Hd showing superimposable diffraction
(Fig. 5) and DSC (Fig. 6) proles. However, in the DSC curves,
the melting point of spray dried Hd showed a small shift to a
lower value (Tm 255.29C) and a half value of heat of fusion
(Hm) by comparison with neat Hd (Tm 259.17C). This shift

and the lower melting peak intensity (Fig. 6) may be due to the
size reduction of Hd crystals obtained by spray drying
Batches #1a, #1b, and #1c. The PXRD analysis (Fig. 7) of
batch #1a showed an high reduction in crystalline state of the
drug. The diffraction peaks already present, with very low
intensity with respect to the Hd raw material, are probably
due to the few amount of uncoated drug in crystalline state, as
conrmed by FM and SEM analyses. FTIR spectra did not show
any new peaks with respect to raw materials. Similar results
were obtained for batches #1b and #1c (data not shown).
In conclusion, results from SEM, FM, DSC, and PXRD
analyses showed that spray drying of the feeds containing Hd
and CAP (batches #1a, #1b, and #1c) produced a high
reduction in drug crystallinity and the encapsulation of the
drug in an amorphous state.
Dissolution/Release Studies
The release prole of CAP/Hd gastroresistant microsys
tems obtained with 3:1, 1:1, and 5:1 polymer/avonoid weight
ratio are reported in Fig. 8 with respect to the prole of neat Hd.
Fig. 3. Scanning electron microscopy micrographs of the CAP raw material (a 500) and CAP spray dried
(batch #5, b, 3,000)
396
Fig. 4. Scanning electron microscopy micrograph (2,000) of batch

#1a (25% hesperidin loaded CAP microparticles)
CAP seems to be able to protect Hd in the simulated

gastric uid, and the dissolution proles of all CAP/Hd
microparticles (batches #1a, #1b, and #1c), were almost
superimposable at pH 1.0 (Fig. 8). The amount of Hd
released/dissolved from these three batches in 2 h was about
3.0 4.0% , while about 20.0% of Hd alone was dissolved at
the same time.
At pH 6.8, the drug release proles from all CAP/Hd
microparticles (batches #1a, #1b, and #1c) were almost
superimposable with the dissolution prole of neat Hd with
slight difference due to microparticles composition. About
23.0% and 25.0% of drug was dissolved/released from
batches #1b and #1a, respectively, and about 28.0% from
#1c at 45 min after pH change, in comparison with about
25.0% of neat Hd that dissolved at the same time. Thus, CAP
was able to protect Hd in gastric environment, but the release
of avonoid was incomplete in the intestinal uid from all
CAP/Hd gastroresistant formulations.
For this reason, considering that the best results in term
of yield of the process, actual drug content and encapsulation
efciency (Table I) were obtained for batch #1a (CAP/Hd 3:1
weight ratio), in the second set of experiments, microparticles
were formulated in the presence of enhancers of the
dissolution rate.
Fig. 5. X ray diffraction patterns on hesperidin (Hd) raw material

and hesperidin spray dried (batch #9)
Sansone et al.
Fig. 6. Differential scanning calorimetry thermograms of hesperidin

raw material (Hd) and hesperidin spray dried (batch #9)
Gastroresistant CAP/Hd/Enhancers Microparticles

The feed liquids were prepared by dispersing Hd in a 2%
CAP buffer solution (3:1 polymer/drug weight ratio) and in
the presence of two different surfactants such as SDBS or
Tween 85 or a superdisintegrant such as CMC. These feeds
were spray dried to give batches #2 (CAP/Hd/CMC), #3
(CAP/Hd/SDBS), and #4 (CAP/Hd/Tween).
SDBS was selected as anionic surfactant, analogue to
SLS sodium lauryl sulfonate yet used to enhance the
dissolution rate of other drugs such as propanolol and
theophilline (21); Tween 85 (an esteried and polyethoxy
lated derivative of sorbitan) as non ionic surfactant used in
pharmaceutical systems for its compatibility, stability, and
minimal binding to proteins (22). Carboxymethylcellulose
sodium (CMC) is an insoluble, hydrophilic, highly absorbent
material reported to possess excellent water swelling capac
ities with an enhancement of the drug dissolution rate due to
its brous nature (23).
As control, CAP free (batches Hd/CMC #10, Hd/SDBS
#11 and Hd/Tween #12) and drug free (batches CAP/CMC
#6, CAP/SDBS #7 and CAP/Tween #8) microparticles were
also prepared.
Fig. 7. X ray diffraction patterns on batch #1a (CAP/Hd), Hd raw

material and CAP raw material
Fig. 8. Dissolution prole of CAP microparticles loaded with 50

(#1b), 25 (#1a), 16.6% (#1c) of hesperidin prepared by spray drying
buffer feed solutions (IF, pH 7.5) in comparison with the dissolution
prole of hesperidin alone (Hd)
397
ized by spherical microparticles, some typically wrinkled, with
a particle size lower than 5 m (Fig. 9b).
To produce batch #2, Hd was preliminarily loaded on
CMC (CMC/Hd weight ratio 2.5:1) by mixing with microniz
ing spheres until uniformity and then this physical mixture
was suspended into the CAP liquid feed and spray dried. FM
analysis of CMC/Hd mixture showed that Hd is well loaded
on CMC, almost keeping its crystalline state (Fig. 10a). As
control, the mixture CMC/Hd (batch #10), previously pre
pared, was also spray dried and morphologically character
ized by FM. The results showed that the spray drying process
is able to reduce the crystals dimension and improve the
loading of Hd on the polymer, also promoting the formation
of some spherical microparticles (Fig. 10b).
The micrographs of batch #2 conrmed the presence of
bigger microparticles (size about 80 m) with respect to the
microsystems prepared without enhancers, and showed a clear
content of Hd crystals well coated by CAP (Fig. 10c and d).
Microparticles Production
The actual and theoretical drug contents of each batch,
production yields and encapsulation efciency are reported in
Table I. The production yields were higher for CAP/Hd/
SDBS (batch #3, 93.0%) than for CAP/Hd/Tween (batch #4,
76.0%) or CAP/Hd/CMC (batch #2, 75.0%). Satisfactory
results were obtained for CAP free microsystems (batches
#10, 11, and 12).
Microparticles Morphology
Laser scattering particle size analysis showed that Hd
microparticles formulated with CAP in the presence of CMC
or SDBS or Tween (batches #2, 3, and 4) had a much higher
particle size (80.96, 42.86, and 43.18 m, respectively) with
respect to the microsystems prepared without enhancers (#1a,
12.280.07 m). These results suggest the presence of bigger
microparticles or aggregates.
Batch #2. Neat CMC crosslinked showed an irregular
and lamellar structure with dimensions of about 50 m (1.12;
Fig. 9a), whereas spray dried CMC crosslinked is character
Batches #3 and #4. Observing the FM micrographs of

batch #3 (CAP/Hd/SDBS), the absence of aggregates is
evident; CAP is able to coat drug giving big and well
separated microparticles (size about 40 m) characterized
by a spherical or stretched shape with a smooth and exible
surface and crystals of the drug inside (Fig. 11a and b).
The micrographs of batch #4 (CAP/Hd/Tween85) showed
the presence of smaller microparticles (Fig. 11c) with dimen
sions of about 2 m but with a dimensional distribution by
laser scattering analysis of 42.86 m1.07, due to the presence
of a few amount of crystals embedded on the surface as well as
aggregates. The laser diffraction technique, in fact, interprets
the particles as spheres, causing a heterogeneous particle
distribution for batches containing crystals. Some micropar
ticles are pierced from the uncovered crystals and partially
collapsed.
Thermal, FTIR, and X Ray Analyses

The absence of Hd melting point in the thermal prole of
batches #2 and #3 suggested that the avonoid was well
encapsulated as previously shown by FM analysis. Only for
Fig. 9. Scanning electron microscopy micrographs of neat (a 1,000) and spray dried CMC crosslinked
(9b, 1,000)
Sansone et al.
398
Fig. 10. Fluorescence microscopy images of the physical mixture of Hd/CMC (a, 63); Hd/CMC spray dried
(batch #10, b, 40); CAP/Hd/CMC microparticles (batch #2, c, 40) and their inside (d, 63)
Fig. 11. Fluorescence microscopy images of CAP/Hd/SDBS (batch #3) microparticles at different
magnication (a, 40; b, 100). Scanning electron microscopy (c, 2,500) micrographs of CAP/Hd/Tween
microsystems (batch #4)
399
Fig. 12. Differential Scanning calorimetry thermograms of raw materials Hd, CMC, and
CAP and CAP/Hd/CMC microparticles (batch #2)
batch #4, a small peak in correspondence of the Hd melt

value, ascribable to few drug crystals not completely encap
sulated, was observed. In addition, the DSC proles of
batches #3 and #4, showed the thermal trend of CAP shifted
to lower temperature; the shift may be due to the physical
interaction between the polymer and the surfactants (Figs. 12
and 13). These results do not exclude the presence of drug in
residual crystalline state, because, sometimes, the DSC
technique could not detect any residual crystalline structure,
due to the ability of low melting point polymers to act as
solvent for a drug (16). In fact, by increasing the temperature,
CAP which melts at lower temperature than avonoids starts
to melt and dissolve drug crystals. For this reason, XRD
analysis was necessary for further investigation.
All PXRD proles of batches #2, #3, and #4 showed an

increase in crystallinity of the drug compared to PXRD
prole of batch #1a (drug in the amorphous state), but
however reduced with respect to crystalline Hd raw material
(Fig. 14).
Furthermore, PXRD of batch #2 showed the presence of
other two peaks (2/degree 36.65 and 43) corresponding to
part of CAP that probably was not involved in the Hd
microspheres formation. This behavior suggests the hypothesis
of a competition between CMC and CAP in the formation of
Hd microsystems during the spray drying process.
FTIR spectra of batch #2, Hd and raw polymers (CAP
1,735, 1,587, 1,249, and 1,126 cm 1; CMC 1,587, 1,386, and
1,070 cm 1), did not show any differences in the peak patterns
conrming the absence of chemical interaction between the
coating materials and the drug. Also, in the case of batches #3
and #4, FTIR analysis showed neither new bands ascribable
to any chemical interaction. All these results showed that the
major difference between the gastroresistant microsystems
Fig. 13. Differential Scanning calorimetry thermograms of raw

materials SDBS, Hd, and Tween 85, CAP blank (batch #5), CAP/
Hd/SDBS (batch #3), and CAP/Hd/Tween (batch #4) microparticles
Fig. 14. X ray diffraction patterns on Hd raw material, CAP raw

material, batches #1a (CAP/Hd), #2 (CAP/Hd/CMC), #3 (CAP/Hd/
SDBS), and #4 (CAP/Hd/Tween 85)
Sansone et al.
400
formulated with CMC or SDBS or Tween 85 resides on their
morphology, particle size and on the solid state of Hd. All the
above data indicated that spray drying technique was not
suitable to change solid crystalline state of neat Hd processed
alone, but was able to reduce only the crystal size. Spray
drying of gastroresistant polymer (CAP) and Hd (batch #1a)
produced small microparticles with the drug and the polymer
in amorphous state, some aggregates and very few crystals of
the drug not coated by the polymer. Spray drying CAP and
Hd in the presence of the enhancers produced well formed
microparticles with different sizes, bigger in the presence of
CMC (80.962.40 m) than in the presence of SDBS (43.18
1.83 m), or Tween 85 (42.861.07 m; Table I), containing
the drug keeping its crystalline state and generally well coated
by the polymer. Gastroresistant microsystems formulated
with SDBS showed exible coating and absence of
aggregates with respect to those formulated in the presence
of Tween 85 in which aggregates and a few amount of crystals
not coated and embedded on particles surface were observed.
In addition, CMC seems to compete with CAP in the Hd
microsystem formation.
Dissolution/Release Studies
The dissolution/release prole of Hd from batches #2
(CAP/Hd/CMC), #3 (CAP/Hd/SDBS), and #4 (CAP/Hd/
Tween) are reported in Fig. 13 in comparison with dissolu
tion/release proles of neat Hd and batch #1a (CAP/Hd)
prepared without the enhancers.
The dissolution proles of batches #1a, #2, and #4 were
almost superimposable at pH 1.0 (Fig. 15); in fact, the amount
of Hd released/dissolved from these three batches in 2 h, was
about 3.0%. At the same time, about 6.0% of Hd was
released/dissolved from CAP/Hd/SDBS (batch #3) micro
particles, in comparison with about 20.0% of neat Hd. As
expected, a very high amount of avonoids was released from
the non gastroresistant microparticles (batches #10, #11, and
#12) into the gastric uid (Fig. 16). In particular, the highest
Hd release from CAP free microparticles was observed in the
presence of Tween (batch #12, 87.0%) with respect to batches
#11 (Hd/SDBS, 46.0%) and #10 (Hd/CMC, 31.0%). Also,
after pH change, the release of Hd increased from all these
three non gastroresistant batches with respect to neat Hd,
due to the action of enhancers (Fig. 16).
Furthermore, after pH change, an improvement of Hd
dissolution rate was observed from gastroresistant micro
Fig. 16. Release prole of non gastroresistant (CAP free) micro

particles (batches #10, #11, #12) in comparison with Hd raw (Hd)
particles formulated with all different enhancers of the

dissolution rate (Fig. 15). At pH 6.8, while dissolution release
from batch #1a was incomplete, about 100% of Hd dissolved/
released from the microparticles formulated in the presence
of CMC (batch #2), Tween 85 (batch #4) and SDBS (batch
#3) in 45 min after pH change. This behavior may be
explained by an increase of the drug water interaction due
to the presence of the enhancers of the dissolution rate that
improved wettability and solubility of Hd. Moreover, the
enhancement of the drug dissolution rate in IF was higher for
all gastroresistant microparticles than that observed for
microsystems prepared without CAP (batches #10, #11, and
#12; 61.0 91.0%). This result may be due to the action of the
enhancers of the dissolution rate of CAP itself.
These results indicated that CAP was able to protect Hd
in gastric environment for all gastroresistant microparticles
prepared, with slight differences due to microsystem compo
sition also in the presence of the enhancers. These excipients
were able to increase the dissolution rate of Hd in the
simulated intestinal environment without altering CAP ability
to protect Hd in the acidic uid.
Long-Term Stability Studies
To examine the effect of the microencapsulation process
on the stability of Hd, long term studies were performed on all
microsystems produced. The formulations were analyzed for
hesperidin content over 12 months, including four time points
(0, 3, 6, and 12 months) by HPLC method. No decrease of Hd
content as well as no drug degradation product were recorded.
CONCLUSION
Fig. 15. Release prole of all CAP/Hd gastroresistant microparticles

(batches #1a, #2, #3, #4) in comparison with Hd raw (Hd)
CAP may be used as polymeric carrier with adequate

coating properties able to protect Hd in the gastric medium.
The spray drying technique and process conditions selected
were effective in microencapsulating and stabilizing the labile
avonoid giving satisfactory encapsulation efciency, product
yield, microparticles morphology, and drug stability over
12 months. The presence of the enhancers such as CMC,
SDBS, and Tween 85 in the formulation seems to be essential
to obtain complete drug release in the simulated intestinal
uid. This approach might be suitable for delivering avo
noids such as Hd to the intestine, avoiding gastric degradation
of the molecule and improving its bioavailability from solid
oral dosage forms.

REFERENCES
1. Barthe GA, Jourdan PS, McIntosh CA, Mansell RL. Radioim
munoassay for the quantitative determination of hesperedin and
analysis of its distribution in Citrus sinensis. Phytochemistry
1988;27:249 54.
2. Garg A, Garg S, Zaneveld LJD, Singla AK. Review article:
chemistry and pharmacology of the citrus bioavonoid hesper
edin. Phytother Res. 2001;15:655 69.
3. Bok SH, Lee SH, Park YB, Bae KH, Son KH, Jeong TS, et al.
Plasma and hepatic cholesterol and hepatic activities of 3
hydroxy 3 methyl glutaryl CoA reductase and acyl CoA: choles
terol transferases are lower in rats fed citrus peel extract or a
mixture of citrus bioavonoids. J Nutr. 1999;129:1182 5.
4. Hirata A, Murakami Y, Shoji M, Kadoma Y, Fujisawa S. Kinetics
of radical scavenging activity of hesperetin and hesperidin and
their inhibitory activity on COX 2 expression. Anticancer Res.
2005;25:3367 74.
5. Horcajada MN, Coxam V. Hesperidin, a citrus avanone,
improves bone acquisition and prevents skeletal impairment in
rats in nutritional aspects of osteoporosis. 2nd ed. New York:
Elsevier; 2004. p. 103 20.
6. Yang M, Tanaka T, Hirose Y, Deguchi T, Mori H, Kawada Y.
Chemopreventive effects of diosmin and hesperidin on N butyl
N (4 hydroxybutyl)nitrosamine induced urinary bladder carcino
genesis in male ICR mice. Int J Can. 1997;73:719 24.
7. Franke AA, Cooney RV, Custer LJ, Mordan LJ, Tanaka Y.
Inhibition of neoplastic transformation and bioavailability of
dietary avonoid agents. Adv Exp Med Biol. 1998;439:237 28.
8. Choi JS, Park KY, Moon SH, Rhee SH, Young HS. Antimuta
genic effects of plant avonoids in the Salmonella assay system.
Arch Pharm Res. 1994;17:71 5.
9. Calomme M, Pieters L, Vlietinck A, Berghe DV. Inhibition of
bacterial mutagenesis by Citrus avonoids. Planta Med.
1996;62:222 6.
10. Tanaka T, Makita H, Kawabata K, Mori H, Kakumoto M.
Modulation of N methyl N amylnitrosamine induced tumorigen
esis by dietary feeding of diosmin and hesperidin, alone and in
combination. Carcinogenesis 1997;18:957 65.
11. Beninati S. Transglutaminase activity and protein polyamine
binding capacity in animal and plants cells. In: Pandalai G, editor.
Recent developments in phytochemistry. vol 1. Kerala: Research
Signpost; 1997. p. 243 53.
401
12. Lentini A, Forni C, Provengano B, Beninati S. Enhancement of
transglutaminase activity and polyamine depletion in B16 F10
melanoma cells by avonoids naringenin and hesperitin correlate
to reduction of the in vivo metastatic potential. Amino Acids.
2007;32:95 100.
13. Thacher SM, Rice RH. Keratinocyte specic transglutaminase
of cultured human epidermal cells: relation to cross linked
envelope formation and terminal differentiation. Cell 1985;40:
685 95.
14. Benedetti L, Grignani F, Scicchitano BM, Jetten AM, Diverio D,
Lococo F, et al. Retinoid induced differentiation of acute
promyelocytic leukemia involves PML RAalpha mediated in
crease of type II transglutaminase. Blood 1996;87:1939 50.
15. Arani I, Adler Storthz K, Trying SK, Brysk H, Brysk MM.
Differentiation markers in oral carcinoma cell lines and tumors.
Anticancer Res. 1997;17:4607 10.
16. Kanaze FI, Kokkolou E, Niopas I, Georgarakis M, Stergiou A,
Bikiaris D. Thermal analysis study of avonoid solid dispersion
having enhanced solubility. J Therm Anal Calorim. 2006;
83:283 90.
17. Lauro MR, Maggi L, Conte U, De Simone F, Aquino RP.
Quercetin gastro resistant microparticles obtained by spray
drying technique. J Drug Del Sci Tech. 2005;15:363 9.
18. Lauro MR, De Simone F, Sansone F, Iannelli P, Aquino RP.
Preparation and release chacteristics of naringin and naringenin
gastro resistant microparticles by spray drying. J Drug Del Sci
Tech. 2007;17:119 24.
19. Palmieri GF, Bonacucina G, Di Martino P, Martelli S. Gastro
resistant microspheres containing ketoprofene. Journal of Mi
croencapsulation. 2002;19:111 9.
20. Giunchedi P, Conte U. Spray drying as preparation method of
microparticulate drug delivery systems: an overview. STP Pharm
Sci. 1990;5:276 90.
21. Gaylord N, Schor LM. Controlled release solid drug dosage
forms based on mixture of water soluble non ionic cellulose
ethers and anionic surfactants. US Patent 1989;4849229.
22. Tommasini S, Calabr ML, Raneri D, Ficarra P, Ficarra R.
Combined effect of pH and polysorbates with cyclodextrins on
solubilization of naringenin. J Pharm Biomed Anal. 2004;36:
327 33.
23. Sangalli ME, Giunchedi P, Colombo P, Gazzaniga A, La Manna
A. Cross linked sodium carboxymethylcellulose as a carrier for
dissolution rate improvement of drugs. Boll Chim Farm.
1989;128:242 7.

DOI: 10.1208/s12249 009 9220 7
Research Article
Development of Microsponges for Topical Delivery of Mupirocin
Netal Amrutiya,1 Amrita Bajaj,1,2 and Madhu Madan1
Abstract. The goal of the present study was to develop and evaluate microsponge based topical delivery
system of mupirocin for sustained release and enhanced drug deposition in the skin. Microsponges
containing mupirocin were prepared by an emulsion solvent diffusion method. The effect of formulation
and process variables such as internal phase volume and stirring speed on the physical characteristics of
microsponges were examined on optimized drug/polymer ratio by 32 factorial design. The optimized
microsponges were incorporated into an emulgel base. In vitro drug release, ex vivo drug deposition, and
in vivo antibacterial activity of mupirocin loaded formulations were studied. Developed microsponges
were spherical and porous, and there was no interaction between drug and polymer molecules. Emulgels
containing microsponges showed desired physical properties. Drug release through cellulose dialysis
membrane showed diffusion controlled release pattern and drug deposition studies using rat abdominal
skin exhibited signicant retention of active in skin from microsponge based formulations by 24 h. The
optimized formulations were stable and nonirritant to skin as demonstrated by Draize patch test.
Microsponges based emulgel formulations showed prolonged efcacy in mouse surgical wound model
infected with S. aureus. Mupirocin was stable in topical emulgel formulations and showed enhanced
retention in the skin indicating better potential of the delivery system for treatment of primary and
secondary skin infections, such as impetigo, eczema, and atopic dermatitis.
KEY WORDS: emulgel; ex vivo drug deposition; microsponges; mouse surgical wound model;
mupirocin.
INTRODUCTION
Microparticles and nanoparticles have been increasingly
investigated to achieve targeted and sustained release of
drugs (1). Microsponges are porous, polymeric microspheres
that are mostly used for prolonged topical administration.
Microsponges are designed to deliver a pharmaceutically
active ingredient efciently at minimum dose and also to
enhance stability, reduce side effects, and modify drug release
proles (2). These attributes have been successfully demon
strated in the FDA approved Retin A Micro (0.1% or
0.04% tretinoin) and Carac (0.5% 5 urouracil) products for
acne treatment and actinic keratoses, respectively.
Many of conventional delivery systems require high
concentrations of active agents to be incorporated for
effective therapy because of their low efciency as delivery
1
Department of Pharmaceutical Technology, C.U. Shah College of

Pharmacy, S.N.D.T Womens University, Santacruz (W), Mumbai,
400049 Maharashtra, India.
2
To whom correspondence should be addressed. (e mail: bajajamri
ta@rediffmail.com)
ABBREVIATIONS: CFU, colony forming units; EE, entrapment
efciency; ESD, emulsion solvent difusion; JSS, steady state ux; KP,
permeability coefcient; MP emulgel, emulgel containing mupirocin;
MP, mupirocin; MP ointment, marketed mupirocin ointment
formulation; MP/EC, mupirocin/ethylcellulose; MS emulgel, emulgel
containing mupirocin loaded microsponges; PVA, polyvinyl alcohol;
PY, production yield.
systems. Thus, the need exists for delivery systems to

maximize the period of time that an active ingredient is
present, either on the skin surface or within the epidermis
while minimizing its transdermal penetration into the body.
The microsponge based polymeric microspheres uniquely
fulll such requirements (3).
Microsponges are prepared by several methods utilizing
emulsion systems as well as by suspension polymerization in a
liquid liquid system. The most common emulsion system
used is oil in water (o/w), with the microsponges being
produced by the emulsion solvent diffusion (ESD) method (4).
Mupirocin (MP), a topical antibiotic agent is effectively
used for treatment of primary and secondary skin infections
(5). MP is produced by Pseudomonas fluorescens and has in
vitro activity against a range of Gram positive and some
Gram negative bacteria. However, MP also inhibits the
growth of a number of pathogenic fungi in vitro, including a
range of dermatophytes and pityrosporum (6). MP inhibits
bacterial protein synthesis by binding to the enzyme, iso
leucyl t RNA synthetase, which prevents the incorporation of
isoleucine into proteins (7). Its chemical structure and
mechanism of action are unique among antibiotics in clinical
use; therefore, cross resistance between MP and other anti
biotics is not a concern (8). MP reaching the systemic
circulation is rapidly converted to monic acid, an inactive
metabolite that is rapidly excreted in the urine. Penetration
into the deeper epidermal and dermal layers is enhanced in
traumatized skin and under occlusive dressings. Mupirocin is
402

slowly metabolized by the skin to the antimicrobially inactive
metabolite monic acid (6). Therefore, controlling the release
of drug will improve the efcacy of formulation and decrease
the frequency of application.
The mupirocin 2% ointment with polyethylene glycol
applied three times daily has demonstrable activity and is
currently marketed. Topical ointment preparations are less
acceptable to patients, since ointments have high viscosities
leading to inappropriate application to skin lesions, and
patients may report garment soiling from greasy residues
(9). To enhance patient acceptance and compliance, novel
formulations could be developed such as creams, gels,
emulgels containing particulate drug carriers like micro
particles and nanoparticles. These carriers could protect and
release the active in a controlled manner. Benzoyl peroxide
microsponges have already been developed for topical
treatment of acne and athletes foot (10).
Literature search revealed that no study carried out to
formulate once a day sustained release medication containing
MP. Thus, the aim of the present investigation was to design
microsponges as novel carriers for topical delivery of MP.
This investigation consisted of preparation, optimization, and
evaluation of MP microsponges and incorporation of opti
mized microsponges in semisolid vehicle base to obtain
cosmetically acceptable products. Factorial design or re
sponse surface methodology was used to study the effect of
factors (formulation and process variables) inuencing
responses [entrapment efciency (EE), production yield,
and mean particle size] by carrying out limited number of
experiments (11). The optimized microsponges were incorpo
rated into an emulgel base that possesses the advantages of
both emulsions and gels with exclusively accomplished
requirements of vehicle for topical drug delivery systems (12).
403
Drug Content Analysis

The quantitative determination of MP in microsponges
was carried out using a reversed phase high performance
liquid chromatograph (HPLC) analysis method. The HPLC
instrument was equipped with a model series CCPE pump
(Tosoh, Tokyo, Japan), with a 20 L loop injector and a
linear model UV absorbance detector. Data were processed
using EZChrom Chromatography Data System, Version
6.7 (Scientic Software Inc., San Ramon, CA, USA).
Briey, the chromatographic separations were per
formed using Hypersil GOLD C18, 5 m, 2504.6 mm i.d.
column, eluted with mobile phase at the ow rate of 0.5 mL/
min, and the eluent was monitored using a UV detector at
221 nm. The mobile phase consisting of methanol water
(80:20, v/v) was adjusted to pH 5.0 with ortho phosphoric
acid. The samples were prepared in mobile phase and ltered
through 0.2 m cellulose acetate membrane lter. All the
determinations were performed at ambient temperature, and
the retention time was found to be 8.2 min for MP. A linear
relationship (r2 >0.9995) was found between the peak area vs.
concentration. External standardization by peak area was
used for quantitative determination of MP. The developed
HPLC method was validated for linearity, accuracy, precision,
and specicity as per ICH guidelines.
The actual drug content and EE were calculated as given
below:
Actual drug content % Mact =Mms 100
Entrapment efficiency % Mact =Mthe 100
where Mact is the actual amount of MP in weighed quantity of

microsponges, Mms is the weighed quantity of microsponges,
and Mthe is the theoretical amount of MP in microsponges.
Materials
MP was procured from Mctony Biotech (Tianjin) Co.
Ltd. (Tianjin, China). Ethyl cellulose 46 cp was obtained from
Sigma Aldrich (St. Louis, USA). Polyvinyl alcohol (PVA;
MW =approximately 1:25,000), methanol, and acetonitrile
were procured from S. D. Fine Chem Limited (Mumbai,
India). Sepigel 305 (polyacrylamide, C13 14 isoparafn
laureth 7) was purchased from Seppic (Paris, France). All
other chemicals and solvents were of analytical reagent grade.
Animal studies were performed in accordance with protocols
approved by Institutional Animal Ethics Committee (IAEC,
C. U. Shah College of Pharmacy, Mumbai, India).
Methods
MP microsponges were prepared by an ESD method.
The organic internal phase containing MP and ethyl cellulose
in dichloromethane was gradually added into external phase,
which contained PVA as emulsifying agent. The mixture was
stirred at 1,000 2,000 rpm for 3 h at room temperature to
remove dichloromethane from the reaction ask. The formed
microsponges were ltered, washed with distilled water, and
dried at room temperature. Microsponges were weighed, and
production yield (PY) was determined.
Optimization of Formulation Parameters and Process

Variables
Preliminary trials were undertaken to establish the effect
of mupirocin/ethylcellulose (MP/EC) ratios on the physical
characteristics of microsponges. In all formulations, the total
amount of polymer, internal phase volume, PVA concentra
tion, and stirring rate were kept constant.
To optimize the dependent variables such as production
yield, EE, and mean particle size of microsponges, a series of
formulations were prepared by 32 factorial design using
volume of internal phase (X1) and stirring rate (X2) as
independent variables.
Thermal analysis of MP, ethyl cellulose, and MP loaded
microsponge based formulations were studied employing
differential scanning calorimeter (Mettler Toledo DSC,
USA). Samples (5 mg) were accurately weighed into alumi
num pans and sealed. All samples were run at a heating rate
of 10C/min over a temperature range 25 300C in atmo
sphere of nitrogen.
404
The morphology of microsponges was examined using a
scanning electron microscope (GEOL 5400, USA) operating
at 20 kV. Dried microspheres were coated with gold
palladium alloy for 45 s under an argon atmosphere before
observation. SEM photograph was recorded at magnication
of 500.
Particle Size Studies and Porosity Determination
Particle size studies were carried out using laser light
scattering technique using Mastersizer 2000 (Malvern Instuments
Ltd., Worcestershire, UK). The porous properties were deter
mined by Mercury Intrusion Porosimeter (Autoscan 60,
Quantachrome, USA) in the pressure range 0 4,000 kg/cm2.
Preparation of Mupirocin Loaded Emulgel
Oil in water emulsion was prepared using Tween 20 and
Span 20 (3 4% w/w) as emulsifying agents and liquid parafn
(5% w/w) as an oily phase with stirring at 1,800 2,000 rpm for
20 min. The obtained emulsion was added to sepigel with
gentle stirring to get the emulgel base. At this point, either
MP (2% w/w) or MP microsponges (equivalent to 2% w/w of
MP) were incorporated to get homogenous emulgel based
topical delivery systems.
In vitro release studies were carried out using Franz
diffusion cells with a receptor compartment volume of 20 mL
and an effective diffusion area of 3.14 cm2. Cellulose dialysis
membrane (Himedia, Mumbai, India) was soaked in receptor
media (phosphate buffer, pH 5.8) for 24 h before experiment.
A predetermined amount of emulgel containing MP
microsponges (MS emulgel) was placed on the donor side.
The receptor medium was continuously stirred at 600 rpm
and thermostated at 320.5C with a circulating jacket. At
predetermined time intervals, 2 mL samples were withdrawn
from the receiver compartment and replaced with an equal
volume of fresh buffer. The collected samples were analyzed
by HPLC. The release kinetics of conventional emulgel
containing mupirocin (MP emulgel) and marketed mupirocin
ointment formulation (MP ointment) were used for
comparison. The drug release data were analyzed to
determine the release kinetics (zero order and rst order) as
well as diffusion controlled mechanism (Higuchi model) using
linear regression analysis.
Ex Vivo Drug Deposition Studies
Drug deposition study was performed on the excised rat
abdominal skin using Franz diffusion cell (13). Epidermal side
of the skin was exposed to ambient condition, while dermal
side was kept facing the receptor solution. Receptor com
partment containing 20 mL phosphate buffer pH 5.8 was
thermostated at 320.5C and stirred at 600 rpm. Skin was
saturated with diffusion medium for 1 h before the applica
tion of sample. A 200 mg of sample was applied on the donor
compartment. For determination of drug deposited in the
Amrutiya, Bajaj and Madan

skin, the diffusion cell was dismantled after a period of 4, 8,
16, and 24 h. The skin was carefully removed, and drug
present on the skin surface was cleaned with distilled water.
Quantification of Mupirocin in Skin Samples
MP was extracted from skin using a modication of the
procedure described by Echevarria et al. (14). Briey, the skin
was cut into small pieces and homogenized with 10 mL
methanol by tissue homogenizer. The homogenized sample
was subjected to ultrasonication for 10 min for complete
extraction of drug. This methanolic extract was centrifuged at
5,000 rpm for 10 min. The supernatant was collected,
evaporated, and reconstituted with mobile phase. The sample
was ltered through a 0.2 m membrane lter and analyzed
by developed HPLC method as outlined earlier. To deter
mine recovery from skin, the intact skin was spiked with
known quantities of the compound and homogenized,
extracted, and analyzed as described above.
Primary Skin Irritation Studies
Primary skin irritation studies of the optimized formula
tions were performed using New Zealand white albino rabbits
in accordance with the guidelines of the Consumer Product
Safety Commission (15). Formalin and placebo emulgel
formulation were used as positive and negative controls,
respectively. The scores were recorded as per the Draize
patch test.
In Vivo Studies
Swiss albino mice, weighing 18 20 g were used to
compare the efcacy of developed MP loaded formulations
with that of marketed MP ointment formulation using mouse
surgical wound model infected with Staphylococcus aureus
(9). The surgical wound model was constructed as previously
described by McRipley et al. (16). Briey, supercial surgical
wounds were produced on the backs of mice by making a
longitudinal midline incision 2.30.2 cm in length and extending
down to the panniculus carnosus. A contaminated suture was
inserted through the skin to infect the wound and secured by
knotting. Therapy was administered topically 0.1 mL/mouse.
Treatment was initiated at 4 h after surgery and continued for
further 3 days. On day 5 after surgery, 16 to 20 h after the last
topical application, the animals were killed by diethyl ether
asphyxiation. A 1 2 cm area of skin, including the wound, was
excised and homogenized in 1 mL of saline solution. The
homogenates were serially diluted, plated on agar plates, and
incubated at 37C. After 24 h, the number of colony forming
units (CFU) per wound was counted. Bacterial counts were
expressed in terms of meanSD.
Stability Studies
The prepared emulgels were packed in aluminum collaps
ible tubes (5 g) and subjected to stability studies at 5C, 25C/
60% RH, 30C/65% RH, and 40C/75% RH for a period of
3 months. Samples were withdrawn at 15 day time intervals and
evaluated for physical appearance, pH, rheological properties,
drug content, and drug release proles.
405
Table I. Optimization of Drug:Polymer Ratio for Preparation of MP Microsponges

Formulation
code
Drug: polymer
ratio (by weight)
MP1
MP2
MP3
MP4
MP5
MP6
MP7
MP8
a
Theoretical
drug content (%)
Actual drug
contenta (%SD)
Production
yielda (%SD)
Entrapment
efciencya (%SD)
Mean particle
sizea (mSD)
16.66
28.57
37.50
44.44
50.00
54.55
58.33
61.54
10.631.25
20.380.57
30.661.39
38.800.80
45.551.56
50.890.93
53.121.12
54.981.72
67.031.36
68.111.78
69.282.77
72.411.26
78.243.94
81.971.98
80.432.46
79.621.85
63.813.49
71.332.01
81.762.34
87.321.79
91.102.43
93.291.18
91.071.49
89.342.59
30.364.28
32.846.35
35.284.95
38.413.34
40.324.72
43.703.45
47.516.83
52.765.24
0.25:1
0.5:1
0.75:1
1:1
1.25:1
1.5:1
1.75:1
2:1
Each observation is the meanSD of three determinations
The data obtained from each experiment was subjected
to statistical analysis using one way analysis of variance
followed by Bonferroni multiple comparisons test. P<0.05
was considered to be signicant.
volume and stirring rate had a signicant antagonistic

inuence on % EE without producing any interaction. The
regression Eq. 3 for % EE is as follows:
Y%EE 93:75
1:60X1 2:19X2
A linear regression Eq. 4 generated for PY showed the

negative inuence of both the independent variables with a
R2 value of 0.9805.
RESULTS
Formulation and Optimization of Microsponges
The methods of preparation of microsponges are limited
in terms of complexity and cost. Some commercially available
microsponges are prepared by suspension polymerization
method, but ESD method serves as an alternative for
preparing microsponges (2,4). ESD method was found to be
an easy, reproducible, rapid technique for the preparation of
MP microsponges. Moreover, it had an advantage of avoiding
solvent toxicity.
The effect of MP/EC ratio on production yield, EE, and
mean particle size is shown in Table I. It is indicated that MP/
EC ratio (1.5:1) had the optimum capacity for drug entrap
ment. With further increase in drug/polymer ratio from 1.5:1
to 2:1, no signicant change in EE and production yield was
observed. Therefore, further optimization of formulation
MP6 was carried out using internal phase volume (X1) and
stirring rate (X2) as the independent variables by 32 factorial
design (13,17).
Responses of different batches obtained using factorial
design are shown in Table II. A linear model generated for %
EE as dependent variable revealed that the internal phase
Ypy 81:95
3:34X1
1:54X2
A polynomial regression Eq. 5 revealed that the negative

inuence of both the independent variables with positive
interaction on mean particle size of the microsponges.
Ymps 42:74
23:36X1
10:25X2 2:2 X1 X2
Results of multiple regression analysis of all the param

eters investigated are summarized in Table III. Response
surface methodology study indicated that formulation IS
8 was better in terms of EE, production yield, and particle
size. Thus, optimized microsponges batch (IS 8) was sub
jected to further characterization studies and incorporated
into an emulgel base.
Characterization of Microsponges
The thermal behavior of MP, ethyl cellulose, and
optimized MP microsponge formulations is shown in
thermograms (Fig. 1). MP presented a sharp endothermic
Table II. Optimization of Internal Phase Volume and Stirring Rate by 32 Factorial Design
Formulation
code
IS
IS
IS
IS
IS
IS
IS
IS
IS
a
1
2
3
4
5
6
7
8
9
Variable X1 internal
phase volume (mL)
5
5
5
7.5
7.5
7.5
10
10
10
(1)
(1)
(1)
(0)
(0)
(0)
(+1)
(+1)
(+1)
Variable X2 stirring
rate (rpm)
1,000
1,500
2,000
1,000
1,500
2,000
1,000
1,500
2,000
(1)
(0)
(+1)
(1)
(0)
(+1)
(1)
(0)
(+1)
Each observation is the meanSD of three determinations
Production
yielda (%SD)
Entrapment
efciency a (%SD)
Mean particle
sizea (mSD)
88.241.13
85.932.42
83.341.68
82.191.46
81.971.98
80.462.34
80.121.68
79.871.97
77.461.82
92.182.13
95.682.36
97.281.15
90.862.19
93.291.18
95.321.68
89.541.76
92.871.63
93.121.38
79.36.38
65.25.24
55.14.83
54.75.06
43.73.45
32.84.19
28.33.64
18.22.93
12.93.14
406

Table III. Summary of Regression Analysis Results for Measured Responses
Coefcients
Parameters
11
22
12
r2
Entrapment efciency
Production yield
Mean particle size
93.75
81.95
42.74
1.60
3.34
23.36
2.19
1.54
10.25
0.28
0.953
0.56
0.89
0.621
1.48
0.38
0.56
2.2
0.9871
0.9805
0.9990
peak at 63.4C corresponding to its melting point, while at

that temperature, optimized MP microsponges presented a
broad endothermic peak with the loss of its sharp
appearance. SEM of MP loaded microsponges as shown
in Fig. 2 revealed that the microsponges were uniform,
spherical in shape, and porous in nature. No intact MP
crystals were seen visually. The result obtained from
porosimetry experiments of different microsponges formu
lations is depicted in Table IV. The porosity of micro
sponges formulations was found to increase with increasing
the drug amount.
In Vitro Drug Release Studies

The inuence of composition and vehicle on release
prole of different formulations was investigated using a
cellulose dialysis membrane. In vitro release proles of MP
from different formulations shown in Fig. 3 indicated that MP
ointment and MP emulgel released the drug within 4 and
10 h, respectively, while MS emulgel was able to show a
sustained release up to period of 24 h. Drug release from
emulgels was found to be diffusion controlled mechanism
with R2 value ranging from 0.9738 to 0.9790. The steady state
ux (Jss) and permeability coefcient (Kp) were calculated,
and the results are tabulated in Table V.
Characterization of Emulgels
Ex Vivo Drug Deposition Studies
The developed emulgel formulations were white, vis

cous preparations with smooth, and homogeneous appear
ance. They were thixotropic, easily spreadable, and had
good aesthetic appeal. The pH and viscosity values of
prepared emulgels were found to be in the range of 5.5 6.0
and 7500 9000 cps, respectively. The drug content was 99
0.5%. The drug content uniformity test for all the emulgels
indicated that the drug was uniformly dispersed in emulgel
formulations.
Amount of MP deposited in excised rat abdominal skin

from different formulations at different time intervals has
been shown in Fig. 4. The amount of MP deposited in skin
was higher from the MS emulgel (204.46.9 g/cm2) as
compared to that with the MP emulgel (87.579.7 g/cm2)
or MP ointment (37.037.6 g/cm2) at the end of 24 h. This
indicates that microsponges improved the drug residence in
skin.
Primary Skin Irritation Studies
The scores for erythema and edema were totaled for
intact and abraded skin for all rabbits at 24 and 72 h. The
primary irritation index was calculated based on the sum of
the scored reactions divided by 24 (two scoring intervals
multiplied by two test parameters multiplied by six rabbits).
The developed formulation showed no erythema or edema on
the intact and abraded rabbit skin. The primary irritation
Fig. 1. DSC thermograms of a ethylcellulose, b MP, and c MP loaded

microsponges (1.5:1)
Fig. 2. SEM of MP loaded microsponges under 500
407
Table IV. The Effect of Drug/Polymer Ratio on Porosity of Microsponges Formulations

Formulation
code
MP1
MP2
MP3
MP4
MP5
MP6
MP7
MP8
a
Total intrusion
volumea (mL g 1)
Total pore
areaa (m2 g 1)
Average pore
diametera (m)
Bulk
densitya (g mL 1)
True densitya
(g mL 1)
Porositya (%)
1.02
1.08
1.14
1.21
1.25
1.26
1.31
1.33
20.21
25.63
29.84
34.73
42.15
53.46
56.71
62.15
0.20
0.17
0.15
0.14
0.12
0.09
0.09
0.08
0.29
0.32
0.33
0.41
0.43
0.39
0.44
0.43
0.42
0.48
0.56
0.69
0.80
0.83
1.05
1.18
30.64
32.14
39.73
41.26
46.81
53.44
58.23
63.71
Each observation is the mean of three determinations
index of the formulation was calculated to be 0.00. Thus, the

formulations found to be nonirritant to the rabbit skin.
Stability Studies
The developed MP loaded formulations were found to
be stable upon storage for 3 months. No change was observed
in their physical appearance, pH, rheological properties, drug
content, and drug release proles.
In Vivo Studies
MP ointment (three times a day), MP emulgel (twice in a
day) and MS emulgel (once a day) administered topically
demonstrated signicant efcacy (P<0.001) compared with
untreated as well as placebo control groups (Fig. 5). There
were no signicant differences between untreated and
placebo controls (P>0.05).
MS and MP emulgel formulations reduced mean bacte
rial counts to 3.370.49 and 3.500.71 log10 CFU/wound,
respectively. The efcacy of developed MS emulgel formula
tions applied once a day was not signicantly different from
that of conventional MP ointment applied three times a day
(P>0.05).
Fig. 3. Comparative in vitro release proles of MP from different

formulations. Each value represents meanSD (n 6)
DISCUSSION
Factorial design (32) approach was used to optimize
formulation and process variables in preparation of
microsponges. The higher drug entrapment efciencies
obtained at high drug/polymer ratios can be explained
through the fact that more amount of drug was present per
unit polymer. The reason for increased production yield at
high MP/EC ratios could be due to the reduced diffusion rate
of dichloromethane from concentrated solutions into aqueous
phase. This provides more time for the droplet formation and
may improve the yield of microparticles (2). The viscosity of
the polymer solution at high MP/EC ratio was responsible for
formation of larger particles.
The data shown in Table II was subjected to multiple
regression analysis, and results were tted in a statistical
model, Y=0 +1X1 +2X2 +11X1X1 +22X2X2 +12X1X2 incor
porating interactive and polynomial terms to evaluate the
responses, where Y is the dependent variable, 0 is the
arithmetic mean response of the nine runs, and i is the
estimated coefcient for the factor Xi (17).
The negative inuence of internal phase volume and
positive inuence of stirring rate on % EE were attributed to
the lower concentration of drug in higher volume of dichloro
methane and decreased drug loss at higher stirring rate,
respectively.
The negative inuence of both the independent variables
as shown by Eq. 4 may be attributed to the decreased
concentration of MP at high level of internal phase volume. It
was also observed that at the higher stirring rates employed,
due to the turbulence created within the external phase,
polymer adhered to the paddle and PY decreased (10).
The negative inuence of internal phase volume on
mean particle size indicated that increasing the internal phase
volume decreased the particle size of microsponges. Particle
sizes of microsponges were directly proportional to apparent
viscosity of internal phase. When the internal phase with
lower viscosity was poured into continuous phase, the
globules of the formed emulsion could easily divide into
smaller droplets and mean particle size decreased. The
negative correlation of stirring rate with mean particle size
may be ascribed to the high shearing effect of stirring blades
(18).
The data shown in Table III clearly indicate that the
output responses were strongly dependent on the selected
independent variables. The high values of correlation coef
408

Table V. Permeation Characteristics of MP from Different Formulations Across Cellulose Dialysis Membrane
Parameters
Fluxa (J, g cm 2 h 1)
Permeability coefcienta (P, 10
Zero order (R2)
First order (R2)
Higuchi (R2)
cm h 1)
MP ointment
MP emulgel
MS emulgel
151.424.01
7.570.20
0.6735
0.9113
0.8957
84.791.7
4.240.08
0.9021
0.9557
0.9790
49.891.25
2.490.06
0.9175
0.9640
0.9738
MeanSD (n 3)
a
Signicant difference between all formulations (P<0.05)
cient indicated a good t (i.e., good agreement between the

dependent and independent variables).
Differential scanning calorimetry (DSC) provides infor
mation on the physical properties of the sample and its
crystalline or amorphous nature and demonstrates a possible
interaction between drug and other compounds in micro
sponges. A broad endothermic peak for MP microsponges, as
shown in Fig. 1, indicated a signicant reduction in MP
crystallinity during the microsponge formation without pro
ducing any interaction between drug and polymer.
Both SEM and porosity data show the presence of pores
on the surface of microsponges. SEM not only shows the
authentic presence of pores but also gives additional infor
mation regarding 3 D structure of the microcarriers. Un
doubtedly, porosity can be established as in process control
for routine estimation of presence of pores for well charac
terized microsponge preparations.
To know the mechanism of drug release from these
formulations, the data were treated according to zero order
(cumulative amount of drug released vs. time), rst order
(log cumulative percentage of drug remaining vs. time),
Higuchis (cumulative percentage of drug released vs. square
root of time) model equations. The in vitro release proles of
MP from emulgels could be best expressed by Higuchis, as the
plots of percent cumulative drug release vs. square root of time
were found to be linear with R2 values ranging from 0.9738 to
0.9790.
The cumulative amount of drug permeated per unit area
was plotted against time, and the slope of the linear portion of
the plot was estimated as steady state ux (JSS). The
permeability coefcient (Kp) was calculated as Kp =Jss/C0,
Fig. 4. Amount of MP deposited in rat skin from different

formulations. Each value represents meanSD (n 3)
where C0 is the initial concentration of drug in donor

compartment.
The slower ux of MP from MS emulgel (49.891.25 g
cm 2 h 1) indicated the controlled release of entrapped drug
from polymeric carrier systems (10).
Effective topical drug therapy requires sufcient amount
of drug uptake into skin over a particular period of time for
maximal pharmacological activity. The amount of MP depos
ited in skin from the MS emulgel was higher as compared to
that with the MP emulgel or MP ointment at the end of 24 h.
This indicates microsponges improved the drug residence in
skin. This result is in agreement with previous studies,
reporting that use of particulate drug carriers like micro
particles and nanoparticles improved the drug residence in
the skin without increasing transdermal transport (19). The
high concentration of drug in the skin after application of
microparticles could be explained by the occlusive effect,
since microparticles produced a lm on the skin surface,
which reduced the transepidermal water loss and favored the
drug penetration into the skin. De Jalon et al. (1) reported
that acyclovir loaded PLGA microparticles deposited higher
amount of drug in basal epidermis. The localization and
quantication of particulate drug carriers are therefore of
Fig. 5. a Mouse surgical wound infected with S. aureus. b Compar

ative efcacy of MP loaded formulations against an experimental
surgical wound infection in mice caused by S. aureus. Each value
represents meanSD (n 6)

substantial interest in furthering our understanding of their
mechanism of action.
Mouse surgical wound model infected with S. aureus has
been used to compare the efcacy of developed formulations
with that of marketed MP ointment. The outcome measure of
the efcacy was expressed in terms of mean number of
bacteria remaining in the treated wounds (log10 CFU/wound).
Although the developed MP loaded formulations were as
effective as marketed MP ointment formulations, they exhibited
benets like reduction in frequency of application, improved
patient compliance, and good aesthetic appeal. The in vivo
efcacy results were also supported by ex vivo drug deposition
studies, which demonstrated that higher amount of MP was
deposited in skin from microsponge based emulgel formulations,
and therapeutic drug concentration were maintained up to 24 h.
CONCLUSION
Microsponge based novel delivery system has been
developed to provide once a day sustained release medication
for topical delivery of mupirocin. The formulations showed
enhanced retention of drug in skin, indicating better potential
of delivery system as compared with marketed mupirocin
ointment and conventional mupirocin emulgel. On the
grounds of efcacy and improved patient compliance due to
reduced frequency of application, microsponge based emul
gel formulations will have signicantly better role in treat
ment of primary and secondary skin infections.
ACKNOWLEDGMENT
The authors are grateful to the Indian Institute of
Technology (IIT), Bombay, India for help in performing
characterization studies.
REFERENCES
1. De Jalon EG, Blanco Prieto MJ, Ygartua P, Santoyo S. Topical
application of acyclovir loaded microparticles: quantication of
the drug in porcine skin layers. J Control Release. 2001a;75:191 7.
2. Orlu M, Cevher E, Araman A. Design and evaluation of colon
specic drug delivery system containing ubiprofen micro
sponges. Int J Pharm. 2006;318:103 17.
409
3. Nacht S, Katz M. The microsponge: a novel topical programma

ble delivery system. In: Osborne DW, Amann AH, editors.
Topical drug delivery formulations. New York: Marcel Dekker;
1990. p. 299 325.
4. Comoglu T, Gonul N, Baykara T. Preparation and in vitro
evaluation of modied release ketoprofen microsponges. Il
Farmaco. 2003;58:101 6.
5. Pappa KA. The clinical development of mupirocin. J Am Acad
Dermatol. 1990;22:873 9.
6. Ward A, Campoli Richards DM. Mupirocin: review of its
antibacterial activity, pharmacokinetic properties and therapeu
tic use. Drugs 1986;32:425 44.
7. Conly JM, Johnston BL. Mupirocin are we in danger of losing
it? Can J Infect Dis. 2002;13:157 9.
8. Parenti MA, Hateld SM, Leyden JJ. Mupirocin: a topical
antibiotic with a unique structure and mechanism of action. Clin
Pharm. 1987;6:761 9.
9. Gisby J, Bryant J. Efcacy of a new cream formulation of
mupirocin: comparison with oral and topical agents in experi
mental skin infections. Antimicrob Agents Chemother.
2000;44:255 60.
10. Jelvehgari M, Siahi Shabad MR, Azarmi S, Martin GP, Nokhod
chi A. The microsponge delivery system of benzoyl peroxide:
preparation, characterization and release studies. Int J Pharm.
2006;308:124 32.
11. Kincl M, Turk S, Vrecer F. Application of experimental design
methodology in development and optimization of drug release
method. Int J Pharm. 2005;291:39 49.
12. Mohamed MI. Optimization of chlorphenesin emulgel formula
tion. AAPS J. 2004;6:1 7.
13. Padamwar MN, Pokharkar VB. Development of vitamin loaded
topical liposomal formulation using factorial design approach:
drug deposition and stability. Int J Pharm. 2006;320:37 44.
14. Echevarria L, Blanco Prieto MJ, Campanero MA, Santoyo S,
Ygartua P. Development and validation of a liquid chromato
graphic method for in vitro mupirocin quantication in both skin
layers and percutaneous penetration studies. J Chromatogr B.
2003;796:233 41.
15. Thomas J, Schloemer B. Primary skin irritation test in the rabbit
of water jel burn dressing. Accessed 11/03/06 from http://waterjel.
com/public/SkinIrritationTest.pdf.
16. Mcripley RJ, Whitney RR. Characterization and quantitation
of experimental surgical wound infections used to evaluate
topical antibacterial agents. Antimicrob Agents Chemother.
1976;10:38 44.
17. Shivkumar HN, Desai BG, Deshmukh G. Design and optimiza
tion of diclofenac sodium controlled release solid dispersion by
response surface methodology. Ind J Pharm Sci. 2008;70:22 30.
18. Re MI, Biscans B. Preparation of microspheres of ketoprofen
with acrylic polymers by a quasi emulsion solvent diffusion
method. Powder Technol. 1999;101:120 33.
19. Alvarez Roman R, Naik A, Kalia YN, Guy RH, Fessi H.
Enhancement of topical delivery from biodegradable nano
particles. Pharm Res. 2004;21:1818 24.

DOI: 10.1208/s12249 009 9222 5
Research Article
Intratumoral Delivery of Paclitaxel in Solid Tumor from Biodegradable
Hyaluronan Nanoparticle Formulations
Abeer M. Al-Ghananeem,1,4 Ahmad H. Malkawi,1 Yahya M. Muammer,1 Justin M. Balko,1 Esther P. Black,1
Walid Mourad,2 and Edward Romond3
Received 29 October 2008; accepted 9 March 2009; published online 21 April 2009
Abstract. In the current study, novel paclitaxel loaded cross linked hyaluronan nanoparticles were
engineered for the local delivery of paclitaxel as a prototype drug for cancer therapy. The nanoparticles
were prepared using a desolvation method with polymer cross linking. In vitro cytotoxicity studies
demonstrated that less than 75% of the MDA MB 231 and ZR 75 1 breast cancer cells were viable after
2 day exposure to paclitaxel loaded hyaluronan nanoparticles or free paclitaxel, regardless of the dose.
These results suggest that hyaluronan nanoparticles maintain the pharmacological activity of paclitaxel
and efficiently deliver it to the cells. Furthermore, in vivo administration of the drug loaded nanoparticles
via direct intratumoral injection to 7,12 dimethylbenz[a]anthracene (DMBA) induced mammary tumor
in female rats was studied. The paclitaxel loaded nanoparticles treated group showed effective inhibition
of tumor growth in all treated rats. Interestingly, there was one case of complete remission of tumor
nodule and two cases of persistent reduction of tumor size that was observed on subsequent days. In the
case of free paclitaxel treated group, the mean tumor volume increased almost linearly (R2 0.93) with
time to a size that was 4.9 fold larger than the baseline volume at 57 days post drug administration.
Intratumoral administration of paclitaxel loaded hyaluronan nanoparticles could be a promising
treatment modality for solid mammary tumors.
KEY WORDS: hyaluronic acid; intratumor; mammary tumor; nanoparticles; paclitaxel.
INTRODUCTION
Many common solid tumors including breast, brain, and
prostate tumors do not respond well to conventional systemic
chemotherapy. An alternative approach to systemic adminis
tration of drugs for the treatment of localized tumor could be
through loco regional administration. This can be achieved by
intra arterial infusion of the chemotherapeutic agents up
stream of the tumor and also through direct intratumor
injection (1 3).
Microencapsulation of drugs using biodegradable poly
mers, delivered locally, provides a novel modality to increase
therapeutic concentrations of a drug for a prolonged period
while decreasing the drug systemic levels and the side effects.
For localized delivery, paclitaxel has been formulated in
biodegradable polymeric microspheres, nanoparticles, surgi
cal pastes, and implants (4 6).
1
Department of Pharmaceutical Sciences, College of Pharmacy,

University of Kentucky, 927 Rose Street, Lexington, Kentucky
40536 0082, USA.
2
Department of Pathology and Laboratory Medicine, College of
Medicine, University of Kentucky, Lexington, Kentucky 40536,
USA.
3
Division of Hematology and Oncology, Markey Cancer Center,
University of Kentucky, Lexington, Kentucky 40536, USA.
4
To whom correspondence should be addressed. (e mail: amalg0@email.
uky.edu)
The attractive polysaccharide hyaluronic acid (HA;

Fig. 1a) has been successfully used for several biomedical
applications with the cytotoxic agent paclitaxel such as;
macromolecular drug copolymer conjugate, hydrophilic coat
ing, and hydrogels (7 10), but not as nanoparticle polymer
system. HA is one of the main components of the extracel
lular matrix in the body. Since HA is a naturally occurring
biopolymer, it is biocompatible and biodegradable. Further
more, most malignant solid tumors and their surrounding
stromal tissue contain elevated levels of HA (9). CD44 is the
principal cell surface receptor for HA, and it presents at low
levels on epithelial, hemopoietic and some neuronal cells (11
13), and at elevated levels in various carcinoma, lymphoma,
breast, colorectal, and lung tumor cells (14 17). Overexpres
sion of the HA receptors CD44 and RHAMM on cancer cells
results in enhanced binding and endocytosis uptake of such
compound.
Paclitaxel (PXL), the first taxane in clinical trials, is an
active chemotherapeutic drug against a broad range of
cancers, especially breast and ovarian cancers. PXL has
shown a high potential as an anticancer drug, but its practical
use is limited due to its low solubility in water as well as other
pharmaceutical solvents compatible with intravenous admin
istration. Paclitaxel is currently available as a solution in a
vehicle composed of Cremophor EL and dehydrated
alcohol (Taxol), or as an injectable suspension of albumin
bound paclitaxel nanoparticles (Abraxane). However, these
two commercial formulations had reported side effects
410
Intratumoral Delivery of Paclitaxel in Solid Tumor
411
Preparation of Paclitaxel-Loaded Cross-Linked Hyaluronan
Nanoparticles
Fig. 1. a Structure of hyaluronic acid (HA) polysaccharide which is

glycosaminoglycan copolymer of D glucuronic acid and N acetyl D
glucosamine. b Chemical cross linking of hyaluronic acid using
glutaraldehyde
(18,19). Thus, safe and effective drug delivery systems are still
in need to improve the current clinical chemotherapeutic
treatments with PXL.
To our knowledge, there are no examples of HA
nanoparticles as a delivery system for chemotherapeutic
agents. Therefore, taking into account the favorable biolog
ical and physicochemical properties of HA, it is reasonable to
hypothesize that HA nanoparticles could improve paclitaxel
cytotoxicity to tumor cells. In the current study, novel
paclitaxel loaded cross linked hyaluronan nanoparticles
(PXL HA NPs) were engineered for the local delivery of
paclitaxel (PXL) as a prototype drug for cancer therapy.
Furthermore, the drug loaded nanoparticles efficacy was
tested through direct intratumoral administration to 7,12
dimethylbenz[a]anthracene (DMBA) induced mammary tu
mor in female rats.
Materials
Paclitaxel, sodium sulfate, sodium metabisulfite, Tween
20, glutaraldehyde (25% in water), 7,12 dimethylbenz[a]
anthracene (DMBA), and hyaluronidase enzyme were pur
chased from Sigma Aldrich Chemical (St. Louis, MO, USA).
The hyaluronic acid sodium salt (from streptococcus equi sp.)
with Mw 1.5MDa was purchased from Fluka Chemie GmbH
(St. Louis, MO, USA). MDA MB 231 and ZR 75 1 breast
cancer cell lines were obtained from American Type Culture
Collection ATCC (Manassas, VA, USA). Ki 67 polyclonal
antibody was obtained from AbCam (Cambridge, MA).
Tetrazolium dye (MTS) assay, phosphate buffered saline
(PBS), MTS CellTiter 96 AqueousONE solution assay,
and all other cell culture supplies were purchased from
Promega (Madison, WI, USA). All other chemicals and
reagents were of analytical grade.
HA was subjected to acidic degradation to produce a

low weight average molecular weight polymer (20). Briefly,
1% w/v aqueous solution of HA was degraded in HCl
solution (pH 0.5) at 37C for 24 h. After this time, the pH
was adjusted to 7.0, and the solution was subjected to
extensive dialysis using Spectra/Por tubing with a molecular
cutoff of 3,500 Da. Nanoparticles were prepared using a
desolvation method with polymer cross linking as follows:
HA (200 mg) was dissolved in 10 mL of water containing 2%
Tween 20. The solution was heated to 40C with constant
stirring at 500 rpm to ensure full mixing. To this solution,
2 mL of a 20% aqueous solution of sodium sulfate was added
slowly followed by 1 2 mL of acetone containing 2 or 20 mg
of paclitaxel. A second aliquot of sodium sulfate solution (5
6 mL) was added until the solution turned turbid, which
indicated the formation of HA aggregates. Approximately
1 mL of distilled water was then added until the solution
turned clear. An aqueous solution of glutaraldehyde (25%,
0.6 mL) was added to cross link the hyaluronic acid (Fig. 1b).
Sodium metabisulfite solution (12%, 5 mL) was added 30 min
later to stop the cross linking process. After 2 h, the crude
product was purified on a Sephadex G 50 column, and
nanoparticle containing fraction was lyophilized in a freeze
drier over a 48 h period.
High-Pressure Liquid Chromatography Analysis of Paclitaxel

Analyses of PXL concentration was conducted on an
Agilent 1100 series chromatographic system (Agilent Tech
nologies, Palo Alto, CA, USA). Chromatographic separation
was achieved with a C18 Altima column (4.6 mm ID,
150 mm; 5 m) from Grace (Grace, MA) protected by a
Phenomenex SecurityGuard in line filter frit (Torrance,
CA, USA). The mobile phase consisted of water, methanol,
and acetonitrile (25:35:40, v/v). The mobile phase was filtered
through a 0.45 m filter (Millipore, Bedford, MA, USA) and
degassed before use. The high performance liquid chroma
tography (HPLC) system was run at ambient temperature
with a flow rate of 1.0 mL/min, and quantitative analysis of
peak area under the curve (AUC) was performed at a
wavelength of 230 nm with total chromatographic run time
of 8 min. The calibration curve for the quantification of
paclitaxel was linear over the standard concentration range of
paclitaxel at 50 50,000 ng/mL with a correlation coefficient
of R2 =0.999. The limit of detection was 50 ng/mL.
Characterizations of PXL-HA NPs

The morphology and size distribution of the PXL HA
NPs were observed using Philips Tecnai 12 Biotwin micro
scope with an accelerating voltage of 100 kV at the University
of Kentucky Medical Center Imaging Facility. A dried nano
particle sample was dispersed directly into distilled water, and
then a copper grid coated with a carbon film was put into the
previous suspension several times and left to incubate for
2.0 min at room temperature. After drying and removal of
excess fluid, the samples were negatively stained with 2%
Al-Ghananeem et al.
412
uranyl acetate, and the grids were examined and recorded
with the transmission electron microscope (TEM) images.
The amount of entrapped paclitaxel in nanoparticles was
detected in triplicate by HPLC analysis. Two milligrams of
paclitaxel loaded nanoparticles were dispersed in 0.5 mL of
PBS and digested with 0.5 mL of hyaluronidase (1 mg/mL in
PBS) in a metabolic shaker at 37C. After about 1 h,
extraction was performed with two volumes of 3 mL of ethyl
acetate each. The ethyl acetate layers were pooled, dried
under a stream of nitrogen, and reconstituted in 100 L
acetonitrile. The encapsulation efficiency (EE) of paclitaxel
in nanoparticles was determined as the mass ratio of the
entrapped paclitaxel in nanoparticles to the theoretical
amount of paclitaxel used in the preparation.
In Vitro Release of Paclitaxel from HA NPs
A membraneless dissolution method was used for in vitro
studies (21). PXL HA NPs (10 20 mg) were incubated in
60 mL of PBS (pH 7.4) at 37C. To maintain a sink condition,
2 mL of octanol were placed on top of the PBS layer to
continuously extract the released paclitaxel from the PBS. At
assigned time points, octanol was removed, and a fresh 2 mL
of octanol was layered on top of the PBS. Assay for PXL
concentration was performed after appropriate dilution in
acetonitrile utilizing the HPLC system mentioned earlier.
In Vitro Cytotoxicity of PXL-HA NPs
RPMI 1640 (with 2 mM L glutamine, 10 mM HEPES,
1 mM sodium pyruvate, 4.5% glucose and 1.5% sodium
bicarbonate) and Leibovitzs L15 media were used as culture
media and were supplemented with 10% fetal bovine serum
(FBS), 50 g/mL penicillin, and 50 g/mL streptomycin. In a
96 well plate (Nunclon), the human breast cancer cell lines
MDA MB 231 and ZR75 were maintained in Leibovitzs L15
and RPMI 1640, respectively. All cells were maintained in an
isolated humidified environment with 5% CO2 at 37C.
Cytotoxicity of PXL in PBS, PXL HA NPs, and unloaded
HA nanoparticles (blank HA NPs) were determined by
measuring the cell growth inhibition using a tetrazolium dye
(MTS) assay. MDA MB 231 and ZR75 cells were seeded in
96 well plates at a density of 5103 cells/well and allowed to
attach for 16 h. The following morning, the cells were treated
with various concentrations (1 or10 g/mL) of PXL and HA
PXL NPs in the appropriate complete growth medium. At
initial treatment, 24, 48, and 72 h posttreatment, cell
proliferation was analyzed using the MTS CellTiter 96
AqueousONE solution assay. Absorbance at 490 nm was
measured on an ELx800 Universal microplate BioTek
Instruments reader (Winooski, VT, USA). The results are
expressed as a percentage of the control treated cells.
In Vivo Evaluation of Antitumor Efficacy of PXL-HA NPs
The in vivo efficacy of the PXL HA nanoparticles was
assessed in female Sprague Dawley rats. Throughout the
experiment, all of the rats were housed under a controlled
environment with 12 h light/dark cycles and a temperature of
222C, given a commercial diet with tap water ad libitum, and
weighed every 7 days. At 6 weeks of age, all rats (n=12) were
administered 10 mg of DMBA in 1 mL of sesame oil by

intragastric gavage weekly for 2 weeks. Beginning 6 weeks
after DMBA administration, animals were palpated once
weekly to monitor mammary tumor growth. The survival and
the body weight of the rat were recorded. The tumor width (w)
and length (l) was recorded with a caliper and tumor size was
calculated using the formula (lw2/2). All research and testing
activities, related to this work, was reviewed and approved by
the Institutional Animal Care and Use Committee (IACUC) at
the University of Kentucky Chandler Medical Center,
Division of Laboratory Animal Resources (DLAR).
When tumors had grown to approximately 100 mm3 in
the DMBA treated rats, the animals were randomized into
four groups A, B, C, and D (n=3/group). Animals were
treated with a single 100 L intratumor injection of 20 mg
paclitaxel/kg body weight PXL HA NPs, equivalent PXL
dose in PBS for groups A and B, respectively. The control
groups C and D received a single 100 L intratumor injection
of unloaded HA nanoparticles to assess the polymer safety,
and PBS (0.1 M, pH 7.4), respectively. The tumor size was
calculated as previously reported. The rats were euthanized
when the tumor size reached greater than 15% of body
weight. The study was terminated 57 days posttreatment.
The tumors were collected immediately after euthaniza
tion or reported mortality. Frozen tumors were fixed in 10%
buffered formalin before being embedded in paraffin blocks.
Two 4 m sections were obtained from each tumor. One
section was processed for routine hematoxylin and eosin
staining for histological assessment. The other section was
immunostained using the Ki 67 antibody for assessment of
the proliferation index of each tumor (22).
Hematoxylin and eosin stained sections were microscop
ically assessed for any evidence of necrosis and degeneration
as a reflection of the effect of the drug. Degenerative changes
in the form of apoptosis and ballooning degeneration were
sought. Immunostained sections were evaluated for the
percentage of cells positive for the ki 67 polyclonal antibody.
Positive stained cells showed brown nuclear staining. The
percentage of cells was expressed after counting 500 cells in
five high power fields.
The statistical significance was studied using two way
analysis of variance (ANOVA) and two tailed Students t tests.
Differences were considered to be significant at a level of P<0.05.
RESULTS
Preparation and Characterization of PXL-HA NPs
The resulted nanoparticles were spherical in shape with
moderate uniformity. Figure 2 shows a transmission electron
microscope (TEM) image of HA PXL NPs. The encapsula
tion efficiency of PXL into HA nanoparticles was determined
as the mass ratio of the entrapped paclitaxel in nanoparticles
to the theoretical amount of paclitaxel used in the prepara
tion and was found to be above 90% at the PXL loading
utilized in our experiment (Table I). Both of the 1% and 10%
loading levels of PXL in the HA nanoparticle formulations
413
Fig. 3. In Vitro accumulated release of PXL from PXL HA nano

particles at 1% (filled diamonds) and 10% (filled squares) PXL
loading in PBS at 37.00.1C. Values are means SD (n 3)
Fig. 2. Transmission electron microscope (TEM) image of paclitaxel

loaded hyaluronan nanoparticles
had shown no statistical significant difference in encapsula

tion efficiency.
PXL HA NPs inhibited growth of the cell lines to relatively

similar extents. Nearly less than 75% of the MDA MB 231
and ZR 75 1 breast cancer cells were viable after a 2 day
exposure to PXL and PXL HA treatment regardless of the
dose. Furthermore, two way ANOVA using a repeated
measures design revealed a statistically significant reduction
in cell viability at 72 h in the MDA MB 231 cell line for all
In Vitro Release of Paclitaxel From HA NPs

To evaluate the potential of HA nanoparticles as a drug
carrier of PXL, the release behavior of PXL from HA
nanoparticles was assessed at 37C1 in PBS (pH 7.4).
During these experiments, the sink conditions were main
tained by regularly replacing the octanol medium. The in vitro
release profile of PXL HA (1% drug loading) nanoparticles
is shown in Fig. 3. Paclitaxel was released from the drug
loaded HA nanoparticles in a biphasic pattern: a fast release
rate in the first 10 h followed by a slow uniform release
afterwards. The initial fast drug release can be ascribed to
those drugs located on or near the particle surface; while the
slow and uniform release could be caused by release of the
drugs inside the nanoparticles. In the first 10 h, about 70
0.4% of the loaded drug was released out of the nano
particles. The accumulative drug release in 18 h was 90
1.2%.
In Vitro Cytotoxicity of PXL-HA NPs
As shown in Fig. 4, empty cross linked HA NPs did not
show any toxicity. On the other hand, both the PXL and
Table I. Characterizations of PXL Loaded HA Nanoparticles
Sample
Theoretically
loaded
PXL (wt%)
Measured
loaded
PXL (wt.%)
Encapsulation
efficiency (%)
Size (nm)
HA
PXL HA
PXL HA
1
10
0.830.07
7.80.90
837
789
26916.2
29441.2
304.828.4
Fig. 4. In vitro cytotoxicity study results using MDA MB 231 (top)

and ZR 75 1 (bottom) human breast cancer cell lines
Al-Ghananeem et al.
414
treatment groups (PXL 1, PXL HA 1, PXL 10, and PXL
HA 10) relative to the appropriate control treatment. This
effect was not observed in the ZR 75 cell line.
These results suggest that HA nanoparticles maintain
the pharmacological activity of PXL. Interestingly, however,
no advantage on cell growth inhibition was noted in the
PXL HA NPs treated cells over the free PXL formulation,
regardless of CD44 status. This is likely due to the fact that
the in vitro cell culture model is likely to be incapable of
demonstrating an advantage of cell growth inhibition under
these conditions, since drug exposure and uptake is not a
limiting factor in this model. In vivo tumor models would
represent a more ideal case to demonstrate such an
advantage.
In Vivo Antitumor Effect
There was no significant difference in the initial mean
size of the tumor nodules between the PXL HA NPs and the
PXL suspended in PBS groups before treatment. However,
the PXL HA NPs treated group showed effective inhibition
of tumor growth in all treated rats. Interestingly, there was
one case of complete remission of tumor nodule and two
cases of persistent reduction of tumor size that was observed
on subsequent days. The mean size of the tumor nodules
came close to baseline volume over the 57 day observation
period. In the case of the PXL suspended in PBS group, the
mean tumor volume increased almost linearly (R2 =0.93) with
time to a size that was 4.9 fold larger than the baseline
volume at 57 days post drug administration. The single
intratumor injection of PXL HA NPs therefore produced
significant tumor growth inhibition effect compared to PXL
suspended in PBS at the same paclitaxel dose of 20 mg/kg
(Fig. 5).
For both the PXL HA NPs and the PXL suspended
in PBS treated groups, there was no significant weight
change in the animals after drug administration, suggest
ing an absence of significant toxicity associated with the
two formulations at the paclitaxel dose of 20 mg/kg
(Fig. 6).
Control groups C and D that were treated with a single
100 L intratumor injection of unloaded HA nanoparticles to
assess the polymer safety and PBS exhibited similar tumor
Fig. 6. The change in DMBA treated animals body weight as a

function of time after the rats received intratumor injection of PXL
HA nanoparticles (filled diamonds), PXL in normal saline (filled
squares), HA nanoparticles (filled circles) and PBS (filled triangles).
The intratumor dose of paclitaxel administered was 20 mg/kg. The
data are expressed as mean SD (n 3)
growth profiles that were characterized by a steady growth of

the tumor nodule with time (Fig. 5). The changes in tumor
size at day 10 were comparable for both control groups. In
addition, the rate of tumor growth in these vehicle treated
groups was comparable to that seen in rats treated with PXL
suspended in PBS, suggesting that PXL suspended in PBS at
a paclitaxel dose of 20 mg/kg did not effectively inhibit the
growth of DMBA induced tumor in rats. Within 2 3 weeks,
the control groups which were treated with HA or PBS had
an increase in tumor size up to ten times the initial tumor size
before treatment. Two rats from the PBS treated rats were
euthanized after 2 weeks because their tumor size reached
greater than 15% of their body weight. The rest of rats in
these groups could not survive after 3 weeks, and their tumors
were collected after reporting their death.
Histopathology analysis of collected tumors illustrated
that tumors treated with intratumor injection of free PXL
showed no significant necrosis or degenerative changes. Ki 67
labeling indices in these tumors were 5% to 10% (Fig. 7a b).
On the other hand, tumors treated with intratumor injection
of PXL HA NPs showed areas of degeneration in the form of
apoptosis and vacuolization with areas of necrosis. The ki 67
labeling indices in both tumors was 0% indicating no growth
(Fig. 8a b)
DISCUSSION
Fig. 5. The change in tumor size in DMBA treated female SD rats as

a function of time after a single intratumor injection with PXL HA
nanoparticles (filled diamonds) or PXL in normal saline (filled
squares) at 20 mg paclitaxel/kg (mean SD, n 3; p 0.05)
Our interest in HA was prompted by its potential as a drug

carrier in cancer chemotherapy. Several reviews have previously
discussed the advantages of HA as a drug carrier and a targeting
ligand for cancer, as well as other pathologies (23 28). Direct
intratumoral delivery systems have significant potential advan
tages due to providing higher drug concentration at the target
site compared with systemic routes and circumvention of the
various physiological barriers to tumor drug delivery including
the high interstitial pressure common to most solid tumors (29).
The feasibility of directly injecting tumors has been greatly
advanced recently by the utilization of nanoparticle pharma
ceutical technology, which permit the enhanced delivery of
drugs to the cytoplasm of the cancer cells (30).
415
period. This could be attributed mainly to two properties of
HA nanoparticles that are favorable for drug release. First,
compared with microparticles, the small, submicron size of
the nanoparticles decrease the effective diameter for the
release medium to reach the drug and for outward drug
diffusion. Second, the hydrophilicity of HA polymer enhances
fluid uptake, which would result in enhanced drug release.
As shown in Fig. 3, paclitaxel was released from the
drug loaded HA NPs in a biphasic pattern: a fast release rate
in the first 24 h followed by a slow uniform release
afterwards. The initial fast drug release can be ascribed to
those drugs located on or near the particles surface; while the
slow and uniform release could be caused by diffusion of the
drugs inside the nanoparticles.
In order to determine if the paclitaxel loaded cross
linked hyaluronan nanoparticles retained pharmacologic
activity in vitro, an MTS assay was performed to compare
cell proliferation of two breast cancer cell lines, MDA 231
and ZR 75, in the presence of free paclitaxel or PXL HA
nanoparticles. To ensure that the cytotoxicity was caused by
PXL itself and not by the polymer, cells were also incubated
Fig. 7. a HE (hematoxylin and cosin) stains of tumor harvested from a

rat treated with a single intratumor injection of PXL in PBS (20 mg/kg)
at day 57 posttreatment. The tumor tissue shows a viable adenocarci
noma. b Tumor harvested from a rat treated with a single intratumor
injection of PXL in PBS (20 mg/kg) at day 57 posttreatment showing a
Ki 67 labeling index of 10% (immunoperoxidase 200)
Cross linked hyaluronan nanoparticles are nano sized

and can be attractive delivery candidates for a variety of
biomedical applications. Furthermore, the biocompatibility of
the HA nanoparticles, the high affinity toward cancer cells,
and the suitable physicochemical properties of these HA
nanoparticles make it as a nano drug delivery system being
capable of selectively inhibiting cancer cells of solid tumor
and metastatic tumor. In the current study, the developed
PXL HA NPs particle size and the EE were not significantly
influenced by the PXL loading levels utilized in this study.
This suggests that PXL molecules are entrapped into the
hydrophobic HA inner cores and the entrapped PXL
molecules at the loading levels utilized here did not alter
significantly the average diameter of the HA nanoparticles.
For a drug to be released from nanoparticles, the release
medium must diffuse into the particle and dissolve the
entrapped drug, which then diffuses out of the nanoparticle
(31). The HA nanoparticles released paclitaxel in vitro almost
completely over a 50 h period. The cumulative release of the
drug did not show any significant difference in PXL release
rate between the 1% and 10% PXL loaded nanoparticles in
PBS at 37.00.1C with almost 90% released in 2 days
Fig. 8. a HE (hematoxylin and cosin) stains of tumor harvested from

a rat treated with a single intratumor injection of PXL HA nano
particles (20 mg/kg) at day 57 posttreatment. The tumor tissue shows
degenerative and necrotic changes in the form of cytoplasmic
vacuolization, fatty change, and apoptosis. b Tumor harvested from
a rat treated with a single intratumor of PXL HA nanoparticles
(20 mg/kg) at day 57 treatment showing a Ki 67 labeling index of 0%
(immunoperoxidase 200)
Al-Ghananeem et al.
416
with empty cross linked HA nanoparticles. HA exhibits a
number of properties of a successful drug carrier. This water
soluble, nonimmunogenic polysaccharide has multiple func
tional groups available for chemical conjugation (24,25).
Furthermore, since it is the major CD44 ligand, HA can be
used to target cells on which CD44 is expressed (26). In the
current study, we utilized the HA polymer locally to enhance
the drug delivery to the cancer cells rather than to target the
cells after systemic delivery. The MDA MB 231 and ZR 75 1
cells are both human breast epithelial cancer cell lines.
However, the MDA MB 231 cells express positive CD44
receptor on their surface, while ZR 75 1 cells are deficient in
the receptor (32,33). The in vitro cytotoxicity studies demon
strated that the placebo HA NPs (without paclitaxel encap
sulated) was nontoxic by MTS assay at the studied
nanoparticle concentration levels, which implies that the
polymer is biocompatible. Furthermore, the drug loaded
HA nanoparticles maintained the pharmacological activity
of PXL. Interestingly, however, no advantage on cell growth
inhibition was noted in the PXL HA nanoparticle treated
cells over the free PXL formulation, regardless of CD44
status. This is likely due to the fact that, generally, PXL enters
the tumor cell by diffusion and most nanoparticles by
endocytosis. However, due to the larger NP size in the
current study (>300 nm) and the drug release kinetics at
pH 7.4, PXL is released most likely from the HA PXL NP at
the extracellular matrix and then enters the cell afterwards by
diffusion. On the other hand, in the in vivo scenario, HA NPs
are most likely acting as a shield that keeps the drug for a
more prolonged time intratumorally and, therefore, enhanc
ing the therapeutic effect. Furthermore, the in vitro cell
culture model is likely to be incapable of demonstrating an
advantage of cell growth inhibition under these conditions,
since drug exposure and uptake is not a limiting factor in this
model. In vivo tumor models would represent a more ideal
case to demonstrate such an advantage.
Solid tumors have several potential barriers to drug
delivery that may limit drug penetration, such as alterations in
the distribution of blood vessels, blood flow, and interstitial
pressure (34,35). Therefore, high systemic levels of the
cytotoxic drug often cause a related systemic toxicity without
reaching the tumor at effective concentrations. The focus for
this current study was to deliver the cytotoxic agent locally.
The PXL HA NPs formulation is injectable through a 23
gauge needle to the tumor tissue, and it was examined in
female Sprague Dawley rats bearing DMBA induced tumor.
After treatment, it was found that the average tumor volume
as a function of time was significantly less for the tumor
bearing rats injected with PXL HA NPs compared to the free
PXL intratumor injection. There was no significant difference
in the initial mean size of the tumor nodules between the
PXL HA NPs and the PXL suspended in PBS groups before
treatment. However, the PXL HA NPs treated group showed
effective inhibition of tumor growth in all treated rats. There
was even one case of complete remission of tumor nodule and
two cases of persistent reduction of tumor size that was
observed on subsequent days. The mean size of the tumor
nodules remained at close to baseline volume or even below
through shrinking over the next 57 days. In the case of the
PXL suspended in PBS group, the mean tumor volume
increased almost linearly with time to a size that was 4.9 fold
larger than the baseline volume at 57 days post drug

administration. This was in agreement with the histopathol
ogy analysis of the tumor tissue collected from both treatment
groups. Tumors collected from rats treated with intratumor
injection of free PXL showed a proliferation index reflecting
moderate proliferation, while the tumors collected from rats
treated with intratumor injection of PXL HA NPs had a
Ki 67 proliferative index indicating no growth in such cells
suggesting that they were exposed to an effective cytotoxic
treatment. It is known that the Ki 67 is an antigen expressed
in cells that are not in the G0 phase, but in G1, G2, S, and M
phases and that tumors with a high ki 67 proliferative index
are rapidly growing and aggressive (22). Therefore, the single
intratumor injection of PXL HA NPs produced significant
tumor growth inhibition effect compared to free PXL sus
pended in PBS at the same paclitaxel dose of 20 mg/kg.
Although we did not see any mortality following single dose
intratumor administration of free PXL or PXL HA NPs at
20 mg/kg, the tumor volume results clearly show that the PXL
HA NPs formulation was significantly efficacious in decreasing
the tumor size and propagation relative to free PXL or control
treatment. Furthermore, weight change of an animal after drug
administration is often used to evaluate the toxicity of the drug
treatment (36). For both the PXL HA NPs and the PXL
suspended in PBS treated groups, there was no significant
weight change in the animals after drug administration,
suggesting an absence of significant toxicity associated with
the two formulations at the paclitaxel dose of 20 mg/kg.
CONCLUSION
We have developed cross linked HA nanoparticles loaded
with paclitaxel. The resulting nanoparticles had cytotoxicity
similar to that of free paclitaxel in vitro with superior antitumor
efficacy in vivo when administered with intratumor injection
for the treatment of DMBA induced mammary tumor in rats.
Histological studies proved the existence of the necrotic
process caused by the cytotoxic drug. Thus, the direct
administration of PXL HA NPs at the tumor sites could be a
promising treatment modality for solid tumors, and further
testing is warranted to evaluate the formulation clinically.
ACKNOWLEDGMENT
This research was supported by grant no. 85 001 19 IRG
to AMA from the American Cancer Society.
REFERENCES
1. Deurloo MJ, Kop W, van Tellingen O, Bartelink H, Begg AC.
Intratumoural administration of cisplatin in slow release devices:
II. Pharmacokinetics and intratumoural distribution. Cancer
Chemother Pharmacol 1991;27 5:347 53.
2. Aigner KR. Intra arterial infusion: overview and novel
approaches. Semin Surg Oncol 1998;14 3:248 53.
3. Lee JD, Yang WI, Lee MG. Effective local control of malignant
melanoma by intratumoural injection of a beta emitting radio
nuclide. Eur J Nucl Med Mol Imaging 2002;29 2:221 30.
4. Zhang X, Jackson JK, Wong W. Development of biodegradable
polymeric paste formulations for taxol: an in vitro and in vivo
study. Inter J Pharm 1996;137 2:199 208.

5. Dhanikula AB, Panchagnula R. Development and characteriza
tion of biodegradable chitosan films for local delivery of
paclitaxel. AAPS J 2004;6 3:e 27.
6. Ho EA, Vassileva V, Allen C, Piquette Miller M. In vitro and in
vivo characterization of a novel biocompatible polymer lipid
implants system for the sustained delivery of paclitaxel. J Cont
Rel 2005;104 1:181 91.
7. Ramaswamy M, Zhang X, Burt HM, Wasan KM. Human plasma
distribution of free paclitaxel and paclitaxel associated with
diblock copolymers. J Pharm Sci 1997;86 4:460 4.
8. Luo Y, Prestwich GD. Synthesis and selective cytotoxicity of a
hyaluronic acid antitumor bioconjugate. Bioconjug Chem
1999;10 5:755 63.
9. Lokeshwar VB, Selzer MG. Differences in hyaluronic acid
mediated functions and signaling in arterial, microvessel, and
vein derived human endothelial cells. J Biol Chem 2000;275
36:27641 9.
10. Obara K, Ishihara M, Ozeki Y. Controlled release of paclitaxel
from photocrosslinked chitosan hydrogels and its subsequent
effect on subcutaneous tumor growth in mice. J Control Release
2005;110 1:79 89.
11. Hardingham T, Fosang A. Proteoglycans: many forms and many
functions. FASEB J 1992;6 3:861 70.
12. Hoare K, Savani RC, Wang C, Yang B, Turley EA. Identification
of hyaluronan binding proteins using a biotinylated hyaluronan
probe. Connect Tissue Res 1993;30 2:117 26.
13. Turley EA, Belch AJ, Poppema AJ, Pilarski LM. Expression and
function of a receptor for hyaluronan mediated motility on
normal and malignant lymphocytes B. Blood 1993;81 2:446 53.
14. Vazquez E, Dewitt DM, Hammond PT, Lynn DM. Construction
of hydrolytically degradable thin films via layer by layer deposi
tion of degradable polyelectrolytes. J Am Chem Soc 2002;124
47:13992 3.
15. Benkirane Jessel N, Schwinte P, Falvey P. Build up of polypep
tide multilayer coatings with anti inflammatory properties based
on the embedding of piroxicam cyclodextrin complexes. Adv
Funct Mater 2004;14 2:174 82.
16. Guyomard A, Nysten B, Muller G, Glinel K. Loading and
release of small hydrophobic molecules in multilayer films based
on amphiphilic polysaccharides. Langmuir 2006;22 5:2281 7.
17. Berg MC, Zhai L, Cohen RE, Rubner MF. Controlled drug
release from porous polyelectrolyte multilayers. Biomacromole
cules 2006;7 1:357 64.
18. Massey PJ, Chen TT, Caudle MR. Biphasic steroidogenic effects
of taxol and cremophor EL. Endocrine 1994;2 10:891 7.
19. Cella D, Peterman A, Hudgens S, Webster K, Socinski MA.
Measuring the side effects of taxane therapy in oncology: the
functional assessment of cancer therapy taxane (FACT taxane).
Cancer 2003;98 4:822 31.
20. Shu XZ, Liu Y, Palumbo FS, Prestwich GD. Disulfide
crosslinked hyaluronan gelatin hydrogel films: a covalent
417
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
mimic of the extracellular matrix for in vitro cell growth.

Biomaterials 2003;24 21:3825 34.
Barichello JM, Morishita M, Takayama K, Nagai T. Encapsula
tion of hydrophilic and lipophilic drugs in PLGA nanoparticles
by the nanoprecipitation method. Drug Dev Ind Pharm 1999;25
4:471 6.
Isik Gonul I, Akyurek N, Dursun A, Kupeli B. Relationship of
Ki67, TP53, MDM 2 and BCL 2 expressions with WHO 1973
and WHO/ISUP grades, tumor category and overall patient
survival in urothelial tumors of the bladder. Pathol Res Pract
2008;204 10:707 17.
Pouyani T, Prestwich GD. Functionalized derivatives of hyalur
onic acid oligosaccharides: drug carriers and novel biomaterials.
Bioconjug Chem 1994;5 4:339 47.
Prestwich GD, Marecak DM, Marecek JF, Vercruysse KP,
Ziebell MR. Controlled chemical modification of hyaluronic
acid: synthesis, applications, and biodegradation of hydrazide
derivatives. J Controlled Release 1998;53 1 3:93 103.
Vercruysse KP, Prestwich GD. Hyaluronate derivatives in drug
delivery. Crit Rev Ther 1998;15 5:513 55.
Jaracz S, Chen J, Kuznetsova LV, Ojima I. Recent advances in
tumor targeting anticancer drug conjugates. Bioorg Med Chem
2005;13 17:5043 54.
Liao YH, Jones SA, Forbes B, Martin GP, Brown MB.
Hyaluronan: pharmaceutical characterization and drug delivery.
Drug Delivery 2005;12 6:327 42.
Yadav AK, Mishra P, Agrawal GP. An insight on hyaluronic acid
in drug targeting and drug delivery. J Drug Targ 2008;16:91 107.
Thorpe PE. Vascular targeting agents as cancer therapeutics.
Clin Cancer Res 2004;10:415 27.
Zamboni WC. Liposomal, nanoparticle, and conjugated formu
lations of anticancer agents. Clin Cancer Res 2005;11 23:8230 4.
Alonso MJ, Cohen S, Park TG, Gupta RK, Siber GR, Langer R.
Determinants of release rate of tetanus vaccine from polyester
microspheres. Pharm Res 1993;10 7:945 53.
Culty M, OMara TE, Underhill CB, Yeager H Jr, Swartz RP.
Hyaluronan receptor (CD44) expression and function in human
peripheral blood monocytes and alveolar macrophages. J Leukoc
Biol 1994;56 5:605 11.
Herrera Gayol A, Jothy S. Effects of hyaluronan on the invasive
properties of human breast cancer cells in vitro. Int J Exp Pathol
2001;82 3:193 200.
Jain RK. Delivery of molecular medicine to solid tumors. Science
1996;271 5252:1079 80.
Curnis F, Sacchi A, Corti A. Improving chemotherapeutic drug
penetration in tumors by vascular targeting and barrier alter
ation. J Clin Inv 2002;110 4:475 82.
Harper E, Dang W, Lapidus RG, Garver RI. Enhanced efficacy of
a novel controlled release paclitaxel formulation (PACLIMER
delivery system) for local regional therapy of lung cancer tumor
nodules in mice. Clin Cancer Res 1999;5 12:4242 8.

DOI: 10.1208/s12249 009 9223 4
Research Article
Colon Specific Delivery of Indomethacin: Effect of Incorporating pH Sensitive
Polymers in Xanthan Gum Matrix Bases
Laila F. A. Asghar,1 Chetan B. Chure,1 and Sajeev Chandran1,2,3
Abstract. In the present study, an attempt has been made to design controlled release colon specic
formulations of indomethacin by employing pH responsive polymers Eudragit (L100 or S100) in matrix
bases comprised of xanthan gum. The prepared tablets were found to be of acceptable quality with low
weight variation and uniform drug content. In vitro release studies indicated rapid swelling and release of
signicant percentage of drug in the initial period from matrix tablets composed of xanthan gum alone.
Addition of pH responsive polymers Eudragit (L100 or S100) to xanthan gum matrix resulted in
negligible to very low drug release in the initial period in acidic to weakly acidic medium. Furthermore,
with increase in pH of the dissolution medium due to dissolution of Eudragit L100/Eudragit S100 that
resulted in the formation of a porous matrix, faster but controlled drug release pattern was observed.
Thus, a sigmoidal release pattern was observed from the designed formulations suitable for colonic
delivery. Drug release mechanism in all cases was found to be of super case II type, indicating erosion to
be the primary cause of drug release. Since the drug release from almost all the matrix bases in the initial
phase was negligibly low and followed with controlled release for about 14 16 h, it was concluded that a
matrix design of this composition could have potential applications as a colon specic drug delivery
device with additional advantage of easy scale up and avoidance of all or none phenomenon associated
with coated colon specic systems.
KEY WORDS: colon specic delivery; controlled release; Eudragits; matrix; pH sensitive polymers;
xanthan gum.
INTRODUCTION
Indomethacin is a non steroidal anti inammatory drug,
commonly indicated in the treatment of osteo and rheumatoid
arthritis. Recent reports have implicated its use as an anti
cancer agent against various in vitro and in vivo models of
colorectal cancer (1). It has been reported to cause growth
inhibition, induction of apoptosis, and reduction in proliferation
rates of HT 29 colon cancer cells, along with down regulation of
survivin (an apoptosis inhibitor) (2 4). Oral administration of
indomethacin has been reported to cause dose dependent
systemic and local upper gastrointestinal side effects in 35%
to 50% patients (5). A formulation of indomethacin with
negligible to no release in upper gastrointestinal (GI) tract
and controlled release in colonic region would achieve thera
peutically effective concentration of drug locally in colon. At
the same time, such a formulation would minimize systemic or
upper GI tract related side effects of indomethacin.
1
Formulation Development & Pharmacokinetics Laboratory, Phar

macy Group, Birla Institute of Technology and Science, Pilani, 333
031 Rajasthan, India.
2
Innovation Cell, Pharma Research, Lupin Ltd. (Research Park),
46A/47A, Nande Village, Mulshi Taluka, Pune, 411042 Maharashtra,
India.
3
To whom correspondence should be addressed. (e mail: sajeevchan
dran@lupinpharma.com)
Various approaches used for targeting drugs to the colon

include prodrug based techniques (including azo polymer and
hydrogel systems), time dependent delivery systems (includ
ing osmotic pump, swelling, and coating controlled), pH
sensitive polymer based coating systems, and bacterial
enzyme controlled systems (6,7). Several researchers have
reported various colon targeted formulations of indometha
cin. Some of these formulation approaches include using pH
sensitive polymers for coating drug loaded pellets (8,9),
compression coating of tablets using either guar gum (10)
and pectin and chitosan mixtures (11) or guar gum and
Eudragit FS 30 D coated pellets (12), and drug embedding in
HPMC/pectin/calcium chloride matrix bases (13).
pH based colon specic drug delivery systems have been
developed by coating drug embedded polysaccharide matri
ces (both single unit and multi unit systems) with pH
dependent polymers (14,15). A summary of some observations
and important ndings with respect to Eudragit based
coating polymers is presented in Table I. It was shown
through in vitro release studies that these polymers (either
alone or in combination) exhibit excellent protection in
gastric pH followed by gradual or sudden release in alkaline
environment in different pH conditions (6.0 7.4; Table I).
Some coated formulations, based on Eudragit FS 30D, have
shown to resist disintegration/dissolution in upper GI tract
but have been reported to disintegrate after colonic arrival
(12,22 24). However, the use of polymers that release the
418
Eudragit FS30D
Eudragit FS 30 D
Eudragit S100
Eudragit S100
Eudragit S100
Eudragit S100
Eudragit S100
Single polymer coated systems

Eudragit L100
pH sensitive polymer
Important ndings
In case of guar gum with methacrylic acid graft copolymer,

up to 70% drug release occurred during initial 45 h from
The release was reduced to 1824% after enteric coating
with EL100
Pectin matrix of theophylline coated with Eudragit S100
Presence of Eudragit coat (27% w/w) prevented initial drug release
(near zero percent release for rst 300340 min) in pH 1 1 (2 h),
pH 6 8 (2 h), pH 7 4 (10 h)
Indomethacin pellets coated with Eudragit S100
In vitro testing in pH 1 2 (2 h) followed by pH 6 8 18% drug release
in rst 3 h with 3% coat
Matrix tablets prepared by direct compression of mixtures
In vitro release studies in simulated gastric uid pH indicated that the
Eudragit S100 enteric-coated matrix tablets showed both gastric
of hydroxyethylcellulose and ethylcellulose or micro-crystalline
resistance and an adequate lag time followed by a controlled-release
cellulose and coated with Eudragit S100
phase Tablets coated with 1 0 3 0 7 (w/w) drug/MCC/HEC tablets,
released drug after a 260 min lag time and completed 90%
release in 10 h
Tablet cores were coated with Eudragit S dissolved in ethanol
Tablets coated with Eudragit S (aqueous) disintegrated in all
(organic), or aqueous dispersion (aqueous), or Eudragit FS
volunteers mainly in the proximal to mid small intestine Eudragit S
aqueous dispersion and administered to human subjects
(organic) tablets failed to disintegrate in three out of eight volunteers,
while disintegration was in the ileocecal junction and ascending
colon in all others Eudragit FS coated tablets disintegrated in 14 out
of the 16 administrations with the site of disintegration focused on the
ileocecal junction and ascending colon
Tablets (radiolabeled with Technetium 99m) coated with
Gastrointestinal pH showed variability between and within individuals
Eudragit S were administered to eight healthy subjects in
but no differences were seen between pre-feed and fasted states
a three-way crossover study after overnight fasting
Three tablets failed to disintegrate in pre-feed and fed regimens and
one in the fasted state; this was attributed to ileocecal pH and
ileocecal junction residence time
Meloxicam-loaded cores prepared by layering drug-binder
The in vitro drug release from the pellets was pH-dependent with
(HPMC)-solubilizer (beta-cyclodextrin) solution onto
sufcient gastric resistance (pH 1 2 no release; pH 6 8 6%;
nonpareils and then coated with Eudragit FS 30 D
pH 7 0 52%; pH 7 2 100%; pH 7 4 100%, after 3 h incubation)
The onset of meloxicam absorption from the coated pellets with
15% (w/w) Eudragit FS 30 D (3 00 8 h) was signicantly delayed
compared to that from the uncoated drug-layered cores (0 60 3 h)
The combination of Eudragit RL30D and RS30D were used as
The release prole of tablets was studied in three phosphate buffers with
sustained-release lm, and Eudragit FS30D used as enteric lm,
the pH 6 5, 7 0 or 7 4 for 12 h after a simulated gastric presoak for 2 h
which was expected to release drug depending on pH and time
in 0 1 M HCl For the in vitro study, there was no drug released in
0 1 M HCl for 2 h, and release occurred slowly when pH was above
6 5 For the in vivo study, the coated tablets remained intact in the upper
gastrointestinal tract, and drug release began after the colonic arrival
Matrix tablets of metronidazole prepared with

polysaccharides and guar gum grafted with methacrylic
acid copolymer and coated with Eudragit L100
Technique employed
Table I. Use of pH Responsive Polymers in Coating of Single-unit Dosage Forms for Colon Specic Release
(23)
(22)
(21)
(20)
(19)
(18)
(17)
(16)
Reference
Indomethacin Matrix Tablets for Colon Specific Delivery

419
Important ndings
Eudragit L100 and Eudragit S100 Mesalazine (5-aminosalicylic acid) tablets coated with mixture
of both polymers
Eudragit S100, starch
Tablets coated with a mixture of pH-responsive enteric polymer
(Eudragit S100) and biodegradable polysaccharide
(resistant starch) in a single layer matrix lm
Eudragit L100 and Eudragit S100 Indomethacin pellets coated with Eudragit S100 and Eudragit
L100 as pH-dependent polymers and Eudragit RS was used
as a time-dependent polymer as a single coating formulation
Eudragit L100 and Eudragit S100 Indomethacin pellets coated with Eudragit S100 Eudragit L100
(1 4, 1 1 and 1 0) at different level of coating (10%, 15%
and 20%, w/w), respectively
5-ASA release from the coated pellets depended upon both the combination
ratio of the Eudragit L100 and ES100 and the thickness of the coating
layer For tablets coated with 1 4 (EL100 ES100) mixture, drug release
was <1 0% in pH 1 2 (2 h), <3 0% in pH 6 0 (1 h) and about 80%
release in pH 7 2
Pellets released no indomethacin at pH 1 2 (simulating stomach pH)
and pH 6 5 (simulating proximal part of small intestine pH); drug release
was slow at pH 6 8 (simulating lower part of small intestine pH), but it
was fast at pH 7 2 (simulating terminal ileum pH)
Dissolution studies of pellets in the media with different pH
(1 2, 6 5, 6 8, and 7 2) showed that drug release in colon could be
controlled by addition of Eudragit RS to the pH-dependent polymers
The lag time before drug release could be controlled by coating level
Initial tablet disintegration occurred in the ileocecal region in three subjects
and the ascending colon in ve subjects (5 650 86 h after dosing)
In vivo studies involving human subjects showed that the coated tablets were
able to resist breakdown in the stomach and small intestine
Consistent disintegration of the dosage form was seen at the ileocecal
junction/large intestine
Under gradient pH conditions (pH = 1 2, 6 8, 7 4, and 6 5 for 2, 2, 1,

and 15 h, respectively), indomethacin was released from Eudragit
FS30D-coated pellets quickly after changing pH to 7 4 For guar
gum/Eudragit FS30D double-coated pellets, only about 5% of the
drug was released after another 1 h, showing retarding effect by
guar gum coating After changing pH to 6 5 and addition of
galactomannanase, enzyme-dependent drug release was observed
Pharmacokinetic study in beagle dogs showed good lag times for
release of drug
Enteric coating of HPMC capsules containing paracetamol with
Capsules coated with Eudragit L 30 D-55 were gastro resistant for
two pH responsive polymers separately
2 h at pH 1 2 in vitro and disintegrated completely in small intestine
in vivo Those coated with Eudragit FS 30 D were resistant for a further
1 h at pH 6 8 and disintegrated in the proximal colon in an average
time of 6 9 h post dose
Tablet cores were coated with Eudragit S100 dissolved in ethanol Dissolution studies done in pH 1 2 and two compendial buffers in the
(organic), Eudragit S100 aqueous dispersion (aqueous), or
pH range 6 87 4 and Hanks physiological buffer (pH 7 4 ) All tablets
Eudragit FS aqueous dispersion and Eudragit P4135
except those coated with P 4135 prevented drug release in acid medium
At pH 7 4, drug release was in the following order Eudragit S100
(aqueous dispersion)>Eudragit FS>Eudragit S100 (organic solution)
Results indicated that tablets coated with Eudragit FS would be
appropriate for ileocolonic delivery
A pH- and enzyme-dependent system obtained by coating guar

gum and Eudragit FS30D sequentially onto drug-loaded pellets
in a uidized bed coater
Technique employed
Dual polymer coated systems

Eudragit L100 and Eudragit S100 5-Amino salicylic acid loaded pellets coated with both polymers
in combination
Eudragit S100, Eudragit FS

30 D, Eudragit P4135
Eudragit L30 D-55

and Eudragit FS 30 D
Eudragit FS30D
pH sensitive polymer
Table I. (continued)
(28)
(27)
(9)
(8)
(26)
(25)
(24)
(12)
Reference
420
Asghar, Chure and Chandran

drug at higher pH values (>7.0) may fail to give reproducible
results, since pH in the lower GI lumen (ileum and colon)
may fail to exceed the dissolution pH of the polymer in
some patients, for example, in case of inammatory bowel
disease (29). Therefore, coated systems, in general, suffer
from the drawback of non reproducible release in vivo.
Extensive studies carried out by Ibekwe et al. in the recent
past have shown that tablets coated with Eudragit polymers
demonstrated erratic performance in vivo, and many tablets
failed to disintegrate inside the human body (20,21). This has
been previously attributed to the narrow pH gradient between
the small and large intestine, intersubject variability in GI pH,
residence time of dosage form at ileocecal junction, pH changes
that occur in diseased conditions, and fasted or fed states
resulting in variable performance of these systems (30,31).
The combination of pH dependent polymers with time
based polymers could offer a means for achieving controlled
release of drug from the coated system (9). Furthermore, pH
based polymers in combination with biodegradable guar gum
(12) and starch (28) have also been attempted and proven as
better triggers by microbial degradation for colon specic
release. In the present investigation, a novel method of matrix
tablet design possessing bimodal release prole was envisaged.
The use of pH dependent polymers (Eudragits L100 or S100) in
polymer matrix of hydrophilic polymer (xanthan gum) was
envisaged. Previously, our group has reported the use of pH
sensitive polymers alone in matrix (32) and in combination with
other polymers ethyl cellulose (32) or polycarbophil and
carbopol (33) for potential colon specic delivery. Application
of Eudragits (L100 and S100) in combination with ethyl
cellulose as multiunit matrix microsphere for colon specic
delivery has been reported (34). It is expected that a dual
polymer matrix embedded system comprising of a combination
of swelling controlled and pH dependent polymers can offer a
suitable means of achieving a pH and time dependent system
that releases the drug in a bimodal (sigmoidal fashion). A matrix
system, upon exposure to alkaline environment of the colon will
result in partial or complete dissolution of pH responsive
polymers and will therefore generate a porous system that will
facilitate entry of dissolution medium into the pores of the
matrix and affect drug release by diffusion and matrix erosion in
high pH region. Such a system will result in a release prole
suitable for colonic delivery and may help to reduce the
improbability in drug release from a coated system, wherein
the core is unexposed and drug release can occur only after all
the layers of the coat are dissolved.
Xanthan gum, a polysaccharide based natural gum, has
been widely employed as a hydrophilic polymer to prepare
controlled release matrices because of its cost effectiveness
and regulatory acceptance (35). It has been used as a release
retardant polymer alone (36) or in combination with other
polymers [galactomannan (37), chitosan (38)] in controlled
release matrices. It has also been used as a copolymer with
guar gum to form matrix bases (39) and compression coats
(10) in colonic delivery. When used as a matrix base, xanthan
gum forms a time dependent swelling controlled system. The
drug release from such matrix is through diffusion from the
swollen xanthan gum matrix (40,41).
Therefore, in the present investigation, it was envisaged
that the presence of pH sensitive polymers in a matrix would
control the rapid initial swelling of xanthan gum based
421
matrices and thereby minimize drug release in the acidic to
weakly acidic conditions of the upper GI tract while
enhancing the drug release in the neutral to slightly alkaline
environment of the colon. The objective of the study was to
investigate the effect of pH responsive polymers Eudragit
(L100 or S100) on indomethacin release from xanthan gum
matrix based formulations and evaluate their potential for
controlled release as well as colon specicity. The effect of
varying the polymer proportions (xanthan gum) alone and in
combination with EL100 or ES100 was studied. The transit of
formulations was investigated in Wistar rat, and percentage
drug recovered at different time points was analyzed. The
effect of storage on the stability and release prole of selected
formulations was also investigated. The stored batches were
also evaluated for the absence of physical and chemical
interactions.
Indomethacin (micronized form) was obtained as gift
sample from Ajanta Pharma Ltd, Aurangabad, India. Xan
than gum was purchased from Signet Chem, Mumbai, India.
Eudragit polymers were obtained from Rohm Pharma,
Germany. All other chemicals and reagents used were either
of analytical or pharmaceutical grade.
Analytical Method
Indomethacin in pure form and designed formulation was
analyzed using in house developed and validated UV Visible
spectrophotometric method using Jasco V 570 double beam
UV Visible spectrophotometer (Jasco Corporation, Tokyo,
Japan) accompanied with Spectra Manager software. The
method involved analysis of the drug at 320 nm in phosphate
buffer pH 7.4 in the range of 5 50 g/ml using 1 cm matched
quartz cells.
Tablet Manufacturing
Matrix embedded tablets (each containing 75 mg of
indomethacin) using either XG alone or in combination with
EL100/ES100 were prepared by wet granulation technique.
Batch quantities of drug and polymer(s) pre sieved through
no. 120 mesh (ASTM) and dried at 55C were mixed. The dry
blend was granulated with ethyl alcohol (q.s.) and passed
through no. 40 mesh and dried at 55C on a tray drier. The
dried granules were passed through no. 60 mesh and the
passings blended with 1% w/w talc and 0.5% w/w magnesium
stearate and compressed using 7 mm punches on a 16 station
rotary tablet compression machine (Cadmach, Ahmedabad,
India). Three batches of tablets were prepared for each
formulation. Formulas of prepared matrix embedded tablets
containing XG are presented in Table II, respectively.
Physical Characterization of Designed Tablets
The designed formulations were studied for their phys
ical properties like weight variation, thickness, crushing
strength, friability, and drug content uniformity. For estimat
ing weight variation, 20 tablets of each formulation were
weighed using a Mettler Toledo balance (AG135, Mettler
422

Table II. Composition, Physical Characterization, and Release Rate Kinetics of XG based Formulations
Drug contenta
Batches
XG (mg)
EL100 (mg)
ES100 (mg)
mg/tablet
Power law correlation
Correlation time
span (h)
rf
2 8
1 10
2 12
0.9504
0.9204
0.9167
1.2710
1.8310
1.2310
2 12
2 14
0.9130
0.9750
3.7510
1.1610
2
2
2
2
4
4
2
2
2
1
1
1
0.9942
0.9827
0.9926
0.9852
0.9876
0.9967
0.9950
0.9979
0.9822
0.9758
0.9234
0.9425
3.2410
3.4210
2.6710
1.0510
2.1310
7.2210
9.4510
3.5710
1.3010
3.5310
1.7310
1.6310
Effect of XG only
IXG5
3.75
73.50.2
IXG10
7.5
74.70.1
IXG20
15
72.60.3
Effect of EL100 or ES100 alone on indomethacin matrix
IEL20
15
73.00.1
IES20
15
72.60.2
Effect of EL100 or ES100 on XG matrix
IXG5EL5
3.75
3.75
76.30.4
IXG5EL10
3.75
7.5
73.80.3
IXG5EL20
3.75
15
75.00.1
IXG5EL40
3.75
30
75.40.3
IXG10EL10
7.5
7.5
74.10.2
IXG10EL20
7.5
15
74.20.1
IXG5ES5
3.75
3.75
74.70.2
IXG5ES10
3.75
7.5
74.60.2
IXG5ES20
3.75
15
72.60.3
IXG5ES40
3.75
30
76.90.1
IXG10ES10
7.5
7.5
76.30.4
IXG10ES20
7.5
15
73.50.1
12
14
12
14
12
12
10
12
14
12
10
10
MSSRg
2
2
3
4
3
3
4
4
4
2
3
2
2
Kh (%/hn)
ni
14.277
5.220
15.707
1.13
1.09
0.55
0.5
1.2
1.1
5.1
13.7
23.9
2.659
1.137
1.48
1.40
2.1
3.5
10.8
26.7
7.327
1.346
1.047
1.402
5.945
2.409
12.159
5.417
4.950
0.938
6.276
6.358
1.08
1.61
1.55
1.27
1.05
1.30
0.79
0.78
0.79
1.26
0.40
0.86
2.9
3.6
4.2
5.2
3.7
4.6
2.1
4.1
5.2
6.0
1.0
2.0
10.2
13.6
17.7
26.5
13.3
16.2
12.6
36.7
39.3
37.4
9.2
9.1
10%
90%
Each tablet contains 75 mg of indomethacin. Also contains 1% w/w talc and 0.5% w/w magnesium stearate as formulation additives. The
diameter of the tablets was 0.700.01 cm
a
MeanSD (n 10)
b
SD from the mean value (n 20)
c
MeanSD (n 10)
d
Mean of ten tablets
e
MeanSD (n 5)
f
Correlation coefcient
g
Mean sum of squared residuals
h
Release rate constant
i
Diffusional exponent indicative of the release mechanism
j
Time for 10% of the drug release (in h)
k
The predicted or calculated time for 90% of the drug release (in h from Eq. 2)
Toledo, GmbH, Greifensee, Switzerland). The crushing

strength of ten tablets was measured using Monsanto
(standard type) tablet hardness tester. Friability was deter
mined on ten tablets in a Campbell Electronic Friabilator for
4 min at 25 rpm. For estimation of drug content, ten tablets
were crushed, and the aliquot of powder equivalent to 10 mg
of drug was extracted in methanol/phosphate buffer pH 7.4
(1:9), suitably diluted using phosphate buffer pH 7.4 and
analyzed spectrophotometrically at 320 nm.
In vitro Release Studies
In vitro dissolution studies were carried out using USP
Type II (paddle method) apparatus (Electrolab TDT 08L
with autosampling unit, Mumbai, India) at 75 rpm. The
dissolution was carried out for the rst 2 h in distilled water
(500 ml, pH 6.8 7.0). Then, 200 ml of phosphate buffer
concentrate (4.75 g of KH2PO4 and 1.07 g of NaOH in
distilled water) was added to raise the total media volume to
700 ml and the pH to 7.4 for the remaining period. At
predetermined time intervals, a 10 ml sample was withdrawn
and replaced with fresh dissolution media. The samples were
ltered, suitably diluted using phosphate buffer pH 7.4, and
analyzed spectrophotometrically at 320 nm. The release

studies were conducted in duplicate per batch for three
batches, and the mean values from three batches along with
the SD were plotted against time (Figs. 1, 2, and 3) and used
for all further calculations. Dissolution of pure indomethacin
alone was also recorded and used as a control for in vitro
release study (shown in Fig. 1). Effect of Eudragit on drug
release from XG matrix was compared against an indometh
acin matrix formulation prepared with only EL100 or ES100
(20% w/w of drug; IEL20 and IES20) as a control (shown in
Figs. 2 and 3).
Effect of Simulated GI Fluid pH (Without Enzymes)
on Release
Selected formulations from previous study were studied
in a medium of changing pH. The initial condition was 350 ml
of 0.1 N HCl (pH 1.2) for 0 2 h. From 2 4 h, the pH of the
media was raised to 4.5 (for simulation of duodenum), with
total dissolution media volume of 600 ml. From the fourth
hour onwards, the pH was raised to 7.4 by adding 300 ml
phosphate buffer concentrate (2.18 g of KH2PO4 and 1.46 g of
NaOH in distilled water) to get dissolution volume of 900 ml.
423
Fig. 1. Release prole of indomethacin from matrices containing

varying proportions of xanthan gum. Each data point represents
meanSD (n 6)
Fig. 3. Release prole of indomethacin from xanthan gum matrix

showing effect of varying proportion of ES100. Each data point
represents mean SD (n 6)
The study was further continued until the end in 900 ml

volume. At predetermined time intervals, a 10 ml sample was
withdrawn and replaced with fresh dissolution media. After
appropriate dilutions, the samples were analyzed by the UV
method discussed in previous section. The corresponding
release proles against pure indomethacin dissolution in the
respective media are presented in Fig. 4.
for Fickian release (diffusion controlled), >0.45 and <0.89 for

non Fickian release (diffusion and polymer relaxation), 0.89
for case II release (only relaxation and swelling), and >0.89
for super case II release (relaxation and erosion) for swel
lable systems. For cylindrical systems like tablets, the n values
of 0.45 and 0.89 represent pure diffusion or erosion
controlled release, respectively. The values of the coefcient
were calculated using linear regression analysis between
log Mt =M1 and log t data obtained from drug release
studies on MS Ofce Excel 2003 software. The value of n was
obtained as slope of the regression equation, and K was
calculated as antilog of the intercept value. The t10% (time
required for 10% drug release) was determined directly from
the plot of cumulative percentage drug released versus time,
while the t90% (time required for 90% drug release) was
calculated as
Characterization of Sigmoidal Release Profiles by Power

Law Equation
In order to understand the mechanism of drug release
from these formulations, the cumulative percentage drug
release data (after 2 h) was tted into the power law equation
given by Korsemeyer et al. (42) and Ritger and Peppas (43).
Mt =M1 Kt n
t90% anti log flog 90
log K =ng
where Mt =M1 is percentage of drug released at any time t; K

is release rate constant incorporating the structural and
geometric characteristics of the polymeric system and the
drug; and n is the diffusion exponent indicative of the release
mechanism of the drug. The value of n for a cylinder is <0.45
The values of correlation time span, K, n, t10%, and t90%, r

(correlation coefcient of the regression analysis), and mean
sum of squared residuals (MSSR) as obtained from the
dissolution data of designed formulations are given in Table II.
Fig. 2. Release prole of indomethacin from xanthan gum matrix

showing effect of varying proportion of EL100. Each data point
represents meanSD (n 6)
Fig. 4. Release prole of selected formulations in simulated GI uid

pH. Each data point represents meanSD (n 6)
424

range of 25 200C under constant purge of nitrogen gas (ow
rate of 30 ml/min) using differential scanning calorimeter
(Shimadzu, Japan, Model DSC 60). For Fourier transform
infrared (FTIR), the samples were appropriately diluted with
dried potassium bromide, and IR spectra were acquired in the
range of 400 to 4,000 cm 1 with a resolution of 4 cm 1. The
data was processed using Kubelka Munk method for baseline
correction.
The correlation time span is the period of drug release phase

taken for calculation of release kinetics. Using the calculated
values of K and n, the release proles were predicted beyond
14 24 h for each formulation and are shown as dotted trend line
(s) in the respective gures.
The dissolution proles of selected formulations in
changing pH medium (without enzymes) were compared
with the target dissolution prole (negligible to no release in
rst 6 h followed by controlled release up to 14 16 h) using f1
(dissimilarity) and f2 (similarity) factor (44) as shown below.
("
f1
n
X
jRt
Tt j
#,"
n
X
t 1
In vivo Evaluation of Formulations
#)
100
Rt
The GI transit of selected formulation (IXG10EL10) was

carried out in vivo in rats. The protocol was previously
approved by the Institutional Animal Ethics Committee and
was in conformance with Principles of Laboratory Animal
Care (NIH publication no. 85 23, revised in 1985). Healthy
male Wister rats (350 400 g) were selected for the study. Mini
tablets were prepared using 4 mm punches, and formulation
parameters were optimized suitably to approach as close as
possible to the drug release pattern of the actual formulation.
Before tablet administration, the animals were kept on
overnight fast. The formulation was administered in triplicate
for GI transit studies. The tablet was placed in the throat of
the animal with a pair of forceps, and about 1.5 2 ml of water
was ushed down the throat slowly to facilitate entry of tablet
into esophagus with the help of syringe. The animals were
killed at xed time intervals, and the position of tablet was
located. The recovered tablets at various time points were
analyzed for residual drug content to estimate the amount of
drug released at each time point.
t 1
8"
n
<
X
f2 50: log
1 1=n
R t
:
t 1
#0:5
Tt
100
9
=
where n is the number of sampling points, Rt and Tt is the drug

release from reference and test sample at sampling point t,
respectively. The corresponding data are presented in Table III.
Batch Reproducibility and Stability on Storage
Three batches of each formulation were prepared, and
the release studies were done using the same conditions for
estimating batch reproducibility. In order to assess the long
term stability of the various formulations prepared, selected
formulations from each batch were sealed in cellophane
packets, placed in hermetically sealed vials and separately
stored at ambient conditions (25C/60% RH) and accelerated
stability test conditions (40C/75% RH) for 6 months. At the
end of the study period, the formulations were observed for
change in physical appearance, drug content, and in vitro drug
release characteristics. The initial (zero time) results were
compared with post stability testing period results for statis
tical differences. The powdered samples of indomethacin
matrix tablets were also subjected to differential scanning
calorimetry (DSC) study and determination of IR spectrum.
For DSC, pure drug and formulation (IX10EL10) equal to
2.5 mg of drug were accurately weighed onto standard
aluminum pans, hermetically sealed and thermograms
obtained at a scanning rate of 10C/min over a temperature
Data Analysis
The difference in the release data between the different
formulations was compared using paired t test for means and
one way analysis of variance at 5% level of signicance using
Microsoft Ofce 2003, Excel package.
Physical Characteristics
Physical appearance, crushing strength, weight variation ,
and drug content uniformity of different tablet formulations
Table III. Release Kinetics Data from Different Plots for Selected Formulations in Simulated GI Fluid pH
Power law correlation
Batches
IXG5ES5
IXG5EL10
IXG10EL10
a
Correlation time span (h)

4 14
4 14
4 14
ra
0.9613
0.9812
0.9012
MSSRb
3.2710
2.2110
4.8610
3
3
3
Similarity factorg
Kc
nd
t10%e
t90%f
f1
f2
10.114
4.016
10.972
0.86
1.12
0.91
4.2
4.5
4.9
12.7
15.9
14.1
13.95
12.95
2.03
41.84
48.35
71.04
Correlation coefcient
Mean sum of squared residuals
Release rate constant
d
Diffusional exponent indicative of the release mechanism
e
f
g
Comparison with theoretical target release prole. For similarity f2 should be>50 and f1 <15
b
Dissimilarity factorg

were found to be within satisfactory limits. The crushing
strength was found to vary between 4.5 and 5.0 kg. The
percentage friability in all the formulations was observed to
be 0.5%. The manufactured tablets showed low weight
variation (SD within 5% of the average weight of the tablet)
and high degree of drug content uniformity (within 7% of
the theoretical value) indicating that wet granulation is an
acceptable method for good quality matrix tablets of indo
methacin for pH modulated delivery.
In vitro Release of Matrix Tablets
The in vitro release proles of drugs from Eudragit
based systems are normally investigated in buffers like
phosphate buffers with pH ranging from 6.5 to 7.5 after a
pretest in acid medium (45). Indomethacin, an indole acetic
acid derivative with a pKa of 4.5, has been reported to have
solubility of 0.01 mmol/L (3.66 g/ml) in pH 1.2 and
5.52 mmol/L (1,975 g/ml) in pH 7.2 at 37C (46). Our
studies revealed that drug (present in a micronized form) had
a solubility of ~53 g/ml in distilled water at 25C and ~80 g/ml
at 37C, which could be due to its ionization at the pH of
distilled water (6.8 7.0).
Based on this information, dissolution was carried out in
distilled water for the rst 2 h. Saturation solubility will not be
achieved even when 55 60% of labeled claim is released in
the rst 2 h in distilled water. Furthermore, Eudragits are
insoluble in water (47,48), and their dissolution depends on
ionic strength and buffer capacity of the medium (49).
Although the pH of distilled water (6.8 7.0) was near the
threshold pH for dissolution of Eudragit, the polymer is not
expected to dissolve due to its negligible ionic strength.
In addition, previous reports have shown that, as the
patient consumes a tablet with a good quantity of water,
dissolution of poorly water soluble drugs can be done in
distilled water for the initial period (50,51). During prelimi
nary studies, it was observed that drug release from the
formulations could be differentiated well in this medium, as
release never exceeded the saturation solubility of the drug in
any case (Figs. 1, 2, and 3). The release studies were then
investigated in phosphate buffer of pH 7.4 (ionic strength of
0.129 and osmolality of 228 mOsm/kg) (25). The drug was
freely soluble at this pH, and this medium simulates the
alkaline environment of distal small intestine and colon. The
dissolution proles of all the formulations were compared
with respect to t10%, t90%, and n values.
A plot of cumulative percentage released versus time for
matrix tablets of indomethacin prepared using varying
proportions of xanthan gum alone (5%, 10%, and 20% w/w of
drug) against pure indomethacin is shown in Fig. 1. It was
observed that the initial percentage released from all the
formulations was quite high (38% for 5% XG matrix and 15
20% for the others) in the rst 2 h followed by a slower and
more controlled release during the later stages depending on
the proportion of the polymer in the matrix. On the other
hand, pure indomethacin, being highly hydrophobic in nature
(LogP of 4.4) dissolved slowly as a powder. The release
kinetics data when tted to the power equation (Table II) to
get the calculated values of t90% indicated 5.1 h for IXG5,
which was extended to 23.9 h for IXG20 when the proportion
of xanthan gum was increased from 5 to 20% w/w of drug,
425
respectively. The use of higher proportions of xanthan gum
resulted in the formation of a thick polymeric gel layer, which
acted as a barrier to drug diffusion. The values of 1.13 and
0.55 for the diffusional exponent n indicated a change in the
release mechanism from super case II (n>1.0) to anomalous
non Fickian type (0.45<n<0.89). In case of IXG5 and IXG10,
where the indomethacin proportion was relatively higher,
drug release took place due to erosion of tablet surface, due
to limited swelling of xanthan gum in the presence of a
hydrophobic drug. On the other hand, due to the relatively
higher proportion of xanthan gum in IXG20, the drug release
mechanism was elucidated as anomalous due to swelling of
xanthan gum and diffusion of drug through the swollen layer
(40,41). These results implied that xanthan gum alone in
matrix form was not suitable for colonic delivery.
Effect of EL100 on Xanthan Gum Matrices
For the matrix tablets prepared using xanthan gum in
5% w/w of drug with varying proportions of Eudragit L100
(5%, 10%, 20%, and 40% w/w of drug), the in vitro drug
release proles against drug release from indomethacin with
20% EL100 (IEL20) are shown in Fig. 2. The initial
percentage of drug release from all the formulations in the
rst 2 h was almost negligible (less than 5%) in distilled water
followed by an linear increase in release rate post 2 h in
pH 7.4 that depended on the proportion of EL100. The
release kinetics data for the various formulations revealed
t 10% (ranging from 2.9 h for IXG5EL5 to 5.2 h for
IXG5EL40), implying signicant inhibition in the initial drug
release (Table II). It was observed that, after 2 h, the release
of drug from the formulations was extended from 10.2 h for
IXG5EL5 to about 26.5 h for IXG5EL40, indicating exten
sion in duration of release with corresponding increase in
relative proportion of EL100. The drug release from these
formulations was observed to show strong pH dependency in
their release proles and a sigmoidal character that depended
on the relative proportion of EL100. The carboxylic acid
group present in Eudragits reacts with the phosphate bases
(HPO42 ) in the buffer resulting in increase in Eudragit
dissolution rate in pH 7.4 (52). The release proles were also
signicantly different from indomethacin+EL100 (IEL20)
matrix, due to the time dependency in release that was
conferred by the swelling of xanthan gum (40). Thus,
regulating the amount of EL100 in 5% xanthan gum matrix
base could confer desired retardation in initial phase followed
by controlled release ranging from 12 to 28 h.
With increase in the relative proportion of EL100 from
10% (IXG5EL10) to 20% (IXG5EL20), a proportionate
retardation was observed in the corresponding initial release
rates resulting in enhanced t10% values (from 3.6 h for
IXG5EL10 to 4.2 h for IXG5EL20). The drug release
duration was similarly extended from 13.6 h for IXG5EL10
to 17.7 h for IXG5EL20. This was attributed to increase in
total polymer content that resulted in the formation of a
relatively strong matrix with decreased porosity and increased
tortousity. Similarly, with increase in the relative proportion
of EL100 from 10% (IXG10EL10) to 20% (IXG10EL20), a
similar effect on the initial drug release rate (t10% ranging
from 3.7 h for IXG10EL10 to 4.6 h for IXG10EL20) was
observed. The drug release duration was extended from
426
13.3 h for IXG10EL10 to 16.2 h for IXG10EL20. Alternately,
when the relative proportion of xanthan gum was increased
from 5% (IXG5EL10) to 10% w/w of drug (IXG10EL10),
there was insignicant change in the corresponding release
kinetics (Table II). This was also true for IXG5EL20 and
IXG10EL20, which were statistically similar to each other
with respect to the release data. This implied that change in
relative proportion of xanthan gum from 5% to 10% in
20% EL100 matrix did not inuence matrix properties
signicantly.
It has been shown previously that high initial swelling of
xanthan gum based matrices resulted in the release of a
signicant drug load from formulations during the early drug
release phase (41). From the present study, it was inferred that
the presence of pH based polymer EL100 was able to control
the initial rapid swelling of xanthan gum based matrices and
thereby prevent the high percentage of drug release, which was
previously observed for formulations prepared with xanthan
gum alone (nearly 40% drug release for IXG5 in 2 h). A
possible explanation for this could be that, during granulation,
the granulating solvent (ethyl alcohol) dissolved a portion of
EL100, which not only imparted the necessary adhesion
between the matrix components but also formed a layer over
the xanthan gum particles. This may have inhibited the swelling
of the hydrophilic gum in water. The inhibition of drug release in
the initial phase (<2 3% drug release; Fig. 2) is comparable to
that reported earlier for EL100 based coated systems in
simulated gastric uid (16).
Secondly, EL100 in a polymeric base could impart a pH
responsive drug release character. With increase in the pH of
dissolution medium to 7.4, an increase in the drug release rate
was observed on account of matrix erosion due to dissolution
of EL100. The formation of a porous matrix then facilitated
enhanced diffusion of the drug through the pores. This
hypothesis was drawn from a previous report that Eudragit
L100 acted as a pore former in a matrix at a higher pH range
(53). The values of n from Peppas equation for XG+EL100
series ranged from 1.05 to 1.61, indicating release mechanism
to be super case II type due to increase in matrix erosion,
which can be attributed to dissolution of Eudragit after 2 h
in pH 7.4 medium. With the exception of IXG5EL40, all
other formulations demonstrated signicantly pH and time
dependent sigmoidal drug release characteristics suitable for
colonic delivery. Thus, it was concluded that regulating the
relative proportion of EL100 would help attain the desired drug
release pattern from a xanthan gum based matrix.
Effect of ES100 on Xanthan Gum Matrices
A plot of cumulative percentage released versus time for
matrix tablets of indomethacin prepared using xanthan gum
in 5% w/w of drug with varying proportion of Eudragit S 100
(ES100) (5%, 10%, 20%, and 40% w/w of drug) when
compared to indomethacin with ES100 (IES20) is shown in
Fig. 3. It was observed that, on increasing the relative
proportion of ES100 from 5% to 40% w/w of drug, there
was proportionately greater retardation in the initial release
rate as indicated by the t10% (ranging from 2.1 h for IXG5ES5
to 6.0 h for IXG5ES40). Similarly, drug release was extended
from 12.6 h for IXG5ES5 to 37.4 h for IXG5ES40, resulting in
a release prole similar to that obtained for indomethacin+

ES100 only matrix (IES20) matrix (Table II). The release
kinetics calculated for these formulations were signicantly
different when compared to the matrix base IXG5 (Fig. 1).
The t10% and t90% values were found to be higher in the case
of ES100 based xanthan gum matrices when compared to
EL100 probably due to the difference in pH solubility (pH 6.0
for EL100 and 7.0 for ES100) of the two polymers. With
increase in relative proportion of hydrophilic gum in matrix,
the release proles differed signicantly from indomethacin+
ES100 matrix (IES20) matrix as observed for IXG5ES10 and
IXG10ES10, and also for IXG5ES20 and IXG10ES20. The
hydrophilic component may have aided in the penetration of
dissolution uid that explains the increase in release rate with
increase in percentage of hydrophilic polymer, thereby facilitat
ing the release of a highly hydrophobic drug from the matrix.
When the relative proportion of xanthan gum in the
matrix was increased from 5% (IXG5ES10) to 10% of drug
(IXG10ES10), the release rates were found to increase
(Fig. 3). The signicantly lowered t10% (1.0 h) and t90% (9.2 h)
values for IXG10ES10 are indicative of this. Similar release
kinetics were observed for IXG10ES20 (t10% of 2.0 h and t90% of
9.1 h), which were signicantly different from IXG5ES20.
As observed for xanthan gum EL100 formulations, the
retardation in drug release rate was found to depend on the
relative proportion of ES100, as shown by the increase in t10%
and t90% values with increase in proportion of ES100 (Table
II). Although good retardation in the initial release phase was
observed, there was considerable deviation from the theoret
ical target of 80 90% release in 14 16 h. As observed in case
of EL100 based xanthan gum formulations, ES100 based
formulations could successfully retard the initial release, and
except for IXG10ES20, all formulations showed less than 3%
release in rst 2 h, implying that ES100 was as effective in a
matrix form in preventing drug release in vitro as it is when
used as a coating polymer (17 19).
The values of n for XG+ES100 series ranged from 0.40
to 1.26, indicating that with an increase in relative proportion
of ES100 in the matrix, release mechanism shifted from
anomalous (matrix swelling and diffusion) to super case II
(erosion type), implying that drug release could have
occurred by a combination of several processes like diffusion,
swelling of hydrophilic component (polymer relaxation), and
erosion of matrix (due to dissolution of Eudragit S100) in
alkaline media. Therefore, it was concluded that matrices with
10% xanthan gum with varying proportions of ES100 demon
strated desirable release kinetics in vitro and indicate good
potential for site specic controlled drug delivery to the colon.
High values of correlation (r ranging from 0.9425 to
0.9979) and very low values of MSSR (2.6710 4 to 1.73
10 2) indicate goodness of t of dissolution data to the power
law equation for xanthan gum and Eudragit (EL100 and
ES100) matrices (Table II).
Effect of Simulated GI uid pH (Without Enzymes)
on Release
For an ideal colon targeted drug delivery system, the
drug release should be prevented in the stomach and small
intestine. Release of drugs must be completed within the
residence time of the dosage form in the colon. Since colonic
residence is highly variable (10 to 30 40 h) (54,55), it was

thought that a drug release program designed for intermedi
ate range of 14 16 h would ensure that maximum drug release
would occur even in cases when colonic transit time is on the
lower side as is the case in various pathologies of the bowel.
Therefore, in the case of the present study, it was assumed that
for colon targeting purpose, a 14 to 16 h extended release
formulation with a delay in onset of about 4 6 h would be
suitable. This time lag would ensure the passage of the
formulation intact through the stomach and small intestine
without appreciable drug loss. These two assumptions were
used to dene a theoretical target release prole shown in Fig. 4.
The in vitro release studies conducted in the initial
dissolution conditions were intended to characterize and
understand the effect of Eudragit on hydrophilic matrix
swelling and initial drug release (in distilled water medium
for 2 h) and also to investigate the potential of the various
formulations to complete drug release in the stipulated time
frame of 14 16 h in the alkaline environment of colon (pH 7.4
medium). The performance of selected designed formula
tions, IXG5ES5, IXG5EL10, and IXG10EL10 was also
evaluated in a pH gradient system in order to investigate
the suitability of formulations in real time changing pH
situation existing in GI tract (Fig. 4). The choice of pH
conditions was pH 1.2 for a duration of 2 h (simulated gastric
uid), pH 4.5 for 2 h (simulated duodenum) followed by
pH 7.4 (simulated distal ileum and colon) for the remaining
period of study. The drug release from the various formula
tions was compared with the theoretical target values using f1
(dissimilarity) and f2 (similarity) factors (Table III). It was
observed that the t10% and t90% for IXG10EL10 were 4.9 and
14.1 h, thereby approaching close to target values (f2 >50 and
f1 <15). The drug release from other formulations (IXG5ES5
and IXG5EL10) was also characterized by a sigmoidal
pattern and showed only a minor deviation from the target
prole (Fig. 4). A signicant pH and time dependent release
pattern was observed for these formulations implying suit
ability for colon specic release.
In combination with Eudragit, it was possible to obtain
desirable release kinetics from xanthan gum even when
employed in very low polymer proportions of 5% and 10%.
Furthermore, such matrices can work on the principle of a
dual trigger mechanism pH dependency and time depen
dent swelling thereby ensuring bimodality in release. By
utilizing a suitable blend of hydrophilic (XG) and slightly
hydrophobic (EL100 or ES100) polymers, it was possible to
regulate drug release from a matrix to achieve desirable
release kinetics (Fig. 4).
Therefore, from the present study, it can be concluded
that the use of pH based polymers in combination with
hydrophilic polymer(s) like xanthan gum to form a polymeric
matrix base controls the initial swelling of these polymers to a
good extent, which could prevent early drug loss from their
matrices during upper GI transit. It also confers matrix
strength and rigidity to the formulations, thereby enabling
lower proportions of these polymers to be used in matrix bases.
427
reliable and reproducible. There was no change in the
physical appearance or in the drug content of the different
formulations at the end of the sixth month storage period at
40C/75% RH (data not shown). Furthermore, in vitro
release studies carried out on the formulations stored at
accelerated test conditions indicated no statistically signicant
change in the drug release proles when compared to
formulations stored at ambient conditions (data not shown).
These results imply good stability of product on long term
storage. DSC thermograms obtained for pure drug and
formulations before and after storage revealed that the
melting endotherm and enthalpy of fusion of drug were well
preserved in all cases. Furthermore, FTIR studies showed
that there was no change in the IR spectrum of drug in
formulation (IXG10EL10), and all peaks of pure drug were
well preserved. This implied absence of physical and chemical
interaction between drug and formulation excipients.
In vivo Evaluation of Formulations
The total length of isolated rat intestine from the
stomach was found to be 1302.5 cm. The formulation
IXG10EL10 (administered as mini tablet) was recovered at
regular time intervals at a distance of 19.252.47 cm (duode
nal region) at 2 h, 67.5010.61 cm (small intestine) at 4 h,
115.503.54 cm (cecum) at 6 h and 122.503.24 cm (colon) at
8 h. The percentage drug released in GI tract at each time
point was calculated by subtracting the percentage drug
recovered from each tablet from 100. The percentage tablet
at the end of fourth hour was nearly 80%, indicating that drug
release was minimal (20%) in upper GI tract (Fig. 5).
However, release may have been rapid afterwards, as percent
age drug recovered at 6 h (from cecal region) was only 25%
and that from the colon was less than 15%. The high amount
of drug loss in cecal region is attributed to the relatively higher
pH of the cecum (6.580.4) that could have dissolved the
Eudragit polymers and enhanced drug release (56). Thus, it
was concluded that, as drug loss during transit through
stomach and small intestine was minimal, the formulation
could act as a potential colon specic drug delivery device.
CONCLUSION
Controlled release systems for colon specic drug were
developed successfully and were found to possess acceptable
Batch Reproducibility and Stability on Storage

No signicant difference was observed in the release
prole of different batches of each matrix formulation,
indicating that the manufacturing process employed was
Fig. 5. Percentage drug released from formulation IXG10EL10 at

various time points in male Wistar rats (n 3)
428
physical characteristics. Drug release from almost all the
matrix bases was characterized by negligible release in the
initial phase followed by controlled release for a time period
of 14 16 h, which is the normal residence time of a solid
dosage form in the colon. Formulations when subjected to
stability studies indicated no signicant change in physical
appearance, drug content, and in vitro release pattern.
Furthermore, no physical and chemical interaction was
evident from DSC and FTIR studies, indicating stability of
indomethacin in the prepared matrices. An advantage of such
a matrix design that comprises of pH dependent polymers in
polysaccharide matrices is that it can overcome the drawbacks
of coated systems wherein there is a possibility of the coat
remaining insoluble during its passage through the colon.
REFERENCES
1. Hull MA, Gardner SH, Hawcroft G. Activity of the non
steroidal anti inammatory drug indomethacin against colorec
tal cancer. Cancer Treat Rev. 2003;29:309 20.
2. Mandayam S, Huang R, Tarnawski AS, Chiou SK. Roles of
survivin isoforms in the chemopreventive actions of NSAIDS on
colon cancer cells. Apoptosis. 2007;12:1109 16.
3. Kapitanovi S, aev T, Antica M, Kralj M, Cavri G, Kreimir
P, et al. Effect of indomethacin on E cadherin and catenin
expression in HT 29 colon cancer cells. Exp Mol Pathol.
2006;80:91 6.
4. Fujino H, Chen XB, Regan JW, Murayama T. Indomethacin
decreases EP2 prostanoid receptor expression in colon cancer
cells. Biochem Biophys Res Commun. 2007;359(3):568 73.
5. Goodman GA, Goodman LS. Analgesics antipyretics; pharma
cotherapy of gout. In: Brunton, editor. The pharmacological
basis of therapeutics, 11th edn. New York: Mcgraw Hill; 2006. p.
699 700.
6. Kinget R, Kalala W, Vervoot L, Mooter GU. Colonic drug
targeting. J Drug Target. 1998;6:129 49.
7. Rubinstein A. Approaches and opportunities in colon specic
drug delivery. Crit Rev Ther Drug Carr Syst. 1995;12:101 49.
8. Akhgari A, Garekani HA, Sadeghi F, Azimaie M. Statistical
optimization of indomethacin pellets coated with pH dependent
methacrylic polymers for possible colonic drug delivery. Int J
Pharm. 2005;305:22 30.
9. Akhgari A, Sadeghi F, Garekani A. Combination of time
dependent and pH dependent polymethacrylates as a single
coating formulation for colonic delivery of indomethacin pellets.
Int J Pharm. 2006;320:137 42.
10. Krishnaiah YSR, Satyanarayana S, Rama Prasad YV, Narasimha
RS. Evaluation of guar gum as a compression coat for drug
targeting to colon. Int J Pharm. 1998;171:137 46.
11. Fernndez Hervs MJ, Fell JT. Pectin/chitosan mixtures as coat
ings for colon specic drug delivery: an in vitro evaluation. Int J
Pharm. 1998;169:115 9.
12. Ji C, Xu H, Wu W. In vitro evaluation and pharmacokinetics in
dogs of guar gum and Eudragit FS 30D coated colon targeted
pellets of indomethacin. J Drug Target 2007;15(2):123 31.
13. Wu B, Deng D, Lu Y, Wu W. Biphasic release of indomethacin
from HPMC/pectin/calcium matrix tablet: II. Inuencing varia
bles, stability and pharmacokinetics in dogs. Eur J Pharm
Biopharm. 2007;67(3):707 14.
14. Sinha VR, Kumria R. Binders for colon specic drug delivery: an
in vitro evaluation. Int J Pharm. 2002;249:23 31.
15. Chourasia MK, Jain SK. Design and development of multi
particulate system for targeted drug delivery to colon. Drug
Deliv. 2004;11(3):201 7.
16. Mundargi RC, Patil SA, Agnihotri SA, Aminabhavi TM.
Development of polysaccharide based colon targeted drug
delivery systems for the treatment of amoebiasis. Drug Dev
Ind Pharm. 2007;33(3):255 64.

17. Mura P, Maestrelli F, Cirri M, Gonzlez Rodrguez ML, Rabasco
Alvarez AM. Development of enteric coated pectin based ma
trix tablets for colonic delivery of theophylline. J Drug Target.
2003;11(6):365 71.
18. Sinha VR, Kumria R. Coating polymers for colon specic drug
delivery: a comparative in vitro evaluation. Acta Pharm. 2003;53
(1):41 7.
19. Alvarez Fuentes J, Fernndez Arvalo M, Gonzlez Rodrguez
ML, Cirri M, Mura P. Development of enteric coated timed
release matrix tablets for colon targeting. J Drug Target. 2004;12
(9 10):607 12.
20. Ibekwe VC, Fadda HM, Parsons GE, Basit AW. A comparative
in vitro assessment of the drug release performance of pH
responsive polymers for ileo colonic delivery. Int J Pharm.
2006;308:52 60.
21. Ibekwe VC, Liu F, Fadda HM, Khela MK, Evans DF, Parsons
GE, et al. An investigation into the in vivo performance
variability of pH responsive polymers for ileo colonic drug
delivery using gamma scintigraphy in humans. J Pharm Sci.
2006;95(12):2760 6.
22. Gao C, Huang J, Jiao Y, Shan L, Liu Y, Li Y, et al. In vitro release
and in vivo absorption in beagle dogs of meloxicam from Eudragit
FS 30 D coated pellets. Int J Pharm. 2006;322(1 2):104 12.
23. Li Y, Li HJ, Yang GR, Gu WP, Ma YK, Zhang MH, et al. Colon
specic delivery tablets of sodium 4 aminosalicylic acid. Yao Xue
Xue Bao. 2006;41(10):927 32.
24. Cole ET, Scott RA, Connor AL, Wilding IR, Petereit HU,
Schminke C, et al. Enteric coated HPMC capsules designed to
achieve intestinal targeting. Int J Pharm. 2002;231(1):83 95.
25. Ibekwe VC, Fadda HM, McConnell EL, Khela MK, Evans DF,
Basit AW. Interplay between intestinal pH, transit time and feed
status on the in vivo performance of pH responsive ileo colonic
release systems. Pharm Res. 2008;25(8):1828 35.
26. Cheng G, An F, Zou MJ, Sun J, Hao XH, He YX. Time and pH
dependent colon specic drug delivery for orally administered
diclofenac sodium and 5 aminosalicylic acid. World J Gastro
enterol. 2004;10(12):1769 74.
27. Brunner M, Lackner E, Exler PS, Fluiter HC, Kletter K,
Tschurlovits M, et al. 5 aminosalicylic acid release from a new
controlled release mesalazine formulation during gastrointestinal
transit in healthy volunteers. Aliment Pharmacol Ther. 2006;23
(1):137 44.
28. Ibekwe VC, Khela MK, Evans DF, Basit AW. A new concept in
colonic drug targeting: a combined pH responsive and bacterial
ly triggered drug delivery technology. Aliment Pharmacol Ther.
2008;28(7):911 6.
29. Ewe K, Schwartz S, Petersen S, Press AG. Inammation does
not decrease intraluminal pH in chronic inammatory bowel
disease. Dig Dis Sci. 1999;44:1434 9.
30. Friend DR. Colon specic drug delivery. Adv Drug Deliv Rev
1991;7:149 99.
31. Rubinstein A. Colonic drug delivery. Drug Discov Today
Technol. 2005;2(1):33 7.
32. Asghar LF, Chandran S. Design and evaluation of pH modulated
controlled release matrix systems for colon specic delivery of
indomethacin. Pharmazie. 2008;63:736 42.
33. Asghar LFA, Chandran S. Design and evaluation of matrices of
Eudragit with polycarbophil and carbopol for colon specic
delivery. J Drug Target 2008;16:741 57.
34. Chandran S, Kango SS, Asghar LFA. Microspheres with pH
modulated release: Design and characterization of formulation
variables for colonic delivery. J Microencapsul 2008;22:1 11.
doi:10.1080/02652040802424021.
35. J. Varshosaz, N. Tavakoli, F. Kheirolahi. Use of hydrophilic natural
gums in formulation of sustained release matrix tablets of tramadol
hydrochloride. AAPS PharmSciTech 2006;7(1) Article 24.
36. Andreopoulos AG, Tarantili PA. Xanthan gum as a carrier for
controlled release of drugs. J Biomater Appl. 2001;16(1):34 46.
37. Vendruscolo CW, Andreazza IF, Ganter Ferrero JLC, Bresolin
TM. Xanthan and galactomannan (from M. scabrella) matrix
tablets for oral controlled delivery of theophylline. Int J Pharm.
2005;296(1 2):1 11.
38. Phaechamud T, Ritthidej GC. Sustained release from layered
matrix system comprising chitosan and xanthan gum. Drug Dev
Ind Pharm. 2007;33(6):595 605.

39. Sinha VR, Kumria R. Polysaccharide matrices for microbially
triggered drug delivery to the colon. Drug Dev Ind Pharm 2004;30
(2):143 50.
40. Talukdar MM, Rombaut P, Kinget R. The release mechanism of
an oral controlled release delivery system for indomethacin.
Pharm Dev Technol. 1998;3(1):1 6.
41. Yeole PG, Galgatte UC, Babla IB, Nakhat PD. Design and
evaluation of xanthan gum based sustained release matrix tablets
of diclofenac sodium. Indian J Pharm Sci 2006;68(2):185 9.
42. Korsmeyer RW, Gurny R, Doelker E, Buri P, Peppas NA.
Int J Pharm 1983;15:25 35.
43. Ritger PL, Peppas NA. A simple equation for the description of
solute release. I. Fickian and non Fickian release from non
swellable devices in the form of slabs, spheres, cylinders or discs.
J Control Release. 1987;5:23 36.
44. Moore JW, Flanner HH. Mathematical comparison of dissolution
proles. Pharm Technol. 1996;20:64 74.
45. Li SP, Feld KM, Kowarski CR. Preparation and evaluation of
Eudragit acrylic resin for controlled drug release of pseudoephe
drine hydrochloride. Drug Dev Ind Pharm. 1991;17:1655 83.
46. Valizadeh H, Nokhodchi A, Qarakhani N, Zakeri Milani P,
Azarmi S, Hassanzadeh D, et al. Physicochemical characteriza
tion of solid dispersions of indomethacin with PEG 6000, Myrj
52, lactose, sorbitol, dextrin, and Eudragit E100. Drug Dev Ind
Pharm. 2004;30(3):303 17.
47. Ammar HO, Khalil RM. Preparation and evaluation of sus
tained release solid dispersions of drugs with Eudragit polymers.
Drug Dev Ind Pharm. 1997;23(11):1043 54.
429
48. Kislalioglu MS, Khan MA, Blount C, Goettch RW, Bolton S.
Physical characterization and dissolution properties of ibuprofen:
Eudragit coprecipitates. J Pharm Sci 1991;80:799 804.
49. Fadda HM, Basit AW. Dissolution of pH responsive formula
tions in media resembling intestinal uids: bicarbonate vs.
phosphate buffers. J Drug Deliv Sci Technol 2005;15:273 9.
50. Sheu MT, Chou HL, Kao CC, Liu CH, Sokoloski TD.
Dissolution of diclofenac sodium from matrix tablets. Int J
Pharm. 1992;85:57 63.
51. Saha RN, Sajeev C, Sahoo J. A comparative study of controlled
release matrix tablets of diclofenac sodium, ciprooxacin hydro
chloride and theophylline. Drug Deliv. 2001;8(3):149 54.
52. Chan WA, Boswell CD, Zhang Z. Comparison of the release
proles of a water soluble drug carried by Eudragit coated
capsules in different in vitro dissolution liquids. Powder Technol.
2001;119:26 32.
53. Akiyama Y, Yoshioka M, Horibe H, Hirai S, Kitamori N,
Toguchi H. pH independent controlled release microspheres
using polyglycerol ester fatty acids. J Pharm Sci. 1994;83:1600 7.
54. Follonier N, Doelker E. Biopharmaceutical comparison of oral
multiple unit and single unit sustained release dosage forms.
STP Pharma Sci. 1992;2:141 55.
55. Hinton JM, Lennard Jonnes JE, Young AC. A new method for
studying gut transit times using radioopaque markers. Gut.
1969;10:842.
56. McConell EL, Basit AW, Murdan S. Measurement of rat and
mouse gastrointestinal pH, uid and lymphoid tissue, and
implications for in vivo experiments. J Pharm Pharmacol.
2008;60:63 70.

DOI: 10.1208/s12249 009 9225 2
Research Article
Comparison of the Halving of Tablets Prepared with Eccentric and Rotary
Tablet Presses
T. Sovny,1 P. Ksa Jr.,1 and K. Pintye-Hdi1,2
Received 17 December 2008; accepted 1 March 2009; published online 21 April 2009
Abstract. The aim of this study was to compare the densification of powder mixtures on eccentric and
rotary tablet presses and to establish relationships with the halving properties of the resulting scored
tablets. This is an important problem because the recent guidelines of EU require verification of the
equal masses of tablet halves. The models of Walker, Heckel, and Kawakita were used to describe the
powder densification on the two machines. The calculated parameters revealed that the shorter
compression cycle of rotary machines results in poorer densification and lower tablet hardness at a given
compression force. This is manifested in poorer halving properties, which are influenced mainly by the
hardness. Better densification improves the halving even at lower tablet hardness. This demonstrates that
these parameters can be good predictors of tablet halving properties.
KEY WORDS: direct compression; halving; Heckel analysis; Kawakita analysis; Walker analysis.
INTRODUCTION
Tablets are the most common dosage form in medicine.
For the individual therapy of patients, it is important to vary
the dose. In the pharmaceutical industry, this problem is often
solved through the production of scored tablets. A difficult
problem associated with the use of such tablets is to ensure
their breaking into equal halves. This is a multi factorial
problem, the solution of which must be verified in accordance
with the EU guidelines.
Numerous parameters can influence the structure of
tablets, and this in turn exerts effects on the breaking. The
properties of the compressed materials, the shape of the
punches, and the type of the tablet press, for instance, all
influence the halving properties. Powder densification is not
achieved in a uniform manner with the different tablet
presses: With eccentric machines, only the upper punch is
active during the compression, whereas with rotary presses,
both punches penetrate in the die, and the compression time
is shorter as compared with eccentric presses. For both types
of machines, the measurement of axial forces is relatively
easy through the application of strain gauges; however, the
determination of the displacement is easier with eccentric
machines because it is influenced only by the movement of
the upper punch. The productivity of eccentric machines is
much less than that of rotary machines, so they are not use in
industrial production. However, the deformation and densifi
cation properties of materials can be studied very well with
these presses, and it is therefore very important to establish
relationships between the densification performances of the

machines and to predict the behavior of materials on rotary
presses.
In the past century, many different models have been
proposed with which to describe the densification of materials
during compression. The first model was developed by
Walker (1), whose analysis is based on the reduction in
volume of the powder bed as a function of the logarithm of
the applied compression force.
The model devised by Heckel (2,3) is based on the
theory that there is a linear relationship between the behavior
of materials and the compression force during the compaction
and thus between the densification and the pressure. Despite
some critical observations (4), this model is probably the most
widely used in pharmaceutical technology (5 8), often in
comparison with other models (9 12), such as the Kawakita
equation, developed 10 years later (13,14). In this latter
model, the volume reduction at the applied pressure is
described as a function of the pressure. In most of the
publications, the data measured in the die are used in the
calculations. Busignies et al. (15) studied the compaction
behavior of granular lactoses, comparing the results of the
Heckel equation measured with in the die method with those
calculated from the properties of the ejected tablets, i.e., with
the out of the die method. In our study, we have also used
this latter model to compare the densification on eccentric
and rotary presses, for which Palmieri et al. (16) applied the
in the die method.
Department of Pharmaceutical Technology, University of Szeged,

Etvs str. 6, Szeged, 6720, Hungary.
2
To whom correspondence should be addressed. (e mail: klara.
hodi@pharm.u szeged.hu)
Binary mixtures of microcrystalline cellulose (Vivapur

102, J. Rettenmeier & Shne, Germany) and spray dried
mannitol (Pearlitol SD 200, Roquette Pharma, France),
430
Comparison of the Halving of Tablets Using Different Mathematical Models

lubricated by the addition of 1% of magnesium stearate (Ph.
Eur.), were compressed (Table I).
The powders were mixed with a Turbula mixer (Willy A.
Bachofen Maschienenfabrik, Switzerland; 8 min+2 min after
the addition of the magnesium stearate, 50 rpm).
The flow properties of the materials and mixtures were
determined with a PharmaTest PTG 1 powder rheological
tester (Pharma Test Appartebau, Germany).
The compaction behavior of the materials was tested with
an Engelsmann stampfvolumeter (JRS Pharma, Germany).
A Korsch EK0 (E. Korsch Maschienenfabrik, Germany)
eccentric and a Ronchi AM8S (Officine Meccanice F.lli
Ronchi, Italy) rotary tablet press mounted with strain gauges
and an eccentric press with a displacement transducer of the
upper punch were applied for tablet compression, with flat
single punches 8 mm in diameter and with a bisecting line.
The hardness of the resulting tablets was measured with
a Heberlein tablet hardness tester (Heberlein & Co. AG,
Switzerland).
For measurement of the force required to break the
tablets into halves, a laboratory constructed hardness tester
was utilized (Fig. 1), using three bend tablet hardness testing.
The tablet must be centered under the breaking item, which
moves vertically down. The load is detected with a computer
connected measuring cell, which is placed under the sample
holder.
The true densities of powders and tablets were deter
mined with a Quantachrome helium stereopycnometer
(Quantachrome GmbH., Germany).
RESULTS
The primary aim of this work was to study the
densification behavior of binary powder mixtures of two
widely used pharmaceutical excipients on eccentric and rotary
tablet presses. We used the models of Walker, Heckel, and
Kawakita to describe the powder densification and sought
relationships with the halving properties of the tablets. The
dry binder microcrystalline cellulose Vivapur 102 and the
filler material spray dried mannitol (Pearlitol SD 200) were
mixed in different ratios, as indicated in Table I. The results
of the preformulation tests for these materials and their
mixtures are presented in Table II. The anisometric particles
of Vivapur 102 (Fig. 2) exhibited poor flow properties during
powder rheological tests, and this resulted in lower bulk
density values. The rearrangement of the particles was
irregular but took place quickly in response to tapping, giving
rise to an exponential type rearrangement profile. The results
with Pearlitol indicated that this material can greatly improve
the poorer properties of Vivapur. It displays excellent flow
properties, and the isometric particles (Fig. 3) demonstrate
Table I. Compositions of Powder Mixtures
Samplea
Vivapur 102 (%)
Pearlitol SD 200 (%)
1
2
3
4
5
90
70
50
30
10
10
30
50
70
90
The mixtures were lubricated by the addition of 1% Mg stearate
431
Fig. 1. Schematic figure of laboratory constructed tablet hardness

tester
linear rearrangement behavior. The properties of the differ

ent powder mixtures varied between the end points of the
two component materials. An increasing quantity of Pearlitol
improved the compactibility of the powder, but the cohesive
ness decreased, which resulted in lower inter particulate
binding forces. This may be due to a lower number of contact
points between the isometric particles. These parameters
were calculated from Eq. 1, described by Kawakita et al. (17):
N=CN=a1=ab
where N is the number of taps applied for powder densifica

tion, while a and 1/b are constants referring to the compress
ibility and cohesiveness, respectively. C is the volume
reduction, which can be calculated via the following equation:
C V0
V =V0
where V0 is the initial volume of the powder bed, and V is the

current volume of the powder after a given number of taps.
Equation 1 is a modification of Eq. 3, which was
proposed by Kawakita and Ldde in 1971 to describe powder
densification behavior during compression:
P=C P=a 1=ab
where P is the applied pressure and C is calculated according

to Eq. 2, where V is now the volume of the powder bed at the
applied pressure.
We determined the Kawakita constants from the data on
samples compressed at 5, 10, or 15 kN, plotted according to
the out of the die method. The values of constant a decreased
on elevation of the Pearlitol quantity, as an indication of the
better rearrangement of the particles during compression,
which corresponds to the preformulation data (Table III).
The values of the constant a are very similar for the two
tablet machines, suggesting that it depends only on the
properties of the materials. In contrast, the values of 1/b,
which varied characteristically with the composition in the
preformulation tests, displayed a minimum at a mass ratio of
50:50. This means that, at this ratio, the lowest energy is
needed to reduce the volume of the powder to half of the
original. This may be due to the better utilization of the
432
Sovny, Ksa and Pintye-Hdi

Table II. Preformulation Results on Materials and Mixtures
Parameter
Vivapur 102
Pearlitol SD 200
Flow time (s)

Angle of repose (deg)
Bulk density (g/cm3)
Hausner factor
Carr index (%)
Compactibility
Cohesiveness
4.9
31.7
0.39
1.37
26.76
0.03
2.93
4.5
30.1
0.46
1.24
19.34
0.02
1.99
4.2
28.8
0.53
1.25
20.27
0.02
1.89
4.3
26.7
0.65
1.19
15.67
0.02
1.68
3.6
24.8
0.76
1.15
12.67
0.01
0.59
3.1
23.5
0.83
1.07
6.27
0.01
0.31
3.1
22.8
0.73
1.12
10.40
0.01
0.92
transmitted energy, which can be a result of the lower

adhesion and friction of the particles on the die wall. This
reduces the loss of the energy during the process, so a bigger
proportion of the demanded energy is assigned to the volume
reduction. Overall, the less energy investment is needed at
this ratio to given volume reduction. This effect is
corresponding with the results of other calculations discussed
below. Figure 4 reveals that the intercepts of the plots
obtained with the two machines are different: the calculated
1/b values are two to threefold higher for the rotary press.
This suggests that the shorter compression time and the
simultaneous action of the two punches result in a lower
volume reduction for rotary presses at a given compression
pressure. With the rotary press, a local minimum can be
observed in factor 1/b also at ratio 10:90. A possible
explanation of this phenomenon can be that the different
compression mechanism of the rotary press results to a better
rearrangement of particles at this ratio as on eccentric press.
Optimum points are also observed in the results calcu
lated from the Walker equations:
log P
LV C1
W log P C
pressure and the initial volume of the powder bed, and C and
C1 are constants. The coefficient L is the pressing modulus,
which can be calculated from Eq. 4. It likewise exhibits a
maximum at a ratio of 50:50 (Table III), reflecting the
smallest volume reduction at a given pressure at this ratio.
This also may be due to the better utilization of the
transmitted energy. As a result of the already low energy
investment, strong bonds are formed between the particles.
This is supported by the small value of the coefficient W at
this ratio (Table III), which shows the percentage volume
reduction when the pressure changes on a logarithmic scale.
Thus, the increase of the compression force causes only a
minor further volume reduction because almost the biggest
possible bonding forces have developed already at lower
ones. A difference can be seen in the slopes of the Walker
plots relating to the eccentric and rotary presses (Figs. 5
and 6).
We also used the equation developed by Heckel to
acquire more information concerning the powder densifica
tion (Fig. 7):
ln 1=1
D KP A
where P is the applied pressure, V is the relative volume,

calculated as V/V0, i.e., the ratio of the volume at the applied
where D is the relative density of the tablet (calculated as the

ratio of the apparent density of the tablet and the true density
of the powder) at pressure P, while K and A are constants.
The reciprocal of constant K is the mean yield pressure (Py).
Constant A gives the densification of the powder due to the
Fig. 2. SEM picture of Vivapur 102
Fig. 3. SEM picture of Pearlitol SD 200
100V
433
Table III. Parameters of the Different Equations Calculated from Linear Regression Analysis
Heckel
Sample
Eccentric press
1
2
3
4
5
Rotary press
1
2
3
4
5
Kawakita
Walker
Py
Da
Db
1/b
217.391
196.078
217.391
256.410
312.500
0.775
0.782
0.775
0.785
0.779
0.506
0.489
0.465
0.450
0.417
0.732
0.709
0.697
0.689
0.665
6.907
6.088
6.267
6.863
7.649
6.409
5.468
5.363
5.877
6.218
15.549
17.517
18.447
17.016
15.176
232.558
212.766
263.158
270.270
277.778
0.618
0.615
0.633
0.657
0.665
0.348
0.322
0.323
0.322
0.303
0.727
0.701
0.687
0.691
0.671
15.834
15.685
10.546
19.531
17.872
14.098
13.776
9.596
14.164
14.394
6.534
7.116
10.364
7.035
6.278
Fig. 4. Kawakita plot calculated with the out of the die method
Table IV. Tensile Strengths of the Tablets (MPa)

Compression pressure (MPa)
Sample
Eccentric press
1
2
3
4
5
Rotary press
1
2
3
4
5
100
200
300
3.626
2.384
1.952
1.168
1.231
5.702
3.759
2.670
2.020
2.056
6.038
4.642
3.385
2.861
2.919
1.280
0.941
0.591
0.605
0.537
3.576
2.508
1.748
1.511
1.803
3.807
3.269
2.456
2.275
2.595
434
Fig. 5. Walker plot based on Eq. 4
initial rearrangement of the powder bed (Da) according to the

following equation:
Da 1
eA
With the application of Da, the densification due to the

fragmentation of particles (Db) can be calculated:
Db Da
D0
where D0, related to the initial die filling, is defined as the

apparent density of the powder bed at zero pressure.
As compared with the other methods, Heckel analysis
does not give such unanimous results (Table III). The mean
yield pressure does not display a characteristic change; the
differences between the machines are clear. The larger values
of Da and Db demonstrate that a greater proportion of the
particle densification is due to the initial rearrangement and
particle fragmentation in the eccentric press. The differences
in initial rearrangement and fragmentation may result from
the differences in the method of die filling and the compres
sion cycle.
The parameters discussed above all influence the post

compressional properties of tablets. Table IV demonstrates
that the tensile strength is almost two times higher for the
tablets formed with the eccentric press. It can be seen that the
tensile strength decreases with increasing amount of Pearlitol,
the small adhesion and cohesion forces and the fewer contact
points between the particles greatly reducing the tablet
hardness. This parameter is probably the most important
influencing factor of the breaking. A strong relationship can
be observed between tensile strength of tablets and halving
properties (Table V). The results suggest that tensile strength
at about 3.00 MPa is required to the acceptable halving on
eccentric press, and even higher values are necessary on
rotary one.
DISCUSSION
The primary aim of this study was to establish relation
ships between powder densification on different tablet
presses and the halving properties of the resulting tablets.
The calculations with the equations suggest that the longer
Fig. 6. Walker plot based on Eq. 5
435
Fig. 7. Heckel plot based on the out of the die method
time of compression causes a two to threefold higher stress

on use of the eccentric machine at the same compression
force. This causes a greater proportion of particle fragmen
tation, but the bonding between particles becomes much
stronger, resulting to greater hardness.
The tablet hardness and the properties of the materials
are strongly related to the breaking properties of the tablets.
Postcompressional examination of the halving of the scored
tablets is very important because it is necessary to verify it in
the pharmaceutical file. The axial stress acting on the
bisecting line gives rise to elastic stresses in the tablets,
resulting in different degrees of deformation, depending on
the properties, deformability, and bonding of the tablets. In
this case, elastic deformation is advantageous. The change in
the deformation can be measured through the application of
strain gauges. When the elasticity predominates, the force
time curve rises with increasing axial stress and collapses at
the moment of breaking of the tablet (Fig. 8a). The tablets
break well when the masses of the two halves are closely
similar. We chose a 505% tolerance limit in this study. The
breaking result depends mainly on the hardness of the tablets
and was better at higher tensile strength. However, it is
additionally influenced by the internal structure of the tablets.
When the tablet hardness is low or internal structural defects

exist, the breaking surface crumbles or the tablets break into
more than one piece. The new breaking surfaces are
associated with extra stress that does not act on the bisecting
line. This type of breaking often proceeds through several
steps, which are seen in the curves as extra peaks (Fig. 8b).
When the structure of the tablet is free from defects, a lower
tensile strength may be sufficient for good breaking, as in the
case of sample 3. The results allow the conclusion that the
tablet hardness must be relatively high for good halving.
However, when the internal structure of the tablets is free
from defects, a lower hardness may be sufficient. Neverthe
less, the application of a higher compression force on a rotary
Table V. Amounts of Well halved Tablets

Compression pressure (MPa)
Sample
Eccentric press
1
2
3
4
5
Rotary press
1
2
3
4
5
100
200
300
50%
50%
60%
50%
30%
90%
100%
100%
80%
80%
100%
60%
80%
70%
90%
10%
0%
0%
0%
0%
30%
10%
10%
0%
0%
40%
40%
30%
0%
10%
Fig. 8. Breaking curves of a well halved (a) and a not well broken
tablet (b)
436
machine is needed to equal the better results on tablets

compressed on eccentric machines. No structural defects are
formed during compression when the powder is well com
pressed and undergoes low adhesion to the die wall, and the
bonding between the particles is formed quickly. The models
of Walker and Kawakita describe these processes better than
the Heckel model and furnish a better prediction of the
halving properties. The results obtained with the Heckel
equation should be utilized with caution.
CONCLUSIONS
Overall, it can be stated that the properties and the
densification behavior of the powders strongly influence the
halving properties of scored tablets. The relationships be
tween the densification behavior of powders and the applied
compression force can be characterized well with the use of
the equations of Heckel, Walker, and Kawakita. They also
can give useful information to the comparison of the behavior
of materials on different tablet presses. Comparing these
results, with the properties of compressed tablets, conclusions
can be drawn about the probable halving properties.
ACKNOWLEDGMENT
The work was supported by a Sanofi Aventis scholarship.
REFERENCES
1. Walker EE. The properties of powder. Part VI. The compress
ibility of powders. Trans Faraday Soc 1923;19:73 82.
2. Heckel RW. Density pressure relationships in powder compac
tion. Trans Metall Soc AIME 1961;221:671 5.
3. Heckel RW. An Analysis of powder compaction phenomena.

Trans Metall Soc AIME 1961;221:1001 8.
4. Sonnergaard JM. A critical evaluation of Heckel equation. Int J
Pharm 1999;193:63 71.
5. Ilkka J, Paronen P. Prediction of the compression behaviour of
powder mixtures by the Heckel equation. Int J Pharm 1993;94:
181 7.
6. Kuentz M, Leuenberger H. A new model for the hardness of a
compacted particle system, applied to tablets of pharmaceutical
polymers. Powder Techol 2000;111:145 53.
7. Kuny T, Leuenberger H. Compression behaviour of the enzyme
galactosidase and its mixture with microcrystalline cellulose.
Int J Pharm 2003;260:137 47.
8. Rees JE, Grant DJ. Influence of elastic deformation of particles
on Heckel analysis. Pharm Dev Technol 2001;6:193 7.
9. Patel S, Kaushal AM, Bansal AK. Effect of the particle size and
compression force on compaction behavior and derived mathe
matical parameters of compressibility. Pharm Res 2006;24:111 24.
10. Sonnegaard JM. Impact of particle density and initial volume on
mathematical compression models. Eur J Pharm Sci 2000;11:307
15.
11. Patel S, Kaushal AM, Bansal AK. Compaction behavior of roller
compacted ibuprofen. Eur J Pharm Biopharm 2008;69:743 9.
doi:10.1016/j.ejpb.2008.01.005
12. Denny PJ. Compaction equations: a comparison of the Heckel
and Kawakita equations. Powder Technol 2002;127:162 72.
13. Kawakita K, Ldde KH. Some considerations on powder
compaction equations. Powder Technol 1971;4:61 8.
14. Kawakita K, Hattori KI, Kishigami M. Charactheristic constants
in Kawakitas powder compression equations. J Powder Bulk
Solids Technol 1977;1:3 8.
15. Busignies V, Tchoreloff P, Leclerc B, Besnard M, Couarraze G.
Compaction of crystallographic forms of pharmaceutical granular
lactoses. I. Compressibility. Eur J Pharm Biopharm 2004;58:569 76.
16. Palmieri GF, Joiris E, Bonacucina G, Cespi M, Mercuri A.
Differences between eccentric and rotary tablet machines in the
evaluation of powder densification behaviour. Int J Pharm
2005;298:164 75.
17. Yamashiro M, Yuasa Y, Kawakita K. An experimental study on
the relationships between compressibility, fluidity and cohesion
of powder solids at small tapping numbers. Powder Technol
1983;34:225 31.

DOI: 10.1208/s12249 009 9224 3
Research Article
Development and Evaluation of Ethyl Cellulose-Based Transdermal Films
of Furosemide for Improved In Vitro Skin Permeation
Dhaval P. Patel,1,4 Chitral Mallikarjuna Setty,2 Gaurav N. Mistry,1 Santnu L. Patel,1 Tarun J. Patel,1
Pritesh C. Mistry,1 Amar K. Rana,1 Pritesh K. Patel,3 and Rishabh S. Mishra3
Abstract. Transdermal lms of the furosemide were developed employing ethyl cellulose and
hydroxypropyl methylcellulose as lm formers. The effect of binary mixture of polymers and penetration
enhancers on physicochemical parameters including thickness, moisture content, moisture uptake, drug
content, drug polymer interaction, and in vitro permeation was evaluated. In vitro permeation study was
conducted using human cadaver skin as penetration barrier in modied Keshary Chein diffusion cell. In
vitro skin permeation study showed that binary mixture, ethyl cellulose (EC)/hydroxypropyl methylcel
lulose (HPMC), at 8.5:1.5 ratio provided highest ux and also penetration enhancers further enhanced
the permeation of drug, while propylene glycol showing higher enhancing effect compared to dimethyl
sulfoxide and isopropyl myristate. Different kinetic models, used to interpret the release kinetics and
mechanism, indicated that release from all formulations followed apparent zero order kinetics and non
Fickian diffusion transport except formulation without HPMC which followed Fickian diffusion
transport. Stability studies conducted as per International Conference on Harmonization guidelines did
not show any degradation of drug. Based on the above observations, it can be reasonably concluded that
blend of EC HPMC polymers and propylene glycol are better suited for the development of transdermal
delivery system of furosemide.
KEY WORDS: chemical penetration enhancers; ethyl cellulose; furosemide; hydroxypropyl methylcellulose;
in vitro skin permeation; transdermal lms.
INTRODUCTION
Transdermal delivery of drugs provides many advantages
over conventional administration including enhanced efcacy,
increased safety, greater convenience, and improved patient
compliance. This can avoid the peak and valley effect of
oral or injectable therapy and can enable more effective
treatment by delivering drugs at a steady rate into blood
stream over an extended period of time. It also reduces the
dosage related side effects because the amount of drug
delivered into the biological system in a very controlled
manner and avoids rst pass metabolism (1,2). This route of
administration may be particularly signicance in infants and
children because of their greater surface area to weight ratio
(3). The system designs for transdermal patches include
matrix, microreservoir, reservoir, adhesive, and membrane
Formulation Development Department, Amneal Pharmaceutical

Company (I) Pvt ltd, R & D Centre, Vil. Rajoda, Tal; Bavla,
Ahmedabad, 382220 Gujarat, India.
2
Department of Pharmaceutics, N.E.T. Pharmacy College, Raichur,
584103 Karnataka, India.
3
Department of Pharmaceutics, Manoharbhai Patel Institute of
Pharmacy, Gondia, India.
4
To whom correspondence should be addressed. (e mail: dhavalpha@
gmail.com)
matrix hybrid. Matrix type transdermal patches remain the

most popular as they are easy to manufacture (4).
Furosemide (5 (aminosulfonyl) 4 chloro 2 [(2 furanyl
methyl) amino] benzoic acid) is potent diuretic agent that
induces a powerful diuresis, followed by the loss of sodium,
potassium, and chloride into urine, by acting on thick ascending
limb of the loop of henle (5). Its usual daily dose for adults is 20
80 mg, while for pediatric use ranges form 1 mg/kg up to a
maximum of 40 mg daily. It is commonly used in the treatment of
the cardiac and pulmonary disorders in premature infants and
neonates. The half life of furosemide is about 2 h and its
bioavailability has been reported to be about 60 70% (6).
Furosemide is administrated peroral or parenterally al
though its physicochemical and pharmacokinetic characteristics
like low molecular weight, lipid solubility, elimination half life,
and low melting point are in agreement with the ideal properties
of molecule for effective penetration of the stratum corneum
(7). The furosemide containing hydroxypropyl cellulose gel (8)
and ethylene vinyl acetate matrix (9) for transdermal delivery
purposes has been reported.
Ethyl cellulose (EC) is regarded as nontoxic, nonallergic,
and nonirritating material and has good lm forming prop
erties that enable it to form tougher lms. However, the water
permeability of pure EC is very low. The benets of EC have
been utilized by mixing with other hydrophilic polymers. In
the present study, we developed a suitable matrix lm for
furosemide by employing EC and hydroxypropyl methylcel
437
Patel et al.
438
lulose (HPMC) and to study the effect of polymers and
chemical penetration enhancers on the physicochemical
properties of transdermal lm and permeation of drug across
human cadaver skin.
until used. The composition of different lms is given in

Table I.
Thickness
The thickness of lm before and after the permeation
study was determined using micrometer gauge (Mitoyoto,
Japan). Film was measured at different places and mean
value was calculated.

Materials
Furosemide (mol. wt. 330.75, pKa 3.9, logP = 0.74;
Hemdeep organic Pvt. Ltd., Ankleshwar, India), ethyl
cellulose (45 cps; Colorcon Pvt Ltd., Goa, India), and hydroxy
propylmethyl cellulose (15 LV; Aurobindo Pharma Ltd.,
Hyderabad, India) were received as gift samples. Propylene
glycol, isopropyl myristate, and dimethyl sulfoxide (DMSO)
were purchased from S. D. Fine Chem. Ltd., Mumbai, India.
Di n butyl phthalate was purchased from Central Drug
House Pvt. Ltd., Mumbai. All other reagents were of
analytical grade.

The uniformity of drug distribution was determined by
taking known area of the lms at different places of the lm.
The lms were dissolved in 2 ml of casting solvent and
subsequently diluted with phosphate buffer pH 8.0. After
appropriate dilution, solutions were analyzed spectrophoto
metrically (UV Pharmaspec 1700, Shimadzu, Japan) for
furosemide at 229 nm using solution of lms prepared without
drug as reference to neglect the absorption of components of
the formulation if any.
Moisture Content
Methods
Preparation of Transdermal Film
Transdermal lms of furosemide (4.52 mg/cm2) containing
different ratio of EC and HPMC were prepared on mercury
surface. The required amount of drug and polymers were
dissolved in methanol dichloromethane (1:1) solvent system.
Di n butyl phthalate (15% w/w of polymer) was used as
plasticizer. Propylene glycol (PG), isopropyl myristate
(IPM), and DMSO were added to the polymer drug solution.
The resultant homogeneous solution was poured into a
glass ring placed on mercury substrate. The lms were
dried for a period of 24 h, and the rate of evaporation was
controlled by inverting funnel over the Petri dish. The dry
lms were wrapped in aluminum foil and kept in desiccators
The prepared lms were weighed individually and kept

in desiccators containing activated silica at room temperature
(30C) for 24 h till a constant weight was attained. The
percentage of moisture content was calculated as the differ
ence between initial and nal weight with respect to nal
weight (10).
Moisture Uptake
A weighed lm kept in desiccator at room temperature
(30C) for 24 h was taken out and exposed to 84% relative
humidity (RH) in a stability chamber (Lab Care, Mumbai, India)
until a constant weight of lm was obtained. The percentage
moisture uptake was calculated as the difference between nal
and initial weight with respect to initial weight (10).
Table I. Composition of Furosemide Transdermal Films

Penetration enhancer (10 % w/w of polymer)
Formulation
Drug (% w/w)
EC (% w/w)
HPMC (% w/w)
H
H1
H2
H3
H4
H5
H6
H7
H8
H9
H10
H11
H12
26.22
26.22
26.22
26.22
24.64
24.64
24.64
24.64
24.64
24.64
24.64
24.64
24.64
100
95
90
85
95
90
85
95
90
85
95
90
85
00
05
10
15
05
10
15
05
10
15
05
10
15
PG
DMSO
IPM
In all formulations, plasticizer at a level of 15% w/w of polymer is added

% w/w weight of drug to total weight of dry lm, EC ethyl cellulose, HPMC hydroxypropyl methylcellulose, PG propylene glycol, DMSO
dimethyl sulfoxide, IPM isopropyl myristate
Ethyl Cellulose-Based Transdermal Films of Furosemide

In Vitro Skin Permeation Study
439
RESULT AND DISCUSSION
Permeation studies were performed for different formu

lations across cadaver skin in modied Keshary Chein
diffusion cell at 32 0.5C. The diameter of the donor
compartment cell was 2 cm, providing 3.14 cm2 effective
constant area. The lms with area 2.83 cm2 were applied to
the skin using adhesive tape (cellophane) as backing layer.
The phosphate buffer pH 8.0 (70 ml) was used as receptor
compartment medium to ensure sink conditions and stability
of the drug (11). This whole assembly was kept on a magnetic
stirrer and the solution was stirred continuously using a
magnetic bead. The samples were withdrawn at different time
interval and replaced with equal volume of diffusion medium.
Samples were analyzed spectrophotometrically at 229 nm. To
ascertain whether the components of the skin or other
excipients of the lm interfere in the drug analysis, blank
experiment (lms without drug) was run using skin as barrier
membrane using phosphate buffer pH 8.0. When the solution
was analyzed at 229 nm for any interfering constituents, the
released constituents were amounting to an average of 0.06
0.02%.
Stability Studies
The stability studies were conducted according to
International Conference on Harmonization guidelines by
storing the TD lms at 402C/75% RH in stability chamber
(Lab Care, Mumbai, India) for 6 months. The samples were
withdrawn after 6 months and analyzed for drug content in a
UV spectrophotometer.
Statistical Analysis of Data

The results were analyzed by one way analysis of
variance with Tukey post t test using Graph Pad Prism
software 5 version (Graph Pad software Inc., San Diego, CA,
USA).
When all the dried lms were subjected to physical

examination, lms appeared to be translucent suggesting that
the drug was not completely solubilized rather dispersed/
suspended in the matrix. In preliminary studies, the EC lms
were found to be brittle and di n butyl phthalate was used as
plasticizer to reduce the brittleness of the lms. The studies
revealed that addition of di n butyl phthalate at 15% w/w of
polymer produces smooth, uniform, and exible lms. Hence,
further studies were carried out using plasticizer at 15%
w/w level in the lms.
Thickness
Thickness of lms varied between 0.083 and 0.117 mm
(Table II) suggesting that formulation variables used in the
study did not produce any signicant effect (p>0.05) on the
thickness of lms.
The drug content of all the formulations (Table II) was
4.43 mg/cm2 with a low standard deviation (0.07). The
results of drug content analysis have shown that the method
employed to prepare lms in this study was capable of giving
lms with uniform drug distribution with an insignicant
batch variability (p>0.001).
Moisture Content and Moisture Uptake
Moisture content and moisture uptake studies provide
information regarding stability of the formulation (4). The
results revealed that the moisture content and moisture
uptake were found to increase with increasing concentration
of hydrophilic polymer (HPMC). The presence of penetra
tions enhancers PG and IPM did not show any major changes
in moisture content and moisture uptake values. In case of
DMSO, slight increment in both parameters was observed
(Table II). This may be due to the water afnity of DMSO
Table II. Drug Content and Physicochemical Parameters of Films
Formulation
Drug content
(mg/cm2)
%Moisture uptake
SD, n 4
%Moisture content
SD, n 4
Thickness before
permeation study
(mm) SD, n 4
Thickness after
permeation study
(mm) SD, n 4
H
H1
H2
H3
H4
H5
H6
H7
H8
H9
H10
H11
H12
4.440.04
4.450.03
4.480.06
4.490.05
4.450.03
4.470.03
4.440.01
4.440.04
4.430.04
4.510.07
4.450.01
4.450.04
4.510.06
02.270.06
05.520.08
08.500.02
13.600.02
06.060.09
09.620.05
14.000.07
05.750.20
08.680.18
13.820.09
05.970.03
09.250.08
13.890.42
01.270.09
03.270.08
05.770.15
09.290.43
03.520.06
06.040.09
09.790.07
04.53016
06.800.12
11.330.07
03.770.05
06.290.09
10.050.03
0.0830.006
0.0970.012
0.1070.032
0.1030.021
0.1100.020
0.1030.015
0.1100.010
0.1070.021
0.1170.015
0.1030.012
0.1100.026
0.1100.010
0.1170.015
0.08450.002
0.0990.004
0.1050.001
0.1050.006
0.1120.004
0.1070.004
0.1120.003
0.1080.001
0.1200.002
0.1060.001
0.1120.004
0.1120.003
0.1200.001
SD standard deviation
Patel et al.
440
Table III. Permeation Parameters
Formulation
H
H1
H2
H3
H4
H5
H6
H7
H8
H9
H10
H11
H12
Drug permeated
(8 h) mcg/cm2 Flux (mcg/cm2/h)
SD, n 4
SD, n 4
63.922.20
90.221.52
209.64 1.04
263.942.86
204.611.44
283.081.35
428.871.72
125.391.89
231.852.34
299.892.18
171.152.72
239.601.64
403.113.56
5.781.25
8.382.11
22.051.17
28.853.14
19.421.30
27.191.72
45.092.86
12.832.45
23.741.89
30.081.83
17.181.24
23.951.68
36.913.50
Enhancement
factor (E)
1
1.44
3.81
4.99
3.35
4.70
7.80
2.21
4.10
5.20
2.97
4.14
6.38
Fig. 2. Effect of PG, IPM, and DMSO on the permeation of
furosemide from EC/HPMC (9.5:0.5) lm
(12). The small moisture content in the formulation helps

them to remain stable and from being a completely dried and
brittle lms (13), and low moisture uptake protects the
material from microbial contamination and bulkiness of the
lms (10). Thus, the results of physicochemical studies
conducted on different polymeric lms containing furosemide
favored the combination of these polymers for preparation of
transdermal lms.
An in vitro skin permeation study is predictive of in vivo

performance (4) and also valuable and necessary for studying
the rate and mechanisms of percutaneous absorption of drugs
(14). In the present study, in vitro permeation study was
carried out using human cadaver skin as penetration barrier.
The cumulative amount of drug permeated per centimeter
squared was plotted against time and the steady state
permeation ux was calculated from slope of linear portion

of the curve. The control (H) containing furosemide in
polymeric matrix of ethyl cellulose lm showed a ux of
5.781.25 mcg/cm2 h and release of 63.922.20 mcg/cm2 in
8 h (Table III). A few of workers have used penetration
enhancers in HPMC gels (8) and plasticizers in ethylene vinyl
acetate matrix (9) to improve the permeation of furosemide.
In our work, the concentration of HPMC, a hydrophilic
polymer, was used as a variable. It was observed that as the
concentration of HPMC was increased in the formulations
(H1, H2, and H3), the mean cumulative amounts of drug
permeated (90.22, 209.64, 263.94 mcg/cm2) and ux (8.38,
22.05, 30.08 mcg/cm2/h) increased substantially (Table III;
Fig. 1). The hydrophilic nature of HPMC seems to have
contributed toward increase in penetration of the solvent
molecules into the polymeric matrix and disturbed the
compactness of polymeric matrix resulting in faster release.
Fig. 1. Effect of HPMC concentrations on the permeation of

furosemide from EC lm

furosemide from EC/HPMC (9:1) lm
In Vitro Skin Permeation Study
Ethyl Cellulose-Based Transdermal Films of Furosemide
441
Table V. Stability Study Data of Films
Drug content(mg/cm2)
Formulation

furosemide from EC/HPMC (8.5:1.5) lm
H
H1
H2
H3
H4
H5
H6
H7
H8
H9
H10
H11
H12
Before (SD),
n 4
After
6 months
(SD), n 4
4.440.04
4.450.03
4.480.06
4.490.05
4.450.03
4.470.03
4.440.01
4.440.04
4.430.04
4.510.07
4.450.01
4.450.04
4.510.06
4.430.00
4.430.08
4.460.03
4.430.00
4.390.02
4.430.09
4.390.01
4.400.01
4.390.04
4.440.10
4.400.03
4.430.01
4.450.03
Further, to enhance the permeation of drug for achieving

the desired therapeutic plasma level, one of the approaches
makes use of penetration enhancers. In the present study,
formulations H1, H2, and H3 were selected to evaluate the
effect of enhancers like PG, IPM, and DMSO on in vitro
permeation of drug through the human cadaver skin. Figures
2, 3, and 4 show that addition of enhancers further increased
(p<0.05) the values of the amount of drug permeated and
ux compared to their counterpart without enhancers.
Among penetration enhancers used, PG has shown the
highest amounts of drug permeated and ux. The calculated
enhancement factor (15) for PG was found to be higher (3.35,
4.70, and 7.80) compared to IPM and DMSO at same
composition (Table III) and the relative order of the release
rates for the enhancers was PG > IPM > DMSO. The higher
enhancement capacity of PG may be attributed to its
cosolvent and enhancer effects. As cosolvent, it may solubi
lize the drug and as penetration enhancer, it may affect the
intracellular lipids (16) or protein (17) and thus increase the
partitioning of the drug in favor of stratum corneum.
Release Kinetics and Mechanism of Drug Release

Different kinetic models (zero order, rst order (18), and
Korsmeyer et al. expression (18,19)) were applied to interpret
the drug release kinetics and to know the mechanism of drug
release from these matrix systems with the help of Eqs. 1, 2,
and 3.
Mt M0 ka t
LnMt LnM0 k1 t
Mt =M0 kk t n :
In these equations, Mt is the cumulative amount of drug

released at any specied time point and M0 is the dose of the
Table IV. Permeation Kinetics Data from Different Transdermal Films
Formulation
H
H1
H2
H3
H4
H5
H6
H7
H8
H9
H10
H11
H12
Zero order kinetics
First order kinetics
Korsmeyer et al. model
R2 (SD, n 4)
R2 (SD, n 4)
R2 (SD, n 4)
n (SD, n 4)
0.99350.0040
0.97270.0152
0.99840.0009
0.99660.0007
0.97280.0138
0.99100.0051
0.97610.0168
0.98870.0082
0.98510.0095
0.97900.0157
0.98410.0127
0.98950.0075
0.99260.0005
0.490.01
0.510.01
0.640.01
0.660.03
0.540.02
0.560.04
0.650.05
0.600.04
0.660.02
0.660.03
0.590.01
0.610.06
0.670.01
0.92710.0127 000.015
0.94860.0188
0.96510.0125
0.95830.0315
0.95910.0224
0.95680.0158
0.98300.0090
0.96380.0182
0.97750.0160
0.96740.0126
0.96950.0230
0.96650.0112
0.98100.0097
0.93040.0124
0.95210.0161
0.97230.0202
0.96800.0153
0.96550.0189
0.96690.0175
0.98560.0083
0.96810.0134
0.98040.0093
0.97320.0156
0.97440.0119
0.97350.0111
0.98790.0080
Patel et al.
442
drug incorporated in the delivery system. k0, k1, and kk are
rate constants for zero order, rst order, and Korsmeyers
model, respectively, and n is the diffusional release exponent
indicative of the operating release mechanism. The correla
tion coefcient values (R2) are presented in Table IV.
The in vitro permeation proles of all formulation
(Figs. 1, 2, 3, and 4) obtained by plotting cumulative amount
of drug permeated against time shows a similar pattern of
drug permeation having initial faster (burst) release followed
by slower release. Hence, the in vitro permeation data neither
t into zero order (R2 =0.9271 to 0.9830) nor rst order (R2 =
0.9304 to 0.9879) kinetics completely. When the polymeric
layer is placed in contact with the skin, the drug compound
migrates through the polymer, partitions across the polymer/
skin interface, and then migrates into skin (20). The initial
faster release may be attributed to the rapid diffusion of the
drug immediate to the surface of the lm. Thus, rapid
depletion of the surface drug and consequent increase in
mean diffusional path length might have caused latter slower
release. In addition, latter slower release of the drug from the
formulations (except H) can also be accounted for the
increase in diffusion path length due to the swelling of
hydrated HPMC. The hypothesis was further conrmed by
the increase (p<0.01) in thickness of the lms after the
permeation study. This assumption was further conrmed by
tting the release data into Eq. 3. The formulation H showed
strong linearity with R2 value 0.9935 with an n value of
0.49. It indicates that diffusion is the mechanism of drug
release from formulation H. However, when HPMC loaded
formulations plotted with Eq. 3, irrespective of drug
concentration showed high linearity (R2 =0.9727 to 0.9984)
with a comparatively higher slope (n) values (p<0.01, except
H1) ranging from 0.51 to 0.66 (Table IV). It indicates that
drug release was leaning toward diffusion and swelling
coupled mechanism so called anomalous diffusion.
Presence of swellable polymer (HPMC) in the matrix might
be responsible for the drug release controlled by more than
one process.
Stability Study
Table V shows drug content of the formulations before
and after stability study. These formulations were stored at
402C/75% RH in stability chamber (Lab Care, Mumbai,
India) for 6 months. After 6 months, visual examination of
the dispersed did not show any changes in particle size. Drug
content of the patches after stability studies was 4.39 to
4.46 mg/cm2 and did not show any signicant variations.
These result indicates that drug remain stable after stability
studies.
CONCLUSION
The furosemide transdermal lms using EC/HPMC
polymer blend were prepared. Among the penetration
enhancers, propylene glycol, dimethyl sulfoxide, and isopro
pyl myristate used, the highest permeation rates were noticed
with propylene glycol. Incorporation of HPMC enhanced the
ux of the drug and also was responsible for the swelling
coupled diffusion controlled drug release.
ACKNOWLEDGMENT
Authors are thankful to Hemdeep Organic Pvt. Ltd,
Colorcon Pvt Ltd., and Aurobindo Pharma Ltd. for providing
furosemide, ethyl cellulose, and hydroxypropyl methyl cellu
lose, respectively, as gift samples.
REFERENCES
1. Cleary GW. Medical applications of controlled release. In:
Langer DL, editor. Transdermal controlled release systems.
Boca Raton FL: CRC; 1984. p. 203 51.
2. Kydonieus AF. Controlled release technologies. Boca Raton,
FL: CRC; 1987.
3. Sintov AC, Krymberk I, Gavrilov V, Gorodischer R. Transdermal
delivery of paracetamol for paediatric use: effect of vehicle
formulation on the percutaneous penetration. J Pharm Pharamcol
2003;55:911 9.
4. Mukherjee B, Mahapatra S, Gupta R, Patra B, Tiwari A, Arora
P. A comparison of povidone ethylcellilose and povidone
eidragit transdermal dexamethasone matrix patches based on in
vitro skin permeation. Eur J Pharm Biopharm 2005;59:475 83.
5. Giebisch G. The use of a diuretic agent as a probeto investigate
site and mechanism of ion transport process. Arzneim Forsch/
Drug Res 1985;35:336 42.
6. Micromedex Healthcare Series for Windows, Micromedex
Thompson Healthcare. 2001;vol. 109.
7. Barry BW. Is transdermal drug delivery research still important
today? Drug Discov Today 2001;6:967 71.
8. Agyralides GG, Dallas PP, Rekkas DM. Development and in
vitro evaluation of furosemide transdermal formulation using
experimental design techniques. Int J Pharm 2004;281:35 43.
9. Cho CW, Choi JS, Shin SC. Controlled release of furosemide
from ethylene vinyl acetate matrix. Int J Pharm 2005;299:127 33.
10. Arora P, Mukherjee B. Design, development, physicochemical,
and in vitro and in vivo evaluation of transdermal patches
containing diclofenac diethyl ammonium salt. J Pharm Sci 2002;
91:2076 89.
11. Bundgaard H, Norgaard T, Nielsen NM. Photodegradation and
hydrolysis of furosemide and furosemide esters in aqueous
solutions. Int J Pharm 1988;42:217 24.
12. Rowe RC, Sheskery PJ, Weller PJ. Hand book of pharmaceutical
excipient, Fourth Edition. London: Pharmaceutical; 2003.
13. Mutalik S, Udupa N. Glibenclamide transdermal patches:
physicochemical, pharmacodynamic, and pharmacokinetic eval
uations. J Pharm Sci 2004;93:1577 94.
14. Singh P, Roberts MS. Skin permeability and local tissue
concentration of nonsteriodal anti inammatory drugs after
topical application. J Pharmacol Expt Ther 1994;268:144 51.
15. Yu JW, Chien T, Chien YW. Transdermal dual controlled
delivery of testosterone and estradiol. I. Impact of system design.
Drug Dev Ind Pharm 1991;17:1883 904.
16. Bouwstra JA, De Vries MA, Gooris GS, Brass W, Brussee J,
Ponec M. Thermodynamic and structural aspects of the skin
barrier. J Control Rel 1991;15:209 20.
17. Goodman M, Barry BW. Action of penetration enhancers on
human stratum corneum as assessed by differential scanning
calorimetry. In: Bronaugh RL, Maibach H, editors. Percutaneous
absorption: mechanisms methodology drug delivery. New York:
Dekker; 1989. p. 567 93.
18. Khatun M, Islam SMA, Akter P, Quadir MA, Reza S. Md.
Controlled release of naproxen sodium from Eudragit RS 100
transdermal lm. Dhaka Univ J Pharm Sci 2004;3:1 2.
19. Kormeyer RW, Gurny R, Doelker E, Buri P, Peppas NA.
Int J Pharm 1983;15:25 35.
20. Kurnik RT, Potts RO. Modeling of diffusion and crystal
dissolution in controlled release system. J Control Rel 1997;45:
257 64.

DOI: 10.1208/s12249 009 9228 z
Research Article
Anhydride Prodrug of Ibuprofen and Acrylic Polymers
Boaz Mizrahi1 and Abraham J. Domb1,2
Abstract. The objective of this study was to synthesize anhydride prodrugs for carboxylic acid bearing
agents such as non steroidal anti inammatory drugs, shield the carboxylic acid group from irritative
effects, and obtain sustained release patterns. Ibuprofen was used as a representative drug for anhydride
derivatization. Conjugates of ibuprofen with carboxylic acid moieties of different acrylic polymers were
prepared by dehydration reaction using acetic anhydride. Products were characterized by infrared
spectroscopy, nuclear magnetic resonance, and scanning electron microscopy followed by preparation of
microspheres with different sizes from the conjugate Eudragit L 100 ibuprofen. The drug release was
monitored by high performance liquid chromatography. Ibuprofen was bound to the polymers via an
anhydride bond in high reaction yields (75 95%) with drug loading of up to 30% (w/w). These anhydride
derivatives hydrolyzed and release the drug at different periods ranging from 1 to 5 days, depending on
the hydrophobicity and the cross linking of the conjugates. The release of drug from the microspheres
was correlated to their size and ranged from 2 to almost 8 days. This study demonstrates the promise of
anhydride prodrug for extending drug action while shielding the carboxylic acid group.
KEY WORDS: anhydride; controlled release; ibuprofen; microspheres; prodrugs.
INTRODUCTION
A wide variety of compounds having carboxylic acid
groups are biologically active, for example, the non steroidal
anti inammatory drugs, such as ibuprofen, naproxen, indo
methacin, and diclofenac. The presence of free acid groups in
these compounds produces local irritation on interaction with
mucosal tissues, and at the same time, they ionize at
physiological pH, which makes the drug poorly absorbed
through biological membranes (1).
The general approach to overcome these limitations is
esterication of the carboxylic acid groups to produce non
irritating prodrug forms, provided that the parent bioactive
agent can be released from the prodrug at its sites of activity
(2). However, several aliphatic and aromatic esters of
carboxylic acid drugs are not sufciently labile in vivo to
ensure a sufciently high rate and extent of prodrug
conversion (3,4). As a result, plasma enzymes may not
hydrolyze there esters fast enough to obtain the necessary
conversion of the ester to the free acid (5). In another
consideration, ester derivative prodrugs have been found
cleavable enzymatically to release their bioactive forms and,
thus, are dependent on the enzymatic activity, which may vary
among individuals or even in the same individuals at various
times during the day or in various sites where the drug is
administered. This fact may result in a large variation of drug
1
Department of Medicinal Chemistry and Natural Products, School

of Pharmacy, Faculty of Medicine, The Hebrew University of
Jerusalem, Jerusalem, 91120, Israel.
2
To whom correspondence should be addressed. (e mail: avid@ekmd.
huji.ac.il)
bioavailability (6,7). To overcome the aforementioned limi

tations, anhydride derivatives of carboxylic acid bearing drugs
are suggested.
Unlike the ester bond used in prodrugs, the anhydride
bond is more susceptible to hydrolysis to its carboxylic acid
counterparts in a predictable rate and pattern and is less
sensitive to enzymolysis than the esters or amides (8). Studies
on polyanhydrides as drug carriers show that they degrade in
a controlled manner and are biocompatible with the human
body tissues (9). Interestingly, not much attention has been
given to the formation of prodrugs based on the hydrolyti
cally degradable anhydride bonds, although biodegradable
polymers based on these bonds have been developed and
used as erodible carriers for drugs both in animals and in
humans (10).
Our previous study (11) focused on the synthesis of
mixed anhydrides of ibuprofen with various fatty acids for
improved bioavailability. It was found that anhydride conju
gated drugs having longer alkyl chain provide higher absorp
tion through lipid water membrane. In addition, this
anhydride prodrug approach can be applied for manipulating
the release rate of a free drug by using different types of fatty
acids.
Polyacrylic acid (PAA) based polymers are largely used
in various technological applications due to their viscoelastic
properties, bioadhesiveness, safety, and easy administration.
They can be found in drug delivery systems (12), mucoadhe
sive forms (13), synthesis of nanoparticles by the sol gel
method (14), and in ophthalmic and vaginal preparations
(15), to name a few of their diverse applications. Eudragit L
100 and Eudragit S 100 are anionic copolymers based on
methacrylic acid and methyl methacrylate. The ratio of the
453
454
Mizhari and Domb
free carboxyl groups to the ester groups is approximately 1:1

in Eudragit L 100 and approximately 1:2 in Eudragit S
100. Their average molecular weight is approximately
135,000. Eudragit acrylic resins are widely used in enteric
coated and controlled release formulations due to their pH
dependent solubility (16,17).
The objective of this work was to design a prodrug that
relies on anhydrides of ibuprofen with carboxylic acid bear
ing polymers. These prodrugs are predicted to be less
irritating to stomach or intestinal mucosal membranes and
can be used for various topical dosage forms.
(10 mL) for 6 h followed by evaporation of the acetic

anhydride using an oil pump (0.3 mmHg at 70C). The dry
residue was washed several times with dry diethyl ether to
remove access of ibuprofen symmetric anhydride. Then, the
powder was left to dry in a desiccator for 3 h. Cross linked
PAA based anhydride was received as off white powder.
Preparation of Ibuprofen Anhydride
Ibuprofen (10 g) was reuxed in acetic anhydride
(100 mL) for 3 h followed by the evaporation of acetic
anhydride to dryness, and the residue was distilled under
vacuum (0.3 mmHg at 70C) to yield (70%) a viscous liquid
of symmetric ibuprofen anhydride.
Materials
Characterization
Ibuprofen was a gift from Ethyl Corp. (NJ, USA). Acetic
anhydride was purchased from Aldrich (Milwaukee, WI,
USA). PAA with Mw of 2,000 and 5,000 Da and phenol
phthalein solution and polyvinyl alcohol (PVA, 30 70 kDa)
were purchased from Sigma Aldrich (Rehovot, Israel).
Eudragit S 100 and L 100 were purchased from Rohm
Pharma (GmbH, Germany). Cross linked polyacrylic acid
(Carbopol 934) was obtained from Goodrich Co., Ltd.
(Cleveland, OH, USA). All solvents were analytical grade
from Biolab (Jerusalem, Israel) and were used as received.
Synthesis of Mixed Anhydrides
Linear Polymers
Polymer powder (1 g) and ibuprofen (0.25 g) were
reuxed in acetic anhydride (10 mL) under dried nitrogen
atmosphere for 3 h followed by evaporation of the acetic
anhydride using an oil pump (0.3 mmHg at 70C). The
products were dissolved in dichloromethane and precipitated
from dry diethyl ether. The nal product was received after
drying in a desiccator for 6 h.
Cross linked Polymer
Cross linked PAA (1 g) and ibuprofen (0.25 g) were
reuxed in acetic anhydride (10 mL) and dichloromethane
Quantitative analysis of carboxylic acid groups was

determined by dissolving 0.25 g of each of the polymers in
sufcient amount of water (Eudragit S 100 and L 100 were
dissolved in 3:1 DDW/ethanol solution) followed by titration
with 0.1 volumetric solution of sodium hydroxide using
phenolphthalein solution as indicator. After a blank titration
was performed to each of the solvents, the amount of free
carboxylic acids was calculated.
Infrared (IR) spectroscopy (Perkin Elmer System 2000
FT IR) was performed on polymers and conjugates. Con
jugates were measured on NaCl plates using dichloromethane
as solvent. Each sample was recorded between 4000 and
500 cm 1 by 16 scans. 1H NMR spectra were recorded on a
Varian 300 MHz instrument using DMSO D6 or CDCl3 as
solvents. Values were recorded as parts per million relative to
internal standard (TMS). Drug loading was detected after
hydrolyses in 1 N NaOH for 24 h. The free drug content was
detected by high performance liquid chromatography
(HPLC) at 270 nm.
Ibuprofen Analysis
Quantitative analysis of ibuprofen was performed by
HPLC using a C18 Lichospher 100 column (Merck, Dorstadt,
Germany, 2504 mm, 5 m) using a 70:30 (v/v) mixture of
acetonitrile/water and 0.2% phosphoric acid as mobile phase
Fig. 1. a Schematic illustration of the methods for the synthesis of anhydrides conjugate
455
Table I. Anhydrides Parameters

Polymer in conjugate
Polymer weight (g)
Ibuprofen (g)
Yield (%)
Drug loadinga (% w/w)
Ibuprofen to carboxylic acidb (%)
PAA 2000
PAA 2000
PAA 2000
PAA 2000
PAA 5000
Eudragit L 100
Eudragit S 100
Cross linked PAA
1
1
1
1
1
1
1
1
0.05
0.25
1
4
0.25
0.25
0.25
0.25
82
73
34
11
77
95
94
88
1.1
7.8
14.1
30.8
6.2
4.2
3.6
7.3
0.5
4.3
8.3
22.7
5
4.1
5.8
4.2
a
b
Calculated after hydrolysis in 1 N NaOH followed by drug detection by HPLC

Calculated after quantitative analysis of carboxylic acid groups of each of the polymers followed by titration with 0.1 volumetric solution of
sodium hydroxide using phenolphthalein solution as indicator
at a ow rate of 1 mL/min. Ibuprofen was detected by UV at

270 nm, and its typical retention time was 4.9 min.
In Vitro Release of Ibuprofen
Thirty milligrams of each conjugate was suspended in a
vial containing 30 mL of 0.1 M, pH 7.4, phosphate buffer. The
vials were incubated at 37C with constant shaking at 75 rpm.
At time intervals, each vial was centrifuged at 3,000 rpm for
5 min and samples (100 L) from the releasing medium were
withdrawn for HPLC analysis. The withdrawn volume was
replaced with fresh buffer followed by resuspending before
continuing incubation. The results are a mean of three
experiments.
Preparation of Microspheres
Ibuprofen loaded microspheres were produced from
Eudragit L 100 anhydride conjugate loaded with 4.2% (w/w)
ibuprofen. One hundred milligrams of the conjugate was
dissolved in 1 mL dichloromethane. The solution was
dispersed in 300 mL of 1% (w/v) aqueous solution of PVA
(pH 2). The resulting emulsion was stirred for 3 h at 750 rpm
by a mechanical stirrer. Microspheres were then collected by
centrifugation and dried at room temperature under vacuum
until required for use.
In a second assay, conjugate of the same batch was
dissolved in dichloromethane (100 mg in 1 mL). This was
then emulsied in an aqueous phase (300 mL, pH 2) of a 1%
(w/v) solution of PVA using a homogenizer Silverson L4R
(Silverson Machines, Ltd., Bucks, UK) at 8,000 rpm for
30 min. Emulsion was left to mix by a magnetic stirrer in a
hood for additional 3 h before centrifugation, collecting the
microspheres and drying at room temperature under vacuum.
Ibuprofen content was over 90%.
Statistical comparisons of the ndings were made by one
way analysis of variance. Comparison of means was per
formed by the least signicant difference test. Data analysis
was performed using a statistical software package (Instat;
GraphPad Software, San Diego, CA, USA). The signicance
level was set at p<0.05.
Synthesis of Mixed Anhydrides
The common way of producing anhydride bonds is by
activating the carboxylic acid with acetic anhydride [18].
Mixed anhydrides of ibuprofen and carboxylic acid moieties
of acrylic polymers were synthesized from the corresponding
polymers in acetic anhydride for 3 h (Fig. 1). All mixed
anhydrides were solid at room temperature. The anhydride
parameters are summarized in Table I.
The production yields of the anhydride conjugate were
between 73% and 95% for the 1:0.25 polymers to drug feed
Characterization
The morphology of microspheres was analyzed by
scanning electron microscopy (SEM, Quanta 2000). Before
the examination, the surfaces were coated with a thin layer of
gold so as to improve the conductivity and prevent charging.
Microsphere diameter and distribution size were calcu
lated using Image Analyzer software (Digimizer MedCalc
Software, Mariakerke, Belgium).
The release pattern of each of the microspheres was
analyzed as described in In Vitro Release of Ibuprofen.
Fig. 2. Infrared spectrum (NaCl method) of a ibuprofen; b poly

acrylic acid (Mw 2,000 Da); and c anhydride based conjugate from
ibuprofen and polyacrylic acid (Mw 2,000 Da)
456
Mizhari and Domb
Fig. 3. 1H NMR spectra (in CDCl3) of ibuprofen
ratio. As can be seen in Table I, increasing the ratio of drug to

polymer considerably decreases the yield and increases the
drug loading and the drug to carboxylic acid ratio. The rank
order of drug loading was PAA > cross linked PAA >
Eudragit L 100 > Eudragit S 100, corresponding to the
amount of carboxylic acids in the polymer, its solubility, and

the Mw of the polymers.
The percentage of drug loading and the ibuprofen to
carboxylic acid ratio increases as drug concentration increases
in the reaction. For the PAA with Mw of 2,000, ibuprofen
Fig. 4. 1H NMR spectra (in DMSO D6) of anhydride based conjugate from ibuprofen and polyacrylic acid 2,000 Da
Fig. 5. In vitro release of ibuprofen from anhydride conjugates of

ibuprofen and polymers as measured by high performance liquid
chromatography. The polymers were: polyacrylic acid 2,000 Da
(diamond), polyacrylic acid 5,000 Da (open circle), cross linked
polyacrylic acid (square), Eudragit L 100 (triangle), Eudragit S
100 (ex), polyacrylic acid 2,000 Da at pH 2 (filled circle). Experiments
were performed in phosphate buffer (pH 7.4) at 37C
loading score increased rapidly at concentrations below 0.25

to 1 g polymer and less at higher concentrations.
Characterization
Figure 2 shows the IR spectra of ibuprofen, PAA 2000,
and their mixed anhydride containing 7.8% ibuprofen. It can
be seen that while the IR spectra of the drug and the polymer
show carbonyl absorbance at 1,710 cm 1 , the mixed
anhydride demonstrate anhydride absorbance at 1,800 and
1,780 cm 1. Additionally, the stronger carboxylic acid
absorbance of the parent polymer at 3,500 2,500 cm 1
comparing to the weak peak of the conjugate conrms
anhydride bond formation. Similar results were also
reported by Jiang et al. (19).
The 1H NMR spectra of ibuprofen and conjugate from
ibuprofen and PAA 2000 are shown in Figs. 3 and 4,
respectively. While the spectrum of the conjugate correlates
with the NMR data of the free ibuprofen, its carboxylic acids
hydrogen peak at 11.6 ppm was not observed, suggesting the
formation of anhydride.
When ibuprofen was exclusively reuxed in acetic
anhydride (10 mL) followed by the evaporation of the acetic
anhydride, a yield of 75% was obtained. This symmetric
457
Fig. 7. In vitro release from anhydride conjugate from Eudragit L

100 loaded with 4.2% (w/w) ibuprofen as measured by high
performance liquid chromatography. Microspheres of 5 m in
diameter (triangle), conjugate as received (diamond), microspheres
of 100 m in diameter (ex). Experiments were performed in
phosphate buffer (pH 7.4) at 37C
ibuprofen anhydride product was received as a viscous liquid

soluble in ether. Similar to the spectra of the conjugate, its
NMR spectra (not shown) revealed the disappearance of the
carboxylic acids hydrogen at 11.6 ppm. This symmetric
ibuprofen anhydride had the following parameters: IR,
(cm 1) 1,812, 1,748; 1H NMR (300 MHz, CDCl3), 7.14 7.01
(m, 8H), 3.71 3.62 (m, 2H), 2.45 (d, J=7.2 Hz, 4H), 1.92 1.79
(m, 2H), 1.43 (dd, J=4.2, 7.2 Hz, 4H), and 0.91 (d, 6.6 Hz,
12H).
Release of Ibuprofen
The release properties of ibuprofen from these mixed
anhydrides were examined in vitro in a 0.1 M, pH 7.4,
phosphate buffer at 37C. As can be seen in Fig. 5, the release
rate of ibuprofen was a function of the polymer chain length,
hydrophilicity, and cross linking. Typically, 80% of the drug
was constantly released for 14 h from the ibuprofen PAA
2000 derivative, whereas only 35% from the Eudragits L 100
and the cross linked PAA during the same time period. PAA
2000 and 5000 showed similar pattern of release with a total
release of the ibuprofen within 30 and 38 h, respectively. In
the same manner, the Eudragit copolymers showed compa
rable patterns with release of almost 90% of the drug content
within 5 days.
Fig. 6. Scanning electron microscopy conjugate anhydride. a Eudragit L 100 loaded with 4.2% (w/w) of ibuprofen.
Microspheres prepared from the same conjugate as in a with a diameter of b 5 and c 100 m
458
A comparison study of these anhydrides in buffer pH 2
was also performed; however, no ibuprofen could be moni
tored (representative PAA 2000 mix anhydride in pH 2 is
shown in Fig. 5). This can be explained by the slower
anhydride hydrolysis (20,21) and the poor solubility of
ibuprofen in acidic solutions (11).
Mizhari and Domb

ACKNOWLEDGMENTS
We thank Dr. Alfonso Bentolila and Dr. Christopher
Levins for the help in NMR data analysis.
REFERENCES
Preparation of Microspheres, Characterization, and Release
Ibuprofen loaded microspheres were produced using
conjugate of Eudragit L 100 loaded with 4.2% (w/w) ibupro
fen. The technique involved oil in water emulsion followed
by the solvent evaporation method (22). Particle size
decreased from 90.11 to 50.3 m with the increase of
stirring speed from 750 to 8,000 rpm, respectively.
SEM photomicrographs of the conjugate and its micro
spheres are depicted in Fig. 6a c. Figure 6a represents the
conjugate before the formation of microspheres. It can be
seen that the particles present nodular cluster morphology
with varying density networks. On the other hand, micro
spheres made from the same conjugate are spherical and
smooth, as seen in Fig. 6b, c.
Figure 7 shows the drug release proles from the
conjugate and the microspheres. The drug in 5 m micro
spheres was rapidly released within 2 days. This was likely
due to the increased surface to volume ratio of the smaller
particles. Non microspheric conjugates moderately sustained
the drug release to 5 days, whereas 100 m particles could
prolong to almost 8 days.
The results obtained suggest that anhydride based pro
drugs may overcome some of the problems of carboxylic acid
bearing drugs mentioned in INTRODUCTION. First,
protecting the carboxylic function can be temporarily masked
by anhydride bond to overcome stomach and mucosal
irritation. Secondly, this approach can be applied for extend
ed action/release of acidic drugs after administration to the
eye, skin, vagina, or oral intake where these polymers are
already widely used. Thus, using these polymers is ideal due
to their contribution to various systemic and topical delivery
systems.
CONCLUSIONS
This work provides a simple synthetic method for
producing anhydride based conjugates of carboxylic acid
bearing drugs and carboxylic acid bearing polymers. In
this approach, the release rate of a free drug can be
manipulated using different types of polymer and/or by
controlling the particle size. For systemic or localized
delivery via oral administration, a sustained release of
drug can be obtained with the only degradation by
product being commonly used biocompatible biopolymer.
Nevertheless, chemical and physical stability in biological
relevant media as well as clinical studies are required in
order to assess the healing potential of these conjugates.
Further in vitro and in vivo evaluations are required to
investigate the therapeutic potential of the above de
scribed anhydride based drug conjugates.
1. Bruesewitz C, Funke A, Kuhland U, Wagner T, Lipp R.

Comparison of permeation enhancing strategies for an oral
factor Xa inhibitor using the Caco 2 cell monolayer model. Eur J
Pharm Biopharm. 2006;64(2):229 37.
2. Bundgaard H. Design of prodrugs. Amsterdam: Elsevier; 1985.
3. Harrison IT, Lewis B, Nelson P, Rooks W, Roszkows A, Tomoloni
A, et al. Nonsteroidal antiinammatory agents. 1. 6 Substituted 2
naphthylacetic acids. J Med Chem. 1970;13(2):203 5.
4. Child RG, Osterberg AC, Sloboda AE, Tomcufcik AS. Fenbu
fen, a new anti inammatory analgesic synthesis and structure
activity relationships of analogs. J Pharm Sci. 1977;66(4):466 76.
5. Hansen LB, Christrup LL, Bundgaard H. Ketobemidone pro
drugs for buccal delivery prediction of the extent of saliva
catalyzed hydrolysis of various ester prodrugs under simulated in
vivo conditions. Int J Pharm. 1992;88(1 3):229 35.
6. Goetz RM, Holtz J. Enhanced angiotensin converting enzyme
activity and impaired endothelium dependent vasodilation in
aortae from hypertensive rats: evidence for a causal link. Clin
Sci. 1999;97(2):165 74.
7. Higashikawa F, Murakami T, Kaneda T, Kato A, Takano M. Dose
dependent intestinal and hepatic rst pass metabolism of midazo
lam, a cytochrome P450 3A substrate with differently modulated
enzyme activity in rats. J Pharm Pharmacol. 1999;51(1):67 72.
8. Kumar N, Edlund U, Teomim D, Rasiel A, Domb AJ.
Biopolymers: polyesters (Part B). Germany: Wiley VCH; 2002.
9. Domb AJ, Israel ZH, Elmalak O, Teomim D, Bentolila A.
Preparation and characterization of carmustine loaded polyanhy
dride wafers for treating brain tumors. Pharm Res. 1999;16(5):762 5.
10. Nair LS, Laurencin CT. Biodegradable polymers as biomaterials.
Prog Polym Sci. 2007;32(8 9):762 98.
11. Shaaya O, Magora A, Sheskin T, Kumar N, Domb AJ.
Anhydride prodrugs for nonsteroidal anti inammatory drugs.
Pharm Res. 2003;20(2):205 11.
12. Mizrahi B, Domb AJ. Mucoadhesive tablet releasing iodine for
treating oral infections. J Pharm Sci. 2007;96(11):3144 50.
13. Mizrahi B, Golenser J, Wolnerman JS, Domb AJ. Adhesive
tablet effective for treating canker sores in humans. J Pharm Sci.
2004;93(12):2927 35.
14. Michalski S. Cleaning, retouching and coatings. Preprints to the
contributions to the Brussels congress. London; 1990.
15. Weyenberg W, Bozdag S, Foreman P, Remon JP, Ludwig A.
Characterization and in vivo evaluation of ocular minitablets
prepared with different bioadhesive Carbopol starch compo
nents. Eur J Pharm Biopharm. 2006;62(2):202 9.
16. Bonacucina G, Palmieri GF. Acrylic polymers as thickening
agents for tetraglycol cosolvent. J Pharm Sci. 2006;95(4):726 36.
17. Wade A, Weller P. Handbook of pharmaceutical excipients.
London: The Pharmaceutical Press; 1994.
18. Fessenden RJ, Fessenden JS. Organic chemistry. 6th ed. Pacic
Grove, CA: Brooks/Cole; 1998.
19. Jiang HL, Zhu KJ. Synthesis, characterization and in vitro degrada
tion of a new family of alternate poly(ester anhydrides) based on
aliphatic and aromatic diacids. Biomaterials 2001;22(3):211 18.
20. Ege SN. Organic chemistry structure and reactivity. Lexington,
MA: Heath & Co.; 1994.
21. Liu WP, Lee HK. Degradation study of unsaturated anhydrides
by reversed phase high performance liquid chromatography. J
Chromatogr A. 1998;805(1 2):109 18.
22. Andersson N, Kronberg B, Corkery R, Alberius P. Combined
emulsion and solvent evaporation (ESE) synthesis route to well
ordered mesoporous materials. Langmuir 2007;23(3):1459 64.

DOI: 10.1208/s12249 009 9226 1
Research Article
Formulation and Performance Characterization of Radio-Sterilized
Progestin-Only Microparticles Intended for Contraception
Shivanand Puthli1 and Pradeep Vavia1,2
Received 21 June 2008; accepted 1 March 2009; published online 21 April 2009
Abstract. The aim of this study was to formulate and characterize a microparticulate system of progestin
only contraceptive. Another objective was to evaluate the effect of gamma radio sterilization on in vitro
and in vivo drug release characteristics. Levonorgestrel (LNG) microspheres were fabricated using poly
(lactide co glycolide) (PLGA) by a novel solvent evaporation technique. The formulation was optimized
for drug/polymer ratio, emulsier concentration, and process variables like speed of agitation and
evaporation method. The drug to polymer ratio of 1:5 gave the optimum encapsulation efciency. Speed
of agitation inuenced the spherical shape of the microparticles, lower speeds yielding less spherical
particles. The speed did not have a signicant inuence on the drug payloads. A combination of
stabilizers viz. methyl cellulose and poly vinyl alcohol with in water solvent evaporation technique
yielded microparticles without any free drug crystals on the surface. This aspect signicantly eliminated
the in vitro dissolution burst effect. The residual solvent content was well within the regulatory limits.
The microparticles passed the test for sterility and absence of pyrogens. In vitro dissolution conducted on
the product before and after gamma radiation sterilization at 2.5 Mrad indicated no signicant difference
in the drug release patterns. The drug release followed zero order kinetics in both static and agitation
conditions of dissolution testing. The in vivo studies conducted in rabbits exhibited LNG release up to
1 month duration with drug levels maintained within the effective therapeutic window.
KEY WORDS: contraceptive; gamma radiation; in vivo; levonorgestrel; poly(lactide co glycolide).
INTRODUCTION
Poly(lactide co glycolide) (PLGA) is the polymer of
choice for parenteral systems (1 3). Depending upon the
type and ratio of lactide and glycolide, the biodegradation
behavior can be altered. Injectable contraceptives offer
several advantages. This administration is highly effective
with reliable reversibility. The miniaturized systems that
employ biodegradable polymers have attracted the attention
of many researchers since these systems bypass the surgical
complications associated with the implantable devices (4 8).
In comparison to the oral formulation like tablet, the
parenteral system would require lesser amounts of the drug.
Since the injectable system would require a lower dose, this
would in turn lead to fewer unwanted side effects associated
with the molecule. Further, a system that would release the
active agent for prolonged periods would be more patient
adherent (9). One can avoid the problems associated with
missing a dose during the ongoing therapy.
In this study, we employed levonorgestrel (LNG) as the
model progestin drug. LNG is 13 ethyl 17 ethynyl 17
1
Department of Pharmaceutical Sciences, University Institute of

Chemical Technology, University of Mumbai, Nathalal Parikh Marg,
Matunga, Mumbai, 400 019 Maharashtra, India.
2
To whom correspondence should be addressed. (e mail: drugdel@
rediffmail.com)
hydroxy 1,2,6,7,8,9,10,11,12,13,14,15,16,17 tetradecahydrocy

clopenta[a] phenanthren 3 one, a levorotatory enantiomer
derived from 19 nortestosterone. It is the drug of choice by
most medical practitioners for contraception. Microparticles
of LNG using poly(lactide co glycolide) polymer have been
reported (10). LNG microspheres that released the drug
continuously for more than 8 months in female baboons have
been reported. However, no further progress has been
reported on this system (11). However, these studies do not
account the sterilization step and the consequence of post
sterilization effects on the performance of the product.
The primary objective of our study was to develop a
progestin only contraceptive preparation for monthly release
of LNG. The microparticles were prepared using a solvent
evaporation method. Sterility is an important parameter for
parenteral systems. It is imperative to examine the effects of
radiation sterilization on the stability of the system from its
developmental stages. The effect of gamma radiation terminal
sterilization on the in vitro and in vivo release of the steroid
was studied. This signicant study was undertaken to critically
investigate the suitability of the radiation technique for
sterilization of this developed system. The formulation was
optimized for different formulation factors and process
variables. The product was characterized for in vitro drug
release kinetics. The proof of concept of the developed
parenteral system was successfully demonstrated by in vivo
studies in rabbits.
443
444
Materials
LNG was a gift sample from Wyeth Laboratories (Mum
bai, India). Methyl cellulose (Methocel 400, MC) was
obtained from Colorcon Asia Pvt. Ltd. (Mumbai, India).
Polyvinyl alcohol (PVA, partially hydrolyzed 88%), sodium
azide, thiomerosal, potassium dihydrogen orthophosphate,
sodium dihydrogen phosphate (anhydrous), disodium
hydrogen phosphate (anhydrous), sodium dodecyl sulfate,
and sodium chloride were purchased from S.D. Fine
Chemicals (Mumbai, India). Poly(lactide co glycolide) 53/47
(PLGA; Mw 14,000 Da; inherent viscosity 0.2 dL/g) was a
gifted from Purac Biochem (Holland). A regular hypodermic
disposable syringe was purchased from Becton Dickinson and
Company, USA. The uid thyoglycollate medium was
purchased from Hi Media Pvt. Ltd. (Mumbai, India). Sodium
carboxymethylcellulose (USP grade; medium viscosity
polymer, 2% solution at 25C has a viscosity of 400 800 cps)
was purchased from Sigma Chemicals (USA). Bacillus subtilis
ATCC No. 6633 (NCIM 2063) and Candida albicans ATCC
No. 10231 (NCIM 3471) were obtained form National
Chemical Laboratory (Pune, India). Bacteroides vulgatus
ATCC No. 8482 (MTCC 1350) was procured from Institute
of Microbial Technology (India). High performance thin layer
chromatography (HPTLC) precoated silica gel 60 F254 plates
were purchased from Merck (Germany). Potassium bromide
(KBr) was purchased from Ranbaxy Chemicals (Mumbai,
India). Karl Fisher (KF) reagent was purchased from Merck
(Germany). Female albino Belgium rabbits were obtained
from the Institute for Research in Reproduction, Mumbai,
India. Radioimmunoassay (RIA) kit was a gift from the World
Health Organization, London. All other reagents used were of
analytical grade and were purchased from Ranbaxy Chemicals
(Mumbai, India).
Methods
Fabrication and Optimization of LNG Poly(Lactide co Glycolide)
Microparticles (LNG PLGA)
Microparticles containing LNG were formed by the
modied o/w emulsion in water interrupted solvent evapora
tion technique. The procedure involved placing 200 mL
puried water containing polymeric emulsier/s in a beaker.
The stabilizers namely MC and PVA were employed in
varying concentrations. The drug and PLGA (mg 40:200,
80:200, and 160:200) dissolved in 4 mL of dichloromethane
was gradually added into the aqueous continuum. The
resulting emulsion was stirred at a constant rate at a
temperature of 23 24C. The semisolid droplets (embryonic
microparticles) were then separated from the system followed
by resuspension in 30 mL of stabilizer free water. The
product was harvested after the methylene chloride evapora
tion proceeded to completion (about 10 h). The drug loaded
microspheres were further lyophilized at 0.004 mbar pres
sure, 40C temperature (Labconco Corporation, England,
UK). The two step solvent evaporation process eliminated
the formation of free drug crystals in the aqueous phase or on
the microparticle surface. The formulation was optimized by
Puthli and Vavia

a series of microsphere trials as described in subsequent
sections. All the formulation variables and process parame
ters have been listed in tabular form in Table I. All experi
ments were conducted in triplicate runs.
Drug to Polymer Ratio
The drug to polymer ratio combination trials were taken
in order to get optimum drug encapsulation efciency.
Various combinations of drug and polymer were chosen in
the preparation of microparticles by the solvent evapora
tion technique. Drug to polymer ratios tried were 1:5, 2:5,
and 4:5 wt/wt. The ratio that would give optimum en
capsulation efciency with microparticles having the desired
characteristics without crystallization of the free drug was
investigated.
Emulsifier Concentration
Various emulsiers were tested to obtain a product
without any free drug crystals on the microsphere surface.
The inuence of the emulsier concentration in the aqueous
phase on maintaining the sphericity of the microparticles and
on the drug crystallization behavior was studied. The
following formulations were prepared using different concen
trations and mixtures of the emulsiers:
1. PVA: 0.27, 1.0, 2.0 and 3.0% wt/wt
2. MC: 0.1, 0.3 and 0.5% wt/wt
3. PVA (0.27% wt/wt) + MC (0.05% wt/wt)
Speed of Agitation
Stirring speed trials were undertaken with an objective of
getting perfectly spherical microsphere. The speed of agita
tion was altered and its inuence on the sphericity of the
product was studied. Three speeds tried were 270 rpm (speed
A), 550 rpm (speed B), and 910 rpm (speed C). The speed
was monitored using a digital photo tachometer DT 2234A
(Lutron, Taiwan).
Mode of Solvent Evaporation
Two processes were followed for the evaporation of the
organic solvent. In one method, the solvent was evaporated to
completion as the microparticles were still in the emulsier
containing aqueous phase. This method gave rise to crystalli
zation of the drug. Hence, this method was not followed in
subsequent optimization trials. The second method involved
decanting the nascent microparticles before total solvent
evaporation and resuspending them in plain water (without
the emulsier). The remainder of the residual solvent was then
evaporated to completion in this medium. This technique was
employed in further optimization of the formulation.
Gamma Radiation Sterilization
The LNG PLGA system (batch code LP 10) was
subjected to cobalt 60 radiation at 2.5 Mrad. The samples to
445
be irradiated were lled into 5 mL glass vials, stoppered with

rubber closure, and sealed with aluminum overseal closure.
Vials were packed in dry ice into polyurethane container to
assure a low temperature during the irradiation process.
Although gamma radiation causes a minimal temperature
rise, keeping the temperature low avoids the possibility of
hydrolytic degradation of the PLGA polymer. Placebo micro
spheres, plain LNG, physical mixture of polymer and LNG,
and drug loaded microparticles were subjected to gamma
radiation sterilization. The gamma radiation sterilization
process was carried out in the presence of air and at ambient
temperature. The essential parameter in gamma radiation
sterilization is the measurement of radiation dose. During the
gamma radiation, the dosimetric control was performed using
Fricke dosimeters (ferrous sulfate dosimeter, Bhabha Atomic
Research Centre, India) which were placed inside the gamma
chamber. Non irradiated samples served as control. The study
utilized a gamma chamber GC 900 (Bhabha Atomic Re
search Centre, Bombay, India). A dose rate of 0.13 Mrad/h
was employed. The drug content after radiation was
determined and compared with the non radiated system.
The cumulative effect of all the parameters discussed above
on the nal quality of the microspheres viz. the sphericity,
absence of free drug crystals, process yields, particle size as
well as the drug encapsulation efciency was critically
investigated.
chloride. The solution was further diluted suitably and

spotted on HPTLC plates. The blank consisted of micro
particulate placebo system treated in a similar manner.
Representative standard curve of LNG was constructed by
plotting the peak area as well as peak height versus drug
concentration.
Characterization of the LNGPLGA System

Particle Size Analysis
The particle size was determined by measuring the
diameter of individual particle using an optical microscope
(Nikon YS100, Nikon Instruments Inc., USA). The diameter
of 100 particles was measured and the mean particle size
determined.
Drug Encapsulation Efficiency
Microspheres (10 mg) were accurately weighed and
dissolved in 10 mL of dichloromethane. The polymer was
precipitated using ethanol (volume made up to 25 mL). After
centrifugation, the clear supernatant was subjected to HPTLC
analysis. The analysis was conducted in triplicates. The
encapsulation efciency was calculated by the actual and
theoretical drug loading values.
Analytical Method (HPTLC Method)
The drug loading in microspheres was estimated by the
HPTLC method. Briey, the stationary phase used was
HPTLC precoated silica gel 60 F254 plates (Merck, Germany;
size 10 cm10 cm). An autosampler (Camag Linomat IV,
Switzerland) was used. The mobile phase consisted of
benzene/methanol (9:1 vol/vol). Plate development was linear
ascending in Camag twin trough chamber (Switzerland).
Spectrodensitometric analysis (Camag TLC scanner II, Swit
zerland) was done with a scanning speed of 1 mm/s at a
wavelength of 240 nm. Integration was done in Perkin Elmer
integrator system LCI 100 (Switzerland). Drug loaded micro
spheres (100 mg) were dissolved in 10 mL of methylene
Sphericity
The sphericity of the microspheres was computed by the
Lovgren and Lundberg technique (12). Briey, individual
microparticle was viewed on a projection microscope (model
MP3Nr 3725, Poland). The longest diameter (L) and that
perpendicular to it at its mid point (b) were measured. Fifty
readings were taken. The ratio of L/b was put into class
intervals and the
sphericity calculated using the
percent

formula, S 1 b2 rf 100 , where b is the lower class
interval limit, and rf is the relative frequency.
Moisture Content
The residual water content in the microparticles was
found by the KF technique using a Karl Fischer/Autotitrator
(model 831 KF Coulometer, Metrohm, UK). Dehydrated
methanol (Merck, 20 mL) was titrated to the electrometric
end point with the KF reagent (Merck). The microsphere
sample was then carefully transferred to the titration vessel
and after stirring for 1 min titrated again using the KF reagent
till the characteristic end point.
In Vitro Drug Release Kinetics
In the in vitro release studies, 25 mg of microspheres
(batch LP 10) was placed into a 100 mL stoppered conical
ask containing 20 mL dissolution medium that consisted of
0.9% wt/vol sodium chloride solution in distilled water. To
maintain the sink conditions, 0.5% wt/vol sodium dodecyl
sulfate was added, and to prevent microbial growth, sodium
azide (0.02% wt/vol) was employed. Dissolution testing was
done on the sample subjected to gamma radiation as well as
on the non radiated system. Two methods used in the study
included the static and the shaking method. In the shaking
technique, samples were agitated (80 strokes/min) using a
constant temperature shaker water bath. The temperature
was maintained at 370.5C. The test involved withdrawing
2 mL of microsphere free samples at predetermined time
intervals of 4, 8, 12, 16, 20, 24, 28, and 32 days (replacing with
fresh medium after every sampling) and measuring the drug
release by a sensitive high performance liquid chromatogra
phy (HPLC) method.
The drug release data were tted to various kinetic
models viz. zero order, rst order, and Higuchi kinetics (13).
Higuchi equation describes the release of a drug from an
insoluble matrix as the square root of a time dependent
process. This is essentially based on Fickian diffusion as given
below:
Qt 2DS"A
p
0:5S"0:5 t0:5 kH t
Where Qt is the amount of drug released in time t, D is

the diffusion coefcient, S is the solubility of the drug in the
446
Puthli and Vavia
dissolution medium, is the porosity, A is the drug content

per cubic centimeter of matrix tablet, and kH is the release
rate constant for the Higuchi model. The release data were
also tted to Baker Lonsdale kinetic model (14). This model
is for diffusion controlled release and more specic from
spherical matrix systems and is given below.
h
3=2 1
2=3
F kt
Where F is the fraction of drug released, t is the time,

and k is the rate constant. It has been reported that drug
release from controlled drug delivery systems using typically
hydrophilic matrices shows a time dependent prole. Further,
the drug release decreases with time due to increased
diffusion path length due to a gel layer formed on the surface
that retards further the ingress of uid and subsequent drug
release (15). This leads to rst order release kinetics depicted
in following equation:

F 100 1
kt
Where F is the percentage of drug released at time t.

Zero order kinetics is obtained when the drug release from a
system is independent of the concentration of the drug (16).
When the percent drug release was plotted as a function of
time, a linear relationship was obtained with the r value closet
to unity.
Analytical Method (HPLC)
A validated stability indicating HPLC method was used.
Analysis was conducted on Jasco intelligent unit (Japan)
using Lichrospher RP 18 column (5 m, length 125 mm,
Merck) column and mobile phase consisting of acetonitrile/
water/acetic acid (40:5:5 vol/vol/vol). The ow rate was kept
at 1 mL/min. Detection wavelength was 360 nm using Jasco
UV 975 UV/VIS detector coupled with Borwin V 1.21
chromatography software. The analysis was conducted in
triplicates. Calibration curves were obtained from the plots of
peak area versus drug concentration for the concentration
range of 10 to 80 g/mL. The limit of detection for the drug
was estimated to be 100 ng, the level at which the drug could
be detected without any noise. The limit of quantitation was
found to be 350 ng.
interaction was investigated by infra red absorption spectro

photometry. A FT IR spectrophotometer (Jasco FT/IR 5300
instrument, Japan) was employed in this study (KBr disc
method was used).
Differential scanning calorimetric thermograms were
obtained for LNG, polymer PLGA, physical mixture of drug
and polymer (1/5 wt/wt ratio), and the LNG PLGA micro
particulate system. Microparticulate samples were separately
sealed in aluminum cells and set in a Shimadzu thermal
analyzer DT 40 apparatus (Japan). Thermal analysis was
performed at a heating rate maintained at 10C/min in
nitrogen atmosphere. Alumina was employed as the refer
ence substance.
X ray Diffraction Studies
X ray diffractograms were obtained for LNG, PLGA,
and LNG PLGA microsphere system at a scanning rate of 2/
min. The polymorphic state of the drug in the microparticu
late system was examined by X ray diffraction studies. An
automatic X ray diffractometer (Siemens 5000, Germany)
equipped with an X ray generator was used. Nickel ltered
Cu K1 radiation having a wavelength of 1.5106 , operating
at 35 kW and 20 mA in the range (2) of 5 to 70 was
employed.
Residual Solvent Content
Scanning electron microscopy was performed to examine

the surface morphology of the microparticles (model S 570,
Hitachi, Japan). The microparticles were placed on a metal
stud coated with adhesive label. The sample was sputter
coated with conductive gold palladium (Edwards Sputter
Coater).
The residual methylene chloride content in the LNG

PLGA microspheres was determined by head space gas
chromatography using CHEMITO 8610 HT gas chromato
graph using BPX5 capillary column equipped with the head
space system. The oven, injection port, and detector temper
atures were kept at 40C, 80C, and 240C, respectively.
Nitrogen (ow rate of 5 mL/min) was used as carrier gas and
a ame ionization detector was employed in the study. The
microparticles (200 mg) were dissolved into 5 mL of distilled
water and sonicated for 5 min. The sample was then placed in
the head space tube in the heating block and heated at 80C
for 10 min, and the head space was subsequently injected.
The standard employed in the analysis consisted of 0.1 mL
(132.5 mg) of methylene chloride (99.5%) suitably diluted in
distilled water to obtain a nal concentration of 53 and
8.6 ppm. The above two standard solutions (4 mL) were
taken separately in the head space tube in the heating block
and heated to 80C for 10 min. The head space was injected
and gas chromatograms of the samples and standards were
recorded. The residual solvent content was determined from
the peak areas.
Infra red Spectroscopy
Sterility Testing
IR spectra were taken for LNG, gamma sterilized LNG

at 2.5 Mrad dose, LNG PLGA microspheres, physical
mixture of LNG + PLGA (1:5 wt/wt ratio), PLGA polymer,
and blank placebo microspheres prepared by the same
manufacturing process. The possibility of excipient drug
The test for sterility was done on the optimized batch

according to the United States Pharmacopoeia method (17).
Briey, the sample was added to the uid thyoglycollate
medium and incubated at 32.52.5C for 14 days. The
contents were observed for any microorganism growth. The
Surface Morphology
447
formulation subjected to gamma radiation was subjected to

the sterility testing.
with male rabbits and tested periodically for indication of

conception/pregnancy.
Test for Pyrogens

A test was carried out to conrm the absence of
pyrogens (18). Three healthy, adult rabbits of either sex
weighing not less than 1.5 kg were selected for the study.
Basal body temperature was documented. Sterilized micro
spheres were diluted with pyrogen free saline solution and
injected slowly into the marginal ear vein of each rabbit.
Rectal temperature of each rabbit was monitored every
30 min for 3 h.
Pharmacokinetic Study
The study protocol was approved by the Institutional
Animal Ethics Committee (IAEC, India) prior to the study.
The guidelines of the Committee for the Purpose of Control
and Supervision of Experiments on Animals (CPCSEA,
India) were followed during the study. To evaluate the
efcacy of the formulations in biological milieu, in vivo
studies were conducted on non rodent species. Healthy
female albino Belgium rabbits (Institute for Research in
Reproduction, Mumbai, India) selected from the same colony
with body weight of 1.5 2.0 kg were used in the study. The
performance of the product was assessed by monitoring blood
levels of the drug released from the system. The animals were
quarantined in a well ventilated room with dened room
temperature, humidity, and lighting conditions. Each group
contained six animals. The randomization technique was used
in grouping the animals. For identication purpose, the
animals were tattooed using a suitable color coding scheme.
Each animal received one formulation. For the purpose of
comparison, a control group comprising of equal number of
animals as that of the test group was taken. The animals did
not receive any formulation. They were subjected to mating
with male rabbits. This served as the positive control. All the
animals received standard balanced diet and free access to
water throughout the study.
Based on the in vitro release proles, the LNG PLGA
microsphere (batch LP 10) was selected to evaluate the
therapeutic performance of the designed dosage form. The
LNG loaded microparticles (previously sterilized by gamma
radiation at 2.5 Mrad) were suspended in sterile saline (1 mL
volume). The product after reconstitution was injected
intramuscularly into proven fertile female rabbits. The
injection was done using sterile disposable hypodermic
syringe equipped with a 22 gauge needle. Before injecting
the product, blood samples were taken and this served as the
initial zero day reading. The serum separated served as
serum blanks during the RIA study. Blood samples were
withdrawn at predetermined time intervals of 1, 3, 4, 8, 12, 16,
20, 24, and 28 days from the marginal ear vein of the rabbit.
The serum separated was collected and stored in stoppered
vials at 20C until further analysis. The samples were
analyzed by a sensitive RIA technique. In addition to the
measurement of the drug in the body, the performance of the
formulation as an effective contraceptive was assessed by
examination of conception in rabbits. The animals were kept
Drug Analysis (Radioimmunoassay Technique)

A sensitive radioimmunoassay method was employed in
estimating the drug in the serum samples. For the standard
curve, ethanolic solution of LNG was suitably diluted with
anhydrous ethanol and buffer to obtain concentration levels
of 1,500, 750, 375, 188, 94, 47, and 23 fmol/tube. The solvent
was evaporated and 200 L of blank serum was added to each
of the tubes. The extraction was done with 2 mL of ether and
vortex mixing for 2 min. The tubes were then transferred to a
mixture of solid carbon dioxide and acetone. The ether layer
was carefully decanted into assay tubes after the aqueous
layer was frozen. The tubes were left standing inside the fume
cup board till the solvent evaporated. A buffer (200 L) was
added to each of the assay tube and vortex mixed, allowing
the buffer to roll around the tubes. The tubes were put in a
water bath at 40C and allowed to stand for 5 min. The
contents were then removed and vortexed. To each of the
tube, 50 L of antiserum and 50 L of the working tracer
were added and mixed.
The serum samples collected from the rabbits (200 L
each) were subjected to extraction procedure as described for
the standard curve. To each extraction tube, 50 L of
antiserum and 50 L of the working tracer were added and
vortexed. The vials were then transferred into a beta counter
(Wallac 1409 liquid scintillation counter). An equilibration
time of 24 h was allowed before the counting. Background
counts of the vials were taken prior to use. The counting time
was 180 cpm (beta spectrum). For the calibration curve, a
dose response curve was plotted. The method consisted of
plotting the logits on the y axis and log dose on the x axis.
The concentration of the drug in the serum samples was read
from the calibration plot. Pharmacokinetic parameters viz.
Cmax, AUC0t, AUCinf, and Kel were calculated by the
WINNONLIN program version 5.2 (Pharsight). Mean
residence time (MRT) was computed by the statistical
moment theory method.
All the results are expressed as mean values standard
deviation (SD). For curve tting to different kinetic models in
dissolution studies, the polynomial regression method was
employed. To check whether there was any signicant
difference in the mean of the treated groups, comparison of
the mean values of various groups was performed by one way
analysis of variance. Statistical signicance was dened at
P>0.05.
Optimization of LNGPoly(Lactide-co-Glycolide)
Microparticles (LNGPLGA)
Two methods of solvent evaporation, i.e., continuous
method and interrupted evaporation process, were studied. It
was observed that crystal formation occurred when solvent
evaporation was performed according to the continuous
448
Puthli and Vavia

Table I. The Inuence of Formulation and Process Parameters on the Quality of LNG PLGA Microparticles
Batch
code
D/P ratio
wt/wt
Emulsier concentration
(% wt/wt)
Speed of
agitation
Mean particle
size (m)
Free drug
crystals
EE (%)
mean SD
Sphericity (%)
LP
LP
LP
LP
LP
LP
LP
LP
LP
LP
LP
LP
1:5
1:5
1:5
1:5
1:5
1:5
1:5
1:5
1:5
1:5
2:5
4:5
PVA (0.27)
PVA (0.27)
PVA (0.27)
PVA (1.0)
PVA (2.0)
PVA (3.0)
MC (0.1)
MC (0.3)
MC (0.5)
PVA (0.27) and MC (0.05)
PVA (0.27) and MC (0.05)
PVA (0.27) and MC (0.05)
C
B
A
C
C
C
C
C
C
C
C
C
45.30
49.25
51.82
55.89
48.56
50.55
49.99
46.26
59.23
54.25
+
+
+
+
+
+
+
82.451.25
79.821.09
78.591.33
80.951.40
79.551.23
82.011.09
81.461.19
80.991.38
81.331.27
82.991.42
93.63
84.96
Aggregation
90.61
100
Precipitation
93.66
92.97
91.89
100
88.80
85.92
1
2
3
4
5
6
7
8
9
10
11
12
Moisture
content
(% wt/wt)
0.1402
0.1621
0.1713
0.1321
0.1384
0.1223
0.1408
0.1404
0.1335
0.1218
0.1409
0.1556
Speed of agitation: A=270 rpm, B=550 rpm, and C=910 rpm. Values given are for triplicate samples
LNG PLGA levonorgestrel poly(lactide co glycolide), D/P drug/polymer, EE encapsulation efciency, PVA polyvinyl alcohol, MC methyl
cellulose, + presence of drug crystals, absence of drug crystals
evaporation technique. In the interrupted solvent evaporation

(when the microspheres were removed from the aqueous
continuum and further evaporation was carried out in
stabilizer free water), this crystallization phenomenon was
eliminated. However, in the latter case, if microparticles were
removed at an early stage, they tend to agglomerate due to
excess of methylene chloride. On the other hand, if the
emulsier was removed too late, then this led to the
formation of microcrystals. Hence, it was necessary to dene
the precise time at which the transfer should be executed. The
typical average time window for microsphere transfer was
4510 min. An optimized ratio of polymer and drug (ratio of
5:1, batch LP 10) gave improved encapsulation efciency of
80.99% without free drug crystallization. An impeller agita
tion speed of 910 rpm gave microparticles with sphericity of
100%. The effect of formulation and process parameters on
the morphology is shown in Table I.
Among all the ratios of the dispersed phase to the
aqueous phase, a ratio of 1:5 vol/vol was found to give
spherical particles, and hence, this ratio was selected for
further optimization of the system. In all the batches D50%
was about 51.82 m and D90% was about 40.82 m. The drug
to polymer ratio affected sphericity and formation of free
drug crystals. Higher drug to polymer ratios viz. drug to
polymer ratios of 2:5 wt/wt and 4:5 wt/wt yielded micro
particles with less sphericity (85 88%) along with free drug
crystal formation. At low speeds of agitation, the sphericity of
the microparticles was disturbed and also the process yield
was less due to agglomeration tendencies. Higher stirring
speeds produced microspheres in good yields (about 80%).
The speed did not have a signicant inuence on the drug
payloads. For a constant rate of agitation, the mean particle
size increased as the PVA concentration increased. This was
due to the increase in viscosity of the aqueous phase. The
type and concentration of the emulsier played a deciding
role in maintaining the sphericity of the microparticles. The
microspheres exhibited perfect sphericity when PVA was used
in 2% wt/wt concentration. However, there was crystallization
of the drug. Lower concentrations of PVA gave ovoid shaped
microparticles. Methocel in concentrations of 0.1% to 0.5%
wt/wt gave a product with sphericity ranging from 91% to

93%. Also, in all the cases, there was drug precipitation. The
emulsifying agents employed in the solvent evaporation
process played a key role in successful formation of spherical
microparticles. Methyl cellulose and partially hydrolyzed
polyvinyl alcohol used in the study gave suitable micro
spheres having the desired qualities. A combination of PVA
(0.27% wt/wt) and MC (0.05% wt/wt) yielded perfectly
spherical microparticles (100% sphericity) without free drug
crystallization. Hence, this formulation was selected for
further studies. Thus, a novel interrupted o/w emulsication
solvent evaporation technology was successfully designed for
the production of PLGA microspheres of LNG. The moisture
content determined in all the formulations is given in Table I.
The in vitro drug release proles of the LNG PLGA
microparticles are shown in Fig. 1 (by the static method) and
Fig. 2 (using the agitation method). There was no major
difference seen in the dissolution proles with both the
methods used for dissolution. Thus, agitation of the dissolu
Fig. 1. In vitro dissolution proles of LNG PLGA microspheres

before and after gamma sterilization (n=6) in static dissolution
conditions. Values indicate mean SD (for triplicate samples)
449
shape of the microspheres remained unaltered after gamma
radiation at 2.5 Mrad. There was no evidence of microscopic
pores, and free drug crystals were absent. The microparticu
late system was free from any structural defects.
Infra-red Spectroscopy
Fig. 2. In vitro dissolution proles of LNG PLGA microspheres

before and after gamma sterilization (n=6) in agitation dissolution
conditions. Values indicate mean SD (for triplicate samples)
tion medium did not affect the dissolution rate and extent.
The release data were tted into different kinetic models viz.
zero order, rst order, Higuchi kinetics and Baker Lonsdale
equation. The corresponding coefcients of determination (r)
and slopes were also computed. The t for various kinetic
models and comparison of calculated coefcient of determi
nation indicated that the drug release kinetics followed
predominantly a zero order release prole with r2 value of
0.9826 for the gamma radiated sample and 0.9854 for the
non irradiated LNG PLGA microspheres. Thus, gamma
radiation sterilization did not affect the in vitro drug release
kinetics. Many research groups have studied the effect of
gamma radiation on the product characteristics. Some
experiments have noted signicant changes in drug release
rate (either an increase or decrease in the drug prole was
observed after subjecting the microparticles to gamma
radiation) (19 23). On the other hand, there are reports
where the in vitro release rates from drug delivery devices
remained unaffected by gamma radiation (24 28). In our
studies, the drug release data from gamma irradiated LNG
PLGA microparticles indicate that the cobalt 60 radiation did
not inuence the dissolution prole characteristics. Thus, it
can be reasonably concluded that the radiation gamma dose
of 2.5 Mrad can be safely used to terminally sterilize the
product.
In case of LNG, gamma sterilized LNG, and LNG

PLGA microspheres, there was no shift or changes observed
in the wave numbers (conjugated C=O stretching at
1,655 cm 1, C=C stretching at 1,618 cm 1, methyl C H
asymmetric bending at 1,444 cm 1, alcoholic C O stretching
at 1,066 cm 1, and acetylene C H bending at 692 cm 1). The
IR spectra of PLGA polymer and blank microspheres also
did not show any shift or additional peaks (1,763 cm 1 for
C=O stretch, 1,387 cm 1 for C H stretch, 1,088 and
1,196 cm 1 for C O stretch). The presence of characteristic
peaks of LNG in the physical mixture and drug loaded
microspheres indicates the chemical stability of LNG in the
developed system. The IR spectra of the gamma radiated
drug showed the same absorption bands as the non radiated
LNG. Further similar absorption peaks in the PLGA polymer
and placebo microspheres indicated that the manufacturing
process did not inuence the polymer characteristics. Thus,
the above data suggest that there is a fairly good compatibility
of the drug and polymer in the developed system.
LNG exhibited a characteristic sharp endotherm at
241.9C (Fig. 4). The bulk PLGA did not show any melting
endotherm as seen in plot B. The physical mixture of the drug
and polymer gave an endotherm at 240.1C. However, the
drug loaded microsphere system showed no peaks originating
from LNG. The above ndings suggest that the drug is
uniformly distributed in the polymer matrix.
Gamma Radiation Studies

The product retained its physical characteristics after
exposure to gamma radiation of 2.5 Mrad. The color of the
product did not change after exposure to cobalt 60 irradia
tion. The microsphere system was free owing and no
clumping and/or aggregation behavior was observed. The
drug content was found to be 99.781.22% which indicated
that the microparticulate system was stable to cobalt 60
radiation. The mean particle size of the product was also
not affected by gamma radiation (48.202.53 m).
Surface Morphology
The scanning electron photomicrographs of micro
spheres of batch LP 10 is illustrated in Fig. 3. The
miniaturized particles had a smooth surface texture. The
Fig. 3. Scanning electron photomicrograph of LNG PLGA micro

spheres. There are no free drug crystals on the microsphere surface
450
Puthli and Vavia

Test for Pyrogens
The product passed the test since the summed response
in the rabbits body temperature did not exceed 1.15C.
Pharmacokinetic Study
Fig. 4. Differential scanning calorimetry. A levonorgestrel, B PLGA,

C physical mixture of LNG and PLGA, D LNG PLGA micro
particles
X-ray Diffraction Studies

The X ray diffractogram of the pure drug shows a
characteristic pattern of crystalline nature. The intensity of
the diffraction peaks of PLGA was not sharp (data not
shown). Taking into consideration the diffraction pattern of
the LNG PLGA system after gamma radiation, X ray
diffraction peaks of the drug were retained. The intensity of
the diffraction peaks of LNG in the microsphere system
coincided with those of the pure drug. It is reasonable to
conclude that the solvent evaporation technique employed in
microsphere formation might have resulted in some amounts
of the drug being present in crystalline form and partly in the
amorphous state. The above results can also be justied by
the fact that methylene chloride is a good solvent for LNG
and even a small quantity could affect crystallization of the
drug in a given microparticle sample. Changes in polymorphic
state may also be attributable to the gradual loss of small
amounts of residual methylene chloride from microspheres
on storage.
In the RIA method, a representative standard curve of

LNG was constructed by plotting the log(concentration)
versus logit value. The polynomial regression for the calibra
tion plots showed good linear relationship with coefcient of
correlation 0.9998, slope 1.0993, and intercept 5.1739 over
the concentration range studied.
The plasma concentration time prole is depicted in
Fig. 5. Just like in the case of in vitro dissolution prole, there
was no dose dumping in the in vivo prole as well. This may
be attributed to the modied formulation process of micro
spheres that gave the product without free drug crystals on
the surface. Pharmacokinetic parameters were calculated
using a non compartment model by processing with WinNon
lin statistical software program version 5.2 (Pharsight). The
terminal elimination rate constant (Kel) was calculated by the
linear least square regression analysis of the log linear phase
of the concentration time curve with the last three experi
mental points. The AUC0t was calculated by the linear
trapezoidal rule from 0 to the last measured plasma concen
tration. AUC0inf was calculated by adding Ct/Kel to AUC0t.
Individual pharmacokinetic parameters were calculated for
each of the plasma concentration time prole for each
animal. The mean and SD were then computed and the
mean plot of the plasma concentration time prole was
obtained. The AUC0t was 10,765.12,150.3 pg/mL day 1
(coefcient of variation, CV=0.19), AUC0inf was 36,058.76
6,490.4 pg/mL day 1 (CV=0.17) with Cmax of 469.684.42 pg/
mL (CV=0.17), and Kel was 0.020.004 day 1 (CV=0.2). The
MRT was calculated to be about 15 days. The control group
exhibited signs of conception. No pregnancies occurred in all
test animals injected with the microparticles during the 1
month study period. Also, there was reversibility of fertility in
the rabbits after about 5 6 months time.
It has been reported in the literature that inhibition of
ovulation occurs at LNG concentration of 500 800 pg/mL
Residual Solvent Content

The residual methylene chloride content was found to be
14 ppm. This was well below the reported limit of 500 ppm
(29).
Sterility Testing
The product (in all the experimental doses of gamma
radiation) complied with the test for sterility.
Fig. 5. Plasma concentration time prole LNG release from the

microparticle system after intramuscular injection in female rabbits
(n=6). Values indicate mean SD (value for triplicate samples)

(30). In a comparative study, the article claims that serum
LNG levels during the 12th month of use were 500 800 pg/
mL in the polydimethylsiloxane (silastic) rod users, while in
the capsule users, they were 150 300 pg/mL (31). However, it
should be noted that suppression of ovulation may not be
critical for contraceptive efcacy. Progestin only contracep
tive systems have achieved acceptable rates of contraception
without inhibition of ovulation but by rendering the endome
trium inhospitable to nidation or cervical mucus impermeable
to sperm migration (32). A randomized clinical trial provides
evidence that the minimum LNG concentration necessary to
protect against pregnancy is below 200 pg/mL and possibly
below 175 pg/mL. This study also indicates that drug
concentration in the upper part of the range 151 200 pg/mL
is protective against pregnancy (33). In our studies, there was
no tissue inammation or erythemia observed at the site of
injection, which indicates that this dosage form is biocompat
ible with the formulation excipients. Comparison of drug
release observed in the 1 month period in the in vitro
dissolution studies and similar plasma prole where the drug
levels start to decrease after day 28 indicates good in vitro/in
vivo correlation. More signicantly, the gamma irradiation
effects did not inuence the in vivo release kinetics and
performance of the formulation. It has been reported in the
literature that there is good correlation of the in vivo drug
release from microsphere systems in rabbits and humans
(34,35).
There are very few commercially available products for
contraceptives that utilize long term controlled release tech
nology. A silicone based implantable device using non
biodegradable polymer is available for the delivery of LNG
by the trade name Norplant (marketed by Population
Council). The formulation needs surgical manipulations for
the insertion and retrieval of the product. Since we have
employed a biodegradable polymer to develop the
formulation, it is not required to be removed after cessation
of therapy. Our attempt was successful in formulating a
pharmaceutically acceptable system that avoids the above
shortcomings with ease of single intramuscular administration
of the product. Oral route of administration of contraceptive
steroids requires higher quantities of the active moieties. This
leads to increase in side effects. The undesired side effects
observed after oral administration of LNG in clinical trials
range from minor ones like headache/dizziness, nausea,
abdominal pain, vomiting, to more signicant like high blood
pressure (severe headache, ushing, blurred vision), fatigue,
menstrual changes, and liver damage. Oral contraceptives have
been associated with thrombophlebitis, arterial thromboembolism,
pulmonary embolism, myocardial infarction, cerebral hemorrhage,
cerebral thrombosis, hypertension, gallbladder disease, hepatic
adenomas, carcinomas, or benign liver tumors. The daily dose of
oral tablet is 0.1 0.15 mg which is generally given in combination
with ethinyl estradiol. The serum concentration of LNG after daily
oral administration of the tablet formulation is much higher than
that required for contraception. Here, we present an injectable
system with lesser dose in the formulation as compared to the oral
tablets. Thus, a transformation from oral delivery system to a long
term parenteral product using lower amounts of the drug would
lead to reduced unwanted side effects, which is an additional
feature of the developed system.
451
CONCLUSIONS
The present investigation demonstrates that gamma
radiation sterilization can be a method of choice for steroid
based microparticulate systems of poly(lactide co glycolide).
The drug release can be controlled for a 1 moth period with
reliable reversibility of fertility. Further, the dose dumping
phenomena observed with these systems can be avoided by a
change in the preparative technique of the microspheres.
ACKNOWLEDGMENTS
The authors would like to thank Purac Biochem for
providing poly(lactide co glycolide), Wyeth Laboratories for
LNG, and World Health Organization (London) for provid
ing radioimmunoassay kit. The authors acknowledge the
valuable help of Dr. Dixit in preparation of this manuscript.
REFERENCES
1. Jiang W, Gupta RK, Deshpande MC, Schwendeman SP.
Biodegradable poly(lactide co glycolide) microparticles for in
jectable delivery of vaccine antigens. Adv Drug Deliv Rev
2005;57:391 410.
2. Panyam J, Labhasetwar V. Biodegradable nanoparticles for drug
and gene delivery to cells and tissues. Adv Drug Deliv Rev
2003;55:329 47.
3. Mundargi RC, Ramesh Babu V, Rangaswamy V, Patel P,
Aminabhavi TM. Nano/micro technologies for delivering
macromolecular therapeutics using poly(D,L lactide co glyco
lide) and its derivatives. J Control Release 2008;125:193
209.
4. Heya T, Okada H, Ogawa Y, Toguchi H. Factors inuencing the
proles of TRH release from copoly(d,l lactic/glycolic acid)
microspheres. Int J Pharm 1991;72:199 205.
5. Sah HK, Chien YW. Evaluation of a microreservoir type
biodegradable microcapsule for controlled release of proteins.
6. Urata T, Arimori K, Nakano M. Modication of release rates of
cyclosporin A from polyl(L lactic acid) microspheres by fatty
acid esters and in vivo evaluation of the microspheres. J Control
Release 1999;58:133 41.
7. Chattaraj SC, Rathinavelu A, Das SK. Biodegradable micro
particles of inuenza viral vaccine: comparison of the effects of
routes of administration on the in vivo immune response in mice.
J Control Release 1999;58:223 32.
8. Viswanathan NB, Thomas PA, Pandit JK, Kulkarni MG,
Mashelkar RA. Preparation of non porous microspheres with
high entrapment efciency of proteins by a (water in oil) in oil
emulsion technique. J. Control Release 1999;58:9 20.
9. Okada HY, Heya IT, Uneo H, Ogawa Y, Toguchi H. Pharma
cokinetics of once a month injectable microspheres of leuprolide
acetate. Pharm Res 1991;8:787 91.
10. Wang SH, Zhang LC, Lin F, Sa XY, Zuo JB, Shao QX, Chen GS,
Zeng S. Controlled release of levonorgestrel from biodegradable
poly(d,l lactide co glycolide) microspheres: in vitro and in vivo
studies. Int J Pharm 2005;301:217 25.
11. Beck LR, Pope VZ, Tice TR, Gilley RM. Long acting injectable
microsphere formulation for the parenteral administration of
levonorgestrel. Adv Contracept 1985;1:119 29.
12. Lovgren K, Lundberg P. Determination of sphericity of
pellets prepared by extrusion/spheronization and the impact
of some process parameters. Drug Dev Ind Pharm 1989;15:
2375 92.
13. Higuchi T. Mechanism of sustained action medication: theoret
ical analysis of rate of release of solid drugs dispersed in solid
matrices. J Pharm Sci 1963;52:1145 9.
452
14. Doshi CC, Bhalla HL. In vitro release studies of levonorgestrel
loaded biodegradable microspheres. Ind J Pharm Sci 1999;61:39
43.
15. Shah MV, De Gennaro MD, Suryakasuma H. An evaluation of
albumin microcapsules prepared using a multiple emulsion
technique. J Microencapsul 1987;4:223 38.
16. Ranga Rao KV, Padmalatha DK, Buri PK. Cellulose matrices for
zero order release of soluble drugs. Drug Dev Ind Pharm
1988;14:2299 320.
17. United States Pharmacopoeia, Microbiological tests: Sterility
Tests, USP 29 NF 24, The United States Pharmacopoeial
Convention, Inc., Rockville, 2006.
18. European Pharmacopoeia, Appendix XIVD Test for pyrogens,
Ph. Eur. method 2.6.8, Council of Europe, 2005.
19. Beck LR, Pope VZ. Controlled release delivery systems for
hormones. A review of their properties and current therapeutic
use. Drugs 1984;27:528 47.
20. Sanders LM, Kent JS, McRae GI, Vickery BH, Tice TR, Lewis
DH. Controlled release of a luteinizing hormone releasing
hormone analogue from poly(d,l lactide co glycolide) micro
spheres. J Pharm Sci 1984;73:1294 7.
21. Spenlehauer G, Vert M, Benoit JP, Chabot F, Veillard M.
Biodegradable cisplatin microspheres prepared by the solvent
evaporation method: morphology and release characteristics. J
Control Release 1988;7:217 29.
22. Spenlehauer G, Vert M, Benoit JP, Boddaert A. In vitro and in
vivo degradation of poly(D,L lactide/glycolide) type micro
spheres made by solvent evaporation method. Biomaterials
1989;10:557 63.
23. Ruiz JM, Busnel JP, Benoit JP. Inuence of average molecular
weights of poly(DL lactic acid co glycolic acid) copolymers 50/50
on phase separation and in vitro drug release from microspheres.
Pharm Res 1990;7:928 34.
24. Wise DL, Trantolo DJ, Marino RT, Kitchell JP. Opportunities
and challenges in the design of implantable biodegradable
polymeric systems for the delivery of antimicrobial agents and
vaccines. Adv Drug Deliv Rev 1987;1:19 39.
25. Hartas SR, Collett JH, Booth C. The inuence of gamma irradiation
on the release of melatonin from poly(lactide coglycolide) micro
spheres. Proc Int Symp Control Release Bioact Mater 1992;19:321 2.
Puthli and Vavia

26. Crossan IM, Whateley TL. Microspheres of dexamethasone as a
biodegradable controlled release drug delivery system. Proc Int
Symp Control Release Bioact Mater 1994;21:184 5.
27. Fallon PA, Whateley TL. The development and evaluation of
poly(lactic glycolic acid) implants for sustained delivery of
dexamethasone to the brain. Proc Int Symp Control Release
Bioact Mater 1994;21:264 5.
28. Yoshioka S, Aso Y, Otsuka T, Kojima S. The effect of
irradiation on drug release from poly(lactide) microspheres.
Radiat Phys Chem 1995;46:281 5.
29. ICH harmonized tripartite guideline, Q3C(R3) Impurities:
guideline for residual solvents (http://www.ich.org/lob/media/
media423.pdf).
30. Roy S, Stanczyk FZ, Mishell DR, Lumkin M, Gentzschein E.
Clinical and endocrinologic study of continuous levonorgestrel
administration from subcutaneous solid polydimethylsiloxane
rods. Contraception 1980;21:595 615.
31. Moore DE, Roy S, Stanczyk FZ, Mishell D. Bleeding and
serum d norgestrel, estradiol and progesterone patterns in
women using d norgestrel subdermal polysiloxane capsules
for contraception. Contraception 1978;17:315 28.
32. Ory SJ, Hammond CB, Yancy SG, Hendren RW, Pitt CG. The
effect of a biodegradable contraceptive capsule (Capronor)
containing levonorgestrel on gonadotropin, estrogen, and pro
gesterone levels. Am J Obstet Gynecol 1983;145:600 5.
33. Sivin I, Lhteenmki P, Ranta S, Darney P, Klaisle C, Wan L,
Mishell DR, Lacarra M, Viegas OA, Bilhareus P, Koetsawang S,
Piya Anant M, Diaz S, Pavez M, Alvarez F, Brache V,
LaGuardia K, Nash H, Stern J. Levonorgestrel concentrations
during use of levonorgestrel rod (LNG ROD) implants. Contra
ception 1997;55:81 5.
34. Lancranjan I, Bruns C, Grass P, Jaquet P, Jervell J, Kendall
Taylor P, Lamberts SW, Marbach P, Orskov H, Pagani G,
Sheppard M, Simionescu L. Sandostatin LAR: a promising
therapeutic tool in the management of acromegalic patients.
Metabolism 1996;45:67 71.
35. Comets E, Mentre F, Kawai R, Nimmerfall F, Marbach P,
Vonderscher J. Modeling the kinetics of release of octreotide
from long acting formulations injected intramuscularly in rabbits.
J Pharm Sci 2000;89:1123 33.

DOI: 10.1208/s12249 009 9231 4
Research Article
Development and Evaluation of Sustained Release Gastroretentive Minimatrices
for Effective Treatment of H. pylori Infection
Atul C. Badhan,1,2 Rajashree C. Mashru,1 Punit P. Shah,1 Arti R. Thakkar,1 and Nitin B. Dobaria1
Abstract. In the present work, sustained release gastroretentive minimatrices of amoxicillin have been
designed and optimized using central composite design. Effect of amount of xanthan gum, rate
controlling polymers (HPMC K100M CR/PEO coagulant (1:1)), carbopol 974P, and gas generating
couple (sodium bicarbonate/citric acid (3:1)) was studied on dependent (response) variables, i.e.,
buoyancy lag time, drug release at 1 h, time required for 95% drug release, swelling index, and
bioadhesive strength. Minimatrices were prepared by non aqueous granulation method using solution
of PVP K30 in isopropyl alcohol. All the formulations were found to contain 99.2% to 100.9% of
amoxicillin per minimatrix. Optimum formulation (Formulation number AGT09) containing high level of
the independent variables was having buoyancy lag time of 7 min and drug release at 1 h was 32.5%. It
required 9.39 h for 95% drug release while swelling index and bioadhesive strength were 341 and
17.9 dyn/cm2, respectively. This formulation was said to be optimum because it has minimum buoyancy
lag time, requires maximum time for 95% drug release, and has higher bioadhesive capabilities. In vitro
results of an optimized formulation indicate its sustained drug release and gastric retention capability,
which may be very useful for effective treatment of H. pylori infection.
KEY WORDS: central composite design; gastroretentive drug delivery system (GRDDS); Helicobacter
pylori (H. pylori).
INTRODUCTION
For certain drug candidates, prolonging the gastric
retention is desirable for achieving greater therapeutic
benet. For example, drugs that are absorbed in the proximal
part of the gastrointestinal tract (1) and drugs that are less
soluble in or are degraded by the alkaline pH may benet
from prolonged gastric retention (2 4). In addition, for local
and sustained drug delivery to the stomach and proximal
small intestine to treat certain conditions, prolonged gastric
retention of the therapeutic moiety may offer numerous
advantages including improved bioavailability and therapeu
tic efcacy and possible reduction of dose size (5 7).
Since its discovery in 1982 by Warren and Marshall
(leading to their recent Nobel Prize in Medicine) and its
conrmation as a pathogen at the end of the 1980s, re
searchers have attempted in several ways to efciently
eradicate Helicobacter pylori from the stomach. It is well
known that long lasting H. pylori infections can lead to severe
diseases such as gastric cancers and mucosa associated
lymphoid tissue lymphomas. In most countries, H. pylori
infection is associated with a four to sixfold increased risk of
Centre of Relevance and Excellence in Novel Drug Delivery

systems, Pharmacy Department, The M.S. University of Baroda,
Fatehgunj, Baroda 390002, India.
2
To whom correspondence should be addressed. (e mail: atulbadhan@
rediffmail.com)
gastric cancer. This means that the majority of gastric

carcinomas in the world are related to H. pylori infection
(8,9). Since 1994, the International Agency for Research on
Cancer and the World Health Organization has been
considering that H. pylori infection is carcinogenic to humans
(group 1 carcinogen). Because of the high level of antibiotic
resistance to H. pylori and the poor patient compliance, new
medicines with better effectiveness and simpler regimens are
required. It has been suggested that prolonged local avail
ability of antibacterial agents may augment their effectiveness
in treating H. pylori infection (10). In particular, H. pylori
lives deep within the gastric mucus layer, and prolonged local
application of drug is needed for its sufcient diffusion to the
bacteria. A logical way to improve the effectiveness of
therapy is to develop a drug delivery system which can reside
in the stomach for longer duration and release drug as long as
possible in the ecological niche of the bacterium (11), and
Gastroretentive Drug Delivery System (GRDDS) is an
ultimate solution for this.
Extensive efforts have been made in both academia and
industry towards the development of GRDDS (7). These
efforts resulted in GRDDS that were designed in large part
based on various approaches which include (a) low density
dosage form that causes buoyancy above gastric uid (12); (b)
high density dosage form that is retained in the bottom of the
stomach (13); (c) bioadhesion to the stomach mucosa (14,15);
(d) slowed motility of the gastrointestinal tract by concomi
tant administration of drugs or pharmaceutical excipients
(16); (e) expansion by swelling or unfolding to a large size
459
Badhan et al.
460
which limits emptying of the dosage form through the pyloric
sphincter (17).
An important discrepancy has always been noted be
tween the very potent activity of amoxicillin against H. pylori
when tested in vitro by conventional methods such as the
minimum inhibitory concentration method and the results of
H. pylori eradication in vivo. Eradication is achieved only in
approximately 10% to 20% of cases. H. pylori infection is a
mucosal infection, with bacteria lying in the mucous layer and
being strongly attached to the cells. This attachment could
modify the susceptibility of bacteria to antibiotics. Moreover,
H. pylori lives in an environment which does not seem to be
favorable to phagocytic cells and therefore, a bactericidal
instead of a bacteriostatic effect must be considered (18).
Conventional drug delivery systems cannot maintain
effective drug concentration for longer time in stomach due
to their short gastric residence time. Hence, objective of the
present research work was to develop a GRDDS in the form
of minimatrices of amoxicillin. Multiparticulate drug delivery
systems usually based on subunits such as minimatrices,
granules, or pellets show numerous advantages over mono
lithic devices (undivided forms) such as higher degree of
dispersion in the gastrointestinal tract and reduced risk of
systemic toxicity due to dose dumping (19,20). Due to
unpredictable gastric emptying associated with migrating
myoelectric complex motility pattern, multiparticulate sys
tems are more advantageous than the single unit systems, as
the later ones experience all or none emptying pattern
Table I. Formulation Designing by Central Composite Design

Factor levelsa
Formulation no.
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
a
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
20
19
21
22
23
24
25
26
X1
X2
X3
X4
0
1
1
0
1
0
1
0
1
0
1
1
1
1
0
1
1
0
1.48
1
1
1
1.48
1
1
0
0
1
1
0
1
0
1
0
1
1.48
1
1
1
1
0
1
1
1.48
0
1
1
1
0
1
1
0
1.48
1
1
0
1
0
1
0
1
0
1
1
1
1
0
1
1
0
0
1
1
1
0
1
1
1.48
0
1
1
1.48
1
0
1
1.48
1
0
1
1
1
1
0
1
1
0
0
1
1
1
0
1
1
0
1.48, 1, 0, 1, 1.48 are the coded values for level of formulation

variables
Table II. Coded Values and Actual Values of Formulation Variables

in Central Composite Design
Actual valuesa
Coded values
1.48
1
0
1
1.48
a
X1
X2
X3
X4
6.83
9
12
18
20.17
4.55
6
9
12
13.45
4.55
6
9
12
13.45
2.28
3
4.5
6
6.72
Actual values indicate % w/w of nal weight of minimatrix
from the stomach (13). Advantageously, minimatrices mostly

having a diameter of 2 3 mm can be manufactured with
higher reproducibility compared to pellets, especially, regard
ing their weight and equal dimension (21). The amoxicillin
minimatrices developed in the present work would have
longer gastric residence time due to oating as well as gastric
mucoadhesive property.
Materials
Amoxicillin trihydrate was received as a gift sample from
Aristo Pharmaceuticals Ltd. (Mumbai, India). Polyethylene
oxide (PEO) coagulant and hydroxypropylmethylcellulose
(HPMC) K100M CR were gifted by Colorcon Asia Pvt. Ltd.
(Goa, India). Microcrystalline cellulose (Avicel PH102) and
Carbopol 974P were obtained from Signet Chemical Corpo
ration (Mumbai, India) and BF Goodrich Co. (Clevelend,
OH), respectively. Xanthan gum, talc, magnesium stearate,
and polyvinylpyrrolidone (PVP) K30 were purchased from S.
D. Fine Chemicals (Mumbai, India). Sodium bicarbonate,
citric acid, and isopropyl alcohol were purchased from
Qualigens Fine Chemicals (Mumbai, India).
Methods
Experimental Design
Design of experiment has been widely used in pharma
ceutical eld to study the effect of formulation variables and
their interaction on dependent (response) variables (22 24).
In the present study, central composite design (orthogonal)
was used for formulation designing and optimization. The
experimental design consists of total 26 experiments (see
Table I), which include 16 factorial, eight axial, and two
center points. Level of xanthan gum (X1), rate controlling
polymers (HPMC K100M CR/PEO coagulant (1:1)) (X2),
carbopol 974P (X3), and gas generating couple (sodium
bicarbonate/citric acid (3:1)) (X4) were selected as formula
tion (independent) variables. The formulation variables and
their levels as shown in Table II were chosen from the
knowledge obtained from the preliminary studies in our
laboratory. In addition to formulation variables, each mini
matrix contained amoxicillin trihydrate 40.97% w/w (equiva
lent to 12.5 mg of amoxicillin), PVP K30 5% w/w, talc 0.5%
Effective Treatment of H. pylori Infection
461
w/w, magnesium stearate 0.5% w/w, and microcrystalline

cellulose as diluent to adjust nal weight to 35 mg.
Buoyancy lag time (Y1), drug release at 1 h (Y2), time
required for 95% drug release (Y3), swelling index (Y4), and
bioadhesive strength (Y5) were studied as response (depen
dent) variables.
All the response variables were tted to quadratic model
and regression analysis was carried out to get a quantitative
relationship between dependent and the analyzed indepen
dent variables. The equation can be given as
Yi b0 b1 X1 b2 X2 b3 X3 b4 X4 b11 X12
Buoyancy Lag Time

It is the time interval between introduction of the
minimatrices in the dissolution vessel to the time when these
start oating towards the surface of dissolution medium. It
was determined simultaneously during drug release study.
Drug Release Study
b22 X22 b33 X32 b44 X42 b12 X1 X2 b13 X1 X3

b14 X1 X4 b23 X2 X3 b24 X2 X4 b34 X3 X4
ate dilutions were done. Drug content was estimated at max

of 229 nm.
where b0 is arithmetic mean of 26 runs; bi is an estimated

coefcient for factors X1, X2, X3, and X4. All experimental
results were computed by statistical software DOE v6.0.5
(Stat Ease Inc., Minneapolis, MN, USA). Response surface
plots, showing effect of formulation variables on various
response variables, were generated using JMP software v5.1
(SAS Institute Inc., Cary, NC, USA)
Preparation of Minimatrices
Required quantities of amoxicillin trihydrate, xanthan
gum, PEO coagulant, HPMC K100M CR, sodium bicarbon
ate, citric acid, and microcrystalline cellulose (Avicel PH102)
were properly mixed and passed through sieve no.30 (Jayant
Scientic Sieves, Mumbai, India). PVP K30 solution (5% w/v)
was prepared by dissolving it in isopropyl alcohol. This solution
was slowly added to dry powder blend and kneading was done
by hand to obtain a granular mass of sufcient strength.
Granulated mass was air dried at room temperature for 15
20 min and then was dried at 40C for 20 min in tray dryer
(Shree Kailash Industries, Baroda, India). Dried mass was
passed through sieve no. 30 and resultant granules were lu
bricated by adding carbopol 974P, talc, and magnesium stearate.
All the lubricants were previously passed through sieve no.
40. Bulk density of the lubricated granules was determined
by using density test apparatus (Electrolab, Mumbai, India)
and angle of repose was determined by funnel method.
Lubricated granules were compressed into minimatrices
on eight station rotary tablet compression machine (General
Machinery Co., Mumbai, India) using 4 mm circular multi tip
punches. Compressed minimatrices were evaluated for weight
variation, thickness, hardness, and friability as in process
quality control parameters.
Drug Content
UV spectrophotometric method (UV 1700, Pharmaspec,
Shimadzu, Japan) was developed for estimation of amoxicillin
content. This method was validated for linearity, specicity,
accuracy, and precision. Twenty minimatrices were nely
powdered. Powder equivalent to weight of one minimatrix
was taken in 100 ml volumetric ask. About 70 80 ml of
0.1 N HCl was added to it and sonication (Vetra, Italy) was
done for 20 min. Volume was made up to 100 ml. The solution
was ltered using Whatman lter paper type I and appropri
Drug release study was carried out in 900 ml of 0.1 N

HCl at 370.5C using USP type II dissolution test apparatus
(VDA 6 DR, Veego Instruments Corporation, Mumbai, India)
at 50 rpm. Sample (5 ml) was withdrawn at 1, 2, 3, 4, 6, 8, 10,
and 12 h and was replenished with equal volume of dissolution
medium. After suitable dilution, amount of drug released was
estimated by UV spectrophotometric method at 229 nm.
Fluid Uptake Study
This study was carried out by a new and convenient
method using the baskets of dissolution test apparatus. Five
minimatrices were placed in a basket which was immersed in
a Petri dish having 100 ml of 0.1 N HCl. Baskets were
removed every hour, excess of the 0.1 N HCl was soaked by
tissue paper and nal weight was measured. The study was
carried up to 12 h. Fluid uptake capacity was expressed as
swelling index, and it was calculated by Eq. 2.
Swelling Index W2
W1 =W1 100
where W1 =Initial weight of minimatrices and W2 =Weight of

wet minimatrices at 12 h
Bioadhesion Study
This study was performed using Instron tensiometer
(Instron 1121, UK). On the upper jaw of tensiometer, single
minimatrix was stuck using adhesive tape and on the lower
jaw goat stomach tissue (which was freshly collected from
local slaughterhouse) was xed. Upper jaw, having 10 g load
cell, was lowered until it came in proper contact with the
tissue and was kept as such for 20 s. Afterwards, upper jaw
was moved in upward direction at speed of 5 mm/min until
the minimatrix was completely detached from the tissue.
During the test, goat stomach tissue was wetted by adding
20 l of 0.1 N HCl. Force in dyn/cm2 required for this
detachment was measured.
Evaluation of Granule Properties
For all the designed formulations, bulk density was found
between 0.42 and 0.58 gm/cm3. Angle of repose was between
30 and 40 which indicates good ow properties (25). Good
ow properties are important for avoiding weight variation
problems during compression of the minimatrices.
Badhan et al.
462
Table III. Values of the Response Variables
Buoyancy lag time
(min) SDa
Drug Release at
1 h (%) SDa
Time for 95%drug

release (h) SDa
Swelling index SDa
Bioadhesion
(x103 dyn/cm2) SDa
Y1
Y2
Y3
Y4
Y5
171.0
162.6
121.0
322.0
212.0
151.0
102.0
81.0
71.0
192.0
232.0
222.0
141.7
252.6
152.0
182.0
211.0
163.0
211.7
222.0
241.0
152.6
181.0
221.0
141.0
161.7
43.40.98
41.10.95
45.10.89
46.31.21
48.61.39
45.61.85
38.11.11
46.91.64
32.51.93
47.12.00
42.21.82
35.21.49
39.51.32
53.31.75
45.61.28
48.10.92
40.91.83
42.31.87
38.51.21
31.91.56
45.20.95
49.20.90
50.61.41
47.30.79
52.61.87
45.41.14
4.390.36
5.930.16
4.510.10
4.420.14
4.440.22
4.440.27
6.250.19
4.450.20
9.390.17
3.340.22
7.580.23
9.200.16
7.880.55
3.310.08
4.440.11
5.860.14
6.210.16
5.850.56
7.470.45
5.940.19
4.510.16
4.360.44
4.330.20
5.970.18
3.320.14
6.900.17
345.810.74
501.220.91
317.25.79
342.212.43
204.98.36
296.65.98
453.913.30
329.58.11
341.016.46
258.68.81
200.23.99
412.612.25
179.15.20
314.011.52
296.612.50
215.55.25
443.020.32
393.27.51
382.56.72
466.820.91
423.712.13
225.511.08
159.05.13
289.312.93
250.36.97
310.06.70
6.80.26
7.90.36
7.20.26
9.30.44
12.90.26
10.10.46
12.20.61
9.70.46
17.90.52
9.20.40
15.30.70
17.10.44
16.90.26
7.00.26
8.90.44
8.70.26
17.50.40
11.80.53
11.20.50
8.20.30
7.10.26
10.20.26
8.80.36
8.30.40
7.00.26
18.20.78
Formulation no.
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
AGT
a
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Values represent Average Standard Deviation of three experiments
Experimental Design
Preliminary experiments in the laboratory revealed that
independent variables X1 and X2 play signicant role in
sustaining drug release while X3 was important for maintain
ing matrix integrity, sustaining drug release and for its

bioadhesive feature. Variable X4 had prominent role for
achieving minimum buoyancy lag time. Total buoyancy time
depends on the overall entrapment of the gas in the matrix
network formed by X1, X2, and X3. Hence, these four
Table IV. Estimation of Regression Coefcients for Different Response Variables

Y2
Y1
Y3
Term
EC
Prob > F
EC
Prob > F
bo
X1
X2
X3
X4
X12
X22
X32
X42
X1X2
X1X3
X1X4
X2X3
X2X4
X3X4
R2
16.80
0.64
0.17
1.00
5.32
1.01
0.13
0.59
1.01
0.19
0.44
0.81
0.19
0.44
0.56
0.8442
0.4007
0.8220
0.1981
< 0.0001
0.3657
0.9031
0.5921
0.3657
0.8250
0.6076
0.3474
0.8250
0.6076
0.5109
45.36
4.51
2.73
2.06
0.20
1.81
0.24
0.38
0.62
0.29
0.09
0.06
0.39
0.06
0.64
0.9012
<0.0001
0.0007
0.0048
0.7371
0.0565
0.7811
0.6647
0.4797
0.6720
0.8971
0.9264
0.5696
0.9264
0.3556
EC
4.12
0.64
1.28
0.89
0.06
0.54
0.30
0.78
0.23
0.08
0.18
0.09
0.20
0.07
0.13
0.9676
Y4
Y5
Prob > F
EC
Prob > F
< 0.0001
< 0.0001
< 0.0001
0.5929
0.0037
0.0693
0.0003
0.1564
0.4792
0.1529
0.4542
0.1094
0.5523
0.2809
310.40
90.84
4.33
14.05
10.07
2.34
3.57
4.49
8.10
6.88
17.83
5.59
17.36
5.68
4.01
0.8466
< 0.0001
0.7316
0.2781
0.4310
0.8982
0.8454
0.8068
0.6595
0.6304
0.2260
0.6951
0.2378
0.6907
0.7786
EC
9.17
0.54
1.28
3.55
0.38
0.22
0.69
1.60
0.24
0.13
0.44
0.44
0.41
0.46
0.18
0.9383
The terms having Prob > F values very small (<0.0001) indicate that these have signicant effect on the response variables
EC estimated coefcient;
Prob > F
0.1197
0.0020
< 0.0001
0.2527
0.6424
0.1623
0.0052
0.6200
0.7336
0.2473
0.2473
0.2737
0.2230
0.6346
463
formulation variables were selected for systematic optimiza

tion studies. Results of the experiments carried out as
per central composite design are shown in Table III.
The dependent and independent variables related using the
mathematical relationships are shown in Table IV. The
polynomial equations can be used to draw conclusions by
considering sign (positive or negative) and magnitude of the
coefcient. High values of coefcient of determination (R2)
indicate good t. The prediction proler correlating
independent and response variables is shown in Fig. 1.
Drug Content
All the formulations were found to contain 99.2% to
100.9% of added amount of amoxicillin per minimatrix. Drug
content was estimated as per the procedure described in
Methods section. Sonication was necessary in the proce
dure as the minimatrices contained polymers which have
tendency to form matrix and inhibit drug release. Preliminary
experiments conrmed 20 min of sonication time to be
sufcient for complete drug extraction from the matrix
network.
Buoyancy Lag Time
Aim of the present research work was to develop a
formulation having gastroretentive capabilities which can be
achieved by imparting oating and gastric mucoadhesive
properties. Hence buoyancy lag time is very important
parameter for the developed formulations. Short lag time
may ensure immediate oating of the minimatrices and may
further avoid settling of the formulation in lower part of
stomach and ultimately avoid escape of the formulation from
Fig. 1. Prediction proler correlating independent variables and response

variables
Fig. 2. a Response surface plot and b contour plot showing effect of

X1 and X4 on buoyancy lag time
pyloric sphincter. Buoyancy lag time varied from 7 to 32 min

for the developed formulations. Least lag time of 7 min and
maximum lag time of 32 min was observed for Formulation
no. AGT 09 and AGT 04, respectively. Least lag time might
be observed in Formulation no. AGT 09 due to presence of
higher amount of gas generating couple (X4) as compared to
other formulations.
As can be seen from the results of regression analysis in
Table IV and prediction proler in Fig. 1, gas generating
couple (sodium bicarbonate/citric acid (3:1)) (X4) signicantly
decreased buoyancy lag time. Sodium bicarbonate and citric
acid react in presence of acidic dissolution medium and
generates carbon dioxide which gets entrapped in polymer
matrix and decreases density of the minimatrix (26). Sodium
bicarbonate alone can react with gastric uid to produce
carbon dioxide. But citric acid was also included in the
formulation to assure that an acidic microenvironment within
the swelling matrix is maintained. This may contribute to
continuous generation of carbon dioxide in the matrix
independent of external changes in the pH environment.
Xanthan gum has the tendency to form a viscous gel.
Formation of viscous gel entraps the gas bubbles inside the
matrix and minimizes chances of bubbles getting escaped
from the polymer network channels. This in turn led to
oating behavior of the minimatrices for longer duration.
464
Badhan et al.
Hence all the developed formulations were found oating up

to 12 h. Response surface plot in Fig. 2a and corresponding
contour plot in Fig. 2b show effect of different levels of
xanthan gum (X1) and gas generating couple (X4) on buoy
ancy lag time (Y1). Physical integrity of all the formulations
was maintained due to presence of carbopol which becomes
viscous in presence of water and tends to bind the mixed
polymeric system together and reduces matrix erosion (27).
Drug Release Study
Sustaining drug release is very important aspect for
maintaining drug concentration for longer time in the
stomach, which is residence site of H. pylori. Maintaining
effective drug concentration for longer time may completely
eradicate H. pylori infection (11). Hence, core goal of the
present research work was to prepare a formulation having
gastroretentive capability with sustained drug release feature.
Drug release can be sustained for longer time by
retarding initial hour release to maximum possible extent.
Xanthan gum (X1), rate controlling polymers (HPMC,

X1 and X2 on time required for 95% drug release

X1 and X2 on percentage of drug release at 1 h
K100M CR, and PEO; X2) and carbopol 974P (X3) were
found to play important role in decreasing drug release at
initial hour (Fig. 1). Drug release at 1 h (Y2) was 32.5% for
Formulation no. AGT 09 which contains high level of gum,
polymers, and carbopol while it was 53.3% for Formulation
no. AGT 14 containing lowest level of these formulation
variables. Results of regression analysis (Table IV), from its
negative sign and magnitude and smaller value of Prob>F,
indicate that xanthan gum has signicant role in retarding
drug release at rst hour as compared to HPMC, PEO, and
carbopol.
HPMC is a neutral hydrophilic polymer. The polymer
molecular chains of HPMC hydrate in contact with water
entangle and form a gel matrix. When exposed to water,
carbopol becomes viscous and, thus, tends to bind the mixed
polymeric system together. During hydration process, chan
nels are formed in the matrix networks which are responsible
for drug diffusion. After coming in contact with water,
xanthan gum forms very viscous network. This network is
particularly built up in the drug diffusion channels formed by
465
Table IV indicate that X1, X2, and X3 have signicant effect
on Y3 while magnitude of regression coefcient shows
maximum inuence of X2. Similar observation is presented
as a prediction proler in Fig. 1.
Xanthan gum, HPMC, PEO, and carbopol together
played crucial role in sustaining drug release. Xanthan gum
decreased drug release at initial hour due to its rapid
viscolysing property (Fig. 1). HPMC and PEO were particu
larly responsible for sustaining drug release at later period. So
the drug release was found to be high initially and then
gradually decreased. The diffusional spaces inside the gelling
system are controlled by the molecular weight of the polymer.
Diffusion is the predominant drug release mechanism from
high molecular weight HPMC and PEO matrices which swell
to a higher extent. Swelling phenomenon increases matrix
size; therefore, diffusional path length is increased (28). Drug
entity present in the matrix core may ultimately be requiring
more time to travel towards the matrix surface. This

X1 and X2 on swelling index
polymeric network and ultimately may be responsible for

retarding drug release in initial hour. The channel blockage
might have enhanced with increasing gum level which might
be responsible for decrease in initial hour drug release.
Response surface plot in Fig. 3a shows effect of xanthan
gum (X 1) and rate controlling polymers (X 2) on the
percentage drug release at 1 h (Y2). It can be clearly
interpreted from contour plot in Fig. 3b that for decreasing
drug release at initial hour, higher level of X1 and X2 is
necessary.
Time required for 95% drug release (Y3) was also
increased due to xanthan gum, rate controlling polymers
(HPMC, K100M CR, and PEO), and carbopol 974P. Formu
lation nos. AGT 09 and AGT14 required 9.39 and 3.31 h,
respectively, for 95% drug release. Less time may be required
for Formulation no. AGT 14 as the matrix may not be
capable to sustain drug release for longer time due to low
level of X1, X2, and X3 and the observation was vice versa for
Formulation no. AGT 09 containing high levels of these
formulation variables. Results of regression analysis in

X2 and X3 on bioadhesion
Badhan et al.
466
phenomenon may be responsible for increased time required
for 95% drug release. HPMC and PEO take some time to get
hydrated and swell. As this process is time dependent, drug
release might be sustained in later hours due to these
polymers. Sine the swelling capacity depends on amount of
polymer present in the formulation, as concentration of X2
increased from 1 level to +1 level, time required for 95%
drug release (Y3) was signicantly decreased as shown in
Fig. 1. Effect of combination of xanthan gum (X1) and rate
controlling polymers (X2) on Y3 is shown in response surface
plot (Fig. 4a) and contour plot (Fig. 4b).
Carbopol is a water insoluble but water swellable cross
linked polymer with molecular weight approximately 2
106 Da. Swelling occurs due to the uncharged COOH
group that hydrates by forming hydrogen bonds with the
imbibing water, thus extending the polymer chains. Swelling
of this polymer contributes partially to the oating behavior
of the GRDDS. When exposed to water, carbopol becomes
viscous and, thus, tends to bind the mixed polymeric
system together and reduces erosion of GRDDS (27).
This viscous network ultimately results in sustained drug
delivery phenomenon.
Fluid Uptake Study
The degree of hydration of the polymer is one of the
factors determining the degree and velocity of drug release
from the swellable matrices (29). Mobility of the polymer
chains and, thus, drug diffusion signicantly depends on the
water content of the matrix system. At high water content,
polymer chain relaxation takes place with volume expansion
giving high swelling of the system (30). Xanthan gum, HPMC,
PEO, and carbopol have the property to absorb water and get
hydrated. Thus, percentage of uid uptake depends on the
amount of these components present in the formulation.
Results of the uid uptake study indicate that amount of
xanthan gum has prominent effect on this parameter.
Formulation containing highest amount of xanthan gum
(Formulation no. AGT 20) was having swelling index 466.8
while formulation containing least amount (Formulation no.
AGT 23) has value of 159. Signicance of the effect of
xanthan gum on swelling index can be interpreted from
regression analysis values in Table IV and can be observed
from prediction proler in Fig. 1. This effect may be due to
water holding and viscolyzing property of xanthan gum. In
case of the formulations containing lower amount of xanthan
gum, swelling index values may be less because the matrix
may not be capable to hold water for longer duration.
Maximum swelling index was observed at highest level X1
and X2 (Fig. 5).
Bioadhesion Study
Gastroretention can be achieved by imparting oating
property to the formulation, but to further strengthen this
feature gastric mucoadhesion is also very important. For
introducing this feature, carbopol 974P was added in the
formulation which is widely used as a bioadhesive polymer.
Positive sign and magnitude of regression coefcients in
Table IV indicates signicant inuence of carbopol 974P
(X3) on bioadhesion parameter. The concentration dependent
increase in bioadhesive strength can be clearly observed from

the prediction proler in Fig. 1. Carbopols are commonly
used as mucoadhesives. These polyacrylates interact with
mucus by hydrogen and van der Waals bonds, created
between the carboxylic groups of polyacrylates and the sialic
acid residues of mucin glycoproteins (31). HPMC, a long
chained and nonionic polymer, has also limited bioadhesive
property. It could be due to formation of physical or hydrogen
bonding with the mucus components. Presence of this compo
nent also enhances overall bioadhesion of the formulation.
Response surface plot (Fig. 6a) and corresponding contour
plot (Fig. 6b) show effect of combination of X2 and X3 on
bioadhesive strength. Ultimately, bioadhesive property of the
optimized formulation could assist the tablet to stay in the
upper part of gastrointestinal tract and enhance the gastro
retention along with the oating feature (32).
CONCLUSION
In the present research work, gastroretentive minima
trices, having oating and gastric bioadhesive capabilities,
have been developed. The formulation was optimized by
using central composite design approach. This is very
systematic approach to study inuence of level of various
formulation variables on different response variables. As
minimatrix formulation is multiparticulate drug delivery
system approach, it overcomes drawback of all or none
principle of gastric emptying of single unit drug delivery
system, and also, there is no risk of burst drug release. The
technique implemented for preparation of the minimatrices is
cost effective, and formulation can be scaled up on large scale
using existing tablet manufacturing facility. Hence, the
developed GRDDS may be explored as an effective tool in
the management of H. pylori associated gastric complications
as it has therapeutic as well as manufacturing advantages.
ACKNOWLEDGEMENTS
Authors are thankful to Aristo Pharmaceuticals Ltd. for
providing gift sample of amoxicillin and Colorcon Asia Pvt
Ltd., Signet Chemical Corporation and BF Goodrich Co. for
providing gift sample of excipients.
REFERENCES
1. Rouge N, Buri P, Doelker E. Drug absorption sites in the
gastrointestinal tract and dosage forms for site specic delivery.
Int J Pharm. 1996;136:117 39.
2. Fell JT, Whitehead L, Collet H. Prolonged gastric retention
using oating dosage forms. Pharm Technol. 2000;24:82 90.
3. Nur AO, Zhang JS. Captopril oating and/or bioadhesive
tablets: design and release kinetics. Drug Dev Ind Pharm.
2000;26:965 969.
4. Streubel A, Siepmann J, Bodmeier R. Gastroretentive drug
delivery systems. Expert Opin Drug Deliv. 2006;3:217 33.
5. Conway BR. Drug delivery strategies for the treatment of
Helicobacter pylori infections. Curr Pharm Design. 2005;11:
775 90.
6. Davis SS. Formulation strategies for absorption windows. Drug
Discov Today. 2005;10:249 57.
7. Deshpande AA, Rhodes CT, Shah NH, Malick AW. Controlled
release drug delivery systems for prolonged gastric residence: an
overview. Drug Dev Ind Pharm. 1996;22:531 9.

8. Marshall B. Helicobacter pylori: 20 years on. Clin Med. 2002;2:
147 52.
9. Warren JR. Unidentied curved bacilli on gastric epithelium in
active chronic gastritis. Lancet 1983;1:1273 5.
10. Cooreman MP, Krausgrill P, Hengels KJ. Local gastric and serum
amoxicillin concentrations after different oral application forms.
Antimicrob Agents Chemother. 1993;37:1506 9.
11. Bardonnet PL, Faivre V, Pugh WJ, Piffaretti JC, Falson F.
Gastroretentive dosage forms: Overview and special case of
Helicobacter pylori. J Control Release. 2006;111:1 18.
12. Singh BN, Kim KH. Floating drug delivery systems: an approach
to oral controlled drug delivery via gastric retention. J Control
Release. 2000;63:235 59.
13. Talukdar R, Fassihi R. Gastroretentive delivery systems: a mini
review. Drug Dev Ind Pharm. 2004;30:1019 28.
14. Durig T, Fassihi R. Evaluation of oating and sticking extended
release delivery systems: an unconventional dissolution test. J
Control Release. 2000;67:37 44.
15. Moes AJ. Gastroretentive dosage forms. Crit Rev Ther Drug
Carrier Syst. 1993;10:143 95.
16. Rubinstein A, Friend DR. In: Domb AJ (ed) Polymeric site
specic pharmacotherapy. Wiley, Chichester; 1994, p. 267 313.
17. Klausner EA, Lavy E, Friedman M, Hoffman A. Expandable
gastroretentive dosage forms. J Control Release. 2003;90:143 62.
18. Megraud F, Trimoulet P, Lamouliatte H, Boyanova L. Bacteri
cidal effect of amoxicillin on helicobacter pylori in an in vitro
model using epithelial cells. Antimicrob Agents Chemother.
1991;35:869 72.
19. Bechgard H, Nielsen GH. Controlled release multiple units and
single unit doses. a literature review. Drug Dev Ind Pharm.
1978;4:53 67.
20. Follonier N, Doelker E. Biopharmaceutical comparison of oral
multiple unit and single unit sustained release dosage forms. S.
T. P. Pharma Sci. 1992;2:141 58.
21. De Brabander C, Vervaet C, Fiermans L, Remon JP. Matrix
minitablets based on starch/microcrystalline wax mixtures. Int J
Pharm. 2000;199:195 203.
467
22. Lewis GA, Mathieu D, Phan Tan Luu R. Pharmaceutical
experimental design. New York: Marcel Dekker Inc.; 1999.
23. Li S, Lin S, Daggy BP, Mirchandani HL, Chein YW. Effect of
HPMC and carbopol on the release and oating properties of
gastric oating drug delivery system using factorial design. Int J
Pharm. 2003;253:13 22.
24. Narendra C, Srinath MS, Babu G. Optimization of bilayer
oating tablet containing metoprolol tartrate as a model drug
for gastric retention. AAPS PharmSciTech. 2006;7(2): article 34.
25. Marshall K. Compression and consolidation of powdered solids.
In: Lachman L, Lieberman HA, Kanig JL, editors. The theory
and practice of industrial pharmacy. Mumbai: Varghese; 1987. p.
66 8.
26. Li S, Lin S, Chien YW, Daggy BP, Mirchandani HL. Statistical
optimization of gastric oating system for oral controlled
delivery of calcium. AAPS PharmSciTech 2001;2(1): article 1.
27. Li S, Lin S, Daggy BP, Mirchandani HL, Chien YW. Effect of
formulation variables on the oating properties of gastric
oating drug delivery system. Drug Dev Ind Pharm. 2002;28:
783 93.
28. Miranda A, Millan M, Caraballo I. Study of the critical points of
HPMC hydrophilic matrices for controlled drug delivery. Int J
Pharm. 2006;311:75 81.
29. Michailova V, Titeva ST, Kotsilkova R, Krusteva E, Minkov E.
Inuence of hydrogel structure on the processes of water
penetration and drug release from mixed hydroxypropylmethyl
cellulose/thermally pregelatinized waxy maize starch hydrophilic
matrices. Int J Pharm. 2001;222:7 17.
30. Siepmann J, Peppas NA. Modeling of drug release from delivery
systems based on hydroxypropyl methylcellulose (HPMC). Adv
Drug Del Rev. 2001;48:139 57.
31. Dodou D, Breedveld P, Wieringa PA. Mucoadhesives in the
gastrointestinal tract: revisiting the literature for novel applica
tions. Eur J Pharm Biopharm. 2005;60:1 16.
32. Chavanpatil MD, Jain P, Chaudhari S, Shear R, Vavia PR. Novel
sustained release, swellable and bioadhesive gastroretentive drug
delivery system for ooxacin. Int J Pharm. 2006;316:86 92.

DOI: 10.1208/s12249 009 9232 3
Research Article
Inuence of the Selected Antioxidants on the Stability of the Celsior Solution
Used for Perfusion and Organ Preservation Purposes
Aneta Ostrka-Cielik,1 Barbara Doliska,1,3 and Florian Ryszka2
Received 27 March 2008; accepted 17 March 2009; published online 21 April 2009
Abstract. The purpose of the following research was to improve the original Celsior solution in order to
obtain a higher degree of stability and effectiveness. The solution was modied by the addition of
selected antioxidants such as vitamin C, cysteine, and fumaric acid in the following concentrations: 0.1,
0.3, and 0.5 mmol/l. The solutions stability was estimated using an accelerated stability test based on
changes in histidine concentrations in the solution using Paulys method for determining concentrations.
Elevated temperatures, the factor accelerating substances decomposition reaction rate, were used in the
tests. The research was conducted at four temperatures at intervals of 10C: 600.2C, 700.2C, 800.2C,
and 900.2C. It was stated that the studied substances decomposition occurred in accordance with the
equation for rst order reactions. The function of the logarithmic concentration (log%C) over time was
revealed to be rectilinear. This dependence was used to determine the kinetics of decomposition reaction
rate parameters (the rate constant of decomposition k, activation energy Ea, and frequency factor A). On
the basis of these parameters, the stability of the modied solution was estimated at +5C. The results
obtained show that the proposed antioxidants have a signicant effect on lengthening the Celsior
solutions stability. The best results were reached when combining two antioxidants: vitamin C and
cysteine in 0.5 mmol/l concentrations. As a result, the Celsior solutions stability was lengthened from
22 to 299 days, which is 13.5 times. Vitamin C at a concentration of 0.5 mmol/l increased the solutions
stability by 5.2 times (t90 115 days), cysteine at a concentration of 0.5 mmol/l caused a 4.4 times stability
increase (t90 96 days), and fumaric acid at a concentration of 0.5 mmol/l extended the stability by 2.1
times (t90 48 days) in relation to the original solution.
KEY WORDS: accelerated stability test; antioxidants; ascorbic acid; Celsior; cysteine; fumaric acid.
INTRODUCTION
The Celsior solution is one of the standard solutions used
in perfusion and organ preservation purposes. It has primarily
been utilized in cardioplegia. The solution enables heart
storage for a period of 4 6 h. It can be used to preserve the
heart using the simple hypothermia method exclusively. The
advantages of this method include fast cardiac standstill and a
decrease in cardiac enzyme activity. Celsior is an extracellular
solution with a low potassium concentration and a high
concentration of sodium. Its composition enables the reduc
tion of growth in the concentration of calcium ions in cardiac
muscles, prevents cellular ederma and intracellular acidosis,
as well as protects against the negative effects of free radicals.
Glutamic acid prevents increases in Ca2+ concentrations in
cells. Lactobionate and mannitol exhibit anti edermal
Department of Applied Pharmacy, Medical University of Silesia, 41

205, Sosnowiec, Kasztanowa 3, Poland.
2
Biochefa Pharmaceutical Research and Production Plant, 41 205,
Sosnowiec, Kasztanowa 3, Poland.
3
To whom correspondence should be addressed. (e mail: b.dolinska@
biochefa.pl)
properties, whereas histidine prevents the acidication of

the environment. Apart from histidine, glutathione also acts
as an eliminator of free radicals. The tests conducted show
that Celsior improves the functioning of the right atrium of
the transplanted heart in a signicant manner, restoring heart
action with a potentially lower risk of rejection. The
conducted tests indicate that the solution may also be used
to store the liver, kidneys, and lungs (1 11).
The reperfusion period (the restoration of blood ow to
tissues and organs) leads to oxygen deciency in cells and,
consequently, to irreversible changes in the transplanted
organ. In the course of this process, the transformation of
xanthic dehydrogenase into xanthic oxidase takes place.
Electrons, which have been previously transported by NAD,
will be transported by oxygen using oxidase, which is
associated with the formation of free radicals (12). Antiox
idants are often added to preservative solutions in order to
reduce damage resulting from anoxia and reperfusion (13).
The antioxidizing properties of vitamin C, fumaric acid, and
cysteine were used in the test for histidine stabilization.
Histidine histidine hydrochloride forms a strong buffer
system. This amino acid shows great buffer capacity when
compared to other buffers used in solutions. It is also a
substance of relatively signicant stability. The energy of
activation value is approximately 20,000 cal/mol (4).
468
Celsior Solution Used for Perfusion/Organ Preservation Purposes

Glutathione (GSH) is one of the most important cellular
antioxidants. Together with its oxidized forms (glutathione
disulde, GSSG), glutathione peroxidase and glutathione
reductase, it eliminates both free and lipid radicals, protecting
cellular membranes, DNA, and proteins against oxidative
stress and xenobiotics (14,15). It acts as a hydrosulde buffer
and is indispensable to the maintenance of correct thiolic
redox potential by means of sustaining hydrosulde groups
(SH) of proteins in reduced form. Reduced glutathione is
necessary to maintain the proper structure of erythrocytes
and to keep hemoglobin in non oxidized form. Cells with an
insufcient volume of glutathione undergo hemolysis. During
the period of organ storage and reperfusion, a decrease in the
concentration of reduced GSH and an increase in the volume
of its reduced form (GSSG) in cells have been observed
(16,17); therefore, its supplementation is very signicant. The
introduction of glutathione as a component of the solution
contributes to a signicant increase in manufacturing costs,
which is additionally accompanied by the issue of low cellular
absorption of glutathione (17). It was also proven that low
concentrations of glutathione in the transplanted organs
engender an immunosuppressive effect (14). The replacement
of glutathione by cysteine may be a good solution. This
results from the fact that the basic substrates of GSH
synthesis are three amino acids: L glutamic acid, L cysteine,
and glycine (14). The antioxidizing properties of cysteine are
related to the opportunity of complexing metal ions and,
therefore, protecting against hazardous Fenton reactions. It
also has a radioprotective effect and protects lymphocytes
against chromosome aberrations (17). In high concentrations,
vitamin C is also a powerful antioxidant. It is classied as a
water soluble antioxidant. Vitamin C forms a strong oxidizing
and reducing system: L ascorbic acid/dehydro L ascorbic acid
with the potential of enabling the reduction of inactive forms
of oxygen. It neutralizes hydroxylic, oxyalcohol, superoxide,
and nitric radicals (18 27). The application of fumaric acid in
cardoplegia improves myocardial contractility and decreases
the concentration of lactate dehydrogenase, which is corre
lated to increases in the number of dead cells at higher levels
of activity. Its protective impact is related to the transforma
tion of fumaric acid into succinic acid via a dehydrogenation
reaction in Krebs cycle (28). It has also been shown to
positively impact the restoration of proper function in
immature myocytes (29).
The introduction of antioxidants as solution components
may signicantly reduce adverse changes in stored organs.
Taking into consideration patients waiting time, it is also
recommended that the manufactured perfusion and preserva
tion solutions are of the longest possible stability and, therefore,
of greatest effectiveness. The introduction of antioxidants into
the solutions may signicantly increase their stability, thereby
contributing to a reduction in manufacturing costs.
The aim of the study was to determine the inuence of
selected antioxidants on the stability of the Celsior solution.
The accelerated stability test, enabling the rapid determination
of solution stability, as well as the evaluation of stabilizers
applied at different concentrations, was used to analyze
stability. This method is based on quantitative dependence
between the reaction rate and temperature, which can be
evaluated using the laws of chemical reaction kinetics.
469

The following were used for test purposes: the Celsior
solution [prepared in accordance with the original com
position of SangStat Medical Corporation USA (30)] with
the following composition: histidine, 30 mmol/l; mannitol,
60 mmol/l; lactobionic acid, 80 mmol/l; glutamic acid,
20 mmol/l; reduced glutathione, 3 mmol/l; Na+, 100 mmol/l;
K+, 15 mmol/l; Mg2+, 13 mmol/l; Ca2+, 0.25 mmol/l; Cl ,
41.25 mmol/l; solution pH 7.20 7.40; osmolarity 320
360 mOsm/l. All reagents in the original solution, as well as
the applied antioxidants (vitamin C, cysteine hydrochloride,
and fumaric acid), were purchased from Sigma (St. Luis, MO,
USA). All other chemicals were of analytical grade.
The solution was subsequently modied by replacing
glutathione with the selected antioxidants, ascorbic acid,
cysteine, and fumaric acid, in the following concentrations:
0.1, 0.3, and 0.5 mmol/l, respectively. The solution was
prepared in such a manner that the components, in the
amounts from least to greatest, have been dissolved in
approximately 700 ml of water pro injection. pH values of
7.20 7.40 in the prepared solution were achieved by adding
1 M solution of NaOH and adding 1,000 ml of water pro
injection. The solutions were then ltered using a Sarto
pure GF 2 sterilizing inltration trench made of cellulose
acetate with pore sizes of 0.22 m at a rate of 50 ml/min. In
the next stage, the solution was poured into sterilized 100 ml
glass bottles and closed with sterilized gum stoppers
equipped with caps. The bottles were made of rst class
transparent hydrolytic glass. The hydrolytic resistance of
internal surfaces is of HC 1 class. The solutions were
prepared in compliance with the GMP principles in a
laminar airow cabinet.
The modied solutions were prepared analogically, with
a single difference consisting in the addition of an antioxidant
at a relevant concentration at the beginning and adding the
remaining components after the said antioxidant dissolves.
The physical and chemical properties of the prepared
solutions were analyzed. Measurements of pH values were
made using a microcomputer pH meter of CP 315 model,
with combined electrode, manufactured by Elmetron
(Poland). The accuracy of measurements amounted to
0.01 pH. Density measurements were made on the basis of
determining the masses of three solution samples of identical
capacity. The viscosity of the tested solutions was determined
using a method based on Stokes law, the so called falling ball
method. A rotatory Hppler viscosimeter, model VEB MLW
Prfgerte Werk (Germany), was used for test purposes. The
light refraction coefcient was measured using Abbys
refractometer, RL1 type, manufactured by PZO (Poland).
The measurement range, nd, amounted to 1.3 1.7. The
accuracy of the refraction coefcient measurement was in
the range of 1.3 1.42, amounting to 0.0004, whereas in the
1.42 1.7 range, it amounted to 0.0002. Buffer capacity was
determined by the number of HCl hydrogen ion moles
indispensable to change the pH value of 1 ml of solution by
1 unit. Osmolarity measurements were made using a 800 cl
osmometer manufactured by the Trident Med Company
(Poland). The measurement accuracy of the osmometer
amounted to 1 mOsm/kg H2O (0.4%). Solution color was
470
Ostrka-Cielik, Doliska and Ryszka
determined for each series before and after an accelerated

senescence test. The color was determined in daylight, in glass
test tubes placed on a white matted background. The solution was
deemed colorless if no change of color compared to a test tube
containing distilled water, used as a control liquid, was observed.
Determination of Histidine Content in Celsior Solution Using
Paulys Method
An analysis of histidine in the course of processing
chemical reactions during the accelerated stability test was
carried out using Paulys spectrophotometric method (31).
Test tubes were lled with 1 ml of histidine working solution
at concentrations of 5, 10, 20, 30, 35, 40, 50, and 60 g/ml and
then lled with 2 ml of sulfonic acid and 1 ml of 5% sodium
nitrite solution; the solution was mixed and 3 ml of 10%
Na2CO3 was subsequently added. The absorption of the
resulting azo complex was measured using =530 nm wave
length with a blind trial as reference. The calibration curve of
the dependence between absorption and the concentration of
histidine, dened by the equation y = 0.120x, was then
produced. The correlation coefcient amounted to R2 =
0.999, whereas the linear regression error equaled 0.000753.
The accuracy of the measurement of 0.63% was positively
evaluated on the basis of standard deviation values, the
relative standard deviation, as well as the condence interval
for the average value. Method sensitivity amounted to 2.27
104 l mol 1 cm 1. The measurements were made in glass
cuvettes of 1 cm thickness using a CECIL 3021 UV vis
spectrophotometer manufactured by the Chemist Instruments
Company (Poland). The spectrophotometers photometric
accuracy amounted to 0.005 A.
Substances used in the determination of histidine were of
analytical grade and derived from Sigma and PPH POCh S.A.
Gliwice (Poland).
Solution Stability Analysis Using Accelerated Stability Test
Solution stability was determined on the basis of the
kinetics of the rate of histidine content change at raised
temperatures (32 34). The examinations were preceded by a
unitary packaging tightness test by immersing them in a 1%
water solution of methylene blue. Five unitary packages

containing the relevant liquid were placed in a ultrathermo
stat (U7; Remontom, Poland) lled with water (the degree to
which a constant temperature was maintained amounted
0.1C). The test was conducted at four temperatures at
intervals of 10C, i.e., 600.2C (333 K), 700.2C (343 K),
800.2C (353 K), and 900.2C (363 K). The temperature
range was determined on the basis of initially conducted tests
of the rate of histidine degradation. Sampling was started at
the moment of temperature equalization of the tested
solution with an ultrathermostat temperature (time t=0,
assumed content of determined substance is 100%). The
samples taken were cooled in ice water directly after being
removed in order to prevent further degradation. The tests
were continued until more than 50% of the degradation of
the determined substance was obtained.
Determination of Order of Histidine Degradation
The reaction order was dened on the basis of the results
of analytically determining the histidine concentration in the
sampled solutions after the accelerated stability test. It was
stated that the degradation of the examined substances was
compliant with the rst order reactions equation. Graphs of
the concentration logarithm (log%C) over time are rectilin
ear. The above dependence was used to calculate the rate
constant.
Determination of Kinetic Parameters of Degradation Rate
Rate constants (k) were calculated for the rst order
reaction at four different temperatures using the following
equation (32,33):
k 2:303 lg C0 =C0
Cx =t

min1
in which C0 is the output content of the degraded substance

determined over time t=0 (100%) and Cx is the content of
determined substance after a specic period of time.
The results obtained for each temperature were pre
sented as an average of ve results (k). Furthermore, the
standard deviation was calculated (SD).
Table I. Reaction Rate Constants for Histidine Degradation in Solutions at Various Temperatures
k 10
(min 1)
Solution
60 (C)
70 (C)
80 (C)
90 (C)
Celsior
+0.1 (mmol/l) vit. C
+0.1 [mmol/l] cysteine
+0.3 (mmol/l) cysteine
+0.5 (mmol/l) cysteine
+0.5 (mmol/l) vit. C +0.5 (mmol/l) cysteine
+0.1 (mmol/l) fumaric acid
1.8600.016
1.4300.018
1.0600.013
0.8500.009
1.5200.023
1.1800.013
0.9200.010
0.4500.006
1.7300.021
1.4800.022
1.3100.016
3.3200.017
2.6600.014
2.1000.010
1.7300.010
2.8400.017
2.3200.016
1.8700.013
0.9700.006
3.1300.022
2.7500.022
2.4800.020
5.8400.018
4.7900.011
4.0040.016
3.4200.015
5.0800.014
4.3900.016
3.6500.011
2.0100.016
5.4500.014
4.9700.021
4.6200.018
9.9900.012
8.3300.026
7.4700.014
6.5300.017
8.7300.026
8.1900.015
6.8400.014
3.9900.017
9.4500.027
8.7700.020
8.3200.020
471
On the basis of the obtained results (k), the graphs of lnk

dependence versus reciprocal absolute temperature were
made. It was stated that the dependence lnk=f(1/T) is a linear
function, allowing for the assumption that in the said cases, the
Arrhenius equation in logarithm form was complied with:
Ea 1
ln A:
R
T
ln k
The results obtained conrm the thesis that within the

temperature range analyzed, the activation energy (Ea) and
frequency coefcient (A) may be considered as constant
values.
For linear dependence lnk=f(1/T), the straight gradient
coefcient (a) and intersection point with ordinate axis (b)
was calculated using the linear regression method, as well as,
based on it, the activation energy (Ea) and frequency
coefcient (A).
ln k a
Ea
1
b
T
n
n
P
i 1
ln ki
1
Ti
n
n 2
P
1
n
b ln A
ln ki
i 1
n 2
P
1
Ti
i 1
n
n
P
i 1
n
P
i 1
1
Ti
2
1
Ti
Ti
i 1
n
P
Ea 1
ln A
R
T
n
P
i 1
n
P
i 1
1
Ti
n
P
i 1
1
Ti
n
P
i 1
ln ki
5
1
Ti
i 1
v
u n 2
u1 X 1
b a t
:
n i 1 Ti
Ea 1
ln A:
R
T
Time, in which 10% of the examined substance will be

degraded at a temperature of 5C, was calculated according
to the following equation:
t90 0:1053=k278 :
1
Ti
v
u
n
n
n
P
P
P
u
1
1
ln ki 2 a
b
u
Ti ln ki
Ti
u n
i
1
i
1
i
1
6
a u
n 2
un 2
n 2
P1
P
t
1
n
Ti
Ti
i 1
Reaction rate constant at a temperature of 5C was calculated

using the determined Ea and A values:
ln k278
ln ki
i 1
2
2
n
P
Fig. 2. Arrhenius plot for the rst order rate constant of histidine
degradation in Celsior solution and in Celsior solution modied by
adding cysteine over the temperature range of 60 90C
Because the calculated values are a function of the other

measured values, error t90 was determined using the total
differential method.

n
X
f xi

f x1 ; x2 ; ::; xn
x xi
i 1

t90

k 0:1053 k
t90

2
k
k

k
k
0:1053

a b
k2
a
b
10
11
adding vitamin C over the temperature range of 60 90C
adding fumaric acid over the temperature range of 60 90C
472

Table II. Activation Energy (Ea) and Frequency Coefcient (A) Determined for Solutions
Solution
Celsior
+0.1 (mmol/l)
+0.3 (mmol/l)
+0.5 (mmol/l)
+0.1 (mmol/l)
+0.3 (mmol/l)
+0.5 (mmol/l)
+0.5 (mmol/l)
+0.1 (mmol/l)
+0.3 (mmol/l)
+0.5 (mmol/l)
vit. C
vit. C
vit. C
cysteine
cysteine
cysteine
vit. C + 0.5 (mmol/l) cysteine
fumaric acid
fumaric acid
fumaric acid

1
1
1
ea T b a ea T b b
T

0:1053
1

a b :
k
T
t90
0:1053

k2
Ea (kJ mol 1)
lnA
56.300.16
59.100.14
65.600.14
68.400.14
58.500.18
65.000.17
67.100.13
73.000.16
56.800.20
59.800.17
62.100.16
11.700.05
12.500.05
14.500.05
15.300.05
12.300.06
14.400.06
15.000.05
16.400.06
11.900.07
12.800.06
13.500.06
increases for the Celsior solution, ranging between 1.71 and

2.15. It was also stated that the temperature coefcient,
changing together with the temperature, has greater values at
lower temperature ranges than at higher temperature ranges
and grows together with the increase in the concentration of
antioxidants.
The determined reaction rate constants have enabled
settling of their dependence on temperature. Graphic pre
sentation of this dependence in the form of lnk=f(1/T) is
presented in Figs. 1, 2, and 3. The observed linear depen
dence between the reaction rate constant and temperature is
determined by the Arrhenius equation. Deviations from the
linear relation of the lnk to 1/T function occurred in many
studies (34). Its image is a straight Ea/RT gradient, which
intersects the ordinate axis at point lnA where 1/T=0. For
linear dependence lnk=f(1/T), the straight gradient coef
cient (a) and intersection point with the ordinate axis (b)
were calculated using the linear regression method, in
addition to, based on it, the activation energy (Ea) and
frequency coefcient (A). The determined Ea and lnA values
are presented in Table II. Activation energy values express
the stability of the examined solutions. The highest values of
this parameter are exemplied by the Celsior solution
modied by adding vitamin C and cysteine at concentrations
of 0.5 mmol/l. The resulting value of the activation energy
amounts to 73 kJ mol 1. The lowest activation energy value
(56.3 kJ mol 1) was obtained by adding cysteine at a
12

Based on the analyses conducted using Paulys method,
both the decrease in the histidine content of the Celsior
solution and the order of the reaction were determined.
Initial concentrations of histidine and reducing compounds in
the examined solutions were assumed to be 100%, whereas
subsequent concentrations were expressed as a percentage of
the initial value. On the basis of the obtained results, the
graphs of the logarithm of concentration dependence (log%
C) of non degraded substances over time were made. Trans
formations at all of the examined temperatures are congruent
with the rst order reaction equations. Table I presents the
calculated degradation rate constants (k) for rst order
reactions at four different temperatures. The obtained results
for each temperature were presented as an average of ve
results (k) together with the standard deviation. The chemical
reaction rate resulting from the vant Hoff equation (propor
tion of chemical reaction rate after temperature change and
rate of the same reaction before temperature change)
Table III. The Stability of Celsior Solution and Modied Solution at 20C and 5C
20 (C)
Solution
Celsior
+0.1 (mmol/l)
+0.3 (mmol/l)
+0.5 (mmol/l)
+0.1 (mmol/l)
+0.3 (mmol/l)
+0.5 (mmol/l)
+0.5 (mmol/l)
+0.1 (mmol/l)
+0.3 (mmol/l)
+0.5 (mmol/l)
vit. C
vit. C
vit. C
cysteine
cysteine
cysteine
vit. C + 0.5 (mmol/l) cysteine
fumaric acid
fumaric acid
fumaric acid
k 10
(min 1)
11.551.38
7.740.81
4.170.44
2.900.30
8.541.16
4.750.60
3.360.33
1.230.15
10.431.54
7.711.02
6.060.73
5 (C)
t90 (days)
6.30.8
9.51.0
17.61.9
25.22.6
8.61.2
15.41.9
21.72.1
59.37.3
7.01.0
9.51.2
12.11.5
k 10
(min 1)
3.300.41
2.100.23
0.980.11
0.640.07
2.300.33
1.100.15
0.760.08
0.250.03
3.000.45
2.100.28
1.500.19
t90 (days)
22.02.7
35.03.8
75.08.1
114.712.2
31.34.4
65.08.4
96.29.8
298.537.7
24.73.8
35.74.8
47.85.9
+0 5 (mmol/l) vit C +0 5
(mmol/l) cysteine
+0 5 (mmol/l) fumaric acid
+0 5 (mmol/l) cysteine
+0 5 (mmol/l) vit C
+0 3 (mmol/l) vit C
+0 1 (mmol/l) vit C
Celsior
Solution
Before test
After test
Before test
After test
Before test
After test
Before test
After test
Before test
After test
Before test
After test
Before test
After test
Before test
After test
Before test
After test
Before test
After test
Before test
After test
Stability test
7 30
6 98
7 30
6 98
7 20
6 84
7 20
6 77
7 40
7 33
7 22
7 03
7 20
6 61
7 20
6 82
7 20
6 84
7 20
6 77
7 40
7 33
pH
1 013
1 000
1 020
0 993
1 013
0 993
1 013
0 993
1 013
1 013
1 066
1 013
1 027
1 013
1 020
1 020
1 020
1 020
1 013
1 006
1 013
1 013
Density
(g/ml)
1 3417
1 3420
1 3420
1 3419
1 3429
1 3420
1 3417
1 3420
1 3420
1 3415
1 3415
1 3415
1 3415
1 3415
1 3415
1 3415
1 3415x
1 3415
1 3415
1 3411
1 3420
1 3415
Light refraction
coefcient
333
332
339
337
337
333
333
332
341
326
342
338
340
337
340
329
332x
329
332
325
341
326
Osmolarity
(mOsm/l)
0 019
0 019
0 019
0 021
0 023
0 021
0 022
0 025
0 021
0 022
0 023
0 021
0 021
0 021
0 022
0 022
0 021
0 019
0 021
0 023
0 021
0 022
Buffer capacity
(mol/l)
Table IV. Physical and Chemical Properties of the Prepared Solutions after the Accelerated Stability Test at 90C
1 112
1 102
1 095
1 079
1 068
1 062
1 102
1 101
1 201
1 157
1 286
1 090
1 112
1 067
1 106
1 043
1 128
1 073
1 179
1 096
1 201
1 157
Viscosity
(Pa s)
Colorless
Yellow strain
Colorless
Yellow strain
Colorless
Yellow strain
Colorless
Yellow strain
Colorless
Yellow strain
Colorless
Yellow strain
Colorless
Yellow strain
Colorless
Yellow strain
Colorless
Yellow strain
Colorless
Yellow strain
Colorless
Yellow strain
solution color

473
474
concentration of 0.1 mmol/l to the Celsior solution. The
calculated values are contained in the 50 90 kJ mol 1 range,
which reects the average for most medicinal products.
The presence of constant Ea and A values in the examined
temperature scope allows for the determination of reaction
rate constants at the storage temperature of the tested
solutions, as well as the determination of t90 degradation times.
The calculated values have been presented in Table III. The
chemical stability of an active substance constitutes the core
element of tests of the stability of medicinal products.
However, it is still important for the medicinal products
physical and chemical properties to remain unchanged. Table
IV contains the physical and chemical properties of the Celsior
solution as well as the properties of modied solutions
following accelerated stability tests at a temperature of 90C.
At this temperature, the greatest changes in the analyzed
parameters are to be assumed. pH, density, buffer capacity,
osmolarity, light refraction coefcient, and viscosity values
were examined. These values did not change signicantly after
conducting the accelerated stability test. The yellow strain
color observed after the test may be a result of the Maillard
reaction (35) taking place between the rst order amine groups
(cysteine, histidine) and reducing sugars (emerging in liquids
via hydrolysis). The change in the solutions color proves the
lowered effectiveness of such solutions.
Having analyzed the data presented in Tables I, II, and
III, it may be stated that the addition of an antioxidant has
a signicant impact on the reaction rate constant. The
lowest k value at a storage temperature of 5C for the
Celsior solution and its modied versions was observed in
the original solution with the addition of vitamin C and
cysteine at 0.5 mmol/l (k= 0.25 10 6 min 1) concentrations.
The determinant of the solutions stability at a given
temperature is given by the t90 value. Having analyzed the
data from Table III, it may be stated that a combination of
two optimal antioxidants, ascorbic acid and cysteine (at
0.5 mmol/l concentrations), resulted in extending the
Celsior solutions stability from 22 to 299 days, provided
that the physical and chemical parameters are not changed
signicantly (pH 7.33, osmolarity 326 mOsm/l). The remaining
antioxidants also extended the solutions stability (Table III).
Vitamin C at a concentration of 0.5 mmol/l increased stability
by 5.2 times (t90 =115 days). Cysteine at a concentration of
0.5 mmol/l resulted in increasing the stability by 4.4 times (t90 =
96 days), whereas fumaric acid at a concentration of 0.5 mmol/
l increased stability by 2.1 times (t90 =48 days) compared to the
original solution. The antioxidizing properties of the selected
antioxidants decrease in the following order: vitamin C >
cysteine > fumaric acid. This dependence results from the
redox potential of reactions present in the solution during the
storage period.
Schemes of the chemical reactions present during the
course of the tests have been dened. Histidine in the Celsior
solution may be present in two tautomeric forms. Therefore,
we may assume the presence of a tautomerization reaction.
Its structure contains an imidazole ring which participates in
four types of reactions: as an electrophile, giving and
receiving electron pairs; it may also additionally provide a
single electron or react chemically with two chemical
compounds at the same time. This property arises from
molecular orbital theory and is related to the presence of two

nitrogen atoms (secondary pyrrole and tertiary pyridine) of
different physical and chemical potentials (35). In the
examined solutions, histidine is most likely subject to
decarboxylation reactions, transforming into histamine [ (4
imidazol) ethylamine] (36,37).
Vitamin C is the most efcient antioxidant. It is capable
of reacting with metal cations contained in liquids, generating
monodent, and dissolving complex compounds of relatively
low stability. The lactone ring and side chain are engaged in
the complexion process. They initially dissociate the protons
from the lactone ring at the O 3 position and, consequently at
the O 2 position, which is related to the presence of C(2)
OHOC(1) hydrogen bonding, hindering dissociation of the
OH(2) group proton. As a consequence, the following
compounds are formed: K(HAs)2H2O, Na(HAs)2H2O, Mg
(HAs)24H2O, and Ca(HAs)24H2O. The complexation reac
tion has a stabilizing effect on the examined solutions, as
lower concentrations of cations prevent precipitation. Cyste
ine and histidine exhibit similar properties and form soluble
complex compounds with K+, Na+, Mg2+, and Ca2+ ions
(38,39).
We were unable to nd information in the available
literature on antioxidants effects on organ perfusion and the
preservation of solution stability. The conducted tests on
HTK solution stability, which depend on the applied anti
oxidants (40), conrm the hypothesis that antioxidants have a
signicant impact on extending the stability of solutions for
perfusion and organ preservation purposes. An HTK solution
stored at +5C is stable for 260 days with no antioxidants,
450days with vitamin C, 265days with cysteine, and 242
days with fumaric acid (40).
CONCLUSIONS
Antioxidants have a signicant impact on the stability of
solutions used for perfusion and organ preservation. The
conducted laboratory tests prove that a combination of two
optimal antioxidants, ascorbic acid and cysteine (0.5 mmol/l
concentrations), resulted in an increase in the Celsior solutions
stability from 22 to 299 days. It is our belief that a modied
Celsior solution manufactured under industrial conditions will
exhibit greater stability.
ACKNOWLEDGMENTS
This study was funded by the State Committee for
Scientic Research, Warsaw, Poland (grant PBZ KBN 048/
PO5/2001).
REFERENCES
1. Budziski G, Cierpka L. Methods of storing organs for
transplantation purposes. In: Smorg Z, Somki R, Cierpka
L, editors. Biotechnological and medical foundations of
xenotransplantation. Poznan: Orodek Wydawnictw Nauko
wych; 2006. p. 279 90.
2. Kosieradzki M, Danielewicz R. Ischemic damage of organs and
their storage. In: Rowiski W, Waaszewski J, Pczka L, editors.
Clinical transplantology. Warsaw: Wydawnictwo Lekarskie
PZWL; 2004. p. 105 21.

3. Ryszka F, Doliska B, Ostrka Cielik A. Solutions for organs
storage purposes and evaluation of their efciency. In: Smorg Z,
Somski R, Cierpka L, editors. Biotechnological and medical
foundations of xenotransplantation. Poznan: Orodek Wydaw
nictw Naukowych; 2006. p. 291 303.
4. Ryszka F, Ostrka Cielik A, Doliska B. Impact of solutions
components for organs storage purposes on their stability and
biochemical properties. Biotechnologia 2006;1(72):97 102.
5. Roberts RF, Nishanian GP, Carey JN, Sakamaki Y, Starnes VA,
Barr ML. A comparison of the new preservation solution Celsior
to EuroCollins and University of Wisconsin solutions in lung
reperfusion injury. Transplantation 1999;67(1):152 5.
6. Xiong L, Legagneux J, Wassef M, Oubenaissa A, Detruit H,
Mouas C, et al. Protective effects of Celsior in lung transplanta
tion. J Heart Lung Transplant 1999;18:320 7.
7. Janen H, Janen PH, Broelsch CE. Celsior solution compared
with University of Wisconsin solution (UW) and histidine
tryptophan ketoglutarate solution (HTK) in the protection of
human hepatocytes against ischemia reperfusion injury. Transpl
Int 2003;16:515 22.
8. Howden BO, Jablonski P. Liver preservation: a comparison of
Celsior to colloid free University of Wisconsin solution. Trans
plantation 2000;70:1140 2.
9. Carrier M, Trudel S, Pelletier LC. Effect of Celsior and
University of Wisconsin solutions on myocardial metabolism
and function after warm ischemia. J Cardiovasc Surg (Torino).
1999;40(6):811 16.
10. Mohara J, Takahashi T, Oshima K, Aiba M, Yamagishi T,
Takeyoshi I, et al. The effect of Celsior solution on 12 hour
cardiac preservation in comparison with University of Wisconsin
solution. J Cardiovasc Surg (Torino). 2001;42(2):187 92.
11. Abrahamse STL, Dinant S, Pfaffendorf M, Gulik TM. In vitro
function of porcine carotid arteries preserved in UW, HTK and
Celsior solutions. Fund Clin Pharmacol 2002;16(6):503 11.
12. Wodek L. Oxygen reactive species in physiological and patho
logical conditions. Cellular antioxidant systems. Farm Pol
2004;9:404 18.
13. Southard JH, Den Butter B, Marsh DC. The role of oxygen free
radicals in organ preservation. Klin Wochenschr 1991;21:1073 6.
14. Bilska A, Kryczyk A, Wodek L. The different aspects of the
biological role of glutathione. Post Hig Med Dosw 2007;61:
438 53.
15. Wu G, Fang YZ, Yang S, Pluton JR. Glutathione metabolism
and its implications for health. J Nutr. 2004;134:489 92.
16. Markuszewski L, Okoski P, Banach M, Wierzbiski P,
Pietruszyski R. Role of oxidative stress and reactive oxygen
species in pathogenesis of stunning myocardium. Glutathione
as a substance which prevents stunning. Folia Cardiol. 2006;13
(1):9 18.
17. Wodek L. Biothols in physiological and pathological conditions
and in therapy. Krakow: Wydawnictwo Uniwersytetu Jagiello
skiego; 2003. p. 7 15.
18. Guz J, Dziaman T, Szpila A. Do antioxidant vitamins inuence
carcinogenesis process? Post Hig Med Dosw. 2007;61:185 98.
19. Konopacka M. Function of vitamin C in oxidant damages of
DNA. Post Hig Med Dosw. 2004;58:343 8.
475
20. Sies H, Stahl W, Sundquist AR. Antioxidant functions of

vitamins. Vitamins E and C, beta carotene, and other carote
noids. Ann NY Acad Sci. 1992;669:7 20.
21. Padayatty SJ, Katz A, Wang Y, Eck P, Kwon O, Lee J, et al.
Vitamin C as an antioxidant: evaluation of its role in disease
prevention. J Am Coll Nutr. 2003;22:18 35.
22. Sroka Z, Gamian A, Cisowski W. Low molecular antioxidant
compounds of natural origin. Post Hig Med Dosw 2005;59:34 41.
23. Lloberas N, Torras J, Herrero Fresneda I, Cruzado JM, Riera M,
Hurtado I, Grinyo JM. Postischemic renal oxidative stress
induces inammatory response through PAF and oxidized
phospholipids. Prevention by antioxidant treatment. FASEB J
2002;16:908 10.
24. Deutsch JC. Ascorbic acid oxidation by hydrogen peroxide. Anal
Biochem 1998;255:1 7.
25. Kulikowska Karpiska E, Moniuszko Jakoniuk J. The antiox
idative barrier in the organism. Pol J Environ Stud 2004;13
(1):5 13.
26. Ziemlaski , Wartanowicz M. Function of nutritional antiox
idants in health and disease. Pediatr Wspcz Gastroenterol
Hepatol ywienie Dziecka 1999;1:97 105.
27. Piekutowski K, Roszkowski K. Free oxygen radicals. Protective
function of antioxidant vitamins in tumors prevention and
treatment. Wspcz Onkol 1999;4:143 4.
28. Laplante A, Vincent G, Poirier M, Rosiers C. Effects and
metabolism of fumarate in the perfused rat heart; A 13C mass
isotopomer study. Am J Physiol 1997;272:E74 82.
29. Pearl JM, et al. Fumarate enriched blood cardioplegia results in
complete functional recovery of immature myocardium. Ann
Thorac Surg 1993;57:1636 41.
30. Sangstat Medical Corporation; 1999. http://www.sangstat.com/
docs/celsior pl.pdf. Accessed June 22, 2004.
31. Alexander RJ. A laboratory manual of analytical methods of
protein chemistry. New York: Pergamon; 1970. p. 234 56.
32. Shi L, Schoeld T. Pharmaceutical stability testing conference.
Expert Opin Drug Saf 2004;3:153 8.
33. International Conference on Harmonization (ICH). Guidance
for Industry. Q2B. Validation of Analytical Procedures: Meth
odology. 1996. Available at http://www.fda.gov/cder/guidance/
index.htm. Accessed June 25, 2003.
34. Truhlar DG, Kohen A. Convex Arrhenius plots and their
interpretation. PNAS. 2001;98(3):848 51.
35. Joule JA, Smith GF. Chemistry of heterocyclic compounds.
Warsaw: PWN; 1984. p. 43.
36. Stanosz M, Mach Szczypiski J, Stanosz S. Biochemical and
pharmacological aspects of histamine. Gin Prakt. 2005;86
(5):37 42.
37. Tomasik P. Mechanism of organic reactions. Warsaw: PWN;
1998. p. 26.
38. Bartosz G. Second nature of oxygen. Warsaw: PWN; 1995. p. 27.
39. Ciszewska Jdrasik M, Pertkiewicz M. Mixtures for perentenal
nutrition purposes. Warsaw: PZWL; 2004. p. 12 27.
40. Doliska B, Ryszka F, Ostrka Cielik A. The effect of selected
antioxidants on the kinetics of changes in the stability of an HTK
Solution. A technical note. AAPS PharmSciTech 2006;7(2)
Article 51.

DOI: 10.1208/s12249 009 9233 2
Research Article
Microemulsion-Based Vaginal Gel of Clotrimazole: Formulation, In Vitro
Evaluation, and Stability Studies
Yogeshwar G. Bachhav1 and Vandana B. Patravale1,2
Abstract. The objective of the present investigation was to develop and evaluate microemulsion based gel for
the vaginal delivery of clotrimazole (CMZ). The solubility of CMZ in oils and surfactants was evaluated to
identify components of the microemulsion. The ternary diagram was plotted to identify the area of
microemulsion existence. Various gelling agents were evaluated for their potential to gel the CMZ
microemulsion without affecting its structure. The bioadhesive potential and antifungal activity of the CMZ
microemulsion based gel (CMZ MBG) was determined in comparison to the marketed clotrimazole gel
(Candid V gel) by in vitro methods. The chemical stability of CMZ in CMZ MBG was determined as per
the International Conference on Harmonization guidelines. The CMZ microemulsion exhibited globule size
of 48.4 nm and polydispersity index of 0.75. Carbopol ETD 2020 could successfully gel the CMZ
microemulsion without disturbing the structure. The CMZ MBG showed signicantly higher (P<0.05) in
vitro bioadhesion and antifungal activity as compared to that of Candid V gel. The stability studies indicated
that CMZ undergoes acidic pH catalyzed degradation at all the storage conditions at the end of 3 months.
KEY WORDS: clotrimazole; microemulsion; microemulsion based vaginal gel; stability studies; vaginal delivery.
INTRODUCTION
Vulvovaginal candidiasis is one of the most common
gynecological disorders. Approximately 75% of women experi
ence vulvovaginal candidiasis during their life and about 40% to
50% of them suffer from multiple episodes (1,2). For the
treatment of vulvovaginal candidiasis, local treatment with
antifungal agents is a rst resort. The local (vaginal) delivery
not only gives site specic treatment but also avoids toxic side
effects of antifungal agents that are encountered on oral
administration. The most commonly prescribed treatment for
vaginal candidiasis has been the topical application of clotrima
zole (CMZ), an imidazole antifungal agent. CMZ is known to be
very effective locally and presents no major side effects (3).
Clotrimazole is available in several conventional dosage forms
such as creams, gels, pessaries, and ovules for vaginal applica
tion. However, these conventional dosage forms do not offer
prolonged duration of action which compromises the efcacy of
the CMZ (4). Furthermore, poor water solubility of CMZ
(0.49 g/ml) (5) also presents a hindrance for the local
availability of CMZ and limits the effective antifungal therapy.
In order to overcome these disadvantages, several delivery
strategies such as liposomes (4), microspheres (6), sustained
release bioadhesive tablets (7), and polycarbophil gels (8) have
been proposed for the vaginal delivery of CMZ. However,
potential of microemulsions has not been explored for the
delivery of CMZ.
1
Department of Pharmaceutical Sciences and Technology, Institute

Chemical Technology, Matunga, Mumbai 400019, India.
2
To whom correspondence should be addressed. (e mail: vbpatravale@
udct.org)
Microemulsions have demonstrated a great potential for

improving the systemic and local bioavailability of an array of
hydrophobic therapeutic agents (9 11). In view of this, the
solubilization of CMZ in microemulsions would improve its
vaginal availability. However, as discussed earlier, it is also
essential to have a dosage form which adheres to the vaginal
mucosa and increases the residence time of CMZ in vagina.
This functionality can be imparted by gelling of the CMZ
microemulsion using bioadhesive agent. The success of
mucoadhesive microemulsions has already been described in
the literature for the nasal delivery (12). Thus in the present
investigation, formulation of CMZ microemulsion based gel
(CMZ MBG) was attempted for vaginal delivery. The devel
oped CMZ MBG was evaluated for in vitro release, in vitro
bioadhesive time and in vitro antifungal activity. Furthermore,
the chemical stability of CMZ in the formulated gel was
evaluated as per the International Conference on Harmoniza
tion (ICH) guidelines.
Materials
Clotrimazole was kindly gifted by Cipla Pharmaceuticals
Ltd., Mumbai, India. Cremophore EL (polyoxyl 35 castor oil),
Solutol HS 15 (macrogol 15 hydroxystearate; BASF India Ltd.,
Mumbai, India), Carbopol ETD 2020 (Noveon India Ltd.,
Mumbai, India), Gattefosse excipients such as Capryol 90
(propylene glycol monocaprylate), Plurol Oleique (polyglyceryl
oleate), Lauroglycol 90 (propylene glycol monolaurate), Lab
racfac CC (caprylic capric acid triglycerides), Labral 1944 CS
476
Microemulsion-Based Vaginal Gel of Clotrimazole: Formulation, In Vitro Evaluation, and Stability Studies
(oleoyl macrogolglycerides), Methocel K4M (hydroproyl meth
ylcellulose; Colorcon Asia Ltd., Mumbai, India), Manugel DMB
(sodium alginate; Anshul Agencies Ltd., Mumbai, India), and
Captex 8000 (glyceryl tricaprylate, Indchem International,
Mumbai, India) were received as gift samples. Methanol (high
performance liquid chromatography (HPLC) grade), Tween 80,
citric acid anhydrous, disodium hydrogen phosphate, chloroc
resol, dimethyl sulfoxide (DMSO), and benzyl alcohol (all AR
grade) were purchased from s.d. Fine Chemical Ltd., Mumbai,
India. Sabaraud dextrose agar was purchased from HiMedia
Ltd., Mumbai India.
Candid V gel (clotrimazole 2% w/w, Glenmark Phar
maceuticals Ltd., Mumbai, India), a marketed vaginal gel of
clotrimazole was purchased from local market. Double
distilled water was used whenever required.
Solubility Studies
The solubility of CMZ in various oils and surfactants was
determined by using shake ask method (n=3). Briey, an
excess amount of CMZ was added to each vial containing
5 ml of the selected vehicle, i.e., either oil or surfactant. After
sealing, the mixture was vortexed using a cyclomixer for
10 min in order to facilitate proper mixing of CMZ with the
vehicles. Mixtures were shaken for 72 h in an isothermal
shaker (Remi, Mumbai, India) maintained at 37 1C.
Mixtures were centrifuged at 5,000 rpm for 15 min, followed
by ltration through membrane lter (0.45 m, 13 mm, Pall
Life sciences, Mumbai, India). The concentration of CMZ in
the supernatant was determined by HPLC method.
HPLC Analysis of CMZ
The solubility of CMZ in various excipients was deter
mined by a stability indicating validated reverse phase HPLC
method developed in house. The HPLC apparatus consisted of
Jasco PU 2080 Plus Intelligent HPLC pump (Jasco, Japan)
equipped with a Jasco UV 2075 Intelligent UV/VIS detector
(Jasco, Japan), a Rheodyne 7725 injector (Rheodyne, USA), a
Jasco Borwin Chromatography Software (version 1.50) integra
tor software, and a Hi Q Sil RP 18 (4.6250 mm and 10 m
particle size) column. The mobile phase consisted of a mixture
of methanol: dipotassium hydrogen phosphate (0.025 M) buffer
(75:25 v/v) at a ow rate of 1.5 ml/min that led to retention time
of 12.5 min when detection was carried out at 254 nm. The assay
was linear (r2 =0.999) in the concentration range 10 250 g/ml
with the lowest detection limit of 1.33 g/ml of CMZ. The
method was validated with respect to accuracy and interday and
intraday precision as per ICH guidelines and the relative
standard deviation was less than 2% in both the cases.
Phase Diagram
An oil titration method was employed in the present
investigation to construct phase diagrams (13). Briey, mixtures
of the double distilled water with Cremophore EL were
prepared at ratios (% w/w) of 9:1, 8:2, 7:3, 6:4, 5:5, 4:6, 3:7, 2:8,
and 1:9 into different vials. A small amount of Capryol 90 in
0.5% (w/w) increments was added into the vials. Following each
addition, the mixtures in vials were vortexed for 2 to 3 min and
were allowed to equilibrate at 25C for 30 min. After
477
equilibration, the mixtures were examined visually for phase

separation, transparency, and ow properties. In addition, the
mixtures were observed through crossed polarizers (fabricated
in house by using polarizing lenses, Nikkon, Japan) for deter
mining the optical isotropy of the systems. The point at which
the mixture became turbid or showed signs of phase separation
was considered as the end point of the titration. The area of
microemulsion existence was determined and denoted as ME.
Formulation of Microemulsion
From the phase diagrams, suitable composition was chosen
for further studies. The composition is shown in Table I. Briey,
CMZ (200 mg) was dissolved in Capryol 90 (14.0 g) by using
overhead stirrer at 1,000 rpm. To this solution, Cremophore EL
(43.5 g), benzyl alcohol (2.0 g), and chlorocresol (0.1 g) were
added and the mixture was stirred further to yield a homoge
nous solution. To this solution, water (38.4 g) was added and
stirred further to yield microemulsion.
Globule Size Analysis of the Microemulsion

The average globule size and polydispersity index of the
CMZ microemulsion were determined in duplicate by the
photon correlation spectroscopy (Beckman Coulter N4 plus,
Wipro, India). Measurements were carried at an angle of 90
at 25C. Microemulsion was diluted with double distilled
water to ensure that the light scattering intensity was within
the instruments sensitivity range. Double distilled water was
ltered through 0.45 m membrane lters (Pall Life Sciences,
Mumbai, India) prior to globule size determination.
Formulation of Microemulsion-Based Gel of CMZ

Various gelling agents, namely, sodium alginate (Manugel
DMB), hydroxypropyl methylcellulose (Methocel K4 M), and
Carbopol ETD 2020, were evaluated for their ability to gel
CMZ microemulsion. Briey, the nonaqueous part of the
microemulsion was prepared as described above (Formulation
of Microemulsion); gelling agent was dispersed/dissolved in
water by using overhead stirrer at 1,000 rpm. Aqueous part was
mixed with nonaqueous part by stirring. The suitable gelling
agent was selected on the basis of compatibility with micro
emulsion structure, feel, and ease of spreadability. The opti
mized composition of CMZ MBG is shown in Table II.
Table I. Composition of CMZ Microemulsion

Ingredient
Content (% w/w)
Clotrimazole
Capryol 90
Cremophor EL
Benzyl alcohol
Chlorocresol
Water to make (g)
2
14
43.5
2
0.1
100
478
Bachhav and Patravale

Table II. Composition of CMZ MBG
Ingredient
Content (% w/w)
Clotrimazole
Capryol 90
Cremophor EL
Benzyl alcohol
Chlorocresol
Carbopol ETD 2020
Water to make (g)
2
14
43.5
2
2.0
0.8
100
Characterization of the CMZ-MBG
0.5 C with a rotating speed of 25 rpm in buffer pH 4.5 citrate

phosphate buffer as a dissolution medium. A watch dish
containing 1.0 g of the developed formulation was tightly
secured with a stainless steel wire screen (350 mesh size
sinker). The dish was then dipped in 500 ml pH 4.5 citrate
phosphate buffer, contained in a vessel of USP dissolution test
apparatus. During the study, 2 ml of aliquots were removed at
predetermined time intervals (0.5, 1, 2, 3, 4, 6, 8, 10, and 12 h)
from the dissolution medium and replaced with fresh media.
The amount of CMZ released in the dissolution medium was
determined by a HPLC method described earlier. Samples were
ltered through 0.45 nylon membrane lter.
In Vitro Antifungal Activity
Determination of CMZ Content, pH, and Spreadability

For determination of drug content, about 1 g of the gel was
weighed in a 100 ml volumetric ask and dissolved in methanol; it
was diluted appropriately and analyzed by the HPLC method
described earlier. The spreadability of the gel was determined
using the following technique: 0.5 g gel was placed within a circle
of 1 cm diameter premarked on a glass plate over which a second
glass plate was placed. Aweight of 500 g was allowed to rest on the
upper glass plate for 5 min. The increase in the diameter due to
spreading of the gels was noted. The pH of the 10% (w/w) gel was
determined using Equip tronic Digital pH meter Model EQ 610,
standardized using pH 4.0 and 7.0 standard buffers before use.
Rheological Studies on the MBG
Antifungal activity of CMZ MBG, Candid V gel, and

CMZ standard (CMZ dissolved in DMSO) was evaluated
against Candida albicans ATCC 10231 by using a cup plate
method. Briey, the concentration of C. albicans ATCC 10231
in inocula was equivalent to 51015 CFU/ml. CMZ MBG,
Candid V gel (500 mg each), and 1 ml of CMZ dissolved in
DMSO (10 mg/ml) was added to agar plate and plates were
kept in the dark conditions at room temperature for 48 h. After
incubation, the mean zone of inhibition was recorded for all
the test samples (n=3).
Stability Studies
Chemical stability of CMZ in CMZ MBG was assessed
at various storage conditions viz. 25C/60% relative humidity
(RH), 30C/65% RH, and 40C/75% RH for a period of
3 months. CMZ MBG was packed in 10 g aluminum oint
ment tubes. Samples (n=3) were removed at 0, 30, 60, and
90 days and were assessed for CMZ content by a stability
indicating HPLC method described earlier. The statistical
signicance of differences in the data was analyzed utilizing
analysis of variance followed by Bonferronis test (GraphPad
InStat Demo Version). Differences were considered statisti
cally signicant at P<0.05.
Brookeld Synchro Lectric Viscometer (Model RVT)

with helipath stand was used for rheological studies. The
sample (30 g) was placed in a beaker and was allowed to
equilibrate for 5 min before measuring the dial reading using a
T C spindle at 0.5, 1, 2.5, and 5 rpm. At each speed, the
corresponding dial reading on the viscometer was noted. The
spindle speed was successively lowered and the corresponding
dial reading was noted. The measurements were carried in
duplicate at ambient temperature. Direct multiplication of the
dial readings with factors given in the Brookeld viscometer
catalog gave the viscosity in centipoises.
In Vitro Bioadhesion Study
Solubility Studies
The bioadhesive potential of the CMZ MBG was evaluat

ed in comparison with the marketed clotrimazole gel (Candid
V gel) by an in vitro method reported by Nakamura et al. (14).
Briey, an agar plate (1% w/w) was prepared in pH 4.5 citrate
phosphate buffer. Test sample of 50 mg was placed at the center
of plate. After 5 min, the agar plate was attached to a US
Pharmacopeia disintegration test apparatus (Fig. 1) and moved
up and down in pH 4.5 citrate phosphate buffer at 371C. The
sample on the plate was immersed into the solution at the lowest
point and was out of the solution at the highest point. The
residence time of the test samples on the plate was noted
visually. The experiments were performed in triplicate.
The results of the solubility studies are shown in

Table III. It is evident from Table III that the Capryol 90
exhibited the highest solubilizing potential for the CMZ as
compared to the other oils. Among the various surfactants
In Vitro Dissolution
In vitro release proles of CMZ MBG and Candid V
were studied using modied USP XXIII apparatus I at 37
Fig. 1. Apparatus used for in vitro bioadhesion study
Table III. Solubility of CMZ in Various Oils and Surfactants (n 3)
479
Table IV. Results of In vitro Bioadhesion Studies (n 3)
Excipient
Solubility (mg/ml)
Formulation
In vitro bioadhesion time (min)
Capryol 90
Lauroglycol 90
Captex 8000
Plurol Oleique
Labral 1944 CS
Cremophor EL
Solutol HS 15
Tween 80
Labrasol
103.16
97.45
10.62
3.00.5
57.01.5
45.02.0
48.73.2
36.43.5
58.04.0
FLZ MBG
Candid V gel
483.5*
241.5
used in the study, Labrasol exhibited the highest solubilizing

potential for CMZ followed by Solutol HS 15, Cremophore
EL, and Tween 80 (Table III). However, Labrasol has poor
emulsication ability for Capryol 90 as compared to the other
surfactants (15). Among the surfactants employed in the
study, Cremophor EL exhibited excellent emulsication
ability for Capryol 90 (15) and fairly good solubilizing
potential for the CMZ. Hence, Cremophore EL was selected
for the further studies.
Phase Diagram and Globule Size Analysis of the Microemulsion
It is reported that nonionic surfactants alone can yield
microemulsion without help of cosurfactant (9 11). The phase
diagram of Cremophore EL Capryol 90 water system is shown
in Fig. 2. Black region in the gure indicates microemulsion
region (ME region) and white region is non ME region. It is
evident from the gure that Cremophore alone could give
considerable microemulsication region (>30%). White circle
in the black region of Fig. 2 indicates ME composition selected
for formulation. Selection of this ME composition was based
on solubility of CMZ in Capryol 90 and amount of Cremophor
Fig. 2. Ternary phase diagram of Cremophore EL Capryol 90 Water

system
*P<0.05 as compared to Candid V gel
EL required for emulsication to produce a stable micro

emulsion. The selected microemulsion system was composed
of maximum amount of water possible. The microemulsion for
further studies was selected from the phase diagram which had
globule size of 48.4 nm and polydispersity index of 0.75. The
incorporation of CMZ did not have considerable inuence on
the globule size of the microemulsion.
Formulation and Characterization of the MBG

Various gelling agents such as sodium alginate, hydrox
ypropyl methylcellulose, and Carbopol ETD 2020 were
evaluated for the gelling of CMZ microemulsion. It was
observed that sodium alginate affected the structure of the
microemulsion and resulted in separation of oily phase. This
observation could be attributed to that fact that salts like sodium
alginate can affect the structure of the microemulsion (9 11).
Hydroxypropyl methylcellulose was unable to yield viscosity
desirable for the gel formulations. Only Carbopol ETD 2020
could yield clear gel without disturbing the microstructure of the
CMZ microemulsion. Furthermore, Carbopols are known to
have mucoadhesive properties and have been used in the
formulation vaginal delivery systems (7,8,16). Hence, Carbo
pol ETD 2020 was selected for the formulation of MBG.
The CMZ content of the MBG and Candid V gel was
found to be 98.43.2% and 99.60.8%, respectively. The pH
value of CMZ MBG was 4.52 which is equivalent to the vaginal
pH; for candid V, pH value was 6.8. Spreadability is an
important property of topical formulation from patient compli
ance point of view. The diameter for CMZ MBG and Candid
V was found to be 7.2 and 6.20 cm, respectively. High
spreadability value of CMZ MBG compared to Candid V
indicates better spreading ability at the site of application. Both
CMZ MBG and Candid V showed pseudoplastic behavior
and the viscosity values of CMZ MBG and Candid V at 5 rpm
were 9.0106 and 7.8106 mPa s, respectively.
Fig. 3. In vitro dissolution prole of CMZ formulations (n 3)
480
Table V. Antifungal Activity of Various Samples Against C. albicans

ATCC 10231 (n 3)
Formulation
Zone of inhibition (mm)
CMZ standard
Candid V gel
CMZ MBG
4.20.1*
3.00.1*
5.40.2
CMZ MBG showed signicantly higher in vitro antifungal activity

*P<0.05 as compared to CMZ MBG
Fig. 4. Chromatogram of the CMZ MBG subjected to 40C/75% RH

at the end of 3 months
In Vitro Bioadhesion Study

The bioadhesive potential of CMZ MBG and commercial
formulation (Candid V gel) was evaluated by in vitro method.
The results of the study are shown in Table IV. The CMZ MBG
showed signicantly higher retention time as compared to
Candid V gel (P<0.05). This clearly indicates that the CMZ
MBG may have higher residence time in vagina as compared to
Candid V gel. The increased bioadhesivity of CMZ MBG can
be attributed to the presence of Carbopol.
In Vitro Dissolution
In vitro release prole of CMZ MBG and Candid V
gel is shown in Fig. 3. More than 85% of CMZ was released
over the period of 12 h from CMZ MBG and it followed rst
order release kinetics. Candid V gel showed maximum 70%
release of CMZ up to 12 h. The in vitro release pattern of
CMZ MBG and Candid V gel was equivalent up to 3 h.
After this initial phase, Candid V gel produced higher
release of CMZ up to 6 h and it was constant at the end of
12 h. CMZ MBG provided controlled release of CMZ up to
12 h. The difference between the release pattern of MBG and
Candid V gel was nonsignicant as determined by F2 test
(F2 =61.99 with standard error of 7.049).
In Vitro Antifungal Activity
Stability Studies
The results of stability studies are shown in Table VI. It
is evident from the table that CMZ showed a signicant
degradation (P<0.05) at all the storage conditions at the end
of 3 months. This observation was totally unexpected. The
rate of degradation increased with the increase in the
temperature. Hence, at 40C/75% RH, around 38% of
CMZ was degraded at the end of the 3 months. The
chromatogram of the sample (stored at 40C/75% RH) at
the end of the third month is shown in Fig. 4 and it clearly
shows the well resolved degradation product of the CMZ.
Interestingly, this chromatogram is similar to the chromato
gram obtained after forced degradation of CMZ in 1 N HCl
(Fig. 5). This clearly indicates that degradation of CMZ in
MBG is due to the acidic pH of the formulation. The
acidic pH of the MBG could be attributed to the presence
of the free fatty acids (such as caprylic acid) in the Capryol
90.
It was observed that commercial formulation of CMZ
(Candid V gel) has pH value of 6.81 and it claims stability
of up to 1.5 years. This corroborates the inferences drawn
about the acid catalyzed degradation of CMZ.
Our preliminary investigations on the stabilization of the
CMZ in CMZ MBG indicated that the increase in the pH of
the gel results in the loss of transparency and also disturbs the
The results of antifungal studies are shown in Table V. It

is evident that CMZ MBG showed higher antifungal activity
as compared to the marketed Candid V gel and CMZ
standard (P<0.01). The enhanced in vitro antifungal activity
of CMZ MBG may be attributed to enhanced penetration of
oil globules containing CMZ through fungal cell walls to
inhibit ergosterol synthesis.
Table VI. Chemical Stability of CMZ in CMZ MBG During Stability

Testing (n 3)
% CMZ content in gel (SD)
Month
0.5
1
2
3
25C/60% RH
101.2
98.1
92.1
84.6
(4.2)
(2.5)
(4.2)
(2.9)
30C/60% RH
98.7
95.1
93.7
87.2
(0.5)
(4.1)
(0.7)
(1.36)
40C/75% RH
94.7
84.8
71.1
62.4
(0.9)
(1.7)
(1.1)
(2.6)
Fig. 5. Chromatogram of forced degradation of CMZ in acidic con
ditions (1 N HCl)
structure of the microemulsion (data not shown). Hence, it
may be necessary to reformulate the developed MBG. The
future efforts would be directed towards identifying an oily
phase that preserves the integrity of CMZ on long term
storage and has good solubilizing potential for CMZ.
Nonetheless, this investigation clearly indicates that MBG
could be a viable alternative to the conventional vaginal
formulations.
CONCLUSION
The microemulsion based gel could be successfully
developed for the vaginal delivery of clotrimazole. The
developed gel showed promising in vitro performance with
respect to bioadhesivity and antifungal activity. However,
CMZ showed a considerable degradation on long term
storage in the developed microemulsion based gel which
was attributed to the pH of the formulation.
ACKNOWLEDGEMENTS
YGB thanks the University Grants Commission, New
Delhi India for the nancial support. The authors are
thankful to Cipla Pharmaceuticals, BASF, Indchem Interna
tional, Noveon India, Anshul Agencies and Colorcon Asia
Pvt. Ltd. for the gift samples of the drug and excipients.
Authors would like to acknowledge Mr. Abhijit Date for his
help in the preparation of the manuscript and also for the
technical discussion.
REFERENCES
1. Lanchares JL, Hernandez ML. Recurrent vaginal candidiasis
changes in etiopathogenical patterns. Int J Gynecol Obstet.
2000;71:S29 35.
481
2. Ferrer J. Vaginal candidosis: epidemiological and etiological

factors. Int J Gynecol Obstet. 2000;71:S21 7.
3. Ritter W, Patzschke K, Krause U, Stettendorf S. Pharmacoki
netic fundamentals of vaginal treatment with clotrimazole.
Chemotherapy 1982;28:37 42.
4. Pavelic Z, Skalko Basnet N, Jalsenjak I. Characterisation and in
vitro evaluation of bioadhesive liposome gels for local therapy of
vaginitis. Int J Pharm. 2005;301:140 8.
5. Pedersen M, Bjerregaard S, Jacobsen J, Srensen AM. A
genuine clotrimazole cyclodextrin inclusion complex isolation,
antimycotic activity, toxicity and an unusual dissolution rate. Int
J Pharm. 1998;176:121 31.
6. Richardson JL, Whetstone J, Fisher AN, Watts P, Farraj NF,
Hinchcliffe M, et al. Gamma scintigraphy as a novel method to
study distribution and retention of a bioadhesive vaginal delivery
in sheep. J Control Rel. 1996;42:133 42.
7. Sharma G, Jain S, Tiwary K, Kaur G. Once daily bioadhesive
vaginal clotrimazole tablets: design and evaluation. Acta Pharm.
2006;56:337 45.
8. Knuth K, Amiji M, Robinson J. Hydrogel delivery systems for
vaginal and oral applications. Adv Drug Deliv Rev. 1993;11:137 67.
9. Tenjarla S. Microemulsions: an overview and pharmaceutical
applications. Crit Rev Ther Drug Carrier Syst. 1999;16:461 521.
10. Lawrence MJ, Rees GD. Microemulsion based media as novel
drug delivery systems. Adv Drug Del Rev. 2000;45:89 121.
11. Date AA, Patravale VB. Microemulsions: applications in trans
dermal and dermal delivery. Crit Rev Ther Drug Carrier Sys.
2007;24:547 96.
12. Vyas TK, Babbar AK, Sharma RK, Misra A. Intranasal
mucoadhesive microemulsions of zolmitriptan: preliminary stud
ies on brain targeting. J Drug Target. 2005;13:317 24.
13. Corswant CV, Engstrom S, Sderman O. Microemulsions based
on soybean phosphatidylcholine and triglycerides. Phase behav
ior and microstructure. Langmuir 1997;13:5061 70.
14. Nakamura F, Ohta R, Machida Y, Nagai T. In vitro and in vivo
nasal mucoadhesion of some water soluble polymers. Int J
Pharm. 1996;134:173 81.
15. Date AA, Nagarsenker MS. Design and evaluation of self
nanoemulsifying drug delivery systems (SNEDDS) for cefpodox
ime proxetil. Int J Pharm. 2007;329:166 72.
16. Baloglu E, zyazc M, Hzarcoglu SY, Karavana HA. An in
vitro investigation for vaginal bioadhesive formulations: bioad
hesive properties and swelling states of polymer mixtures. Il
Farmaco. 2003;58:391 6.

DOI: 10.1208/s12249 009 9234 1
Research Article
SMEDDS of Glyburide: Formulation, In Vitro Evaluation, and Stability Studies
Yogeshwar G. Bachhav1 and Vandana B. Patravale1,2
Received 22 July 2008; accepted 17 March 2009; published online 21 April 2009
Abstract. The objective of the present investigation was to develop and evaluate self microemulsifying
drug delivery system (SMEDDS) for improving the delivery of a BCS class II antidiabetic agent,
glyburide (GLY). The solubility of GLY in oils, cosurfactants, and surfactants was evaluated to identify
the components of the microemulsion. The ternary diagram was plotted to identify the area of
microemulsion existence. The in vitro dissolution prole of GLY SMEDDS was evaluated in comparison
to the marketed GLY tablet and pure drug in pH 1.2 and pH 7.4 buffers. The chemical stability of GLY in
SMEDDS was determined as per the International Conference on Harmonisation guidelines. The area of
microemulsion existence increased with the increase in the cosurfactant (Transcutol P) concentration.
The GLY microemulsion exhibited globule size of 133.5 nm and polydispersity index of 0.94. The stability
studies indicated that GLY undergoes signicant degradation in the developed SMEDDS. This
observation was totally unexpected and has been noticed for the rst time. Further investigations
indicated that the rate of GLY degradation was highest in Transcutol P.
KEY WORDS: degradation; glyburide; SMEDDS; stability studies; Transcutol P.
INTRODUCTION
Glyburide (GLY) or glibenclamide is a second generation
sulfonylurea used in the treatment of noninsulin dependent
diabetes. It is one of the most prescribed long acting antihyper
glycemic agents (1). GLY is classied as BCS class II drug, which
means it has high permeability and poor water solubility (2).
The poor water solubility of GLY is responsible for its poor
dissolution rate, which ultimately leads to variable absorption of
GLY. Furthermore, there are reports which have documented
that GLY shows large variations in interindividual bioavailabil
ity and bioequivalence of the marketed products (3,4). Thus, it
can be concluded that the bioavailability and in vivo perfor
mance of GLY is dependent on its dissolution rate. In view of
this, researchers have attempted various techniques to improve
the dissolution rate and, ultimately, the in vivo performance of
GLY. To date, the potential of various drug delivery approaches,
such as amorphization (5), solid dispersion (6 8), cyclodextrin
complexation (9,10), and permeation enhancers (10), has been
established in improving in vitro and/or in vivo performance of
GLY. However, to date, the potential of microemulsions has not
been established in the delivery of GLY.
The potential of microemulsions and/or self microemulsify
ing drug delivery systems (SMEDDS) or anhydrous form of
microemulsion in the improvement of the bioavailability and
therapeutic performance of the hydrophobic agents has been
very well established (11 14). In view of this, the present
investigation was aimed at developing SMEDDS for the
1
Department of Pharmaceutical Sciences and Technology, Institute of

Chemical Technology, Matunga, Mumbai 400019, India.
2
To whom correspondence should be addressed. (e mail: vbpatravale@
udct.org)
delivery of GLY. SMEDDS of GLY with globule size <150 nm

were successfully developed. The results of stability studies of
GLY SMEDDS are also presented in this investigation.
Materials
Glyburide was kindly provided as a gift sample by Cipla
Pharmaceuticals (Mumbai, India). Capryol 90, Lauroglycol 90,
Transcutol P, Plurol Oleique CC 497, Labral M 1944 CS,
Labrasol (Colorcon Asia, Mumbai, India), Capmul MCM
(Indchem International, Mumbai, India), hydroxypropyl cellu
lose (Signet Chemicals, Mumbai, India), and Solutol HS 15
(BASF India, Mumbai, India) were received as gift samples.
Tween 80, Tween 20, and PEG 400 (all AR grade) were
purchased from Merck (Mumbai, India). Hydrochloric acid,
potassium dihydrogen phosphate (all AR grade), and acetoni
trile (high performance liquid chromatography [HPLC] grade)
were purchased from s.d. Fine Chemicals (Mumbai, India).
Solubility Studies
The solubility of GLY in various oils, cosurfactants, and
surfactants was determined by using the shake ask method.
Briey, an excess amount of GLY was added to each vial
containing 5 mL of the selected vehicle, i.e., oil, cosurfactant,
and surfactant. After sealing, the mixture was vortexed using
a cyclomixer for 10 min in order to facilitate proper mixing of
GLY with the vehicles. Mixtures were shaken for 72 h in an
isothermal shaker (Remi, Mumbai, India), maintained at 37
1C, and afterwards, mixtures were centrifuged at 5,000 rpm
for 15 min, followed by ltration through membrane lter
482
SMEDDS of Glyburide
483
(0.45 m, 13 mm; Pall Life Sciences, Mumbai, India). The

concentration of GLY in the supernatant was determined by
an in house HPLC method.
HPLC Analysis of GLY
The solubility of GLY in various excipients was deter
mined by a stability indicating and validated in house HPLC
method. The HPLC apparatus consisted of a Jasco PU 2080
Plus Intelligent HPLC pump (Jasco, Japan) equipped with a
Jasco UV 2075 Intelligent UV/VIS detector (Jasco, Japan), a
Rheodyne 7725 injector (Rheodyne, USA), a Jasco Borwin
Chromatography Software (version 1.50) integrator software,
and a Hi Q Sil RP 18 (4.6 mm250 mm and 10 particle
size) column. The mobile phase consisted of a mixture of
acetonitrile/monobasic ammonium phosphate (0.05 M) buffer
(55:45 v/v) at a ow rate of 1.5 mL/min that led to a retention
time of 7.1 min when detection was carried out at 225 nm.
The assay was linear (r2 =0.9991) in the concentration range
1 5 g/mL with the lowest detection limit of 30 ng/mL of
GLY. The method was validated with respect to accuracy and
interday and intraday precision as per the International
Conference on Harmonisation (ICH) guidelines, and the
relative standard deviation was <2% in both the cases.
The developed method could successfully resolve the
acid, base, and oxidative degradation products of GLY (data
not shown). The same method was used for determining
chemical stability of GLY in SMEDDS. For determining the
solubility of GLY and for determining the GLY content in
SMEDDS, the various samples were suitably diluted by the
mobile phase and injected in the HPLC system. There was no
need for any extraction from the sample as all the compo
nents were soluble in the mobile phase.
Phase Diagrams
A water titration method was employed in present
investigation to construct phase diagrams (15). Briey, mixtures
of Capryol 90 and Tween 20+Transcutol P were prepared at
ratios (in percent w/w) of 9:1, 8:2, 7:3, 6:4, 5:5, 4:6, 3:7, 2:8, and
1:9 into different vials. The phase diagrams were plotted at two
different Km values (ratio of surfactant to cosurfactant), viz.,
0.67 and 0.55. A small amount of double distilled water in 0.5%
(w/w) increment was added into the vials. Following each
Fig. 2. Solubility of GLY (in milligrams per milliliter) in various

surfactants and cosurfactants (n 3; data are expressed as meanSD)
addition, the mixtures in the vials were vortexed for 2 3 min

and were allowed to equilibrate at 25C for 30 min. After
equilibration, the mixtures were examined visually for phase
separation, transparency, and ow properties. In addition, the
mixtures were observed through crossed polarizers (fabricated
in house by using polarizing lenses; Nikkon, Japan) for deter
mining the optical isotropy of the systems. The point at which
the mixture became turbid or showed signs of phase separation
was considered as the endpoint of the titration. The area of
microemulsion existence was determined and denoted as ME.
The phase diagrams were constructed at two different ratios of
Tween 20/Transcutol P, namely, 0.55 and 0.67.
Formulation of SMEDDS
The optimized SMEDDS formulation consisted of gly
buride/Capryol 90 (oily phase)/Transcutol P (cosurfactant)/
Tween 20 (surfactant)/hydroxypropyl cellulose (viscosity
builder) (0.75:15.1:53:30.3:0.75) in percent w/w. Briey, the
calculated amounts of GLY that was dissolved in Capryol 90
and Transcutol containing hydroxypropyl cellulose and
Tween 20 were added to this solution under vortexing. The
homogenous SMEDDS was used for further studies.
Globule Size Analysis
SMEDDS (0.5 g) was diluted with 50 mL double distilled
water. The average globule size and polydispersity index of the
GLY microemulsion were determined in duplicate by the
photon correlation spectroscopy (Beckman Coulter N4 plus,
Wipro, India). Measurements were carried out in triplicate at an
angle of 90 at 25C. If required, microemulsion was diluted with
double distilled water to ensure that the light scattering intensity
was within the instruments sensitivity range. Double distilled
water was ltered through 0.45 m membrane lters (Pall Life
Sciences, Mumbai) prior to globule size determination.
In Vitro Dissolution Profile
Fig. 1. Solubility of GLY (in milligrams per milliliter) in various oils

(n 3; data are expressed as meanSD)
SMEDDS of GLY (equivalent to 5 mg of GLY) was

lled in size 0 hard gelatin capsules. In vitro release proles
of SMEDDS of GLY, pure GLY powder, and marketed GLY
formulation (Euglucon Tablets, Nichols Piramal India, Mum
484

Stability Studies
Chemical stability of GLY in SMEDDS was assessed under
various storage conditions, viz., 25C/60% RH, 30C/65% RH,
and 40C/75% RH as per ICH guidelines. SMEDDS equivalent
to 5 mg of GLY was lled in size 0 hard gelatin capsules. Ten
such capsules were packed in Alu Alu strip packs and stored at
various aforementioned storage conditions up to 1 month.
Samples were removed at 0, 15, and 30 days of interval and
the GLY content in the samples was analyzed by an in house
HPLC method described earlier.
Stability of GLY in Individual SMEDDS Components
In order to understand the unusual and unexpected
degradation of the GLY, it was decided to monitor the
stability of GLY in the components of SMEDDS. Briey,
GLY (5.0 mg) was dissolved in various components of
SMEDDS (1.0 mL), viz., Transcutol P, Tween 20, and
Capryol 90, whereas in case of hydroxypropyl cellulose,
GLY powder was thoroughly mixed with equal amount of
hydroxypropyl cellulose to physical mixture. All these
samples and a pure GLY powder were stored at 60C/ambient
RH for 15 days. The GLY content of the samples was
analyzed by an in house HPLC at the end of 15 days.
Solubility Studies
Fig. 3. Ternary phase diagram of Tween 20+Transcutol P Capryol

90 water system at a Km 0.67 and b Km 0.55
bai, India) were studied using United States Pharmacopeia

(USP) XXIII apparatus I at 370.5C with a rotating speed of
100 rpm in buffer pH 1.2 and pH 7.4 as the dissolution media.
During the study, 2 mL of aliquots were removed at
predetermined time intervals (5, 10, 20, and 30 min) from
the dissolution medium and replaced with fresh media. The
amount of GLY released in the dissolution medium was
determined by an in house HPLC method described earlier.
The results of solubility studies of the GLY in various

oils, cosurfactants, and surfactants are shown in Figs. 1 and 2.
It is evident that GLY has considerably low solubility in oils,
surfactants, and cosurfactants despite a high log P value of
4.2. Such a behavior has already been seen with drugs like
amphotericin B and carbamazepine. Hence, GLY can be
considered as a drug which is poorly soluble in both lipid and
water. However, due to the very small dose of GLY, it would
still be possible to formulate a SMEDDS formulation that can
be lled in a unit dosage form. Among the various oils
screened, Capryol 90 showed the highest solubilizing poten
tial for GLY (5.850.45 mg/mL) compared to Capmul MCM
(2.30.5 mg/mL), Labral M 1944 CS (1.70.3 mg/mL),
Lauroglycol 90 (1.60.3 mg/mL), and Plurol Oleique CC 497
(0.70.2 mg/mL). All the surfactants tested (Cremophor EL,
Tween 20, and Tween 80) showed similar solubilization for
GLY (6.3 6.5 mg/mL). In all the solubilizers studied, Trans
cutol P showed the highest solubilizing potential for GLY
(151.2 mg/mL) compared to Solutol HS 15 (6.50.8 mg/mL),
Labrasol (5.30.6 mg/mL), and PEG 400 (2.30.6 mg/mL).
Table I. In Vitro Dissolution Proles of Various GLY Formulations

Cumulative percent GLY released with time
pH 1.2 buffer
pH 7.4 buffer
Formulation
5 min
10 min
20 min
30 min
5 min
10 min
20 min
30 min
Pure GLY powder

Marketed GLY tablets
GLY SMEDDS
0.80.2
1.91.5
91.27.6
1.00.3
2.20.7
100.54.6
1.10.3
2.40.9
90.16.3
1.30.1
3.01.3
86.45.2
1.150.75
6.120.57
95. 02.8
2.11.5
5.90.8
95.53.0
5.11.5
7.41.3
96.81.3
7.33.0
8.50.9
97.31.8
SMEDDS of Glyburide
485
Table II. Chemical Stability of GLY in SMEDDS During Stability

Testing (n 3)
Percent GLY content in SMEDDS (SD)
Day
25C/60% RH
30C/65% RH
40C/75% RH
0
15
30
98.11.1
90.61.7*
81.31.9*
98.11.1
84.72.5*
75.34.3*
98.11.1
61.60.7*
50.22.5*
both the dissolution media, indicating that the release was

independent of the pH of the dissolution medium. However,
there was decrease in cumulative percent GLY release in
pH 1.2 buffer with the increase in time. The reason for this is
the extreme insolubility of the GLY in pH 1.2 buffer. Because
of the same reason, pure GLY and marketed GLY tablet
showed very poor release in pH 1.2 buffer compared to
SMEDDS. The pH 1.2 buffer cannot provide sink condition
*P<0.05, signicantly different compared to 0 day when evaluated by

analysis of variance
Capryol 90, Tween 20, and Transcutol P were selected for

further studies.
Phase Diagrams and Globule Size Analysis
Pseudoternary phase diagrams were constructed to
investigate the effect of surfactant to cosurfactant ratio (Km)
on the area of microemulsion existence (Figs. 3a, b). It is a
well known fact that the Km value has considerable effect on
the area of microemulsion existence. In the present study, it
was observed that the area of microemulsion existence was
higher at Km =0.55 compared to that of Km =0.67. Hence, for
the formulation of SMEDDS, a lower Km value (0.55) was
employed. The optimized microemulsion had a globule size of
133.5 nm with the polydispersity index of 0.94.
Formulation of SMEDDS
Various SMEDDS compositions were screened (data not
shown) and optimal composition which could solubilize the
therapeutic dose of GLY was selected.
Hydroxypropyl cellulose was added to increase the
viscosity of SMEDDS so as to enable unit dosage form of
GLY SMEDDS. There are reports in the literature where
cellulosics have been used for this purpose (16,17).
In Vitro Dissolution Study
The results of in vitro dissolution proles of various GLY
formulations are shown in Table I. It is evident from the table
that GLY SMEDDS showed a dramatic improvement in the
in vitro dissolution prole compared to the marketed GLY
tablet and pure GLY in both the dissolution media. GLY
SMEDDS showed more than 90% GLY released in 5 min in
Table III. Percent GLY Content in Individual SMEDDS Excipients

after 15 Days at 60C of Storage (n 3)
Sample
Percent GLY content
Pure GLY powder

GLY+hydroxypropyl cellulose
GLY dissolved in Capryol 90
GLY dissolved in Transcutol P
GLY dissolved in Tween 20
99.51.5
98.72.5
68.23.5*
2.00.5*
44.11.3*
*P < 0.05, signicantly different compared to pure GLY when

evaluated by analysis of variance
Fig. 4. Chromatogram of the GLY dissolved in various SMEDDS

excipients at the end of 15 days when stored at 60C: a Transcutol P, b
Tween 20, c Capryol 90
486
for GLY. The USP recommends pH 7.4 buffer as a dissolution
medium for GLY. But since any oral dosage form would be
rst exposed to acidic conditions, therefore, the in vitro
release of GLY was evaluated in pH 1.2 buffer. It is also
noteworthy that the in vitro dissolution of GLY formulations
in pH 1.2 buffer has been evaluated for the rst time as there
are no reports in the literature where pH 1.2 buffer is used to
evaluate enhanced dissolution rate of GLY. It is also an
important observation that decreased dissolution of GLY over
the period of time was not observed in the pH 7.4 buffer. The
marketed formulation and pure GLY showed very less
release (<10%) in both media compared to the GLY
SMEDDS.

CONCLUSION
The SMEEDS of GLY could be successfully developed,
and the developed SMEDDS showed dramatic improve
ment in the dissolution rate of GLY compared to the
marketed formulation and pure GLY powder. However,
stability studies indicated that GLY showed signicant
degradation in the developed SMEDDS. This clearly
indicates the need to investigate the stability of the API in
the SMEDDS components selected for formulation devel
opment, an aspect which has not been focused so far in the
literature.
ACKNOWLEDGEMENT
Stability Studies
The results of stability studies of GLY SMEDDS are
shown in Table II. Surprisingly, the content of GLY in
SMEDDS started to decline signicantly 15 days after the
storage at all storage conditions. The degradation was higher
at the 40C/75% RH. On the 30th day, GLY in SMEDDS
showed around 50% degradation at 40C/75% RH. This was
a totally unexpected result. There are no reports on the
stability of the GLY in the oily phases, surfactants, or the
solubilizers like Transcutol P. Furthermore, the GLY is not
reported to be a molecule which can easily undergo
degradation.
Lipid excipients, surfactants, and solubilizers are usually
thought to be inert. The stability aspects of the active
pharmaceutical ingredient (API) in SMEDDS formulations
have not been reported in most of the investigations.
Paradoxically, all these excipients are routinely used for the
formulation of SMEDDS. This observation clearly indicates
the need to evaluate the stability of the API in the SMEDDS
or SMEDDS components before proceeding to formulation
development. In order to understand the degradation behav
ior of GLY in SMEDDS, it was decided to monitor its
stability in the individual SMEDDS components.
Stability of GLY in Components of SMEDDS

The results of the stability of GLY in the individual
SMEDDS components are shown in Table III. It is evident
from the table that GLY showed a signicant degradation in
all the components except hydroxypropyl cellulose at the end
of the 15 days. The chromatograms of the samples stored at
60C/ambient RH for 15 days are shown in Fig 4a c. It is
evident that the rate of degradation of GLY was very fast in
Transcutol P compared to the other excipients (Table III).
Chemically, Transcutol P is diethylene glycol monoethyl
ether, and so far, there are no reports on the degradation of
APIs in Transcutol P. At the same time, it should be noted
that GLY was not compatible with other components, as
degradation of GLY was observed in contact with Capryol 90
and Tween 20. It can be seen from the chromatograms that
the degradation product was the same in all the excipients,
although the extent of degradation was variable. In the
future, studies would be undertaken to understand the
mechanism of the degradation of GLY by conducting liquid
chromatography mass spectrometry analysis.
YGB thanks the University Grants Commission, New

Delhi, India for the nancial support. The authors would like
to thank Cipla Pharmaceuticals, Colorcon Asia Pvt. Ltd.,
Indchem International, BASF India, and Signet Chemicals
for the gift samples of the drug and the excipients. The
authors would also like to acknowledge Mr. Abhijit Date for
his help in the preparation of the manuscript and also for the
technical discussion.
REFERENCES
1. Groop L, Wahlin Boll E, Totterman KJ, Melander A, Tolppanen
EM, Fyhrqvist F. Pharmacokinetics and metabolic effects of
glibenclamide and glipizide in type 2 diabetes. Eur J Clin
Pharmacol 1985;28:697 704.
2. Wei H, Lobenberg RL. Biorelevant dissolution media as a
predictive tool for glyburide a class II drug. Eur J Pharm Sci
2006;19:45 62.
3. Chalk JB, Patterson M, Smith MT, Eadie MJ. Correlations
between in vitro dissolution, in vivo bioavailability and hypo
glycaemic effect of oral glibenclamide. Eur J Clin Pharmacol
1986;31:177 82.
4. Blume H, Ali SL, Siewert M. Pharmaceutical quality of
glibenclamide products: a multinational postmarket comparative
study. Drug Dev Ind Pharm 1993;19:2713 41.
5. Cordes D, Mueller BW. Deactivation of amorphous glibencla
mide during dissolution. Eur J Pharm Sci 1996;4:S187.
6. Varma MM, Jayaswal SB, Singh J. In vitro and in vivo evaluation
of fast release solid dispersions of glibenclamide. Indian Drugs
1992;29:608 11.
7. Betageri GV, Makaria KR. Enhancement of dissolution of
glyburide by solid dispersion and lyophilization techniques. Int
J Pharm 1995;126:155 60.
8. Tashtoush BM, Al Qashi ZS, Najib NM. In vitro and in vivo
evaluation of glibenclamide in solid dispersion systems. Drug
Dev Ind Pharm 2004;30:601 7.
9. Sanghavi NM, Venkatesh H, Tandel V. Solubilizaiton of gliben
clamide with cyclodextrin and its derivative. Drug Dev Ind
Pharm 1994;20:1275 83.
10. Zerrouk N, Corti G, Ancillotti S, Maestrelli F, Cirri M, Mura P.
Inuence of cyclodextrins and chitosan, separately or in combi
nation, on glyburide solubility and permeability. Eur J Pharm
Biopharm 2006;62:241 6.
drug delivery systems. Adv Drug Deliv Rev 2000;45:89 121.
12. Date AA, Patravale VB. Microemulsions: applications in trans
dermal and dermal delivery. Crit Rev Ther Drug Carrier Syst
2007;24:547 96.
13. Yang SC, Gursoy RN, Lambert G, Benita S. Enhanced oral
absorption of paclitaxel in a novel self microemulsifying drug
delivery system with or without concomitant use of P glycoprotein
inhibitors. Pharm Res 2004;21:261 70.
SMEDDS of Glyburide
14. Gursoy RN, Benita S. Self emulsifying drug delivery systems
(SEDDS) for improved oral delivery of lipophilic drugs. Biomed
Pharmacother 2004;58:173 82.
15. Li P, Ghosh A, Wagner RF, Krill S, Joshi YM, Serajuddin ATM.
Effect of combined use of nonionic surfactant on formation of
oil in water microemulsions. Int J Pharm 2005;288:27 34.
487
16. Piclin PS, Homar M, Valant AZ, Perlin MG. Sodium ascorbyl
phosphate in topical microemulsions. Int J Pharm 2003;256(1
2):65 73.
17. Kongan A, Garti N. Microemulsions as transdermal drug
delivery vehicles. Adv Colloid Interface Sci 2006;123 126:369
85.

DOI: 10.1208/s12249 009 9236 z
Research Article
Effect of Sugars, Surfactant, and Tangential Flow Filtration on the Freeze-Drying
of Poly(lactic acid) Nanoparticles
Samuli Hirsjrvi,1,2,3 Leena Peltonen,1 and Jouni Hirvonen1
Received 20 November 2008; accepted 29 March 2009; published online 21 April 2009
Abstract. Poly(D,L lactic acid) nanoparticles were freeze dried in this study. With respect to drying, effect
of protective excipients and purication from excess surfactant were evaluated. The nanoparticles were
prepared by the nanoprecipitation method with or without a surfactant, poloxamer 188. The particles
with the surfactant were used as such or puried by tangential ow ltration. The protective excipients
tested were trehalose, sucrose, lactose, glucose, poloxamer 188, and some of their combinations. The best
freeze drying results in terms of nanoparticle survival were achieved with trehalose or sucrose at
concentrations 5% and 2% and, on the other hand, with a combination of lactose and glucose.
Purication of the nanoparticle dispersion from the excess surfactant prior to the freeze drying by
tangential ow ltration ensured better drying outcome and enabled reduction of the amount of the
protective excipients used in the process. The excess surfactant, if not removed, was assumed to interact
with the protective excipients decreasing their protective mechanism towards the nanoparticles.
KEY WORDS: freeze drying; nanoparticles; poly(lactic acid); protective excipients; tangential ow
ltration.
INTRODUCTION
Polymeric nanoparticles, poly(lactic acid) (PLA) and its
derivates being recognized examples, are a high potential
platform for efcient exploitation of different drug delivery
formulations and routes because of the properties provided
by their small size (1). These possible benets include
controlled release, protection of the encapsulated drug, and
drug targeting. Nanoparticles are generally applied as inject
able dispersions, but due to the stability problems encoun
tered in the liquid state, they have to be dehydrated for
storage (2). Freeze drying is one of the most convenient
methods enabling liquid removal and further reconstitution of
nanoparticles for therapeutical use (3).
Freeze drying is a complicated process involving changes
in temperature and physical state of materials as well as
concentrations of different substances in the liquid environ
ment, which easily disturb the stability of nanoparticle
dispersion. Therefore, appropriate protective excipients are
generally regarded as indispensable for successful freeze
drying cycles, and their choice needs to be evaluated for
different nanoparticulate formulations. Approved protective
sugars are, e.g., trehalose (2 12), sucrose (6,10 13), lactose
(6,7,14,15), glucose (2,4,7,14,15), saccharose and maltose (10),
and mannitol (7,11,12). The protective mechanism of these
1
Division of Pharmaceutical Technology, Faculty of Pharmacy,

University of Helsinki, Helsinki, Finland.
2
Present address: INSERM U646, 10 rue Andr Boquel, 49100
Angers, France
3
To whom correspondence should be addressed. (e mail: samuli.
hirsjarvi@helsinki.)
excipients originates from the amorphous matrix of the frozen

sugar in water (16). The amorphous matrix forms hydrogen
bonds with the nanoparticles acting as a water substitute
inhibiting the destructive effect of ice crystals.
Certain potentially toxic residues may remain in the nano
particulate dispersion after the preparation process including
organic solvents, surfactants/stabilizers, coating agents, and,
possibly, monomers/initiators from the polymerization process.
Tangential ow ltration (TFF), or cross ow ltration, has been
found as a very practical method to purify nanoparticulate
dispersions from these impurities (5,13,17 20). It is a more gentle
and continuous process and more easily scaled up than, e.g., the
traditional dead end ultraltration (21) or ultracentrifugation
(22). In TFF, the dispersion is pumped tangentially along the
surface of the membrane. Filtrate is formed when an applied
pressure forces part of the dispersion medium through the
membrane. The retained particles do not block the membrane
because they are swept away with the tangential ow. TFF is still
a less used method in the nanoparticle purication, probably due
to the very small volumes used in the nanoparticle research in
laboratory scale and limited selection of commercial products
available for the small volumes.
The aim of this study was to evaluate protective
excipients, i.e., different sugars (trehalose, sucrose, glucose
and lactose) and a surfactant (poloxamer 188), on the
outcome (quality parameters like redispersibility and main
tained size distribution) of freeze dried poly(D,L lactic acid)
nanoparticles. The protective excipients were tested on
surfactant free nanoparticles and nanoparticles prepared with
the surfactant. Additionally, suitability of the tangential ow
ltration was assessed in the purication of the PLA nano
particles (removal of excess surfactant) prior to the freeze
488
Freeze-drying of PLA Nanoparticles
489
drying. Functionality of the different excipients was evaluated

after the drying by visual inspection, electron microscopy, and
size determinations (photon correlation spectroscopy).
be adjusted by the clamps (7 and 8 in Fig. 1). The

nanoparticle dispersion volume was 20 ml and the volume
of water used in purication was 120 ml.
Freeze-drying
Materials
Nanoparticle dispersions were dispersed in vials (1 ml of

nanoparticles + 250 l of sugars/surfactant) and freeze dried
with a Lyostar II (FTS, Stone Ridge, USA). The freeze
drying cycle was as follows: freezing at 40C (240 min);
primary drying at 35C and 150 mTorr (1,020 min); second
ary drying at 20C and 50 mTorr (180 min).
The nanoparticles were freeze dried with different pro
tective excipients at two concentrations, namely 0.4% or
0.2% (w/v nanoparticles in water). When single excipients
(sugars and poloxamer 188 added after preparation) were
used, their concentrations in the dispersions were 5%, 2%,
and 1%. Lactose and glucose were used also in combination
with mass ratio of 2:1 (lactose glucose), at three different
concentrations (lactose and glucose 4% and 2%, 1.6% and
0.8% and 0.8% and 0.4%), based on our previous positive
experiences (15). For comparison purposes, lactose and
glucose were used with poloxamer 188 in such manner that
the sugar concentration was kept constant (2.5%) and the
poloxamer 188 concentration was 2.5%, 1%, or 0.5%. All the
concentrations are percent w/v.
PURASORB PDL 02A poly(D,L lactic acid) (a dona

tion from PURAC Biomaterials, Gorinchem, The Nether
lands; IV 0.20 dl/g) was used in the nanoparticle preparation.
Other excipients used in the preparation and drying were
acetone (Riedel de Han, Seelze, Germany), poloxamer 188
(Lutrol F 68, BASF, Ludwigshafen, Germany), D trehalose,
sucrose, D glucose (Sigma Aldrich Chemie, Steinheim, Ger
many), lactose (DMV International, Veghel, The Nether
lands), and ultrapuried water (Millipore, Molsheim, France).
Eleven micrometer paper lters (Whatman, Brentford, UK)
were used for the raw purication of the nanoparticle
dispersions. 0.2 m Isopore membrane lters (Millipore,
Molsheim, France) aided in the visualization by electron
microscopy.
Nanoparticle Preparation and Purification
PLA nanoparticles were prepared by the nanoprecipita
tion method (23). Twenty ve milligrams of PLA were
dissolved in 2 ml of acetone. The polymeric solution was
added with a syringe and a gauge (0.45 mm) directly into 4 ml
of the outer phase (water or water with the surfactant, 20 mg
of Poloxamer 188) under mild stirring. The organic solvent
was evaporated for 40 min in fume hood and the nanoparticle
dispersion was diluted with water and ltered (paper lter) to
remove possible undesired aggregates formed during the
nanoprecipitation.
Nanoparticle batches prepared with poloxamer 188 were
dried as such (with excipients), or the excess surfactant was
rst removed by TFF. The ltration was performed with a
continuous dialtration mode using a Minimate TFF
System and a Capsule with Omega TI300K Membrane (Pall
Corporation, Ann Arbor, USA) run by a peristaltic pump
(Fig. 1). Briey, the nanoparticle dispersion circulated
through the ltration capsule with the help of a peristaltic
pump. The ow rate was 40 ml/min, ltration rate 1 2 ml/min,
and operating pressure 10 psi. The operating pressure could
Nanoparticle Characterization
Size distributions of the nanoparticles were determined
with Malvern Zetasizer 3000HS (Malvern, Worcestershire,
UK). Particle sizing was based on photon correlation
spectroscopy; the results were analyzed by CONTIN algo
rithm and the sizes presented are based on the intensity
distributions.
Freeze dried samples were characterized by the appear
ance of the cake, redispersibility in water, Tyndall effect,
presence of visual aggregates (AS=aggregation scale), poly
dispersity index (PI, a dimensionless number from 0 to 1
describing the homogeneity of the sample), and size relative
to the initial size (Sf/Si). These parameters are commonly
used in the characterization of the freeze dried nanoparticles
(3).
Appearance of the nanoparticle populations was visual
ized by scanning electron microscopy (SEM). Nanoparticu
late samples, collected by ultraltration on membrane
surfaces, were attached by two sided tape on metal plates,
sputtered for 20 s by platinum with an Agar Sputter Coater
(Agar Scientic Ltd., Essex, UK) and analyzed with a DSM
962 SEM (Zeiss, Jena, Germany).
Fig. 1. Schematic illustration of the TFF setup. (1) nanoparticle

dispersion; (2) magnetic stirrer; (3) peristaltic pump; (4) pressure
gauge; (5) ltration capsule; (6) ltrate; (7) a clamp to control the
ltrate rate; (8) a clamp to control the pressure; (9) retentate; (10)
dialtration medium (water)
Role of a surfactant is rarely studied in the freeze drying

process because it is usually included in the nanoparticle
formulation as an indispensable stabilizer. Polymer selection
of this study allowed preparation of surfactant free nano
particles: the D,L PLA used was relatively short chained (IV
0.20 dl/g corresponds to MW of 20,000 g/mol) and possessed
free carboxylic acid chain ends. These properties provide
water miscibility and charge for the hydrophobic polymer,
490
Hirsjrvi, Peltonen and Hirvonen
which enables the surfactant free preparation and stable

nanoparticle dispersion.
Therefore, different PLA nanoparticles were prepared
for the freeze drying experiments: particles without the
surfactant and particles with the surfactant (poloxamer 188).
The surfactant containing particles were used as such or
puried by the TFF. Average diameters of all the nano
particles were approximately 200 nm, which showed that
neither the surfactant nor its removal affected the particle
size. More precisely, the measured particle sizes and PI,
respectively, were as follows: nanoparticles without the
surfactant, 202 nm and 0.03; nanoparticles with the surfactant,
201 nm and 0.01; surfactant containing nanoparticles puried
by the TFF, 199 nm and 0.03. Polydispersity indices of all
types of the particles were low, indicating very homogeneous
size distributions. Particle size of each batch was measured
before the freeze drying to ensure precise comparison with
the particle size after the drying (Sf/Si).
The freeze drying cycle was a generic one providing
sufcient conditions for successful drying: the freezing was
performed with low temperature shelves ( 40C), and the
primary drying was performed in a relatively low temperature
( 35C) for an extensively long period (1,020 min). A one
for all cycle was used because of the different properties of
the different excipients (see later).
Most of the studied nanoparticle protectant combina
tions expressed the Tyndall effect (opalescence due to the
light scattering of the colloidal nanoparticle dispersion).
These batches were characterized to possess the AS 0 or
1 (Table I). However, the batches with AS 1 could not be
considered as successfully freeze dried ones because of the
presence of some visible particles in the dispersion. In

addition to aggregation, other typical signs of poor freeze
drying processes are increased size (increase in Sf/Si) and
increase in polydispersity index indicating a broader size
distribution. Therefore, a freeze drying process of nano
particles can be considered successful, if the particles are
easily redispersed in liquid without aggregation, the Sf/Si
remains close to 1.0 and the PI does not change.
Effect of Nanoparticle Concentration
Increase in the nanoparticle concentration has been

reported to favor the freeze drying (20). However, drying of
these PLA nanoparticles at higher concentration (0.4%) was
not very successful. Proper cakes were formed, but generally
the redispersed nanoparticle dispersions were highly aggre
gated. The only acceptable aggregate free drying results were
achieved with the nanoparticles prepared with the poloxamer
188 and protected by 5% sucrose (Sf/Si 1.3; PI<0.1) and a
combination of lactose 4% and glucose 2% (Sf/Si 1.0; PI<0.1).
Interestingly, this high nanoparticle concentration was aggre
gated instantly once it was mixed with the highest tested
trehalose concentration (5%) before drying. PLA nano
particles have been found to aggregate when the electrolyte
concentration of the dispersion medium increases to a certain
level (24). This disturbs the interparticle electrostatic repul
sion, which is the most important stabilization mechanism of
these PLA nanoparticles. Trehalose molecules as well as
other sugars are not charged but very polar ones instead. In
conditions of extreme dehydration (such as freeze drying),
Table I. Freeze drying of the PLA Nanoparticles Dried at Concentrations of 0.2% (w/v)
Sf/Si
PI
AS
5% of protectant
a
Trehalose
Trehaloseb
Trehalosec
Sucrosea
Sucroseb
Sucrosec
Lactosea
Lactoseb
Lactosec
Glucosea
Glucoseb
Glucosec
Poloxamerb
Lactose and glucosea
Lactose and glucoseb
Lactose and glucosec
Lactose and poloxamerb
Glucose and poloxamerb
1.0
< 0.1
0
1.2
< 0.1
0
1.1
< 0.1
0
1.0
< 0.1
0
1.0
< 0.1
0
1.0
< 0.1
0
nd
nd
1
nd
nd
1
1.3
< 0.1
0
nd
nd
2
1.1
< 0.1
0
1.0
< 0.1
0
1.1
< 0.1
0
Lactose 4% and glucose 2%
1.0
< 0.1
0
1.2
< 0.1
0
1.0
< 0.1
0
Sugar 2.5% and poloxamer 2.5%
1.1
<0.1
0
1.0
<0.1
0
Sf/Si
PI
AS
2% of protectant
1.0
< 0.1
0
1.2
0.2
0
1.2
< 0.1
0
1.0
< 0.1
0
1.0
< 0.1
0
1.0
< 0.1
0
nd
nd
1
nd
nd
1
nd
nd
1
nd
nd
2
nd
nd
2
nd
nd
2
nd
nd
1
Lactose 1.6% and glucose 0.8%
1.0
< 0.1
0
1.3
< 0.1
0
1.1
0.1
0
Sugar 2.5% and poloxamer 1%
1.1
<0.1
0
1.0
<0.1
0
Sf/Si
PI
AS
1% of protectant
nd
nd
2
nd
nd
1
1.2
0.8
0
nd
nd
2
nd
nd
1
1.1
0.2
0
nd
nd
1
nd
nd
1
nd
nd
1
nd
nd
2
nd
nd
2
nd
nd
2
nd
nd
1
Lactose 0.8% and glucose 0.4%
nd
nd
1
nd
nd
1
nd
nd
1
Sugar 2.5% and poloxamer 0.5%
1.2
<0.1
0
1.0
0.1
0
Concentrations of the additives are expressed as percent w/v

Sf/Si ratio of nanoparticle size after and before freeze drying, PI polydispersity index, AS aggregation scale (0 no aggregation, 1 uniform visible
small aggregates, 2 strong aggregation), nd not determined due to observed aggregation
a
Nanoparticles prepared without the surfactant
b
Nanoparticles prepared with poloxamer 188, no TFF purication
c
Nanoparticles prepared with poloxamer 188, puried by TFF

e.g., trehalose and sucrose have similar protective action,
but in an aqueous environment trehalose is bound (hydro
gen bonds) much more effectively to biomaterials (25). This
is due to the different orientation of the hydroxyl groups of
the sugars. Probably, in the cases of high nanoparticle and
trehalose concentrations, the concentration increase (al
though not caused by electrolytes) and the strong dehydra
tion effect of the sugar towards the nanoparticles led to the
destruction of dispersion stability.
Using a nanoparticle concentration of 0.2%, freeze
drying could be performed successfully with the different
combinations and concentrations of excipients. Results of
those PLA nanoparticles freeze dried with the different
protective excipients are presented in Table I and are the
topic of the following discussion. Concentrations in the low
range are frequently encountered in the nanoparticle studies
(20,26), especially when the nanoprecipitation has been used
as a nanoparticle preparation method (6,10).
Effect of the Protective Excipients
The three kinds of nanoparticles were freeze dried

without any excipients for comparison purposes. These
batches produced appropriate cakes during drying, but the
formulations were highly aggregated upon redispersion in
water.
When the nanoparticles were dried at a concentration of
0.2% with a single protectant, the higher the protectant
concentration was the better were the quality parameters of
the nanoparticles. This is in good agreement with previous
data, as 5% sugar concentrations have often provided enough
material for cake formation and nanoparticle protection
(2,4,26). In this case, 5% resulted in 25:1 sugar to nano
particle mass ratio, which is relatively high in terms of the
protectant amount. The high ratio might be one reason for
the positive freeze drying outcomes. However, considering
the amount of the sugars, concentrations up to 30% have
been frequently studied and reported (2,13,14). In this study,
on the other hand, the ranges of mass ratios of the sugars to
nanoparticles of 10:1 and 5:1 (originating from 2% and 1% of
the sugar, respectively) are common (4,6,10,14,20,26).
The best protective sugars were trehalose and sucrose
used as concentrations of 5% and 2% (Sf/Si close to 1.0 and
PI around 0.1). These concentrations protected adequately all
the three kinds of nanoparticles. Slight increase in size (Sf/Si),
typical for the freeze drying of nanoparticles, was observed in
some samples with trehalose, but the polydispersity indices
still remained low. Instead, 1% of sugar was not enough to
protect the nanoparticles leading to aggregation or at least
increase in polydispersity. The slightly better results with
sucrose compared to trehalose originate from the different
orientations of the hydroxyl groups of the sugars, as discussed
earlier. As expected based on our previous study, lactose
alone could not provide protection for the nanoparticles due
to its poor protective action during the freezing step (15).
Generally, all the tested formulations produced decent
cakes except the ones with glucose as the only protectant.
Collapse of a product can take place if sublimation (primary
drying) takes place above the collapse temperature (Tc) of
the given composition (27). Characteristics for collapsed
491
products are high residual water contents and long redis
persion times. Indeed, the glucose batches were not immedi
ately redispersed in water after the freeze drying but required
a longer time to be reconstituted. Tc of glucose is around 42C
(4), which was below the primary drying temperature ( 35C).
Tcs of trehalose, sucrose, and lactose are approximately 29C,
31C, and 31C, respectively (28). Naturally, the primary
drying could be performed at very low temperatures, but that
might not be relevant in terms of process duration and energy
consumption.
Despite the poor appearance of the glucose protected
formulations, the nanoparticles could be recovered in the
cases of surfactant containing nanoparticles with 5% of
glucose. When glucose and poloxamer 188 were added in
the nanoparticle formulations as a combination, high quality
nanoparticles could be reconstituted without aggregation.
The amount of glucose was kept constant (2.5%), while the
amount of poloxamer 188 varied from 2.5% to 0.5%. Even
the lowest amount of poloxamer 188 enhanced the appear
ance of the cake and enabled successful drying. On the other
hand, if poloxamer 188 was used as the only added
protectant, a concentration of 5% was needed for good
reconstitution results. As glucose is known to be an efcient
cryoprotectant during the freezing step (15), these results
emphasize the role of freezing as the crucial part of the
freeze drying process.
In our previous study, we have used lactose and glucose
successfully together compensating each others limitations in
the freeze drying process (15): lactose has provided good
appearance for the dried product whereas glucose has
protected the formulation during the freezing step. Therefore,
in this study, lactose and glucose were tested as a combination
using three different sugar concentrations. The two higher
concentrations gave positive results (Sf/Si close to 1.0 and PI
around 0.1), but the combination lactose 0.8% and glucose
0.4% did not help in the drying. Taking into account the
results with single sugars, it is obvious that a total concentra
tion of at least 2% of protectants is needed to ensure
successful freeze drying of these PLA nanoparticles. This
means a 10:1 mass ratio of sugar(s) to nanoparticles.
As observed, poloxamer 188 enhanced the protective
effect of glucose. Even higher enhancement occurred when
poloxamer 188 was used with lactose: the nanoparticles could
be reconstituted with Sf/Si slightly above 1.0 and PI around
0.1. Although the lowest combination concentration of the
sugar and poloxamer 188 in samples was 3% (higher than the
lowest concentration of single sugars and lactose and
glucose), it can be assumed that poloxamer 188, used at a
high concentration relative to the nanoparticle concentration,
acts as a coprotectant during the freeze drying. As the
surfactant enhanced the effect of poor cryoprotectant lactose,
it can be considered as a protectant for the freezing step. On
the contrary to our ndings, poloxamer 188 and similar
surfactants are reported to be susceptible to crystallization
during freeze drying, which leads to aggregation of nano
particles (13,29). Another explanation is that poloxamer is
more soluble in cold water (which is encountered during
freezing) due to increased hydrogen bonding leading to
destabilization (26). However, in these cases, poloxamer is
an indispensable part of the nanoparticle structure (aggrega
tion occurs if the surfactant is removed from the surface)
492
Fig. 2. Example of successful freeze drying process: TFF ltrated

PLA nanoparticles were freeze dried with 2% of sucrose
compared to the nanoparticles in our current study (the

surfactant is not necessary). At the same time, combination of
a sugar and poloxamer in freeze drying has been reported to
give positive results because the sugar dehydrates the
surfactant in the bulk solution forcing it to the particle
surface (to stabilize the particles) (3). It is also agreed that
poloxamer 188 forms a hydrophilic layer at the particle
surface helping redispersion after the freeze drying (26).
Probably, in our current study, the added surfactant remains
on the nanoparticle surface providing protection during
freeze drying. Similar results have been reported recently
considering poloxamer 188 as an indispensable protectant
used together with sugars (10). However, in the current study,
because satisfactory results can be achieved using the single
sugars, use of the poloxamer 188 as an additional protectant is
not relevant.
Effect of Tangential Flow Filtration
Poloxamers are widely used as nanoparticle stabilizers and

they are generally considered as safe excipients in human use
(30). However, removal of the excess surfactant material remain
ing in the nanoparticle formulations after preparation should be
considered because it may modify the physiological, physico
chemical, and drug release properties of the system(s). Therefore,
the TFF was applied as a purication method in this study.
The volume used in the purication, diavolume, was 6: 120
ml of water on 20 ml of nanoparticle dispersion. This diavolume
should reduce the amount of the excess poloxamer 188 remain
ing in the dispersion close to 0.1% of the initial amount (31).
Freeze drying results of the tangential ow ltrated
batches were better than those of the nonltrated. Aggre
gate free dispersions were observed after redispersion even
when 1% of trehalose or sucrose was used. Also, in this case,
the outcome with sucrose (PI 0.2) was better than with
trehalose (PI 0.8). Among the tested lactose concentrations,
the only satisfactory result was achieved with a tangential
ow ltrated batch protected with 5% of lactose. The origin
for the better freeze drying results after TFF may be the
following: in the cases of nonltrated batches, the excess

poloxamer 188 forms hydrogen bonds with the protective
sugars (32) thus disturbing their protective mechanism. The
amount of poloxamer 188 used in the nanoparticle prepara
tion was 20 mg. On the other hand, 1% of a protective sugar
in the dispersion was 12.5 mg (250 l of 5% sugar solution
added in 1 ml of nanoparticle dispersion). Although a part of
the poloxamer 188 was included in the nanoparticle structure,
excess surfactant still existed capable of forming the hydrogen
bonds with the sugars (i.e., the amounts of poloxamer 188 and
the sugar were in the same range). This obviously competed
with the hydrogen bonding between the sugar and the
nanoparticle surface, which is the assumed prerequisite for
the protection during the freeze drying process (16). This
effect could be avoided by removing the excess surfactant by
TFF. It should be remembered, however, that both the TFF
puried nanoparticles and the surfactant containing non
ltrated nanoparticles survived better during the drying than
the surfactant free nanoparticles.
Summarizing the three different types of PLA nano
particles, the best freeze drying results were achieved with
the tangential ow ltrated batches, then with the polox
amer 188 containing nonltrated batches and, nally, the
poorest results with the surfactant free nanoparticles. Thus,
these results are in agreement with the current understanding
that the use of surfactant provides extra stability for the
prepared nanoparticles towards further processing and result
ing changes in the dispersion state (10,24,26).
Imaging by Electron Microscopy
Reconstituted nanoparticle dispersions after freeze

drying are usually deposited (and let to dry) to microscopy
plates for electron microscopy visualization. Because the
protective excipients are not removed, the nanoparticles are
covered by a sheet formed by the excipients, which makes
observations of the particles difcult (4,15,33). To avoid
this, the nanoparticle dispersion can be deposited on an
ultraltration membrane of proper size: the freeze drying
excipients are removed by the ltration and the nano
Fig. 3. Example of a poor freeze drying process: PLA nanoparticles

were freeze dried with 2% of lactose

particles remaining on the membrane surface can be
visualized by a conventional electron microscopy.
An example of a successfully freeze dried nanoparticle
population is presented in Fig. 2. For comparison, Fig. 3
shows a typical failure of the freeze drying cycle: Tyndall
effect could be observed, but with some visible aggregation
and increases in particle size and polydispersity.
CONCLUSIONS
Poly(D,L lactic acid) nanoparticles could be successfully
freeze dried when the concentrations of the nanoparticles
and the protective excipients were optimized. A sugar
concentration of at least 2% was needed for the successful
drying, sucrose being the best choice for protection. Also,
the combination of lactose and glucose gave positive
results. In the case of these poly(lactic acid) nanoparticles,
the use of poloxamer 188 surfactant was not necessary
during the formulation and preparation steps, but it clearly
enhanced the readiness of the nanoparticles for the freeze
drying. Furthermore, purication of the nanoparticles from
the excess surfactant using tangential ow ltration enabled
even better drying results when the different sugars were
studied.
ACKNOWLEDGEMENTS
The Electron Microscopy Unit of the Institute of
Biotechnology (University of Helsinki) is acknowledged for
providing laboratory facilities and analytical equipment.
REFERENCES
1. Couvreur P, Vauthier C. Nanotechnology: intelligent design to
treat complex disease. Pharm Res. 2006;23:1417 50.
2. Chacn M, Molpeceres J, Berges L, Guzmn M, Aberturas
MR. Stability and freeze drying of cyclosporine loaded poly
(D,L lactide glycolide) carriers. Eur J Pharm Sci. 1999;8:99
107.
3. Abdelwahed W, Degobert G, Stainmesse S, Fessi H. Freeze
drying of nanoparticles: formulation, process and storage con
siderations. Adv Drug Deliv Rev. 2006;58:1688 713.
4. Abdelwahed W, Degobert G, Fessi H. A pilot study of freeze
drying of poly( caprolactone) nanocapsules stabilized by poly
(vinyl alcohol): formulation and process optimization. Int J
Pharm. 2006;309:178 88.
5. de Jaeghere F, Allmann E, Leroux J C, Stevels W, Feijen J,
Doelker E, et al. Formulation and lyoprotection of poly(lactic
acid co ethylene oxide) nanoparticles: inuence on physical
stability and in vitro cell uptake. Pharm Res. 1999;16:859 66.
6. Jeong Y I, Shim Y H, Kim C, Lim G T, Choi K C, Yoon C.
Effect of cryoprotectants on the reconstitution properties of
surfactant free nanoparticles of poly(D,L lactide co glycolide). J
Microencapsul. 2005;22:593 601.
7. Konan YN, Gurny R, Allmann E. Preparation and character
ization of sterile and freeze dried sub 200 nm nanoparticles. Int J
Pharm. 2002;233:239 52.
8. Cavalli R, Caputo O, Carlotti ME, Trotta M, Scarnecchia C,
Gasco MR. Sterilization and freeze drying of drug free and drug
loaded solid lipid nanoparticles. Int J Pharm. 1997;148:47 54.
9. Schwarz C, Mehnert W. Freeze drying of drug free and drug
loaded solid lipid nanoparticles (SLN). Int J Pharm.
1997;157:171 9.
493
10. Layre A M, Couvreur P, Richard J, Requier D, Ghermani NE,
Gref R. Freeze drying of composite core shell nanoparticles.
Drug Dev Ind Pharm. 2006;32:839 46.
11. Anhorn MG, Mahler H C, Langer K. Freeze drying of human
serum albumin (HSA) nanoparticles with different excipients. Int
J Pharm. 2008;363:162 9.
12. Zillies JC, Zwiorek K, Hoffmann F, Vollmar A, Anchordoquy
TJ, Winter G, et al. Formulation development of freeze dried
oligonucleotide loaded gelatin nanoparticles. Eur J Pharm Bio
pharm. 2008;70:514 21.
13. Saez A, Guzmn M, Molpeceres J, Aberturas MR. Freeze
drying of polycaprolactone and poly(D,L lactic glycolic) nano
particles induce minor particle size changes affecting the oral
pharmacokinetics of loaded drugs. Eur J Pharm Biopharm.
2000;50:379 87.
14. de Chasteigner S, Fessi H, Cav G, Devissaguet JP, Puisieux F.
Gastro intestinal tolerance study of a freeze dried oral dosage
form of indometacin loaded nanocapsules. STP Pharma Sci.
1995;5:242 6.
15. Hirsjrvi S, Peltonen L, Kainu L, Hirvonen J. Freeze drying of
low molecular weight poly(L lactic acid) nanoparticles: effect
of cryo and lyoprotectants. J Nanosci Nanotechnol.
2006;6:3110 7.
16. Abdelwahed W, Degobert G, Fessi H. Investigation of nano
capsules stabilization by amorphous excipients during freeze
drying and storage. Eur J Pharm Biopharm. 2006;63:87 94.
17. Dalwadi G, Sunderland VB. Purication of PEGylated nano
particles using tangential ow ltration (TFF). Drug Dev Ind
Pharm. 2007;33:1030 9.
18. Dalwadi G, Benson H, Chen Y. Comparison of dialtration and
tangential ow ltration for purication of nanoparticle suspen
sions. Pharm Res. 2005;22:2152 62.
19. Limayem I, Charcosset C, Fessi H. Purication of nanoparticle
suspensions by a concentration/dialtration process. Sep Purif
Technol. 2004;38:1 9.
20. de Jaeghere F, Allmann E, Feijen J, Kissel T, Doelker E, Gurny
R. Freeze drying and lyopreservation of diblock and triblock
poly(lactic acid) poly(ethylene oxide) (PLA PEO) copolymer
nanoparticles. Pharm Dev Technol. 2000;5:473 83.
21. Hirsjrvi S, Peltonen L, Hirvonen J. Layer by layer polyelectro
lyte coating of low molecular weight poly(lactic acid) nano
particles. Colloid Surf B. 2006;49:93 9.
22. Sukhorukov GB, Donath E, Lichtenfeld H, Knippel E,
Knippel M, Budde A, et al. Layer by layer self assembly of
polyelectrolytes on colloidal particles. Colloid Surf A.
1998;137:253 66.
23. Fessi H, Puisieux F, Devissaguet JP, Ammoury N, Benita S.
Nanocapsule formation by interfacial polymer deposition follow
ing solvent displacement. Int J Pharm. 1989;55:R1 4.
24. Hirsjrvi S, Peltonen L, Hirvonen J. Surface pressure measure
ments in particle interaction and stability studies of poly(lactic
acid) nanoparticles. Int J Pharm. 2008;348:153 60.
25. Luzardo MdC, Amalfa F, Nuez AM, Daz S, Biondi de Lopez
AC, Disalvo EA. Effect of trehalose and sucrose on the
hydration and dipole potential of lipid bilayers. Biophys J.
2000;78:2452 8.
26. Quintanar Guerrero D, Ganem Quintanar A, Allmann E, Fessi
H, Doelker E. Inuence of the stabilizer coating layer on the
purication and freeze drying of poly(D,L lactic acid) nano
particles prepared by an emulsion diffusion technique. J Micro
encapsul. 1998;15:107 19.
27. Carpenter JF, Pikal MJ, Chang BS, Randolph TW. Rational
design of stable lyophilized protein formulations: some practical
advice. Pharm Res. 1997;14:969 75.
28. Adams GDJ, Ramsay JR. Optimizing the lyophilization cycle
and the consequences of collapse on the pharmaceutical
acceptability of Erwinia L asparaginase. J Pharm Sci.
1996;85:1301 5.
29. Zambaux MF, Bonneaux F, Gref R, Dellacherie E, Vigneron
C. MPEO PLA nanoparticles: effect of MPEO content on
some of their surface properties. J Biomed Mater Res.
1999;44:109 15.
494
30. Singh Joy SD, McLain VC. Safety assessment of poloxamers 101,
105, 108, 122, 123, 124, 181, 182, 183, 184, 185, 188, 212, 215, 217,
231, 234, 235, 237, 238, 282, 284, 288, 331, 333, 334, 335, 338, 401,
402, 403, and 407, poloxamer 105 benzoate, and poloxamer 182
dibenzoate as used in cosmetics. Int J Toxicol. 2008;27:93 128.
31. Millipore Corporation. Technical brief: protein concentration
and dialtration by tangential ow ltration. Billerica: Millipore
Corporation; 2003.

32. Kokol V. Interactions between polysaccharide polymer thick
ener and bifunctional bireactive dye in the presence of
nonionic surfactants. Part 1: surface tension and rheological
behavior of different polysaccharide solutions. Carbohydr
Polym. 2002;50:227 36.
33. Choi MJ, Briancon S, Andrieu J, Min SG, Fessi H. Effect of
freeze drying process conditions on the stability of nanoparticles.
Dry Technol. 2004;22:335 46.

DOI: 10.1208/s12249 009 9227 0
Research Article
Disintegration of Highly Soluble Immediate Release Tablets: A Surrogate
for Dissolution
Abhay Gupta,1 Robert L. Hunt,1 Rakhi B. Shah,1 Vilayat A. Sayeed,2 and Mansoor A. Khan1,3
Abstract. The purpose of the work was to investigate correlation between disintegration and dissolution
for immediate release tablets containing a high solubility drug and to identify formulations where
disintegration test, instead of the dissolution test, may be used as the acceptance criteria based on
International Conference on Harmonization Q6A guidelines. A statistical design of experiments was used
to study the effect of ller, binder, disintegrating agent, and tablet hardness on the disintegration and
dissolution of verapamil hydrochloride tablets. All formulation variables, i.e., ller, binder, and dis
integrating agent, were found to inuence tablet dissolution and disintegration, with the ller and
disintegrating agent exerting the most signicant inuence. Slower dissolution was observed with
increasing disintegration time when either the ller or the disintegrating agent was kept constant.
However, no direct corelationship was observed between the disintegration and dissolution across all
formulations due to the interactions between different formulation components. Although all tablets
containing sodium carboxymethyl cellulose as the disintegrating agent, disintegrated in less than 3 min, half
of them failed to meet the US Pharmacopeia 30 dissolution criteria for the verapamil hydrochloride tablets
highlighting the dependence of dissolution process on the formulation components other than the
disintegrating agent. The results identied only one formulation as suitable for using the disintegration test,
instead of the dissolution test, as drug product acceptance criteria and highlight the need for systematic
studies before using the disintegration test, instead of the dissolution test as the drug acceptance criteria.
KEY WORDS: disintegration test; dissolution test; ICH Q6A; specication.
INTRODUCTION
Immediate release oral dosage forms, i.e., tablets and
capsules, are most widely used drug delivery systems available.
These products are designed to disintegrate in the stomach
The opinions expressed in this work are only of authors and do not
necessarily reect the policy and statements of the FDA.
1
Division of Product Quality Research, Ofce of Testing and Research,

Ofce of Pharmaceutical Science, Food and Drug Administration,
10903 New Hampshire Avenue, LS Bldg 64, Silver Spring, Maryland
20993 002, USA.
2
Division of Chemistry III, Ofce of Generic Drugs, Food and Drug
Administration, 7500 Standish Place, Rockville, Maryland 20855,
USA.
3
To whom correspondence should be addressed. (e mail: Mansoor.
khan@fda.hhs.gov)
ABBREVIATIONS: DCP, Dicalcium phosphate dihydrate; HPMC,
Hypromellose; ICH, International Conference on Harmonization;
LMH, Lactose monohydrate; MCC, Microcrystalline cellulose;
NaCMC, Sodium carboxymethyl cellulose; PVP, Polyvinyl pyrollidone;
USP, United States Pharmacopeia; VLHX1, Tablets containing LMH,
HPMC, without disintegrating agent, and hardness of 6 Kgf; VLHX2,
Tablets containing LMH, HPMC, without disintegrating agent, and
hardness of 10 Kgf; VLPA1, Tablets containing LMH, PVP, NaCMC,
and hardness of 6 Kgf; VLPA2, Tablets containing LMH, PVP, NaCMC,
and hardness of 10 Kgf; VLHA1, Tablets containing LMH, HPMC,
NaCMC, and hardness of 6 Kgf; VLHA2, Tablets containing LMH,
HPMC, NaCMC, and hardness of 10 Kgf.
followed by their dissolution in the uids of the gastrointestinal

tract (Fig. 1). Dissolution of the drug substance, under
physiological conditions, is essential for its systemic absorption.
For this reason, dissolution testing is typically performed on
solid dosage forms to measure the drug release from the drug
product as a test for product quality assurance/product perfor
mance and to determine the compliance with the dissolution
requirements when stated in the individual monograph (1). In
limited number of cases, an in vitro in vivo correlation is
established between the drug release and drug product absorp
tion necessary for therapeutic effects. Disintegration test is also
a standardized test and is primarily used as a quality assurance
tool to conrm complete disintegration of solid oral dosage
forms within the prescribed time when placed in a liquid
medium under the experimental conditions described in their
respective ofcial monographs (1). Disintegration test neither
implies nor tests for the complete solution of the drug or the
dosage form. The difference between these two tests is that in
case of the disintegration test, the test measures the time
required for a product to disintegrate and de aggregate into
multiparticulate system in a given medium, while in case of the
dissolution test, the test measures the concentration of the drug
product in a given medium at a specied time. Thus, disinte
gration of a solid dosage form may not be a measure of
dissolution of the drug substance in the dosage form.
The recent International Conference on Harmonization
(ICH) Q6A guideline provides specications and acceptance
criteria which should be established for all new drug
495
Gupta et al.
496
solubility <250 mL throughout the physiological pH range

(1.2 6.8), was used (4). Filler, binder, and disintegrating agent
were selected as the formulation variables, while manufac
turing conditions were varied by compressing the resulting
formulations at two compression pressures.
Chemicals
Fig. 1. Relationship between dissolution and disintegration of an oral
solid dosage form
substances and new drug products that have not been

registered previously in the ICH region (2). The guidance
recommends using a single point measurement test to mea
sure the release of drug substance from immediate release
drug products. This guidance also provides an option where
dissolution testing may be replaced by disintegration testing
for some immediate release solid oral drug products, con
taining high solubility drug over a pH range. In such cases,
ICH allows disintegration time with an upper time limit to be
used as the drug release acceptance criteria if following
conditions are satised (2):
1. The drug product is not designed to produce modied
release.
2. The solubility of drug is high enough at 370.5C so
that dose/solubility <250 mL through out the physio
logical pH range (1.2 6.8).
3. Greater than 80% dissolution is achieved in 15 min at
pH 1.2, 4.0, and 6.8.
4. A relationship has been determined and established
between disintegration and dissolution or disintegra
tion is more discriminating than dissolution.
5. Dissolution does not affect bioavailability.
6. Changes in formulation or manufacturing variables do
not affect dissolution.
There was a concern of using disintegration as a
universal surrogate for dissolution for immediate release
dosage forms (3). The current study focuses on determining
the relationship, if any, between the disintegration time and
the dissolution time, and identifying formulations where
dissolution testing may be replaced by disintegration testing
as the acceptance criteria for the release of the drug product.
Verapamil hydrochloride, a high solubility drug with dose/
Verapamil hydrochloride (Fermion Oy, Espoo, Finland)

was used as the model drug. Lactose monohydrate (LMH;
Foremost Farms USA, Rothschild, WI, USA) or dicalcium
phosphate dihydrate (DCP; Di Cafos, CFB KG, Budenheim,
Germany) was used as the ller. Hypromellose (HPMC;
Methocel E15 LV, Dow Chemical, Midland, MI, USA) or
polyvinyl pyrollidone (PVP; Kollidon 30, BASF Corp, Florham
Park, NJ, USA) was used as the binder. Microcrystalline
cellulose (MCC; Emcocel 50 M, JRS Pharma, Cedar Rapids,
IA, USA) or sodium carboxymethyl cellulose (NaCMC;
Ac di sol, FMC Biopolymer, Newark, DE, USA) was used as
the disintegrating agent. Talc (Spectrum Chemicals Mfg. Corp.,
Gardena, CA, USA) and magnesium stearate (Riedel de
Haen, Seelze, Germany) were used as glidant and lubricant,
respectively. All materials were used as received after
passing through a standard US standard No. 20 sieve. The
concentrations are listed in Table I.
Preparation of Formulations Using Design of Experiments
A 2232 full factorial design was used to evaluate the
inuence of ller, binder, disintegrating agent, and hardness
on the dissolution and disintegration time of tablets. Appro
priate quantities of drug and ller were weighed and blended
together in a low shear bench top mixer (Model N50, Hobart
Canada, North York, Ontario, Canada) for 5 min and
transferred to a Strea 1 uid bed granulator (Niro Inc.,
Columbia, MD, USA). The batch size was 1,500 g. Granula
tion was performed by spraying a 6.0% w/v aqueous binder
solution on to the uidized powder blend at spray rate of 7
1 g/min. The inlet air temperature and air ow were kept
constant at 55 1C and 140 10 m 3 /h, respectively.
Subsequent to the binder addition, granules were dried to
moisture content of less than 2.0% w/w in the uid bed. The
dried granules were screened though a US standard No. 20
sieve and blended with the disintegrating agent (if present),
Table I. Concentration of the Formulation Components Used to Prepare Tablets

Ingredient
Function
Concentration (%)
Verapamil HCl
DCP/LMH
PVP/HPMC
NaCMC/MCC/none
Talc
Magnesium stearate
Drug
Filler
Binder
Disintegrating agent
Glidant
Lubricant
10.0
75.0
4.0
10.0/10.0/0.0
0.5
0.5
Code (V*1*2*3 #)a

*1 D/L
*2 P/H
*3 A/M/X
HCl hydrochloride, DCP/D dicalcium phosphate dihydrate, LMH/L lactose monohydrate, PVP/P polyvinyl pyrollidone, HPMC/H
hypromellose, NaCMC/A sodium carboxymethyl cellulose, MCC/M microcrystalline cellulose, X none
a
# 1 or 2; used to distinguish tablets from each formulation with low or high hardness, respectively
Disintegration Test as a Surrogate for Dissolution Test

lubricant, and glidant in a 2.0 L V blender (GlobePharma,
New Brunswick, NJ, USA) at 30 rpm for 5 min.
The 12 formulations, obtained by varying the ller,
binder, and disintegrating agent, were compressed into tablets
on a 10 station instrumented tablet press (MiniPress I,
GlobePharma, New Brunswick, NJ, USA) tted with a single
set of 12 mm at faced, beveled edged, round punches and
cylindrical die. The press speed was kept constant at 20 rpm.
Approximately 60 tablets were compressed from each
formulation at low (6 KgF) and high (10 KgF) hardness
settings, resulting in 24 set of tablets. The die and punch were
cleaned using methanol after compressing each formulation.
The 24 set of tablets were assigned a ve digit alpha numeric
code V*1*2*3#, with *1, *2, *3, and # representing
ller, binder, disintegrating agent, and tablet hardness,
respectively. The codes used for each digit are listed in
Table I. Thus, VDPA1 represent tablets with 10 KgF hardness
and containing DCP, PVP, and NaCMC as ller, binder, and
disintegrating agent, respectively.
497
Fig. 3. Relationship between disintegration time and the tablet

dissolution in 15 min (closed symbols) and at 30 min (open symbols)
as function of ller used in the formulation (red squares LMH; blue
diamonds DCP). Results are expressed as mean standard
deviation for n 6
Dissolution Test
A VanKel VK 7000 dissolution system (Varian, Cary, NC,
USA) automated with a VanKel peristaltic pump that ows
sample from the dissolution vessels to an 18 cell Cary UV
visible spectrophotometer (Varian, Cary, NC, USA) was
used. Testing was performed according to the US Pharmaco
peia (USP)30 procedure for the verapamil hydrochloride
tablets by monitoring the absorbance at 278 nm (5).
Dissolution testing of selected tablets was also performed in
pH 1.2, 4.0, and 6.8 buffer media.
Disks were not mounted on the tubes and the time at which
all six tablets had disintegrated was recorded.
Data Analysis
Data analysis was performed using JMP software (ver
7.0.1, SAS Institute Inc, Cary, NC, USA). A p value of <0.05
was considered as statistically signicant.
Disintegration Test
An Electrolab ED 2 L Disintegration Tester (Globe
Pharma, New Brunswick, NJ, USA) was used to measure the
disintegration time of the tablets, in accordance with USP30
<701> Disintegration procedures for the uncoated tablets
using 900 mL of distilled water at 37C. Six tablets were
dropped into individual tubes of the basket rack assembly.
Fig. 2. Variation in the tablet dissolution in 15 min (closed diamond)

and at 30 min (open square) as function of the tablet disintegration
time. Results are expressed as mean standard deviation for n 6
The 24 set of tablets prepared using different ller,

binder, and disintegrating agent and compressed into tablets
to different hardness showed large variation in the dissolution
and disintegration time. The percentage dissolution ranged
from 16.4% to 93.5 % at 15 min and from 26.7% to 100.5% at
30 min, while the disintegration time ranged from under 1 to
over 20 min. However, no relationship between tablet
dissolution and tablet disintegration time was observed across
the 24 set of tablets (Fig. 2).
Fourteen out of 24 set of tablets failed to meet the USP
dissolution criterion of >80% (Q + 5%) dissolution in 30 min
for the verapamil tablets (4). These included all 12 set of
tablets prepared using DCP as the ller. Only two set of
tablets containing LMH as the ller failed to meet the USP
dissolution requirement. These tablets were formulated using
HPMC as the binder and were without any disintegrating
agent and compressed at the low and the high hardness
settings (VLHX1 and VLHX2). The remaining ten set of
tablets containing LMH as the ller showed >85% dissolution
at 30 min. These results show a strong dependence of
dissolution on the type of ller used to prepare the tablets.
Statistical analysis of the dissolution results conrmed a
signicant faster dissolution (p<0.0001) at 30 min in presence
of LMH as ller. A statistically signicant faster disintegra
tion (p<0.0002) was also observed for tablets prepared using
LMH as the ller as compared to those prepared using DCP
as the ller (Fig. 3).
Gupta et al.
498
A signicant inuence of disintegrating agent was also
observed on the tablet dissolution (p<0.002) and the disinte
gration time (p<0.0001; Fig. 4). Although all tablets formu
lated using NaCMC disintegrated in less than 3 min, those
containing DCP as the ller failed to meet the USP
dissolution criterion of >80% (Q + 5%) dissolution in
30 min. On the other hand, tablets formulated without any
disintegrating agent required at least 5 min for complete
disintegration, yet met the USP dissolution criterion when
formulated using LMH as the ller and PVP as the binder.
Disintegrating agent accelerates tablet disintegration into
smaller fragments increasing the surface area exposing to
the medium for dissolution of the drug to occur. The results
highlight the importance and inuence of other formulation
components, e.g., ller, binder, etc., on the dissolution process
and cautions against relying solely on the disintegrating agent
to accelerate tablet dissolution
Tablets formulated using PVP showed faster dissolution
(p<0.015) and faster disintegration (p<0.004) as compared
with those formulated using HPMC as the binder. Although
no inuence of hardness was observed on the dissolution,
tablets compressed to lower hardness disintegrated faster as
compared to those compressed to higher hardness values (p<
0.02). Signicant inuences of ller, disintegrating agent, and
binder were also observed on the tablet dissolution at 15 min
(p<0.0001, p<0.002, and p<0.012, respectively) as shown in
Pareto plot in Fig. 5 conrming the strong correlation of the
tablet dissolution with these three variables. Signicant two
way interaction effects between ller disintegrating agent,
ller binder, and binder disintegrating agent (p<0.0005, p<
0.01, p<0.03, respectively) were observed on the tablet
disintegration time (Fig. 6). The presence of these two way
interactions resulted in slower disintegration and hence
slower dissolution of the VLPA1 and VLPA2 tablets as
compared to the VLHA1 and VLHA2 tablets, although the
effect analysis of individual variable showed faster dissolution
and faster disintegration for tablets formulated using PVP as
the binder. These results highlight the importance of under
standing the effects of interaction occurring among different
Fig. 4. Relationship between disintegration time and the tablet

dissolution in 15 min (closed symbols) and at 30 min (open symbols)
as function of disintegrating agent used in the formulation (red
diamond NaCMC; black squares MCC; pink triangles none).
Results are expressed as mean standard deviation for n 6
Fig. 5. Pareto plot for dissolution at 15 min. A p value of <0.05 was

considered statistically signicant
formulation components in addition to identifying their

individual effects and are in line with the Food and Drug
Administrations quality by design initiative (6).
Out of the 12 formulations, tablets prepared from two
formulations at both hardness settings (VLHA1, VLHA2,
VLPA1, and VLPA2) had disintegration time of less than
2 min and average dissolution of >80% within 15 min when
tested under the USP30 conditions. However, one tablet each
of VLPA1 and VLPA2 had dissolution <80% at 15 min. One
of the ICH Q6A Decision Trees #7(1) criterion for using
disintegration test for drug release is >80% dissolution in
15 min across the physiological pH range (pH 1.2 6.8).
Dissolution testing of these four set of tablets in pH 1.2, 4.0,
and 6.8 buffer solutions also gave similar results. Dissolution
of all VLHA1 and VLHA2 tablets exceeded 80%, while at
least one tablet of VLPA1 and VLPA2 had dissolution <80%
at 15 min under each of the three pH conditions making only
the VLHA1 and VLHA2 tablets suitable for replacing,
dissolution testing by disintegration testing.
CONCLUSIONS
A signicant inuence of different formulation compo
nents was observed on the tablet dissolution and disintegra
tion with the ller and disintegrating agent exerting the most
signicant inuence. At constant ller or disintegrating agent,
an increase in disintegration time led to slower tablet
dissolution. However, no relationship was observed between
the disintegration and dissolution across all formulations due
to the interactions between different formulation compo
nents. Tablets, prepared at both hardness settings, of only one
Fig. 6. Pareto plot for disintegration time. A p value of <0.05 was

considered statistically signicant
Disintegration Test as a Surrogate for Dissolution Test

out of the 12 formulations met the ICH Q6A criteria of >80%
dissolution across the physiological pH range between 1.2 and
6.8. These tablets were formulated using LMH as ller,
HPMC as binder, NaCMC as disintegrating agent, and may
be suitable for using the disintegration test instead of the
dissolution test as the drug product acceptance criteria. Thus,
systematic studies are needed to determine the relationship
between the disintegration and dissolution before using the
disintegration test, instead of the dissolution test as the drug
acceptance criteria.
REFERENCES
1. United States Pharmacopeia 30/National Formulary 25. 2008.
The U.S. Pharmacopeial Convention, Inc. Rockville, MD.
(www.USP.org).
499
2. ICH Q6A Specications: Test Procedures and Acceptance
Criteria for New Drug Substances and New Drug Products:
Chemical Substances. Available at www.ich.org. Accessed January
24, 2008.
3. Sayeed V. Disintegration test as a surrogate for dissolution: some
practical considerations. In: AAPS Workshop on the Role of
Dissolution in QbD and Drug Product Life Cycle. Arlington, VA:
AAPS Workshop, 2008.
4. Vogelpoel H, Welink J, Amidon GL, Junginger HE, Midha KK,
Moller H, et al. Biowaiver monographs for immediate release
solid oral dosage forms based on Biopharmaceutics Classication
System (BCS) literature data: verapamil hydrochloride, propran
olol hydrochloride, and atenolol. J Pharm Sci 2004;93:1945 56.
5. Verapamil hydrochloride tablets monograph. United States Phar
macopeia 30/National Formulary 25. 2008. The U.S. Pharmaco
peial Convention, Inc. Rockville, MD. (www.USP.org).
6. Chen C. A FDA perspective on Quality by Design. Pharmtech,
Dec 5, 2007. Available at http://pharmtech.ndpharma.com/
pharmtech/Article/A FDA Perspective on Quality by Design/
ArticleStandard/Article/detail/469915. Accessed May 15, 2008.

DOI: 10.1208/s12249 009 9229 y
Research Article
Difference in the Lubrication Efficiency of Bovine and Vegetable-Derived
Magnesium Stearate During Tabletting
Abhay Gupta,1 Mazen L. Hamad,1,3 Mobin Tawakkul,1 Vilayat A. Sayeed,2 and Mansoor A. Khan1,4
Received 17 September 2008; accepted 17 March 2009; published online 24 April 2009
Abstract. The purpose of this work was to evaluate and compare the functionality of bovine fatty acids
derived (MgSt B) and vegetable fatty acids derived (MgSt V) magnesium stearate powders when used
for the lubrication of granules prepared by high shear (HSG) and uid bed (FBG) wet granulation
methods. The work included evaluation of tablet compression and ejection forces during tabletting and
dissolution testing of the compressed tablets. Granules prepared by both granulation methods required
signicantly lower ejection force (p<0.01) when lubricated with the MgSt V powder as compared to
those lubricated with the MgSt B powder. Granules prepared by the HSG method and lubricated with
the MgSt V powder also required signicantly lower compression force (p<0.01) to produce tablets of
similar weight and hardness as compared to those lubricated with the MgSt B powder. The dissolution
proles were not affected by these differences and were the same for tablets prepared by same granulation
method and lubricated with either magnesium stearate powder. The results indicate signicant differences
(p<0.01) between lubrication efciency of the MgSt B and the MgSt V powders and emphasize the
importance of functionality testing of the MgSt powders to understand the impact of these differences.
KEY WORDS: characterization; uid bed granulation; functionality testing; high shear granulation;
lubricant efciency; magnesium stearate.
INTRODUCTION
Magnesium stearate (MgSt) is a widely used lubricant in
the pharmaceutical industry for the manufacturing of tablet
dosage form. The recent increase in the life threatening
bovine diseases, e.g., mad cow disease, foot and mouth
disease, etc., has resulted in MgSt manufacturers switching
over from the bovine derived fatty acids to the vegetable
derived fatty acids as the starting material to synthesize MgSt
powder. This difference in the starting fatty acids results in a
change in the technical grade of the MgSt powder. A change
in the technical grade of this excipient is considered as a
major change by the United States Food and Drug Admin
istration and requires studies to demonstrate equivalence
between the products formulated before and after imple
The opinions expressed in this work are only of authors, and do not
necessarily reect the policy and statements of the FDA.
1
Division of Product Quality Research, Ofce of Testing and

Research, Ofce of Pharmaceutical Science, Food and Drug
Administration, 10903 New Hampshire Avenue, LS building 64,
Silver Spring, Maryland 20993 002, USA.
2
Division of Chemistry III, Ofce of Generic Drugs, Food and Drug
Administration, 7500 Standish Place, Rockville, Maryland 20855,
USA.
3
College of Arts and Sciences, University of Hawaii at Hilo, 200 W.
Kawili St., Hilo, Hawaii 96720, USA.
4
To whom correspondence should be addressed. (e mail: Mansoor.
khan@fda.hhs.gov)
menting this change (1,2). The primary function of MgSt

powder is to lubricate the die walls and assist in the ejection
of the compressed tablet from the die. Therefore, determina
tion of the functional equivalency of the MgSt as an efcient
die wall lubricant for powders derived from these two fatty
acid sources also becomes crucial. The authors have earlier
reported differences in the lubrication efciency of the bovine
and vegetable derived magnesium stearate powders when
used to lubricate dry granulated products (3). Granules
prepared using different granulation methods may not always
show differences in lubrication efciency between the two
grades of magnesium stearate powder; additional studies are
warranted to examine the effect of this variation on the wet
granulated product. Additional, non compendial analytical
techniques were also used to identify potential differences
between these two grades of MgSt powder.
Previous studies have indicated that pharmaceutical
product quality is sensitive to differences in the MgSt
powders, and the product quality may be compromised if it
does not perform as expected, leading to processing prob
lems, e.g., poor tablet hardness, chipping, lamination, etc. or
change in the drug product release prole (4 6). Due to this
reason, for some formulations, a switch from the bovine fatty
acids derived MgSt powder (MgSt B) to the vegetable
derived MgSt powder (MgSt V) may also necessitate a
change in the concentration of MgSt powder to obtain
equivalent lubrication efciency with the new formulation.
As demonstrated in this paper, such determination can easily
be performed on an instrumented tablet press by comparing
500
Efficiency of Bovine and Vegetable-Derived Magnesium Stearate
501
the compression and ejection forces during tabletting from

granules lubricated with either grade of MgSt powder at
different MgSt concentrations. The information gained could
then be used to change the MgSt concentration in the
formulation, if needed, and submitted to the agency as part
of the application documents.
Five percent (w/w) aqueous HPMC solution was sprayed on

the uidized powder blend, followed by drying to a moisture
content of less than 1.5% (w/w). The dry granules were
screened though a standard US no. 20 sieve.
Granules prepared by both granulation methods were

lubricated at 0.25%, 0.50%, 1.00%, and 1.50% (w/w) MgSt
concentration in a 2.0 L V blender for 5 min at 30 rpm.
Material
Two MgSt powders with similar specications but
derived from different fatty acid sources, namely, bovine
(MgSt B) and vegetable (MgSt V), were obtained from
Mallinckrodt (Hazelwood, MO, USA). A test formulation
containing ibuprofen (IBU, Albemarle, Orangeburg, SC,
USA), microcrystalline cellulose (MCC, Emcocel, JRS
Rettenmaier & Soehne, Rosenburg, Germany), and lactose
monohydrate (LMH, Foremost, Rothschild, WI, USA) in a
1:1:2 ratio was used to prepare granules for tablet compres
sion. Hypromellose, 2.4% [dry weight basis; hydroxypropyl
methylcellulose (HPMC), E15 Methocel TM LV, Dow
Chemical, Midland, MI, USA] was used as the binder. All
materials were used as received after passing through
standard US no. 20 (850 m) sieve, unless otherwise noted.
Characterization of Magnesium Stearate Powders
The MgSt B and MgSt V powders were found to be
equivalent based on the certicate of analysis provided by the
manufacturer and complied with the USP30 NF25 require
ments for the MgSt powder. Non compendial tests were
therefore used to identify potential differences between the
two grades. A 2920 CE differential scanning calorimeter
(DSC; TA Instruments, New Castle DE, USA) and a 2950
thermogravimetric analyzer (TGA) (TA Instruments) were
used for the thermal analysis from room temperature to
300C at 20C/min heating rate under constant ow of dry
nitrogen. A symmetrical gravimetric analyzer (model SGA
100, VTI Corporation, Hialeah, FL, USA) was used for
determining the equilibrium moisture content at different
relative humidity (RH) conditions between 5% and 95% RH.
Preparation of Granules
High Shear Granulation
Appropriate quantities of IBU, MCC, and LMH were
weighed and blended together in a high shear granulator
followed by the addition of a 10% (w/w) aqueous HPMC
solution over 5 min. The granules were wet massed for an
additional 5 min and hand screening though a standard US
no. 8 (2.36 mm) sieve. The resulting granules were air dried
overnight to a moisture content of less than 1.5% (w/w). The
dry granules were screened though a standard US no. 20
sieve.
Fluid Bed Granulation
Appropriate quantities of IBU, MCC, and LMH were
weighed, blended, and transferred to a uid bed granulator.
Lubrication
Tablet Compression and Characterization

The lubricated granules were compressed on a ten
station, instrumented tablet press (MiniPress I, Globe
Pharma, New Brunswick, NJ, USA) tted with load cells to
measure the tablet compression and ejection forces. Only one
set of 12 mm at faced, beveled edged, round punches and
cylindrical die was used to minimize variation arising from the
use of multiple punch die sets. The press speed was kept
constant at 20 rpm. Tablets were compressed starting with the
granules lubricated at 0.25% MgSt, followed by 0.50%,
1.00%, and 1.50% MgSt. The granules lubricated with the
MgSt B powder were compressed rst, followed by those
lubricated with the MgSt V powder. Tablet weight and
hardness were kept constant for all tablets. The die and
punch on the tablet press were cleaned using methanol after
each compression run. One hundred compression and ejec
tion force proles were collected on each lubricated granules.
All tablets were characterized for weight variation,
hardness, and dissolution. Ten tablets randomly selected from
each batch were weighed on an analytical balance followed
by measurement of the diametral crushing strength using a
tablet hardness tester (VK 200 Varian, Cary, NC, USA). In
accordance with the dissolution test for the ibuprofen tablets
from USP30, the USP II paddle method was used with
paddles rotating at 50 rpm (7). Dissolution proles were
collected for tablets by monitoring the absorbance of the
dissolution medium at 266 nm.
RESULTS
The two grades of MgSt showed no difference in the
DSC and TGA proles (Fig. 1), and the proles were in
agreement with those reported in the literature for the MgSt
powder (8,9). No difference in the equilibrium moisture
content was observed between the two grades of MgSt
powders below 60% RH (Fig. 2). However, above 60% RH,
a signicantly higher moisture uptake was observed with the
MgSt B powder.
The specications for tablets are typically based on the
tablet weight and hardness, with the target values determin
ing the compression and ejection forces on the tablet press.
Hence, the weight and hardness of the tablets were main
tained at constant values for all tablets (Fig. 3). At constant
tablet weight and hardness, an increase in the MgSt concen
tration resulted in an increase in the tablet compression force
for tablets prepared by the HSG method (Fig. 4a). This
increase, however, did not lead to an increase in the tablet
ejection force, which remained constant. Among the two
grades, signicantly higher compression and ejection forces
Gupta et al.
502
Fig. 1. DSC and TGA plots for the MgSt B and MgSt V powders
(p<0.01) were needed for tablets compressed from the HSG

granules lubricated with the MgSt B powder as compared to
those lubricated with the MgSt V powder. The tablets
prepared by the HSG method and lubricated with either
0.5% MgSt B or 0.5% MgSt V powder were compressed to
similar weights and under similar compression force (Fig. 4a).
However, a signicantly greater force (p<0.01) was needed to
eject tablets lubricated with 0.5% MgSt B powder, yet
resulted in tablets with signicantly lower hardness values
(p<0.01) as compared to those lubricated with 0.5% MgSt V
powder (Figs. 3a and 4a).
The tablets prepared by the FBG method also showed a
signicantly lower ejection force (p<0.01) for the granules
lubricated with the MgSt V powder at all MgSt concentra
tions (Fig. 4b). Similar compression force was required to
compress the FBG granules lubricated with either grade of
MgSt powder at 0.25% and 0.5% concentrations. However,
for both grades, a lower ejection force was observed for the
granules lubricated with 0.5% MgSt. Above this concentra
tion, an increase in the MgSt concentration resulted in an
increase in the tablet compression and ejection forces for
both grades. The ejection forces, although, remained lower
Fig. 3. Weight and hardness data for tablets lubricated with the
MgSt B (square) and MgSt V (triangle) powders. a Tablets prepared
by the HSG method. b Tablets prepared by the FBG method
for the MgSt V lubricated tablets as compared to the MgSt B

lubricated tablets.
Tablets prepared by the two granulation methods
showed different dissolution proles. However, dissolution
proles of the tablets prepared by the same granulation
method but lubricated with either grade of the MgSt powder
showed no difference (Fig. 5). Dissolution proles of tablets
lubricated at different concentration of MgSt also showed no
difference.
DISCUSSION
Fig. 2. Moisture sorption/desorption proles of the MgSt B (square)

and MgSt V (triangle) powders
A signicant difference was observed in the lubrication

efciency of the MgSt B and the MgSt V powders when
tested on granules prepared by the high shear and uid bed
wet granulation methods. When compressed to similar
hardness and weight, granules lubricated with the MgSt V
powder required lower ejection force on the tablet press as
compared to the granules lubricated with the MgSt B powder.
Difference was also observed in the moisture sorption proles
of the two grades, suggesting higher afnity for moisture as
well as the presence of some amorphous material in the
MgSt B powder (10). None of these differences, however,
Efficiency of Bovine and Vegetable-Derived Magnesium Stearate
503
resulted in signicant differences in the release proles of

tablets lubricated with either MgSt powder.
The variation in lubrication efciency of different MgSt
may be formulation dependent. The results of this study are
based on a single formulation containing ibuprofen [a high
permeability and low solubility (BCS class II) compound] as the
active pharmaceutical ingredient and granulated using two wet
granulation techniques. Formulations prepared with different
active ingredients, excipients, and/or using different manufac
turing processes may show different results. The tablet press
used in this study was operated at a relatively slow speed of
15 rpm. On tablet presses running at high speed, which is
normally the case in the production environment, this small
difference in lubrication efciency may also lead to problems, e.
g., capping, lamination, chipping, sticking, etc. Hence, it is
important to study and understand the effect of any such change
before switching between the two grades of MgSt powders.
CONCLUSIONS
This study highlights the importance of the functionality
testing of MgSt powders, in addition to testing for the
compendial requirements, for a comprehensive comparison
Fig. 5. Dissolution proles of tablets lubricated with 0.5% MgSt B

(square) and MgSt V (triangle) powders. a Tablets prepared by the
HSG method. b Tablets prepared by the FBG method
of MgSt powders derived from different starting materials

prior to making the source change. It also demonstrates that
the switch in the MgSt powder derived from different source
may need adjustment in the manufacturing and processing
parameters to meet the establish product specications, e.g.,
hardness, dissolution, stability, etc.
ACKNOWLEDGMENTS
Fig. 4. Compression force and ejection force data for tablets
lubricated with the MgSt B (square) and MgSt V (triangle) powders.
a Tablets prepared by the HSG method. b Tablets prepared by the
FBG method
The authors gratefully acknowledge Robbe C. Lyon,

Christopher D. Ellison, Agnes A Nguyenpho and Rakhi B.
Shah for their technical suggestions, contribution, and help
with some of these studies.
Gupta et al.
504
REFERENCES
5.
1. United States Food and Drug Administration Guidance
Document. Changes to an approved NDA or ANDA. 2004.
http://www.fda.gov/cder/guidance/3516fnl.htm (accessed 07/25/
2006).
2. United States Food and Drug Administration Guidance Docu
ment. Immediate release solid oral dosage forms: scale up and
post approval changes: chemistry, manufacturing and controls, in
vitro dissolution testing, and in vivo bioequivalence documenta
tion 1995. http://www.fda.gov/cder/guidance/cmc5.pdf (accessed
07/25/2006).
3. Hamad ML, Gupta A, Shah RB, Lyon RC, Sayeed VA, Khan
MA. Functionality of magnesium stearate derived from bovine
and vegetable sources: dry granulated tablets. J Pharm Sci.
2008;97(12):5328 40.
4. Roberts M, Ford JL, MacLeod GS, Fell JT, Smith GW, Rowe
PH, Dyas AM. Effect of lubricant type and concentration on the
6.
7.
8.
9.
10.
punch tip adherence of model ibuprofen formulations. J Pharm

Pharmacol. 2004;56:299 305.
Rowe RC. Interaction of lubricants with microcrystalline cellu
lose and anhydrous lactose a solubility parameter approach. Int
J Pharm. 1988;41:223 6.
Iranloye TA, Parrott EL. Effects of compression force, particle size,
and lubricants on dissolution rate. J Pharm Sci. 1978;67:535 9.
United States Pharmacopeia 30/National Formulary 25. 2008.
The US Pharmacopeial Convention, Inc. Rockville, MD. http://
www.USP.org (accessed 02/05/2008).
Wada Y, Matsubara T. Pseudo polymorphism and crystalline
transition of magnesium stearate. Thermochim Acta. 1992;
196:63 84.
Brauner JS, Bradby S, Chow BC, Lindauer RF, White PA.
Magnesium stearate specic surface area and hydration states of
29 commercial samples. PF. 1997;23:889 922.
Swaminathan V, Kildsig DO. An examination of the moisture
sorption characteristics of commercial magnesium stearate.
AAPS PharmSciTech. 2001;2:28.

DOI: 10.1208/s12249 009 9230 5
Research Article
Optimization of Hydrogels for Transdermal Delivery of Lisinopril
by BoxBehnken Statistical Design
Ramesh Gannu,1 Vamshi Vishnu Yamsani,1 Shravan Kumar Yamsani,1
Chinna Reddy Palem1 and Madhusudan Rao Yamsani1,2
Received 5 September 2008; accepted 17 March 2009; published online 28 April 2009
Abstract. The aim of this study was to investigate the combined inuence of three independent variables
on the permeation kinetics of lisinopril from hydrogels for transdermal delivery. A three factor, three
level Box Behnken design was used to optimize the independent variables, Carbopol 971 P (X1),
menthol (X2), and propylene glycol (X3). Fifteen batches were prepared and evaluated for responses as
dependent variables. The dependent variables selected were cumulative amount permeated across rat
abdominal skin in 24 h (Q24; Y1), ux (Y2), and lag time (Y3). Aloe juice has been rst time investigated
as vehicle for hydrogel preparation. The ex vivo permeation study was conducted using Franz diffusion
cells. Mathematical equations and response surface plots were used to relate the dependent and
independent variables. The regression equation generated for the cumulative permeation of LSP in 24 h
(Q24) was Y1 1,443.3 602.59X1 +93.24X2 +91.75X3 18.95X1X2 140.93X1X3 4.43X2X3 152.63X12
150.03X22 213.9X32. The statistical validity of the polynomials was established, and optimized
formulation factors were selected by feasibility and grid search. Validation of the optimization study
with 15 conrmatory runs indicated high degree of prognostic ability of response surface methodology.
The use of Box Behnken design approach helped in identifying the critical formulation parameters in the
transdermal delivery of lisinopril from hydrogels.
KEY WORDS: Box Behnken; factorial design; hydrogel; lisinopril; optimization.
INTRODUCTION
Lisinopril (LSP) is an angiotensin converting enzyme
inhibitor used for the treatment of hypertension and conges
tive heart failure and to alleviate strain on hearts damaged as
a result of a heart attack. LSP is slowly and incompletely
absorbed after oral administration with a bioavailability of
25 30% (1,2). LSP is available only in the form of oral tablets
in the market. However, this formulation has a major
disadvantage since it is incompletely absorbed from the
gastrointestinal tract. To overcome the problem of incomplete
absorption, low oral bioavailability, and for the effective
treatment of chronic hypertension, alternative long acting
formulations could be benecial. Transdermal route of
administration may be a good alternative to circumvent these
problems and is recognized as one of the potential route for
the local and systemic delivery of drugs. LSP was selected as a
model drug for the investigation because of its clinical need,
low oral dose (2.5 20 mg), low molecular mass (405.5 g/mol),
and low oral bioavailability (25%). Avoidance of hepatic rst
pass elimination, decrease in side effects, improved patient
1
National Facilities in Engineering and Technology with Industrial

Collaboration (NAFETIC) Centre, University College of Pharma
ceutical Sciences, Kakatiya University, Warangal, 506 009, Andhra
Pradesh, India.
2
To whom correspondence should be addressed. (e mail: ymrao123@
yahoo.com)
compliance, interruption or termination of treatment when

unnecessary, as well as maintaining suitable plasma concen
tration for longer duration through a non invasive zero order
delivery are the well documented advantages of this route of
administration (3). However, the highly organized structure of
stratum corneum forms an effective barrier to the permeation
of drugs, which must be modied if poorly penetrating drugs
are to be administered. The use of penetration enhancers
would signicantly increase permeation of drugs through skin.
Terpenes present in naturally occurring volatile oils appear to
be clinically acceptable enhancers (4); a wide variety of
terpenes have been shown to increase the percutaneous
absorption of number of drugs (5). In the present study,
menthol was used as penetration enhancer.
In the development of transdermal dosage form, an
important issue was to design an optimized pharmaceutical
formulation with appropriate penetration rate within a short
time period and minimum trials. Traditional experiments
require more effort, time, and materials when a complex
formulation needs to be developed. Various experimental
designs (6 8) are useful in optimizing formulations, requiring
less experimentation and providing estimates of the relative
signicance of different variables. Response surface method
ology (RSM) is a widely practiced approach in the develop
ment and optimization of drug delivery devices (9,10). In this
investigation, we explored the utility of RSM for the
optimization of hydrogels. Based on the principle of design
of experiments, the methodology encompasses the use of
505
Gannu et al.
506
various types of experimental designs, generation of polyno
mial equations, and mapping of the response over the
experimental domain to optimize the hydrogels. The tech
nique requires minimum experimentation and time, thus
proving to be far more effective and cost effective than the
conventional methods of formulating dosage forms.
The work reported in this paper, a Box Behnken design
was used to optimize hydrogels containing lisinopril as drug
and Carbopol 971P as gelling agent. Independent variables
selected were Carbopol 971P (X1), menthol (X2), and
propylene glycol (X3) to evaluate their separate and com
bined effects on permeation of LSP in 24 h (Q24) across rat
abdominal skin, ux, and lag time as dependent variables.
30 min to separate the bers. The supernatant (Aloe vera

juice) was used for the preparation of hydrogels.
The composition of the gels was detailed in Tables I and
II. Carbopol hydrogels were prepared by dispersing in small
aliquots of Carbopol 971P (CP 971P) into aloe juice contain
ing 10% v/v ethanol and propylene glycol. After continuous
stirring at 1,000 rpm for 1 2 min, LSP and menthol were
added, and the contents were stirred for 6 h. The pH was
adjusted to 7 with 1% v/v triethanolamine.
Formulation LSP7W using water as vehicle was prepared
to study the effect of vehicle on permeation of LSP from
hydrogel. It was prepared by dispersing CP 971P in water
containing 10% v/v ethanol and propylene glycol. After
continuous stirring at 1,000 rpm for 1 2 min, LSP and
menthol were added, and the contents were stirred for 6 h.
The pH was adjusted to 7 with 1% v/v triethanolamine.
Materials
pH Evaluation
Lisinopril was a gift sample from M/s Dr Reddys
Laboratories, Hyderabad, India. Menthol was purchased
from Merck, Mumbai, India. Carbopol 971P was gift sample
from Lubrizol Advanced Materials India Pvt. Ltd, Mumbai,
India. All other chemicals used were of analytical grade.
Preparation of Aloe Juice and LSP Hydrogels
Aloe vera juice was prepared from the full size mature
leaves of Aloe vera. They were cut from the plant, and the
green rind was removed. From the cut leaf bases, the yellow
juice was allowed to drain into a container to remove the
exudates. The colorless parenchyma was ground in a blender
(Remi, Mumbai, India) and centrifuged at 3,000 rpm for
The pH of the hydrogels was recorded with pH meter

(Elico, India), by bringing the glass electrode in contact with
the hydrogel and allowing it to equilibrate for 1 min. Experi
ments were performed in triplicate to check for the neutral
ization of gels. pH evaluation was carried out for all
experimental formulations in triplicate.
Rheological Measurements
The rheological properties of hydrogels were measured
using Brookeld Programmable DVIII+ Digital Rheometer
(Brookeld Engineering Laboratories Inc., Massachusetts,
USA). The rheological measurements were performed using a
Table I. Variables and Observed Responses in Box Behnken Design for Hydrogels
Independent variables
Batch
LSP1
LSP2
LSP3
LSP4
LSP5
LSP6
LSP7
LSP8
LSP9
LSP10
LSP11
LSP12
LSP13
LSP14
LSP15
Independent
variables
X1 Carbopol
971P (% w/w)
X2 Menthol
(% w/w)
X3 Propylene
glycol (% v/v)
X1%
(w/w)
X2%
(w/w)
X3%
(v/w)
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
Dependent variables
Y2
Y1 (g) (g h
1,597.3
589.1
1,730.2
646.1
1,943.7
861.3
2,429.7
783.6
1,282.3
1,569.2
1,454.0
1,723.2
1,468.9
1,415.3
1,445.7
Levels used, Actual (Coded)
Gel
Assay
Low
K (cm 1) (1)
cm 2) Y3 (h) pH index (%)
13.89
5.65
15.98
6.50
17.67
9.39
22.67
6.91
10.01
11.10
11.45
14.51
12.29
11.75
11.67
1.86
7.01
1.84
5.58
1.79
4.19
1.17
6.01
2.09
2.12
1.93
1.91
2.14
2.25
1.80
6.4
6.9
6.8
7.3
6.8
7.4
6.8
7.0
6.9
6.7
7.1
7.0
6.8
7.1
6.9
0.89
2.11
1.05
2.40
1.03
2.21
1.11
2.35
1.58
1.52
1.86
1.74
1.65
1.72
1.68
99.8
101.0
98.6
102.3
99.4
99.6
100.8
98.6
100.2
98.6
101.6
99.1
99.0
99.7
100.0
Medium
(0)
High
(+1)
0.25
0.63
12
10
15
1.38
0.56
1.58
0.64
1.75
0.93
2.25
0.68
0.99
1.10
1.13
1.44
1.22
1.16
1.16
507
Table II. Composition of Checkpoint Formulations, the Predicted and Experimental Values of Response Variables, and Percentage Prediction
Error
Optimized formulation
composition (X1/X2/X3)
Response
variable
Experimental value
Predicted value
Percentage
prediction error
0.285:10.12:8.34
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
Y1
Y2
Y3
1,873.47
18.09
1.66
1,551.96
13.99
2.13
1,594.98
13.67
1.58
1,479.38
12.20
1.83
1,626.25
15.04
1.68
2,014.73
16.53
1.43
1,390.42
12.21
1.84
1,659.80
14.24
1.40
1,812.36
16.78
2.08
1,651.27
14.12
1.37
1,785.16
15.62
1.71
1,453.12
12.27
1.60
1,564.55
13.01
1.90
1,896.95
14.82
1.46
1,782.05
15.82
1.52
1,873.40
17.03
1.78
1,481.65
12.17
2.19
1,679.32
14.16
1.49
1,515.63
12.31
1.78
1,700.96
14.38
1.64
1,923.91
17.35
1.37
1,504.84
12.37
1.72
1,720.56
14.64
1.47
1,882.14
17.15
1.99
1,693.58
13.77
1.40
1,840.24
16.85
1.54
1,510.65
12.42
1.70
1,692.00
14.88
1.73
1,768.47
15.31
1.67
1,807.08
16.34
1.65
0.004
5.860
7.229
4.530
13.009
2.817
5.288
3.584
5.696
2.450
0.902
2.732
4.594
4.388
2.381
4.508
4.961
4.196
8.229
1.310
6.522
3.661
2.809
5.000
3.850
2.205
4.327
2.562
2.479
2.190
3.085
7.875
9.942
3.959
1.222
6.250
8.146
14.373
8.947
6.773
3.306
14.384
1.405
3.287
8.553
0.648:11.00:12.30
0.448:8.64:8.12
0.515:5.22:11.12
0.438:10.75:9.71
0.345:9.35:11.75
0.558:6.98:7.82
0.340:4.10:12.48
0.276:11.68:10.66
0.484:9.52:5.90
0.254:5.40:11.45
0.384:4.05:5.90
0.289:5.64:6.62
0.375:10.60:9.50
0.288:7.52:8.84
controlled stress rheometer with the cone (24 mm) and plate
geometry. The viscosity was determined by torque sweep from
10% to 110%. All the measurements were performed in trip
licate at 25C. The equilibrium time before every measurement
was 5 min, and the sample volume used was approximately
1 mL. Calculation of rheological properties was performed
using Rheocalc 32 software (Brookeld Engineering Laborato
ries Inc., USA).
Determination of Drug Content
Weighed quantity of about 1.0 g of hydrogel was
dissolved in 100 mL of distilled water, sonicated for 10 min
using bath sonicator, and ltered through 0.45 m membrane

lter. The ltrate was suitably diluted, and the drug content in
the sample was determined using high pressure liquid chro
matography (HPLC).
Preparation of Rat Abdominal Skin
Albino rats weighing 150 200 g were killed using
anesthetic ether. The hair of test animals was carefully
removed with electrical clippers, and the full thickness of
skin was removed from the abdominal region. The epidermis
was prepared surgically by heat separation technique (11),
which involved soaking the entire abdominal skin in water at
Gannu et al.
508
60C for 45 s, followed by careful removal of the epidermis.
The epidermis was washed with water and used for ex vivo
permeability studies.
found to be within limits [percent coefcient of variation (%

CV) was less than 15%]. Sample preparation briey involved
the ltration of sample through 0.4 membrane lter, diluted
with mobile phase, and 20 L was spiked into column.
Ex vivo Permeation Studies

Franz diffusion cell with a surface area of 3.56 cm2 was
used for ex vivo permeation studies. The rat abdominal skin
was mounted between the compartments of the diffusion cell
with stratum corneum facing the donor compartment. About
1.0 g of gel was placed in donor compartment. The receiver
phase is 12 mL of phosphate buffer saline (PBS) pH 7.4,
stirred at 400 rpm on a magnetic stirrer; the whole assembly
was kept at 370.5C. The amount of drug permeated was
determined by removing 1 mL of sample at appropriate time
intervals up to 24 h; the volume was replenished with an
equal volume of PBS pH 7.4. The drug content in the samples
was determined by HPLC, and the concentration was
corrected for sampling effects according to the Eq. 1 (12).
Cumulative amounts of drug permeated in microgram
per square centimeter were plotted against time, drug ux (g
h 1 cm 2) at steady state was calculated by dividing the slope
of the linear portion of the curve by the area of the exposed
skin surface (3.56 cm2), and the permeability coefcient was
deduced by dividing the ux by initial drug load as shown in
Table I.
Cn1 Cn VT =VT

1
VS Cn1
Cn1
where Cn1 is the corrected concentration of the nth sample,

Cn is the measured concentration of lisinopril in the nth
sample, Cn 1 is the measured concentration of the lisinopril in
the (n 1)th sample, VT is the total volume of the receiver
uid, and VS is the volume of the sample drawn. The
theoretical (required) ux was calculated using Eq. 2 (13).
JTarget
CSS CLT BW
A
A represents the surface area of the hydrogel applied over the

excised skin (i.e., 3.56 cm2), BW the standard human body
weight of 60 kg, CSS the LSP concentration at the therapeutic
level (70 ng/mL), and the CLT the total clearance (6.36 L/h) (14);
the calculated target ux value for LSP was 7.50 g h 1 cm 2.
HPLC Analysis of LSP
Analysis of samples was performed with a Shimadzu
HPLC system equipped with LC 10AT pump, UV Vis
spectrophotometric detector (SPD 10A), and C18 column
(Phenomenex; 25 cm4.6 mm; 5 m) at ambient temperature.
The mobile phase used was a mixture of phosphate buffer
(25 mM potassium dihydrogen ortho phosphate, pH 5.0) and
acetonitrile at a ratio of 88:12. A ow rate of 1 mL/min was
maintained, and the detection wavelength was 215 nm. A
calibration curve was plotted for LSP in the range of 50
2,500 ng/mL. A good linear relationship was observed
between the concentration of LSP, and the peak area of
LSP with a correlation coefcient (r2 =0.999). The required
studies were carried out to estimate the precision and
accuracy of the HPLC method of analysis of LSP and were
Experimental Design
Box Behnken statistical screening design was used to
statistically optimize the formulation factors and evaluate
main effects, interaction effects, and quadratic effects on the
amount permeated in 24 h (Q24), ux, and lag time. RSM,
such as the Box Behnken, model possible curvature in the
response function (15,16). A three factor, three level Box
Behnken design was used to explore quadratic response
surfaces and constructing second order polynomial models
with Design Expert (Version 7.1, Stat Ease Inc., Minneapolis,
MN, USA). The Box Behnken design was specically
selected, since it requires fewer runs than a central composite
design in cases of three or four variables. This cubic design is
characterized by set of points lying at the mid point of each
edge of a multidimensional cube and center point replicates
(n=3), whereas the missing corners help the experimenter
to avoid the combined factor extremes (17). A design matrix
comprising of 15 experimental runs was constructed. The
non linear computer generated quadratic model is given as
Y = b0 + b1X1 + b2X2 + b3X3 + b12X1X2 + b13X1X3 + b23X2X3
+ b11X12 + b22X22 + b33X32, where Y is the measured response
associated with each factor level combination; b0 is an
intercept; b1 to b33 are regression coefcients computed
from the observed experimental values of Y; and X1, X2,
and X3 are the coded levels of independent variables. The
terms X1, X2 and Xi2 (i=1, 2, or 3) represent the interaction
and quadratic terms, respectively (15,16). The dependent and
independent variables selected were shown in Table I along
with their low, medium, and high levels, which were selected
based on the results from preliminary experimentation. The
concentration of CP 971(X1), menthol (X2), and PG (X3)
used to prepare the 15 experimental trials, and the respective
observed responses are given in Table I.
Optimization Data Analysis and Optimization

Model Validation
Statistical validation of the polynomial equations gener
ated by Design Expert was established on the basis of analysis
of variance provision in the software. A total of 15 runs were
generated. The models were evaluated in terms of statisti
cally signicant coefcients and R2 values. Various feasibility
and grid searches were conducted to nd the optimum
parameters. Various 3 D response surface graphs were
provided by the Design Expert software. By intensive grid
search performed over the whole experimental region, 15
optimum checkpoint formulations were selected to validate
the chosen experimental domain and polynomial equations.
The optimized checkpoint formulation factors were
evaluated for various response properties. The resultant
experimental values of the responses were quantitatively
compared with the predicted values to calculate the
percentage prediction error.
509
Y2 Flux 11:90
RESULTS
5:22X1 0:89X2 0:92X3

1:87X1 X3 0:49X2 X3
0:31X1 X2
Characterization of Gels
The prepared hydrogels were slight yellowish in color
with slight aromatic odor. The pH of all the 15 hydrogel
formulations prepared as per Box Behnken design and
optimized formulations were found in the range of 6.4 7.4
after neutralization with triethanolamine. The content of LSP
in hydrogels varied from 98.6% to 102.3%.
1:89X22 1:76X32
0:50X12
Y3 lag time 2:06 2:01X1
0:18X2 0:10X3
0:35X1 X2 0:61X1 X3
1:64X12
0:36X22
0:01X2 X3
0:42X32
Rheological Measurements
CP 52 spindle was used for the viscometric characteriza
tion of hydrogels, as the working range for this spindle as
reported by manufacturers is 400 7,500 cps (at 15 rpm spindle
speed) or 4,500 35,000 cps (at 0.25 rpm). The decrease in
viscosity of the gels observed with an increasing shear rates
can be described well by an exponential function, and hence,
the obtained data were analyzed using the Power Law (18)
as expressed by the following equation, respectively:
k Dn
where is shear stress; K is gel index (GI) or consistency

index; D is shear rate; and n is ow index. Rheocalc 32
software was used to automatically apply the model to
generated data, and the value of GI was recorded. The GI
value for different formulations is presented in Table I. The
gel index was found to be ranging from 0.89 to 2.40.
Ex vivo Skin Permeation Experiments

The amount of LSP permeated from 15 experimental
runs was found to be ranging from 589.1 to 2,429.7 g in 24 h
with a ux of 5.65 to 22.67 g h 1 cm 2. The lag times were
found to be decreased signicantly from 7.01 to 1.17 h.
Experimental run 7 (LSP7) showed maximum amount of LSP
(2,429.7 g) permeated through rat abdominal skin among all
the experimental runs. Formulation LSP7W composed of
water as vehicle in hydrogel preparation was prepared to
investigate the effect of vehicle on permeation. The
formulation showed 1,087.1 g of LSP permeated in 24 h
with a ux of 11.5 g h 1 cm 2 and permeation coefcient of
1.2 cm/h. The rat skin permeation prole of LSP exhibited
zero order permeation at a constant penetration rate (r2,
0.902 0.992). The independent variables and response (Q24,
ux and lag time) of these model formulations were shown in
Table I. Figure 2 shows the permeation proles of hydrogel
formulations through excised rat abdominal skin.
Experimental Design
DISCUSSION
The independent variables and the responses for all 15
experimental runs are given in Table I. The effect of variables
on responses is shown in Fig. 1. The 15 experimental formula
tions of hydrogels were prepared using CP 971P polymer. The
responses, Q24 (Y1) and ux (Y2), were found to be signicantly
higher (Y1, 1,282.3 2,429.7 g; Y2, 9.39 to 22.67 g h 1 cm 2)
only when the CP 971P and menthol were used at 0.25% or
0.63% and 8% or 12% w/w concentration level, respectively.
The lag time (Y3) was found to be signicantly lower (Y3, 1.17
2.25 h) at low to high levels of menthol and PG. The ranges of
other responses, Y1, Y2, and Y3 were 589.1 2,429.7 g, 5.65
22.67 g h 1 cm 2, and 1.17 7.01 h, respectively.
The responses of these model formulations ranged from
a low drug penetration of 589.1 g (LSP2, CP 971 1%,
menthol 4%, and PG 10%) to a higher penetration of
2,429.7 g (LSP7, CP 971P 0.25%, menthol 8%, and PG
15%). For estimation of quantitative effects of the different
combination of factors and factor levels on the Q24, ux, and
lag time, the response surface models were calculated with
Design Expert software by applying coded values of factor
levels. The model described could be represented as:
Y1 Q24 1; 443:3
602:59X1 93:24X2 91:75X3
18:95X1 X2
152:63X12
140:93X1 X3
150:03X22
4:43X2 X3
213:9X32
Ex vivo Skin Permeation Experiments

In our preliminary study, menthol showed a potential
enhancement effect on LSP permeation through rat abdom
inal skin. However, menthol had low solubility in aqueous
systems; a co solvent is required to improve its solubility.
Additionally, some reports (19) have indicated that specic
combinations of vehicles and enhancers such as menthol in
ethanol, and PG had shown an increased in drug penetration.
In our previous investigation Aloe juice had showed an
improvement in percutaneous absorption of LSP (20).
Therefore, in this study, the combination of Aloe juice
containing 10% v/v ethanol and PG was used in the
preparation of Carbopol based hydrogels. PG, menthol, and
Aloe juice were used as multi enhancers to produce the
synergistic enhancement effect on penetration rate of lisino
pril and to decrease the used amount of enhancers. Ethanol
was used to solubilize menthol, and it also possesses
penetration enhancement properties. Formulation LSP7
showed maximum Q24 and ux among the hydrogels and
was also showed statistically signicant (P<0.05) difference
compared to that of Q24 and ux of LSP7W. The results
showed that the use of Aloe juice in hydrogel preparation
showed signicant improvement in skin permeation of LSP.
In order to quickly obtain an optimal formulation with fewer
experimental trials and quantify the effect of these enhancers,
Gannu et al.
510
Table III. Summary of Results of Regression Analysis for Responses Y1, Y2, and Y3 for Fitting to Quadratic Model
Quadratic model
R2
Adjusted R2
Predicted R2
SD
% CV
Regression equations of the tted quadratic model
Response (Y1)
0.9804
0.9451
0.6921
118.05
8.45
Response (Y2)
0.9782
0.9392
0.6638
1.10
9.11
Response (Y3)
0.9783
0.9329
0.6849
0.45
15.54
Y1 1,443.3 602.59X1 +93.24X2 +91.75X3 18.95X1X2

140.93X1X3 4.43X2X3 152.63X12 150.03X22 213.9X32
Y2 11.90 5.22X1 +0.89X2 +0.92X3 0.31X1X2 1.87X1X3 +
0.49X2X3 +0.50X12 1.89X22 +1.76X32
Y3 2.06+2.01X1 0.18X2 +0.10X3 0.35X1X2 +0.61X1X3
0.01X2X3 +1.64X12 +0.36X22 0.42X32
a computer optimization technique, including uniform design,

Design Expert 2007, was used.
Ex vivo Permeation Kinetics
In order to develop a transdermal drug delivery system
for localized and systemic delivery, it is necessary to check the
release/permeation prole. The description of permeation
proles by a model function has been attempted using
different kinetics (zero order, rst order and Higuchi
square root model) and using the following equation (Eq. 7)
derived by Korsmeyer et al. (21).
Mt=M Kt n
where Mt/M is the fractional permeation of drug, Mt is the

amount released at time t, M is the total amount of drug
contained in the hydrogel, t is time, K is the kinetic constant,
and n is the diffusional release exponent indicative of the
operating release mechanism. The drug permeation from
Hydrogels followed zero order as the most appropriate model
describing permeation kinetics (correlation coefcient be
tween 0.902 and 0.992). On the other hand, n values (0.119<n
>0.254) indicated that amount of permeated drug was by
Fickian diffusion (22).
Fitting of Data to the Model
A three factor, three level Box Behnken statistical ex
perimental design as the RSM requires 15 experiments.
Formulation LSP7 showed a signicantly higher amount of
drug permeation (Y1, Q24) and higher ux (Y2) among the
experimental runs. All the responses observed for 15
formulations prepared were simultaneously t to rst order,
second order, and quadratic models using Design Expert
7.1.5. It was observed that the best t model was quadratic
model and the comparative values of R2, SD, and %CV are
given in Table III along with the regression equation
generated for each response. A positive value represents an
effect that favors the optimization, while a negative value
indicates an inverse relationship between the factor and the
response (23). It is evident that all the three independent
variables, viz., the concentration of CP 971P (X1), menthol
(X2), and PG (X3) have positive effects on the three
responses, viz., Q24 (Y1), ux (Y2), and lag time (Y3).
The quantitative effects of the different combination of
factors and factor levels on the Q24, ux, and lag time were
calculated using response surface models. The signicant P
values (P<0.05), R2, adjusted R2, and coefcient of variation
values of this model indicated that the assumed regression
model was signicant and valid for each considered response.

The values of the coefcients in the model are related to the
effect of these variables on the response. From this model,
menthol, PG in Aloe juice containing 10% v/v ethanol as
vehicle system was the best, indicating that the combination
of above system had the greatest potential inuence on the
penetration of lisinopril from hydrogel system. The
penetration enhancers used in the investigation works by
different mechanisms. Menthol acts by modifying the solvent
nature of the stratum corneum and improves drug par
titioning into the tissue (24). PG alters the thermodynamic
activity of the drug from the reservoir, which would in turn
modify the driving force for diffusion, solvent may partition
into the tissue facilitating uptake of the drug into skin, and
there may be some minor disturbance to intercellular lipid
packing within the stratum corneum bilayers (24). The
incorporation of ethanol in Aloe juice also improved the
LSP permeation across skin apart from solubilizing effect.
The probable mechanism for the enhancement could be due
to the permeation of ethanol into the stratum corneum and
can alter the solubility properties of the tissue with a
consequent improvement for drug partitioning into the
membrane (25). In addition, ethanol as a volatile solvent
may extract some of the lipid fraction from the stratum
corneum and improve drug ux through skin. Therefore, the
incorporation of multi enhancers in the hydrogels could
enhance the LSP permeation that could meet the target ux.
The 3 D response surfaces (Fig. 1d f) were drawn to
estimate the effects of the independent variables on response
and to select the optimal formulation. The required ux to
reach therapeutic concentration calculated was found to be
about 7.5 g h 1 cm 2. Hence, the penetration rate of optimal
formulations in the optimization process was set at above
7.50 g h 1 cm 2. Formulation LSP7 showed a ux of 22.67 g
h 1 cm 2 could meet the target ux (7.50 g h 1 cm 2),
calculated from the pharmacokinetic parameters of LSP,
indicating that the concentrations may be enough to elicit
the pharmacological effect.
Data Analysis
Formulations LSP7, LSP5, LSP12, and LSP3 had the
highest Q24 and ux. Table II shows the observed and
predicted values with residuals and percent error of responses
Fig. 1. Contour plot showing effect of a Carbopol 971P (X1) and

menthol (X2); b Carbopol 971P (X1) and propylene glycol (X3); c
menthol (X2) and propylene glycol (X3) on response Y2 (ux);
corresponding response surface plots d f
511
Gannu et al.
512
Fig. 2. a, b Permeation proles of lisinopril from hydrogels
for all the formulations. The Q24 and ux (dependent

variable) obtained at various levels of the three independent
variables (X1, X2, and X3) was subjected to multiple
regression to yield a second order polynomial equation (full
model). The value of the correlation coefcient (r2) of Eq. 4
was found to be 0.9804, indicating good t (Table III). The
Q24 values measured for the different formulations showed
wide variation (i.e., values ranged from a minimum of
589.1 g in LSP2 to a maximum of 2,429.7 g LSP7). The
results clearly indicate that the Q24 value is strongly affected
by the variables selected for the study. The main effects of X1,
X2, and X3 represent the average result of changing one

variable at a time from its low level to its high level. The
interaction terms (X1X2, X1X3, X2X3, X12 , X22 , and X32 ) show
how the Q24 changes when two variables are simultaneously
changed. The negative coefcients for all three independent
variables indicate an unfavorable effect on the Q24, while the
positive coefcients for the interactions between two
variables indicate a favorable effect on Q24. Among the
three independent variables, the lowest coefcient value is for
X1 ( 602.59), indicating that this variable is insignicant in
prediction of Q24.
513
maximum of 22.67 g h 1 cm 2 in LSP7). The interaction

terms (X1X2, X1X3, X2X3, X12 , X22 , and X32 ) show how the
ux changes when two variables are simultaneously changed.
The positive coefcients (X1X3) for the interactions between
two variables indicate a favorable effect on ux. Among the
three independent variables, the lowest coefcient value is for
X1 ( 5.22), indicating that this variable is insignicant in
prediction of ux.
The value of the correlation coefcient (r2) of Eq. 6 was
found to be 0.9783, indicating good t (Table III). Among the
independent variables selected and their interactions, only X1
and X3 (Eq. 6) were found to be signicant (P<0.05),
indicating a major contributing effect of X1 and X3 on lag
time. A positive value of the coefcient for X1 (CP 971P) and
X3 (propylene glycol) indicates a favorable effect on lag time.
The hydrogels composed of high propylene glycol and low
polymer concentration yielded low lag time.
Contour Plots and Response Surface Analysis
Two dimensional contour plots and 3 D response surface
plots are shown in Fig. 1, which are very useful to study the
interaction effects of the factors on the responses. These types
of plots are useful in study of the effects of two factors on the
response at one time. In all the presented gures, the third
factor was kept at a constant level. All the relationships
among the three variables are non linear, although they
exhibit a nearly linear relationship of factor X2 with factors
X1 and X3, in the form of almost straight lines up to the
medium level of menthol concentration (Fig. 1). At higher
concentrations of menthol, these become curvilinear or non
linear. Factors X2 and X3 have curvilinear relationship at all
levels of the two variables on the response Y2. Response
surface plots show the relationship between these factors
even more clearly. The Q24 and ux were found to be
increased with increasing concentrations of either menthol
(up to medium level) or PG at constant concentration of
polymer.
Optimization
Fig. 3. Linear correlation plots a c between actual and predicted

values
The value of the correlation coefcient (r2) of Eq. 5 was

found to be 0.9782, indicating good t (Table III). The ux
values of LSP7, LSP5, LSP12, and LSP3 were found to be
more among the experimental trials. The ux values were
found to be increased from medium to high levels of variables
X2 and low to high levels of X3. The ux values measured for
the different formulations showed wide variation (i.e., values
ranged from a minimum of 5.65 g h 1 cm 2 in LSP2 to a
The optimum formulation was selected based on the

criteria of attaining the maximum value of Q24, target ux,
and low value of lag time by applying constraints on Y1
(1,000Y2,500), Y2 (7.5Y25), and Y3 (0.50Y2.50).
Upon trading of various response variables and comprehen
sive evaluation of feasibility search and exhaustive grid
search, the formulation composition with CP 971P concen
tration of 0.64%, menthol 7.90%, and propylene glycol 14.9%
was found to fulll the maximum requisite of an optimum
formulation because of maximum Q24, target ux, and low lag
time values.
Validation of Response Surface Methodology
Fifteen checkpoint formulations were obtained from the
RSM, the composition and predicted responses of which are
listed in Table II. To conrm the validity of the calculated
optimal parameters and predicted responses, the optimum
formulations were prepared according to the above values of
the factors and subjected to ex vivo permeation studies. From
Gannu et al.
514
the results presented in Table II, the predicted error is below
15%, indicating that the observed responses were very close
to the predicted values. Percentage prediction error is helpful
in establishing the validity of generated equations and to
describe the domain of applicability of RSM model. Linear
correlation plots between the actual and the predicted
response variables were shown in Fig. 3. The linear correla
tion plots drawn between the predicted and experimental
values demonstrated high values of R2 (Q24, 0.967; ux, 0.948;
and lag time, 0.875) indicating goodness of t. Thus, the low
magnitudes of error as well as the R2 values in the present
investigation prove the high prognostic ability of the RSM.
6.
7.
8.
9.
CONCLUSIONS
Optimization of hydrogel formulation is a complex
process that requires one to consider a large number of
variables and their interactions with each other. The present
study conclusively demonstrates the use of a Box Behnken
statistical design is valid for predicting the Q24, ux, and lag
time in optimization of hydrogel formulations. The derived
polynomial equations and contour plots aid in predicting the
values of selected independent variables for preparation of
optimum hydrogel formulations with desired properties.
Hydrogels could be prepared having suitable ux. Further
studies are recommended to prove the therapeutic efcacy by
pharmacokinetic or/and pharmacodynamic studies in humans.
10.
11.
12.
13.
14.
15.
ACKNOWLEDGMENTS
One of the author (Ramesh Gannu) thank All India
Council for Technical Education (AICTE), New Delhi, India
for providing nancial assistance in the form of National
Doctoral Fellowship (NDF). The authors are thankful to Dr
Reddys Laboratories, Hyderabad, India and Lubrizol Ad
vanced Materials India Pvt. Ltd, Mumbai, India for providing
gift samples of lisinopril and Carbopol 971P respectively.
16.
17.
18.
19.
REFERENCES
1. Lancaster SG, Todd PA. Lisinopril: A preliminary review of its
pharmacodynamic and pharmacokinetic properties and thera
peutic use in hypertension and congestive heart failure. Drugs
1988;35:646 69.
2. Marc H. Angiotensin converting enzyme inhibitors, angiotensin
and calcium channel blockers. In: Williams AD, Thamas LL,
editors. Foyes principles of medicinal chemistry. 5th ed.
Baltimore, MD: Lippincott Williams and Wilkins; 2002. p. 533
61.
3. Barry BW. Mode of action of penetration enhancers in human
skin. J Control Release. 1987;68:85 97.
4. Williams AC, Barry BW. Species differences in percutaneous
absorption of nicorandil. Pharm Res. 1991;8:17 24.
5. Moghimi H, Williams AC, Barry BW. A lamellar matrix model
for stratum corneum intercellular lipids. V. Effects of terpene
20.
21.
22.
23.
24.
25.
penetration enhancers on the structure and thermal behaviour of

the matrix. Int J Pharm. 1997;146:41 54.
Chang JS, Huang YB, Hou SS, Wang RJ, Wu PC, Tsai YH.
Formulation optimization of meloxicam sodium gel using
response surface methodology. Int J Pharm. 2007;338:48 54.
Sibanda W, Pillay V, Danckwerts MP, Viljoen AM, Vuuren S,
Khan RA. Experimental design for the formulation and optimi
zation of novel cross linked oilispheres developed for in vitro
site specic release of mentha piperita oil. AAPS PharmSciTech.
2004;5(1):Article 18.
Dayal P, Pillay V, Babu RJ, Singh M. Box Behnken experimental
design in the development of a nasal drug delivery system of
model drug hydroxyurea: characterization of viscosity, in vitro
drug release, droplet size, and dynamic surface tension. AAPS
PharmSciTech. 2005;6(4):Article 72.
Anitha P, Mohammad THN, Khan MA. Response surface
methodology for optimization and characterization of limo
nene based coenzyme Q10 self nanoemulsied capsule dosage
form. AAPS PharmSciTech. 2004;5(4) Article 66.
Lewis GA, Mathieu D, Luu RPT. Pharmaceutical experimental
design. New York: Marcel Dekker; 1999.
Levang AK, Zhao K, Singh J. Effect of ethanol/propylene glycol
on the in vitro percutaneous absorption of aspirin, biophysical
changes and macroscopic barrier properties of the skin. Int J
Pharm. 1999;181:255 263.
Hayton WL, Chen T. Correction of perfusate concentration for
sample removal. J Pharm Sci. 1982;71:820 821.
Suwanpiodokkul N, Thongnopnua P, Umprayan K. Transdermal
delivery of zidovudine (AZT): the effects of vehicles, enhancers,
and polymer membranes on permeation across cadaver pig skin.
AAPS PharmSciTech. 2004;5:1 7, Article 48.
Clarkes Analysis of Drugs and Poisons (2007) http://www.
medicinescomlete.com.
Govender S, Pillay V, Chetty DJ, Essack SY, Dangor CM,
Govender T. Optimisation and characterisation of bioadhesive
controlled release tetracycline microspheres. Int J Pharm.
2005;306:24 40.
Chopra S, Pati GV, Motwani SK. Release modulating hydrophil
ic matrix systems of losartan potassium: optimisation of formu
lation using statistical experimental design. Eur J Pharm
Biopharm. 2007;66(1):73 82.
Box GEP, Behnken DW. Some new three level designs for the
study of quantitative variables. Technometrics 1960;2:455 75.
Bonacucina G, Martelli S, Palmieri GF. Rheological, mucoadhe
sive and release properties of Carbopol gels in hydrophilic
cosolvents. Int J Pharm. 2004;282:115 30.
Jantharaprapap R, Stagni G. Effects of penetration enhancers on
in vitro permeability of meloxicam gels. Int J Pharm.
2007;343:26 33.
Gannu R, Yamsani VV, Yamsani MR. Enhancement potential of
Aloe vera on permeation of drugs with diverse lipophilicites
across rat abdominal skin. Curr Trends Pharm Biotechnol.
2008;2(4):548 54.
Korsmeyer RW, Gurny R, Doelker E, Buri P, Peppas NA.
Mechanism of solute release from porous hydro matrices and
other factors may be responsible. Int J Pharm. 1983;15:25 35.
Peppas NA. Analysis of Fickian and non Fickian drug release
from polymers. Pharm Acta Helv. 1985;60:110 111.
Hamed E, Sakr A. Application of multiple response optimization
technique to extended release formulations design. J Control
Release. 2001;73:329 338.
Williams AC, Barry BW. Penetration enhancers. Adv Drug
Deliv Rev. 2004;56:603 618.
Megrab NA, Williams AC, Barry BW. Oestradiol permeation
across human skin, silastic and snake skin membranes: the effects
of ethanol/water co solvent systems. Int J Pharm. 1995;116:101
112.

DOI: 10.1208/s12249 009 9235 0
Research Article
Extended Release Felodipine Self-Nanoemulsifying System
Pradeep R. Patil,1 Shailesh V. Biradar,1 and Anant R. Paradkar1,2
Received 1 July 2008; accepted 1 March 2009; published online 5 May 2009
Abstract. The purpose of the present study was to formulate a self nanoemulsifying system (SNES)
containing model lipophilic drug, felodipine (FLD), to improve its solubility. The SNES was formulated
using varying amounts of Miglyol 840 (as an oil), Cremophor EL (as a surfactant), and Capmul
MCM (as a co surfactant). The SNES were characterized for turbidity, droplet size and in vitro FLD
release. The SNES containing oil, surfactant, and co surfactant in the weight ratio of 3.5:1.0:1.0,
respectively, showed good emulsication, median droplet size of 421 nm, and rapid FLD release (>90%
release in 15 min). Gelling was induced in the SNES by addition of Aerosil 200 (A 200). Rheological
studies clearly demonstrated the formation of gelled microstructure with enhanced elasticity for SNES
with A 200. Since FLD warrants extended delivery for management of hypertension, the gelled SNES was
further encased within the hydrophobic Gelucire 43/01 (GEL) coat to extend the release of FLD. Caprol
PGE 860 (CAP) was added to this coat as a release enhancer. No interaction was seen between GEL and
CAP in differential scanning calorimetry. The effect of two formulation variables in the encased SNES, viz.,
the gelling agent (A200) and the release enhancer (CAP), on the in vitro FLD release was evaluated using 32
factorial design experiments. CAP by virtue of channel formation in GEL coat favored the FLD release,
while the A200 retarded the FLD release by inducing gelling. At later time points, an interaction between
these two variables was found to govern extended release of FLD. The developed gelled SNES encased
within the GEL coat can be used as an extended release composition for lipophilic drugs.
KEY WORDS: extended release; felodipine; gelucire 43/01; self nanoemulsifying system.
INTRODUCTION
Self emulsifying system is a mixture of oil and surfactants
and/or co solvent. They have the ability to emulsify spontane
ously in aqueous media even under gentle agitation (1). Upon
peroral administration, these systems readily form ne
emulsion in the gastrointestinal tract by virtue of mild
agitation provided by gastric mobility and, thus, present the
contained drug in solublized form in vivo. Therefore, these
1
Department of Pharmaceutics, Poona College of Pharmacy, Erand

wane, Bharati Vidyapeeth University, Pune 411 038 Maharashtra
State, India.
2
To whom correspondence should be addressed. (e mail: arparadkar
@rediffmail.com)
ABBREVIATIONS: A 200, Aerosil 200; CAP, Caprol PGE 860;
CTAB, Cetyl trimethyl ammonium bromide; DSC, Differential
scanning calorimeter; FLD, Felodipine; G, Solid modulus; G,
Viscous modulus; GEL, Gelucire 43/01; HLB, Hydrophilic
lipophilic balance; Hz, Hertz; JCIC, Japanese Cosmetic Ingredients;
LC, Liquid crystal; Log P, Partition coefcient; LVR, Linear
viscoelastic region; mg/g, Milligram per gram; mg/ml, Milligram per
milli liter; N, Normal; n*, Complex viscosity; Pa, Pascal; PEG,
Polyethylene glycol; Ph.Eur, European Pharmacopoeia; R, Co
efcient of correlation; R2, Coefcient of determination; SNES,
Self nanoemulsifying system; SNES II A, Self nanoemulsifying
system II with 25 mg A200; SNES II B, Self nanoemulsifying
system II with 50 mg A200; USP/NF, United State Pharmacopoeia/
National Formulary; m, Micrometer; w/v, Weight by volume.
self emulsifying systems have the potential to improve the rate

and extent of absorption of lipophilc (BCS Class II) drugs
showing dissolution as a rate limiting step for in vivo
absorption (2 4). Renaissances in the use of self emulsifying
system over the past two decades are inviting increasing
attention. Recent trends are focused on the development of
modied self emulsifying solid or semi solid formulations as an
alternative to the conventional liquid self emulsifying system.
These include spray dried emulsion (5), enteric coated solid state
pre microemulsion concentrates (6), self nanoemulsied tablets
containing ubiquinone (7), supersaturable self emulsifying sys
tem of paclitaxel (8), and self emulsiable pellets (9,10). Solid
self emulsifying system comprising goat fat and Tween 65 was
formulated for the delivery of diclofenac (11). Booth et al. (12)
prepared solid self emulsifying system using extrusion sphero
nization technique, wherein lactose and microcrystalline cellu
lose were used as solidifying aids. Gelled self emulsifying system
containing ketoprofen was formulated as an intermediate for
semi solid or solid dosage forms (13).
However, very few reports are available in the literature
that relate to extended drug release systems wherein the self
emulsifying system has been advantageously employed to
increase the solubility of lipophilic drug. Barthelemy and
Benameur (14) invented a self emulsifying system loaded
cellulose matrix composition, which in aqueous media formed
microemulsion. This microemulsion released the drug in a
sustained manner. Schwarz (15) developed solid self emulsi
fying dosage form for sustained delivery of active principle by
515
516
Patil, Biradar and Paradkar
using large amounts of adsorbent and hydrophilic polymers.

Serratoni et al. (16) prepared a self emulsifying system
containing water insoluble methyl and propyl paraben, which
was loaded on pellets. These pellets were further coated with
polymers for controlled release of parabens.
Gelucire is a family of vehicles derived from the
mixtures of mono , di , and triglycerides with polyethylene
glycol esters of fatty acids. These are available with a variety
of properties depending on their hydrophilic lipophilic bal
ance (HLB) range (HLB 1 18) and melting point range (33
70C) (17 19). They have a wide variety of applications in
fast and sustained release pharmaceutical formulations
(20,21). Gelucire 43/01 (GEL) is a hydrophobic lipid with
an HLB value of 1 and melting point of 43C. It is a blend of
saturated triglycerides of different fatty acids, viz., C8 (3%),
C10 (2%), C12 (29%), C14 (2%), C16 (17%), and C18 (36%).
Sutananta et al. (22) reported sustained release GEL matri
ces, where merely 1.7% theophylline was released within
20 h. Sustain release ability of the GEL has been attributed to
its extreme hydrophobicity.
FLD (oral dose, 5 10 mg/day) is a crystalline drug
(melting point 145C) and is practically insoluble in water
(water solubility, 19.7 mg/L; log P, 4). Few attempts revealed
in the literature underline the need for dissolution rate
enhancement of FLD. Major reported approaches are
formation of glassy forms (23), micronization and subsequent
complexation with beta cyclodextrin (24), solid dispersion
(25), and nanodispersions (26). These techniques suffer from
certain disadvantages like generation of high static charges,
poor wettability and ow properties, agglomeration, poly
morphic conversions, and periodic recrystallization. In addi
tion, FLD warrants extended delivery for management of
hypertension (27) and is marketed as extended release tablets
(Plendil; 2.5 , 5 , and 10 mg strengths). Conventional
extended release formulation of FLD can offer the solubili
ty dependent release rather than formulation dependent due
to its poor aqueous solubility. In this study, FLD loaded
SNES was prepared using different oils and surfactants.
Prepared SNES formulations were evaluated for emulsifying
properties, droplet size measurement, rheological properties,
and in vitro drug release. Effect of formulation variables on in
vitro drug release was studied using 32 factorial design.
Optimized SNES was further encased within hydrophobic
GEL coat. The unique advantage of the developed system is

that the drug is in dissolved state, and hence, formulation
dependent extended release can be achieved.

Gelucire 43/01(USP/NF, Ph. Eur.) (waxy solid, melting
point=43C, HLB=01) was a gift from Gattefosse, France. FLD
(USP) was a gift by Wockhardt Ltd., India. Glyceryl mono and
dicaprate (Capmul MCM) and oleic acid esters of decagly
cerol (Caprol PGE 860: CAP) were supplied by Abitec Corp.,
USA as gift samples. Propylene glycol dicaprylocaprate
(Miglyol 840; Ph. Eur., JCIC) was provided by Sasol Corp.,
Germany. Polyoxyl 35 castor oil (Cremophor EL; USP/NF)
and colloidal silicon dioxide (Aerosil 200: A 200; USP/NF;
Degussa Corp., Germany) were gift samples from BASF Corp.,
New Jersey and Get Rid Pharmaceuticals, Pune, India, respec
tively. All other reagents were of analytical grade.
Formulation of FLD SNES

FLD [50 mg, weighed using Mettler Toledo (AB204 S)
balance, Switzerland with readability of 0.1 mg] was dissolved
in Miglyol 840 (700 mg) by heating on temperature
controlled water bath to 50C in a glass vial. To this, varying
amounts of Cremophor EL and Capmul MCM were
added and mixed well (Table I). The resultant mixture was
cooled to ambient temperature (30C) and was observed after
10 h for drug precipitation, if any. Emulsifying properties of
prepared SNES was examined upon dilution with water.
Evaluation of FLD SNES

Turbidimetry
SNES (0.2 ml) was added to puried water (150 ml)
under continuous stirring (50 rpm) on magnetic plate (Ika
Werke, Germany) at 30C and equilibrium turbidity was
measured using Systronic turbidimeter (Systronics, Type
131, Ahmadabad, India).
Table I. Composition and Evaluation of FLD SNES

SNES
Composition/evaluation parameter
Composition (% w/w)
FLD
Miglyol 840
Cremophor EL
Capmul MCM
Evaluation parameter
Visual quality of emulsion produced
Turbiditya (NTU)
Median droplet sizea (nm)
Drug contenta (mg/g)
FLD felodipine, SNES self nanoemulsifying system
a
MeanSD, n 3
II
III
3.7
51.85
29.63
14.81
4.35
60.87
17.39
17.39
3.7
51.85
14.81
29.63
Very Good
104.2114.18
394.4221.24
35.872.13
Very good
86.4318.54
421.4318.54
43.211.89
Moderate
196.332.24
664.867.83
36.122.97
Extended Release SES
517
Droplet Size Measurement

Droplet size measurement was performed by using laser
diffractometer (Mastersizer 2000 version 2.0, Malvern Instru
ments, Malvern, UK). The measurements were based on Mie
theory. SNES were diluted with water (200 times) before
measurement, and all measurements were performed in
triplicates after 2 min stirring. The data was presented in
terms of median diameter (D50), which was calculated by
Malvern Software version 5.22. For each measurement,
obscuration was in the range of 10 20%, while weighted
residual was less than 2%.
Drug Content
SNES (500 mg) was dissolved in 100 ml methanol to
obtain stock solution. This solution, after stepwise dilution
with methanol, was analyzed for FLD content by UV
spectrophotometric analysis (Jasco V 530, Japan) at
362 nm against similarly treated placebo SNES as blank.
In Vitro Drug Release
In vitro drug release studies (in triplicates) were carried
out using USP 24 type II dissolution test apparatus (Electro
lab TDT 06P, Mumbai, India). FLD SNES (equivalent to
5 mg FLD) was added to 900 ml phosphate buffer (pH 6.5)
with 0.5% w/v cetyl trimethyl ammonium bromide (CTAB),
added to effect the sink condition (28). Dissolution media was
thermostated at 370.5C and stirred at 100 rpm. Aliquots
were collected periodically and replaced with fresh and pre
warmed dissolution medium. The aliquots, after ltration
through Whatman paper (pore size 0.45 m), were analyzed
using spectrophotometer at 362 nm (against similarly treated
placebo SNES as blank) for FLD content. The data was
analyzed using PCP Disso version 3.0 software (Pune, India).
Preparation of Blank GEL Carriers
Pre weighed quantities of GEL (350 mg each) were
melted on a temperature controlled water bath to 50C in a
glass vial. To the molten mass, varying amounts of CAP was
added and mixed well (Table II). Approximately 75% of such
mixture was poured into pre chilled specially fabricated

stainless steel mould (internal die length, 15 mm; die
diameter, 10 mm) and then allowed to solidify. Central cavity
of 11 mm length and 6.5 mm diameter was created in this
molten semisolid mass by virtue of stainless steel rod (length,
15 mm; diameter, 6.5 mm). Both speed of penetration (2 mm/
min) and penetration depth (11 mm) of steel rod were
controlled by advanced force gauge (Mecmesin, West Sussex,
England). After 2 min, the rod was removed (2 mm/min),
thus creating an open cavity. Dilute aqueous potassium
permanganate solution (0.25 ml) was poured in the cavity,
which was then sealed by spreading the remaining amount of
molten GEL mixture onto it. This was allowed to further
solidify at cold conditions (2 8C) overnight. Before evalua
tion, GEL carrier was equilibrated to 30C for 5 h.
Evaluation of Blank GEL Carriers
Blank carriers were evaluated for weight uniformity, leak
test, cavity volume, and coat thickness (Table II). Leak test
was performed by placing the blank carriers (lled with dilute
aqueous potassium permanganate solution) in distilled water
at 37C and stirring for 15 min. The appearance of pink color
in outer aqueous media was an indication of leakage from
carrier. Mean cavity volume was measured by breaking ten
blank carriers and emptying their contents in a measuring
cylinder. Coat thickness of ten blank GEL carriers was
measured using vernier caliper.
Differential scanning calorimetry (DSC) thermograms of
untreated GEL (A), blank carrier containing 350 mg GEL
with 25 mg CAP (B), blank carrier containing 350 mg GEL
with 50 mg CAP (C), and blank GEL carriers after 2 and 5 h
hydration (D and E, respectively) were obtained using a
Mettler Toledo DSC 821e instrument equipped with an
intracooler (Mettler Toledo, Switzerland). The hydration
medium was 900 ml of phosphate buffer (pH 6.5) with 0.5%
w/v CTAB maintained at 370.5C and stirred at 100 rpm.
Indium standard was used to calibrate the DSC temperature
and enthalpy scale. GEL carrier (between 5 to 10 mg) was
hermetically sealed into a pierced aluminum crucible and
heated at a constant rate of 10C/min over a temperature
range 0C to 100C. Inert atmosphere was maintained by
purging nitrogen at a ow rate of 50 ml/min.
Table II. Composition and Evaluation of Blank GEL Carriers

Blank GEL carrier
Composition/evaluation parameter
Composition
GEL (mg)
Caprol PGE 860 (mg)
Colored solution (ml)
Evaluation parameter
Weighta (mg)
Leak test
Cavity volumea (ml)
Coat thicknessa (mm)
GEL Gelucire 43/01
MeanSD, n 10
A
350
0.25
350
25
0.25
350
50
0.25
6055.32
Negative
0.240.05
1.320.14
6215.32
Negative
0.250.06
1.290.17
6465.32
Negative
0.250.04
1.300.15
518
Preparation of Gelled SNES Containing FLD

To 500 mg of optimized SNES composition, either 50 mg
[SNES II (A)] or 100 mg [SNES II (B)] of A 200 was
separately added in sealed glass containers. These systems
were mixed thoroughly and kept at 40C for 3 h and then
stored at 30C.
Rheological Studies of FLD SNES
Rheological study of SNES containing FLD was carried
out using a controlled stress rheometer (Viscotech, Rheo
logica Instruments AB, Lund, Sweden). The aim of the study
was to evaluate the effect of gelling agent on viscoelastic
properties of FLD SNES. Data analysis was done using Stress
RheoLogic Basic software, version 5.0. A cone and plate
geometry was used with 25 mm diameter and cone of 1.0.
Fresh sample was used for each measurement, and measure
ments (in triplicate) were carried out at 25C. The sample was
subjected to a stress of 1 to 200 Pa at a xed frequency of
1 Hz. Solid modulus (G), viscous modulus (G), complex
viscosity (n*), and loss tangent (tan ) were recorded over a
range of applied stress to determine the ability of SNES to
resist the deformation under applied stress.
Preparation of Factorial Batches of GEL Carriers with FLD
SNES
The effect of two formulation variables, viz., gelling
agent (A 200) and release enhancer (CAP), on the in vitro
drug release was evaluated using 32 factorial design and
response surface methodology. The batches (using 32 factorial
design) were prepared by similar method employed for blank
carriers, except different FLD SNES compositions [II, II (A)
and II (B)] were added to the GEL cavities formed in lieu of
dilute aqueous potassium permanganate solution. In addition,
the composition of carrier varied with respect to the amounts
of CAP, as described in Table III.
In vitro Drug Release
In vitro FLD release from prepared GEL carrier was
determined in 900 ml 0.1 N hydrochloric acid (pH 1.2) for
rst 2 h and thereafter in 900 ml phosphate buffer (pH 6.5)
with 0.5% w/v CTAB. Both dissolution media were main
tained at 370.5C and stirred at speed of 100 rpm. The

release proles were statistically treated for non linear
regression analysis using PCP Disso version 3.0 software
(Pune, India) to calculate time required for X% (i.e. 25%,
50%, and 75%) drug release and coefcients were recorded.
Preliminary studies were focused on screening of differ
ent pharmaceutically acceptable oils for FLD solubilization.
Various oils and blank SNES compositions thereof were
studied for their ability to solubilize FLD (as determined
visually by observing precipitation on overnight storage at
30C. Miglyol 840 was selected as an oil phase, as blank
SNES compositions thereof showed better drug solubility
(between 35 and 45 mg/g) in trial runs. Hence, Miglyol 840
was used along with different ratio of hydrophilic surfactant
(Cremophor EL) and lipophilic cosurfactant (Capmul
MCM) (Table I). SNES I and SNES II with 2:1 and 1:1
w/w ratio of Cremophor EL to Capmul MCM, respectively,
yielded emulsions of very good visual quality. Median droplet
size for SNES I was around 394 nm and that for SNES II was
about 421 nm. SNES III produced moderate quality emulsion
with median droplet size about 664 nm due to insufcient
amount of hydrophilic surfactant, Cremophor EL (HLB 12
14). The drug content of formulated SNES ranged between
35 and 44 mg/g, depending on the quantity of carrier aids. In
spite of the change in the emulsion droplet size, SNES I, II,
and III showed rapid in vitro drug release (>90% in rst
15 min) with no signicant difference (p>0.05) (data not
shown). SNES II with highest drug content (about 43% mg/g)
and smaller droplet size was selected for gelling and further
evaluation.
Gelled SNES was obtained by addition A 200 in SNES II
and subsequent mixing at 30C. Gelling might have induced
by formation of H bonds between the silanol (Si OH) groups
present on the surface of A 200, in a non polar milieu (here,
the SNES) as reported by Raghwan et al. (29). Gelled SNES
containing FLD showed signicant increase (p<0.01) in
median droplet size [541.7323.55 nm for SNES II (A) and
581.6627.33 nm for SNES II (B)] as compared to liquid
SNES II (421.4318.54 nm). With the addition of A 200, the
drug content for SNES II (A) and SNES II (B) were
observed to decrease to 40.861.89 and 38.322.97 mg/g,
respectively (because of dilution). The drug release in
Table III. Batches Prepared Using 32 Factorial Design with Coded Levels and Actual Values of Variables
Batch no.
1
2
3
4
5
6
7
8
9
Variable X1: Amount of A 200 (mg)

per 250 mg of SNES II
0
0
0
25
25
25
50
50
50
(1)a
(1)
(1)
(0)
(0)
(0)
(+1)
(+1)
(+1)
GEL Gelucire 43/01

Values in parentheses indicate coded levels of variables
Variable X2: Amount of Caprol

PGE 860 (mg) per 350 mg of GEL
0
25
50
0
25
50
0
25
50
(1)
(0)
(+1)
(1)
(0)
(+1)
(1)
(0)
(+1)
519
phosphate buffer (pH 6.5) with 0.5% w/v CTAB from gelled
SNES [II (A) and II (B)] was slower (about 75% in 30 min and
1.5 h, respectively) than that of SNES II due to the presence of
A 200. The increased median droplet size and comparatively
slow drug release from both gelled SNES might be due to
increased viscosity of liquid crystal (LC) phase. Similar results
were observed in our previous study (13).
Rheological properties of material can be evaluated by
dynamic (oscillatory) and static measurements. The dynamic
measurement provides a more direct correlation with micro
structure than static rheology since the materials can be
examined in their at rest state without causing any disruption
of their underlying structures (30,31). The term viscoelastic
ity is used to describe behavior, which falls between the
classical extremes of elastic response by the Hookean solids
and the Newtonian liquids. Viscoelastic materials exhibit
simultaneous existence of viscous and elastic properties. The
knowledge of viscoelastic properties is pivotal for SNES
because the microstructural environment or mobility are
responsible for curvature formation, which further inuence
the observed droplet size and rate of drug diffusion. Thus,
these properties can be indirectly probed using these
viscoelastic measurements. An oscillation stress sweep test is
a dynamic test where the solid modulus G and viscous
modulus G are measured as a function of stress at a constant
frequency. G is a measure of the energy stored in a cycle of
oscillation while G represents the energy dissipated per
cycle. In stress sweep measurement, linear viscoelastic region
(LVR) is determined, where the ratio of stress and strain is a
function of time alone. The results of oscillatory stress sweep
of SNES with and without A 200 are shown in Table IV. From
Table IV, it was evident that A 200 induced gelling in SNES II
(B) as seen from nearly 170,000 fold increase in G and
33,000 fold increase in complex viscosity as compared to
SNES II. SNES II (B) presented higher G and longer LVR in
stress sweep, conrming its elastic nature (32). It was further
observed that induced elasticity increased with an increase in
A 200 concentration. An elastic system will resist the
deformation induced by applied stress and, thus, on emulsi
cation will yield droplets with higher radius of curvature.
Therefore, SNES II (B) with higher elasticity had presented
large size droplets on emulsication as compared to SENS II.
An augment in elasticity of SNES with addition of A 200
might be the result of enhanced particle particle interaction,
especially intermolecular hydrogen bonding (24). This phe
nomenon might be responsible for observed gelling. Loss
tangent (tan ) is the ratio of viscous modulus (G) to solid
modulus (G). The higher the loss tangent, the lesser is the
elasticity of the material (33). The tan for SNES II was
much higher than SNES with A 200. Among SNES II (A)
and SNES II (B), the tan of the later was lower. From
observed tan values, it was further conrmed that, with an

increase in A 200 concentration, the elasticity of the SNES
was increased (Table IV). In drug release studies, SNES II
(B) had exhibited slower drug release. The drug release from
self emulsifying system is reported to be governed by
diffusion (13). The rate of diffusion is function of elasticity
of the system. The higher the elasticity, the lower is the rate of
diffusion. Results of rheological studies and drug release
studies revealed similar phenomenon. Therefore, from rheo
logical studies, it can be concluded that, with the addition of
A 200, the elasticity as well as complex viscosity of the SNES
was increased, which resulted in large droplet size and
retardation of drug diffusion.
Blank GEL carriers containing potassium permagnate
solution did not show any leaking, conrming the ability of
adopted method to produce leak proof sealing. No signicant
variations in the weight, cavity volume, and coat thickness
were observed for different batches of blank carriers (p>0.05;
Table II), presumably due to the similar conditions main
tained during their preparation.
DSC thermograms of pure GEL (Fig. 1A) and blank
carriers with 25 and 50 mg of CAP (Fig. 1B, and 1C,
respectively) did not show any difference in their melting
behavior. This indicated an absence of any physical interac
tion between them. The normalized energy of the melting
endotherm, however, decreased with increasing amount of
highly viscous release enhancer. This may be due to
decreasing weight fraction of melting component (GEL) in
the blank carriers B and C (Table V). In addition, the blank
GEL carriers hydrated in dissolution media for 2 h (Fig. 1D)
and 5 h (Fig. 1E) have not shown any change in melting
temperature of GEL, thus ruling out possibility of potential
polymorphic transitions in GEL during in vitro release.
However, the enthalpy of melting endotherm decreased
signicantly at these time points because of hydration of
GEL domain. The presence of small endotherms above 80C
in thermogram E (Fig. 1) may be ascribed to gradual
evaporation of water of hydration. Thus, from DSC studies, it
was conrmed that there was neither any physical interaction
observed between GEL and CAP nor there was any polymeric
transformation even after 5 h hydration of GEL.
Erosion pattern of blank GEL carriers was studied in
900 ml of phosphate buffer pH 6.5 (similar to In vitro Drug
Release). Erosion was observed after 2 h. The particle size
analysis of dissolution media at initial stage and after 2 and
5 h of hydration was performed using laser diffractometer.
The droplet size measurements were based on Mie theory.
The size range of eroded particles was between 400 and
800 nm. Therefore, the dissolution media appeared turbid
after 2 h. Similar study was carried out for GEL carriers with
increasing amounts of release enhancer (CAP). The coarse
Table IV. Results of Oscillatory Stress Sweep for Different SNES

Batch
LVRa
Ga (Pa)
na (Pa s)
tan a
SNES II
SNES II (A)
SNES II (B)
101.52 1306.11
102.31 1405.50
153.29 200 7.01
0.09710.0035
0.08380.0022
170.86.9174
0.0180.0008
0.1670.0061
33.010.7427
12.50.6237
0.6910.0294
0.6870.0331
LVR linear viscoelastic region, SNES self nanoemulsifying system

a
MeanSD, n 3
520

design was adopted. Such experimental design enabled
evaluation of the effect of these variables (and their
interactions, if any) at three different levels on in vitro drug
release prole from GEL coated formulation (Table III). The
in vitro drug release proles of studied SNES batches are
given in Table VI. The amount of drug released as a function
of time was mathematically treated to t into a simple, semi
empirical model developed by Korsmeyer et al. (34) relating
exponentially the drug release to the elapsed time (t):
Mt =M1 k tn
Fig. 1. DSC thermograms A neat GEL; B GEL/Caprol (350:25 w/w);

C GEL/Caprol (350:50 w/w); D neat GEL after 2 h hydration in
dissolution medium; E neat GEL after 5 h hydration in dissolution
medium
particles (size range 3 to 5 m) were observed in the

dissolution media, indicating leaching of release enhancer
and GEL fraction.
In order to study the effect of two important variables,
viz., amount of gelling agent (A 200) in FLD SNES and
amount of release enhancer (CAP) in GEL coat, 32 factorial
where k is a constant incorporating structural and geometric

characteristics of the dosage form, n is the release exponent,
indicative of the drug release mechanism and Mt/M is
fractional release of drug as a function of time. This model
is more applicable for dosage forms wherein there are more
than one type of release phenomenon that could be involved
(35). The in vitro FLD release data from the factorial batches
were treated for model tting using PCP Disso version 3.0
software (Pune, India). For the best t models of all factorial
batches, the n values ranged between 0.83 and 1.70, while the
k values were between 0.7 and 9.6 (with R values between
0.9908 and 0.9996). Therefore, it can be concluded that there
was no clear discrimination among these batches with respect
to the drug release pattern.
Sutananta et al. (22) investigated the mechanism of drug
release from different grades of GEL. The nature of GEL
was found to strongly affect the underlying drug release
mechanism. The drug release was primarily controlled by
diffusion in case of hydrophobic variants of Gelucire (Gelu
cire 43/01 and 54/02), whereas erosion was the predominant
mechanism in case of hydrophilic Gelucire (Gelucire 55/18
and 50/13). Abdalla and Mader (36) recently evaluated the
release kinetics from the lipid self emulsifying formulation
comprising C18 mono and diglycerides and macrogol 15
hydroxy stearate. The release kinetics, assessed by electron
spin resonance spectroscopy, revealed that the hydrophobic
spin probe used in the study was localized mainly in the lipid
environment indicating the release of lipophilic drugs through
lipid barrier, primarily by diffusion process. In the present
study as well, FLD was present in the lipid (oil) milieu that
comprises surfactants, and such composition was further
encased within the hydrophobic GEL. Deriving analogy
from the reports by Sutananta et al. (22) and Abdalla and
Mader (36), the drug release from studied system can be
Table V. DSC Integrated Data of Blank GEL Carriers

DSC integrated data
Sample
Normalized energy (J/g)
Onset temp (C)
Peak temp (C)
Endset temp (C)
Untreated pure GEL (A)

GEL/Caprol: 350:25
w/w (B)
GEL/Caprol: 350:50
w/w (C)
GEL after 2 h hydrationa (D)
GEL after 5 h hydrationa (E)
74.19
76.90
40.43
40.07
45.16
43.21
47.34
45.76
60.54
41.55
43.45
45.87
43.90
36.07
42.94
43.17
44.98
46.95
47.63
49.03
DSC differential scanning calorimetry, GEL Gelucire 43/01

Hydration of pure GEL was done in dissolution media
521
Table VI. In vitro Release Proles of Factorial Batches
Drug releaseda (%)
Time (h)
Batch 1
Batch 2
Batch 3
Batch 4
Batch 5
Batch 6
Batch 7
Batch 8
Batch 9
0
2
4
8
12
16
20
0
10.56.8
21.55.6
38.06.7
58.45.6
78.46.1
89.74.8
0
16.86.4
29.66.7
45.25.4
64.56.1
81.55.8
89.94.6
0
20.16.1
35.96.3
58.66.7
72.65.9
88.64.9
94.55.2
0
8.96.9
14.66.8
33.26.7
50.35.6
71.85.3
78.05.3
0
10.56.5
16.25.4
35.66.1
57.95.2
72.85.9
81.34.9
0
14.67.3
22.15.9
46.86.1
64.86.4
79.85.8
87.65.4
0
2.566.4
8.865.8
29.45.8
47.66.4
67.25.3
74.34.9
0
4.95.9
14.26.1
32.15.9
52.36.3
68.35.4
78.34.8
0
8.66.8
16.85.2
37.56.4
58.64.9
77.74.6
83.25.9
MeanSD, n 3
attributed to diffusion mechanism. However, this hypothesis

needs further evaluations using advanced and sophisticated
instruments.
The data of time required for X% drug release from
factorial batches were calculated and further subjected to
multiple regression analysis using PCP Disso version 3.0
software (Pune, India). The data were tted into following
equation.
T50%rel h 9:96 1:68X1

R2 0:952; P 0:007
1:53X2
Y 0 1 X1 2 X2 11 X1 X1 22 X2 X2
12 X1 X2
Responses obtained for time required for X% drug

release from factorial batches after multiple regression
analysis and removing the insignicant variables are depicted
in Fig. 2. Time required for 25% and 50% drug release was
found to be directly dependent on the amount of gelling
agent and inversely proportional to the amount of release
enhancer (Fig. 3a, b). However, at these time points, no
interaction between these two variables was observed to
affect the in vitro drug release kinetics, as shown in the
following equations.
T25%rel h 5:09 1:57X1

R2 0:952; P 0:004
T 25 % rel
20
1:17X2
T 50 % rel
T 75 % rel
18
16
Time (h)
14
12
10
8
6
4
2
0
1
Batch No.
Fig. 2. Responses obtained for time required for X% drug release

from factorial batches
Fig. 3. a Effect of variables on time required for 25% drug release. b

Effect of variables on time required for 50% drug release. c Effect of
variables on time required for 75% drug release
522
The retardation of in vitro drug release by addition of

gelling agent in SNES was in line with the earlier report (13).
Moreover, such retardation effect was potentiated by encas
ing the gelled SNES in the hydrophobic GEL coat (22). On
the contrary, leaching of CAP during the release resulted in
the formation of aqueous channels, thus promoting drug
diffusion through hydrophobic GEL barrier. This mechanism
was further conrmed using stereomicroscopic observation of
the coat surface during in vitro release. Fine and irregular
cracks were developed on the blank GEL coat surface after 4
to 5 h. For GEL coats containing release enhancer, similar
cracks were observed at earlier time points (after 2 to 3 h),
which eventually got enlarged at the later time points (data
not shown). Interestingly, as evident form Eq. 5 below, the
magnitude of retardation effect of gelling on drug release at
later times (e.g., T75% rel) was remarkably less (1 =0.81) as
compared to the enhancing effect of channeling agent (2 =
1.22) owing to leaching of CAP and formation of water lled
channels across the GEL wall. This favored the hydration of
SNES and subsequent drug diffusion (Fig. 3c).
T75%rel h 16:67 0:81 X1
0:96 X1X2

R 0:945; P 0:018
1:22 X2
1:99 X1X1
5
Also at this juncture, a potential interaction (12 =0.96)

between gelling agent and release enhancer was observed to
affect drug release as one of the factors. Thus, the rate of
drug release at earlier time points was governed by two
independent variables, and the same, at later time points, was
additionally governed by complex interaction between these
two variables.
Thus, this study demonstrated the feasibility of new
formulation comprising SNES encased in a hydrophobic coat
for extended drug release.
CONCLUSIONS
The formulated SNES containing FLD improved the
dissolution rate markedly and can be utilized for the fabrica
tion of extended release system. Gelled SNES containing FLD
encased in a hydrophobic GEL coat can serve as an alternative
for conventional extended release formulations. Moreover, by
varying the contents of release enhancer and gelling agent in
such composition, the release prole of FLD can be manipu
lated as required. Thus, formulation dependent extended
release of lipophilic FLD was achieved with studied system.
ACKNOWLEDGMENTS
Authors thank Wockhardt Ltd. India and Colorcon Ltd.
India (Gattefosse, France) for the gift samples of FLD and
Gelucire 43/01, respectively. Authors are thankful to Dr.
Rakesh Kapoor, Bioriginal Food and Science Corp., Canada
for providing the gift sample of Caprol PGE 860 through
Abitec Corp, USA. Authors also express gratitude toward
Sasol GmbH and BASF, Mumbai, India for the gift samples
of Miglyol 840 and Cremophor EL, respectively. Pradeep
R. Patil and Shailesh V. Biradar are thankful to CSIR, India
for awarding Senior Research Fellowship.
REFERENCES
1. Charman SA, Charman WN, Rogge MC, Wilson TD, Pouton
CW. Self emulsifying drug delivery systems: formulation and
biopharmaceutical evaluation of an investigational lipophilic
compound. Pharm Res. 1992;9:87 93.
2. Pouton CW. Self emulsifying drug delivery systems: assessment
of the efciency of emulsication. Int J Pharm. 1985;27:335 48.
3. Shah NH, Carvajal MT, Patel CI, Infeld NH, Malick AW. Self
emulsifying drug delivery systems (SEDDS) with polyglycolyzed
glycerides for improving in vitro dissolution and oral absorption
of lipophilic drugs. Int J Pharm. 1994;106:15 23.
4. Gursoy RN, Benita S. Self emulsifying drug delivery systems
(SEDDS) for improved oral delivery of lipophilic drugs. Biomed
Pharmacother. 2004;58:173 82.
5. Hansen T, Holm P, Schultz K. Process characteristics and
compaction of spray dried emulsions containing a drug in lipids.
Int J Pharm. 2004;287:55 66.
6. Kim CK, Shin HJ, Yang SG, Kim JH, Oh YK. Once a day oral
dosing regimen of cyclosporin A: combined therapy of cyclo
sporin a premicroemulsion concentrates and enteric coated
solid state premicroemulsion concentrates. Pharm Res.
2001;18:454 59.
7. Nazzal S, Nutan M, Palamakula A, Shah R, Zaghloul AA, Khan
MA. Optimization of self nanoemulsied tablet dosage form of
ubiquinone using response surface methodology: effect of
formulation ingredients. Int J Pharm. 2002;240:103 14.
8. Gao P, Rush BD, Pfund WP, Huang T, Bauer JM, Morozowich
W, et al. Development of supersaturable SEDDS formulation of
paclitaxel with improved oral bioavailability. J Pharm Sci.
2003;92:2386 98.
9. Newton M, Petersson J, Podczeck F, Clarke A, Booth S. The
inuence of formulation variables on the properties of pellets
containing a self emulsifying mixture. J Pharm Sci. 2001;90:987
95.
10. Tuleu C, Newton M, Rose J, Euler D, Saklatvala R, Clarke A, et
al. Comparative bioavailability study in dogs of a self emulsifying
formulation of progesterone presented in a pellet and liquid form
compared with an aqueous suspension. J Pharm Sci.
2004;93:1495 502.
11. Attama AA, Nzekwe IT, Nnamani PO, Adikwu MU, Onugu CO.
The use of solid self emulsifying systems in the delivery of
diclofenac. Int J Pharm. 2003;262:23 8.
12. Booth SW, Clarke A, Newton JM. Spheronized self emulsifying
system for hydrophobic and water sensitive agents. United States
Patent number 6,630,150, 2003.
13. Patil P, Joshi P, Paradkar A. Effect of formulation variables on
preparation and evaluation of gelled self emulsifying drug
delivery system (SEDDS) of ketoprofen. AAPS PharmSciTech.
2004;5:E 42.
14. Barthelemy P, Benameur H. Composition with sustained release
of active principle, capable of forming microemulsion. United
States Patent number 6,309,665 (2001).
15. Schwarz J. Solid self emulsifying dosage form for improved
delivery of poorly soluble hydrophobic compounds and the
process of preparation thereof. United States Patent Application
Publication number 20030072798 (2003).
16. Serratoni M, Newton M, Booth S, Clarke A. Controlled drug
release from pellets containing water insoluble drugs dissolved in
a self emulsifying system. Eur J Pharm Biopharm. 2007;65:94 8.
17. Sutananta W, Craig DQM, Newton JM. The effects of ageing and
thermal behavior and mechanical properties of pharmaceutical
glycerides. Int J Pharm. 1994;111:51 62.
18. Ainaoui A, Ouriemchi EM, Bidah D, El Amrani MK, Vergnaud
JM. Process of drug release with oral dosage forms with lipidic
gelucire matrix. J Polym Eng. 1997;17:245 57.
19. Ainaoui A, Vergnaud JM. Modeling the plasma drug level with
oral controlled release forms with lipidic Gelucire. Int J Pharm.
1998;169:155 62.
20. Dennis AB, Farr SJ, Kellaway IW, Taylor G, Davidson R. In
vivo evaluation of rapid release and sustained release gelucire
capsule formulations. Int J Pharm. 1990;65:85 100.
21. Barker SA, Yap SP, Yeun KH, McCoy CP, Murphy JR, Craig
DQM. An investigation into the structure and bioavailability of
22.
23.
24.
25.
26.
27.
28.
alpha tocopherol dispersion in gelucire 44/14. J Control Release.

2003;91:477 88.
Sutananta W, Craig DQM, Newton JM. An evaluation of
mechanisms of drug release from glycerides bases. J Pharm
Pharmacol. 1995;47:182 7.
Trojak A, Kocevar K, Musevic I, Srcic S. Investigation of the
felodipine glassy state by atomic force microscopy. Int J Pharm.
2001;218:145 51.
Sharma VK. Preparation of micron size pharmaceutical particles by
micro uidization. United States Patent number 6,555,139, 2003.
Won DH, Kim MS, Lee S, Park JS, Hwang SJ. Improved
physicochemical characteristics of felodipine solid dispersion
particles by supercritical anti solvent precipitation process. Int J
Pharm. 2005;301:199 208.
Karavas E, Georgarakis E, Bikiaris D. Felodipine nanodisper
sions as active core for predictable pulsatile chronotherapeutics
using PVP/HPMC blends as coating layer. Int J Pharm.
2006;313:189 97.
Pannarale G, Puddu PE, Monti F, Collauto F, Serani G,
Schiariti M, et al. 24 hour antihypertensive efcacy of extended
release felodipine. Am J Hypertens. 1995;8:73A.
Eklund M, Lofroth J, Sundgren M. New pharmaceutical
formulation. Patent Cooperation Treaty Application Publication
number WO/2001/005376 (2001).
523
29. Raghavan SR, Walls HJ, Khan SA. Rheology of silica dispersions
in organic liquids: New evidence for salvation forces dictated by
hydrogen bonding. Langmuir 2000;16:7920 30.
30. Barnes HA, Hutton JF, Walters K. An introduction to rheology.
In: Rheology Series, Elsevier, Amsterdam, 1989.
31. Eccleston GM, Barry BW, Davis SS. Correlation of viscoelastic
functions for pharmaceutical semisolids: comparison of creep and
oscillatory tests for oil in water creams stabilized by mixed
emulsiers. J Pharm Sci. 1973;62:1954 61.
32. Kobayashi M, Ishikawa S, Samejima M. Application of nonlinear
viscoelastic analysis by the oscillation method to some pharma
ceutical ointments in the Japanese Pharmacopeia. Chem Pharm
Bull. 1982;30:4468 78.
33. Davis SS. Viscoelastic properties of pharmaceutical semisolids
III: nondestructive oscillatory testing. J Pharm Sci. 1971;60:1351
5.
34. Korsmeyer RW, Gurny R, Doelker EM, Buri P, Peppas NA.
Mechanism of solute release from porous hydrophilic polymers.
Int J Pharm. 1983;15:25 35.
35. Costa P, Lobo JMS. Modeling and comparison of dissolution
proles. Eur J Pharm Sci. 2001;13:123 33.
36. Abdalla A, Mader K. Preparation and characterization of a self
emulsifying pellet formulation. Eur J Pharm Biopharm.
2007;66:220 6.

DOI: 10.1208/s12249 009 9243 0
Research Article
Near-Infrared Analysis of Hydrogen-Bonding in Glass- and Rubber-State
Amorphous Saccharide Solids
Ken-ichi Izutsu,1,2 Yukio Hiyama,1 Chikako Yomota,1 and Toru Kawanishi1
Received 17 October 2008; accepted 9 April 2009; published online 7 May 2009
Abstract. Near infrared (NIR) spectroscopic analysis of noncrystalline polyols and saccharides (e.g.,
glycerol, sorbitol, maltitol, glucose, sucrose, maltose) was performed at different temperatures (30 80C)
to elucidate the effect of glass transition on molecular interaction. Transmission NIR spectra (4,000
12,000 cm 1) of the liquids and cooled melt amorphous solids showed broad absorption bands that
indicate random conguration of molecules. Heating of the samples decreased an intermolecular
hydrogen bonding OH vibration band intensity (6,200 6,500 cm 1) with a concomitant increase in a free
and intramolecular hydrogen bonding OH group band (6,600 7,100 cm 1). Large reduction of the
intermolecular hydrogen bonding band intensity at temperatures above the glass transition (Tg) of the
individual solids should explain the higher molecular mobility and lower viscosity in the rubber state.
Mixing of the polyols with a high Tg saccharide (maltose) or an inorganic salt (sodium tetraborate)
shifted both the glass transition and the inection point of the hydrogen bonding band intensity to higher
temperatures. The implications of these results for pharmaceutical formulation design and process
monitoring (PAT) are discussed.
KEYWORDS: amorphous; glass transition; hydrogen bonding; NIR; PAT.
INTRODUCTION
The development of amorphous solid pharmaceutical
formulations requires thorough characterization of physical
properties (1 4). Optimizing the molecular mobility and the
system viscosity, which both change signicantly at glass
transition temperature (Tg), is essential to ensure their unique
functional characters (e.g., faster dissolution of active ingre
dients, stabilization of lyophilized protein conformation) and
storage stability of these solids. Understanding how the
molecular interactions, particularly hydrogen bonding, in
amorphous carbohydrate solids affect their physical properties
is relevant to the formulation design and process control (5).
Near infrared (NIR) spectroscopy is an advancing ana
lytical method that provides varied information on the
chemical and physical properties of pharmaceutical formula
tions including molecular structure, crystallinity (6 8), crystal
polymorphs (9,10), residual water content (11), component
miscibility (12), protein secondary structure (13), and molec
ular interactions (14 17). Relatively low absorbance that
enables diffuse reection measurement and recent advances
in chemometrics (e.g., principal component analysis [PCA],
partial least squares calibration [PLS]) make the NIR spec
1
National Institute of Health Sciences, Kamiyoga 1 18 1, Setagaya,

Tokyo 158 8501, Japan.
2
To whom correspondence should be addressed. (e mail: izutsu@
nihs.go.jp)
troscopy a powerful nondestructive process analytical tool for

quality control (PAT) (18). The applicability of vacuum
sealed samples in glass vials should be an apparent advantage
of NIR in the analysis of physically unstable amorphous
solids. Proper use of NIR spectroscopy, combined with other
sophisticated in , on , and at line analytical tools (e.g., mid
infrared, far infrared, terahertz, and Raman spectroscopy;
thermal analysis), should improve the product and process
understanding required for better formulation quality (16,19,20).
The purpose of this study was to characterize hydrogen
bonding proles in rubber and glass state amorphous solids
by NIR spectroscopy. The random conguration of molecules
and the accompanying wide variation of molecular interac
tions in amorphous polyol and saccharide solids provide
broad absorption bands in the NIR spectra (6,7). Some of the
broad bands (e.g., OH stretching vibration) indicate different
hydrogen bonding states (e.g., intermolecular, intramolecular,
free) in the solids (21,22). The proles of the hydrogen
bonding in water and some alcohol liquids depend largely on
temperature (23 25). Recent mid infrared studies on the
amorphous saccharide lms indicate different temperature
dependent shifts of OH stretching (3,300 3,400 cm 1) and
bending (1,000 1,100 cm 1) vibration band peaks between their
glass (below Tg) and rubber (above Tg) states (26,27). It is
plausible that NIR spectroscopy provides valuable information
on the molecular interactions and physical properties of various
noncrystalline samples (e.g., liquid, rubber state solid, glass
state solid).
524
NIR Analysis of Hydrogen-Bonding in Glass and Rubber State

Materials
All chemicals employed in this study were of analytical
grades and were obtained from the following commercial
sources: glycerol and sorbitol (Wako Pure Chemical, Osaka,
Japan); maltose monohydrate, sucrose, and sodium tetrabo
rate (Sigma Chemical, St. Louis, MO, USA); deuterium water
(Acros Organics, Geel, Belgium); maltitol and maltotriitol
(Hayashibara Biochemical Laboratories, Okayama, Japan).
Preparation of Freeze-Dried Amorphous Solids
Freeze dried solids were prepared by lyophilizing aque
ous polyol and saccharides solutions (100 mg/mL, 2 mL) in
at bottom borosilicate glass vials (21 mm diameter; SVF 10,
Nichiden rika Glass, Kobe, Japan) using a Freezevac 1C
freeze drier (Tozai Tsusho, Tokyo, Japan). The sample
solutions frozen by immersion in liquid nitrogen were dried
under vacuum without temperature control for 16 h and then
maintained on the shelf at 35C for 8 h. Hydrogen/deuterium
exchanged glycerol, sorbitol, and glucose were obtained by
freeze drying the solutions in D2O twice (200 mg/mL).
Preparation of Cooled-Melt Amorphous Solids
The cooled melt solids subjected to thermal and NIR
analysis were obtained by melting the crystalline powder in a
quartz cuvette (1 mm light path length; Starna Optiglass,
Hainault, UK) or small borosilicate glass tubes (approximate
ly 4 mm internal diameter; TGK, Tokyo, Japan) in a drying
oven (DP23, Yamato Scientic, Tokyo, Japan) at 200C
(sucrose) or 180C (other saccharides) under vacuum for
20 min, then cooled at room temperature. Amorphous
maltose solids were prepared from the monohydrate crystal
with (dehydrated) or without (partially hydrated) vacuum
drying. The cooled melt solids without apparent crack or
bubbles were subjected to the following transmission NIR
analysis.
525
resolution in 128 scans. The absorbance of air was subtracted
as background. The NIR spectra were baseline corrected in
the 8,000 9,000 cm 1 range and smoothed at 25 points. The
possible effect of a temperature induced solid volume change
was not compensated in this study. Diffuse reection NIR
spectra of crystal powders and freeze dried solids in
cylindrical borosilicate glass vials were obtained from the
bottom of the container at room temperature.
RESULTS
Figure 1 shows the diffuse reection NIR spectra of
mannitol, maltose monohydrate, sucrose, glucose, and sorbi
tol in different physical states (crystalline powder, freeze
dried solid) obtained noninvasively from the bottoms of the
glass vials. The crystalline powders showed several unique
sharp bands of OH and CH vibrations in the overtone (5,000
7,500 cm 1) and combination (4,000 5,000 cm 1) spectral
regions. The absorbance of water in the maltose monohydrate
crystal (around 5,150 cm 1) disappeared by lyophilization.
Freeze dried amorphous sucrose and maltose solids showed
broad absorption bands that indicate diversied interactions
between the spatially less ordered molecules. Contrarily, some
sharp peaks (e.g., 4,375 cm 1) indicated partial crystallinity of
the freeze dried mannitol (4,28). Thermal analysis of the freeze
dried solids also indicated different component crystallinity
(data not shown). Freeze drying of glucose and sorbitol resulted
in collapsed solids, which spectra varied largely between the
vials (data not shown). Amorphous saccharide solids prepared
by cooling the heat melt also showed varied diffuse reection
NIR spectra.
Transmission NIR analysis of cooled melt amorphous
solids (e.g., sorbitol, glucose) and a liquid (glycerol) in a
quartz cell (1 mm light path length, 30C) showed similar
spectra with some broad bands (Fig. 2A). The large bands in
Thermal Analysis
A differential scanning calorimeter (DSC Q 10, TA
Instruments, New Castle, DE, USA) and software (Universal
Analysis 2000, TA Instruments) were used to obtain the ther
mal properties of the amorphous solids. Solids (2 5 mg) in
hermetic aluminum cells were scanned from 30C at 5C/min
under N2 gas ow. Glass transition temperatures were
determined from the maximum inection point of the dis
continuities in the heat ow curves.
NIR Spectroscopy
NIR analysis was performed using a FT NIR system
(MPA, Bruker Optik, Germany) equipped with a sample
temperature controller and OPUS software. Transmission
NIR spectra of liquids and cooled melt amorphous solids
were obtained from 30C to 80C at every 5C with 5 min
intervals between the measurements. Absorbance in the
12,000 to 4,000 cm 1 range was obtained with a 2 cm 1
Fig. 1. Diffuse reection NIR spectra of crystalline powder (CP) and

freeze dried solid (FD) containing sugars and sugar alcohols obtained
at the bottoms of glass vials
526
Izutsu, Hiyama, Yomota, and Kawanishi
Fig. 2. Transmission NIR spectra of liquid (glycerol) and cooled melt

solids (sorbitol, glucose) in a quartz cell (1 mm light path length) prepared
with A without and B with hydrogen/deuterium exchange (30C)
the overtone spectral region represent OH stretching vibra

tion rst overtones of intermolecular hydrogen bonding groups
(6,200 6,500 cm 1) and intramolecular or nonhydrogen
bonding groups (6,600 7,100 cm 1) (21,22,25). The addition of
H2O (2 10%, w/w) to glycerol increased the absorbance at
around 4,110, 5,152, and 6,925 cm 1 in the NIR spectra (data not
shown). Hydrogen/deuterium exchanged samples showed a
large band that suggested an OD stretching vibration at
Fig. 3. Effect of heating on the transmission NIR spectra of glycerol

liquid and cooled melt sorbitol solid (1 mm light path length).
Difference spectra were obtained by subtracting the absorbance of
the samples at 30C
Fig. 4. Transmission NIR spectra of cooled melt amorphous glucose,

sucrose, and maltose (dehydrated) solids in glass tubes (approximate
ly 4 mm interior diameter) obtained at 30C, 55C, and 80C
around 5,000 cm 1 (Fig. 2B) (23). Lower absorbance of the

hydrogen/deuterium exchanged amorphous solids at 6,000
7,100 cm 1 conrmed the signicance of the OH stretching
vibration band in the spectral region. Thermal analysis showed
that the cooled melt sorbitol (Tg = 1.2C) and glucose (Tg =
45.8C) are in the rubber and glass states, respectively, at 30C.
The effect of heating on the transmission NIR spectra of
the noncrystalline sorbitol and glycerol are shown as the
difference spectra (Fig. 3). The changes indicated the shift of
the major bands at 6,000 7,000 cm 1 (OH stretching vibration
rst overtone) and at around 4,750 cm 1 (OH stretching/
bending combination) in the original spectra to higher wave
numbers at the elevated temperatures (40C, 60C, and 80C).
The apparent decrease in the absorbance at 6,000 6,500 cm 1
and the concomitant absorbance increase at 6,700 7,100 cm 1
suggested heat induced changes in the hydrogen bonding pro
les of the OH groups from intermolecular to intramolecular or
free bonds.
Transmission NIR spectra (overtone region) of cooled
melt glucose, sucrose, and maltose (dehydrated) obtained at
different temperatures (30C, 55C, and 80C) suggested
varied heat induced changes in the hydrogen bonding band
intensity (Fig. 4). The larger spectral change observed in the
lower glass transition temperature solids (glucose=45.8C)
suggested the contribution of the different hydrogen bonding
proles to the physical properties. The solids were prepared

in small glass tubes (approximately 4 mm internal diameter)
to prevent crack formation in the cooling process. The
browning of sucrose during the preparation indicated its
partial degradation at high temperatures. A small absorption
band at 5,150 cm 1 suggested residual water in the amorphous
saccharide solids. Relatively large absorbance at the 6,200 to
6,500 cm 1 region in the amorphous glucose solid suggested a
larger contribution of the intermolecular hydrogen bonding
compared to those in other saccharides. Different dehydration
process and accompanying changes in the molecular
interactions should explain the partially different spectra of
the cooled melt glucose solid prepared in the quartz optical cell
(Fig. 2) and the glass tube (Fig. 4).
The relationship between the hydrogen bonding proles
and the physical states of the amorphous solids were further
studied. Thermal analysis showed varied glass transition
temperatures of the cooled melt amorphous polyol and
saccharide solids (Table I). The changes in the absorbance
of the intermolecular hydrogen bonding OH band peak
(6,200 6,500 cm 1) obtained every 5C were plotted in Fig. 5.
The spectral region was chosen for comparison because of the
smaller overlapping absorbance of residual water compared to
that in the intramolecular hydrogen bonding band region
(6,600 7,100 cm 1) (11). The noncrystalline samples showed
three types of intermolecular hydrogen bonding band intensity
changes depending on their glass transition temperatures. A
rubber state solid (sorbitol, Tg = 1.2C) and glycerol liquid
showed relatively large (<0.04 U/5C) band intensity drops
from the lowest measurement temperature. Contrarily,
amorphous dehydrated maltose that remained in the glass
state throughout the measurement temperature range (80C<
Tg) showed a much smaller band intensity change. Other solids
(e.g., glucose, sucrose, maltotriitol; 30C<Tg <80C) showed
larger declines of the band intensity above their glass transition
temperatures. Plotting of the band intensity against the
temperature showed apparent inection at the glass
transition. Plateau of the plot above the Tg should indicate
large and constant temperature dependent reduction of the
molecular interactions similar to that of molecular liquids.
527
Fig. 5. Changes in the absorbance of intermolecular hydrogen

bonding OH vibration band obtained by transmission NIR scan of
noncrystalline samples at 5C intervals. Each point of the curves
represents the mean of duplicated measurements. Each symbol
denotes A glycerol (filled circles), sorbitol (open triangles), glucose
(filled squares), sucrose (open circles), maltotriitol (filled triangles)
and B maltose (filled squares), maltose hydrate (open circles), and
maltose+sorbitol at the weight ratios of 2:1 (filled circles), 3:1 (open
triangles), and 5:1 (filled triangles)
Table I. Glass Transition Temperature of Cooled Melt Amorphous

Solids
Glass transition temperatures
Maltose
Maltotriitol
Maltose (w/o vacuum drying)
Sucrose
Maltitol
Glucose
Glycerol (liquid)
Sorbitol
Maltose+sorbitol 5:1a
Maltitol+Na2B4O7 40:1b
Maltitol+Na2B4O7 25:1b
Data are presented as meanSD (n 3)
a
Weight ratio
b
Molar ratio
99.41.2
79.50.8
62.61.7
57.93.2
49.30.5
45.82.2
n.d.
1.20.9
66.51.9
48.60.5
31.00.3
59.02.1
69.21.7
Fig. 6. Effect of temperature on the intermolecular hydrogen

bonding OH vibration band absorbance of maltitol (filled circles)
and its mixture with Na2B4O7 at the molar ratio of 40:1 (open
squares) and 25:1 (filled triangles) obtained at 5C intervals (n 2)
528
Some solids containing multiple components also showed
larger heat induced reduction of the intermolecular hydro
gen bonding band in the rubber state. The mixing of
components (e.g., maltose and sorbitol) or higher residual
water (e.g., cooled melt maltose monohydrate prepared
without vacuum drying) shifts the glass transition tempera
ture of the amorphous solids (Table I). These solids showed
larger heat induced intermolecular hydrogen bonding band
intensity reductions at above their glass transition temper
atures (Fig. 5). Some inorganic salts (e.g., sodium phosphates,
sodium citrates) affect the glass transition temperature of
amorphous polyol and saccharide solids (29,30). The addition
of a small amount of sodium tetraborate (Na2B4O7) signi
cantly raised the Tg of cooled melt maltitol solids (Table I), as
well as the inection temperature of the intermolecular
hydrogen bonding band intensity (Fig. 6) (31,32).
DISCUSSION
The NIR study showed varied hydrogen bonding states
of OH groups in the noncrystalline polyol and saccharide
solids that depended both on their temperatures and their
physical states (glass and rubber). The heating induced
reduction of the intermolecular hydrogen bond was consistent
with the literature on the NIR analysis of water, alcohol, and
polyol liquids (25,33). The limited absorbance of hydrogen/
deuterium exchanged polyols (glucose, sorbitol) at above
6,000 cm 1 indicated the contribution of readily exchangeable
OH groups to the temperature dependent band intensity
change. The appropriate sample temperature control and the
constant light path length made the transmission NIR mea
surement suitable to study the relationship between the physical
states and the molecular interactions. The general trends obse
rved in the transmission mode should be basically applicable to
other data detection modes (e.g., diffuse reection).
The glass transition of the amorphous solids did not
directly affect the NIR spectra, but it altered the extent of the
heat induced hydrogen bonding band intensity change. The
result was in agreement with the different temperature
dependent changes in the mid infrared OH stretching and
bending band peak positions of saccharide lms in the glass
and rubber states (26,27). Loosening of the intermolecular
hydrogen bonding network should induce the higher molec
ular mobility and lower viscosity of the rubber state amor
phous solids. The rubber state amorphous solids are, from
another physical perspective, deeply supercooled liquids with
extremely high viscosity. The similarity in physical states
would explain the large constant heat induced reduction of
the intermolecular band intensity in the polyol liquids (e.g.,
glycerol) and the rubber state solids. The smaller heat
induced spectral changes below the Tg would increase the
mobility of molecules, leading to slow but not negligible
temperature dependent chemical degradation and/or ingredi
ent crystallization in the long term storage of the glass state
amorphous solids (34). Further assignment of the bands in the
NIR spectra and mathematical data processing should
increase the relevance of the analysis.
Information on the molecular interactions (e.g., hydro
gen bonding) that determine the physical properties should
be relevant for the rational design of amorphous formula
tions. NIR spectroscopy should be used to characterize the
Izutsu, Hiyama, Yomota, and Kawanishi

molecular interactions in certain multicomponent amorphous
systems containing the active pharmaceutical ingredients,
excipients (e.g., stabilizer, pH modier), and residual water.
The availability of several detection modes that involve
samples in glass containers without exposure to unfavorable
environments (e.g., humid atmosphere) is a major advantage
of NIR spectroscopy over other analytical methods for the
characterization of the amorphous freeze dried formulations
(15). The proper choice of excipients that form and/or induce
intermolecular hydrogen bonding (e.g., disaccharides, sodium
phosphate) should provide the storage stability and functional
properties required for the amorphous formulations.
NIR spectroscopy is a powerful analytical tool for
process monitoring and raw material inspection (18,20).
Ensuring chemical and physical properties during the process
is important to obtain reliable pharmaceutical formulations.
Clinical functions and storage stabilities of the amorphous
solid formulations depends not only on their glass transition
temperatures (Tg) but also on other physical characters (e.g.,
residual crystallinity, structural relaxation) that are affected
by process parameters (e.g., thermal history) (3). Several
mathematical data processing methods, including PLS and
PCA, have been used to obtain information on the chemical
and physical properties of solids (e.g., crystallinity, collapse,
residual water, chemical degradation) from the NIR spectra.
The preparation of amorphous solids often includes low (e.g.,
freeze drying) or high (e.g., melting, extrusion) temperature
processes. Understanding the effect of temperature and
physical states on the NIR spectra, as well as appropriate
data compensation, should increase the relevance of the
sophisticated analytical methods.
ACKNOWLEDGEMENT
The present study was supported by the Japan Health
Sciences Foundation (KH31029, KHB1006).
REFERENCES
1. Cui Y. A material science perspective of pharmaceutical solids.
Int J Pharm. 2007;339:3 18.
2. Hancock BC, Zogra G. Characteristics and signicance of the
amorphous state in pharmaceutical systems. J Pharm Sci. 1997;86:
1 12.
3. Hilden LR, Morris KR. Physics of amorphous solids. J Pharm
Sci. 2004;93:3 12.
4. Nail SL, Jiang S, Chongprasert S, Knopp SA. Fundamentals of
freeze drying. Pharm Biotechnol. 2002;14:281 360.
5. Tang XC, Pikal MJ, Taylor LS. The effect of temperature on
hydrogen bonding in crystalline and amorphous phases in
dihydropyrine calcium channel blockers. Pharm Res. 2002;19:
484 90.
6. Yonemochi E, Inoue Y, Buckton G, Moffat A, Oguchi T,
Yamamoto K. Differences in crystallization behavior between
quenched and ground amorphous ursodeoxycholic acid. Pharm
Res. 1999;16:835 40.
7. Hogan SE, Buckton G. The application of near infrared
spectroscopy and dynamic vapor sorption to quantify low
amorphous contents of crystalline lactose. Pharm Res. 2001;18:
112 6.
8. Seyer JJ, Luner PE, Kemper MS. Application of diffuse
reectance near infrared spectroscopy for determination of
crystallinity. J Pharm Sci. 2000;89:1305 16.
9. Otsuka M, Kato F, Matsuda Y. Comparative evaluation of the
degree of indomethacin crystallinity by chemoinfometrical Four
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
ier transformed near infrared spectroscopy and conventional

powder X ray diffractometry. AAPS PharmSci. 2000;2:E9.
Tudor AM, Church SJ, Hendra PJ, Davies MC, Melia CD. The
qualitative and quantitative analysis of chlorpropamide poly
morphic mixtures by near infrared Fourier transform Raman
spectroscopy. Pharm Res. 1993;10:1772 6.
Stein HH, Ambrose JM. Near infrared method for determination
of water in aluminum aspirin. Anal Chem. 1963;35:550 2.
Jovanovi N, Gerich A, Bouchard A, Jiskoot W. Near infrared
imaging for studying homogeneity of protein sugar mixtures.
Pharm Res. 2006;23:2002 13.
Bai S, Nayar R, Carpenter JF, Manning MC. Noninvasive
determination of protein conformation in the solid state using
near infrared (NIR) spectroscopy. J Pharm Sci. 2005;94:2030 8.
Izutsu K, Fujimaki Y, Kuwabra A, Aoyagi N. Effect of counter
ions on the physical properties of L arginine in frozen solutions
and freeze dried solids. Int J Pharm. 2005;301:161 9.
Liu J. Physical characterization of pharmaceutical formulations
in frozen and freeze dried solid states: techniques and applica
tions in freeze drying development. Pharm Dev Technol. 2006;11:
3 28.
Rsnen E, Sandler N. Near infrared spectroscopy in the
development of solid dosage forms. J Pharm Pharmacol. 2007;59:
147 59.
Reich G. Near infrared spectroscopy and imaging: basic principles
and pharmaceutical applications. Adv Drug Deliv Rev. 2005;57:
1109 43.
Fevotte G, Calas J, Puel F, Hoff C. Applications of NIR
spectroscopy to monitoring and analyzing the solid state during
industrial crystallization processes. Int J Pharm. 2004;273:159 69.
FDA. Guidance for industry: PAT a framework for innovative
pharmaceutical development, manufacturing and quality assur
ance. http://www.fda.gov/cder/guidance/6419fnl.htm (2004)
De Beer TR, Alles M, Goethals F, Coppens A, Heyden YV, De
Diego HL, et al. Implementation of a process analytical
technology system in a freeze drying process using Raman spec
troscopy for in line process monitoring. Anal Chem. 2007;79:7992
8003.
Ozaki Y, Kawata S. Near infrared spectroscopy. Tokyo: Japan
Scientic Societies Press; 1996.
Shenk JS, Workman JJ Jr., Westerhaus MO. Application of NIR
spectroscopy to agricultural products. In: Burns DA, Ciurczak
WW, editors. Handbook of near infrared analysis. New York:
Taylor & Francis; 2001. p. 419 74.
Czarnecki MA, Czarnik Matusewicz B, Ozaki Y, Iwahashi M.
Resolution enhancement and band assignments for the rst
529
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
overtone of OH(D) stretching modes of butanols by two

dimensional near infrared correlation spectroscopy. 3. Thermal
dynamics of hydrogen bonding in butan 1 (ol d) and 2 methyl
propan 2 (ol d) in the pure liquid states. J Phys Chem A.
2000;104:4906 11.
Czarnecki MA, Ozaki Y. The temperature induced changes in
hydrogen bonding of decan 1 ol in the pure liquid phase studied
by two dimensional Fourier transform near infrared correlation
spectroscopy. Phys Chem Chem Phys. 1999;1:797 800.
Maeda H, Ozaki Y. Near infrared spectroscopy and chemo
metrics studies of temperature dependent spectral variations of
water: relationship between spectral changes and hydrogen
bonds. J Near Infrared Spectrosc. 1995;3:191 201.
Wolkers WF, Oldenhof H, Alberda M, Hoekstra FA. A Fourier
transform infrared microspectroscopy study of sugar glasses:
application to anhydrobiotic higher plant cells. Biochim Biophys
Acta. 1998;1379:83 96.
Wolkers WF, Oliver AE, Tablin F, Crowe JH. A Fourier
transform infrared spectroscopy study of sugar glasses. Carbo
hydr Res. 2004;339:1077 85.
Cao W, Mao C, Chen W, Lin H, Krishnan S, Cauchon N.
Differentiation and quantitative determination of surface and
hydrate water in lyophilized mannitol using NIR spectroscopy. J
Pharm Sci. 2006;95:2077 86.
Ohtake S, Schebor C, Palecek SP, Pablo JJD. Effect of pH,
counter ion, and phosphate concentration on the glass transition
temperature of freeze dried sugar phosphate mixtures. Pharm
Res. 2004;21:1615 21.
Kets EP, IJpelaar PJ, Hoekstra FA, Vromans H. Citrate
increases glass transition temperature of vitried sucrose prep
arations. Cryobiology 2004;48:46 54.
Miller DP, Anderson RE, de Pablo JJ. Stabilization of lactate
dehydrogenase following freeze thawing and vacuum drying in
the presence of trehalose and borate. Pharm Res. 1998;15:1215
21.
Izutsu K, Yomota C, Aoyagi N. Inhibition of mannitol crystal
lization in frozen solutions by sodium phosphates and citrates.
Chem Pharm Bull. 2007;55:565 70.
Watanabe A, Morita S, Ozaki Y. Temperature dependent
structural changes in hydrogen bonds in microcrystalline cellu
lose studied by infrared and near infrared spectroscopy with
perturbation correlation moving window two dimensional corre
lation analysis. Appl Spectrosc. 2006;60:611 8.
Yoshioka M, Hancock BC, Zogra G. Crystallization of indo
methacin from the amorphous state below and above its glass
transition temperature. J Pharm Sci. 1994;83:1700 5.

DOI: 10.1208/s12249 009 9241 2
Research Article
Formulation and Evaluation of Bioadhesive Buccal Drug Delivery of Tizanidine
Hydrochloride Tablets
Gazzi Shanker,1,3 Chegonda K. Kumar,2 Chandra Sekhara Rao Gonugunta,1
B. Vijaya Kumar,2 and Prabhakar Reddy Veerareddy1
Received 2 January 2009; accepted 9 April 2009; published online 8 May 2009
Abstract. The study aim was concerned with formulation and evaluation of bioadhesive buccal drug
delivery of tizanidine hydrochloride tablets, which is extensively metabolized by liver. The tablets were
prepared by direct compression using bioadhesive polymers such as hydroxylpropyl methylcellulose
K4M, sodium carboxymethyl cellulose alone, and a combination of these two polymers. In order to
improve the permeation of drug, different permeation enhancers like beta cyclodextrin ( CD),
hydroxylpropyl beta cyclodextrin (HP CD), and sodium deoxycholate (SDC) were added to the
formulations. The CD and HP CD were taken in 1:1 molar ratio to drug in formulations.
Bioadhesion strength, ex vivo residence time, swelling, and in vitro dissolution studies and ex vivo
permeation studies were performed. In vitro release of optimized bioadhesive buccal tablet was found to
be non Fickian. SDC was taken in 1%, 2%, and 3% w/w of the total tablet weight. Stability studies in
natural saliva indicated that optimized formulation has good stability in human saliva. In vivo
mucoadhesive behavior of optimized formulation was performed in ve healthy male human volunteers
and subjective parameters were evaluated.
KEY WORDS: bioadhesive buccal tablets; in vitro evaluation; in vivo mucoadhesive behavior;
permeation enhancers; stability studies in natural saliva; tizanidine hydrochloride.
INTRODUCTION
Bioadhesive buccal delivery of drugs is one of the
alternatives to the oral route of drug administration, particu
larly to those drugs that undergo rst pass effect. The stratied
squamous epithelium supported by a connective tissue lamina
propria, which is present in buccal mucosa (1), was targeted as
a site for drug delivery several years ago. Problems accompa
nied with oral route of administration such as extensive
metabolism by liver, drug degradation in gastrointestinal tract
due to harsh environment, and invasiveness of parenteral
administration can be solved by administering the drug
through the buccal route (2,3). The buccal route appears to
offer a number of advantages, like good accessibility, robust
ness of the epithelium, usage of the dosage form in accordance
with need, and comparatively less susceptibility to enzymatic
activity. Hence, adhesive mucosal dosage forms were prepared
for oral delivery, in the form of adhesive tablets (4,5), adhesive
gels (6,7), and adhesive patches (8).
The permeation of hydrophilic drug through membrane is
one of the major limiting factors for the development of
1
Department of Pharmaceutics, St. Peters Institute of Pharmaceutical

Sciences, Warangal, India.
2
Department of Pharmaceutics, Jangaon Institute of Pharmaceutical
Sciences, Warangal, Andhra Pradesh 506167, India.
3
To whom correspondence should be addressed. (e mail: shanker
gazzi@yahoo.com)
bioadhesive buccal delivery devices. The epithelium that lines

the buccal mucosa is a main barrier for the absorption of drugs
(9). In order to improve buccal absorption, several approaches
have been introduced. Increased permeation of the drug
through the buccal membrane and prevention of the drug
degradation by enzymes was achieved by changing the physico
chemical properties of the drug (10). Alternatively, improving
the bioadhesion and release characteristics of buccal delivery
devices increases the amount of drug available for absorption
(11). The incorporation of absorption enhancers to the buccal
formulation is one interesting approach. Substances that
facilitate the permeation through buccal mucosa are referred
as permeation enhancers (12). Different types of potential
permeation enhancers have been studied for buccal route to
increase the penetration of drugs (13,14).
The complexation of steroidal hormones with cyclo
dextrins was not effective in increasing the permeation
through buccal route, whereas condensation products of
cyclodextrin with propylene oxide or epichlorohydrins were
able to form complexes with estradiol, testosterone, and
progesterone, thereby enhancing absorption through the
buccal membrane in humans (15).
The delivery of hydrophilic macromolecular drugs via
buccal membrane was made possible by incorporation of
absorption or permeation enhancers, which could reduce
barrier properties of the buccal epithelium (13 20).
Tizanidine hydrochloride (TZD HCL) is an imidazoline
derivative, which acts as agonist on centrally located 2
530
Bioadhesive Buccal Drug Delivery of TZD HCL Tablets

receptors and this leads to myotonolytic effects on skeletal
muscle (21 24). It is structurally and pharmacologically
similar to clonidine and other 2 adrenergic agonists (23,24).
The correct mechanism of tizanidine in decreasing muscle
tone and frequency of spasm is not clearly understood (24).
About 53% to 66% of the dose administered is being
absorbed through the gastrointestinal tract after oral adminis
tration and the peak plasma concentration is reached within 1 to
2 h. Bioavailability of tizanidine is about 34% to 40% and half
life is 2.5 h. The drug is widely distributed throughout the body
and 30% of drug binds to plasma proteins. It undergoes rapid
and extensive rst pass metabolism in the liver (approximately
95% of a dose), leading to the oxidation of the imidazoline
moiety, aromatic system, and the sulfur atom. This leads to
lower bioavailability of tizanidine (25). In order to overcome
such extensive rst pass metabolism, the drug is selected as
suitable candidate for bioadhesive buccal drug delivery.
The aim of the present study was to develop a new
bioadhesive sustained release tablets for buccal drug delivery
of tizanidine hydrochloride.
531
buffer at 4C upon collection. The epithelium was separated
from underlying connective tissues with surgical scissors and tied
to the one side of open tube and this side of the tube (donor
chamber) was brought in contact with the surface of the 50 ml
pH 6.6 buffer solution (28) which was taken in 100 ml glass
beaker (receiver chamber). After the buccal membrane was
equilibrated for 30 min with buffer solution between both the
chambers, the receiver chamber was lled with fresh buffer
solution (pH 6.6), and the donor chamber was charged with 5 ml
(1 mg/ml) of drug solution. Aliquots of 5 ml were collected at
predetermined time intervals up to 6 h and the amount of drug
permeated through the buccal mucosa was then determined by
measuring the absorbance at 319 nm using a UV spectropho
tometer. The medium was replaced with equivalent volume
(5 ml) of buffer, which was prewarmed at 37C (29). After
performing the experiment in triplicate (n=3), mean values were
calculated. The cumulative amount of the permeated drug was
plotted against time. The ux (J) and the permeability coefcient
(P) were calculated by using the following Eqs. 1 and 2:
J dQ=dtA

P dQ=dt $CA

Materials
Tizanidine hydrochloride is a gift sample from Vilin
Biomed Ltd. (Rurki, India). Hydroxylpropyl methylcellulose
K4M, sodium carboxymethyl cellulose (NaCMC), and ethyl
cellulose are gift samples from Zydus Cadila (Ahmedabad,
India). Beta cyclodextrin ( CD) and hydroxylpropyl beta
cyclodextrin (HP CD) are provided by Dr. Reddys labo
ratories (Hyderabad, India). Sodium deoxycholate (SDC)
was purchased from Moly chem (Mumbai, India).
Methods
Solubility Studies
The solubility of TZD HCL in phosphate buffer solution
of pH 6.6 was determined by phase equilibrium method. An
excess amount of drug was taken into 50 ml conical asks
containing 20 ml of pH 6.6 phosphate buffers. These asks
were closed with aluminum foil and constantly agitated at
room temperature for 24 h using rotary shaker (Remi
Instruments, Mumbai, India). After 24 h, the solution was
ltered through a 0.2 m Whatman lter paper. The amount
of drug solubilized was then estimated by measuring the
absorbance at 319 nm using a UV spectrophotometer
(Systronic Pc Based Double Beam Spectrophotometer 2202,
Ahmedabad, India) (26). The studies were repeated in
triplicate (n=3), and mean was calculated.
Ex Vivo Permeation of Drug Solution
Ex vivo permeation study of TZD HCL through the
porcine buccal mucosa was performed using dissolution cell
and membrane assembly (27), at 370.2C and 50 rpm. The
temperature and revolutions per minute were maintained by
using magnetic stirrer (Remi, 2MLH, Mumbai, India). Porcine
buccal mucosa was procured from a local slaughterhouse and
used within 1 h of slaughter. The tissue was stored in Krebs
Where J is ux (mg h 1 cm 2); P is permeability coefcient

(cm h 1); dQ/dt is the slope obtained from the steady state
portion of the curve; C is the concentration difference across
the mucosa and A is the area of diffusion (cm2).
Preparation of Double Layered Buccal Tablets
The formulations were prepared as shown in Table I. Each
tablet contains 4.56 mg of TZD HCl, which is equivalent to 4 mg
of tizanidine base. Before direct compression, all the ingredients
were screened through sieve no. 100 and then thoroughly
blended in glass mortar with pestle. Blending was carried out
separately for core (polymer and drug) and backing layer. The
powder of backing layer was compressed using 8.0 mm at
faced beveled edge punch and dies set, on 16 stage rotary tablet
compress machine (Cadmach, Ahmedabad, India) and blended
powder of core layer was added on previously obtained backing
layer and compressed again (30). In case of formulations F8 and
F9 complexing with beta cyclodextrin ( CD) and hydroxyl
propyl beta cyclodextrin (HP CD), respectively, drug to
cyclodextrins were taken in 1:1 molar ratio. The amount of
cyclodextrins required for single dose was shown in Table I.
Complexes were prepared as follows; rst, cyclodextrins were
taken in glass mortar and little amount of water was added to
make a slurry; later, the powder was added to the slurry by
continuous trituration with pestle and this process is continued
up to 30 min. This slurry containing drug and cyclodextrin was
dried at 60C for 15 min. After drying, the complex was passed
through the sieve no. 100. The residual moisture content was
found to be not more than 1.03% w/w 15 min at 60C in
Sartorius IR Balance moisture analyzer.
Thickness
The thicknesses of buccal tablets were determined using
digital micrometer (Digital Caliper, Aerospace, India). Ten
Shanker et al.
532
Table I. Composition of Double Layer Buccal Tablets of Tizanidine Hydrochloride
Ingredients (milligram per tablet)
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
F12
TZD HCL
HPMCK4M
SCMC
CD
HP CD
SDC
Pearlitol
Aspartame
Aerosil
Backing
Layer (EC)
4.56
95
4.56
4.56
80
15
4.56
65
30
4.56
50
45
4.56
35
60
4.56
20
75
4.56
20
75
17.81
4.56
20
75
4.56
20
75
4.56
20
75
4.56
20
75
1.04
4
1
1
50
2.09
4
1
1
50
3.12
4
1
1
50
95
27.36
4
1
1
50
4
1
1
50
4
1
50
4
1
1
50
4
1
1
50
4
1
1
50
4
1
1
50
4
1
1
50
4
1
1
50
TZD HCL tizanidine hydrochloride, SCMC sodium carboxymethyl cellulose, SDC sodium deoxycholate, EC ethyl cellulose
individual tablets from each batch were used and the average
thickness was calculated.
Weight Variation Test
Weight variation test was performed for ten tablets from
each batch using an electronic balance (Shimadju, Aux*220,
Japan) and average values were calculated.
Hardness
Hardness test was conducted for three tablets from each
batch using Monsanto hardness tester and average values
were calculated
removable weights on right pan and equivalent amount of

clay on other side. The height of the total setup was adjusted
so as to accommodate a glass container of 6.6 cm height. In
order to nd the bioadhesion strength, rst, buccal tablet (n=
3) was stuck to the glass slide with cyanoacrylate adhesive
and balance was set in to weighing mode with the help of a
knob that is situated at the base of the balance. Now, 5 g of
weight on the right pan is removed. This lowered the glass
slide along with the tablet over the membrane with a weight
of 5.0 g. The entire setup was kept undisturbed for 5 min.
Then, the weights on the right hand side were slowly added in
increments of 0.1 g until the tablet just separated from the
membrane surface. The excess weight on the right pan, i.e.,
total weight minus 5 g, was taken as a measure of the
bioadhesive strength.
Assay
Determination of the Ex Vivo Residence Time
Five tablets were selected at random and were powdered
in a mortar; and amount of powder equivalent to single dose
was dissolved in 0.1 N HCL (31) by sonication for 30 min and
ltered through Whatman lter (0.2 m) paper. The drug
content was analyzed spectrophotometrically at 319 nm using
a UV spectrophotometer. Each measurement was carried out
in triplicate and the average drug content was calculated.
The test was performed for buccal tablets without

backing material. Form each batch, six randomly selected
tablets were placed in US Pharmacopeia (USP) disintegration
apparatus baskets (Electrolab ED 2L) and the process of
disintegration was carried out for 4 h. Later, the baskets were
lifted from the uid and observed for complete disintegration
of tablets (32).
The ex vivo residence (ER) time was found using a

locally modied USP disintegration apparatus, which was
applied by Nakamura et al. (34). The disintegration medium
was composed of 800 ml pH 6.6 phosphate buffer maintained
at 37C. The porcine buccal tissue was tied to the surface of a
glass slab, vertically attached to the apparatus. The buccal
tablet was hydrated from one surface using 0.5 ml of pH 6.6
phosphate buffer and then the hydrated surface was brought
in contact with the mucosal membrane. The glass slide was
vertically xed to the apparatus and allowed to run in such
way that the tablet completely immersed in the buffer
solution at the lowest point and was out at the highest point.
The time taken for complete erosion or dislodgment of the
tablet from the mucosal surface was noted. The experiments
were performed in triplicate (n=3) and mean of triplicate was
determined.
Measurement of Bioadhesion Strength
Swelling Studies of Buccal Tablets
Bioadhesive strength (BS) of the disks was measured on

a modied physical balance (33). The apparatus consists of a
modied double beam physical balance in which the right pan
was replaced with a lighter pan and the left pan was replaced
with a glass slide (4 cm length and 2.5 cm width) which was
suspended by means of Teon rings and copper wire. The
setup was balanced in such a way that it consists of 5 g of
Buccal tablets were weighed individually; initial weight

was considered as W1 and placed separately in Petri dishes
(outside dimensions: 100 mm diameter15 mm height; inside
dimensions 88 mm diameter12 mm height) containing 15 ml
of phosphate buffer (pH 6.6) solution in such a way that the
side of tablet which attaches to the buccal membrane was
positioned to the bottom of the Petri dishes with the backing
Disintegration Test
533
membrane being viewable from the top. Tablets were soaked

in such a way that the core tablet completely immersed in the
buffer solution. At regular intervals (0.5, 1, 2, 3, 4, 5, and 6 h),
the buccal tablets were removed from the Petri dishes using
coverslips and excess surface water was removed carefully
using the Whatman lter paper. The swollen tablets were
then reweighed (W2) (35,36). This experiment was performed
in triplicate. The degree of swelling (water uptake) was
calculated according to the following Eq. 3:
Degree of swelling W2
W1 =W1
Surface pH Study
The bioadhesive buccal tablets (n=3) were made in contact
with 1 ml of distilled water and allowed to swell for 2 h at room.
The pH was measured by bringing the pH meter electrode
(Cyber Ph 14l, Cyberlab, India) in contact with the surface of
the tablet and allowing it to equilibrate for 1 min (37).
In Vitro Drug Release of Buccal Tablets
The USP XXIII rotating paddle method was used to
study the drug release from the buccal tablets. The dissolution
medium consisted of 200 ml of phosphate buffer pH 6.6. The
experiment was performed at 370.2C, with a rotation speed
of 50 rpm. The backing layer of buccal tablet was attached to
the glass slide with instant adhesive (cyanoacrylate adhesive).
The slide was placed at the bottom of the dissolution vessel.
Samples (5 ml) were withdrawn at predetermined time
intervals and equivalent amount was replaced with fresh
medium. The samples were ltered through Whatman lter
(0.2 m) paper and analyzed after appropriate dilution by UV
spectrophotometer at 319 nm.
Ex Vivo Permeation of Buccal Tablets
Ex vivo permeation study of TZD HCL tablets through
the porcine buccal mucosa was performed using dissolution
cell and membrane assembly (27), the temperature was
maintained by using a water bath and a thermometer
assembled to it. Simulated buccal movements were main
tained by using magnetic stirrer (Remi, 2MLH, Mumbai,
India). Porcine buccal mucosa was procured from a local
slaughterhouse and used within 1 h of slaughter. The tissue
was stored in Krebs buffer at 4C upon collection. The
epithelium was separated from underlying connective tissues
with surgical scissors and tied to the one side of open tube
and this side of the tube (donor chamber) was brought in
contact with the surface of the 50 ml pH 6.6 buffer solution
(28) which was taken in a 100 ml glass beaker (receiver
chamber). After the buccal membrane was equilibrated for
30 min with buffer solution between both the chambers, the
receiver chamber was lled with fresh buffer solution
(pH 6.6), and the buccal tablet was carefully placed in the
donor chamber using a forceps in such a way that core tablet
can abreast to the buccal membrane in which 1 ml of buffer
solution (pH 6.6) was added (38). Samples (5 ml) were
collected at predetermined time intervals and ltered through
a 0.2 lter, and the amount of drug permeated through the
buccal mucosa was then determined by measuring the
absorbance at 319 nm using a UV spectrophotometer. The

medium in the receiver chamber was replaced with an
equivalent volume of buffer (5 ml), which was prewarmed
at 37C. The experiments were performed in triplicate (n=3)
and mean value was used to calculate the ux and perme
ability coefcient. The enhancement ratio was determined by
dividing the cumulative amount of permeated TZD HCL in
the presence of SDC at the end of 6 h (Q)6 h(enh) by the
amount of TZD HCL alone Q6 h(control), Eq. 4;

Enhancement ratio Q6h enh Q6h control
Stability of Buccal Tablets

Stability studies of buccal tablets were performed for
optimized formulation (F12) in normal human saliva. The
saliva was collected from humans (aged 22 26) and ltered
through Whatman (0.2 m) lter paper. Buccal tablets were
placed in separate Petri dishes containing 5 ml of human
saliva and placed in a temperature controlled oven (Bio
Technics, India) for 6 h at 370.2C. At regular time intervals
(0, 2, 4, and 6 h), the buccal tablets were examined for change
in color, surface area, and integrity (39). The experiments
were repeated in triplicate (n=3) in a similar manner.
In Vivo Mucoadhesive Performance of Buccal Tablets
This study was conducted after obtaining permission
from the institutional human ethical committee and then
informed consent was obtained from all the volunteers before
conducting the study. This study was conducted as per the
guidelines prescribed by the committee under the supervision
of the principal investigator. In vivo studies were conducted
by applying tablets (n=3) on ve healthy volunteers (aged
22 26 years) gums to obtain the residence time, the subjective
parameters, and loss of the fragment, and the possible
production of irritation or pain. Volunteers were given
optimized buccal tablets (F12) along with an instruction sheet
and were asked to press the buccal tablet against the buccal
mucosa for about 1 min. For the purpose of the photography
proof, in one volunteer, buccal tablet was applied to the inner
side of the lower lip and photographs were taken immediately
after application and after 2, 4, and 6 h. In vivo behavior of
the bioadhesive buccal tablet was shown in Fig. 1.Volunteers
were then asked to record the time of application and time of
dislodgment of tablet. After completion of the study, a
questionnaire was given to volunteers to collect information
regarding the parameters such as irritancy, comfort, taste, dry
mouth, salivation, dislodgment of the system during the study,
and heaviness of the system at the place of application.
Though food intake is avoided from 0.5 h prior to the
beginning of the study to the end of the study, water intake is
permitted after the initial 0.5 h (40).
RESULTS
Solubility of the drug in the pH 6.6 phosphate buffer was
found to be 12.85 mg ml 1. The ux and permeability coefcient
of drug solution was found to be 0.0775 mg h 1 cm 2 and
0.0180 cm h 1, respectively. The values of weight variation and
Shanker et al.
534
Fig. 1. In vivo mucoadhesive behavior of optimized formulation (F12)
friability were found to be within the limits of conventional oral

tablets stated in the Indian Pharmacopoeia (1996). In addition
to these evaluation tests, disintegration, thicknesses, hardness,
and percentage of drug content tests were conducted, and the
results were shown in Table II.
in the formulation F7. F1 was found to have low BS and ER

of 4.012 g and 2 h, respectively. Formulations from F10 to F12
have shown similar results as that of the F7. The BS, ER, and
surface pH values were shown in Table III.
Swelling Studies of Buccal Tablets

Measurement of Bioadhesive Strength and Ex Vivo
Residence Time
These evaluation tests were conducted for all the
formulations and there was a gradual increase in both the
bioadhesive strength and the ex vivo residence time form F3
to F7. The maximum BS (21.152 g) and ER (8 h) were found
From F3 to F7, maximum swelling was found at 4 h; from

these, F7 was found to have high 770.67 of swelling at 4 h.
Nearly same degree of swelling was observed for F10, F11,
and F12 as that of the formulation F7, whereas maximum
swelling was found at 3 h for F1 and F9. The comparison of
degree of swelling of all formulations was shown in Fig. 2.
Table II. Physicochemical Properties of Double Layer Buccal Tablets of Tizanidine Hydrochloride
Form code
Weight variation
F (%)
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
F12
202.712.85
203.363.70
201.173.68
202.104.07
202.114.04
203.124.25
202.774.25
222.072.18
233.181.40
204.761.68
204.661.82
205.061.60
0.47
0.23
0.11
0.11
0.05
0.02
0.04
0.04
0.04
0.49
0.23
0.03
F friability, DT is disintegration time
DT (h)
0.830.26
2.500.63
0.920.38
1.250.52
1.330.61
2.920.38
3.920.49
4.830.26
0.830.26
3.920.38
3.920.58
3.920.38
Thickness (mm)
Hardness (kg/cm2)
Percent drug content
2.700.08
2.680.09
2.720.06
2.690.08
2.710.06
2.690.08
2.700.07
3.130.02
3.430.03
2.740.03
2.750.02
2.750.03
3.80.3
5.00.5
4.30.1
4.30.3
4.70.8
4.70.8
5.00.5
8.00.5
5.20.3
5.00.5
5.00.5
5.00.5
99.430.53
99.350.38
99.320.41
99.420.05
99.220.29
99.360.27
99.410.50
99.700.34
99.270.34
99.370.34
99.290.26
99.450.36
535
Table III. Results of Bioadhesion Strength, Ex Vivo Residence Time

Surface pH of All Formulations (n 3)
Form code
Bioadhesion strength (g)
ER (h)
Surface pH
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
F12
4.0130.002
16.5030.003
6.0120.001
6.6740.001
7.6370.001
15.8520.001
21.1530.002
16.0660.001
17.0040.004
21.1590.001
21.1610.001
21.1600.001
2.170.29
5.500.50
2.500.50
2.670.29
3.500.50
4.000.50
6.670.76
4.500.50
1.830.29
6.831.04
7.170.76
7.171.04
5.970.07
6.420.37
6.530.32
6.780.19
6.710.17
6.890.17
6.940.12
6.680.26
6.790.16
6.930.03
6.880.11
6.940.10
Fig. 3. Plot of cumulative percentage drug release vs time for F1 F7

formulations
ER ex vivo residence time
In Vitro Drug Release from Buccal Tablets

F1 and F3 were found to release 99.900.36% and 99.97
0.18% of the drug just within 2 h whereas F2 released 94.96
0.07% of the drug in 6 h. The formulation F7 was found to release
99.850.00% of the drug in 6 h. F9 released about 99.850.14%
drug in 3 h. Formulations F8 and F10 through F12 released the
drug similar to that of F7. The comparison of cumulative drug
release of all formulations was shown in Figs. 3 and 4.
Mathematical Model Fitting of In Vitro Drug Release
The in vitro percentage drug release of optimized
formulation F12 was attempted to t into mathematical
models. The n and R2 values for zero order, rst order,
Higuchi and Peppas, and Hixson Crowell models (41) were
represented in Table IV. The Peppas model is widely used,
when the release mechanism is not well known and when
more than one type of release is involved (42). The
semiempirical equation is shown as Eq. 5:
Mt =M1 kt n
Fig. 2. Plot of degree of swelling vs time for all formulations
Where Mt =M1 is fraction of the drug released at time

t; k represents a constant, incorporating structural and
geometrical characteristics of the buccal devices; and n is
the diffusion exponent, which characterizes the type of
release mechanism during the dissolution process.
For non Fickian release, the value of n falls between 0.5
and 1.0, while in case of Fickian diffusion, n=0.5; for zero
order release (case II transport), n=1; and for supercase II
transport, n is greater than 1. Observation of all the R2 values
indicated the maximum for Higuchi, Peppas, and Hixson
Crowell. The n value of formulation F12 was 0.686 and it also
had the highest R2 (0.9816).
Ex Vivo Permeation of Buccal Tablets
Based on the above results, formulations F7 to F12 were
selected for the ex vivo permeation study. The ux, permeation
coefcient, and cumulative drug permeated from formulation
F7 were found to be 0.3342 mg h 1 cm 2, 0.0844 cm h 1, and
30.06%, respectively. No improvement in drug permeation was
found in case of F8 and F10. There was improvement in
cumulative percentage drug permeated for the formulations
F9, F11, and F12. The values of ux, permeation coefcient, and
cumulative drug permeated from optimized formulation (F12)
Fig. 4. Plot of cumulative percentage drug release vs time for F8 F12

formulations
Shanker et al.
536
Table IV. Estimated Values of n (Diffusional Exponent) and R2 (Correlation Coefcient) for Optimized Formulation
Peppas model
Form code
Zero order R2
First order R2
Higuchi R2
R2
Hixson Crowell R2
F12
0.9130
0.9148
0.9698
0.686
0.9816
0.9972
were found to be 0.8127 mg h 1 cm 2, 0.2052 cm h 1, and 62.73

7.47%, respectively, with an enhancement factor of 1.9. The
slope, ux, and permeability coefcient for formulations F7 to
F12 were shown in Table V. Comparison of cumulative
percentage drug permeated for all (F7 to F12) formulations
was shown in Fig. 5.
Stability of Buccal Tablets
Based on the above result, stability studies were con
ducted only for optimized formulation F12. There was no
change in the color and integrity of the tablets. The change in
surface area (cm2) at 0, 2, 4, and 6 h was found to be 0.5, 1.32,
2.00, and 2.50 cm2, respectively.
In Vivo Mucoadhesive Behavior
This study was conducted for optimized formulation F12.
In bioadhesive buccal drug delivery, comfort of system in oral
cavity is given utmost importance. The result of ve healthy
human volunteers to each subjective parameter was calculat
ed and shown in Table VI.
DISCUSSION
The solubility of drug was conducted in pH 6.6 phos
phate buffer because it is the average pH of the oral cavity. F1
(HPMC K4M alone) and formulations containing high
concentration of the HPMC K4M (F3, F4, and F5) have
shown low disintegration time (DT) and ER due to rapid
disintegration. F2 (NaCMC alone) and formulations F6 to
F12 containing high concentration of NaCMC have shown
higher DT and ER; this might be due to slow disintegration
and slow water uptake by the formulations, whereas formu
lation F9 (contains HP CD) showed lower DT and ER; this
might be due to its high solubility (HP CD) and over
hydration of HP CD that weakens the formed bonds (38).
In 1986, Longer and Robinson (43) dened the term
bioadhesion as the attachment of a synthetic or natural
macromolecule to mucus and/or an epithelial surface. In

general, mucoadhesion/bioadhesion may be dened as the
adhesion between a bioadhesive polymer and mucus.
Mucoadhesion is considered to occur in four major stages:
wetting, interpenetration, adsorption, and formation of sec
ondary chemical bonds between mucus membrane and
polymer. The strength of mucoadhesion is affected by
different factors like molecular weight of the polymer, contact
time with membrane, degree of swelling of the polymer, and
the type of biological membranes used in the study (44). The
adhesion will increase with the degree of hydration until a
point where overhydration leads to a sudden decline in BS,
which might be due to the disentanglement at the polymer/
tissue interface. The degree of swelling was increased with the
increase in the concentration of NaCMC; this led the
formulations F2 and F5 F12 to have higher BS and F9 were
found to have low BS; this might be due to overhydration of
the HP CD. NaCMC is the polyanionic polymer containing
carboxylic groups, which form hydrogen bonds with the
tissue. Rapid rate of hydration of NaCMC led to higher
degree of swelling in a short period of time, which improved
entanglement of polymer chains with the mucus. This
hypothesis was conrmed with that previously reported by
Lerhr et al. (45). All these factors have contributed to the
higher BS of F12.
The surface pH of the buccal tablets was determined in
order to investigate the possibility of any irritation effects in
vivo, as acidic or alkaline pH may cause irritation to the
buccal mucosa. Surface pH of the optimized formulation F12
was found to be 6.85 (near to neutral pH). It was inferred that
neutral pH of the formulation does not cause any irritation to
the mucosa.
Appropriate swelling property of a buccal adhesive
device is required for uniform and prolonged release of drug
Table V. The Slope, Flux, and Permeability Coefcient of F7 to F12

Formulations
Form code
Slope
F7
F8
F9
F10
F11
F12
0.1679
0.1974
0.371
0.2035
0.2531
0.4083
J ux, P permeability coefcient
J (mg h
cm 2)
0.3342
0.3929
0.7385
0.4051
0.5038
0.8127
P (cm h 1)
0.0844
0.0992
0.1865
0.1023
0.1272
0.2052
Fig. 5. Plot of cumulative percentage drug permeated vs time for
formulations F7 to F12
537
Table VI. Response of Healthy Human Male Volunteers to Various Subjective Parameters (n 3)
SI no.
1
Criteria
Volunteers response (%)
Irritation
None
Slight
Moderate
Severe
Taste
Normal
Slightly
Very unpleasant
Pleasant
Very pleasant
Comfort
Very comfortable
Comfortable
Slightly uncomfortable
Moderately uncomfortable
Severely uncomfortable
Dryness of mouth
None
Slight
Moderate
Severe
Salivary secretion
None
Slight
Moderate
Severe
Heaviness at the place of attachment
None
Slight
Moderate
Severe
Dislodgement of the system during study
No
Yes
with proper mucoadhesion (46).The degree of swelling

indicated that the rate of swelling is directly proportional to
NaCMC content and inversely proportional to HPMC K4M
content. The high amount of water intake by NaCMC at a
faster rate might have resulted in the higher rate and extent
of swelling (F7, F10, F11, and F12). Some buccal tablet
formulations (F1, F9, and F3 to F5) did not preserve their
integrity throughout the experimentation and were disinte
grated with in 2 h. The highest loss was observed for the
buccal tablets (F1 and F3) containing a high concentration of
HPMC K4M as mucoadhesive polymer. Some of the buccal
tablet formulations (F1, F3, and F9) were not successfully
recovered and handled from the buffer solution. After
reaching the maximum degree of swelling (3 to 4 h), buccal
tablets (F2, F6, F8, and F9) did not maintain their integrity.
An ideal sustain release system should be able to release
the drug immediately to attain the therapeutic level at a faster
rate and maintain this drug level for a prolonged period of
time. The buccal tablets (F1, F9, and F3 to F5) released the
drug in 3 h, for the same reason as was explained for the
swelling of buccal tablets. The extensive swelling of formula
tion F2 creates a thick gel barrier, which retards and increases
the path length for the diffusion of the drug molecules from
100
80
20
80
20
80
20
20
60
20
90
10
100
buccal tablets; this might be the reason for lower cumulative

percentage of drug release (94.96). In order to overcome this
problem, a combination of the HPMC K4M and NaCMC (F3
to F7) was used. At 19% and 71% concentrations of HPMC
K4M and NaCMC, respectively, F7 released 99.64% of the
drug in 6 h. So this formulation F7 was selected to study the
effect of the permeation enhancers. Hence, formulations F8
to F12 have the same concentration of polymer as that of F7.
Due to high solubility of HP CD, F9 released the total drug
in 3 h. The remaining formulations, i.e., F8 and F10 F12
released the drug similar to that of F7. The CD (F8) has
higher compressibility than HP CD (F9); this led to
increased hardness (8 kg cm 2). This might be the reason
for retarded release for 6 h when compared with F9.
The n and R2 values of Peppas model indicated that the
release of TZD HCL was found to be non Fickian diffusion.
Hixson Crowell model shows (R 2 = 0.9972) that the
formulation mechanism of release depends on the thickness
and diameter.
The ex vivo permeation study was conducted by taking
the 50 ml of 6.6 pH phosphate buffer in receiver chamber in
order to maintain sink conditions. There was no effect of
CD in formulation F8. Addition of HP CD (F9) increased
Shanker et al.
538
the cumulative percentage of drug permeated; the reasons
might be due to forming complex with the individual
molecules which improves the diffusible form of the drug
species at the tablet buccal membrane interface and due to
increasing the solubility of TZD HCL. These results were
similar to those produced by Mira and Mario (38). The HP
CD also has the ability to remove cholesterol and phospho
lipids (especially phosphatidyl choline and sphingomyelin)
from the outer layer of the membrane, thus increasing the
permeability of hydrophilic molecules. The HP CD was
reported to dissolve the membrane components without
penetrating into the membrane. Therefore, the effects were
mild and reversible (47). All these effects might contribute to
enhanced permeation of the drug. From the result, it was
inferred that 1% of SDC (F10) had no effect on permeation
of the drug and 2% of SDC (F11) extracted only mucosal
lipid from the intercellular spaces. Thus, this enhances the
diffusivity of the drug via the paracellular (passing between
the cells) or polar route. Higher concentration of SDC (F12),
i.e., 3%, can extract lipids from the cell membranes, along
with the extraction of mucosal lipid from the intercellular
spaces by the formation of micelles. This resulted in
enhancing passive diffusivity of the drug via transcellular
(crossing the cell membranes and entering the cell) and
paracellular routes (20). It was mentioned that SDC can also
cause the uncoiling and extension of the protein helices,
which leads to opening of the polar pathways for diffusion
(48). All these effects might contribute to enhancing the
permeation of the drug.
From the stability studies, it was known that optimized
formulation F12 had stability in human saliva, which if failed
might have led to color change. It was reported that color of
the omeprazole changed to yellow when it was placed in
human saliva (49). Physical properties of the TZD HCL
buccal tablets such as thickness and diameter slightly changed
owing to swelling of the system in human saliva. Buccal
tablets have maintained their integrity in the natural human
saliva throughout the experiment, exhibiting sufcient
strength of the system.
From the human volunteer studies of optimized formula
(F12), it was observed that slightly bitter taste was found at
4 h, which might be due to higher swelling of the mucoadhe
sive polymers. The excess swelling was responsible for the
increased thickness of the buccal tablet and this led to
improved radial release of TZD HCL, which was negligible
during initial hours. This radial release increased the amount
of the drug into mouth and was responsible for slightly bitter
taste.
CONCLUSION
Development of bioadhesive buccal drug delivery of
tizanidine hydrochloride tablets was one of the alternative
routes of administration to avoid rst pass effect and provide
prolonged release. In addition, these formulations reduce the
need of frequent administration and enhance patient compli
ance. A combination of sodium carboxymethyl cellulose and
hydroxypropyl methylcellulose K4M results in sustained
buccal drug delivery. The in vitro drug release was found to
be non Fickian. The results strongly suggest that increase in
the permeation was due to the effect of sodium deoxycholate
on paracellular and transcellular pathways. From healthy

human volunteers, subjective parameters and mucoadhesive
behavior were found to be satisfactory.
ACKNOWLEDGEMENTS
This study was supported by the St. Peters Institute of
Pharmaceutical Sciences, Warangal, and Jangaon Institute of
Pharmaceutical Sciences, Jangaon, Warangal, India.
REFERENCES
1. Squier CA, Wertz PW. Structure and function of the oral mucosa
and implications for drug delivery. In: Rathbone MJ, editor. Oral
mucosal drug delivery. New York: Marcel Dekker; 1996. p. 1 2.
2. Gibaldi M. The number of drugs administered buccally is
increasing. Clin Pharmacol 1985;3:49 56.
3. Harris D, Robinson JR. Drug delivery via the mucous mem
branes of the oral cavity. J Pharm Sci 1992;81:1 10.
4. Davis SS, Daly PB, Kennerley JW, Frier M, Hardy JG, Wilson CG.
Design and evaluation of sustained release formulations for oral and
buccal administration. In: Bussmann WD, Dries RR, Wagner W,
editors. Controlled release nitroglycerin in buccal and oral form.
Basle: Karger; 1982. p. 17 25.
5. Schor JM, Davis SS, Nigalaye A, Bolton S. Susadrin trans
mucosal tablets. Drug Dev Ind Pharm 1983;9:1359 77.
6. Ishida M, Nambu N, Nagai T. Mucosal dosage form of lidocaine
for toothache using hydroxypropyl cellulose. Chem Pharm Bull
1982;30:980 4.
7. Bremecker KD, Strempel H, Klein G. Novel concept for a
mucosal adhesive ointment. J Pharm Sci 1984;73:548 52.
8. Anders R, Merkle HP. Evaluation of laminated mucoadhesive
patches for buccal drug delivery. Int J Pharm 1989;49:231 40.
9. Gandhi RB, Robinson JR. Bioadhesion in drug delivery. Ind J
Pharm Sci 1988;50:145 52.
10. Shojaei AH. Buccal mucosa as a route for systemic drug delivery:
a review. J Pharm Pharmaceut Sci 1998;1(1):15 30.
11. Nagai T, Machida Y. Advances in drug delivery: mucosal
adhesive dosage forms. Pharm Int 1985;6:196 200.
12. Chattarajee SC, Walker RB. Penetration enhancer classication.
In: Smith EW, Maibach HI, editors. Percutaneous penetration
enhancement. Boca Raton: CRC; 1995. p. 1 4.
13. Aungst BJ, Rogers NJ. Comparison of the effects of various
transmucosal absorption promoters on buccal insulin delivery.
Int J Pharm 1989;53:227 35.
14. Lee VHL, Yamamoto A, Kompella UB. Mucosal penetration
enhancers for facilitation of peptide and protein drug absorption.
Crit Rev Ther Drug Carrier Syst 1991;8:91 192.
15. Pitha J, Harman SM, Michel ME. Hydrophilic cyclodextrin
derivatives enable effective oral administration of steroidal
hormones. J Pharm Sci 1986;75:165 7.
16. Burnside BA, Keith AD, Snipes W. Microporous hollow bers as
a peptide delivery system via the buccal cavity. Proc Int Symp
Control Release Bioact Mater 1989;16:93 4.
17. Zhang J, Niu S, Ebert C, Stanley TH. An in vivo dog model for
studying recovering kinetics of the buccal mucosa permeation
barrier after exposure to permeation enhancers: apparent
evidence of effective enhancement without tissue damage. Int J
Pharm 1994;101:15 22.
18. Steward A, Bayley DL, Howes C. The effect of enhancers on the
buccal absorption of hybrid (BDBB) alpha interferon. Int J
Pharm 1994;104:145 9.
19. Ebert CD, Heiber SJ, Dave SC, Kim SW, Mix D. Mucosal
delivery of macromolecules. J Control Release 1994;28:37 44.
20. Hoogstraate AJ, Wertz PW, Squier CA, Bos Van Geest A,
Abraham W, Garrison MD, Verhoef JC, Junginger HE, Bodde
HE. Effects of the penetration enhancer glycodeoxycholate on
the lipid integrity in porcine buccal epithelium in vitro. Eur J
Pharm Sci 1997;5:189 98.

21. Nance PW, Bugaresti J, Shellenberger K, the North American
Tizanidine Study Group. Efcacy and safety of tizanidine in the
treatment of spasticity in patients with spinal cord injury.
Neurology 1994;44:S44 51.
22. United Kingdom Tizanidine Trial Group. A double blind,
placebo controlled trial of tizanidine in the treatment of spastic
ity caused by multiple sclerosis. Neurology 1994;44:S70 8.
23. Wagstaff AJ, Bryson HM. Tizanidine: a review of its pharmacol
ogy, clinical efcacy, and tolerability in the management of
spasticity associated with cerebral and spinal disorders. Drugs
1997;53:435 52.
24. Acorda Therapeutics, Inc.. Zanaex (tizanidine hydrochloride)
tablets and capsules prescribing information. Hawthorne:
Acorda Therapeutics; 2006.
25. Moffat AC. Clarks isolation and identication of drugs. London:
Pharmaceutical Press; 2006. p. 691.
26. Tanaka M, Yanagibashi N, Fukuda H, Nagai T. Absorption of
salicylic acid through the oral mucous membrane of hamster
cheek pouch. Chem Pharm Bull (Tokyo) 1980;28:1056 61.
27. Zuber M, Chemtob C, Chanmeil J. Sci Tech Pharm 1979;8:1.
28. Bandyopadhyay AK, Metia PK. In vitro evaluation of novel
mucoadhesive buccal tablets of oxytocin prepared with diospyros
peregrina fruits mucilage. Yakugaku Zasshi 2008;128:603 9.
29. Mumtaz AM, Chng HS. Design of a dissolution apparatus
suitable for in situ release study of triamcinolone acetonide from
bioadhesive buccal tablets. Int J Pharm 1995;121:129 39.
30. Remun Lpez C, Portero A, Vila Jato JL, Alonso MJ. Design
and evaluation of chitosan/ethyl cellulose mucoadhesive bilay
ered devices for buccal drug delivery. J Control Release
1998;55:143 52.
31. Sujatha K, Chitra K, Hettiarachchi DS, Vinay Krishna M,
Vasantha J. Spectrophotometric determination of tizanidine
hydrochloride. Indian J Pharm Sci 2003;65:519 20.
32. United States Pharmacopoeia, National Formulary. By authority
of united state Pharmacopoeia Convention, Inc., Meeting at
Washington, 2004, p. 2302.
33. Gupta A, Garg S, Khar RK. Measurement of bioadhesive
strength of mucoadhesive buccal tablets: design of an in vitro
assembly. Indian Drugs 1993;30:152 5.
34. Nakamura F, Ohta R, Machida Y, Nagai T. In vitro and in vivo
nasal mucoadhesion of water soluble polymers. Int J Pharm
1996;134:173 81.
539
35. Tirosh B, Baluom M, Nassar T, Friedman M, Rubinstein A. The
effect of Eudragit RL 100 on the mechanical and mucoadhesion
properties of polycarbophil dosage forms. J Control Release
1997;45:57 64.
36. Ritthidej GC, Phaechamud T, Koizumi T. Moist heat treatment
on physicochemical change of chitosan salt lms. Int J Pharm
2002;232:11 22.
37. Bottenberg P, Cleymaet R, Muynek CD, Remon JP, Coomans
D, Slop D. Development and testing of bioadhesive, uoride
containing slow release tablets for oral use. J Pharm Pharmacol
1991;43:457 64.
38. Becirevic Lacan M, Jug M. Inuence of hydroxypropyl beta
cyclodextrin complexation on piroxicam release from buccoad
hesive tablets. Eur J Pharm Sci 2004;21:251 60.
39. Kashappa Goud Desai H, Pramod Kumar TM. Preparation and
evaluation of a novel buccal adhesive system. AAPS PharmSci
Tech 2004;5:35.
40. Periolia L, Ambrogia V, Angelicia F, Riccia M, Giovagnolia S,
Capuccellab M, Rossia C. Development of mucoadhesive
patches for buccal administration of ibuprofen. J Control
Release 2004;99:73 82.
41. Costa P, Sousa Lobo JM. Modeling and comparison of dissolu
tion proles. Eur J Pharm Sci 2001;13(2):123 33.
42. Peppas NA. Analysis of Fickian and non Fickian drug release
from polymers. Pharm Acta Helv 1985;60:110 1.
43. Longer MA, Robinson JR. Fundamental aspects of bioadhesion.
Pharm Int 1986;7:114 7.
44. Park H, Robinson JR. Mechanisms of mucoadhesion of poly
(acrylic acid) hydrogels. Pharm Res 1987;4:457 64.
45. Lehr CM, Bouwstra JA, Schacht EH, Junginger HE. In vitro
evaluation of mucoadhesive properties of chitosan and some
other natural polymers. Int J Pharm 1992;78:43 8.
46. Peppas NA, Bury PA. Surface interfacial and molecular aspects
of polymer bioadhesion on soft tissues. J Control Release
1985;2:257 75.
47. Challa R, Ahuja A, Ali J, Khar RK. Cyclodextrins in drug delivery:
an updated review. AAPS PharmSciTech 2005;6:E329 57.
48. Gandhi R, Robinson J. Mechanisms of penetration enhancement
for transbuccal delivery of salicylic acid. Int J Pharm 1992;85:129 40.
49. Choi HG, Kim CK. Development of omeprazole buccal adhesive
tablets with stability enhancement in human saliva. J Control
Release 2000;68:397 404.

DOI: 10.1208/s12249 009 9238 x
Research Article
Development and Characterization of 99mTc-timolol Maleate for Evaluating
Efficacy of In Situ Ocular Drug Delivery System
Himanshu Gupta,1,2,4 M. Aqil,1 R. K. Khar,1 Asgar Ali,1 Aseem Bhatnagar,2 Gaurav Mittal,2 and Sanyog Jain2,3
Received 12 August 2008; accepted 9 April 2009; published online 8 May 2009
Abstract. In situ gel forming systems have drawn much attention of current researchers to overcome the
poor bioavailability from the conventional eye drops. The present work described formulation and
pharmacoscintigraphic evaluation of timolol maleate loaded chitosan/hydroxy propyl methyl cellulose
(HPMC) based polymer matrix for enhanced ocular retention. Chitosan and HPMC ratio was optimized
and formulation was characterized for various in vitro parameters. The ocular retention was studied on
New Zealand rabbits by gamma scintigraphy, which is a very simple and noninvasive technique. For
scintigraphy study, the drug timolol maleate was radiolabeled 99mTc by direct labeling method using
SnCl22H2O as reducing agent. The labeling procedure was optimized to get maximum labeling efciency
(>98%). In vitro stability of the radiolabeled drug (99mTc timolol maleate complex) was checked and it
was found to be stable for up to 24 h. Plain drug eliminates rapidly as signicant activity was recorded in
kidney and bladder after 2 h of ocular administration. It was evident from the scintigraphic images and
the time activity curve plotted from the data that the plain drug solution cleared very rapidly from the
corneal region and reached into systemic circulation via nasolachrymal drainage system, as signicant
activity was recorded in kidney and bladder after 2 h of ocular administration. Developed formulation
cleared at a slow rate and remained at corneal surface for longer time duration. No radioactivity was
observed in systemic circulation after 2 h. Ocular irritation of the developed formulation was also
checked by hens egg chorioallantoic membrane test and formulation was found to be practically
nonirritant. The study signied the potential of gamma scintigraphy in evaluation of novel drug delivery
systems in a noninvasive manner.
KEY WORDS: chitosan; gamma scintigraphy; in situ gel; radiolabeling;
INTRODUCTION
Drug delivery in ocular therapeutics is a challenging
problem. Various ocular diseases like glaucoma, conjunctivi
tis, dry eye syndrome, etc. require frequent drug administra
tion. The major problem encountered in drug delivery to eyes
is the attainment of an optimum drug concentration at the site
of action. Poor bioavailability of drugs from conventional eye
drops is mainly due to the precorneal loss factors which
include rapid tear turnover, nonproductive absorption, tran
sient residence time in the cul de sac, and the relative
impermeability of the drugs to corneal epithelial membrane.
Sometimes, systemic absorption of the drug drained through
the nasolachrymal duct may result in some undesirable side
effects (1 3).
1
Department of Pharmaceutics, Faculty of Pharmacy, Jamia Ham

dard, Hamdard Nagar, Delhi 110062, India.
2
Department of Nuclear Medicine, Institute of Nuclear Medicine and
Allied Sciences, Defense Research and Development Organization,
Brig. S.K. Mazumdar Road, Timarpur, Delhi 110054, India.
3
Department of Pharmaceutics, National Institute of Pharmaceutical
Education and Research, Sector 67, SAS Nagar, Punjab 160062,
India.
4
To whom correspondence should be addressed. (e mail: himan
shu18in@yahoo.com)
99m
Tc; timolol maleate.
Due to these physiological and anatomical constraints,

only a small fraction of the administered dose (<1%) is
ocularly absorbed. This forces the clinician to recommend a
frequent dosing at an extremely high concentration and
pulse type dosing results in several side effects of ophthalmic
products. Novel ocular drug delivery systems offer some
improvement over a conventional liquid dosage form but,
because of blurred vision (e.g., ointments) (4,5) or lack of
patient compliance (e.g., inserts), they have not been
universally accepted (6,7). These problems may be overcome
by the use of in situ gel forming systems that are instilled as
drops into the eye and undergo a sol gel transition in the cul
de sac. Different types of in situ gel forming systems based on
different mechanisms (temperature, pH, or ion activated)
using different polymers have been explored for sustained
ocular drug delivery as reviewed by Nanjawade et al. (8).
Chitosan and pluronic F 127 based pH and temperature
triggered in situ gel for sustained ocular drug delivery have
been reported by our group (9).
Present study employed combination of chitosan and
hydroxyl propyl methyl cellulose (HPMC) for the develop
ment of sustained ocular drug delivery system. Chitosan is
reported to act as penetration enhancer that increases trans
corneal permeation of the drug. Besides this, other properties
of chitosan like bioadhesiveness, viscous nature, and ability to
540
Development and Characterization of
99m
Tc-timolol Maleate
541
convert into hydrogel at ocular pH make it the best suitable

candidate for the development of such type of delivery
systems. However, chitosan at higher concentration does not
give clear solution and upon instillation into the eye, it forms
hydrogel (white precipitate) due to precipitation of the
polymer at pH 7 7.0 that can hamper the vision. Further,
due its cationic nature, the polymer can cause ocular
irritation. So we tried to reduce the concentration of chitosan
by the addition of HPMC, a viscosifying polymer.
In the present study, the drug timolol maleate was
radiolabeled with radioactive technetium (sodium pertechne
tate; 99mTcO4 ) and incorporated into the chitosan/HPMC
based developed polymeric system and precorneal retention
and lacrimal clearance were studied by gamma scintigraphy,
which is a noninvasive technique and is a powerful tool in the
evaluation of new drugs/delivery systems and provides an
insight into the in vivo fate of the delivery system (10).
Optimization of Labeling Conditions (pH, SnCl2 Concentration)

The drug timolol maleate was received as gift sample
from M/s Ven Petrochem & Pharma (India) Pvt. Ltd.,
Mumbai, India. Chitosan (practical grade, 75 85% deacety
lated, molecular weight 150 kDa) was obtained as kind gift
from M/s, India Sea Foods, Cochin, India. HPMC K 100 and
all other chemicals and solvents used were purchased from
local suppliers and of analytical grade unless mentioned.
Labeling of Timolol Maleate with
99m
Tc
The timolol maleate was labeled with radioactive nuclide

Tc which was obtained as sodium pertechnetate in normal
saline eluted from molybdenum generator. It was procured
from the regional center of the Board of Radiation and
Isotopes Technology, INMAS, Delhi, India. For labeling, a
2.5 mg of the drug was dissolved in normal saline and mixed
with stannous chloride (SnCl2 1 mg/ml in 10% acetic acid).
The pH of the solution was adjusted by NaHCO3. To this,
99m
Tc (2 3 mci) was added and mixed properly. All the
labeling operation was carried out in hot laboratory under
lead shielding. Ultraviolet spectroscopy of 99mTc labeled
timolol maleate was taken to check any shift/changes from
original molecule.
99m
Determination of Labeling Efficiency

Labeling efciency was checked using instant thin layer
chromatography (ITLC). A drop of the formulation was
applied onto ITLC silica gel coated strips, which was run in
100% acetone as mobile phase. While reduced/hydrolyzed
99m
Tc (colloids) were estimated using pyridine acetic acid
water (PAW=3:5:1.5 v/v) as mobile phase. The strips were
dried and cut into two equal halves and radioactivity was
counted in each half using well type gamma counter
(CAPRAC R, Capintec, USA). Labeling efciency was
calculated from the following formulae:
%Labeling efficiency B 100=T B
%Colloids B 100=T B ;
Where: T = counts at top, B = counts at bottom
Labeling efciency of the drug molecules depends on

amount of reducing agent (SnCl2 concentration) and the pH of
the solution. Hence, to achieve maximum labeling efciency,
the two process variables were optimized. Labeling operation
was carried out using different concentration of SnCl2 while
keeping the pH constant. Labelling efciency was determined
as described previously (Table I). In another set of experi
ments, the amount of stannous chloride was kept constant and
the pH of the solution was varied from 5 to 7.5 by adding 0.5 M
NaHCO3 solution. The labeling efciency at different pH was
determined and optimum pH was selected (Table II).
In Vitro Stability of Labeled Complex
In vitro stability of the labeled formulations was evalu
ated by ITLC. A 100 ml aliquot of the labeled formulation
was mixed with 2.0 ml of phosphate buffered saline (pH 7.4)
and incubated at room temperature; change in labeling
efciency was monitored over a period of 24 h by ITLC as
described above (Table III).
Preparation of Placebo Formulation

Different combinations of placebo formulations were
developed and evaluated for gelling capacity and other
characteristics to identify the best suitable composition for
nal formulation. Chitosan was dissolved in saline solution,
pH adjusted to 5.5 6.0 by 1% v/v acetic acid. HPMC was also
dissolved in normal saline (Table IV). The gelling capacity
was determined by placing a drop of the system in a vial
containing 2 ml of articial tear uid freshly prepared and
equilibrated at 37C and visually assessing the gel formation,
noting the time of gelation and the time taken for the gel
formed to dissolve. Viscosity of the formulation before and
after gelation was measured using Brookelds viscometer
(model DV II, spindle no. 02, at 20 rpm), while clarity was
examined through visual inspection. The formulations are
selected/rejected on the basis of their clarity, turbidity, and
viscosity with change in pH. This method of measuring gelling
capacity by visual inspection and measuring the viscosity
using Brookeld viscometer is widely reported and adopted
by different researchers with appropriate modication as per
their laboratory setup (9,11 12).
Table I. Effect of SnCl 2 Concentration on Labeling Efciency

of Timolol Maleate
Amount of
SnCl2 (g)
Percent of
labeling efciency
Percent
of free Tc
Percent of
colloids
25
50
100
200
400
73.20.4
84.50.3
98.20.2
87.80.4
86.40.2
26.10.3
14.60.13
0.70.1
3.90.5
1.50.2
0.70.2
0.90.4
1.10.1
8.30.3
12.10.2
All values are expressed as meanSD (n 5)
Gupta et al.
542
Table II. Effect of pH on Labeling Efciency of Timolol Maleate
pH
Percent of labeling
efciency
Percent of free Tc
Percent of
colloids
5
5.5
6.0
6.5
7.0
7.5
870.2
900.3
97.50.1
98.20.3
940.2
900.1
12.30.3
80.2
2.00.5
0.80.12
50.1
80.3
0.70.3
20.13
0.50.12
0.20.03
10.3
20.2
Medicated Formulation
For antiglaucoma activity, timolol is prescribed as 0.25%
to 0.5% w/v solution. Hence, a nal drug concentration of
0.25% was used in formulation. Complete formula for the
developed formulation is shown in Table V. The required
quantity of timolol maleate (native or radiolabeled) to give a
nal drug concentration of 0.25% w/v was added to the
previously optimized placebo formulation. Methyl paraben in
a concentration of 0.1% w/v was used as preservative.
Osmolarity of formulation was determined by osmometer
(Fiske Associate, USA) and required quantity of sodium
chloride after calculation was added to make the solution
isotonic. The developed formulation was then lled in 10 ml
capacity amber colored glass vials, with a cap and dropper
with the teat. The formulation in its nal pack was subjected
to terminal sterilization by autoclaving at 121C and 15 psi for
20 min. The sterilized formulations were stored in a
refrigerator (4 8C) until further use. Formulation was tested
for different physicochemical properties as described and
results are shown in Table VI.
Table IV. Combinations of Chitosan/HPMC studied
Formulation
Chitosan (% w/v)
HPMC(% w/v)
1
2
3
4
5
6
7
8
9
10
11
12
0.25
0.5
1.0
0.25
0.5
1.0
0.25
0.5
1.0
0.25
0.5
1.0
2
2
2
1
1
1
0.5
0.5
0.5
0.25
0.25
0.25
Gelling
capacity
+++
+++
+++
++
+++
+++
++
+++
+++
+
++
++
analyzed for drug content at 294 nm using an ultraviolet

spectrophotometer. Results are shown in Fig. 1.
In Vitro Transcorneal Permeation Study
In vitro drug release kinetics from the prepared for

mulations was studied using a modied method reported
earlier (13). Two milliliters of the test solution were placed in
circular plastic cup (2.5 cm internal diameter and 1.2 cm in
depth). This was placed on an inverted US Pharmacopeia
basket which was kept inside a 250 ml beaker. Then, 100 ml
of simulated tear uid was added and stirred with a star
headed magnetic bead. Temperature of 371C was main
tained throughout the study. Five milliliter samples were
withdrawn at regular time intervals and fresh media were
added to replace the withdrawn samples. The samples were
Goat corneas were used to study the permeation of

timolol maleate across the corneal membrane. Whole eye
balls of goat were procured from a slaughter house and
transported to laboratory in cold condition in normal saline
maintained at 4C. The corneas were carefully removed along
with a 5 6 mm of surrounding scleral tissue and washed with
cold saline. The washed corneas were kept in cold freshly
prepared solution of tear buffer of pH 7.4. The study was
carried out in modied Franz diffusion cells following the
method reported by our group previously (9). It consisted of
four cells and each cell consisted of upper and lower
chambers. The upper chamber served as a donar compart
ment in which 100 l of drug solution or formulation under
study was placed. The composition of free drug solution was
exactly the same as the composition of formulation except it
did not contain any polymer. The upper and lower chambers
were separated by goat cornea. The lower chamber served as
a receiver compartment that was infused continuously with
simulated tear uid at the rate of 20 l/min. The whole system
was maintained at 370.5C. The perfusate was collected at
periodic time intervals for up to 4 h in a preweighed micro
centrifuge tubes and subjected to the quantication of timolol
maleate using high performance liquid chromatography method
(14). Timolol maleate was freely soluble in water with octanol/
water partition coefcient 1.8 and has pKa of 9.17.
Table III. In Vitro Stability of Radiolabeled Complex
Table V. Formula of the Developed In Situ Formulation
In Vitro Drug Release Profile
Time (hours)
Percent of labeling efciency
Percent free
Ingredients
0
1
2
4
24
98.80.3
98.70.13
98.60.4
97.40.1
94.50.21
1.20.2
1.30.1
1.20.13
2.60.32
5.50.21
Timolol maleate
Chitosan
HPMC
NaCl
Methyl paraben
Water (q.s.)
Concentration (w/v)
0.25%
0.5%
0.5%
0.45%
0.1%
100%
99m
Tc-timolol Maleate
543
Table VI. Physicochemical Properties of the Developed In Situ Gel

Formulation
India) and having free access to food and water. The study
was carried out under the guidelines compiled by the
Committee for the Purpose of Control and Supervision of
Experiments on Animals (Ministry of Culture, Govt. of India)
and all the study protocols were approved by the local
institutional Animal Ethics committee. Utmost care was
taken to ensure that animals were treated in the most human
and ethically acceptable manner.
Radiolabeled timolol maleate was incorporated into the
optimized placebo formulation along with other ingredients
as described previously. Gamma camera (Millennium VG,
USA), autotuned to detect the 140 KeV radiation of 99mTc
was used for scintigraphy study. Rabbits were anaesthetized
using ketamine HCl injection given intramuscularly in a dose
of 15 mg/kg body weight. The rabbits were positioned 5 cm in
front of the probe and 25 l of the radiolabeled formulation
(100 ci) was instilled onto the left corneal surface of the
rabbits. Recording was started 5 s after instillation and
continued for 20 min using 128128 pixel matrix. Individual
68 frames (6816 s) were captured by dynamic imaging
process. Region of interest was selected on the one frame of
the image and time activity curve was plotted to calculate the
rate of drainage from eye. A single whole body static image
also was taken after 2 h of instillation of drug/formulation.
Each formulation was tested on three rabbits.
Parameter
Inference
Clarity
pH
Osmolarity
Gelation pH
Viscosity (at pH 6.0)
Viscosity (at pH 7.2)
Clear solution
6.0 6.2
298 302 mOsmol
6.9 7.0
405 cps
15010 cps
Values are expressed as meanSD (n 5)
Ocular Irritation Test (HET-CAM Test)

For the present study, modied hens egg chorioallantoic
membrane (HET CAM) test as reported by Velpandian et al.
(15) was carried out. Briey, fertilized hens eggs were
obtained from poultry farm. Three eggs for each formulation
weighing between 50 and 60 g were selected and candled in
order to discard the defective ones. These eggs were
incubated in humidied incubator at a temperature of 37
0.5C for 3 days. The trays containing eggs were rotated
manually in a gentle manner after every 12 h. On day 3, egg
albumin (3 ml) was removed by using sterile techniques from
the pointed end of the egg. The hole was sealed by 70%
alcohol sterilized paralm (American Can Company, USA)
with the help of heated spatula. The eggs were kept in the
equatorial position for the development of CAM away from
the shell. The eggs were candled on the fth day of incubation
and everyday, thereafter, nonviable embryos were removed.
On the tenth day, a window (22 cm) was made on the
equator of the eggs through which formulations (0.5 ml) were
instilled.
A 0.9% NaCl solution was used as a control as it is reported
to be practically nonirritant. The scores were recorded accord
ing to the scoring schemes as shown in Table VII and score
obtained was given in Table VIII.
Gamma Scintigraphy
In vivo precorneal drainage of the developed formula
tion was assessed by gamma scintigraphy. Albino rabbits of
either sex weighing 2 3 kg were used for the study. Animals
were procured from the animal house of INMAS (Delhi,

Gamma scintigraphy is a technique whereby the transit
of a dosage form through its intended site of delivery can be
noninvasively imaged in vivo via the judicious introduction of
an appropriate short lived gamma emitting radioisotope. It
provides an insight in to the fate of the delivery system. In the
present work, we have used pharmacoscintigraphy as a
powerful tool in the evaluation of our developed in situ
formulation. Drug was labeled with radionuclide 99mTc. It was
chosen for the purpose because of its moderate half life (6 h).
Further, it emits gamma rays, which have relatively low
energy as compared to and rays, so it leads to no serious
health hazards to the workers. Drug was instantaneously
labeled with 99mTc. Ultraviolet spectrum of 99mTc timolol
maleate complex did not show any shift/changes from the
original molecule; hence, timolol maleate was not affected
therapeutically by labeling. Labeling efciency was checked
by ITLC using 100% acetone as mobile phase. The Rf value
of free Tc is approximately 0.9, so it reached to the top of the
ITLC strip while the drug Tc complex, due to difference in
Table VII. Scoring Chart for HET CAM Test
Fig. 1. In vitro drug release prole of developed formulation. Values

are expressed as meanSD (n 5)
Effect
Scores
Inference
No visible hemorrhage
Just visible membrane discoloration
Structures are covered partially
due to membrane discoloration
or hemorrhage
Structures are covered totally
due to membrane discoloration
or hemorrhages
0
1
2
Nonirritant
Mild irritant
Moderately irritant
Severe irritant
Gupta et al.
544
Table VIII. Scores Obtained in HET CAM Test
Scores
Time (in minutes)
Formulations
Normal saline as control
Developed formulation
Egg1
Egg2
Egg3
Mean
Egg1
Egg2
Egg3
Mean
15
30
60
120
240
480
1,440
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0.33
molecular weight, retained at the base of ITLC strip. So from

the difference in the top and bottom counts of the ITLC
strips, efciency could be calculated. The labeling procedure
also leads to formation of reduced/hydrolyzed Tc (colloids);
hence, to detect them, ITLC was run in another solvent
system PAW. Colloids were retained at the base while drug
Tc complex traveled to the top of the ITLC strip.
Various labeling parameters, e.g., SnCl2 concentration and
pH, were optimized and it was observed that at 100 g SnCl2
concentration and at pH 6.5 the maximum labeling efciency
(98.2%) was obtained. At these conditions, minimum colloids
(1.1%) were produced. In vitro stability of the labeled complex
was also checked and the complex was found to be stable for up
to 24 h.
The placebo formulations were developed using different
combination of chitosan and HPMC which were evaluated for
various physicochemical characteristics like gelling capacity,
physical appearance, and viscosity at formulation pH (pH 6.0)
and at ocular pH (pH 7.4). It was observed from the results
that increasing the concentration of HPMC imparted viscosity
to the formulation without affecting its clarity. Different
combination of chitosan and HPMC were prepared and
evaluated for its gelling capacity. It is prerequisite for an in

situ gel system that it allows easy instillation into the eye as
liquid drops which undergoes sol to gel transition, triggered
by rise in pH. A concentration of 0.5% of both chitosan and
HPMC was selected as it gave a colorless and transparent
formulation (Table IV), which further gave the formulation
prolonged resident time in cul de sac with no compromise
with vision.
A medicated formulation was prepared from the selected
placebo formulation. A dose of 0.25% of timolol maleate is
prescribed for glaucoma therapy; hence, we also used the
same concentration in our present formulation; 0.1% methyl
paraben was added as preservative and NaCl in calculated
amount was also added to maintain isotonicity of the nal
formulation (Table V). The amber colored bottle closed with
rubber closure and dropper with teat was used for packaging
and was found to be appropriate packaging system for current
formulation. The packaging was tested for resistance for
autoclaving, leakage, and pourability. The packaging passed
all the tests and proved to be a good choice for packaging of
present ocular formulation. Sterilization of the product was
done by autoclaving at 121C for 20 min at 15 psig and test
Fig. 2. In vitro transcorneal permeation prole of plain drug and drug in the developed
formulation. Values are expressed as meanSD (n 5)
99m
Tc-timolol Maleate
Fig. 3. Time activity curve shows precorneal drainage of various

formulations
for sterility was performed on autoclaved packaging accord

ing to IP 1996 standards. There has been no growth/microbial
contamination observed up to 14 days of incubation. Hence,
the formulation passed the sterility test.
The developed formulations were further characterized
for various physiological parameters, like clarity, gelation pH,
viscosity, and osmolarity. The optimized formulation was
isosmotic and gelation pH was found to be near 7.
In vitro drug release prole of the formulation was
determined in simulated tear uid (pH 7.4) and the
formulation demonstrates a slow release rate. Developed
formulation shows 28.43% cumulative drug release after
2 h, 42.44% after 6 h, and 88.23% after 12 h as shown in
Fig. 1.
545
In vitro transcorneal permeation studies were also
conducted and a higher permeation across goat cornea
was observed with chitosan/HPMC based formulation as
compared to plain drug solution (Fig. 2). This might be
attributed to the well known transmucosal enhancer prop
erty of chitosan.
Ocular irritation of the developed formulation was
checked by hens egg chorioallantoic membrane test which
is a rapid, sensitive, and inexpensive test. Testing with
incubated eggs is a borderline case between in vivo and in
vitro systems and does not conict with the ethical and legal
obligations. The chorioallantoic membrane of the chick
embryo is a complete tissue including veins, arteries, and
capillaries and is technically very easy to study. It responds to
injury with a complete inammatory process, a process
similar to that induced in the conjunctival tissue of the rabbit
eyes (15,16). Developed formulation was tested by this
method and the result was compared with those obtained
using normal saline, which was used as control that is
supposed to be practically nonirritant. A means score of 0
was obtained for normal saline. Chitosan/HPMC based
formulation was nonirritant up to 12 h (mean score 0) while
the mean score was found to be 0.33 up to 24 h (Table VII).
The study shows that the formulation is nonirritant to mild
irritant and is well tolerated.
Scintigraphic studies were conducted on Albino New
Zealand rabbits using radiolabeled timolol maleate in the
formulation. The observation of the acquired gamma camera
images showed good spreading over the entire precorneal
area for developed in situ gelling system immediately after
administration as compared with plain drug solution. The
curve of the remaining activity on the corneal surface as a
Fig. 4. Static whole body image after 2 h of drug administration a plain drug solution b developed in situ gel
system
546
function of time (time activity curve) was generated as shown
in Fig. 3. Plain drug solution cleared very rapidly from the
corneal region and reached into systemic circulation via
nasolachrymal drainage system as signicant activity was
recorded in kidney and bladder after 2 h of ocular adminis
tration (Fig. 4a), whereas chitosan and HPMC based formu
lation was cleared at slow rate and retained at corneal surface
for longer duration. No signicant radioactivity was observed
in systemic circulation (kidney and bladder) after 2 h
(Fig. 4b). Chitosan is both viscous and bioadhesive. Further,
viscosity of chitosan is increased as the pH of the formulation
is raised (>7) upon instillation into eye as a result of buffering
action of the tear uid.
CONCLUSION
Drug was successfully radiolabeled with 99mTc for
subsequent evaluation of the efcacy of the developed novel
delivery system by pharmacoscintigraphy. This noninvasive
technique has proved to be an important tool in evaluating
ocular drug delivery system especially in evaluating retention
and precorneal clearance. The developed chitosan/HPMC
based formulation was a nonirritant, enhanced transcorneal
drug permeation, and prolonged the retention at corneal site.
Formulation was found suitable for sustained topical drug
delivery to eyes for rational drug therapy in case of various
ocular diseases.
REFERENCES
1. Maurice DM. Kinetics of topically applied drugs. In: Saettone
MS, Bucci P, Padova S, editors. Ophthalmic drug delivery:
biopharmaceutical, technological and clinical aspects. Fidia
research series. vol 11. Padova: Liviana; 1987. p. 19 26.
2. Schoenwald RD. Ocular drug delivery: pharmacokinetic consid
erations. Clin Pharmacokinet 1990;18:255 69.
Gupta et al.
3. Middleton DL, Leung SS, Robinson JR. In: Lenaerts V, Gurny
R, editors. Bioadhesive drug delivery systems. Boca Raton:
CRC; 1990. p. 179 202.
4. Swarbrick J, Boylan J. Ocular drug formulation and delivery. In
encyclopedia of pharmaceutical technology. New York: Marcel
Dekker; 1995. p. 43 75.
5. Ranade VV, Hollinger MA. Intranasal and ocular drug
delivery. In Drug delivery systems. Boca Raton: CRC; 1996.
p. 209 38.
6. Felt O, Baeyens V. Mucosal drug delivery: ocular. Encyclopedia
of controlled drug delivery, vol II. Hoboken: Wiley; 1999. p. 605
26.
7. Chiou GCY, Watanabe K. Drug delivery to the eyes. In: Ihler
GM, editor. Methods of drug delivery: international encyclope
dia of pharmacology and therapeutics. London: Pergamon; 1986.
p. 203 10.
8. Nanjawade BK, Manvi FV, Manjappa AS. In situ forming
hydrogels for sustained ophthalmic drug delivery. J Control Rel
2007;122:119 34.
9. Gupta H, Jain S, Mathur R, Mishra P, Mishra AK, Velpandian T.
Sustained ocular drug delivery from a temperature and pH
triggered novel in situ gel system. Drug Deliv 2007;14(8):507 15.
10. Davis SS, Hardy JG, Newman SP, Wilding IR. Gamma
scintigraphy in the evaluation of pharmaceutical dosage forms.
Eur J Nucl Med 1992;19(11):971 86.
11. Balasubramaniam J, Kant S, Pandit JK. In vitro and in vivo
evaluation of the Gelrite gellan gum based ocular delivery
system for indomethacin. Acta Pharm 2003;53:251 61.
12. Srividya B, Cardoza RM, Amin PD. Sustained ophthalmic
delivery of ooxacin from a pH triggered in situ gelling system.
J Control Rel 2001;73:205 11.
13. Lin H, Sung KC. Carbopol/pluronic phase change solutions for
ophthalmic drug delivery. J Control Rel 2000;69(3):379 88.
14. Higashiyama M, Inada K, Ohtori A, Tojo K. Improvement of the
ocular bioavailability of timolol by sorbic acid. Int J Pharm
2004;272(1 2):91 8.
15. Velpandian T, Bankoti R, Humayun S, Ravi AK, Kumari SS,
Biswas NR. Comparative evaluation of possible ocular photo
chemical toxicity of uoroquinolones meant for ocular use in
experimental models. Ind J Exp Biol 2006;5:387.
16. Spielmann H. Ocular irritation. In: Castle JV, Gomez MJ,
editors. In Vitro methods in pharmaceutical research. San Diego:
Academic; 1997. p. 265 87.

DOI: 10.1208/s12249 009 9239 9
Research Article
Modulation and Optimization of Drug Release from Uncoated Low Density
Porous Carrier Based Delivery System
Praveen Sher,1,2,4 Ganesh Ingavle,2 Surendra Ponrathnam,2 Pankaj Poddar,3 and Atmaram P. Pawar1,4
Abstract. The purpose of this research work was to explore an application of uncoated porous drug carrier
prepared by single step drug adsorption for a delivery system based on integration of oating and pulsatile
principles intended for chronotherapy. This objective was achieved by utilizing 32 factorial design, solvent
volume (X1) and drug amount (X2) as selected variables, for drug adsorption using solvents, methanol, and
dichloromethane (DCM), of varying polarity. Nitrogen adsorption (N2), scanning electron microscopy of
cross sections, and atomic force microscopy were done to study adsorption patterns and their effect on
release pattern. Drug release study was customized by performing for 6 h in acidic environment to mimic
gastroretention followed by basic environment akin to transit phase. Correlation between porous data from
mercury and N2 adsorption was probably studied for the rst time. Observed regression analysis values for
pore volume, surface area, and drug release indicated the inuence of selected variables. Total release
range in acidic medium was 12.77 24.57% for methanol, 8.79 15.26% for DCM, and nal release of 69.45
92.23% for methanol, and 60.16 99.99% for DCM inuenced by varying internal geometries was observed.
Present form of drug delivery system devoid of any additives/excipients inuencing drug release shows
distinct behavior from other approaches/technologies in chronotherapy by (a) observing desired low drug
release (8%) in acidic medium, (b) overcoming the limitations of process variables caused by multiple
formulation steps and different characteristic polymers, (c) reducing time consumption due to single step
process, and (d) extending as controlled/extended release.
KEY WORDS: chronotherapy; oating pulsatile drug delivery system; low density porous carrier; pore
data; solvent polarity.
INTRODUCTION
During the past decade, there has been an exponential
growth in the investigations related to the application of
porous material in controlling temporal or distributional drug
release by oral, pulmonary, transdermal, and injectables
routes. Inherent attractive features, like stable uniform porous
structure, high surface area, tunable pore sizes with narrow
distribution, and well dened surface properties, make them
suitable for their inclusion in any form of delivery system.
Porous network, characteristic of an individual material, is
important in determining both natural and practical applica
tions such as adsorption, dissolution, and diffusion of drugs.
This allows them to adsorb drugs and release them in a more
reproducible and predictable manner. Until today, different
types of drugs with small and large molecular size have been
1
Department of Pharmaceutics, Poona College of Pharmacy and

Research Centre, Bharti Vidyapeeth University, Erandwane, Pune,
411038, Maharashtra State, India.
2
Polymer Science and Engineering Division, National Chemical
Laboratory, Pune, 411008, India.
3
Nanoscience Group, Material Chemistry Division, National Chem
ical Laboratory, Pune, 411008, India.
4
To whom correspondence should be addressed. (e mail: p atmaram
@rediffmail.com; sherpraveen@gmail.com)
evaluated (1 6). Various methods like stirring in drug solution

or suspension, immersion for long times until equilibrium is
reached, vacuum, emulsion formation, gravimetric method,
and solvent evaporation to entrap the drug within the carrier
are reported (6 11).
In our previous communications, we had demonstrated
the applications of a porous carrier, Accurel MP 1000, made
of isotactic polypropylene, as an uncoated drug carrier (12)
and its utility for the development of oating pulsatile drug
delivery system intended for chronotherapy (13). This thera
py, based on circadian rhythm, is contrary to principles of
present delivery systems based on providing variable/constant
drug amount over a period of time, where timed non uniform
release prole is found more benecial than a uniform one
(14).The essential features of this conceptual delivery system
exhibit as a (a) multiparticulate system, (b) combination of
gastroretentive and pulsatile principles in same dosage form,
(c) single step formulation process using low density porous
carrier as drug loading core, (d) idealistic drug release prole
that is effective at morning time intended for a particular
pathological condition based on chronotherapy, and (e) having
low drug release in stomach suited for non steroidal anti
inammatory drug (NSAID) class of drug (ibuprofen).
The contribution of low density porous carrier in the
development of this drug delivery system associates major
signicance by (a) ensuring the retention of dosage form in
547
Sher et al.
548
the stomach for an extended period without using any excipients
for enhancing oating, (b) simultaneous least release all through
this period (resembling lag phase) suited for NSAID drugs to
avoid gastric irritation, (c) choice of drug loading (melt and
solvent evaporation), and(d) limit/overcome various formula
tion variables by acting as a drug loading core using single
formulation step when compared with other approaches/meth
ods, which need multiple steps by using various polymers and
excipients to achieve such release prole. The implication of last
point seem to be realistic not only when compared with our
previous reported work using hydrogels for formulating the
same delivery system (15) but also with other methods based on
different approaches (10,16 20) and technologies like OROS,
CONTIN, CEFORM, TIMERx, etc. (21) developed for
chronotherapy.
Successful relevance of formulating this conceptual system
reported earlier was done by using just three batches via melt
and solvent evaporation method (13). Based on excellent
observed results, the need to study/explore further drug
adsorption via solvent evaporation for its simplicity along with
optimization by employing factorial design was considered,
thereby stating as the objective of present work. The same
design of experiment was used earlier in another study for
ascertaining as a drug carrier, which in turn acted as a source of
comparison for understanding various aspects of mass transfer
(12). Additional characterization process by drift Fourier
transform infrared analysis (FTIR), scanning electron micros
copy (SEM) of cross sectional areas, N2 adsorption, and atomic
force microscopy (AFM) was done for better understanding.
First attempt to nd any correlation between porosimetry data
obtained from N2 adsorption by Brunauer Emmett Teller
(BET) method and mercury porosimetry (12) for drug loaded
porous carrier was done. Drug release study was tailored for 6 h
as the maximum gastric oating time to evaluate the concept of
oating followed by 3 h in basic medium.
generous gift. Methanol (M), dichloromethane (DCM), and

other reagents were of analytical grade.
Preparation of Drug-Loaded Beads
Drug Loading
Ibuprofen was loaded onto the porous beads by solvent
evaporation. Accurel MP 1000 was closely sieved in the range
of 250 350 m to nullify effect due to variation in particle
size. In a typical study, various amounts of drug was dissolved
in the multiple volumes of solvent (M or DCM) followed by
the constant addition of 100 mg Accurel MP 1000 , kept to
evaporate solvent under ambient conditions.
Factorial Design
A 32 design was employed to nd the interaction
between the selected variables. The variables chosen were
volume of solvent (X1) and the amount of drug (X2) at three
different levels. The coded and the actual values of the
experimental design are given in Table I. The data analysis of
values obtained for pore volume, surface area, pore diameter,
and drug release at various intervals from all batches were
subjected to multiple regression analysis using statistical
software Unistat (Statistic version 3, Meglon, USA). The
equation tted was
Y 0 1 X1 2 X2 11 X12 22 X 22 12 X1 X2
where Y=measured response, X=levels of factors, and =
coefcient computed from the responses of the formulations.
Evaluation and Characterization of Microparticles
Yield and Drug Content
MATERIAL AND METHODS

Materials
Accurel MP 1000 (Membrana ,Germany), a low density
microporous polypropylene microparticles with particle size
<1,500 m, pore size in the range from 5 to 20 m, and void
volume of 70% and ibuprofen (Cipla, India) were obtained as a
The dried weight of microparticles after drug loading was

recorded as nal yield. The drug loaded beads were dissolved in
methanol kept on ultrasonicator, and the content of drug
content was assayed by determining (in triplicate) the absorp
tion at 221 nm using UV VIS Spectrophotometer V 500 (Jasco
International Co. Ltd., Japan). The experiment was repeated
thrice in order to establish accuracy and precision of the method.
Table I. Experimental Variables of Factorial Design with Their Coded Levels and Actual Values Yield and Drug Content
Batch
Coded levels
[solvent (X1), drug (X2)]
Solvent (ml) (X1)
1
2
3
4
5
6
7
8
9
1,1
1,0
1,1
0,1
0,0
0,1
1,1
1,0
1,1
1
1
1
3
3
3
5
5
5
m methanol, d DCM
Practical yield (%)
Drug content (%)
Drug (mg) (X2)
100
200
300
100
200
300
100
200
300
95.01.21
87.330.86
83.501.34
98.351.24
93.561.23
89.631.14
95.621.56
89.501.33
88.001.61
97.171.33
98.521.65
98.981.36
89.881.32
91.261.41
96.501.33
92.501.20
92.122.00
91.230.65
85.432.31
85.852.65
82.743.21
79.901.21
86.371.32
86.201.32
81.162.30
83.271.01
89.261.02
1000.16
1000.21
1000.50
94.631.56
99.251.37
92.71.58
99.251.09
97.701.54
93.521.07
Modulation and Optimization of Drug Release

Thermograms of ibuprophen and Accurel MP 1000
microparticles with and without drug were obtained using
differential scanning calorimeter (Mettler Toledo DSC 821e,
Switzerland) equipped with an intracooler. Indium standard
was used to calibrate the temperature and enthalpy scale. The
powder samples were hermetically sealed in perforated
aluminum pans and heated at constant rate of 10C/min
over the temperature range 25 150C. The system was
purged with nitrogen gas at the rate of 100 mL/min to
maintain inert atmosphere.
549
(Veeco Instruments Inc., Santa Barbara, CA, USA). Height
and deection images were taken and analyzed using the
NanoScope version 5.12r5 software in the ofine mode.
Porosity Measurement
X ray powder diffraction (XRD) patterns of drug and

Accurel MP 1000 beads with and without drug were
recorded by using an X ray diffractometer (Philips PW
1729, The Netherlands). Samples were irradiated with mono
chromatized Cu K radiation (1.542 ) and analyzed at 2
between 2 and 60. The voltage and current used were 30 kV
and 30 mA, respectively. The range and the chart speed were
5103 CPS and 10 mm/2, respectively.
Total surface area and the porosity of the Accurel MP

1000 microparticles with and without drug were measured
by nitrogen adsorption using a porosimeter (Quantachrome
instrument, Autosorb 1TM, gas sorption system, Ontario,
Canada). Briey, weighed amounts of samples were placed
in the glass cells and outgassed with nitrogen at 25C for 3 h
before analysis. Subsequently, the sample and the reference
cells were immersed in liquid nitrogen at 196C, and
absorption isotherm was obtained from the volume of
nitrogen (cm3/g) adsorbed onto the surface as a function to
relative pressure. Total surface area was calculated by BET
method. Various evaluation parameters were obtained by
using Autosorb 1TM software. Mercury intrusion porosimetry
was done using Mercury porosimeter (Autoscan 33 USA).
Samples were rst loaded with mercury in a pressurized cell
under vacuum and later subjected to pressure range of
0 33,000 psi (12).
Thermogravimetric Analysis
Dissolution Release Studies
Appropriately weighed drug loaded microparticles were

subjected to gravimetric assay between 35 and 175C, with
the heat ow of 5C/min using Thermogravimetric Instru
ment (Seiko TG/DTA 32, Japan). All experiments were
performed in the presence of static air.
The dissolution of drug loaded Accurel MP 1000

beads was studied using USP XXIV type II dissolution test
apparatus (Electrolab TDT 06P, India), containing 900 mL of
X ray Powder Diffraction
Fourier Transform Infrared Analysis

FTIR measurements of drug, Accurel MP 1000, and
drug loaded ones were obtained on FTIR spectrophotom
eter (Spectrum One, Perkin Elmer Instruments, UK).
Samples were prepared by mixing with KBr and placing
in the sample holder. The spectra were scanned over the
wave number range of 3,600 to 400 cm 1 at ambient
temperature.
Surface Topography
Microphotographs of the cross sections of beads were
observed using scanning electron microscope (Cambridge
Stereoscan 120, UK) operated with an acceleration voltage of
10 kV. The beads were mounted on the standard specimen
mounting stubs and were coated with a thin layer (20 nm) of
gold in sputter coater unit (VG Microtech, UK).
Atomic Force Microscopy
AFM measurements were performed in ambient condi
tion on coated pellets without any further sample prepara
tion. The AFM measurements were performed repeatedly on
many different pellets and on several different areas.
Although variations were seen, the images presented were
typical of those most frequently observed. All the AFM
measurements were done in the contact mode using a
commercial Multimode with Nanoscope IV controller
Fig. 1. FTIR spectra of Accurel MP 1000 microparticles, drug and

different batches
Sher et al.
550
pH 1.2 HCl and pH 7.2 phosphate buffer maintained at 37
0.5C and stirred at 100 rpm for 6 and 3 h, respectively.
Samples were collected periodically and replaced with a fresh
dissolution medium. Data were analyzed using PCP Disso
software (v2.08, Poona College of Pharmacy, India). All
readings were made in triplicate.
Stability Studies
The stability of few selected drug loaded Accurel MP
1000 batches was monitored up to 3 months at ambient
temperature and relative humidity (30C/60% RH). Samples
were removed and characterized by dissolution studies and DSC.
Fig. 2. SEM of a Accurel MP 1000 microparticles (500), b batch 4 methanol (750), c (750) and d
(1.00K) batch 6 methanol, e (750) and f (2.0K) batch 4 DCM, g (750) and h (2.0K) batch 6 DCM

The observed values for yield and drug content are given
in Table I. Summary of the earlier reported results entail the
inuence of solvent nature, volumes, and drug amount on the
observed values, which was well supported by statistical
analysis. DSC and XRD predicted no change in crystalline
nature during adsorption. Inuence of pore shielding effect
using different solvents suggesting discreet drug adsorption
was observed by TGA.
Evaluation of Drug-Loaded Microparticles
FTIR results are displayed in Fig. 1. Peak values from 2,961
to 2,722 cm 1 of Accurel MP 1000 showing multiple peaks
transformed in to sharp triplet peaks after drug adsorption
without changing the values of other peaks like 2955, 2922,
2869 cm 1 so on. Same was true for particular drug peaks, which
show perfect repetition in the range of 1,507 to 2,173 cm 1.
Apart from them, various other peaks also showed excellent
reproducibility. Another characteristic peaks around 2,625 and
2,730 cm 1 indicated toward the dimmer formation signifying
the interaction of drug over the hydrophobic surface (22).
Overall, these results suggest the physical nature of interaction
rather than chemical one.
Surface Characteristics of Drug-Loaded Microparticles
In the pursuit of studying inner/core adsorption behavior
for descriptive follow up, SEM of cross sections of the
microparticles was considered as shown in Fig. 2. Regular
spread out adsorption pattern using methanol was observed
in Fig. 2b d. This can be related to lower viscosity and slow
phase transformation caused by high boiling point of solvent.
At low solvent volume, the drug adsorption is slightly
551
concentrated in upper layers, which change with increasing
solvent volume by showing migration toward the core as seen
in Fig. 2c and d. This demonstrates the interactive and mobile
behavior of drug, having both hydrophilic and hydrophobic
segments, over hydrophobic polypropylene surface, whereas
polar interactions take place between adsorbed molecules of
methanol and those in the solution resulting in observed
adsorption orientation. The magnitude of these interactions
can show dependence upon free energy of solid and the
dispersed adsorbate (22). On the contrary, DCM, having low
boiling with low polarity, indicating fast phase change, depicts
exactly an opposite pattern than observed with methanol
(Fig. 2e h). These characteristic pictures reveal the concentric
adsorption at the surface in the form of stakes running
throughout. The partial migration can be due to the
decreased polarity of solvent limiting wetting of hydrophobic
surface (23). Higher magnication reveal the hemispherical
adsorption behavior reported earlier using less polar solvent
with hydrophobic surface (24).
AFM was done for further evaluation of the surface
morphology for any difference in adsorption pattern as seen
in Figs. 3 and 4. A contact mode AFM measurement of batch
7 using methanol in air was observed (Fig. 3a f). All the
corresponding images and 3D plot suggest the true character
of the surface. The surface shows wavy nature, which is
indicated by the light and dark colors in the images, having
irregular patterns caused by depressions and counters.
Pictures revealed cracks and height conditions of around
1 m with a non regular pattern. Cracks were found even at
high magnications, which suggest the discontinuity of the
adsorbed drug layer. This type of surface may be due to the
underlying texture of the core structure and/or due to various
drying phenomenon occurring during solvent evaporation
governed by capillary and diffusion (25). Other factor that
surfaced can be due to the interactions taking place using
Fig. 3. AFM gures of batch 7 methanol: a height, b friction, c 3D graph for 10 m, d height, e friction, f 3D graph for 3.4 m
Sher et al.
552
Fig. 4. AFM gures of batch 7 DCM: a deection, b height, c 3D graph for 10 m, d deection, e height, f 3D graph for 5 m, g deection,
h height, i 3D graph for 3.1 m, j deection, k height, l 3D graph for 1 m
high polar solvent with hydrophobic surface as discussed

earlier.
With batch 7, using DCM revealed a large concentric
area of adsorbed drug with a considerable difference in the
layout of adsorbed drug pattern, as seen in Fig. 4a. The large
chunk of adsorbed drug area suggests the formation of stakes,
thereby showing a large depth on all sides. On further
magnication, the wavy nature becomes more prominent.
The size of surface taken at 33 m (Fig. 4g i) and 11 m
(Fig. 4j l) shows the layered form of adsorption phenomenon
suggested by a number of counters with a stepwise formation.
No cracks in the adsorbed drug pattern were observed,
indicating its continuity over the surface. This supports the

claim and explanation of interaction between the low polarity
solvent DCM with the hydrophobic surface/pores on the
adsorption pattern (24).
Porosity Data Evaluation
The drug loading process inuencing surface area and
porosity was studied by adsorbing N2 whose values are given
(Table II). All isotherms were of type III showing hysteresis
displayed due adsorption and desorption in mesopores (24).
Using methanol, recorded values for all evaluation parame
553
Table II. BET Values of Drug Loaded Beads Using Different Adsorption Methods
Surface area (m2 /g)
Pore volume (cc/g)
Pore radius ()
Batch
Methanol
DCM
Methanol
DCM
Methanol
DCM
1
2
3
4
5
6
7
8
9
Aa
4.016E+01
2.867E+01
3.254E+01
2.994E+01
3.626E+01
3.996E+01
2.951E+01
2.874E+01
2.201E+01
7.77E+01
4.139E+01
3.521E+01
3.347E+01
1.484E+01
1.963+01
1.801E+01
1.694E+01
1.818E+01
1.847E+01
1.656E01
5.77303
5.117E02
8.862E02
1.29201
1.425E01
9.135E02
7.586E02
4.988E02
1.98E01
1.785E01
1.4452E01
1.447E01
5.472E02
4.450E02
6.792E02
3.854E02
7.208E02
6.309E02
8.741E+01
8.727E+01
8.706E+01
8.792E+01
8.800E+01
8.719E+01
8.730E+01
8.718E+01
8.805E+01
1.36E+01
1.531E+02
1.547E+02
1.566E+02
1.567E+02
8.752E+01
1.527E+02
8.787E+01
1.562E+02
8.745E+01
BET Brunauer Emmett Teller

a
Accurel MP 1000
ters showed inuence of selected variables. With least solvent

volume (1 ml), observed values decreased with increasing
drug amount, which differed with increasing solvent volume
irrespective of drug amount used. The reason for this can
be attributed to the impact of variable solvent volumes,
wherein capillary (at low solvent volume) or pressure
gradient (at high solvent volumes leading to submerging)
or both work in manipulating the adsorption pattern (26).
This was evident from the graphs, which show increased
activity below 50 with increasing solvent volume
(supplementary data). However, batches using 200 mg drug
showed high values using increased solvent volumes. This
feature was not seen using mercury porosimeter, which
followed a particular trend (12).
Using DCM, recorded values for all evaluation parameters
showed varied patterns at different drug levels. Surface area
decreased with increasing solvent volume at lowest amount of
drug but with minor difference, which behaved oppositely at
other two levels (200 and 300 mg). Negligible difference
between the values was observed. This indicates the continuous
adsorption pattern on the surface at increased drug amount
where less/no accessibility of nitrogen gas was possible, thereby
reconrming our previous results. Once again, this sort of result
was not obtained using mercury porosimetry.
Statistical interpretation of the data by regression
analysis is given in Table III. Using methanol, observed
values display a non signicant negative impact of drug drug
interaction (2) for surface area and positive drug interaction
() for pore volume along with negative impact of drug
solvent () interaction. This distantly points toward the
possibility of interactions of highly polar nature of solvent
with drug containing both hydrophobic and hydrophilic
groups on hydrophobic surface of Accurel MP 1000. Using
DCM, drug amounts () predict negative inuence over
surface area (Fig. 5a) and pore volume (Fig. 5b). The closely
packed drug at the periphery of the surface may have hindered
the nitrogen, thereby showing negative inuence, while with
methanol, the observed adsorption pattern must have worked
conversely. Drug drug interaction shows positive values for
both response variables using DCM and negative response for
surface area using methanol. This can be related to modied
pore sizes and geometry after drug adsorption, which is more
limited to outer surface using DCM and spread throughout

using methanol.
Combined N2 adsorption and mercury intrusion porosim
etry experiments on various porous materials have indicated a
partial agreement concerning various aspects of evaluation
parameters. Drug loaded porous materials showed distinct
difference and similarity of statistical data on comparing the
present data with raw mercury porosimetry data conducted in
our previous work. Values related to the amount of drug
variable () showed its inuence during MPA only while
solvent solvent interaction was absent for both. The difference
between the analysis of mercury porosimetry data (12) and
nitrogen sorption data can be correlated to the (a) different
methodology basis of both techniques for their evaluation, (b)
shape of individual pores after drug loading, (c) type of pore
openings, (d) pore blockages, (e) relation between voids, and
throat (f) the cooperative percolation effect of the porous
network.
In vitro Drug Release
Basic theoretical background of drug release from hydro
phobic porous polymeric system depends on (a) medium
interaction (water) with surface, (b) drug dissolution in water
lled pores governed by pore volume and drug solubility, (c)
diffusion through water lled channels adopting least resistant
channels, (d) geometry and structure of pore network based on
contributing and non contributing porosity, (e) physical and
chemical interactions, (f) partitioning of the diffusant within the
matrix wall, and (g) low drug loading leading to recovery. All
these factors behave independently, opposite, or synergistically
to increase or retard the mass transfer, thereby leading to the
dissolution of drug inuencing drug release (8,27 29).
The present drug delivery system somewhat resembles
the air enclosed system, which showed a maximum oating
time of 9 h depending upon condition of stomach and supine
position of the subject (30). Based on the position of subject
(lying during night), working against the principle of gastro
retention, a oating study for 6 h was undertaken in pH 1.2
HCl IP medium maintaining sink conditions. All batches
showed satisfactory oating characteristics during the time
of experiment. This was followed by 3 h in pH 7.2
Sher et al.
66 86
5 05
17 12
10 44
0 9261
0 0029
phosphate buffer IP taken as an appropriate dissolution

medium to mimic increase in pH after oating. In our
previous work, maximum impact of burst release occurred
within 15 min followed by a very slow release or cutoff
using basic medium for the same formulations (12). In
present the formulation, being a multiparticulate system
where there is no all at once emptying is noted, the analysis
interval was set at 1 h after changing medium to normalize
time for gastric emptying time and simultaneously the burst
effect (13).
59 29
16 296
0 7720
0 0018
73 86
5 05
0 4822
0 0379
76 42
3 423
6 476
4 503
11 64
0 8916
0 0076
d
m
10 00
1 32
3 343
0 865
0 9149
0 0042
From Methanol Adsorbed Microparticles
23 10
4 755
3 781
0 9132
0 0007
1 383
16
2 253
0 9354
0 0003
3 791
0 978
1 30
1 33
0 7514
0 0569
47 06
17 61
0 8398
0 0005
49 59
9 52
5 28
0 8891
0 0014
0 052
0 043
0 049
0 9460
0 0002
0 094
0 015
0 1849
0 2479
m methanol, d DCM
26 35
7 28
0 3427
0 0977
C
2
2
R2
Sig
13 02
7 518
7 625
1 29
0 9923
0 0000
d
m
d
m
d
m
Coefcient
Surface area
Pore volume
Burst release
Percent drug release ( hours)

Response
Table III. Estimation of Regression Coefcients for Different Response Variables
554
Overall drug release show different release proles,

ranging 13 24% at the end of 6 h in acidic medium followed
by the burst release with range of 41 68% to the previously
released drug. The end release ranging 69 94% at 9 h is
shown in Fig. 6. Release prole of all batches at the end of 6 h
decreased with increasing amount of the drug used for
adsorption with maximum intradifference using 3 ml. In
creased surface area phenomenon along with contribution of
open pores seizing more dissolution medium due to low drug
amounts resulted in higher release rates. Batches with high
drug amounts showed the least release irrespective of solvent
volume used, which were nearly half or more than the
starting least amount. Factors for this phenomenon can be
related to (a) restricted/blocked surface, leading to low pore
volume restricting mass transfer, (b) effect of low drug
solubility in medium, (c) formation of multiple dissolution
fronts, (d) roughness of surface, and (e) moment of dissolved
drug packets from within the isolated or interconnected pores
toward the dissolution medium. The ow and interaction of
the dissolution medium inside the porous material appears to
be critical in observing mass transfer. No signicant difference
in drug release was noted considering the same drug amount
at different solvent volumes, indicating similar amount of
drug amount ready for mass transfer.
Batch 6, with least release of 12.77% in acidic medium,
shows signicance in formulating the ideal system for
chronotherapy favoring low release during gastroretentive
period designated as lag period (Fig. 6b). The attainment of
such release without using any release modiers or subjecting
to coating by single or multiple polymers or formulating in
the form of simple or press coated tablets or in the form of
modied capsule consisting of various excipients makes it
exceptional and lucid to formulate.
Statistical interpretations for the drug release were
calculated at various time intervals, which was preferred
over percent release as presented in Table III. The values
conferred for the drug release in acidic medium show the
negative impact of drug coefcient () throughout the time
period. Initially, the negative solvent effect (), along with
positive solvent drug () interaction, was noted, which
vanish afterwards probably due to the behavior/dissolution
of the outermost drug layer. The negative values of drug
drug (2) interaction up to half way indicate the dissolution
of underlying drug layers from the surface inuenced by
adsorption pattern. These values suggest pronounced
impact of drug coefcient at all levels irrespective of
solvent volume.
555
the end of experiment shows the contributions from most
coefcients conferring the positive inuences of the
selected variables, which worked ambiguously in acidic
medium. Inference from this illustrates the inuence of
dissolution medium favoring difference in drug solubility as
in Fig. 7b. The positive values of drug coefcient suggest
its effect in basic medium only. These types of inuences
were similar to our previous study where only basic
medium was used (12).
Fig. 5. Response surface plots showing effect of factorial variables on

a surface area and b pore volume using DCM
Changing to phosphate buffer pH 7.2 IP initiated the

sudden increase in release, with maximum burst release of
68.95% for batch 3. At the end of the experiment, batch 3
showed maximum release, while cut off was observed in some
batches at 8 h mark. Sudden depletion of adsorbed drug
clusters favors dominance of non contributing porosity,
thereby causing new boundary conditions. This leads to the
exposure of underlying rough and hydrophobic nature of
surface and pores to dissolution medium, which in turn assist
in retarding/stopping the drug release. In other batches,
slower drug release was observed after burst release was
inuenced by the continuity of adsorbed drug layer and its
mass transfer momentum from inside to outside (29). Moving
boundary conditions inuencing geometry, described as
Stephan moving problem (31,32), was perfectly observed in
drug release while exposing porous carriers to basic medium
alone than using acidic medium before it.
Statistically, drug as well as solvent solvent ( 2 )
interaction coefcient showed a positive inuence on burst
release fraction (Fig. 7a), while a positive inuence of drug
coefcient () was observed at 7 h. The same observation at
Fig. 6. Cumulative drug release prole from Accurel MP 1000

using methanol
556
Sher et al.
The release in acidic medium showed a characteristic
trend recording the least release of batches using medium
level of drug (200 mg) irrespective of solvent volume used.
Batch 8 recorded the least release of 8%, as in Fig. 8c. This
release is much lesser than any batch even from the methanol
batches. The least/less drug release can be related to the ever
decreasing pore volume caused by adsorption of drug
Fig. 7. Response surface plots showing the effect of factorial

variables on a burst release b drug release at 9 h using methanol
Batch 3 can be considered as optimized formulation

for the purpose of chronotherapy, with 14.79% release in
acidic medium followed by burst of 68.95%. In addition,
batch 6 can be preferred as the subsequent choice with
12.77% release in acidic medium followed by 58.05% burst
release.
From DCM Adsorbed Microparticles
Overall, the drug release ranging 8 15% was much lower
unlike corresponding methanol batches in acidic medium. Burst
effect ranges 31 70%, and the nal release range 55 99% was
observed with a positive difference in some batches, as seen in
Fig. 8. Factors governing the release pattern were more or less
the same, probably differing in magnitude.
Fig. 8. Cumulative drug release prole from Accurel MP 1000

using DCM

inuenced by solvent property. This again supports the
adsorption pattern of drug at meniscus of porous material
using less polar solvent (24). The inuence of the different
solvent volume was not evident in acidic medium. Statistical
interpretation show negative inuence of drug () but a
positive inuence of drug drug interaction (2), as in Fig. 9a.
This can be due to the development of various adsorption
patterns in the form of stakes, which inuence inter pore
connective inuencing pore volume.
Burst release fraction, on changing to basic medium,
observed increased release with increasing drug amount. It
was twice or more for all solvent volumes used. The main
feature, as shown by DCM, was the high burst release
fraction for batch 3 at 70%. The less burst release in batch
1(lowest drug amount) can be due to fast exposure of
underlying hydrophobic pores and surface, thereby retarding
the drug release. These effects seem to be nullied with
increasing amount of drug used for adsorption, where the
adsorbed drug clusters act as the reservoirs for increased drug
release. This difference can be summed up for the interaction
557
of the nature of solvents and mediums affecting drug
adsorption leading to the different rate mass transfer. Positive
drug coefcient inuence was observed statistically, which did
not change when clubbed with total drug release for 7 h.
At the end of the experiment, batch 3 showed maximum
release the same as that of methanol. Comparing with
methanol batches, better drug release was observed using
maximum drug amount, while a lesser drug release was
calculated for other drug amounts. This can be due to
variable mass transfer caused by boundary conditions of the
adsorbed drug layer.
Statistically, at the end of the experiment, positive
contributions from drug variables were prominent along with
negative effect of solvent characteristic of solvent nature, as
in Fig. 9b.
Overall, again, batch 3 can be considered as the most
optimized formulation for effective oating pulsatile drug
delivery with 13.48% release in acidic medium followed by
burst of 70.91%. In addition, batch 9 can be preferred as the
subsequent choice with 11.34% drug release in acidic medium
and 66.95% as burst release.
Stability studies predict no signicant changes in the drug
release. Apart from this, DSC and XRD show the same
results with marginal difference.
CONCLUSION
A lucid and simple oating pulsatile drug delivery system
intended for chronotherapy was achieved using 32 factorial
design for optimizing the formulation. Single step adsorption
techniques, unaided by additives, are promising enough for
further evaluation of present and various other drug delivery
systems. Using this process, four different batches with
desired release parameters can be selected. BET and
mercury data seem to behave differently in accordance with
the principle of instrument working. Cross sections predicted
the inner core geometry of adsorbed porous carrier giving an
insight about the possible drug release phenomenon. The
modied geometry and network structure of pores in Accurel
MP 1000 due to inuence of selected variables govern the
desired drug release pattern. The dissolution medium
showing varying drug solubility shows multiple variations in
pore volume and pore surface, therefore affecting porosity and
permeability. Variable effect was dormant in acidic medium.
This inuenced the drug release rate process, which in turn is
also dependent on static and changing adsorbing pattern of
drugs. This work can be extended for time scheduled drug
release of drugs having low solubility, poor absorption, or
degradation in lower gastrointestinal tract.
ACKNOWLEDGMENTS
Fig. 9. Response surface plots showing effect of factorial variables on

a drug release at 6 h, b drug release at 9 h using DCM
Praveen Sher is thankful to Dr. James R. Benson of

Polygenetics, Inc. for his valuable advice during work.
Praveen Sher and Atmaram Pawar are also thankful to
UGC, India for providing major research project. Authors
are thankful to Dr. Smita Mule (DSC), A.B. Gaikward
(SEM), Baishakhi (AFM), and Supriya Palimkar for help in
evaluation work. Authors also thank Membrana, Germany
for providing a free gift sample of Accurel MP 1000,
particularly Ms. Claudia Gramann for her support.
558
REFERENCES
1. Andersson J, Rosenholm J, Areva S, Linden M. Inuences of
material characteristics on ibuprofen drug loading and release
proles from ordered micro and mesoporous silica matrices.
Chem Mater 2004;16:4160 7.
2. Shivanand P, Sprockel OL. A controlled porosity drug delivery
system. Int J Pharm 1998;167:83 96.
3. Li Z, Wen L, Shao L, Chen J. Fabrication of porous hollow silica
nanoparticles and their applications in drug release control. J
4. Song S W, Hidajat K, Kawi S. Functionalized SAB 15 material as
carrier fro controlled drug delivery: inuence of surface proper
ties on matrix drug interactions. Langmuir 2005;21:9568 75.
5. Salonen J, Laitinen L, Kaukonen AM, Tuura J, Bjorkqvist M,
Heikkila T, Vaha Heikella K, Hirvonen J, Lehto V P. Mesopo
rous silicon microparticles for oral drug delivery: loading and
release of ve model drugs. J Control Release 2005;108:362 74.
6. Ohta KM, Fuji M, Takei T, Chikazawa M. Development of a
simple method for the preparation of a silica gel based controlled
drug delivery system with a high drug content. Eur J Pharm Sci
2005;26:87 96.
7. Otsuka M, Tokumitsu K, Matsuda Y. Solid dosage form
preparations from oily medicines and their drug release. Effect
of degree of surface modication of silica gel on the drug release
from phytonadione loaded silica gels. J Control Release
2000;67:369 84.
8. Streubel A, Sipeman J, Bodmeier R. Floating microparticles
based on low density foam powder. Int J Pharm 2002;241:279 92.
9. Streubel A, Sipemann J, Bodmeier R. Multiple unit gastro
retentive drug delivery systems: a new method for low density
microparticles. J Microencapsul 2003;20(3):329 47.
10. Li Y h, Zhu J b. Modulation of combined release behaviors from
a novel tablets in scapsule system. J Control Release 2004;95
(3):381 9.
11. Gren T, Bjerre C, Camber O, Ragnarsson G. In vitro drug release
from porous cellulose matricies. Int J Pharm 1996;141:53 62.
12. Sher P, Ingavle G, Ponrathnam S, Pawar AP. Low density porous
carrier: drug adsorption and release study by response surface
methodology using different solvents. Int J Pharm 2007;331:72 83.
13. Sher P, Ingavle G, Ponrathnam S, Pawar AP. Low density porous
material based conceptual drug delivery system. Microporous
Mesoporous Mater 2007;102:290 8.
14. Stubbe BG, Smedt S, Demeester CD. Programmed polymeric
devices for pulsed drug delivery. Pharm Res 2004;21:10.
15. Badve SS, Sher P, Korde A, Pawar AP. Development of hollow/
porous calcium pectinate beads for oating pulsatile drug
delivery. Eur J Pharm Biopharm 2007;65:85.
16. Lee B J, Min G H. Oral controlled release of melatonin using
polymer reinforced and coated alginate beads. Int J Pharm
1996;144(1):37 46.
Sher et al.
17. Lee B J, Ryu S G, Cui J H. Controlled release of dual drug
loaded hydroxypropyl methylcellulose matrix tablet using
drug containing polymeric coatings. Int J Pharm 1999;188
(1):71 80.
18. Bussemer T, Bodmeier R. Formulation parameters affecting the
performance of coated gelatin capsules with pulsatile release
proles. Int J Pharm 2003;267(1 2):59 68.
19. Dashevsky A, Mohamad A. Development of pulsatile multi
particulate drug delivery system coated with aqueous dispersion
Aquacoat ECD. Int J Pharm 2006;318(1 2):124 31.
20. Efentakis M, Koligliati S, Vlachou M. Design and evaluation of a
dry coated drug delivery system with an impermeable cup,
swellable top layer and pulsatile release. Int J Pharm 2006;311
(1 2):147 56.
21. Ravishankar H, Patil P, Samel A, Petereit H U, Lizio R, Iyer
Chavan J. Modulated release metoprolol succinate formulation
based on ionic interactions: in vivo proof of concept. J Control
Release 2006;111(1 2):65 72.
22. Gunko VM, Mikhalovskii SV, Melillo M, Voronin EF, Nosach
LV, Pakhlov EM. The effect of the nature and structure of
adsorbents on interaction with ibuprofen. Theor Exp Chem
2004;40(3):137 43.
23. Janczuk B, Chibowski E, Wojcik W. The inuence of n alcohols
on the wettability of hydrophobic solids. Powder Technol
1985;45:1 6.
24. Sarkisov L, Monson PA. Modeling of adsorption and desorption
in pores of simple geometry using molecular dynamics. Langmuir
2001;17:7600 4.
25. Scott DC. An assessment of reasonable tortuosity values. Pharm
Res 2001;18(12):1797.
26. Germann PF, DiPietro L. When is porous media ow preferen
tial? A hydromechanical perspective. Geoderma 1996;74:l 21.
27. Faruk C. Scale effect on porosity and permeability: kinetics,
model and correlation. AIChE J 2001;47(2):271 87.
28. Tongta A, Liapis AI, Siehr DJ. Equilibrium and kinetic
parameters of the adsorption of a chymotrypsinogen A onto
hydrophobic porous adsorbent particles. J Chromatogr A
1994;686:21 9.
29. Siegel RA, Kost J, Langer R. Mechanist studies of macromolec
ular drug release from macroporous polymers. I experiments and
preliminary theory concerning completeness of drug release. J
30. Iannuccelli V, Coppi G, Sansone R, Ferolla G. Air compartment
multiple unit system for prolonged gastric residence. Part II. In
vivo evaluation. Int J Pharm 1998;174:55 62.
31. Collins R. Mathematical modeling of controlled release from
implanted drug impregnated monoliths. Pharm Sci Technol
Today 1998;1(6):269.
32. Collins R, Jinuntuya N, Petpirom P, Wasuwanich S. Mathe
matical model for controlled diffusional release of dispersed
solute drugs from monolithic implants. Ann N Y Acad Sci
1998;858:116.

DOI: 10.1208/s12249 009 9242 1
Research Article
Development of Novel Microemulsion-Based Topical Formulations of Acyclovir
for the Treatment of Cutaneous Herpetic Infections
Shishu,1,2 Sunita Rajan,1 and Kamalpreet1
Received 7 November 2008; accepted 9 April 2009; published online 9 May 2009
Abstract. The present investigation aims at developing microemulsion based formulations for topical
delivery of acyclovir. Various microemulsions were developed using isopropyl myristate/Captex 355/
Labrafac as an oil phase, Tween 20 as surfactant, Span 20 as cosurfactant, and water/dimethylsulfoxide
(1:3) as an aqueous phase. Transcutol, eucalyptus oil, and peppermint oil were used as permeation
enhancers. In vitro permeation studies through laca mice skin were performed using Franz diffusion cells.
The optimum formulation containing 2.5% Transcutol as the penetration enhancer showed 1.7 fold
enhancement in ux and permeation coefcient as compared to marketed cream and ointment
formulation. In vivo antiviral studies were performed in female Balb/c mice against induced herpes
simplex virus I infection. A single application of microemulsion formulation containing 2.5% Transcutol
given 24 h post injection resulted in complete suppression of development of herpetic skin lesions.
KEY WORDS: acyclovir; antiviral; herpes simplex virus; microemulsion; topical.
INTRODUCTION
Acyclovir (ACV), a guanine analogue, is a rst line
antiviral drug for the treatment of infections caused by the
herpes viruses, including herpes simplex 1 and 2 (cold sores
and genital herpes), varicella zoster (shingles and chicken
pox), and the Epstein Barr virus (mononucleosis). Recurrent
herpes labialis and herpes genitalis represent the most
common clinical manifestations associated with herpes sim
plex virus type 1 (HSV 1) and type 2 (HSV 2). Topical
formulations for the treatment of herpetic mucocutaneous
infections have several potential benets over oral and
intravenous administration, including targeting of drug to
the specic sites of infection, higher tissue drug levels,
reduced side effects, lower treatment costs, and better patient
compliance and convenience. The low efcacy of currently
available topical ACV cream formulation may be attributed
to poor penetration of ACV through the stratum corneum
and thus lack of sufcient drug at the basal epidermis target
site requiring ve times a day application (1 3).
Our aim is to develop a therapeutically effective topical
delivery system, which will enable sufcient penetration of
the active drug to the inner skin layers with no signicant
systemic absorption. Improved penetration allows formation
of a reservoir of active drug within the skin (4), thus
maintaining persistent high levels of ACV at the infection
site. Among the many options available in novel colloidal
carriers, microemulsions offer advantage of enhanced drug
solubility, good thermodynamic stability, and enhanced
1
University Institute of Pharmaceutical Sciences, Panjab University,

Chandigarh, 160014, India.
2
To whom correspondence should be addressed. (e mail: shishugoindi@
yahoo.co.in)
effects on skin penetration ability over conventional formu

lations (5). Microemulsions have been visualized as superior
drug carriers because of its several permeation enhancement
mechanisms such as an increased concentration gradient and
thermodynamic activity toward skin and the permeation
enhancement activity of components of microemulsions (6).
In this study, we have developed microemulsion based
topical formulations of ACV and evaluated their efcacy by
in vitro skin permeation studies through mice skin, and nally,
the optimized formulation was evaluated for its antiviral
activity in murine model of cutaneous HSV I infection.
MATERIALS
ACV was obtained as gift sample from Arochem
Industries, Mumbai, India and Captex 355 EP/NF (glycerol
caprylate caprate) was received as a gift from Abitec
Corporation, UK. Labrafac CC (caprylic/capric triglycerides)
and Transcutol (ethoxy diglycol) were gifted by Colorcon
Asia Pvt. Ltd., India. Eucalyptus and peppermint oil were
purchased from Ajanta Chemicals Co., Gurgaon, India. ACV
Cream B.P. (Herpex, 5%) manufactured by Torrent Phar
maceuticals Ltd., India was used as control in the studies. All
other reagents/chemicals used were of AR grade.
METHODS
Solubility Studies. Solubility studies were carried out in
water, water/dimethyl sulfoxide DMSO (1:3), phosphate
buffer pH 6.4, and water/propylene glycol (PG; 1:1) using
the standard method of solubility determination (7). An
559
560
Shishu, Rajan, Kamalpreet
excess of drug was added to a xed volume of each solvent

(10 mL) in different asks, which were kept at 371C in a
thermostatic water shaker bath for 24 h. The asks were
removed after 24 h, contents were ltered through 0.22
lter, and ltrates were analyzed spectrophotometrically at
max 254 nm after appropriate dilution.
Construction of Pseudoternary Phase Diagrams. To in
vestigate the microemulsion region, pseudoternary phase
diagrams were constructed by titration of varied concentra
tions of oil [isopropyl myristate (IPM)/Captex 355/Labrafac
CC], water/DMSO (1:3), Tween 20, and Span 20 with a xed
ratio of Tween 20 and Span 20 (3:1). For each point, the
required quantities of Tween 20, Span 20, and oil were gently
mixed to form a monophasic mixture that was slowly titrated
with aliquots of water/DMSO and stirred at 25C for a
sufcient time to attain equilibrium. After equilibrium was
reached, the mixtures were checked visually for transparency.
Clear and isotropic water rich samples of low viscosity were
deemed to be within the oil in water microemulsion region.
Again Tween 20, Span 20, and water/DMSO mixture was
mixed and titrated with oil and checked for transparency.
Clear region was water in oil microemulsion region.
Preparation of Microemulsion Formulations. After iden
tication of the microemulsion regions in the phase diagram,
the microemulsion formulations were prepared (Table I).
ACV (0.3% w/w) was incorporated into the aqueous phase
(water/DMSO, 1:3) followed by the addition of Tween 20.
Span 20 was dissolved in the oil followed by the addition of
penetration enhancer, and then, the aqueous phase was
added slowly to the oily phase under continuous stirring
using magnetic stirrer at ambient temperature. Homogenous
and stable microemulsions were formed spontaneously.
Preparation of PEG Ointment. For comparative studies,
ACV (5% w/w) was formulated in a conventional ointment
system containing PEG 400 (37.5% w/w) and PEG 4000
(57.5% w/w). It was taken as control II, and ACV Cream B.P.
(Herpex, 5%) was taken as control I.
Characterization of Microemulsions. Drug content was
determined using method given in British Pharmacopoeia (8).
pH of the prepared formulation was determined using pH
meter. Morphology and structure of the microstructures of
drug loaded microemulsion was determined with the aid of
transmission electron microscopy. Droplet size and size
distribution of prepared microemulsion were analyzed using
laser light scattering technique (Malvern Zetasizer). The

developed formulations were studied for their rheological
behavior using Brookeld digital viscometer equipped with
small adaptor and spindle number 21. The microemulsions
were submitted to up and down cycles (0 100 rpm spindle
rotation speed) at 291C, and the rheological behavior of
each disperse system was evaluated by plotting shear rate
against shear stress.
Stability Studies. Three batches of each formulation were
stored in sealed glass containers at 4C, 25C, and 40C for
45 days. The stability of the microemulsions was checked
periodically with respect to transparency, phase separation,
color change, and drug content of ACV at three different
temperature conditions.
In vitro Skin Permeation Study. In vitro permeation
studies of microemulsion formulations of ACV were per
formed through skin of Laca mice using Franz diffusion cell
assembly. The animals were killed by spinal dislocation
method. The hair on the dorsal side of the animal was removed
with the help of surgical blade number 23 in the direction of tail
to head. The shaven part of the animal skin was separated.
Hypodermis, including the blood vessels, was removed using
surgical blade number 23. The excised skin was washed with
normal saline and subsequently used. The skin was mounted
on the diffusion cell assembly, with an effective diffusion area
of 2.84 cm2, keeping the stratum corneum toward donor
compartment. The receptor compartment consisted of 30 mL
phosphate buffer, pH 6.4, maintained at 37C. The prepared
formulations containing an amount equivalent to 3 mg of ACV
were applied on the membrane in the donor compartment. An
aliquot of 1 mL sample was withdrawn at suitable time
intervals and replaced immediately with an equal volume sof
fresh diffusion medium. The samples were analyzed
spectrophotometrically at 254 nm after suitable dilutions. All
experiments were performed in triplicate. The cumulative
amount of ACV permeated through mice skin was plotted as a
function of time.
Skin Retention Study. At the end of the permeation
experiments (after 24 h), the remaining formulation in the
donor phase was scrapped off the skin, and the exposed skin
surface was rinsed with water/DMSO (1:3) to remove excess
drug from the surface. The receptor media was then replaced
with fresh water/DMSO (1:3). Receptor contents were
allowed to stir for the next 24 h. After 24 h, the media was
analyzed for the amount of drug retained in skin (9).
Table I. Composition of Various Microemulsion Formulations of ACV

Ingredients (quantities, % w/w)
ME
E1
E2
E3
T1
T2
T3
P1
P2
P3
Isopropyl Myristate
Tween 20
Span 20
Water: Dimethyl sulfoxide (1:3)
Eucalyptol
Transcutol
Peppermint
38
30
10
22
36.63
32.08
8.49
21.8
1.0
35.13
32.08
8.49
21.8
2.5
32.63
32.08
8.49
21.8
5.0
37.61
27.43
11.6
22.36
36.11
27.43
11.6
22.36
33.6
27.43
11.6
22.36
36.63
32.08
8.49
21.8
35.13
32.08
8.49
21.8
32.63
32.08
8.49
21.8
1.0
2.5
5.0
1.0
2.5
5.0
Novel Microemulsion-Based Topical Formulations of ACV

Statistical analysis. All the data were statistically ana
lyzed by one way analysis of variance followed by multiple
comparison Tukey test. P value of 0.05 was considered to be
signicant.
Histopathological Examination. Laca mice weighing
around 20 25 g were used for the study. The hair on the
dorsal side of animals was removed with surgical blade
number 23 in the direction of tail to head without damaging
the skin. Selected formulation T 2 was applied uniformly on
the dorsal region. The microemulsion was kept in contact
with the skin for 4 h. After that, the animals were killed with
overdose inhalation of chloroform, and exposed dorsal
surface was cut. Then, the specimens were xed in 10%
buffered formalin, embedded in parafn, and microtoned.
The sections were stained with hematoxylin and eosin.
Finally, the specimens were observed under a high power
light microscope.
Skin Sensitivity Testing. Selected formulations were applied
on the forearm skin of 20 healthy human volunteers (age 20
25 years) and observed for any irritation, erythema, skin rash,
and edema for 24 h. A prior written informed consent was
obtained from all the volunteers, and they were apprised of any
possible ill effects of the study.
In vivo Pharmacodynamics Study. Murine model of
cutaneous HSV I infection was used to study the antiviral
effect of ACV microemulsion. HSV I propagated in Vero cells
(African green monkey kidney cells) in minimum essential
medium and female BALB/c mice (5 7 weeks old) were used
in this study (10). The plaque assay was performed to check the
viral infectivity (11). Animals were anesthetized by intraper
itoneal injection of a mixture containing 70 mg of ketamine
hydrochloride and 11.5 mg of xylazine per kilogram of body
weight. Mice were then cutaneously infected with HSV 1 after
scarication (scratching the skin six times in a crossed hatched
pattern) of the shaved right mid ank region with a 27 gauze
needle held vertically (10,12). Treatment was initiated 24 h
post infection (i.e. before the appearance of zostiform rash).
Mice in three groups of four each were treated topically ve
times daily with 15 mg of marketed cream (control I), 15 mg of
PEG ointment (control II), and once daily with 250 l micro
emulsion formulation (T 2) containing 0.75 mg ACV. The
fourth untreated group served as control. The efcacy of
different formulations was evaluated in terms of evolution of
lesions and reduction in number of lesions.
561
Table II. Solubility of ACV in Various Solvents
Medium
Phosphate buffer, pH 6.4
Water
Water/propylene glycol (1:1)
Water/dimethyl sulfoxide (1:3)
Solubility (mg/mL)
1.40.2
2.50.1
3.50.2
13.70.3
buffer, pH 6.4, the drug was found to be sufciently soluble

for being used as a receptor medium and for maintaining sink
conditions during in vitro permeation studies.
Phase Studies. Phase studies were carried out to nd the
area of microemulsion existence and to investigate the effect
of different surfactant/cosurfactant weight ratios on the extent
of stable microemulsion region. The psuedoternary phase
diagrams with various oils (isopropyl myristate, Captex, and
Labrafac) are depicted in Fig. 1a c. The weight ratio of 3:1 of
Tween 20 and Span 20 as surfactant/cosurfactant mixture
resulted in most stable microemulsion. Water/DMSO (1:3)
was used as the aqueous phase. The largest area of micro
emulsion existence was observed in phase diagram employing
isopropyl myristate as an oil phase. Hence, IPM (Fig. 1b) was
selected as oil phase.
Characterization of Microemulsion. Results of our study
indicate development of successful microemulsion formulations
of ACV with optimum characteristics. The drug content of
different formulations was found to be in the range of 980.15 to
99.890.52%. The pH was found to vary between 6.32 and 6.72,
a near neutral range ideal for topical applications. The
formulations were clear and transparent. The transmission
electron microscopic picture revealed spherical nature of the
microemulsion droplets. The mean diameter was found to be
400 nm. The zeta potential of the prepared microemulsion was
found to be 22 mV. Rheological studies showed Newtonian
nature of the microemulsions. The viscosities of different
formulations were found to be in the range of 182.4 372.2 mPa
s at 29C, suggesting suitability for topical application.
Stability Studies. Stability studies data at three different

temperatures showed that all the selected microemulsion
formulations were stable for drug content and other macro
scopic characteristics like phase separation and color change.
All the batches of microemulsion remained clear even after a
period of 45 days at temperature 25C and 40C, but turbidity
occurred in the formulations stored at temperature 4C.
However, when brought to room temperature, all the
formulations returned to their initial states.
Solubility Studies. Solubility studies were carried out in

different media to assess the solubility behavior so as to select
the appropriate medium for in vitro permeation studies and
optimal aqueous medium for developing water in oil micro
emulsion formulations of hydrophilic ACV. In water/DMSO
(1:3), solubility of ACV was found to be highest (13.7
0.3 mg/mL), followed by water/PG (1:1), and water and
phosphate buffer pH 6.4 (Table II). Accordingly, water/
DMSO (1:3) was chosen as the most appropriate aqueous
phase for the development of microemulsion. In phosphate
In vitro Skin Permeation Studies. The permeation proles

of ACV through mice skin from various microemulsions are
shown in Fig. 2a c. As expected, the mean cumulative amount
permeated per unit area of ME in 24 h (Fig. 2a) was found to be
798.486.07 g/cm2, which was signicantly greater (P<0.05)
than control I, the marketed cream (763.182.25 g/cm2), and
control II, the conventional PEG ointment (751.123.1 g/cm2).
The lesser release from ointment can be explained on the basis
of the drug vehicle interaction that results in a lower
thermodynamic activity of the drug (13,14). Other researchers
562
Fig. 1. Pseudoternary phase diagram of the system containing Tween 20, Span 20, aqueous phase and a IPM, b Captex, and c Labrafac
563
Fig. 2. Comparison of mean cumulative amount of ACV released per unit area of mice skin through a formulation ME, control I, and control
II, b formulations ME, E 1, E 2, and E 3, c formulations ME, P 1, P 2, and P 3, d formulations ME, T 1, T 2, and T 3
Fig. 3. Comparison of ux of ACV from different formulations. Mean Flux values for all
the microemulsions were signicantly different from control II at P<0.05. Mean ux values
for all the microemulsions except ME were signicantly different from control I at P<0.05
564
Fig. 4. Comparison of skin retention of ACV from different formulations. Dagger Mean
value is not signicantly different from preceding mean value at P<0.05
have suggested that the retardant effect of PEG is due to its

inability to hydrate the stratum corneum or to a relative osmotic
effect, which tends to dehydrate the stratum corneum (15).
Microemulsions and creams show better drug release because of
the penetration enhancing properties of their components. The
steady state ux (Fig. 3) followed the same suite with ME
having the maximum ux followed by marketed cream (control
I) and PEG ointment (control II).
In order to further improve the permeation rate of ACV
from the microemulsion, various enhancers like eucalyptus
oil, peppermint oil, and Transcutol in the concentration
ranging from 1% to 5% were employed, and their effects
were determined. Results indicate signicant improvement in
the permeation pattern of ACV with the incorporation of
enhancers (Fig. 2b d).
Upon increase in the concentration of eucalyptus oil from
1.0% (E 1) to 5.0% (E 3), there was a signicant improvement
in the permeation characteristics of the drug (804.574.53 to
869.292.29 g/cm2 amount after 24 h; Fig. 2b). A 1.46 fold
enhancement was seen in the ux of formulation E 3
containing 5% eucalyptol. The enhancement achieved with
eucalyptus oil can be the result of combined process of
partition and diffusion and the latter being dominant (16).
Peppermint oil also proved to be a good permeation
enhancer. Formulation P 2 containing 2.5% peppermint oil
showed highest mean cumulative amount permeated per unit
area in 24 h among formulations P 1, P 2, and P 3. A 6.5%
raise was seen in amount permeated in P 2 when compared
with ME. No further signicant improvement was seen in the
permeation on increasing the amount of peppermint oil from
2.5% to 5% (Fig. 2c). Same trend was observed in the ux with
P 2 formulation having an enhancement ratio of 1.52 (Fig. 3).
Peppermint oil containing high proportion of l menthol has
been reported to alter skin permeation by a dual mechanism of
forming a eutectic mixture with the penetrating compound,
thereby increasing its solubility and by altering the barrier
properties of the stratum corneum (17).
The presence of Transcutol in the formulations also
resulted in an increase in the mean cumulative amount
permeated per unit area in 24 h, indicating that Transcutol

also acts as good penetration enhancer for ACV (Fig. 2d).
Formulation T 2 containing 2.5% Transcutol resulted in
maximum enhancement ratio (1.70) as well as the maximum
ux (113.580.85 g cm 2 h 1; Fig. 3). The enhancing ability
of Transcutol has been attributed to its ability to pass through
the skin (18) and get incorporated into the multiple lipid
bilayers, thereby swelling the intercellular lipids (19).
Skin Retention Studies. The amount of ACV retained in
the skin was observed to be signicantly higher (at P<0.05) in
all the microemulsion formulations (with or without enhanc
er) as compared to retention observed with the conventional
cream (control I) or ointment (control II). These results
indicate that the microemulsion formulations penetrate into
the skin and form microdrug reservoirs that can hold the
sufcient drug for a longer period of time and may help in the
suppression of the viral growth more effectively. It was also
observed that, as the concentration of permeation enhancers
was increased, the amount permeated and also the ux was
increased, resulting in the reduction in skin retention of ACV
(Fig. 4).
Histopathological Studies. Dermal tolerance of micro
emulsion formulation T 2 (containing Transcutol as enhancer)
was compared with untreated mouse skin. No anatomical or
pathological changes were observed.
Skin Sensitivity Studies. The results of skin sensitivity
testing using open patch test showed absence of any redness,
irritation, and swelling, indicating the non sensitizing charac
teristics of the tested microemulsion formulations.
In vivo Pharmacodynamic Studies. Exposure to HSV at
mucosal surfaces or abraded skin sites permits the entry of
virus and initiation of its replication in cells of epidermis and
dermis (20). The pharmacodynamic evidence of the prepared
ACV microemulsion formulations was carried out by employ
ing the murine model of cutaneous HSV 1 infection. The

efcacy of the optimized formulation containing 2.5% Trans
cutol as enhancer (T 2) was tested after single application
given 24 h post infection in mice. In the untreated infected
animals, no pathological signs of cutaneous infection were
visible during the rst 5 days following the infection. This can
be explained as when the virus is inoculated in the skin,
where the primary infection occurs; it starts spreading from
this site, probably by retrograde axonal ow, to sensory
ganglia and the central nervous system. Thereafter, the virus
reaches axons that innervate skin within the same dermatome
as the inoculation site, spreads via orthograde ow into the
skin, and produces herpetic lesions within the affected
dermatome (10). On the sixth day, herpetic skin lesions
appeared in the form of small vesicles distant from inocula
tion site. On the seventh day, ulcerations developed in the
infected mice in local region. However, infected mice treated
with all the three drug formulations (T 2, control I, and
control II) developed no lesions, thus suggesting that the
prepared microemulsion containing 2.5% Transcutol (T 2),
which was applied once daily, was as effective as conventional
marketed cream (control I) and PEG ointment (control II)
that were applied ve times daily, in the treatment of herpetic
cutaneous lesions.
CONCLUSION
Results of in vitro permeation studies through mice skin
suggest that, using microemulsion based topical delivery
system, ACV penetrates the skin well, retains in the skin,
and may have efcacy superior to that of topical ointment
and cream. Furthermore, the results of in vivo investigations
using murine model of cutaneous HSV 1 conrms the efcacy
of this formulation against experimental HSV I infection.
These encouraging outcomes of preliminary investigations
strongly warrant further clinical investigations to examine the
efcacy of these formulations in the treatment of human
mucocutaneous herpes simplex virus disease.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the gift sample of
ACV supplied by Arochem Industries (Mumbai, India),
Captex 355 EP/NF (Abitec Corporation, UK), and Labrafac
CC and Transcutol (Colorcon Asia Pvt. Ltd., India). Authors
also thank Prof. Radha Kanta Ratho (Head) and Assistant
Prof. Mini P. Singh, Department of Virology, Postgraduate
Institute of Medical Education and Research, Chandigarh,
India for their help and guidance and for providing laboratory
facilities for antiviral studies.
565
REFERENCES
1. Fiddian P, Yeo JM, Clark AE. Treatment of herpes labialis. J
Infect. 1983;6:41 7.
2. Freeman DJ, Sheth NV, Spruance SL. Failure of topical acyclovir
in ointment to penetrate human skin. Antimicrob Agents
Chemother. 1986;29:730 2.
3. Parry GE, Dunn P, Shah VP, Pershing LK. Acyclovir bioavail
ability in human skin. J Invest Dermatol. 1992;98:856 63.
4. Chikhale P, Bodor N. Improved delivery of acyclovir to the skin
using a dihydrotrigonelline trigonelline REDOX carrier. J
Pharm Sci. 1991;80:801 2.
drug delivery systems. Adv Drug Del Rev. 2000;45:89 121.
6. Peltola S, Savolainen P, Keisvaara J, Suhonen TM, Urtti A.
Microemulsions for topical delivery of estradiol. Int J Pharm.
2003;254:99 107.
7. Aulton ME. Dissolution and solubility. In: Aulton ME, editor.
Pharmaceutics. Edinburgh, UK: Churchill Livingstone; 2002.
p. 15 32.
8. British Pharmacopoeia, Vol. II. The stationary Ofce, London;
1999, p. A 87.
9. Jain S, Jain P, Jain NK, Umamaheshwari RB. Transfersomes a
novel vesicular carrier for enhanced transdermal delivery:
development, characterization, and performance evaluation.
10. Piret J, Desormeaux A, Gourde AP, Juhasz J, Bergeron MG.
Efcacies of topical formulations of foscarnet and acyclovir and
of 5 percent acyclovir ointment (Zovirax) in a murine model of
cutaneous herpes simplex virus type 1 infection. Antimicrob
Agents Chemother. 2000;44:30 8.
11. Schimdt NJ. Cell culture techniques for diagnostics virology. In:
Lennette EH, Schimdt NJ, editors. Diagnostic procedure for
viral, rickettisial and chlamydial infections. Washington: Ameri
can Public Health Association; 1979. p. 100 4.
12. Kurokawa M, Basnet P, Ohsugi M, Hozumi T, Kadota S, Namba
T, Kawana T, Shiraki K. Anti herpes simplex virus activity of
moronic acid puried from Rhus javanica in vitro and in vivo. J
Pharmacol Exp Ther. 1999;289:72 8.
13. Davis SS, Hadgraft J, Al Khamis K. Percutaneous absorption of
methyl salicylate from polyethylene glycol vehicles. J Pharm
Pharmacol. 1981;33:97P.
14. Hadgraft J. Percutaneous absorption: Possibilities and problems.
Int J Pharm. 1983;16:255 70.
15. Barret W, Hadgraft JW, Sarnaky I. The inuence of vehicles on
skin penetration. J Pharm Pharmacol. 1964;16:104T 107T.
16. Abdullah, Ping QN, Liu GH. Enhancing effect of essential oils
on the penetration of 5 uorouracil through rat skin. Yao Xue
Xue Bao 1996;31:214 21.
17. Kaplun Frischoff Y, Touitou E. Testosterone skin permeation
enhancement by menthol through formation of eutectic with
drug and interaction with skin lipids. J Pharm Sci. 1997;86:1394
9.
18. Ganem Quintanar, Lafforgue AC, Falson Rieg F, Buri P. Eval
uation of the transepidermal permeation of diethyleneglycol
monoethylether and skin water loss. Int J Pharm. 1997;147:165
72.
19. Godwin A, Kim NH, Felton LA. Inuence of Transcutol CG on
the skin accumulation and transepidermal permeation of ultra
violet absorbers. Eur J Pharm Biopharm. 2002;53:23 7.
20. Wang F, Kief E. Viral diseases. In: Fauci AS, Braunwald E,
Isselbacher KJ, Wilson JD, Martin JB, Kasper DL, Hauser SL,
Longo DL, editors. Harrisons principles of internal medicine.
Singapore: McGraw Hill; 1998. p. 1076 85.

DOI: 10.1208/s12249 009 9247 9
Research Article
Sugar End-Capped Poly-D,L-lactides as Excipients in Oral Sustained Release
Tablets
Sirpa Vuorinen,1 Jyrki Heinmki,2,4 Osmo Antikainen,2 Mohammed Lahcini,3 Timo Repo,1 and Jouko Yliruusi2
Received 17 November 2008; accepted 23 April 2009; published online 9 May 2009
Abstract. Sugar end capped poly D,L lactide (SPDLA) polymers were investigated as a potential release
controlling excipient in oral sustained release matrix tablets. The SPDLA polymers were obtained by a
catalytic ring opening polymerization technique using methyl D gluco pyranoside as a multifunctional
initiator in the polymerization. Polymers of different molecular weights were synthesized by varying
molar ratios of monomer/catalyst. The matrix tablets were prepared by direct compression technique
from the binary mixtures of SPDLA and microcrystalline cellulose, and theophylline was used as a model
drug. The tablet matrices showed in vitro reproducible drug release proles with a zero order or
diffusion based kinetic depending on the SPDLA polymer grade used. Further release from the tablet
matrices was dependent on the molecular weight of the SPDLA polymer applied. The drug release was
the fastest with the lowest molecular weight SPDLA grade, and the drug release followed zero order
rate. With the higher molecular weight SPDLAs, more prolonged dissolution proles for the matrix
tablets (up to 8 10 h) were obtained. Furthermore, the prolonged drug release was independent of the
pH of the dissolution media. In conclusion, SPDLAs are a novel type of drug carrier polymers applicable
in oral controlled drug delivery systems.
KEY WORDS: direct compression; matrix tablets; poly D,L lactide; release kinetics; sustained release;
theophylline.
INTRODUCTION
Only very few new excipients aimed for controlling oral
drug release have come on the markets during the last
20 years, although there is an increasing need to be able to
modify accurately drug release prole and kinetics for various
types of drug substances. The main material groups used in
drug release control from oral matrices have been cellulose
derivatives, starches, polyethylene glycols, ionic exchangers,
and methacrylates (1 4). Particularly, there is a need for an
excipient group by which drug release rate can be modied in
a wide range.
Poly(D,L lactide) (PDLA) is an amorphous, biodegradable,
and aliphatic polyester polymer. Low molecular weight PDLAs
can be directly prepared from lactic acid, but the most efcient
way to produce high molecular weight polymers in a controlla
ble manner is by catalyst assisted, ring opening polymerization
Laboratory of Inorganic Chemistry, Department of Chemistry,

University of Helsinki, P.O. Box 55, A.I. Virtasen aukio 1, 00014,
Helsinki, Finland.
2
Division of Pharmaceutical Technology, Faculty of Pharmacy,
University of Helsinki, P.O. Box 56, Viikinkaari 5E, 00014, Helsinki,
Finland.
3
Laboratoire de Chimie Bio Organique et Macromolculaire,
Dpartement de Chimie, Fakult des Sciences et Techniques
Marrakech, Universit Cadi Ayyad, BP 549, Marrakech, Morocco.
4
To whom correspondence should be addressed. (e mail: jyrki.
heinamaki@helsinki.)
(ROP) of six membered cyclic diesters (5). Molecular weight

can greatly affect chemical and physical mechanical properties
such as degradation, solubility, viscosity, diffusivity, and glass
transition temperature (6).
Direct compression tablets based on PDLA and a drug
have delivered the drug according to proles suitable either
for implantable (7 10) or for oral dosage forms (11 14). In
literature, however, the results on suitability of low molecular
weight PDLAs for oral drug delivery tablets have been
contradictory. For example, low molecular weight PDLA
(MW 6,000) exhibited good compaction (15) and sustained
release controlling properties in matrix tablets (11). However,
since drug release from tablets containing low molecular
weight PDLA (MW 2,000) was highly pH dependent (12), it
was concluded that low molecular weight PDLA might not be
suitable for peroral application.
By using higher molecular weight PDLAs, limitations
caused by pH dependence can be avoided. One series of
studies with high molecular weight PDLA (Mv 85,000) as a
release controlling excipient in matrix tablets prepared by
direct compression have been performed by Steendam and
Lerk (13). However, a high molecular weight causes difcul
ties in mechanical grinding in order to get the polymer
granules into a powder form. Further, Steendam et al. (14)
reported that sustained drug release was only slightly affected
by molecular weight.
Surprisingly, only two reports about combining PDLA
and other tablet excipient for oral drug delivery application
have been published so far. Omelczuk and McGinity (16,17)
566
Sugar End-Capped Poly-D,L-lactides as Excipients in Oral Tablets
567
Fig. 1. Schematic presentation of Sn(NMe2)4 catalyzed ring opening polymerization of D,

L lactide in the presence of methyl D glucopyranoside. Polymerizations conducted in
toluene at 120C for 17 h (initial monomer concentration [M]0 1.0 M)
applied a wet granulation method for the mixing of PDLA

and microcrystalline cellulose (MCC) or lactose for tablet
preparation. The inuence of molecular weight as well as
thermal treatment on the drug release from PDLA/MCC or
PDLA/lactose sustained release matrix tablets was studied.
The release of the drug slowed down progressively as the
molecular weight increased (16).
The main aim of this study was to evaluate the properties
of sugar end capped poly (D,L lactides) (SPDLAs) as a novel
excipient group for pharmaceutical sustained release applica
tions. The strong hypothesis is that drug release rate can be
modied in a wide range by using SPDLAs with different
molecular weights and water solubilities. The sustained
release matrix tablets were prepared by direct compression
technique from the binary mixtures of SPDLAs and MCC.
Materials
Monomer, D,L lactide was purchased from PURAC
biochem (Purasorb DL) and was recrystallized from toluene.
Microcrystalline cellulose, MCC (Avicel PH 102, FMC
International, Cork, Ireland) was used as a direct compres
sion co ller for SPDLAs. Theophylline anhydrate (Ph.Eur.)
was chosen as a model drug because its solubility in the acidic
and neutral dissolution media is known to be similar (8).
Magnesium stearate (Ph.Eur.) was used as a lubricant.
Analytical Methods
Nuclear Magnetic Resonance Studies
1
H (300 Hz) nuclear magnetic resonance (NMR) spectra
were recorded on a Varian Mercury 300 spectrometer at 27C
in CDCl3. Chemical shifts are expressed as parts per million.
The 1H spectra are referenced relative to CHCl3 (7.27 ppm).
Gel Permeation Chromatography

Molecular weights and molecular weight distributions
were measured at 35C in tetrahydrofuran by size exclusion
gel permeation chromatography (GPC) relative to polysty
rene standards with a Waters 515 HPLC pump; GPC tted
with Styragel columns HR 1, HR 2, and HR 4; a UV detector
Waters 2487; and a refractive index detector Waters 2410.

The glass transition temperatures (Tg) of polymers were
determined with a Mettler 822e differential scanning calorim
etry (DSC) under nitrogen. The measuring temperature
range was from 30C to +100C. The samples of about
5 mg in simply sealed aluminum pans were rst heated from
+25C to +100C to destroy the thermal history and then
cooled to 30C and heated again to +100C. Heating rate
was 15C/min and cooling rate was 30C/min (holding time
was 5 min).
The Fourier transform infrared spectroscopy (FTIR)
spectra (Perkin Elmer Spectrum GX spetrometer) of the
polymers were obtained from potassium bromide disks
(200 mg, 3% polymer) compressed with hydraulic press.
The KBr tablet disk was placed into the IR sample holder and
measured with 64 scans from 4,000 to 370 cm 1 with 0.5 cm 1
nominal resolution.
Polymerization of SPDLAs and Preparation of SPDLA
Powder
In the polymerization of SPDLA polymers (Fig. 1), Sn
(NMe2)4 was applied as a catalyst precursor for the ROP of D,
L lactide and the reaction was initiated with hydroxyl groups
on methyl D glucopyranoside. Three SPDLA batches of
different molecular weights were obtained depending on
molar ratios of monomer/initiator ([M]/[I]) used for polymer
ization (Table I).
The obtained mixture of monomer and polymer was
analyzed by 1H NMR to determine conversion. The polymer
was puried from the monomer by precipitation from
Table I. Conversions Determined by 1H NMR and Molecular
Weights (Mw and Mn) Analyzed with GPC for Three Sugar End
Capped Poly D,L lactide (SPDLA) Polymer Batches
SPDLA batch
[Monomer]/
[initiator]
Conversion
(%)
Mw
(g/mol)
Mn
(g/mol)
SPDLA 5300
SPDLA 11000
SPDLA 17000
25/1
50/1
100/1
96
95
87
5,300
11,000
17,000
4,400
8,700
13,000
Vuorinen et al.
568
Fig. 2. 1H NMR spectrum of sugar end capped poly D,L lactide (SPDLA) polymer (Mw 5300) a before
and b after removal of residual monomer
dichloromethane/hexane solution (1:3) and analyzed with

GPC and DSC to determine molecular weight and glass
transition temperature. By evaporating solvents using a
vacuum pump, a polymer foam was obtained. This foam
was gently crushed in a mortar to grind it into a free owing
powder. For convenience, the polymers will be referred to in
the following discussion by the designation SPDLA Mw, i.e.,
an SPDLA 5300 polymer is a polymer with a molecular

weight of 5,300.
Preparation of Sustained Release Matrix Tablet
The composition of the powder mixture for direct
compression of the tablet matrices was as follows: 46.75%
Fig. 3. Fourier transform infrared spectrum of sugar end capped poly D,L lactide (SPDLA) polymer (Mw 5,300) in the region 3,300 to
500 cm 1

Table II. Peak Assignment for Sugar End Capped Poly D,L lactide
(SPDLA) Polymer (M w 5,300) Fourier Transform Infrared
Spectrum
Assignment
Peak center (cm 1)
CH stretch
C O carbonyl stretch
CH3 bend
CH deformation
C O bend
C O stretch
OH bend
2,997.8; 2,947.2
1,757.2
1,458.6
1,384.8; 1,368.6
1,273.4
1,188.2; 1,133.4; 1,090
1,045.2
569
media were 0.2 M phosphate buffer (pH 6.8), 0.05 M acetate

buffer (pH 4.5), and 0.1 M hydrochloric acid (pH 1.0)
maintained at 370.5C. The basket rotation speed was kept
at 50 rpm and the amount of dissolution media was 900 ml.
During the dissolution test, samples were withdrawn from the
dissolution medium at certain times and replaced with the
same volume of fresh dissolution medium. The concentration
of theophylline in each sample taken was analyzed using an
ultraviolet (UV) spectrophotometer (LKB Ultraspec II).
The analytical wavelength was 270 nm in 0.1 N hydrochloric
acid, 271 nm in pH 4.5 acetate buffer, and 272 nm in pH 6.8
phosphate buffer.
SPDLA, 46.75% MCC, and 6% theophylline with 0.5%

magnesium stearate acting as a lubricant. Physical mixtures
were prepared by mixing the components in a Turbula mixer
(W.A. Bachofen, Switzerland) at 32 rpm during 20 min. After
the addition of magnesium stearate, mixing was continued for
three subsequent minutes.
Cylindrically at faced tablets (250 mg, diameter 9 mm)
were manually compressed using an instrumented single
punch tablet machine (Korsch EK 0, Erweka Apparatebau
GmbH, Germany). To separate the constituent effects from
hardness effects, tablets of each SPDLA grade were com
pressed to a constant hardness of 90 N by varying the mass of
powder mixture for different SPDLA grades.
Study of the Tablet Surface by Scanning Electron Microscopy
Scanning electron micrographs (SEMs) were taken
before and after the dissolution test to estimate the tablet
surface structure and alterations caused by dissolution. The
micrographs were taken with a Zeiss DSM 962 (Carl Zeiss,
Oberkochen, Germany) scanning electron microscope. Be
fore scanning, the samples were coated with platinum using a
vacuum evaporator. SEM images were obtained at an
accelerated voltage of 8 to 10 kV.
Dissolution Studies
The in vitro release tests were performed using an USP
test apparatus type 1 (basket apparatus). The dissolution
Fig. 4. Differential scanning calorimetry data of sugar end capped

poly D,L lactide (SPDLA) polymers
Fig. 5. Scanning electron micrographs (SEM) of sugar end capped

poly D,L lactide (SPDLA) polymers in powder form. Key: a a
SPDLA 5300 polymer, b SPDLA 11000 polymer, and c a
SPDLA 17000 polymer. Magnication 100
Vuorinen et al.
570
The series of three SPDLA polymers with different
molecular weights were prepared for direct compressed
controlled release tablets. For the synthesis of SPDLAs, Sn
(NMe2)4 was applied as a catalyst precursor and was
prepared according to a literature procedure (18). A Sn
amido bond is susceptible for the hydrolysis, and as depicted
in Fig. 1, a treatment with an alcohol (here methyl D
glucopyranoside) selectively generates the corresponding tin
alkoxide complex. In comparison with other published tin
based catalyst precursors, Sn(IV)alkoxides and Sn(II)octoate,
this is the marked benet of Sn(NMe2)4. Due to the
generated tin oxygen bond, the tin alkoxide complex is the

actual catalyst for the ROP. In the polymer, the initial
alkoxide moiety appears as a chain end group (19).
Applied tin initiator is active toward polymerization of D,
L lactide in toluene at 120C and gives high conversion (96%,
95%, and 87%) after 17 h. Monomer conversion was
determined by integrating the relative intensities of the
methyl resonances attributable to the monomer and polymer.
The 1H NMR of SPDLA 5300 (Fig. 2a) shows the peaks at
1.68 (doublet) and 1.5 ppm, which are assigned to methyl
protons in the monomer unit and in the polymer. The
complete removal of monomer by precipitation from
CH2Cl2/hexane solution was veried by 1H NMR (Fig. 2b).
Fig. 6. Scanning electron micrographs (SEM) on the surface of initial sustained release sugar end capped poly D,L
lactide (SPDLA) matrix tablet before and after exposition to the dissolution test at pH 6.8. Key: a1 2 a
SPDLA 5300 polymer, b1 2 a SPDLA 11000 polymer, and c1 2 a SPDLA 17000 polymer. Magnication 100
571
Powder Properties and Tableting

For preparing direct compression sustained release ma
trix tablets, the release rate controlling polymer should
Fig. 7. Stereomicroscopic photographs of sustained release sugar

end capped poly D,L lactide (SPDLA) matrix tablets before and after
exposition to the dissolution test at pH 6.8. Key: a a SPDLA 5300
polymer, b a SPDLA 11000 polymer, and c a SPDLA 17000 polymer
The molecular weight of the polymers can be controlled

with high precision due to the living nature of the polymer
ization reaction. For this study, a series of three sugar end
capped polymers with molecular weights of 5,300, 11,000, and
17,000 g/mol were selected for future studies. After purica
tion, the polymers were dissolved into CH2Cl2, and subse
quent removal of the solvent with vacuum pump produced a
brittle polymer foam. This turned out to be a benecial pre
treatment method for the polylactides as gentle crushing of
the foam in a mortar gave ne and free owing polymer
powder.
Characterization of SPDLAs
FTIR spectrum of SPDLA 5300 polymer is shown in
Fig. 3. The assigned peaks for spectrum are listed in Table II
and are congruent with the frequencies previously reported in
literature for polylactides (20,21,22). FTIR spectra of
SPDLA 11000 and SPDLA 17000 polymers were similar to
that of SPDLA 5300 (not shown).
In order to characterize the physical state of SPDLAs,
DSC analysis was performed. SPDLAs are amorphous and
Tgs could be clearly detected at 39C, 43C, and 41C. The
differences in Tg were not signicant as the molecular weight
increased from 5,300 to 17,000. The glass transition temper
atures of all SPDLAs were above the dissolution temperature
(37C; Fig. 4).
Fig. 8. Percentage of theophylline released from the compressed

sugar end capped poly D,L lactide (SPDLA) tablet matrices contain
ing S PD LA 5300 poly mer, SPD LA 110 00 polym er, o r
SPDLA 17000 polymer. The dissolution tests of the tablet matrices
are performed at a pH 1.0, b 4.5, and c 6.8
Vuorinen et al.
572
exhibit good owing, consolidation, and compaction proper
ties. The SEM photographs in Fig. 5 show different grade
SPDLA powders which were prepared as described in the
experimental section. The shape of the particles is irregular
and particle size varies from 10 to 200 m for all SPDLA
grades. It was advantageous that SPDLA polymers were
directly synthesized to a free owing and readily compress
ible powder form so that additional grinding stages could be
avoided. According to the literature, formulation of direct
compression sustained release matrix tablets based on
PDLAs has failed so far due to the fact that mechanical
grinding of the present polymers is not possible (13).
From a technological point of view, no difculties or
limitations were met related to matrix tablet compression
procedure. According to SEM photographs, the tablets are
homogenous and form a rm matrix, with a surface free from
any clearly visible particulate borderlines (Fig. 6). Compari
son of Fig. 5 with Fig. 6 shows the plastic deformation of the
particles as a result of compression. Further, the hardness of
the tablets was high enough to resist handling and potential
subsequent processing such as packaging.
Drug Release
Stereomicroscopic photographs (Fig. 7) illustrate the
initial sustained release tablet matrices and the residues after
the dissolution test. Upon immersion of the tablets into the
dissolution medium, it was observed that the tablet volume
largely increased without forming a gel layer. The swelling of
the tablets also increased as the molecular weight of SPDLA
increased. However, the swelling in these formulations could
be mainly attributed to the MCC which has a propensity to
swell. During the dissolution tests, no erosion occurred (visual
inspection) and tablets did not disintegrate. Instead, reference
tablets prepared using MCC alone as the matrix former
disintegrated immediately.
Figure 8a c shows the dissolution proles of tablet
matrices tested at three different pHs of the dissolution
media (pH 1.0, pH 4.5, and pH 6.8). The dissolution proles
suggest that pH changes do not inuence any of these
preparations. pH independency is an important property
because sustained release preparations should not be affected
by pH changes in the digestive tract (23). The drug release
was the fastest with the lowest molecular weight SPDLA
grade and all of the drug load was released within approxi
mately 2 h. These tablets were the only ones for which the
drug release followed zero order rate. A linear release prole
as a function of time during the rst 90 min was obtained.
As the dissolution proles in Fig. 8a c indicate, the tablet
matrices containing 50% SPDLA 11000 exhibited clearly
prolonged drug release in three different pH media. The
drug release was complete after 5 h of immersion of the tablet
matrices in the dissolution medium. For these tablets, a linear
release relationship was obtained up to 80% when the
percentage released was plotted versus the square root of
time. This means that the release prole followed the
diffusion based kinetics. Variation in the dissolution results
of individual tablets was higher compared to that obtained
with the respective matrices based on the lower molecular
weight SPDLA (SPDLA 5300). As expected, the tablet
matrices containing 50% SPDLA 17000 exhibited prolonged
drug release, and 100% of the drug load was released at 8 h

after the start of the dissolution test (Fig. 8a c). For these
tablets, the drug release up to 80% exhibited a square root of
time dependency.
CONCLUSIONS
Sustained release tablets were successfully prepared by a
simple direct compression technique from a mixture of
SPDLA polymer and MCC. Dissolution studies showed
signicant correlations between the properties of the SPDLA
and the release proles of the tablets. The release of
theophylline slowed down progressively as the molecular
weight of the SPDLA increased. The preparations were not
inuenced by pH in a dissolution test, suggesting that they are
pH independent. Finally, release prole of different kinds of
drugs incorporated into these matrices can be modied
specically by choice of a particular polymer grade (Mw)
and the proportions of SPDLA and MCC in the mixture.
ACKNOWLEDGEMENT
The study has been carried out with nancial support
from Helsinki Universitys Research Funds.
REFERENCES
1. Ishikawa T, Watanabe Y, Takayama K, Endo H, Matsumoto M.
Effect of hydroxypropylmethylcellulose (HPMC) on the release
proles and bioavailability of a poorly water soluble drug from
tablets prepared using macrogol and HPMC. Int J Pharm
2000;202:173 8.
2. Williams RO III, Sykora MA, Mahaguna V. Method to recover a
lipophilic drug from hydroxypropyl methylcellulose matrix
tablets. AAPS PharmSciTech 2001;2(2):8. article.
3. Al Taani BM, Tashtoush BM. Effect of microenvironment pH of
swellable and erodible buffered matrices on the release charac
teristics of diclofenac sodium. AAPS PharmSciTech 2003;4(3):43.
article.
4. Brouillet F, Bataille B, Cartilier L. High amylose sodium
carboxymethyl starch matrices for oral, sustained drug release:
formulation aspects and in vitro drug release evaluation. Int J
Pharm 2008;356:52 60.
5. Stridsberg KM, Ryner M, Albertsson A C. Controlled ring
opening polymerization: polymers with designed macromolecular
architecture. Adv Polym Sci 2002;157:42.
6. Sdergrd A, Stolt M. Properties of lactic acid based polymers
and their correlation with composition. Prog Polym Sci
2002;27:1123 63.
7. Brannon Peppas L, Vert M. Polylactic and polyglycolic acids as
drug delivery carriers. In: Wise DL, editor. Handbook of
pharmaceutical controlled release technology. New York: Marcel
Dekker; 2000. p. 99 130.
8. Bodmeier R, Chen H. Evaluation of biodegradable poly(lactide)
pellets prepared by direct compression. J Pharm Sci 1998;78:819
22.
9. Murakami H, Kobayashi M, Takeuchi H, Kawashima Y.
Utilization of poly(DL lactide co glycolide) nanoparticles for
preparation of mini depot tablets by direct compression. J
Controlled Release 2000;67:29 36.
10. Takahashi M, Onishi H, Machida Y. Development of implant
tablet for a week long sustained release. J Controlled Release
2004;100:63 4.
11. Mank R, Kala M, Richter M. Darstellung peroralel retardarzneiformen
auf der basis von biologisch abbaubaren polymeren. 2. Mitteilung:
darstellung von matrixtabletten auf der basis von polymilchsure.
Pharmazie 1989;44:328 30.

12. Moll F, Kller G. Biodegradable tablets having a matrix of low
molecular weight poly L lactic acid and poly DL lactic acid. Arch
Pharm 1990;323:887 8.
13. Steendam R, Lerk CF. Poly(DL lactic acid) as direct compression
excipient in controlled release tablets. Part I. Compaction
behaviour and release characteristics of poly(DL lactic acid)
matrix tablets. Int J Pharm 1998;175:33 46.
14. Steendam R, van Steenbergen MJ, Hennink WE, Frijlink HW,
Lerk CF. Effect of molecular weight and glass transition on
relaxation and release behaviour of poly(DL lactic acid) tablets. J
Controlled Release 2001;70:71 82.
15. Mank R, Kala M, Richter M. Darstellung peroralel retardarzneiformen
auf der basis von biologisch abbaubaren polymeren. 1. Mitteilung:
darstellung und charakterisierung von polymilchsure. Pharmazie
1989;44:276 9.
16. Omelczuk MO, McGinity JW. The inuence of polymer glass
transition and molecular weight on drug release from tablets
containing poly(DL lactic acid). Pharm Res 1992;9(1):26 32.
17. Omelczuk MO, McGinity JW. A comparative investigation of the
compaction and dissolution properties of tablets containing poly
18.
19.
20.
21.
22.
23.
573
(D,L lactic acid) as a binder and retardant polymer. STP Pharm
Sci 1995;5(3):181 6.
Thomas IM. The preparation of alkoxides and triethylsilanolates
of Ti, Zr, V, Nb, Ta and Sn from the dialkylamides. Can J Chem
1961;39:1386 9.
Kalmi M, Lahcini M, Castro P, Lehtonen O, Belfkira A, Leskel
M, Repo T. Tetrakis Sn(IV) alkoxides as novel initiators for
living ring opening polymerization of lactides. J Polym Sci
2004;42:1901 11.
Garlotta D. A literature review of poly(lactic acid). J Polym
Environ 2001;9:63 84.
Kister G, Cassanas G, Vert M. Effect of morphology, conforma
tion and conguration on the IR and Raman spectra of various
poly(lactic acid)s. Polym 1998;39:267 73.
Agarwal M, Koelling KW, Chalmers JJ. Characterization of the
degradation of polylactic acid polymer in a solid substrate
environment. Biotechnol Prog 1998;14:517 26.
Streubel A, Siepmann J, Dashevsky A, Bodmeier R. pH
independent release of a weakly basic drug from water insoluble
and soluble matrix tablets. J Controlled Release 2000;67:101 10.

DOI: 10.1208/s12249 009 9237 y
Research Article
Formulation Design and Optimization of Novel Taste Masked Mouth-Dissolving
Tablets of Tramadol Having Adequate Mechanical Strength
Ashwini R. Madgulkar,1,2 Mangesh R. Bhalekar,1 and Rahul R. Padalkar1
Received 12 October 2008; accepted 9 March 2009; published online 14 May 2009
Abstract. The purpose of this work was to develop novel taste masked mouth dissolving tablets of
tramadol that overcomes principle drawback of such formulation which is inadequate mechanical
strength. Tramadol is an opioid analgesic used for the treatment of moderate to severe pain. Mouth
dissolving tablets offer substantial advantages like rapid onset of action, benecial for patients having
difculties in swallowing and in conditions where access to water is difcult. The crucial aspect in the
formulation of mouth dissolving tablets is to mask the bitter taste and to minimize the disintegration time
while maintaining a good mechanical strength of the tablet. Mouth dissolving tablets of tramadol are not
yet reported in the literature because of its extreme bitter taste. In this work, the bitter taste of Tramadol
HCl was masked by forming a complex with an ion exchange resin Tulsion335. The novel combination of
a superdisintegrant and a binder that melts near the body temperature was used to formulate
mechanically strong tablets that showed fast disintegration. A 32 full factorial design and statistical
models were applied to optimize the effect of two factors, i.e., superdisintegrant (crospovidone) and a
mouth melting binder (Gelucire 39/01). It was observed that the responses, i.e., disintegration time and
percent friability were affected by both the factors. The statistical models were validated and can be
successfully used to prepare optimized taste masked mouth dissolving tablets of Tramadol HCl with
adequate mechanical strength and rapid disintegration.
KEY WORDS: tramadol; taste masking; mechanical strength; mouth dissolving tablets; optimization.
INTRODUCTION
Tramadol is a centrally acting opioid analgesic structur
ally related to codeine and morphine used in the treatment of
moderate to severe pain in diverse conditions. Combined with
low dependence/abuse potential, it has proven to be of
signicant advantage over other agents, especially in the
elderly (1). Routes other than oral, like intravenous, which
are used to administer tramadol are used in acute conditions
like postoperative neuralgia or situations when the patient is
hospitalized. Unlike insulin, tramadol must be injected under
the supervision of a physician only. Patients suffering from
arthritis or neuralgia have to take the therapy for longer
duration of time. In such cases, oral route is the preferred
route. Thus, the problems like bitter taste and ease of
swallowing need to be solved for tramadol therapy. Mouth
dissolving tablets (MDT) are very benecial for patients with
difculties in swallowing and in conditions where access to
water is difcult. MDT dissolves fast and exerts rapid onset of
action (2).
MDT can be formulated using various methods. Some of
them involve increasing the porosity of the tablet and
decreasing the disintegration time (DT) (3). Superdisinte
1
AISSMS College of Pharmacy, Kennedy Road, Near RTO, Pune,

411001, India.
2
To whom correspondence should be addressed. (e mail: ashwini.
madgulkar@indiatimes.com)
grants are used that swell or absorb water rapidly to

disintegrate the tablet (4). Technologies like Zydis based on
lyophilization yield tablets that dissolve in a few seconds.
Most of the techniques aim at lowering the DT, but doing this
always compromises the mechanical strength. Zydis tablets
need special packaging and patient counseling for removing
the tablets from the strip (5). Hence, the objective of this
work was to formulate and optimize mouth dissolving tablets
of Tramadol HCl that disintegrate in a few seconds and still
have good mechanical strength. Mouth dissolving tablet of
tramadol is not reported yet because of its extreme bitter
taste. Currently, very few methods have been reported for
taste masking of tramadol and very few literature are
available for the same. Methods like polymer coating increase
drug particle size and are not useful for formulation into
mouth dissolving tablets, as they affect the mouth feel.
Commonly used resins for taste masking like Indion 234 or
Amberlite IRP 64 fail to mask the bitter taste of tramadol.
This work describes an optimized method of masking the
bitter taste of tramadol by ion exchange process.
Generally, a binder increases the mechanical strength but
delays DT. Hence, in the present work, a novel approach is
used which combines a binder having melting point near to
body temperature with a superdisintegrant. Initial screening
of various materials like polyethylene glycols, hydrogenated
vegetable oils, and combinations of various glycerides showed
that Gelucire 39/01 is the binder that ts for this purpose.
Crospovidone, which acts by very rapid water wicking
574
Formulation Design and Optimization of Mouth Dissolving Tablets

mechanism, was used as a superdisintegrant. Effects of
crospovidone and Gelucire on DT and percent friability of
the tablets were optimized using a 32 factorial design, and the
mathematical models were validated.
Materials
Tramadol HCl (Lupin Labs, Pune), Tulsion 335 (Ther
max India Ltd., Pune), crospovidone (ISP Corp, Hong Kong),
Gelucire 39/01(Gattefosse, France), orange avor(Keva Fla
vors, Mumbai), Aerosil, and sodium saccharine(Iatros
Pharma, Pune) were received as gift samples. Menthol,
mannitol, and citric acid (reagent grade) were purchased
from Loba Chemie.
Methods
Taste Masking of Tramadol HCl
Determination of threshold bitterness concentration of
Tramadol HCl (6). A panel of ten healthy human volunteers
(age 20 25) was selected. A series of solutions of Tramadol HCl
in phosphate buffer of pH 6.8 of concentrations 10, 20, 30, 40,
and 50 g/ml was prepared. The volunteers were asked to hold
10 mL of each solution in oral cavity for 30 s and rate the taste
on a scale from 0 to 4 (0: no bitterness, 1: threshold bitterness, 2:
bitter, 3: moderate bitterness, and 4: strong bitterness). Rinsing
the mouth by distilled water and a gap of 30 min were applied
between successive tests. Based on the opinion of the volun
teers, threshold bitterness concentration of Tramadol HCl was
judged.
Preparation of DRC (7,8). Batch method was used to
prepare drug resin complex (DRC). Five grams of Tulsion
335 was allowed to swell for 90 min in a beaker containing
200 mL of deionized water of conductivity less than 2. Then,
5 g of Tramadol HCl was added to it and stirred for 300 min.
The mixture was ltered through Whatman lter paper no.
41. The residue was washed with 100 mL of deionized water
and dried.
Optimization of various parameters for maximum drug
loading (9,10). Drug loading process was optimized for
maximum drug loading considering parameters like effect of
activation, resin swelling time, and pH. The effect of
activation of resin on drug loading was studied as follows.
Five grams of Tulsion 335, placed on a Whatman lter paper
in a funnel, was washed with deionized water and subse
quently with 1 N HCl (100 mL). The resin was rewashed with
water until neutral pH was reached. DRC was prepared in
the same way as discussed earlier using 5 g each of Tramadol
HCl and acid activated resin. Similarly, alkali activation of
Tulsion 335 was done, replacing 1 N HCl with 1 N NaOH.
Finally, Tulsion 335 was also activated with combined
treatment of 1 N HCl and 1 N NaOH solutions. Drug loading
efciency in each case was determined.
To study the effect of other parameters, activated resin
was used.
575
Optimization of resin swelling time was carried out by

keeping 5 g of Tulsion 335 in each of the ve beakers
containing 200 mL of deionized water for 30, 60, 90, 120, and
180 min, respectively. DRC was prepared as described above
using 5 g of Tramadol HCl and percent drug loading was
estimated. To assess the effect of pH on drug loading, drug
solution was prepared in deionized water and pH was
adjusted to 2, 3, 4, 5, 6, 7, and 8 using standard solutions of
hydrochloride and sodium hydroxide. Equal quantity of resin
was swollen separately and added to drug solution and
loading efciency determined at these conditions.
Characterization of DRC
In vitro taste evaluation (6,8). A quantity of DRC equiva
lent to 50 mg of Tramadol was added to each of the six
volumetric asks containing 10 mL of phosphate buffer of
pH 6.8. The mixtures were shaken for 0, 15, 30, 60, 120, and 300 s
and ltered. Content of Tramadol in each ltrate was deter
mined. For satisfactory taste masking, the amount of drug
dissolved at the end of 120 s should not be more than the
threshold bitterness concentration of the drug.
Confirmation of Complexation (9,11,12)
FTIR studies. Tramadol HCl, Tulsion 335, and physical
mixture of both and DRC were subjected to Fourier transform
infrared spectroscopy (FTIR) studies. Samples were prepared
using KBr disc method and spectra were recorded over the
range 400 to 4,000 cm 1. Spectra were analyzed for drug resin
interactions and functional groups involved in the complexation
process.
Powder X ray diffraction studies. X ray diffractograms of
Tramadol HCl, Tulsion 335, and DRC were recorded using
Philips PW 3710 deffractometer and analyzed for interactions
between drug and resin and conrmation of complexation.
Thermal analysis. Differential scanning calorimetry
(DSC) was carried out using Mettler Toledo 823e instrument
equipped with intracooler. Indium zinc standards were used
to calibrate the temperature and enthalpy scale. The samples
were hermetically sealed in aluminum pans and heated over
the temperature range 30C to 300C with heating rate of
10C/min. Inert atmosphere was provided by purging nitro
gen gas owing at 40 mL/min.
Estimation of drug content. One hundred milligrams of
DRC was stirred with 100 mL of 0.1 N HCl for 60 min so as to
release the entire drug from DRC. The mixture was ltered
and 1 mL of the ltrate was diluted to 100 mL using 0.1 N
HCl. The absorbance of this solution was measured at
272.3 nm using 0.1 N HCl as blank and the content of
Tramadol estimated.
In vitro dissolution studies (11). A quantity of DRC
equivalent to 50 mg of Tramadol was subjected to dissolution
studies using USP type II dissolution test apparatus at 372C
at 50 rpm speed. Nine hundred milliliters of 0.1 N HCl was
used as dissolution medium. Aliquot equal to 5 mL was
576
Madgulkar, Bhalekar and Padalkar
withdrawn at specic time intervals and amount of Tramadol

released from DRC was determined.
Formulation development. DRC was used to prepare
MDT by direct compression technique. Composition of tablets
is mentioned in Table I. All materials except Gelucire 39/01
were passed through sieve number 85. Crospovidone was
divided into two equal parts by weight. DRC, one part of
crospovidone, sodium saccharine, citric acid, and mannitol were
mixed uniformly and heated to 40C. Gelucire 39/01 was melted
separately at the same temperature and added to this powder
blend with continuous mixing. The mass was passed through a
sieve number 18 to form granules and cooled. After cooling, the
granules were mixed with remaining crospovidone, menthol,
orange avor, and lubricated with aerosil. Tablets were
compressed using 8 mm at punch on Minipres II MT Rimek
16 station compression machine. Tablet weight was maintained
at 250 mg. Target tablet hardness was between 3 and 4 kg/cm2.
Optimization of Formulation (13 16)
Selection of suitable experimental design. In a full facto
rial design, all the factors are studied in all the possible
combinations, as it is considered to be most efcient in
estimating the inuence of individual variables (main effects)
and their interactions using minimum experimentation. In the
present study, tting a cubic model is considered to be better
as the values of the response surfaces are not known from the
previous ndings. Hence, 32 factorial design was chosen for
the current formulation optimization study. Amounts of
crospovidone and Gelucire 39/01 were selected as indepen
dent factors, whereas DT and percent friability (%F) were
measured as responses. Based on initial trials, levels of
crospovidone were selected as 7.5, 15, and 22.5 mg, whereas
Gelucire 39/01 levels were 10, 15, and 20 mg. Nine formula
tions were prepared according to 32 factorial design and
evaluated. The responses were analyzed for analysis of
variance (ANOVA) using Design Expert version 7.1.5
software. Statistical models were generated for each response
parameter. The models were tested for signicance.
Validation of statistical model. Levels of crospovidone
and Gelucire 39/01 were selected at six different points and
responses predicted by the statistical models were calculated.

Mouth dissolving tablets were prepared using these levels and
responses were measured practically. The predicted responses
were compared against observed responses and closeness
between them was checked.
Response surface plots. Response surface plots were
generated for each response to study the effect of both
factors on each response.
Evaluation of Tablets
General parameters (17,18). Tablets were evaluated for
hardness (Monsanto hardness tester), friability (Roche fria
bilator), and weight variation.
Uniformity of content. Five tablets were selected ran
domly and dissolved in 100 mL of 0.1 N HCl, stirred for
60 min, and ltered. One milliliter of the ltrate was diluted
to 100 mL with 0.1 N HCl. Absorbance of this solution was
measured at 272.3 nm using 0.1 N HCl as blank and content
of Tramadol was estimated.
Disintegration time. Many reports suggest that conven
tional DT apparatus may not give correct values of DT for
MDTs (19 21). The amount of saliva available in the oral
cavity is very limited (usually less than 6 mL), whereas the
conventional DT apparatus uses a large amount of water with
very rapid up and down movements. MDT is required to
disintegrate in such small amounts of saliva within a minute
without chewing the tablet. In a simplest method to
overcome these problems, 6 mL of phosphate buffer of
pH 6.8 was taken in a 25 mL measuring cylinder. Tempera
ture was maintained at 372C. A MDT was put into it and
time required for complete disintegration of the tablet was
noted.
Wetting time (22). A Petri dish containing 6 mL of
distilled water was used. A tissue paper folded twice was
kept in the dish and a tablet was placed on it. A small
quantity of amaranth red color was put on the upper surface
of the tablet. Time required for the upper surface of the tablet
to become red was noted as the wetting time of the tablet.
Table I. Composition of Mouth Dissolving Tablets as per 32 Factorial Design

Formulation code
Ingredient
A1
A2
A3
A4
A5
A6
A7
A8
A9
DRC (eq. to 50 mg of Tramadol)

Crospovidone
Gelucire 39/01
Citric acid
Sodium saccharine
Menthol
Orange avor
Aerosil
Mannitol
Total
103.9
7.5
10
10.5
1.5
1.0
1.5
3.5
110.6
250
103.9
7.5
15
10.5
1.5
1.0
1.5
3.5
105.6
250
103.9
7.5
20
10.5
1.5
1.0
1.5
3.5
100.6
250
103.9
15
10
10.5
1.5
1.0
1.5
3.5
103.1
250
103.9
15
15
10.5
1.5
1.0
1.5
3.5
98.1
250
103.9
15
20
10.5
1.5
1.0
1.5
3.5
93.1
250
103.9
22.5
10
10.5
1.5
1.0
1.5
3.5
95.6
250
103.9
22.5
15
10.5
1.5
1.0
1.5
3.5
90.6
250
103.9
22.5
20
10.5
1.5
1.0
1.5
3.5
85.6
250
All quantities in milligrams
577
Dissolution studies (23). Dissolution test was carried out

using USP type II dissolution test apparatus at 372C and
50 rpm speed. Nine hundred milliliters of 0.1 N HCl was used
as dissolution medium. Aliquot equal to 5 mL was withdrawn
at specic time intervals and amount of Tramadol released
from DRC was determined.
RESULTS
Threshold Bitterness
Most of the volunteers reported 20 g/mL as the
threshold bitterness concentration for Tramadol HCl.
Optimization of Parameters for Maximum Drug Loading
The percent drug loading with inactivated resin, treated
with acid, alkali, and combination thereof was found to be
88.051.2%, 92.541.08%, 93.611.31%, and 95.780.26%,
respectively. Highest drug binding on resin was achieved
when activated with both acid alkali treatments. It was noted
that the resin requires proper swelling time for maximum
drug loading. The loading efciencies for resin swelling time
of 30, 60, 90, 120, and 180 min were found to be 80.34
1.28%, 92.421.56%, 95.471.19%, 95.891.24%, and 95.97
1.47%, respectively. It was observed that a swelling time of
90 min was sufcient for maximum drug loading. When
Tramadol HCl was loaded in different pH environments like
2, 3, 4, 5, 6, 7, and 8, loading was found to be 81.670.89%,
83.12 0.78%, 84.42 1.56%, 91.201.15%, 92.62 1.31%,
95.241.06%, and 95.360.82%, respectively. It was ob
served that optimum drug loading was achieved at neutral
pH and was not much increased at pH higher than this.
Characterization of DRC
Fig. 1. FTIR spectra of Tramadol HCl (A), Tulsion 335 (B) and DRC (C)
diffractograms of Tulsion 335 showed diffused peaks, indicat

ing its amorphous nature, while the diffraction pattern of
DRC represents complete disappearance of crystalline peaks
of drug (Fig. 2).
DSC studies. The thermogram of Tramadol HCl shows a
sharp endothermic peak at 182.51C, whereas no sharp peaks
were observed in case of Tulsion 335. The endothermic peak
of tramadol was absent in the thermogram of DRC (Fig. 3).
Determination of drug content. When DRC was pre
pared using all of the optimized parameters for drug loading,
the percent drug loading was found to be 96.24%, and hence,
the drug content was 48.14% (w/w).
In vitro Taste Evaluation
In vitro Dissolution Studies
No detectable amount of Tramadol HCl dissolved in the

phosphate buffer of pH was detected at the end of 300 s.
Thus, the DRC did not release any drug at salivary pH and
taste masking of Tramadol HCl by making an ion exchange
complex with Tulsion 335 was complete and satisfactory.
The dissolution prole of DRC showed complete drug

release within 20 min. Thus, the loaded drug is quickly
released in acidic conditions.
Confirmation of Complexation (9,10,24)
Tablet blends of all formulations showed good owabil

ity (angle of repose <30 and Carrs index 12). Bulk
densities varied from 0.8 to 0.9 (Table II).
FTIR studies. The IR spectrum of Tramadol HCl showed

C N str at 3,250 cm 1, C H str at 2,950 cm 1, O H str at
1,380 cm 1, C O str at 1,257 cm 1, and peak around 1,680 cm 1
corresponding to hydrohalide salt associated with tertiary
amine. The peaks corresponding to C N str and C H str
disappeared in the spectrum of DRC. Also, the characteristic
peak at 1,680 cm 1 representing HCl was not seen in the
spectrum of DRC. The spectrum of Tulsion 335 showed distinct
C=O str of the COOH functional group of the resin polymer
matrix which was not seen in the spectrum of DRC (Fig. 1).
Physical Properties of Tablet Blend
Evaluation of Tablets
The outcomes of various evaluation parameters are
shown in Tables III and IV. The dissolution studies indicated
100% drug release within 20 min from all batches (Fig. 4). All
batches showed DT less than a minute and friability less than
1.02%.
32 Factorial Design
Powder X ray diffraction studies. The X ray diffracto

grams of Tramadol HCl conrmed its crystalline nature, as
evidenced from the number of sharp and intense peaks. The
A statistical model, Y b0 b1 X1 b2 X2 b12 X1 X2

b11 X1 X1 b22 X2 X2 , incorporating interactive and polynomi
578
Fig. 2. X ray diffraction patterns of Tramadol HCl (A), Tulsion 335 (B), and DRC (C)
al terms was used to evaluate the responses, where Y is the

dependent variable, b0 is the arithmetic mean response of the
nine runs, and bi is the estimated coefcient for the factor Xi.
The main effects (X1 and X2) represent the average result of
changing one factor at a time from its low to high value. The
interaction terms (X1X2) show how the response changes
when two factors are simultaneously changed. The polyno
mial terms (X1X1 and X2X2) are included to investigate
nonlinearity. The DT and %F for the nine batches (A1 to A9)
showed a wide variation (i.e., 12 to 54 s and 0.41% to1.02%,
respectively). The data clearly indicate that the DT and %F
values are strongly dependent on the selected independent
variables. The tted full equations relating the responses DT
and %F to the transformed factors are as:
DT 31:56
15:83X1 0:61X1 X1
5:33X2 0:11X2 X2
1
%F
0:68
0:21X1 0:008333X1 X1
0:08X2 0:003333X2 X2 :
2
Table V shows the results of the ANOVA which was used
to generate statistical models.
Response Surface Plots

It was observed that DT and %F were dependent on
both the factors. There was a linear decrease in the
disintegration time with increase in the levels of both factors.
The same effect was observed with Gelucire 39/01 (Fig. 5).
Validation of Statistical Model
The predicted responses of the six formulations and
corresponding actual experimentally observed values were
found to be in close agreement as indicated in Table VI.
Thus, the models developed to predict the responses were
not only signicant statistically but also were found to be
valid to predict values that were very close to the practical
observations.
DISCUSSIONS
Loading of Tramadol HCl on Tulsion 335 was carried out
by batch process. Batch process is the preferred method for
loading a drug into nely divided ion exchange operations
commonly used with ion exchange resins. Due to its ne
Fig. 3. DSC thermograms of Tulsion335 (A), Tramadol HCl (B), and DRC (C)
579
Table II. Evaluation of Physical Properties of Tablet Blends

Formulation
code
Bulk density
Tapped density
Carrs
index
Angle of
repose (deg)
A1
A2
A3
A4
A5
A6
A7
A8
A9
0.82940.02
0.86260.03
0.84390.04
0.83210.02
0.82840.02
0.86750.02
0.8450.02
0.8640.02
0.82210.03
0.92960.03
0.90360.04
0.91560.01
0.92210.02
0.92860.04
0.91350.02
0.99160.01
0.92340.04
0.92470.02
12.08
4.75
8.49
10.81
11.96
5.30
9.67
11.73
12.48
27.780.03
28.080.02
26.060.04
24.670.01
25.550.02
27.080.01
25.060.01
23.670.02
24.850.01
n 3
particle size, Tulsion 335 does not lend itself to conventional

columnar operations commonly used with ion exchange
resins. Higher swelling efciency in the batch process makes
more surface area available for ion exchange. So batch
process was selected. Treating an ion exchange resin with
combined acid and alkali treatment is considered as rejuve
nation of the resin. This ensures removal of the adsorbed
impurities associated with industrial scale manufacture or
absorbed during storage or handling and activation of the
exchangeable groups. The ion exchange process is pH
dependent, and thus, treating the resin at two extremities of
pH ensures that more exchangeable groups are available for
maximum and rapid drug loading. Hence, activation of the
resin is required. Swelling and hydration increases the rate
and extent of ion exchange process. In unswollen resin
matrix, the exchangeable groups are latent and coiled
towards the backbone. Swelling increases the surface area
and these groups get oriented towards outside. pH affects the
extent of drug loading process. The mobile or exchangeable
cation in Tulsion 335 is the hydrogen ion. In acidic environ
ments (generally pH below 4), the resin exists as free acid in
an essentially nonionic state. Loading of drug onto Tulsion
335 occurs maximum at pH higher than the acidic range.
Thus, it was observed that Tulsion 335 activated by combined
acid and alkali treatment, swollen for 90 min in deionized
water, showed maximum loading of Tramadol HCl at pH
near neutral. The formation of ion exchange complex
between Tramadol HCl and Tulsion 335 was conrmed by
Table IV. Evaluation of Tablet Formulations

Physical evaluation
Trial no.
Weight variation Hardness

Diameter Thickness
(mg)SD
(kg/cm2)SD (mm)SD (mm)SD
A1
A2
A3
A4
A5
A6
A7
A8
A9
Broad range
250.311.61
250.191.32
250.421.47
249.921.46
250.892.11
250.331.68
251.061.20
249.861.42
250.631.63
249.86 250.89
3.20.23
3.50.52
3.40.29
3.90.52
3.80.59
3.30.76
3.70.72
3.20.53
3.10.28
3.1 3.9
80.05
80.04
80.03
80.00
7.90.05
80.03
80.02
80.04
80.02
7.9 8
2.50.02
2.70.03
2.80.04
2.70.02
2.50.01
2.60.02
2.80.05
2.80.03
2.70.04
2.5 2.8
n 3
FTIR, X ray, and DSC studies. It was observed in the FTIR

studies that functional groups involved in the complexation
process were COOH of Tulsion 335 and HCl along with the
tertiary amine of tramadol HCl. The absence of other peaks
of Tramadol HCl discussed earlier in the spectrum of DRC
indicated that the drug was completely embedded in the resin
polymer matrix. X ray studies indicated the crystalline nature
of tramadol HCl and amorphous nature of Tulsion 335 and
DRC. The drug was entrapped in the resin and hence the
sharp crystalline peaks of the drug were not observed in the
X ray diffraction pattern of the DRC. The sharp endothermic
peak at 182.51C corresponds to melting of pure drug and its
crystalline nature. The thermogram of Tulsion 335 and DRC
indicated amorphous nature. The formation of ion exchange
complex and entrapment of Tramadol HCl in the polymer
matrix of Tulsion 335 was thus conrmed from the ndings of
these three studies. All tablet blends showed good ow
properties. Dissolution studies of DRC and mouth dissolving
tablets showed similar prole of drug release. It was observed
that the release of tramadol was controlled by desorption
from DRC and independent of formulation. The exchange
able ion for Tulsion 335 is the H+ ion. Tulsion 335 readily
loses the adsorbed species in H+ ion environment. The tablet
formulations were optimized using 32 factorial design. The
Table III. Evaluation of Various Parameters of Tablets
Trial no.
Disintegration
time (s)
Friability
% Assay
Wetting
time (s)
A1
A2
A3
A4
A5
A6
A7
A8
A9
Broad range
541.02
480.89
421.21
360.67
300.96
250.82
210.69
160.77
120.81
12 54
1.020.02
0.880.01
0.80.04
0.730.02
0.660.03
0.6000.02
0.540.01
0.480.03
0.410.01
0.41 1.02
98.641.02
99.211.26
99.420.85
98.410.62
98.531.08
98.020.96
97.931.21
99.781.05
97.230.85
97.23 99.78
400.94
380.57
341.21
280.68
210.59
190.74
160.81
140.47
90.52
9 40
n 3
Fig. 4. Release prole of tramadol from tablet formulations
580

Table V. ANOVA for Selected Statistical Model
Response model
DT
%F
Sum of squares
df
Mean square
F value
P value
R2
1,681.78
0.31
4
4
420.44
0.077
688.00
98.51
<0.0001
0.0003
0.9985
0.9992
outcomes for response parameters, i.e., DT and friability,

were subjected to regression analysis, and statistical models
were found to be signicant. The high values of correlation
coefcient for DT and %F indicate a good t, i.e., good
agreement between the dependent and independent
variables. The equations may be used to obtain estimates of
the response as a small error of variance was noticed in the
replicates. The F value in the ANOVA table is the ratio of
model mean square to the appropriate error mean square.
The larger the ratio, the larger the F value and the more
likely that the variance contributed by the model is
signicantly larger than random error. If the F ratio the
ratio of variances lies near the tail of the <F> distribution,
then the probability of a larger F is small and the variance
ratio is judged to be signicant. Usually, a probability less
than 0.05 is considered signicant. The F distribution is
dependent on the degrees of freedom <df> for the variance
in the numerator and the <df> of the variance in the
denominator of the F ratio. The model F value of 688 for
DT and 98.51 for friability and high R2 values suggest that
these models are signicant. There is only 0.01% chance that
a model F value this large could occur due to noise. Values
of p less than 0.0500 indicate that model terms are
signicant. In this case, both the models generated for DT
and %F are signicant. As there are no insignicant terms,
model reduction is not required. Adequate precision
measures the signal to noise ratio. A ratio greater than 4 is
desirable. The ratios of 72.654 and 27.963, respectively, for
DT and %F models indicate an adequate signal for each.
These models can be used to navigate the design space. It was
observed that disintegration time was dependent on both
factors. A linear decrease in the disintegration time was
observed with an increase in the levels of both factors.
Adeq. precision
72.654
27.963
Crospovidone, being a superdisintegrant, enhances faster

disintegration. In addition, Gelucire 39/01 also helps in the
disintegration process by melting near body temperature. It
was observed from the response surface and contour plots
that both the factors had inuence on the friability of the
tablets. There was a linear decrease in the values of friability,
as the levels of crospovidone and Gelucire were increased.
Gelucire 39/01, being a binder, increases the mechanical
strength of the tablet, and crospovidone is also known to
produce mechanically strong tablets. Optimum values of
friability were observed at high levels of both factors. Thus,
it was observed that the novel combination of a super
disintegrant and a binder that melts near body temperature
produced mouth dissolving tablets having adequate mechanical
strength and rapid disintegration. Thus, the major problem in
formulating mouth dissolving tablets, i.e., poor mechanical
strength, was overcome. A good evaluation of a statistical
model is not how well it ts the data but how well it predicts the
points. Comparison of predicted responses and observed values
for the same showed close agreement, and the models were
found to be valid.
CONCLUSION
It was concluded that the bitter taste of tramadol
hydrochloride can be masked by forming an ion exchange
complex with Tulsion 335. Mouth dissolving tablets of
tramadol having rapid disintegration and good mechanical
strength can be prepared using a novel combination of
crospovidone and Gelucire 39/01. Thirty two factorial design
and statistical models can be successfully used to optimize the
formulations.
Table VI. Comparision of Predicted Values and Experimental Values

Formulation
code
Predicted
valuesSD
Experimental
valuesSD
Residuals
A10
DT 500.5
%F 0.930.02
DT 441.1
%F 0.850.01
DT 390.76
%F 0.770.04
DT 280.56
%F 0.630.03
DT 170.98
%F 0.480.01
DT 120.84
%F 0.410.02
DT 510.76
%F 0.910.02
DT 421.02
%F 0.820.01
DT 400.69
%F 0.740.02
DT 260.58
%F 0.610.02
DT 190.74
%F 0.500.01
DT 130.49
%F 0.430.02
1
0.02
2
0.03
1
0.03
2
0.02
2
0.02
1
0.02
A11
A12
A13
A14
A15
Fig. 5. Response surface plots for DT and % Friability
n 3

ACKNOWLEDGMENTS
We are thankful to Lupin Labs, Pune, Gattefosse,
France, ISP Corp., Hong Kong, Iatros Pharma, Pune, and
Keva avors, Mumbai for providing necessary materials as
gift samples. We also express sincere thanks to Dr. K.G.
Bothara, Principal, AISSMS College of Pharmacy, Pune for
providing necessary facilities to carry out this research work.
REFERENCES
1. Grond S, Sablotzi A. Clinical pharmacology of tramadol. Clin
Pharmacokinet 2004;43(13):879 923.
2. Bagul U et al. Manufacturing technologies for mouth dissolving
tablets. www.pharmainfo.net, 2006; May 31.
3. Biradar S, Bhagavati S and Kuppasad. Fast dissolving drug
delivery system: a brief overview. Internet J Pharmacol. 2006;4(2).
4. Chang R, Guo X, Burnside B, Couch R. Fast dissolving tablets.
Pharm Technol 2000;24(6):52 8.
5. Bogner R, Meghan F. Fast dissolving tablets. US Pharmacist
2005;27:03.
6. Sohi H, Sultana Y, Khar R. Taste masking technologies in oral
pharmaceuticals: recent development and approaches. Drug Dev
Ind Pharm 2004;30:429 48.
7. Borodkin S, Sundberg D. Polycarboxylic acid ion exchange
adsorbates for taste coverage in chewable tablets. J Pharm Sci
1971;50:1523.
8. Avari J, Bhalekar M. Cation exchange resins for taste masking and
rapid dissolution of sparoxacin. Indian Drugs 2004;41(1):19 23.
9. Pisal S, Zainnuddin R, Nalawade P, Mahadik K, Kadam S.
Molecular properties of ciprooxacin indion 234 complexes.
AAPS PharmSciTech 2004;5(4):e62.
10. Chaudhari P, Chaudhari S, Lanke S, Patel N. Formulation and in
vitro evaluation of taste masked orodispersible dosage form of
Levocetirizine dihydrochloride. Indian J Pharm Edu Res 2007;41
(4):319 28.
11. Kalsi P. Spectroscopic identication of organic compounds, 5th ed.
New Delhi: New Age International Publishers; 2002. p. 84 102.
581
12. Skoog D, Holler F, Nieman T. The principles of instrumental

analysis, 5th ed. Singapore: Thomson Book/Cole; 1998. p. 805 8.
13. Bolton S. Pharmaceutical statistics, 2nd ed. New York: Marcel
Decker; 1990. p. 532 34.
14. Singh B, Ahuja N. Response surface optimization of drug
delivery system. In: Jain NK, editor. Progress in controlled and
novel drug delivery systems, 1st ed. New Delhi: CBS Publishers
and Distributors; 2004. p. 470 509.
15. Singh B, Ahuja N. Development of controlled release buccoad
hesive hydrophilic matrices of diltiazem hydrochloride: optimi
zation of bioadhesion, dissolution and diffusion parameters.
16. Gohel M, Patel M, Agarwal R, Amin A, Dave R, Bariya N.
Formulation design and optimization of mouth dissolving tablets
of nimesulide using vacuum drying technique. AAPS PharmSci
Tech 2004;5(3):36.
17. Lachmann L, Liberman H, Kanig J. The theory and practice of
industrial pharmacy, 3rd edn. Mumbai: Varghese; 1987. p. 296
300.
18. Indian Pharmacopoeia 1996, vol I. The controller of publications,
New Delhi.
19. Morita Y, Tsushima Y, Termoz R, Ajioka J, Takayama K.
Evaluation of disintegration time of rapidly disintegrating tablets
via a novel method using a CCD camera. Chem Pharm Bull
(Tokyo) 2002;50:1181 6.
20. Watanabe Y, Yamamoti Y, Fujii M, Kondoh M, Shibata Y. A
novel method for predicting disintegration time in the mouth of
rapid disintegrating tablet by compaction analysis using Tab All.
Chem Pharm Bull (Tokyo) 2004;52:1394 5.
21. Narazki R, Harada T, Takami N, Koto Y, Ohwaki T. A new
method for disintegration studies of rapid disintegrating tablets.
Chem Pharm Bull (Tokyo) 2004;52:704 7.
22. Kuchekar B, Badhan A, Mahajan H. Mouth dissolving tablets of
salbutamol sulphate: a novel drug delivery system. Indian Drugs
2004;41(10):592 8.
23. United States Pharmacopoeia 29 NF 24. The Ofcial Compendia
of Standards. Asian ed. Rockville, MD: United States Pharma
copoeial Convention Inc.; 2006. p. 2679 2681.
24. Khan S, Kataria P, Nakhat P, Yeole P. Taste masking of
Ondansetron hydrochloride by polymer carrier system and
formulation of rapid disintegrating tablets. AAPS PharmSciTech
2007;8(2):e46.

DOI: 10.1208/s12249 009 9245 y
Research Article
Freeze-Dry Microscopy: Impact of Nucleation Temperature and Excipient
Concentration on Collapse Temperature Data
Eva Meister,1,3 Slobodan ai,2 and Henning Gieseler1,4
Received 12 September 2008; accepted 23 April 2009; published online 14 May 2009
Abstract. The objective of this study was to investigate the impact of nucleation temperature (Tn) and
excipient concentration on the collapse temperature data obtained from freeze dry microscopy (FDM)
experiments. Tn, the temperature of the onset of collapse (Toc), and the full collapse temperature (Tfc)
were determined for aqueous solutions of polyvinylpyrrolidone (PVP) 40 kDa and 2 (hydroxypropyl)
cyclodextrin. Concentrations were varied from 1% to 20% (w/w) for PVP and from 1% to 30% (w/w) for
the 2 (hydroxypropyl) cyclodextrin. Mutual correlation coefcients were calculated for the observed
Tn, Toc, and concentrations of the solutions. In addition, outliers were detected and eliminated by
applying the leaving one out routine and calculating correlation coefcients without it. Tn was found to
be non correlated with concentrations and only weakly correlated with Toc. The correlation between
these two temperatures was particularly poor for the solutions of the highest and lowest concentrations.
In contrast, Toc correlated much better with the corresponding concentrations, resulting in a quadratic t
for PVP and a linear t for 2 (hydroxypropyl) cyclodextrin.
KEY WORDS: collapse temperature; concentration; freeze dry microscopy; nucleation.
INTRODUCTION
Freeze drying is known as an important and frequently
used drying technology for biopharmaceutical products which
are not stable in aqueous solution over an acceptable shelf
life (1). Biopharmaceuticals are often used to ght cancer,
cardiovascular or diabetic diseases, or viral infections. To
increase production turnover while assuring a high quality of
such costly medication, the critical process parameters of the
freeze drying cycle must be optimized (2). Optimization of a
freeze drying cycle means to reduce cycle time. This can be
achieved by optimizing in particular primary drying condi
tions by selecting appropriate chamber pressure and shelf
temperature settings. As a general rule, the product temper
ature at the ice sublimation interface must be maintained just
below the critical formulation temperature throughout this
process step (3). The critical formulation temperature can be
either determined by freeze dry microscopy (FDM) or
differential scanning calorimetry (DSC) and is denoted
collapse temperature (Tc) for the FDM or glass transition
temperature of the maximally freeze concentrated solute
(Tg) for DSC experiments. However, it has been reported in
1
Freeze Drying Focus Group, Division of Pharmaceutics, University

of Erlangen, Cauerstrasse 4, 91058, Erlangen, Germany.
2
Analytical Research and Development, Pzer Inc., Eastern Point
Road, Groton, 06340, Connecticut, USA.
2
Present address: Dept. R&D Coordination, Boehringer Ingelheim
GmbH, 55218, Ingelheim, Germany.
4
To whom correspondence should be addressed. (e mail: gieseler@
freeze drying.eu)
the literature that the measure of Tg is not necessarily

representative for the sublimation procedure in a vial during
a freeze drying run (4). Therefore, freeze dry microscopy has
become the method of choice for critical temperature
measurements over the last couple of years (5). Collapse in
a given region of the product results from surface tension
induced viscous ow of the amorphous phase after the ice
vapor interface has moved past that particular region (4).
When performing primary drying above Tc, one may observe
loss of structure (denoted as shrinkage or collapse) in the
dried region adjacent to the ice vapor interface due to a glass
transition in the amorphous product (6). Shrinkage or
collapse may compromise the (predened) product quality
attributes, e.g., product stability, residual moisture content of
the cake, reconstitution times, etc. (7).
Recent studies focused on a rational design of FDM
methodology, based on various collapse temperature measure
ments and different model systems (8). For a representative Tc
measurement and subsequent optimization of the freeze
drying process, the collapse detection must be as exact and
representative as possible and the inuencing parameters must
be known and taken into consideration during the experiment.
When applying the cooling step during an FDM experi
ment, nucleation of the sample can neither be controlled nor
predicted. As a result, differences in the size of ice crystals
among samples may be observed in the course of experimenting,
even under identical experimental conditions. This would also
be expected during a conventional freeze drying run and has
already been discussed excessively in the literature: the higher
the supercooling of the solution, the less time remains for ice
crystal growth and the smaller is their size. In 2001, Searles et al.
582
FDM: Nucleation Temperature and Excipient Concentration
583
published a study which conrmed that the stochastic nature of

nucleation led to heterogeneity among samples concerning the
freezing mechanism, morphology and primary drying rate (9).
Out of this reason and coincidental factors (particles, damaged
glass cover slides), replicates of FDM measurements are
unlikely to ever produce exactly the same frozen structure.
When the formation of ice crystals is different between replicate
runs in an FDM experiment, the formation of different pore
sizes during the ice sublimation phase is expected to have an
inuence on resistance of the dried structure to vapor ow
which, in turn, might inuence drying rate and collapse behavior
(4,9 10). Apart from the pore size distribution of the dried
matrix, collapse behavior depends on the total solid content of
the excipient (4).
In the present study, mutual correlation between the nucle
ation temperature (Tn), the temperature of the onset of collapse
(Toc), and the total solid content of the excipient was for the rst
time evaluated by FDM experiments. Based on the excessive
number of data points and the use of two different model
excipients, it was possible to conduct such a detailed investiga
tion of interrelation between Tn, Toc, and concentration.
ramp rate used in the FDM software (Linksys 32, Linkam,

Surrey, UK). Note that the rational behind the use of a
10C/min cooling rate in this study is to simply make the
experimental time more appropriate for the amount of data
required for statistical analysis. Further, a much slower cooling
rate was not expected to show a signicant difference in the
results obtained, based on a recent report which showed for
5% (w/w) aqueous solutions of sucrose no difference in collapse
temperature measured by FDM when using a cooling rate of
1C/min and of 10C/min (11). These results might be perfectly
plausible if the variability of the nucleation temperature during
an FDM experiment is taken into consideration, even when
using the same sample and identical experimental conditions
(i.e., Tn data were found exactly in the same temperature range
for a 1C/min and 10C/min cooling rate) (12).
The heating rate was 1.0C/min in the temperature range
of interest to allow a representative measure of the rst
structural changes in the product. The pressure in the stage
(Ps) was measured using a calibrated Pirani gauge and was
controlled by a motor valve below 0.03 mbar (22.5 mTorr).
Pictures of the sample were taken in 1 s intervals by using a
digital camera system which was mounted on top of the
microscope. Pictures were then analyzed for the nucleation
(Tn), onset of collapse (Toc), and full collapse temperature
(Tfc), using the Linksys 32 software. The nucleation temper
ature Tn was determined using the rst recorded picture
illustrating a frozen structure of the measured solute. The Toc
was dened as the temperature at which rst gaps and ssures
were visible adjacent to the sublimation interface. Full
collapse (Tfc) was dened as the temperature at which a full
loss of structure (i.e., no connection between dried product
matrix and sublimation interface) could be determined.

Two excipients commonly used in freeze drying formu
lations were selected as model substances: polyvinylpyrroli
done (40 kDa), a polymer used as a cryo and lyoprotectant,
and 2 (hydroxypropyl) cyclodextrin, a commonly used
collapse temperature modier. Polyvinylpyrrolidone (further
denoted as PVP) and 2 (hydroxypropyl) cyclodextrin (fur
ther denoted as HP CD) were of highest analytical grade
(Fluka, Switzerland) and used as received. Aqueous solutions
were prepared as stock solutions with double distilled water
from an all glass apparatus (Destamat Bi 18 T, Heraeus,
Hanau, Germany) and aliquots ltered through a 0.22 m
membrane (Millipore, Billerica, MA, USA) prior to use.
Total solid contents ranged from 10 to 200 mg/g (1 20% w/w)
for PVP and from 10 to 300 mg/g (1 30% w/w) for HP CD.
Freeze-Dry Microscopy
All solutions were studied by performing up to ve FDM
measurements (n=5). The freeze dry microscope consisted of
a microscope (Axio Imager.A1, Zeiss, Germany) with a
lambda plate plus analyzer and a FDCS 196 freeze drying
stage (Linkam, Surrey, UK). The used magnication was 200
fold. Calibration of the thermocouple in the FDM stage was
performed with 10% (w/w) solutions of KCl ( 10.7C), NaCl
( 21.1C), and MgCl2 ( 33.6C) by detecting the known
melting points of the frozen solutions (4). Custom made,
precision cut spacers (height 0.025 mm) were used to
maintain a constant thickness of the sample layer. The
preparation of the samples was as follows: a 15 mm round
glass cover slide was placed on the silver block of the freezing
stage. This block was coated with a 5 L drop of silicon oil to
improve heat transfer from the silver block to the glass cover
slide. Then, about 2 L of solution were pipetted onto the
15 mm glass and subsequently covered by a second 9 mm
cover glass slide, using the spacers in between both slides.
The cooling rate used throughout this study was 10C/min.
The cooling rate in this context represents the preprogrammed
Statistical Data Analysis

The interrelation between the critical parameters of the
formulations was assessed via correlation coefcient calcu
lations using MATLAB (MathWorks, Natick, MA, USA)
which was also used for plotting all gures. In addition, driven
by the issues spotted in preliminary analysis of the experi
mental data, a validation routine known as leave one out
(L O O) has been applied. The L O O concept is often used
in multivariate analysis, e.g., in multivariate calibration
studies with near infrared spectra from various materials
(13). In short, in the L O O approach one of the samples is
left out from the rest of the samples that are used to build a
calibration model which is then used for prediction of a
variable of interest (e.g., concentration) in future samples.
The effectiveness of the obtained model is validated by
predicting the concentration of the sample that was left out,
thus L O O can be considered a validation approach. This
operation is then repeated n times, with n being the number
of samples and during each cycle, another sample is left out
and used for validation. This is a popular way to validate a
calibration model on a rather limited number of samples.
All L O O calculations in this study were performed in
MATLAB without any special routine being created. More
specic reasons for introducing the L O O approach into the
present analysis are given in the text below. The reader is
referred to the literature for more in depth details about the
L O O routine (13).
Meister, ai and Gieseler
584
Interrelation between Toc and Tfc: PVP 40 kDa and HP--CD
It is common practice to report the collapse temperature
for a given system as the onset temperature, i.e., the
temperature where the rst structural changes are visually
detected (Fig. 1a). If the sample temperature is increased
further beyond the Toc boundary, the dried area loses its
structure increasingly up until a temperature where no coherent
sample matrix forms adjacent to the sublimation interface
(Fig. 1b). The temperature interval between Toc and Tfc is
becoming of increasing interest in investigations of the temper
ature tolerance (i.e., the robustness of the formulation) during
the primary drying phase of a freeze drying cycle. Clearly, a
great difference between Toc and Tfc may be translated into a
higher temperature tolerance of the product matrix. For
example, Fonseca and coworkers reported in 2004 that a lactic
acid bacterial suspensions showed a surprisingly strong bias
between Tg and Tc (up to 10C) which implied a signicant
robustness of the formulation to temperature during the
freeze drying process (11). A statistical analysis for both PVP
40 kDa and HP CD showed that Toc and Tfc are highly
correlated for both datasets, with correlation coefcients of 0.91
for PVP and 0.98 for HP CD. For HP CD, Toc ranges
between 10.4C and 6.7C and Tfc between 9.1C and 5.8C.
Toc for PVP was found between 25.2C and 17.8C and Tfc
between 23.9C and 15.4C. The results clearly illustrate that
irrespective of the total solid content and Tn of the solutions,
the two collapse temperatures vary very similarly. This
observation is important since it implies that the temperature
interval between Toc and Tfc is based on material properties
rather than the applied measurement methodology. Based on
the investigated similarity between Toc and Tfc, there was no
need for any further involvement of Tfc; hence, it is not
addressed in the following text.
Interrelation between Tn, Toc, and Total Solid Content: PVP
40-kDa Solutions
Correlations with Tn involved were found nowhere near
as simple as the correlation between Toc and Tfc. Figure 2 and
Table I illustrate the essence of the problem with regard to
the correlation between Tn and Toc for the PVP 40 kDa data.
The experimental data appear to be broadly spread, and it is

difcult to construe any correlation that may exist in this
dataset simply because of the large variability of the data
points. The standard deviation of Tn for all the concentrations
is signicant, reaching, in the worst case, even 30% for the
10% concentration (four repeats, Table I). Such a signicant
variability might be expected in the rst instance. This
observation of variability is caused by spontaneous nucleation
of the solution which, in turn, can be controlled, neither
during an FDM experiment nor during the freezing step in a
freeze drying cycle (14). Consequently, this will cause signif
icant (inherent) variations in Tn (4). There are literature
references available which report on procedures that can
enhance homogeneity of the nucleation within a given batch
in a freeze dryer (14 16). However, these methods are mostly
limited to laboratory scale freeze dryers (14) or simply
inapplicable for commercial scale freeze drying due to the
issues of sterility and adulteration of the compound, e.g.,
adding a nucleating agent like silver iodide to the formulation
(16). Because of design and technical limitations neither of
these methods can be applied to freeze dry microscopy where
a 2 L volume of solution is frozen and examined.
In the present set of experiments with PVP 40 kDa, Tn
data were found to vary between 22C and 12C. This was
expected, based on reports of nucleation temperatures
measured in vials in a freeze dryer (14 16). Despite the noisy
appearance of the data illustrated in Fig. 2, the distribution of
the points must be more closely inspected because of the
possibility that there are outliers or systematic (but explica
ble) variations the removal of which may result in emergence
of a correlation pattern or, at least, some sort of trend. Since
there were 32 samples in the PVP study, L O O correlation
coefcient calculation was repeated 32 times with 31 samples
in each cycle. The idea behind this procedure is rather simple:
if there is a sample with a Tn/Toc value which is perceivably
different from the values for the rest of the samples, its
omission should lead to increase (in the absolute values) of
the correlation coefcient between Tn and Toc. The correla
tion coefcient for all the data displayed in Fig. 2 is 0.07
which conrms the visual impression of the lack of mutual
dependence between Tn and Toc. However, when the three
points which appear to be furthest away from the bulk of the
data (marked in brackets) are excluded, the correlation
coefcients signicantly increase to 0.37. Correlation coef
Fig. 1. Freeze dry microscopy images of a 12% (w/w) PVP 40 kDa solution: a (left) illustrates the onset of
collapse (Toc) at 19.7C (Ps 0.011 mbar), b (right) the full collapse of the same sample determined
at 19.0C (Ps 0.010 mbar)
Fig. 2. Set of Tn and Toc data for the PVP 40 kDa measurement
series at different total solid content. All data points which were
identied as outliers in the L O O analysis are marked with a bold
cross. The linear t is shown for guiding the eye; its correlation
coefcient is 0.7
cients vary between 1 and +1 with +1 denoting perfectly

synchronized increase or decrease of two variables and 1
denoting perfectly opposite variation, i.e., increase of one of
the variables is followed by decrease of the other. Repetition
of the L O O correlation coefcient routine with the remain
ing 29 samples led to further removal of samples increasing
thereby the correlation coefcient to 0.58 (ve samples
removed). The variation of correlation coefcients during this
L O O procedure is shown in Fig. 3. The third and nal L O
O iteration led to a Tn vs. Toc correlation coefcient of 0.7
(ve more samples removed). Figure 2 shows the nal Tn vs.
Toc dependence with the total of 19 remaining samples in the
dataset. The nal calculated correlation coefcient of 0.7
depicts a trend, rather than a well dened correlation. Fitting
of the remaining points was performed with linear regression
analysis rather than using a nonlinear approach. However,
from the data illustrated in Fig. 2, a quadratic t might also be
appropriate.
Table I. Tn Data for the PVP 40 kDa Solutions, Illustrated with

Increasing Total Solid Content
Concentration
(%, w/w)
Tn (average; C)
SD, Tn
Rel SD, Tn (%)
1
2
3
5
10
12
15
20
16.3
18.3
19.1
17.3
17.6
19.8
20.0
18.0
1.0
3.5
2.3
2.7
5.2
1.3
1.2
1.9
6
19
12
15
30
6
6
11
Values represent average Tn data and corresponding standard (SD)

as well as relative standard deviations (Rel SD)
585
Fig. 3. PVP 40 kDa: variation of the correlation coefcients in the

leave one out procedure after the removal of three samples from the
rst iteration. The line marks the threshold below which the samples
are considered removable. They all belong to either low or highly
concentrated solutions
The outliers removed by L O O analysis were, in most

cases, solutions of either low or high concentration. This is
valuable information for the FDM measurement methodolo
gy. Formulations with low total solid content form small holes
which can even be observed in the dried noncollapsed
structure at temperatures much lower than Toc. As a result,
reliable determination and interpretation of Toc data become
much more difcult and are dependent on the experience of
the user. As for the higher concentrations, the accurate
determination of the Toc is difcult as well, based on (1) the
relatively low amount of light provided by the light source of
the microscope, (2) the high density of the dried structure,
and, thus, 3) the limited transmission of light through the
sample which then allows less precise observation of struc
tural changes. In particular for PVP 40 kDa (i.e., a polymer),
the observed remaining lyophilized structure is very dense
compared to other excipients used in freeze drying. As a
result, Toc measurements of solutions of broadly varying
concentrations are prone to show variability and are subjective
to a signicant degree, i.e., dependent on the individual
assessment of a user performing the experiment. Consequently,
a rational removal of some points from those measurements
may be considered as justied.
In the next phase of the PVP study, Toc was correlated to
the corresponding total solid content (Fig. 4a). This correla
tion is much more straightforward since only a couple of
points were identied to be out of the main trend (with all the
measurements of the 3% sample again being outliers). Those
points are ignored in Fig. 4a; thus, a distinct quadratic
dependence of Toc vs. total solid content is obtained. It is
important to underline at this point that the Toc vs. total solid
content dependency was found to be clearly different from
the correlation between Tn and Toc. The Toc vs. total solid
content correlation plot illustrates a much stronger impact of
concentration on Toc which applies even for small variations
in concentration. Since only a weak correspondence is found
between Tn and Toc, and assuming that Toc is clearly
586
Fig. 4. PVP 40 kDa: a (left) represents an exponential t for the variation of Toc with total solid content for PVP solutions, b (right) represents
the variation of Tn with the total solid content for the PVP solutions. The correlation coefcient for these points is close to 0
correlated with the total solid content, one may expect that
there would be only a weak correlation between Tn and total
solid content. An additional support for this assumption may
be found in the work of Miyata et al. (17) who reported that
solutions of disaccharides showed highly comparable super
cooling behavior over the entire concentration range ana
lyzed. The results of the present study largely conrm these
expectations: Tn vs. total solid contents plot shows that Tn
does not depend on the concentration (Fig. 4b).
In summary, the correlation analysis of the PVP coupled
with the assessment of the quality of the data via use of L O
O routine shows that: (1) Toc increases with an increase of the
total solid content and reaches a plateau for highly concen
Fig. 5. PVP 40 kDa: 3D plot of Tn, Toc, and total solid content. Most
of the experimental points are in the plane spanned by Toc and total
solid content with Tn axis being orthogonal to that particular plane
trated (i.e., >20%) solutions, the points being best tted with
a quadratic model (Fig. 4a). (2) Only a weak and negative
correlation is observed for Toc and Tn after excessive data
treatment that eliminated signicant number of the data
points, most of them referring to extreme concentrations
(<3% and >10%). Toc measurements of those samples are
believed to be prone to errors due to the inherent weaknesses
in the concept of the measurement, and (3) Tn is found to be
independent of the total solid content of the freeze dried
solution. Additionally, a simple multivariate analysis of the
PVP dataset was performed which entirely supported the
univariate analysis described above. Multivariate analysis in
this context means that Toc is simultaneously tted with both
concentrations and the parameter Tn. It is shown that Tn is
practically independent of Toc and as such it does not
contribute to the characterization of Toc. On the other hand,
the concentrations are found to be correlated with Toc.
Figure 5 illustrates a 3 D representation of Toc, total solid
Fig. 6. HP CD: the set of Toc over Tn data for the series of
solutions of various total solid content. The original correlation
coefcient is 0.48, the one after L O O analysis is 0.85 (linear t).
Outliers are marked with a bold cross
Fig. 7. HP CD: variation of the correlation coefcients in the leave

one out procedure after one of the samples from Fig. 6 with 1%
concentration is left out. On the basis of this gure, ve more samples
are eliminated from the original dataset
content, and Tn. Here, most of the experimental points are in

the plane spanned by Toc and total solid content with Tn axis
being orthogonal to that plane.
Interrelation between Tn, Toc, and Total Solid
Content: HP--CD
The original data points for relation for 2 (hydroxy
propyl ) cyclodextrin (HP CD) are shown in Fig. 6. A
negative trend is more perceivable for these data and, again,
there are a few points that appear to be far from the bulk of
the data (Fig. 6, right hand side). At the rst glance, these
data appear to be more coherent in comparison to those in
Fig. 2, and this is also reected in the correlation coefcient
of 0.48 with all the experimental points considered. Similarly
to the previous case, the rst cycle of L O O correlation
coefcient calculations immediately identies one of the
samples with concentration of 1% as an outlier (not plotted).
587
The correlation improves to 0.59 after the elimination of
that particular sample. The second iteration of the L O O
correlation coefcient analysis, however, reveals ve more
samples which seem to spoil the trend (Fig. 7). This analysis
appears to be more convincing than that with the PVP
solutions because of the clearer distinction of the ve
identied samples in Fig. 7. With their elimination (indicated
with a bold cross in Fig. 6), the remaining total of 12 samples
shows an obviously negative correlation that again may be
tted by either a linear or a quadratic function. A linear t is
chosen in order to be more consistent with the employed
correlation terminology that better corresponds to linear
dependencies. The dispersion of the points of the low
concentration samples (that is the highest Tn and the lowest
Toc) prevents the nature of dependence to be more clearly
determined, i.e., both ts are similarly accurate.
As mentioned above, the uncertainties in the measure
ment technique are believed to be the major cause of the
dispersion of the data. For this set of solutions, ve out of
the six removed points deviate quite signicantly from the
samples with similar concentrations, or replicate measure
ments for the same concentration show a high variance. For
example, the three Tn measurements for a 1% (w/w)
concentration of HP CD vary from 11C to 18C which
spans almost the entire range of Tn values for all the
evaluated concentrations. The removed data points that were
obtained from the 10% solutions have an unreasonably high
Tn that places the corresponding Toc data points to the far
right hand side in the plot. This may even appear to be a
systematic error for that particular sample. Finally, Tn values
for the 20% solutions vary between 16C and 21C which
is a signicant relative deviation, keeping in mind that the
average value would be 18.5C. The results obtained for HP
CD, hence, strengthen the observation made for the PVP data:
the measurements of solutions with low and high concentrations
seem to be rather unreliable or irreproducible so that it is
difcult to ascertain exact correlation between Toc and Tn. For
the HP CD data, total of four outliers belong to those
boundary concentrations (two from 1% and two from 20%).
Figure 8a, b illustrates correlations between Toc and Tn vs.
total solid content, respectively. These two plots essentially
Fig. 8. HP CD: a (left) shows variations of Toc over total solid content for HP CD solutions and reveals a strong linear correlation, b (right)
illustrates Tn vs. total solid content for HP CD solutions where the correlation coefcient for these points is close to 0
588
concur with the corresponding correlations for the PVP data
(Fig. 4a, b). The correlation between Toc and the corre
sponding concentrations is unquestionable, although in this
particular case, it appears to be more linear than quadratic
(two obviously outlying points were excluded to produce
Fig. 6). Tn again appears to be independent of total solid con
tent (Figs. 4a, 8a).
To summarize the results for the HP CD experiments,
Tn clearly exerts more impact on Toc than for the PVP 40 kDa
solutions but still has a marginal effect when compared with
the effect of variation in concentration on Toc. PVP is a mixture
of molecules with different chain lengths while HP CD has a
dened chemical structure. Hence, for PVP higher variations
in Toc (and Tfc) were expected reecting more variations in the
composition of the frozen structure observed. Because of these
higher variations in Toc for PVP 40 kDa, Tn has a higher impact
on Toc for HP CD.
CONCLUSION
The results obtained during this study suggest that Tn has
an inuence on the frozen structure and consequently on the
corresponding Toc and Tfc. Since Tn cannot be controlled
during FDM measurements, a user should consider (at least)
three to ve individual measurement of a single solution
(resulting in different Tn), and only the mean value out of
those measurements should be taken into account for further
cycle development and optimization. Signicant deviations in
Tn are probably unavoidable. Furthermore, low (<3% w/w)
and high (>10% w/w for PVP 40 kDa, >20% w/w for 2
(hydroxypropyl ) cyclodextrin) total solid contents are
found to be less reproducible and much more prone to
outliers. For solutions with a low total solid content, the holes
in the dried structure perturb the possibility for accurate
detection, while for highly concentrated solutions, the high
density of the dried area causes observation problems of the
collapse. These two effects are essentially thought to be
behind the outliers of Tn vs. Toc correlations. Finally, there is
a clear correlation between Toc and the corresponding
concentrations. For PVP 40 kDa solutions, quadratic relation
of Toc over concentration is found, whereas a linear correla
tion is determined for HP CD. These dependences may be
useful for further optimizations of formulations by varying the
contents of the excipients. The existing data are rather well
tted by both linear and nonlinear functions, so it remains
unclear (at this point) which of these two is recommendable.
REFERENCES
1. Franks F. Freeze drying of bioproducts: putting principles into
practice. Eur J Pharm Biopharm. 1998;45:221 9.
2. Abduhl Fattah A, Kalonia DS, Pikal MJ. The challenge of
drying method selection for protein pharmaceuticals: product
quality implications. J Pharm Sci. 2006;96(8):1886 916.
3. Schwegman J, Hardwick L, Akers M. Pratical formulation and
process development of freeze dried products. Pharm Dev Tech.
2005;10:151 73.
4. Pikal MJ, Shah S. The collapse temperature in freeze drying:
dependence on measurement methodology and rate of water
removal from the glassy phase. Int J Pharm. 1990;62:165 86.
5. Passot S, Fonseca F, Alarcon Lorca M, Rolland D, Marin M.
Physical characterization of formulations for the development of
two stable freeze dried proteins during both dried and liquid
storage. Eur J Pharm Biopharm. 2005;60:335 48.
6. Bellows RJ, King CJ. Freeze drying of aqueous solutions: maxi
mum allowable operating temperature. Cryobiology 1972;9:559 61.
7. Passot S, Fonseca F, Barbouche N, Marin M, Alarcon Lorca M,
Rolland D, et al. Effect of Product temperature during primary
drying on the long term stability of lyophilized proteins. Pharm
Dev Technol. 2007;12:543 53.
8. Meister E, Gieseler H. Evaluation of collapse temperatures by
freeze dry microscopy: impact of excipient concentration on
measured transition and the overall dependence on measure
ment methodology. Proc. 5th World Meeting on Pharmaceutics
and Pharmaceutical Technology, Geneva, March 27 30, 2006.
9. Searles J, Carpenter J, Randolph T. The ice nucleation
temperature determines the primary drying rate of lyophilization
for samples frozen on a temperature controlled shelf. J Pharm
Sci. 2001;90(7):860 71.
10. Searles J. Freezing and annealing phenomena in lyophilization.
In: Rey L, May J, editors. Freeze drying/lyophilization of
pharmaceutical and biological products. New York: Marcel
Dekker; 2004. p. 109 45.
11. Fonseca F, Passot P, Cunin O, Marin M. Collapse temperature of
freeze dried lactobacillus bulgaricus suspensions and protective
media. Biotechnol Prog. 2004;20:229 38.
12. Meister E. Methodology, data interpretation and practical
transfer of freeze dry microscopy. PhD Thesis, University of
Erlangen Nuremberg, Erlangen, Germany, 2009.
13. Vandenginste BGM, Massart DL, Buydens LMC, de Jong S,
Lewi PJ, Smeyers Verbeke J. Handbook of Chemometrics and
Qualimetrics B, vol 3. Amsterdam: Elsevier; 1998. p. 87 160.
14. Rambhatla S, Ramot R, Bhugra C, Pikal MJ. Heat and mass
transfer scale up issues during freeze drying: II. Control and charac
terization of the degree of supercooling. AAPS PharmSciTech.
2004;5(4), e58:1 9.
15. Pikal MJ. Freeze drying. In: Swarbrick J, Boylan JC, editors.
Encyclopaedia of Pharmaceutical Technology. New York: Marcel
Dekker; 2002. p. 1299 326.
16. Pikal MJ, Costantino HR. Lyophilization of biopharmaceuticals.
In: Borchardt RT, Middaugh CR, editors. Biotechnology:
Pharmaceutical Aspects. Arlington: AAPS; 2004. p. 113 38.
17. Miyata K, Kanno H. Supercooling behavior of aqueous solutions
of alcohols and saccharids. J Mol Liquids 2005;119:189 93.

DOI: 10.1208/s12249 009 9251 0
Research Article
Microemulsions Containing Medium-Chain Glycerides as Transdermal Delivery
Systems for Hydrophilic and Hydrophobic Drugs
Jaclyn Hosmer,1 Rachel Reed,1 M. Vitria L.B. Bentley,2 Adwoa Nornoo,1,3 and Luciana B. Lopes1,4
Abstract. We evaluated the ability of microemulsions containing medium chain glycerides as penetration
enhancers to increase the transdermal delivery of lipophilic (progesterone) and hydrophilic (adenosine)
model drugs as well as the effects of an increase in surfactant blend concentration on drug transdermal
delivery. Microemulsions composed of polysorbate 80, medium chain glycerides, and propylene glycol
(1:1:1, w/w/w) as surfactant blend, myvacet oil as the oily phase, and water were developed. Two
microemulsions containing different concentrations of surfactant blend but similar water/oil ratios were
chosen; ME lo contained a smaller concentration of surfactant than ME hi (47:20:33 and 63:14:23
surfactant/oil/water, w/w/w). Although in vitro progesterone and adenosine release from ME lo and ME
hi was similar, their transdermal delivery was differently affected. ME lo signicantly increased the ux
of progesterone and adenosine delivered across porcine ear skin (4 fold or higher, p<0.05) compared to
progesterone solution in oil (0.050.01 g/cm2/h) or adenosine in water (no drug was detected in the
receptor phase). The transdermal ux of adenosine, but not of progesterone, was further increased (2
fold) by ME hi, suggesting that increases in surfactant concentration represent an interesting strategy to
enhance transdermal delivery of hydrophilic, but not of lipophilic, compounds. The relative safety of the
microemulsions was assessed in cultured broblasts. The cytotoxicity of ME lo and ME hi was
signicantly smaller than sodium lauryl sulfate (considered moderate to severe irritant) at same
concentrations (up to 50 g/mL), but similar to propylene glycol (regarded as safe), suggesting the
safety of these formulations.
KEY WORDS: adenosine; medium chain glycerides; microemulsion; progesterone; transdermal delivery.
INTRODUCTION
An efcient transdermal delivery of compounds is
generally difcult to achieve due to the barrier function of
the skin, provided mainly by the highly organized structure of
the skins outermost layer, the stratum corneum (SC) (1).
This explains why, in spite of the many advantages of topical
and transdermal administration of drugs, there are still few
products commercially available for these routes. Among
many different strategies and formulations studied to over
come the barrier function of the stratum corneum and
increase the skin penetration of drugs, the use of micro
emulsions has generated considerable interest over the past
years (2,3). Microemulsions present multiple advantages over
other dermatological formulations, including (a) thermody
namic stability, (b) ease of preparation, (c) possibility to
incorporate both hydrophilic and lipophilic drugs (at the same
1
Department of Pharmaceutical Sciences, Albany College of Phar

macy and Health Sciences, 106 New Scotland Ave., Albany, New
York 12208, USA.
2
Faculdade de Cincias Farmacuticas de Ribeiro Preto, Universi
dade de So Paulo, Ribeiro Preto, So Paulo, Brazil.
3
Gregory School of Pharmacy, Palm Beach Atlantic University, Palm
Beach, Florida, USA.
4
To whom correspondence should be addressed. (e mail: Luciana.
lopes@acphs.edu)
time, if desirable), and (d) possible skin penetration enhancing

ability. This last property may be related not only to the
structure of the system (small droplet size associated with a
high surface area) but also to the possibility of incorporating
skin penetration enhancers in the oil phase and/or in the
surfactant blend (3,4).
The ability of medium chain mono and diglycerides
(which will be referred to as MCG throughout the text) to
enhance the absorption of drugs through the intestinal
mucosa has been demonstrated (5,6). This effect is associated
with the ability of MCG to interact with membrane lipids and
proteins, increasing membrane permeability (6,7). Although
the skin differs from the intestinal mucosa in many aspects,
some compounds that increase the intestinal permeability
have also been demonstrated to increase the skin penetration
of drugs (1,5,6). Hence, the incorporation of MCG in micro
emulsions has the potential to enhance the efcacy of such
formulations as transdermal delivery systems. In this context,
the rst goal of this study was to evaluate the ability of
microemulsions containing MCG as part of the surfactant
blend to improve transdermal delivery of lipophilic (proges
terone (PGT), logP = 4.04) and hydrophilic (adenosine
(ADN), logP= 0.05) model drugs (8,9).
It is well known that the structure and composition of
microemulsions play an important role on the transdermal
delivery of drugs. Many studies have demonstrated the
589
Hosmer et al.
590
inuence of the type of oil, of surfactant/cosurfactant ratio,
and of the microemulsion structure (o/w versus w/o) on drug
release and skin penetration (2,10,11). Less studied, however,
is the effect of the concentration of the surfactant blend used. In
the face of these facts, the second goal of this study was to
evaluate whether and how the concentration of the surfactant
blend (and thus, the concentration of MCG) inuences the
transdermal delivery of the model drugs. Because of the
importance of developing systems with low irritation potential,
the third goal of this study was to evaluate the relative safety of
the microemulsions containing different amounts of the surfac
tant blend by evaluating their concentration dependent effects
on the viability of cultured broblasts in comparison to those of
propylene glycol (a commonly used compound in topical
formulations) and of sodium lauryl sulfate (considered a
moderate to severe irritant).
Materials
Propylene glycol, Polysorbate 80 (P80), PGT, and ADN
were obtained from Sigma (St. Louis, MO, USA). Medium
chain mono and diglycerides (Capmul) were a kind gift from
Abitec Corporation (Janesville, WI, USA). Acetonitrile and
methanol were purchased from Mallinckrodt Baker (Phillips
burg, NJ, USA) and myvacet 9 45 oil (diacetylated mono
glycerides from soybean oil; for the sake of simplicity, we will
refer to this component solely as oil) was obtained from
Quest (Norwich, NY, USA).
Methods
Pseudoternary Phase Diagram Construction and Sample
Preparation
Ternary phase diagrams were constructed using the
water titration method at room temperature. Because of its
penetration enhancing potential, MCG was chosen as one of
the surfactants of the microemulsions. Since MCG is very
lipophilic (hydrophilic lipophilic balance (HLB)=5 6, which
allows its use as surfactant and as oil solvent) (6,12), we
combined MCG with a more hydrophilic surfactant, polysor
bate 80 (HLB=15), in the surfactant blend so that we could
increase water incorporation and obtain o/w systems (5). In
selected systems, propylene glycol was used as cosurfactant
due to its ability to increase water incorporation in P80 based
microemulsions (2). As a result, four different surfactant
blends were used in the preparation of the microemulsion:
P80/MCG (1:1 w/w), P80/MCG/propylene glycol (3:3:1 w/w/w),
P80/MCG/propylene glycol (1:1:1 w/w/w), and P80/MCG/
propylene glycol (1:2:1 w/w/w). The oil phase (myvacet oil)
was added to the surfactant blend at ratios varying from 1:9 to
9:1 (w/w, surfactant blend/oil). These mixtures were titrated
with water under vortexing, and the systems were rst
characterized using visual inspection to determine phase
separation, uidity, and transparency. Formulations that were
uid, clear, and did not undergo phase separation were
classied as microemulsions.
Two microemulsions were chosen and subjected to further
characterization: ME lo (47:20:33 surfactant/oil/water, w/w/w)
contained a smaller concentration of surfactant blend than

ME hi (63:14:23 surfactant/oil/water, w/w/w). These formula
tions were chosen because they have different concentration of
the surfactant blend but similar water/oil ratios. PGT was
incorporated in these microemulsions at a nal concentration
of 1% (w/w, as this concentration was used in previous
transdermal delivery systems), whereas ADN was incorporat
ed at 0.5% (w/w).
Microemulsion Characterization
Since microemulsions are isotropic systems by denition
(i.e., they present same optical properties when probed in all
directions), ME lo and ME hi were characterized using a
polarized light microscope (Axiotop, Zeiss, Thornwood, NY,
USA). Light scattering assays were performed to measure the
droplet size, and a Zetasizer nano series instrument (Zetasizer
Nano Series, Malvern, Westborough, MA, USA) was used for
this aim. The mean droplet size for each microemulsion was
determined at room temperature. The electrical conductivity of
the two selected microemulsions was measured using the same
equipment (Zetasizer Nano Series, Malvern, Westborough,
MA, USA) and compared to the conductivity of the oil phase
or water separately to determine whether the microemulsions
are oil continuous or water continuous (4).
In Vitro Skin Permeation Assay
Transdermal delivery of PGT and ADN were studied in
vitro using porcine ear skin mounted on Franz diffusion cells.
Briey, the skin from the outer surface of a freshly excised
porcine ear was carefully removed, stored at 20C, and used
within a month. On the day of the experiment, the skin was
thawed and mounted in a Franz diffusion cell (diffusion area
of 1 cm2; Laboratory Glass Apparatus, Berkeley, CA, USA),
with the SC facing the donor compartment and the dermis
facing the receptor compartment. The test formulation
(100 mg) was placed in the donor compartment. The
receptor compartment of the cells was lled with phosphate
buffer (pH 7.4, 100 mM) maintained at 370.5C with
magnetic stirring at 350 rpm throughout the assay. The
receptor phase contained 20% of propylene glycol when the
delivery of PGT was studied; previous studies have used
receptor phases containing up to 50% of propylene glycol
without compromising the skin barrier (13). ME hi, ME lo, or
control formulations (myvacet oil containing 1% PGT or
water containing 0.5% ADN) were allowed to interact with
the skin for 2, 4, 6, 8, or 12 h, after which the experiment was
terminated for sample collection. To determine whether any
of the surfactants have an effect of its own on the skin
penetration of the model drugs, formulations of MCG in
propylene glycol (15%, w/w) and P80 in propylene glycol
(15%, w/w) were also studied; these formulations contained
ADN or PGT at the same concentrations as the others, and
they were left in contact with the skin for 8 h.
At the end of the experiment, skin samples were rinsed
to remove excess formulation. The tape stripping technique
was used to separate SC from the epidermis (E) and dermis
(D); 15 pieces of tape were used, and the pieces were placed
in conical tubes containing 5 mL methanol (for PGT) or 5 mL
of water/methanol (1:1 v/v for ADN). The remaining skin
Microemulsions for Transdermal Delivery

(viable epidermis + dermis, E+D) was cut in small pieces,
placed into conical tubes containing 2 mL methanol (for
PGT) or water/methanol (1:1 v/v for ADN), and homoge
nized using a hand held tissue homogenizer (Biospec prod
ucts, Bartlesville, OK, USA). The SC and [E+D] samples
were then sonicated for 20 min, ltered through a 0.45 m
pore membrane and assayed for PGT and ADN. Aliquots of
the receptor phase were collected, ltered, and assayed for
the drugs. The concentrations of drugs in SC and [E+D] are
indices of topical delivery, whereas the concentration in the
receptor phase is an index of transdermal delivery. The ux
of drug across the skin was calculated using linear regression
analysis; the amount of drug delivered across the skin was
plotted as a function of time, and the slope of the linear
portion of the curve was determined.
In Vitro Drug Release from Microemulsions
The release of PGT and ADN from ME lo and ME hi
was assessed using the Franz diffusion cells and the setup
described in In Vitro Skin Permeation Assay section, except
that a cellulose membrane was used instead of the skin.
Samples of the receptor phase were collected at 2, 4, 8, and
12 h postapplication, ltered and assayed for PGT and ADN.
591
trile/water at 8:2 (v/v, ow rate of 1 mL/min) and detection
wavelength of 240 nm.
Standard solutions of PGT were prepared in methanol or in
propylene glycol/phosphate buffer (1:4, w/w), whereas solutions
of ADN were prepared in methanol/water (1/1, v/v) or
phosphate buffer. Calibration curves of PGT in methanol or
ADN in methanol/water were used to assay the amount of
drugs in the skin; linearity was observed over 0.03 500 g/mL
(r2 =0.999). Calibration curves of PGT in propylene glycol/
phosphate buffer or ADN in phosphate buffer were used to
assay the drugs in the receptor phase of the diffusion cell;
linearity was achieved over 0.03 100 g/mL (r2 =0.999). The
HPLC method for PGT and ADN were reproducible with
within day and between day variations of less than 10%.
Recovery of PGT and ADN from Skin Samples. To
standardize the recovery of PGT and ADN from skin tissue,
skin sections (1.0 cm2) were spiked with 1, 10, 20, and 50 g
of PGT (in a methanolic solution) or 0.5, 1, and 10 g of
ADN (in a water/methanol solution). Fifteen minutes later,
the drugs were extracted from the skin sections as described
in the in vitro permeation studies. PGT recovery was linear
over the concentration range of 1 50 g/cm2 (r2 =0.997) and
varied from 88% to 92%. ADN recovery was linear over the
range of 0.5 10 g/cm2 (r2 =0.988) and varied from 72% to 80%.
Evaluation of Electrical Resistance of Skin

To evaluate the effect of the microemulsions on the
barrier function of the skin, the electrical resistance of this
tissue was measured before and after application of water
(control), ME lo or ME hi using a LCR multimeter (Mod
5300, accuracy 0.8%, Sperry Instruments, Milwaukee, WI,
USA). Skin samples were mounted in diffusion cells, and the
donor and receptor compartments were lled with phosphate
buffered saline (PBS); after 20 min of equilibration, the
electrodes were inserted in the donor and receptor compart
ments for measurement of baseline skin resistance (14,15).
Immediately thereafter, PBS in the donor compartment was
replaced with 100 mg of water (control), ME lo or ME hi for
8 h. This time point was chosen since most of the differences
between ME lo and ME hi skin penetration enhancement
were observed after 8 h. By the end of the experiment, skin
samples were washed with water and blotted dry. The donor
compartment was then relled with PBS, and electrical
resistance was measured. The change in electrical resistance
( resistance) was calculated by subtracting the values of
resistance measured 8 h after treatment from the baseline
resistance values.
Quantification of Drugs
HPLC Methodology. Both drugs were assayed by high
performance liquid chromatography (HPLC) using an equip
ment consisting of a Waters 600 controller, a Waters 717plus
autosampler, and a Waters 996 photodiode array detector. The
separation was performed by a Prevail C 18 column (5 m),
equipped with a C 18 precolumn. ADN was assayed using a
mobile phase composed of water/acetonitrile at 9:1 (v/v, ow
rate of 1 mL/min) and detection wavelength of 260 nm, whereas
PGT was assayed using a mobile phase composed of acetoni
Evaluation of Cellular Viability

To evaluate the relative safety of the microemulsions, we
compared their cytotoxic effects to those of propylene glycol
(a commonly used compound in topical formulations) and
sodium lauryl sulfate (considered a moderate to severe
irritant) (16,17). Murine Swiss 3T3 mouse broblasts were
obtained from American Type Culture Collection (ATCC;
Manassas, VA, USA) and grown at 37C and 5% CO2
atmosphere in Dulbeccos modication of Eagles medium
(ATCC, Manassas, VA, USA) containing 10% fetal bovine
serum (Gibco, Carlsbad, CA, USA) and additional penicillin
and streptomycin (1%). For the cellular viability assay, cells
were plated in 96 well plates (6,000 cells/well) and treated for
24 h with either PBS, propylene glycol, ME lo, ME hi, or
sodium lauryl sulfate at concentrations ranging from 1 to
500 g/mL in cell culture medium.
Cell survival was evaluated using a cell proliferation
assay reagent (CellTiter 96 Aquous One solution, Promega,
Madison, WI, USA) consisting of 3 (4,5 dimethylthiaziazol 2
yl) 5 (3 carboxymethoxyphenyl) 2 (4 sulfophenyl) 2H tetra
zolium salt (MTS) and an electron coupling reagent. The
MTS salt is reduced to a colored formazan product, and the
amount of this product is directly proportional to the number
of living cells. After treatment with PBS, propylene glycol,
ME lo, ME hi, or sodium lauryl sulfate, cells were washed
with PBS, and 100 L of cell culture medium plus 20 L of
the cell proliferation assay reagent were added to each well.
The plates were incubated for 2 h at 37C in a humidied
atmosphere of 5% CO2 in an incubator, and the absorbance
was recorded at 490 nm using a plate reader (SpectraMax,
Molecular Devices Corp., Sunnyvale, CA, USA). These
experiments were performed in triplicate using cells between
passages 2 and 6.
Hosmer et al.
592
Statistical Analyses
The results are reported as means standard deviation.
As in previous skin penetration studies (16), data were
statistically analyzed using nonparametric tests. The Krus
kal Wallis test (followed by Dunns post hoc test) was used to
compare more than two experimental groups. Values were
considered signicantly different when p<0.05.
RESULTS
Microemulsion Development and Characterization
Phase diagrams representing the phase behavior of
mixtures containing different amounts of surfactant, oil, and
water are depicted in Fig. 1.When P80 and MCG were
combined as the surfactant blend, the area of existence of
microemulsions (black shaded area) corresponded to 23% of
the phase diagram (Fig. 1a). Addition of propylene glycol
increased the amount of incorporated water (Fig. 1b d).
When the ratio between P80/MCG/propylene glycol was 3:3:1
(w/w/w), the area of microemulsion existence was slightly
increased, but transparent and highly viscous systems were
observed in the phase diagram (gray shaded area, Fig. 1b).
Further increase in propylene glycol (P80/MCG/propylene
glycol at 1:1:1, w/w/w) abolished the existence of such
systems, and the area of microemulsion was further increased
(Fig. 1c). An increase in the amount of MCG in the surfactant
blend (P80/MCG/propylene glycol at 1:2:1, w/w/w) decreased
the amount of water that could be incorporated and the size
of the area of microemulsion existence (Fig. 1d). Based on
these results, a surfactant blend containing P80/MCG/propyl
ene glycol at 1:1:1 (w/w/w) was chosen. Addition of either
drug did not change water incorporation or the size of the
area of microemulsions existence.
The microemulsions chosen for further characterization

were ME lo and ME hi, which contained ratios of surfactant
blend/oil/water equal to 47:20:33 and 63:14:23 (w/w/w),
respectively (Fig. 1c; Table I). The physicochemical character
istics of ME lo and ME hi are depicted in Table I. The
droplet size of the selected microemulsions was 25.2 and
21 nm (for ME lo and ME hi, respectively; Table I). Both
microemulsions were isotropic when observed under a
polarized light microscope, with no specic texture being
observed even after drug incorporation. As can be observed
in Table I, the conductivity of ME lo and ME hi were closer
to the conductivity of water than that of oil, suggesting that
the systems are oil in water microemulsions (18). This is in
accordance with the structure formed based on the HLB
values of the surfactants used: The calculated HLB for ME lo
and ME hi is 10, which generally gives rise to oil in water
structures (5). The HLB may be even higher since glycols can
increase the effective HLB of nonionic surfactants (19).
In Vitro Skin Permeation Assay
Figure 2 shows the skin penetration and transdermal
delivery of PGT. Compared to the control formulation (drug
solution in the oil), both ME lo and ME hi signicantly (p<
0.05) increased topical (SC and [E+D]) and transdermal PGT
delivery after periods of application as short as 4 h: enhance
ments of 3 and 8 fold on the amount of PGT delivered to
the whole skin (sum of SC and [E+D]) and across this tissue,
respectively, were observed at this time point. Comparing
ME lo and ME hi, there was no signicant increase in the
amount of PGT delivered to the receptor phase or in PGT
ux across the skin (Table II) using ME hi with respect to
ME lo during the studied time period; however, signicantly
higher drug retention in the SC was observed using ME hi at
12 h postapplication.
Both microemulsions also increased ADN delivery into
and across the skin compared to the control solution. However,
unlike PGT, ADN penetration into [E+D] and across the skin
was higher using ME hi than ME lo at 8 h postapplication and
up. After 8 h, ADN delivery to [E+D] was increased from
undetected (using the control solution, Fig. 3b) to 0.190.03 and
0.35 0.04 g/cm2 using ME lo and ME hi (Fig. 3e f),
respectively. Similarly, the transdermal delivery of ADN was
increased from undetected (using control solution, Fig. 3j) to
0.110.07 and 0.300.08 g/cm2 using ME lo and ME hi,
respectively (Fig. 3k l). ADN ux across the skin was two
times higher using ME hi than ME lo (Table II).
In order to determine whether the surfactants affected
the skin penetration of model drugs, solutions of MCG or P80
Table I. Characteristics of the Microemulsions
Formulation
Fig. 1. Ternary phase diagrams. The black shaded areas in a d
represent regions where microemulsions are formed, whereas the
gray shaded area in b represents a region of clear but viscous systems.
Points in c represent the composition of ME lo and ME hi
ME lo
ME hi
Water
Oil
Composition
(surfactant/oil/
water, w/w/w)
Diameter (nm)/
polydispersity
Conductivity
(S/cm)
47:20:33
63:14:23
25.2/0.27
21/0.30
75.4
96.2
80.2
13.3
593
suggesting that MCG may not be the only factor responsible
for the penetration enhancing effect of the microemulsion.
In Vitro Drug Release from Microemulsions
Having demonstrated that skin penetration into and
across the skin can be inuenced differently by ME lo and
ME hi (especially the delivery of ADN), we next evaluated
whether these results are related to differences in drug
release from these formulations. Figure 5 shows the release
proles of PGT and ADN from the two microemulsions
studied. The cumulative amount of PGT or ADN released
from either ME lo or ME hi was plotted as function of time,
and a linear relationship was observed for both drugs and
both systems. No difference was observed on the release
prole of either PGT or ADN when ME lo and ME hi were
used, suggesting that the observed differences in delivery of
Fig. 2. PGT delivery to the skin (topical delivery) and across this
tissue (transdermal delivery) using the two different microemulsions
compared to the control formulation (drug solution in oil) as a
function of time. a c PGT penetration in the SC; d f PGT
penetration in the viable skin layers (E+D); g i PGT penetration in
the whole skin (SC + [E+D]); j l transdermal delivery. *p<0.05
compared to the control formulation, #p<0.05 compared to ME lo
in propylene glycol were prepared and their effect on the

topical and transdermal delivery of PGT and ADN was
evaluated (Fig. 4). MCG, but not P80, signicantly increased
the delivery of PGT and ADN to the SC (3 and 1.5 fold,
respectively), [E+D] (3 and 1.5 fold, respectively), and to the
receptor phase (5 and 2 fold, respectively). This enhance
ment was smaller than that caused by the microemulsions,
Table II. Flux of PGT and ADN Across the Skin

Formulation
PGT ux (g/cm2/h)
ADN ux (g/cm2/h)
Control
ME lo
ME hi
0.0580.008
0.2760.051
0.2620.057
0.0310.002
0.0610.001
PGT progesterone, ADN adenosine
Fig. 3. ADN delivery to the skin (topical delivery) and across this
tissue (transdermal delivery) using the two different microemulsions
compared to the control formulation (drug solution in oil) as a
function of time. a c ADN penetration in the SC; d f ADN
penetration in the viable skin layers (E+D); g i ADN penetration
in the whole skin (SC + [E+D]); j l transdermal delivery. *p<0.05
compared to the control formulation, #p<0.05 compared to ME lo
Hosmer et al.
594
Fig. 4. Delivery of PGT or ADN to the stratum corneum (SC), viable skin layers (E+D),
and across the skin using 15% MCG or polysorbate 80 (P80) in propylene glycol after 8 h.
Control: solution of PGT (1%, w/w) or ADN (0.5%, w/w) in propylene glycol. *p<0.05
compared to the control formulation
the drugs to and across the skin do not result from different
drug release from ME lo compared to ME hi.
Evaluation of Electrical Resistance of Skin
Because penetration enhancers (and delivery systems
containing such compounds) can reversibly decrease the skin
barrier function and its electrical resistance as function of
their concentration, we studied the inuence of ME lo and

ME hi treatment in the electrical resistance of the skin (20).
The results are depicted in Fig. 6. Baseline measurements of
skin resistance varied from 11.0 to 15.8 k/cm2, which is
consistent with previous reports (14). Compared to treatment
with water (control), ME lo and ME hi promoted a 2.5 and
4.2 fold (p< 0.05) decrease in skin electrical resistance,
respectively, suggesting that microemulsion treatment
decreased the barrier function of the skin. Additionally,
there was a signicant difference (p<0.05) between ME lo
and ME hi, suggesting that the ME hi induced barrier
disruption was stronger.
Evaluation of Cellular Viability
Fig. 5. Cumulative in vitro release of a progesterone and b adenosine

from ME lo and ME hi as function of time
Because of the importance of developing topical for

mulations with low toxicity, the cytotoxic potential of ME lo
and ME hi was assessed in comparison to that of propylene
glycol and sodium lauryl sulfate. Whereas propylene glycol is
considered safe and is widely used in topical formulations,
sodium lauryl sulfate is considered a moderate to severe
irritant (16,17). As expected, broblasts viability (expressed
as percent of control, i.e., untreated cells) was not affected
when increasing amounts of PBS (up to 500 g/mL) were
added to the culture medium (Fig. 6). Compared to PBS, a
signicant decrease in cell viability was observed when
sodium lauryl sulfate was used at a concentration as small as
1 g/mL, whereas the same concentration of either the
microemulsions or propylene glycol produced no signicant
effect on cell viability. At a concentration of 50 g/mL or
higher, both microemulsions signicantly (p<0.05) reduced
cell viability, as did propylene glycol. However, the cellular
viability after treatment with ME hi was signicantly smaller
than after treatment with either ME lo or propylene glycol,
Fig. 6. Effect of treatment with water, ME lo, and ME hi for 8 h on

skin electrical resistance. electrical resistance resistance values
after 8 hbaseline resistance values; *p<0.05 compared to water; #p<
0.05 compared to ME lo
but still signicantly higher compared to sodium lauryl sulfate

(Fig. 7).
DISCUSSION
One of the requirements for microemulsion formation is
the existence of a low surface tension at the oil water
interface, achieved by the use of surfactant blends and
cosurfactants. Addition of propylene glycol to P80 and
MCG blends increased the area of existence of microemul
sions in a concentration dependent manner. This is in
accordance with other studies showing that combination of
P80 and propylene glycol increases the amount of water
incorporated in microemulsions (2). The microemulsions ME
lo and ME hi were selected as delivery systems for the model
drugs since they have different concentrations of the surfac
tant blend but similar water/oil ratios. Both systems were
clear, uid, and isotropic with a droplet size in order of
nanometers.
ME lo (lower surfactant concentration) and ME hi (higher
surfactant concentration) increased the skin penetration and
transdermal delivery of PGT and ADN compared to control
solutions of the drugs. This may be attributed to the possibility
of larger drug transfer from the system to the skin due to the
larger surface area of microemulsions (which is associated with
the low interfacial tension and the small droplet size) (2,3). Our
results also demonstrated that the penetration enhancing effect
of ME hi was signicantly higher than ME lo at 8 12 h
postapplication, but affected PGT and ADN delivery different
ly. A possible explanation for the superiority of ME hi may rely
on differences in the internal structure of the two micro
emulsions, which leads to different drug release from the
formulations and, consequently, different skin penetration
(2,11,21). This possibility, however, is not supported by our
observation that both microemulsions provided similar release
proles of PGT and ADN.
Another possible explanation relates to the presence of
penetration enhancer, its concentration difference between
ME lo and ME hi and its effect on skin barrier (22,23). MCG,
but not P80, signicantly increased the skin penetration and
transdermal delivery of the model drugs. Therefore, it is
reasonable to suggest that MCG plays a role on the
595
penetration enhancing effect of the microemulsions. The
reduction of skin electrical resistance after treatment with
ME hi for 8 h was higher than after treatment with ME lo,
suggestion that the ME hi induced barrier disruption was
stronger. This is consistent with increases in the skin
penetration and transdermal delivery of ADN using ME hi,
but not with PGT retention within supercial skin layers,
suggesting that other factors (beside skin permeability)
inuence penetration of PGT. We have previously demon
strated that MCG concentration plays a major role in topical
versus transdermal delivery of a lipophilic drug (7). Com
pared to a solution of PGT in propylene glycol, MCG at 10%
enhanced the topical and transdermal delivery of PGT by 2.5
and 7 fold, respectively. MCG concentrations higher than 10%
further increased PGT retention in the skin but not its
transdermal delivery. This effect was also observed with other
lipophilic drugs and monoglycerides and attributed to the
afnity between them (24,25). Therefore, increased PGT
retention in the skin using ME hi may be a result of its afnity
with MCG present at a higher concentration. The reason why
the difference between ME lo and ME hi became signicant
only at later time points is not clear, but it was observed in other
studies investigating microemulsions as transdermal delivery
systems (2,18). This observation might suggest that the effects of
the microemulsions (and their components) are cumulative and,
as such, are more easily detected at later time points.
Cell cultures have been widely used to evaluate irritation
potential of formulations and their components (26,27). A
good correlation between in vitro cytotoxicity assays and in
vivo skin irritation has been demonstrated for surfactants of
different irritation potential, and since then, cytotoxicity
assays became largely used to predict the irritation potential
of substances (28). Although this method does not determine
the exact concentration of a substance that may be toxic to
the skin (since it does not mimic the complex structure of the
skin), it allows comparing the cytotoxic potential of new
formulations to that of compounds considered safe or irritant
(26). Similar levels of cellular viability were observed after
treatment with propylene glycol or the tested microemulsions
at 1 g/mL. However, when the concentration of formulations
was increased to 50 g/mL, the cellular viability after
Fig. 7. Concentration dependent effect of PBS, propylene glycol,

sodium lauryl sulfate (SLS), ME hi, and ME lo on the viability of
broblasts. Each point represents means standard deviation of
three replicates. *p<0.05 compared to PBS, #p<0.05 compared to
propylene glycol
Hosmer et al.
596
treatment with ME hi was signicantly lower than after
treatment with either ME lo or propylene glycol, but still
signicantly higher compared to sodium lauryl sulfate. Based
on these cytotoxicity results, three conclusions can be drawn.
First, both microemulsions studied are safer compared to
sodium lauryl sulfate. Second, enhancement of surfactant
concentration in the microemulsion was associated with an
increased cytotoxicity. However, it should be pointed out that,
for the hydrophilic compound studied, the enhancement of
surfactant concentration is also associated with a signicant
improvement of transdermal drug delivery. Third, the simi
larity between the irritant potential of the microemulsions
(especially ME lo) and propylene glycol suggest that selected
microemulsions may be safe transdermal delivery systems.
9.
10.
11.
12.
13.
CONCLUSION
In conclusion, microemulsions containing MCG can be
considered effective transdermal delivery systems for hydro
philic and lipophilic drugs. The concentration of the surfac
tant blend affected differently the permeation of the studied
hydrophilic and lipophilic drugs across the skin: The trans
dermal delivery of the hydrophilic drug, but not of the
lipophilic one, was signicantly enhanced when the concen
tration of the surfactant blend was increased.
14.
15.
16.
ACKNOWLEDGMENTS
17.
The authors would like to thank Dr. Alexandre A. Steiner

(Albany College of Pharmacy and Health Sciences, Albany, NY,
USA) for critical comments on the manuscript and Jos Orestes
Del Ciampo (Faculdade de Cincias Farmacuticas de Ribeirao
Preto, Universidade de So Paulo) for the valuable assistance.
This work was supported by a Scholarship of Discovery Grant
Program from Albany College of Pharmacy and Health
Sciences to L.B. Lopes and CNPq and FAPESP (Brazil) to M.
V.L.B. Bentley. J. Hosmer was the recipient of a Research
Award from Albany College of Pharmacy and Health Sciences.
REFERENCES
1. Williams AC, Barry BW. Penetration enhancers. Adv Drug
Deliv Rev 2004;56:603 18.
2. El Maghraby GM. Transdermal delivery of hydrocortisone from
eucalyptus oil microemulsion: effects of cosurfactants. Int J
Pharm 2008;355:285 92.
3. Kogan A, Garti N. Microemulsions as transdermal drug delivery
vehicles. Adv Colloid Interface Sci 2006;123 126:369 85.
4. Heuschkel S, Goebel A, Neubert RH. Microemulsions modern
colloidal carrier for dermal and transdermal drug delivery. J
Pharm Sci 2008;97:603 31.
5. Nornoo AO, Zheng H, Lopes LB, Johnson Restrepo B, Kannan K,
Reed R. Oral microemulsions of paclitaxel: In situ and pharmaco
kinetic studies. Eur J Pharm Biopharm 2008;71(2):310 7.
6. Constantinides PP, Scalart JP, Lancaster C, Marcello J, Marks G,
Ellens H, Smith PL. Formulation and intestinal absorption
enhancement evaluation of water in oil microemulsions incorpo
rating medium chain glycerides. Pharm Res 1994;11:1385 90.
7. Lopes LB, Murphy N, Nornoo A. Enhancement of transdermal
delivery of progesterone using medium chain mono and diglycerides
as skin penetration enhancers. Pharm Dev Techol. 2009. doi:10.1080/
10837450902814180.
8. Magnusson BM, Cross SE, Winckle G, Roberts MS. Percutane
ous absorption of steroids: determination of in vitro permeability
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
and tissue reservoir characteristics in human skin layers. Skin

Pharmacol Physiol 2006;19:336 42.
Frum Y, Khan GM, Sefcik J, Rouse J, Eccleston GM, Meidan
VM. Towards a correlation between drug properties and in vitro
transdermal ux variability. Int J Pharm 2007;336:140 7.
Kantarci G, Ozguney I, Karasulu HY, Arzik S, Guneri T.
Comparison of different water/oil microemulsions containing
diclofenac sodium: preparation, characterization, release rate,
and skin irritation studies. AAPS PharmSciTech 2007;8:E91.
Teichmann A, Heuschkel S, Jacobi U, Presse G, Neubert RH,
Sterry W, Lademann J. Comparison of stratum corneum
penetration and localization of a lipophilic model drug applied
in an o/w microemulsion and an amphiphilic cream. Eur J Pharm
Biopharm 2007;67:699 706.
Bachhav YG, Date AA, Patravale VB. Exploring the potential
of N methyl pyrrolidone as a cosurfactant in the microemulsion
systems. Int J Pharm 2006;326:186 9.
Biruss B, Valenta C. The advantage of polymer addition to a
non ionic oil in water microemulsion for the dermal delivery of
progesterone. Int J Pharm 2008;349:269 73.
Novotny J, Kovarikova P, Novotny M, Janusova B, Hrabalek A,
Vavrova K. Dimethylamino acid esters as biodegradable and
reversible transdermal permeation enhancers: effects of linking chain
length, chirality and polyuorination. Pharm Res 2009;26:811 21.
Rachakonda VK, Yerramsetty KM, Madihally SV, Robinson RL
Jr., Gasem KA. Screening of chemical penetration enhancers for
transdermal drug delivery using electrical resistance of skin.
Pharm Res 2008;25:2697 704.
Verhulst C, Coiffard C, Coiffard LJ, Rivalland P, De Roeck
Holtzhauer Y. In vitro correlation between two colorimetric
assays and the pyruvic acid consumption by broblasts cultured
to determine the sodium laurylsulfate cytotoxicity. J Pharmacol
Toxicol Methods 1998;39:143 6.
Zar T, Graeber C, Perazella MA. Recognition, treatment, and
prevention of propylene glycol toxicity. Semin Dial 2007;20:217 9.
Huang YB, Lin YH, Lu TM, Wang RJ, Tsai YH, Wu PC.
Transdermal delivery of capsaicin derivative sodium nonivamide
acetate using microemulsions as vehicles. Int J Pharm 2008;349:
206 11.
Marszall L, Van Valkenburg JW. The effect of glycols on the
hydrophile lipophile balance and the micelle formation of
nonionic surfactants. J Am Oil Chem Soc 1982;86:84 7.
Venuganti VV, Perumal OP. Poly(amidoamine) dendrimers as
skin penetration enhancers: Inuence of charge, generation, and
concentration. J Pharm Sci. 2008. doi:10.1002/jps.21603.
Fini A, Bergamante V, Ceschel GC, Ronchi C, De Moraes CA.
Control of transdermal permeation of hydrocortisone acetate
from hydrophilic and lipophilic formulations. AAPS PharmSci
Tech 2008;9(3):762 8.
Baroli B, Lopez Quintela MA, Delgado Charro MB, Fadda AM,
Blanco Mendez J. Microemulsions for topical delivery of 8
methoxsalen. J Control Release 2000;69:209 18.
Chen H, Chang X, Weng T, Zhao X, Gao Z, Yang Y, Xu H, Yang
X. A study of microemulsion systems for transdermal delivery of
triptolide. J Control Release 2004;98:427 36.
Lopes LB, Collett JH, Bentley MV. Topical delivery of cyclo
sporin A: an in vitro study using monoolein as a penetration
enhancer. Eur J Pharm Biopharm 2005;60:25 30.
Herai H, Gratieri T, Thomazine JA, Bentley MV, Lopez RF.
Doxorubicin skin penetration from monoolein containing pro
pylene glycol formulations. Int J Pharm 2007;329:88 93.
Benavides T, Mitjans M, Martinez V, Clapes P, Infante MR,
Clothier RH, Vinardell MP. Assessment of primary eye and skin
irritants by in vitro cytotoxicity and phototoxicity models: an in
vitro approach of new arginine based surfactant induced irrita
tion. Toxicology 2004;197:229 37.
Korting HC, Herzinger T, Hartinger A, Kerscher M, Anger
pointner T, Maibach HI. Discrimination of the irritancy potential
of surfactants in vitro by two cytotoxicity assays using normal
human keratinocytes, HaCaT cells and 3T3 mouse broblasts:
correlation with in vivo data from a soap chamber assay. J
Dermatol Sci 1994;7:119 29.
Osborne R, Perkins MA. An approach for development of
alternative test methods based on mechanisms of skin irritation.
Food Chem Toxicol 1994;32:133 42.

DOI: 10.1208/s12249 009 9252 z
Research Article
Comparison of Three Dissolution Apparatuses for Testing Calcium Phosphate
Pellets used as Ibuprofen Delivery Systems
Emilie Chevalier,1 Marylne Viana,1,4 Aymeric Artaud,2 Lisette Chomette,3 Samir Haddouchi,2
Gille Devidts,3 and Dominique Chulia1
Received 15 September 2008; accepted 29 March 2009; published online 14 May 2009
Abstract. Porous calcium phosphate pellets were produced according to two granulation processes (low
and high shear wet granulations) and drug loaded with ve ibuprofen contents (1.75%, 7%, 12.5%, 22%,
and 36%) in order to ensure both bone defect lling and local drug delivery. The drug release kinetics
from the two types of pellets was studied using three dissolution apparatuses: paddle apparatus,
reciprocating cylinder, and ow through cell. The paper compared the three dissolution methods and
considered the effect of the granulation process on the ibuprofen release kinetics. Dissolution data were
analyzed using the Weibull function as well as the difference (f1) and similarity (f2) factors. Dissolution
kinetics was not inuenced by the granulation process, regardless of the dissolution apparatus and of the
drug content. The comparison of the three dissolution devices indicated that ibuprofen was released
faster from granules loaded with 36% of drug content with the reciprocating apparatus, due to the
disintegration of the granules occurring during the dissolution test. For the other drug contents,
dissolution proles were not signicantly different from one apparatus to another. However, the ow
through cell seemed to be more suitable for the drug release study of implantable materials.
KEY WORDS: calcium phosphate pellets; ow through cell; ibuprofen delivery system; in vitro drug
release; paddle apparatus; reciprocating cylinder.
INTRODUCTION
In pharmaceutical industry, in vitro dissolution test is
performed early in order to validate initial screening among
potential formulations to detect the inuence of critical
manufacturing variables and to help in the selection of the
candidate formulation (1,2). The use of dissolution test can
speed up the formulation development, enabling a prompt
identication of potential problems in drug release (3). In
vitro release testing is also a very important tool for batch to
batch quality control (1,2,4). In Europe as well as in the USA,
more than 30 years of research have been devoted to the
characterization of the biopharmaceutical product properties
(5 15). Several guidelines have been then published (16 18)
and all pharmacopeias include recommendations concerning
dissolution tests (19,20). Moreover, dissolution tests have
become one of the primary pharmacopeial tests performed to
ensure the dosage form compliance to quality standards (21).
Furthermore, in vitro studies are the latest tests performed
1
Laboratoire de Pharmacie Galnique, Facult de Pharmacie, CNRS

SPCTS UMR 6638, Universit de Limoges, 2 rue du Dr Marcland,
87025, Limoges Cedex, France.
2
SPS Pharma Services, 28, Place Henri Dunant, 63000, Clermont
Ferrand, France.
3
Varian S.A., 7, avenue des Tropiques, ZA Courtaboeuf 2, BP 12,
91941, LES ULIS Cedex, France.
4
To whom correspondence should be addressed. (e mail: marylene.
viana@unilim.fr)
before the biological evaluations (18). However, the selection

of the appropriate method and data interpretation are not
easily affordable due to the inuence of technological differ
ences especially inducing varied hydrodynamic conditions
(22,23).
From the last decades, various medical devices have
been developed. Among these products, bone implants are
common due to people aging all around the world. In fact,
surgeons are confronted to more and more traumatic and
degenerative bone diseases. Because of the drawbacks of
autologous grafts, calcium phosphate substitutes, such as
hydroxyapatite or tricalcium phosphate, have been increas
ingly used for bone defect lling due to their chemical
composition close to bone mineral phase (24 28). Indeed,
these bioceramics are biocompatible, bioactive, osteoconduc
tive, and resorbable (25,26,29,30). Moreover, they can offer
great interest as drug delivery systems (31 34), achieving a
therapeutic drug concentration directly at the site to be
treated, while maintaining a low systemic drug level (35 38).
In order to evaluate the drug substance release as well as the
biocompatibility and the osteoconductivity of such delivery
systems, in vivo studies must be carried out. These tests
require animals and cell cultures (39) and are consequently
often long and expensive. Therefore, in vitro dissolution tests
could be used in the rst development stages prior to the in
vivo experiments, as for drug products. However, dissolution
devices described in the pharmacopeias relate to pharmaceu
tical dosage forms, but no standard apparatus has been
dened so far for the study of drug release from biomaterials.
597
Chevalier et al.
598
Considering the variety of methods described in the literature
to try to simulate the implanted conditions and characterize
the drug delivery systems (40 44), it is proposed, in this
paper, to investigate the three dissolution apparatuses de
scribed for testing oral dosage forms in US and European
Pharmacopeias (19,20) because of the lack of biomaterial
dissolution test harmonization.
The most common device is the paddle apparatus
(apparatus 2). This assembly consists of a 1000 mL capacity
glass vessel, which may be covered to limit evaporation
phenomenon, and a paddle made up of a blade and a shaft,
used as the stirring element. The vessel is immersed in a
suitable water bath maintained at 37C during the test (19).
The dosage unit form is poured into the vessel containing a
xed volume of dissolution medium. Samples are withdrawn
at regular intervals in order to assay the drug dissolved.
The second apparatus is the reciprocating cylinder,
introduced in the United States Pharmacopoeia in 1991 as
USP Apparatus 3 (45). Its design is based on the disintegra
tion tester (46). The apparatus is composed as follows:
a set of cylindrical, at bottomed glass outer vessels;
a set of glass reciprocating inner cylinders;
screens designed to t the tops and the bottoms of
the reciprocating cylinders.
The outer vessels are immersed in a water bath main
taining the 250 mL of dissolution medium at 37C during the
run (19). The operation involves programming the agitation
rate (in dip per minute, dpm) of the ups and downs for the
inner tube inside the outer tube. At the upstroke, the bottom
mesh in the inner tube moves upward to contact the tested
form and at the downstroke, the sample leaves the mesh and
oats freely within the inner tube, allowing the tested form to
be studied through a moving medium (3,46). Dened
volumes of dissolution medium are withdrawn at regular
intervals prior to drug substance dosage.
The last dissolution method tested in this paper is the
ow through cell. The assembly consists of (19):
a reservoir and a pump ensuring a constant ow rate
of the dissolution medium;
a ow through cell adapted to the dosage unit form
and mounted vertically. A 5 mm diameter ruby bead
and 1 mm glass beads are respectively positioned at
the apex and at the bottom cone of the cell in order
to ensure a laminar ow of the dissolution medium
entering the cell;
a water bath maintaining the dissolution medium at
37C.
One signicant advantage of the ow through cell is that
sink conditions can be maintained, whatever the drug
solubility, using an open loop. However, the run can also be
performed in a closed loop mode, in which a small volume of
medium circulates through the system to provide sample
concentration levels sufcient for the assay.
This paper describes the ability of the three dissolution
apparatuses mentioned in US and European Pharmacopeias
to study the release of ibuprofen, an anti inammatory agent,
from calcium phosphate granules. These implantable biocer
amics were elaborated by two granulation processes (low and
high shear wet granulation) and loaded with ve drug
contents. The effect of granulation process and drug content

on the dissolution proles is also considered for the three
dissolution apparatuses.

Materials
Two types of phosphocalcic granules (710 1,000 m)
were produced, from calcium phosphate (CaP, batch number
G8138/3, Cooper, France) with pregelatinized starch (Sepis
tab ST 200, batch number 80551, Seppic, France) as a binder.
Either high shear wet granulation followed by spheronization
in a Mi Pro granulator (Pro C epT, Zelzate, Belgium) (47) or
low shear wet granulation in a Kenwood mixer (model
KM201, Kenwood Ltd, England) was used. Granules, thus,
prepared were, respectively, called Mi Pro pellets and Kenwood
granules thereafter. Granules were then submitted to a heat
treatment (up to 900C) (48) in order to create porosity by the
removal of the binder also used as a pore former. After that,
granules were drug loaded with 17.5, 67, 125, 222, and 364 mg of
ibuprofen per gram of granules, corresponding, respectively, to
1.75%, 7%, 12.5%, 22%, and 36% of ibuprofen, by a solvent
evaporation technique from an ethanolic ibuprofen solution
(Ibuprofen 50, batch number IB1M738, BASF, Ludwigshafen,
Germany) in a Rotavapor (Bchi, Switzerland).
In vitro Release of Ibuprofen
Dissolution tests were performed on loaded granules
previously described, corresponding to the two granulation
processes and the ve ibuprofen contents. Unloaded granules
were also tested and used as reference. Trials were carried
out, in triplicate, up to the total release of ibuprofen (49),
using a phosphate buffer solution (pH 7.48) at 37C as the
dissolution medium. Ibuprofen quantity was determined, at
regular intervals, by UV absorption spectrophotometry at
264 nm, corresponding to the wavelength of maximal absorp
tion. Three dissolution apparatuses (Fig. 1) were used (19,20):
rotating paddle apparatus (apparatus 2, Prolabo
Dissolution Tester, France) equipped with a paddle
stirrer rotating at 100 rpm (Fig. 1a). The vessels
contained 500 mL of dissolution medium and about
550 mg accurately weighted of loaded granules.
Three milliliters of dissolution medium were with
drawn and ltered before ibuprofen quantity was
determined with an ofine UV Vis spectrophotome
ter (Uvikon 930, Kontron Instrument, UK);
reciprocating cylinder (apparatus 3, Bio Dis, Varian,
Cary, CA, USA) equipped with outer tubes containing
250 mL of medium (Fig. 1b). The top and the bottom of
the inner tubes were polypropylene sieves of 405
meshes. About 275 mg accurately weighted of loaded
granules were placed on the sieve of the inner tube,
agitated at 15 dpm. Three milliliters of dissolution
medium were withdrawn and ltered before ibuprofen
dosage was performed ofine with a UV Vis spectro
photometer (Uvikon 930, Kontron Instrument, UK);
ow through cell (apparatus 4) equipped with tablet
cells of 12 mm (Fig. 1c). A ruby bead of 5 mm
Comparison of Three Dissolution Apparatuses
599
Fig. 1. Dissolution apparatuses a paddle apparatus (dissolution tester, Prolabo) b reciprocating cylinder
(Bio Dis, Varian), and c ow through cell (CE 7smart, Sotax)
diameter and glass beads of 1 mm diameter were placed

in the apex of the ow through cell in order to ensure
laminar ow of the 250 mL of dissolution medium,
previously deaerated by ultrasonic waves, entering into
the cell with a ow rate of 8 mL min 1. About 275 mg
accurately weighted of granules were placed on the
glass bead bed. The automated system CE 7smart
(Sotax, Basel, Switzerland) was linked to a piston
pump CP7 35 (Sotax, Basel, Switzerland) and a UV
VIS spectrophotometer (Lambda 20, Elmer Perkin,
USA) for a direct online analysis of the ltered sample.
Dissolution Data Treatment
Dissolution proles, i.e., cumulative percentage of drug
release (Q, %) versus time (t, min), obtained with the three
congurations were plotted for Mi Pro pellets and Kenwood
granules for the ve tested drug contents.
They were analyzed using the Weibull function (50) in

order to determine the characteristic time TW 80% (min),
corresponding to the time necessary to dissolve 80% of the drug.
Then, release kinetics were compared one to one using
the difference (f1) and similarity factors (f2), as indicated by
the FDA Center of Drug Evaluation and Research (16,17) in
order to evaluate, for each drug content:
the inuence of the dissolution method on ibuprofen
release;
the effect of the granulation process on the release
kinetics.
Two dissolution proles are considered similar when the
f1 value varies from 0 to 15 and f2 is greater than or equal to
50. It should be noted that at least three points are necessary
to compare dissolution proles with the f1f2 procedure
(51,52). Moreover, when 85% or more of the drug substance
is dissolved in 15 min or less, the prole comparison with
Chevalier et al.
600
f1f2 is unnecessary (53), and the drug release proles are
considered as similar.
Results were then modeled according to two mathemat
ical equations, respectively characterizing diffusion or erosion
prevalence, Higuchis square root of time equation (54):
Q% at 1=2 b
According to this equation, if diffusion/erosion ratio A/

B=1, release mechanism is equally controlled by diffusion
and erosion. If A/B>1, then diffusion prevails and if A/B<
1, then erosion predominates.
The initial rate of release was also calculated considering
the quantity of ibuprofen (Q, mg) effectively dissolved during
the rst 5 min.
where a is the release rate (% min 1/2) and b a constant; and

Hixson Crowells cube root of time equation (55):
p
3
100
p
3
100
Q ct
where c is the release rate.

Finally, dissolution data were tted to the Kopchas
empirical equation (56) in order to quantify, when both
mechanisms occur, the respective diffusion (A) and erosion
(B) contributions:
p
Q A t Bt
3

The dissolution proles obtained with each dissolution
apparatus for the two types of granules are presented in
Fig. 2. They highlight the ability of the three dissolution
methods to allow the total release of the ibuprofen quantity
deposited during the loading procedure.
Results indicated that only the ow through cell could be
used to test granules loaded with the lower ibuprofen contents
Fig. 2. Ibuprofen dissolution proles from the two types of granules in each apparatus
36
RC
(drug contents smaller than 22%). In fact, in the case of the

paddle apparatus and of the reciprocating cylinder, the ibupro
fen concentrations were too low to be accurately quantied by
UV spectroscopy due to the device congurations and to the
experimental conditions (vessel sizes, volumes of dissolution
medium). To limit these drawbacks, two kinds of adjustment
might be considered:
F
RC
F
F
RC
RC
22
P RC
36
Dissolution apparatus
Drug content (%)
Fig. 3. Evolution of the characteristic time TW 80% as a function of

granule ibuprofen content in a paddle apparatus, b reciprocating
cylinder, and c ow through cell
Granulation process
RC
22
RC
Kenwood
Table I. Comparison of Dissolution Proles Obtained with the Three Dissolution Apparatuses
Mi-Pro
RC
decreasing the dissolution volume, but smaller equip

ments would be required (vessel/paddle and tube,
respectively);
increasing the sample quantity, but this could be
difcult in development stages
P paddle apparatus, RC reciprocating cylinder, F ow-through cell
601
f1 difference factor
91
68
10 2
21 7
66
19 0
55
42
98
21 3
10 3
26 0
f2 similarity factor
55 3
60 6
49 8
41 1
60 3
42 0
68 4
73 1
53 9
44 6
54 4
36 1
Statistical signicance No difference No difference No difference Difference No difference Difference No difference No difference No difference Difference No difference Difference
Chevalier et al.
17 4
47 1
Difference
32 8
25 0
Difference
TW 80% could not be calculated in all cases, either

because dissolution data were scattered, with a
coefcient of variation larger than 15% (low drug
contents in paddle apparatus and reciprocating
cylinder as seen previously), or because the release
was too fast to be modeled (granules loaded with
1.75% of ibuprofen, tested in ow through cell);
f1f2 comparison could not be achieved for the lower
drug contents (1.75% and 7%) as they did not satisfy
the requirements mentioned in Dissolution Data
Treatment.
17 4
42 9
Difference
75
59 0
No difference
55
62 3
No difference
16 6
40 0
Difference
25 5
30 6
Difference
13 8
49 8
Difference
15 9
43 7
Difference
Consequently, it should be noticed that for dissolution

data analyses:
21 1
36 8
Difference
Statistical signicance
36
12 5
36
12 5
36
36
22
36
22
22
Drug content (%)
36
Kenwood
Granulation process
Mi-Pro
Kenwood
22
Mi-Pro
12 5
22
Kenwood
22
36
12 5
Flow through cell

Reciprocating cylinder
Paddle apparatus
Table II. Comparison of Dissolution Proles Obtained from Granules Loaded with Various Ibuprofen Contents
22
Mi-Pro
22
36
602
Fig. 4. Initial rate of the drug release as a function of granule

ibuprofen content in a paddle apparatus, b reciprocating cylinder, and
c ow through cell
603
Table III. Modeled Dissolution Characteristics

Paddle apparatus
Drug content (%)
22
Granulation process
Higuchi
Hixson Crowell
Kopcha
r
r2
R2
A/B
36
22
Flow through cell
36
22
36
MP
MP
MP
MP
MP
MP
0.994
0.996
0.999
1.60
0.998
0.999
0.999
2.94
0.995
0.995
0.999
5.12
0.992
0.992
0.999
4.01
0.986
0.984
0.999
2.32
0.983
0.989
0.997
1.99
0.980
0.980
0.999
6.95
0.995
0.996
0.992
1.42
0.972
0.980
0.999
15.36
0.992
0.991
0.999
5.35
0.981
0.982
0.998
15.41
0.981
0.977
0.995
33.47
K Kenwood, MP Mi Pro
Ibuprofen appeared to be faster released in the recipro

cating cylinder (Fig. 2) as indicated by the lower TW 80%
values (Fig. 3). Nevertheless, despite these lower TW 80%
values, the pairwise procedure indicated that the dissolution
kinetics was signicantly faster with the reciprocating cylin
der, only for 36% of ibuprofen content (Table I). It should be
noted that TW 80% value increased, for the three appara
tuses, when the drug content increased (Fig. 3), but this
phenomenon was less obvious in the case of the reciprocating
cylinder. In fact, the vertical up and down motion of the inner
tube favored ibuprofen dissolution by shaking the granules.
Furthermore, the design of the apparatus, based on the
disintegration tester, also induced granule degradation as
highlighted by the clouding of the dissolution medium
occurring during the dissolution test. This required to take
into account the absorbance due to the partial disintegration
of the CaP granules (see In vitro release of Ibuprofen).
Considering the specic application as a bone implant,
hydrodynamic conditions involved in the reciprocating cylin
der were not well adapted. Moreover, these particular
agitation conditions were less discriminating to study the
inuence of drug content. Indeed, f1f2 procedure indicated
no signicant difference between the dissolution kinetics
corresponding to 22% and 36% drug content, as opposed to
the results obtained with the two other apparatuses (Table
II). Figure 4 presents the initial rate of drug release for the
three apparatuses. First, it appeared that both ow through
cell and paddle apparatus exhibited slower initial dissolution
for Mi Pro pellets than for Kenwood granules, which was in
accordance with the discussion concerning TW 80% values.
The reciprocating cylinder reversed the granulation process
effect, probably due to its specic hydrodynamic conditions
which submitted the granules to up and down motions. Such
a way, the outer ibuprofen layer would be easier to be
reached. The ibuprofen surface quantity was higher in the

case of Mi Pro pellets, explaining this higher initial rate
(Chevalier et al. Ibuprofen loaded calcium phosphate gran
ules: combination of innovative characterization methods to
rely mechanical strength to drug location. Acta Biomater,
submitted). Secondly, it could also be noticed that the initial
kinetics decreased from paddle apparatus to reciprocating
cylinder and nally ow through cell. However, a downside
should be pointed out for the paddle apparatus which
involved large dissolution volumes compared to physiological
conditions. Therefore, all the results promoted the use of the
ow through cell that appeared to be more suitable to
evaluate drug release over a long period.
Dissolution data were modeled in order to determine the
release mechanism occurring in the three dissolution devices.
The correlation coefcients (r2) for Higuchis and Hixson
Crowells equations, given in Table III, were similar in all
cases, preventing diffusion or erosion prevalence to be
assumed. A strong correlation with Kopchas equation (r2 >
0.99) conrmed the coexistence of diffusion and erosion.
Furthermore, the A/B ratio, higher than 1, indicated that
diffusion prevailed (56), regardless of the granulation process
and of the drug content. Thus, even in the case of the recipro
cating cylinder where granules were partially disintegrated
during the test, release mechanism was preserved, conrming
its relation with the granule formulation. As diffusion preva
lence was better highlighted by ow through cell compared to
paddle apparatus and reciprocating cylinder (Table III), the
ow through cell was recommended for this specic application.
Based on the previous considerations concerning disso
lution tests, some conclusions may be drawn about the
inuence of the granulation process. TW 80% values were
always higher for Mi Pro granules (Fig. 3); this might be
related to their higher rupture strength, their lower porosity,
Table IV. Comparison of Dissolution Proles of the Two Types of Granules

Paddle apparatus
Drug content (%)
22
Granulation process
Statistical signicance
MP
10.4
50.9
No difference
K Kenwood, MP Mi Pro
36
K
MP
6.5
67.9
No difference
22
K
MP
10.3
50.4
No difference
Flow through cell
36
K
MP
9.5
56.3
No difference
12.5
K
MP
9.7
51.3
No difference
22
K
MP
11.0
54.6
No difference
36
K
MP
14.7
51.4
No difference
Chevalier et al.
604
and their higher sphericity, previously studied (data not
published). Nevertheless, the dissolution kinetics of ibuprofen
from granules obtained by the two granulation processes
were not statistically different, whatever the drug content and
the dissolution apparatus (Table IV). In fact, even in the
reciprocating cylinder, where samples were submitted to a
signicant motion, no difference was observed in the behavior
between the two types of granules.
CONCLUSION
Although discriminating dissolution conditions are inter
esting to develop drug delivery systems, difculties may arise
in routine quality control as well as in bioequivalence studies
when dosage forms present high sensitivity to external
dissolution conditions (2). The present work compared
ibuprofen release from two types of granules prepared either
by low shear or by high shear granulation process, intended
for bone implantation. In vitro dissolution studies were
performed with three compendial apparatuses used in phar
maceutical eld (19,20). Release kinetics was not inuenced
by the granulation process, regardless of the dissolution
apparatus. The three devices were able to exhibit how
dissolution time increased with the ibuprofen content. This
paper demonstrated the importance of the apparatus selec
tion, as dissolution method and conditions inuenced ibupro
fen release from the two types of granules. In the case of the
reciprocating cylinder, dissolution kinetics was faster, and the
effect of the granulation process could hardly be taken into
account. This dissolution method was less discriminating,
probably due to the specic design and motion of the
reciprocating cylinder, inducing an undesirable disintegration
of the granules. Therefore, despite the short time and the
small volume required for dissolution test in the reciprocating
cylinder, both interesting for routine tests, the paddle
apparatus and the ow through cell should be preferred in
this specic application. Nevertheless, the volume of dissolu
tion medium used in the paddle apparatus, even though
reduced in this study, was still too important in comparison
with the in vivo conditions. Therefore, to develop bone
implantable materials used as drug delivery systems, the
compendial ow through cell seems to be more suitable.
Furthermore, the dissolution medium ow rate could be
adjusted in order to better mimic bone uid hydrodynamic
conditions. In this purpose, it would be also interesting to test
either the dissolution apparatus 7 which is a compendial small
volume apparatus recently developed for testing medical
devices (57) and the T apparatus which was designed from
the ow through cell to control the convection and diffusion
processes all around the dosage form (58,59).
However, in order to support the in vitro dissolution data
obtained, in vivo experiments have to be performed to
establish in vitro/in vivo correlations and to conclude to the
relevance of the dissolution test.
ACKNOWLEDGMENTS
The authors thank the Rgion Limousin for its nancial
support.
REFERENCES
1. Pillay V, Fassihi R. Evaluation and comparison of dissolution
data derived from different modied release dosage forms: an
alternative method. J Contr Rel. 1998;55:45 55.
2. Missaghi S, Fassihi R. Release characterization of dimehydrinate
from an eroding and swelling matrix: selection of appropriate
dissolution apparatus. Int J Pharm. 2005;293:35 42.
3. Joshi A, Pund S, Nivsarkar M, Vasu K, Shishoo C. Dissolution
test for site specic release isomiazid pellets in USP apparatus 3
(reciprocating cylinder): optimization using response surface
methodology. Eur J Pharm Biopharm. 2008;69:769 75.
4. Siewert M, Dressman J, Brown CK, Shah VP. FIP/AAPS
Guidelines to dissolution in vitro release testing of novel/special
dosage forms. AAPS PharmSciTech. 2003;4:1 10.
5. Mattok GL, McGilveray IJ, Hossie RD. Technical problems of
the USP/NF dissolution test. J Pham Sci. 1971;61:460 2.
6. Smolen VF, Weigand WA. Optimally predictive in vitro drug
dissolution testing for in vivo bioavailability. J Pharm Sci.
1976;65:1718 24.
7. Nasir SS, Wilken Jr LO, Nasir SM. New in vitro dissolution test
apparatus. J Pharm Sci. 1978;68:177 81.
8. Roseman TJ, Derr GR, Nelson KG, Lieberman BL, Butler SS.
Continuous ow bead bed dissolution apparatus for supposito
ries. J Pharm Sci. 1980;70:646 51.
9. Brossard C, Lefort des Ylouses D, Duchne D, Puisieux F,
Carstensen JT. Dissolution of a soluble drug substance from
vinyl polymer matrices. J Pharm Sci. 1982;72:162 9.
10. Aiache JM, Islasse M, Beyssac E, Aiache S, Renoux R, Kantelip
JP. Kinetics of indomethacin release from suppositories. In vitro
in vivo correlation. Int J Pharm. 1987;39:235 42.
11. Aiache JM. French and /or European perspectives on biophar
maceutical characterization of drug dosage forms. J Pharm
Biomed Anal. 1990;8:499 506.
12. Mehta AC. Dissolution testing of tablet and capsule dosage
forms. J Clin Pharm Ther. 1993;18:415 20.
13. Jorgensen E, Bhagwat D. Development of dissolution tests for
oral extended release products. PSTT 1998;1:128 35.
14. Beyssac E, Lavigne J. Dissolution study of active pharmaceutical
ingredients using the ow through apparatus USP 4. Dissolution
Technologies 2005;12:23 5.
15. Graffner C. Regulatory aspects of drug dissolution from a
European perspective. Eur J Pharm Sci. 2006;29:288 93.
16. US Department of Health and Human Services, Food and Drug
Administration, Center of Drug Evaluation and Research
(CDER). Guidance for industry: Dissolution testing of immedi
ate release solid dosage forms. Rockville: FDA; 1997.
17. US Department of Health and Human Services, Food and Drug
Administration, Center of Drug Evaluation and Research
(CDER). Guidance for industry: Extended release oral dosage
forms: development, evaluation and application of in vitro/in vivo
correlations. Rockville: FDA; 1997.
18. International Conference of Harmonization, ICH Q8 Pharma
ceutical Development, www.ich.org; 2 September 2008.
19. Council of Europe. European Pharmacopoeia. 6th ed. Stras
bourg: Council of Europe; 2006.
20. The United States Pharmacopoeia 29th ed, United States
Pharmacopoeial Convention, Rockville, 2006.
21. Maggio RM, Castellano PM, Kaufman TS. A new principle
component analysis based approach for testing similarity of
drug dissolution proles. Eur J Pharm Sci. 2008;34:66 77.
22. Drig T, Fassihi R. Evaluation of oating and sticking extended
release delivery systems: an unconventional dissolution test. J
Contr Release 2000;67:37 44.
23. Kukura J, Baxter JL, Muzzio FJ. Shear distribution and
variability in the USP Apparatus 2 under turbulent conditions.
Int J Pharm. 2004;279:9 17.
24. De Groot K. Bioceramics consisting of calcium phosphate salts.
Biomaterials 1980;1:47 50.
25. LeGeros RZ. Calcium phosphate materials in restorative den
tistry: a review. Adv Dent Res. 1988;2:164 80.
26. Daculsi G, LeGeros RZ, Nery E, Lynch K, Kerebel B.
Transformation of biphasic calcium phosphate ceramics in vivo:
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
ultrastructural and physicochemical characterization. J Biomed

Mat Res. 1989;23:883 94.
Passuti N, Daculsi G, Martin S, Basle M, Roher S. In Heimke G,
Soltesz U, Lee AJC (eds). Macroporous calcium phosphate
ceramics for long bone surgery in human and dogs Clinical and
histological studies. Clinical Implants Materials. Advances in
Biomaterials. BV, Amsterdam: Elsevier Sciences Publishers,
1990, pp. 255 258.
Vallet Regi M, Gonzalez Calbet JM. Calcium phosphates as
substitution of bone tissues. Prog Solid State Chem. 2004;32:1
31.
Daculsi G, LeGeros RZ, Heughebaert M, Barbieux I. Formation
of carbonate apatite crystals after implantation of calcium
phosphate ceramics. Calcif Tissue Int. 1990;46:20 7.
LeGeros RZ, Daculsi G, Orly I, Gregoire M, Heughebaert M,
Gineste M, Kijkowska R. In Ducheyne, Kokubo, Van Blitters
wijk (eds). Formation of carbonate apatite on calcium phosphate
materials: Dissolution/precipitation processes. Bone bonding.
Reed Healthcare Communications, 1992, pp. 201 212.
Arcos D, Ragel CV, Vallet Regi M. Bioactivity in glass/PMMA
composites used as drug delivery system. Biomaterials 2001;22:
701 8.
Meseguer Olmo L, Ros Nicolas MJ, Clavel Sainz M, Vicente
Ortega V, Alcaraz Banos M, Lax Perez A, Arcos D, Ragel CV,
Vallet Regi M. Biocompatibility and in vivo gentamicin release
from bioactive sol gel glass implants. J Biomed Mater Res Part A
2002;61:458 65.
Ginebra MP, Traykova T, Planell JA. Calcium phosphate
cements as drug delivery systems: a review. J Control Release
2006;113:102 10.
Balcik C, Tokdemir T, Senkyl A, Koc N, Timucin M, Akin S,
Korkusuz P, Korusuz F. Early weight bearing of porous HA/TCP
(60/40) ceramics in vivo: a longitudinal study in a segmental bone
defect model of rabbit. Acta Biomaterialia 2007;3:985 96.
Corry D, Moran J. Assessment of acrylic bone cement as a local
delivery vehicle for the application of non steroidal anti inam
matory drugs. Biomaterials 1998;19:1295 301.
Mendez JA, Fernandez M, Gonzalez Corchon A, Salvado M,
Collia F, De Pedro JA, Levenfeld BL, Lopez Bravo A, Vazquez
B, San Roman J. Injectable self curing bioactive acrylic glass
composites charged with specic anti inammatory/analgesic
agent. Biomaterials 2004;25:2381 92.
Gonzalez Corchon MA, Salvado M, De la Torre BJ, Collia F, De
Pedro JA, Vazquez B, San Roman J. Injectable and self curing
composites of acrylic/bioactive glass and drug systems. A
histomorphometric analysis of the behaviour in rabbits. Bio
materials 2006;27:1778 87.
Vallet Regi M, Balas F, Arcos D. Mesoporous materials for drug
delivery. Angew Chem Int. 2007;46:7548 58.
Detsch R, Mayr H, Ziegler G. Formation of osteoclast like cells
on HA and TCP ceramics. Acta Biomaterialia 2008;4:139 48.
Guicheux J, Grimandi G, Trcant M, Faivre A, Takahashi S,
Daculsi G. Apatite as carrier for growth hormone: in vitro
characterization of loading and release. J Biomed Mater Res.
1997;34:165 70.
605
41. Pillay V, Fassihi R. Unconventional dissolution methodologies. J
Pharm Sci. 1999;88:843 51.
42. Krajewski A, Ravaglioli A, Roncari E, Pinasco P. Porous ceramic
bodies for drug delivery. J Mater Sci Mater Med. 2000;12:763 71.
43. Sunder M, Babu NR, Victor SP, Kumar KR, Sampath Kumar TS.
Biphasic calcium phosphates for antibiotic release. Trends
Biomater Artif Organs 2005;18:213 8.
44. Melville AJ, Rodriguez Lorenzo LM, Forsythe JS. Effects of
calcination temperature on the drug delivery behaviour of
ibuprofen from hydroxyapatite powders. J Mater Sci Mater
Med. 2008;19:1187 95.
45. The United States Pharmacopoeia 22nd ed, United States
Pharmacopoeial Convention, Rockville, 1991.
46. Yu LX, Wang JT, Hussain AS. Evaluation of USP Apparatus 3
for dissolution testing of immediate release products. AAPS
Pharm Sci. 2002;4:1 5.
47. Chevalier E, Viana M, Pouget C, Chulia D. Inuence of process
parameters on pellets elaborated in a Mi Pro high shear
granulator. Pharm Dev Technol. 2007;12:133 44.
48. Chevalier E, Viana M, Pouget C, Cazalbou S, Champion E,
Chulia D. In Bioceramics: Properties, Preparations and Appli
cation. From porous pellet fabrication to drug loading and
release: the case of calcium phosphate matrix loaded with
ibuprofen. 2009; in press.
49. Chevalier E, Viana M, Cazalbou S, Chulia D. Validation of a
fabrication process of pellets for bone lling and drug delivery. J
Drug Deliv Sci Technol. 2008;18(6):438 444.
50. Gibassier D, Sado P, Le Verge R, Devissaguet JP. Test de
dissolution et fonction de Weibull. Labo Pharma Pro Tech. 1982;
30:250 5.
51. Moore JW, Flanner HH. Mathematical comparison of curves with an
emphasis on in vitro dissolution proles. Pharm Tech. 1996;20:64 74.
52. Shah VP, Tsong Y, Sathe P. In vitro dissolution prole
comparison statistics and analysis of the similarity factor, f2.
Pharm Res. 1998;15:889 96.
53. Costa P, Sousa Lobo JM. Modelling and comparison of
dissolution proles. Eur J Pharm Sci. 2001;13:123 33.
54. Higuchi T. Mechanism of sustained action medication. Theoret
matrices. J Pharm Sci. 1963;52:1145 9.
55. Hixson AW, Crowell JH. Dependence of reaction velocity upon
surface and agitation. Ind Eng Chem. 1931;23:923 31.
56. Kopcha M, Lordi N, Tojo KJ. Evaluation of release from
selected thermosoftening vehicles. J Pharm Pharmacol. 1991;43:
382 7.
57. Zhou MX, Shoudt D, Calderon G, Feng M. Application of USP
Apparatus 7 to in vitro drug release in scopolamine transdermal
systems. Dissolution Technol. 2007;14:25 9.
58. Amyot F, Boudy V, Jurski K, Counord JL, Guiffant G, Dufaux J,
Chaumeil JC. A new experimental method for the evaluation of
the release proles of drug loaded microbeads designed for
embolisation. ITBM RBM 2002;23:285 9.
59. Borovac T, Pelage JP, Kasselouri A, Prognon P, Guiffant G,
Laurent A. Release of ibuprofen from beads for embolization: in
vitro and in vivo studies. J Control Rel. 2006;115:266 74.

DOI: 10.1208/s12249 009 9221 6
Research Article
Preparation and Characterization of Co-Grinded Mixtures of Aceclofenac
and Neusilin US2 for Dissolution Enhancement of Aceclofenac
Ambarish H. Vadher,1,2,3 Jolly R. Parikh,1,5 Rajesh H. Parikh,4 and Ajay B. Solanki1
Received 6 October 2008; accepted 13 February 2009; published online 15 May 2009
Abstract. The present study was carried out with a view to enhance the dissolution of poorly water soluble
BCS class II drug aceclofenac by co grinding with novel porous carrier Neusilin US2. (amorphous
microporous granules of magnesium aluminosilicate, Fuji Chemical Industry, Toyama, Japan). Neusilin US2
has been used as an important pharmaceutical excipient for solubility enhancement. Co grinding of
aceclofenac with Neusilin US2 in a ratio of 1:5 was carried out by ball milling for 20 h. Samples of co
ground mixtures were withdrawn at the end of every 5 h. and characterized for X ray powder diffraction,
differential scanning calorimetry, and Fourier transform infrared spectroscopy. The analysis revealed the
conversion of crystalline aceclofenac to its amorphous form upon milling with Neusilin US2. Further, in vitro
dissolution rate of aceclofenac from co ground mixture was signicantly higher compared to pure aceclofenac.
The accelerated stability study of co ground mixture was carried out at 40C/75%RH for 4 weeks, and it
showed that there was no reversion from amorphous to crystalline form. Thus, it is advantageous to use a
porous carrier like Neusilin US2 in improvement of dissolution of poorly soluble drugs.
KEY WORDS: aceclofenac; co grinding; dissolution; Neusilin US2; porous carrier.
INTRODUCTION
The drug must dissolve in the gastrointestinal uid before it
can permeate across the gastrointestinal barriers. Thus, dissolu
tion of poorly water soluble drugs becomes the rate limiting
step in their absorption. Enhancement of the rate of dissolution
of such poorly water soluble drug can facilitate formulation
design of immediate release dosage forms. Dissolution enhance
ment can result in improved bioavailability, a critical determi
nant in the success of new chemical entity (1).
Several physical approaches which have been employed to
improve drug dissolution include size reduction, melt adsorption,
melt quenching, solvent deposition, spray drying, and freeze
drying. These techniques often lead to the formation of
amorphous solids which have higher dissolution rates and
therefore higher bioavailability (2). Hancock and Parks reported
two to four times higher solubilities of amorphous solids
compared to crystalline solids (3). Gupta et al. reported
enhancement of drug dissolution rate upon storage with three
A. R. College of Pharmacy and G. H. Patel Institute of Pharmacy,

Vallabh Vidyanagar, 388 120, India.
2
Institute of Science and Technology for Advanced Studies and
Research (ISTAR), Valllabh Vidyanagar, 388 120, India.
3
Present address: Formulation and Development Department,
Alembic Research Centre, Alembic road, Vadodara, 390 003, India.
4
Ramanbhai Patel College of Pharmacy, Education Campus, Changa,
482 321, India.
5
To whom correspondence should be addressed. (e mail: jollyparikh
87@yahoo.co.in)
component granules (4,5) and by co grinding ketoprofen,

indomethacin, progesterone, and naproxen with Neusilin (6).
Watanabe reported higher solubility of indomethacin due to
amorphization by co grinding with silica, (7,8) with cross
povidone (9) and with PVP (10). Otsuka et al. proposed that
non crystalline form of indomethacin has 60% higher solubil
ity than polymorphs (11). The importance of humidity and
ratio of indomethacin to neusilin in amorphization of crystalline
drug was reported by Bogner et al. (2). Ali et al. used a vibration
mill to prepare amorphous co ground mixtures of ufenamic
acid with amorphous calcium silicate and silicon dioxide (12).
Volatile compounds such as naphthalene, d camphor, and
p cresol amorphized and reportedly lost their volatility when
co ground with microcrystalline cellulose (MCC) (13). Amobar
bital amorphized in the presence of variety of excipients such as
carbon black, ethyl cellulose, precipitated silica, and activated
charcoal (14). Poorly water soluble drug TAS 301 was melt
adsorbed onto calcium silicate to prepare amorphous state of the
drug, which was physically stable with improvement of solubility
and oral bioavailability (15). Roland et al. has studied the effect
of co grinding of poorly soluble drug nifedipine with various
hydrophilic carrier such as partially hydrolyzed gelatin (PHG),
polyvinylpyrrolidone (PVP), sodium dodecyl sulfate (SDS),
hydroxypropyl methylcellulose (HPMC), polyethylene glycol
(PEG), urea or Pluronic F108 and reported the enhancement of
dissolution rate. PHG ground mixtures resulted in the fastest
dissolution rate followed by PVP, SDS, HPMC, pluronic, urea,
and PEG (16).
Neusilin consist of amorphous microporous granules of
magnesium aluminosilicate with a high specic surface area
(300 m2 g). Neusilin has silanol groups on its surface, which
606
Aceclofenac and Neusilin US2 for Enhancement of Aceclofenac

make it a potential proton donor as well as acceptor. The
hydrogen bonding potential of silanol in the local
environment on silica surfaces is well documented (17 19).
Aceclofenac (AC) 2 [[2 [2 [(2,6 dichlorophenyl) amino)
phenyl)acetyl)oxy)acetic acid is a nonsteroidal analgesic,
antipyretic, anti inammatory drug. It possesses free carbox
ylic acid in the structure (20,21). The molecular weight of
aceclofenac is 354.2. It belongs to a category of poorly water
soluble BCS class II drug, which exhibits poor solubility and
dissolution (as shown in Fig. 1).
Milling is frequently used in the pharmaceutical industry
to reduce the particle size of drugs. Other than the reduction in
the particle size of a drug on milling, concurrent changes in the
crystal structure of the drug have also been reported (22 24).
A combination of impact and attrition during ball milling can
bring about changes in the polymorphs and hydrates of a drug
and can induce amorphization as well (10).
Solid drugs may exist as crystalline substances or amor
phous particles without identiable structure. The amorphous
or crystalline character substance can affect the stability and
activity of the drug within the formulation. The amorphous
form often presents greater solubility, dissolution velocity, and
bioavailability than the crystalline structure, being that, in the
amorphous state, the necessary energy for molecule separation
is less than that of the crystalline form (25).
The amorphous state is characterized by the absence of the
long range, three dimensional molecular order characteristic of
the crystalline state. Practically, an amorphous material can be
obtained in two ways: (1) by cooling the molten liquid until the
molecular mobility is frozen in, thus producing the glass
(vitrication) and (2) by gradually inducing defects in the crystal
until the amorphous form is attained (amorphization) (26).
Some of the useful properties of amorphous material are
higher solubility, higher dissolution rate, and sometimes better
compressibility. From thermodynamic point, it possesses higher
energy, entropy, and free energy than corresponding crystals (27).
The problem associated with amorphous states generated
by milling are prone to relax toward more stable state and can
recrystallize more or less rapidly upon storage, which leads to
less stable polymorphic form than nonmilled form. Sometimes,
Fig. 1. Schematic diagram of Aceclofenac
607
milling processes induces nonequilibrium transformation which

can result in critical stability issues (28). However it is not
possible to predict the nature of transformation induced by
milling, but it seems the two physical parameters that mainly
drive the transformations are milling intensity (29,30) and
milling temperature (31,33).
Increasing milling intensity has been found to drive some
drug toward increasingly metastable states, e.g., indomethacin
and cimetidine. Concerning milling temperature, it seems that
amorphization mainly occurs when milling is performed
well below the glass transition temperature (Tg) of the
corresponding liquid, while polymorphic transformation
mainly occurs when the milling is performed above Tg.
Aceclofenac exhibits very slight solubility in water, and as a
consequence, it exhibits low bioavailability after oral adminis
tration (34,35). Therefore, the improvement of aceclofenac
dissolution from its oral solid dosage forms is an important issue
for enhancing its bioavailability and therapeutic efcacy.
Different approaches were used to enhance the dissolution
of aceclofenac that include preparation of solid dispersion by
using mixed surfactant system of SLS and alkyl polyglucosides
(Patel et al. (36)), preparation of co crystals of chitosan with
aceclofenac (Mautik et al. (37)), preparation of spherical
agglomerates of aceclofenac with HPMC (Usha et al. (38)),
and preparation of inclusion complex with HP beta cycodextrin
(Pathak and Dahiya (39)). Milling with porous carrier like
Neusilin has been reported to enhance the dissolution of several
drugs. But the same approach was not applied for aceclofenac
till date. So, present study aims to enhance the dissolution of
poorly water soluble drug aceclofenac by co grinding it with
Neusilin in ball mill. The conversion of crystalline aceclofenac to
amorphous form as well as the physical stability of the resulting
amorphous state of the drug was monitored by X ray powder
diffraction (XRD) and differential scanning calorimetry (DSC).
Fourier transform infrared (FTIR) spectroscopy was used to
investigate the presence of any interaction between drug and
Neusilin. The product obtained by ball milling was subjected to
dissolution rate and stability studies.
There have been contradictory reports in the literature
regarding the physical stabilization of amorphous solids using
silicates. Kinoshita et al., showed that melt adsorption of a drug
on Florite (amorphous calcium silicate) led to the formation of
its amorphous state which was stable for 3 days at 60C/80% RH
and at least 2 years at ambient temperature and humidity (1).
Watanabe et al., however, reported the reversion of amorphous
indomethacin co ground with silica at 30 C/11% RH in 10 days
(7) (2). Gupta et al. found that the amorphous drugs formed by
co grinding with Neusilin US2 (magnesium aluminometasili
cate) were physically stable for at least 4 weeks at 40C/75% RH
(6). Konno et al. have reported spontaneous amorphization of
crystalline drugs stored (as a physical mixture) with silica (40).
Since the amorphous form of the drug is associated with higher
energy than its crystalline counterpart, spontaneous amorphiza
tion is an intriguing phenomenon. Attempts have been made to
correlate the amorphization of the drugs by co grinding, simple
mixing, and fusion with the porous nature of the excipients
(41,42). Kim et al. used porous and nonporous silicas to make
amorphous solids by physical mixing and fusion. They suggested
that amorphization and the subsequent stabilization of the
amorphous state is dependent upon the porosity of the silicate
(41). However, the silicates used in their study differed from
608
Vadher, Parikh, Parikh, and Solanki
each other with respect to pH and surface area as well as

porosity. However, in this article, stability study is carried out as
time and condition suggested by Gupta et al. (6).
Materials
Neusilin US2 (Fuji Chemical Industry, Toyama, Japan)
was obtained as a gift sample from Gangwal Chemicals,
Mumbai. Aceclofenac was obtained as a gift sample from
Alembic, Vadodara. All other chemicals and solvents were of
analytical grade.
Methods
Preparation of Co Ground Mixture of Aceclofenac
and Neusilin US2
A powder mixture comprising of aceclofenac and neusilin
in a ratio of 1:5 by weight was milled at 25C using a modied
ball mill for a period of 20 h (6). A jar rolling mill with a plastic
jar (outer diameter=5.5 in.) and glass balls (outer diameter=
1.2 in.) was used to perform the ball milling operation. The
speed of the cylindrical jar was maintained at 80 rpm, which
allows for signicant attrition with some impact. Physical
mixture of (6 g) of aceclofenac and neusilin was milled up to
20 h. (Gupta et al.) to effect conversion to amorphous states of
the drug. Milling was performed at room temperature of 25C,
and no increase in the temperature of the milled material was
detected at the end of the process. The milled material was
sieved through mesh no. 30 (600 m opening) and stored in glass
vials at room temperature until used for further analysis.
Characterization of Co Ground Mixtures of Aceclofenac
and Neusilin US2
The co ground mixture was characterized with respect to
the following methods (43,44):
1.
2.
3.
4.
5.
FTIR studies
Differential scanning calorimetry studies
X ray diffraction studies
In vitro drug dissolution studies
Stability studies
time intervals, respectively, were separately weighed and

hermetically sealed in the aluminum pans. A Thermal
Analysis System instrument (Perkin Elmer DSC pyris 1,
USA) with intracooler, a refrigerated cooling system, was
used. Indium standard was used to calibrate the DSC
temperature and enthalpy scale. The system was purged with
nitrogen gas at a ow rate of 80 mL/min. Initially, the samples
were held at 50C for 1 min and after that heating was
performed from 50C to 320C at a scan rate of 10C/min.
X-Ray Diffraction Studies
X ray diffraction (XRD) pattern of the sample of the
Neusilin US2 and co ground samples were studied by placing
a thin layer of the powder in conventional cavity mounts of an
X ray diffractometer (PHILLIPS, XPert, Holland).
The sample scan was started at 5.0175 angle at 2 to
100.00250 end angle with step size at 2 of 0.0550. The
source was X ray tube with Cu target operated at 2 kW X ray
power that generated K Alpha 1 X ray radiation and detector
was Xe lled counteract or proportional detector.
In vitro Drug Dissolution Studies
The dissolution studies of pure aceclofenac (100 mg), co
ground mixture collected at the end of 5 h and at the end of
20 h (equivalent to 100 mg drug), respectively, were
performed using USP XXIV type II dissolution apparatus
(Veego Make, India) for 8 h. The dissolution medium used
was 500 ml of distilled water maintained at 37C0.5C. The
paddle speed was 100 rpm. Five milliliter samples were
collected at 15, 30, 45, 60, 80, 100, 120, 150, 180, 240, 300,
360, 420, and 480 min and replaced with equal quantity of
dissolution medium. After ltration through Whatman lter
paper 42 (Whatman, Middlesex, UK), samples were analyzed
using double beam UV spectrophotometer at 275.5 nm.
Stability Studies of Co Ground Mixture of Aceclofenac
and Neusilin US2
The milled powders were stored at 40C and 75% RH
for 4 weeks. FTIR spectra of the fresh sample and stored
samples were compared to examine the interaction of the
aceclofenac with Neusilin US2, if any. XRD and DSC data of
the fresh and stored samples were compared to evaluate any
changes in drug crystallinity.
FTIR Studies
IR spectra of aceclofenac, Neusilin US2, and co grinded
mixture of aceclofenac Neusilin US2 collected at different
time intervals was taken on a FTIR (Perkin Elmer, Spectrum
GX FTIR, USA). The pellets were prepared using KBr press
(Spectra Lab, Mumbai, India) using a mixture of sample and
KBr in 1:10 ratio. The spectra were recorded over the wave
number range of 4,000 to 400 cm 1.
Differential Scanning Calorimetry Studies
Samples of aceclofenac, Neusilin US2, and co grinded
mixture of aceclofenac Neusilin US2 collected at 0 and 20 h
FTIR Studies
The FTIR study was carried out to nd out the
interaction taking place between the Neusilin US2 and
aceclofenac at a structural level during co grinding. The
Fig. 2 shows the FTIR spectra of the six different samples
taken at different time intervals.
IR spectrum of pure aceclofenac shows characteristic
peaks at 965.21 cm 1 (O H bending out of plane),
3,319.73 cm 1 (secondary amine N H stretching or O H
stretching), 1,771.89 cm 1 ( C=O stretching) 1,577.9 cm 1,
609
Fig. 2. FTIR analysis of a Aceclofenac, b Neusilin, c Mixture at 0 h, and d Mixture at 20 h
1,508.07 cm 1 ( C=C stretching of aromatic), and some

prominent bands such as 1,344.6 1,256.58 cm 1 ( C O
stretching). Characteristic peaks of drug are also present in IR
spectrum of mixture of the aceclofenac neusilin at different time
intervals with some broadening and reduction in intensity. The
mixture of aceclofenac neusilin gives peak near 3,469.63 cm 1
indicating formation of H bond between OH of COOH and Si
of SiO2. This leads to formation Si O C bridging bond
formation. Further, there is reduction in the intensity of the
peaks 1,771.89 and 1,717 cm 1 indicating interaction but not
disappearance of the drug peak. Hence, it is conrmed that
neusilin binds with the aceclofenac by forming H bond and
converts from its crystalline form to its amorphous form.
Formation of H bond was previously reported by Gupta
et al. on co grinding carboxylic acid containing drugs such as
indomethacin, ketoprofen, naproxen, and progesterone with
Neusilin US2 (6). Aceclofenac is a carboxylic acid containing
drug, so it has an acidic nature; thus, there is a possibility of
an acid base interaction between the drug and Neusilin US2.
Watanable et al. reported that pKa values of SiO2 and
indomethacin are so close that a silanol group may become
amphoteric, functioning as a Bronsted acid and as a Bronsted
base (45). Further, Neusilin US2 is amorphous and practically
insoluble in water. The pH of the 4% w/v slurry of Neusilin
US2 in water is 7.4 indicating its neutral nature (19).
The acid base reaction taking place can also be explained
from the FTIR spectra of the milled powder. As expected in an
acid base reaction, free acid carbonyl peak and the drug dimer
or oligomer peaks disappeared, and the peak for the carbox
ylate ion appeared in the region of 1,540 1,650 cm 1 (in this
case, the peak appeared at 1,652 cm 1) in the FTIR spectrum
(46). Also, the free acid carbonyl peak becomes weaker in the
amorphous neusilin bound states of the carboxylic acid
containing drug. The changes in the FTIR spectra indicate an
acid base interaction between the carboxylic acid containing
drugs and neusilin to form their salts.
It was suggested that water mediates the acid base
reaction between the crystalline states of MgO and aceclofenac
to result in crystalline magnesium salt of aceclofenac. FTIR
data showed a peak of carbonyl at 1,717 cm 1 in the spectrum
of aceclofenac. This carbonyl peak becomes signicantly
weaker and is accompanied by the appearance of a new

carboxylate peak at 1,652 cm 1. Disappearance of the carbonyl
peak and the appearance of the carboxylate peak in the FTIR
spectra of aceclofenac milled with amorphous neusilin suggest
amorphous salt formation in the present study. While FTIR
provides the evidence of salt formation, XRD and DSC data
suggest that the salt is amorphous.
Electrostatic forces between COO group present in
aceclofenac and counterions, such as Mg2+ and Al3+ present in
neusilin, are responsible for causing amorphous salt formation
of drug through hydrogen bonding. These interactions seem to
drive or are responsible for amorphization of the drug in the
present study.
Differential Scanning Calorimetry Studies
DSC patterns of four different samples are shown in the
Fig. 3, and the following parameters were observed. Aceclo
fenac shows peak at 153.683C with a peak height of
19.3266 mW, and Neusilin US2 shows peak at 221.406 C
with a peak height of 0.6928 mW.
While observing the DSC pattern of mixture at 0 h, it was
observed that the peak of the aceclofenac appears at 152.370 C,
which is near to the original peak, but peak height reduces
signicantly to 1.6275 mW. Further, the peak of the Neusilin US2
also shifts toward a higher melting point 251.03C, but there is
reduction in the peak height.
The DSC pattern of Mixture collected at the end of 20 h
shows disappearance of the drug melting point peak from
153.683C indicating complete amorphization of aceclofenac at
the end of 20 h. As previously reported by Gupta et al., the
melting point is considered as important criteria for milling of
the drug with carrier. Analogous to heat providing the energy to
disrupt the crystal lattice during melting, mechanical energy
leads to amorphization during ball milling. So, the drug with
higher melting point would require longer time for amorphiza
tion on milling. The melting point of aceclofenac is 153C; hence,
it was ball milled for 20 h to achieve complete amorphization (6).
It was shown by Tong et al. that stronger electrostatic
interactions between the carboxylate group of indomethacin
and counterions, such as sodium and potassium, can increase the
610
Fig. 3. DSC analysis of a Aceclofenac, b Neusilin, c Mixture at 0 h, and d Mixture at 20 h
glass transition temperature, Tg, of the amorphous salts,

resulting in higher physical stability of the salt in comparison
with the acid at a particular storage temperature. The Tg of the
salt of indomethacin with monovalent lithium (139C) (46) was
much higher than that of the melt quenched acid (44C) (47),
and it was suggested that the Tg might be even higher for salts
with divalent magnesium (present in Neusilin), thereby improv
ing the physical stability of the amorphous salt of indomethacin.
Similarly for aceclofenac, it was not possible to determine the Tg
of the amorphous neusilin bound drugs by DSC, which may
very well be higher than the melting point of the drugs.
Supporting evidence for amorphization of the aceclofe
nac on milling is obtained from the absence of any peaks in
the XRD patterns.
X-Ray Diffraction Studies

The XRD analysis was carried out for the four different
samples, and the plots are shown in Fig. 4 after overlapping
on single scale to study the difference in crystallinity of
aceclofenac. The XRD of pure aceclofenac is shown by black
lines, and it shows characteristics peaks at 18.32, 19.23.,
22.07, 24.30, and 25.76 and 31.97 at (2 ). The peak at
25.76 at (2 ) was used to compare the XRD pattern of
Neusilin US2 and co ground mixture of aceclofenac neusilin
collected at 0 and 20 h intervals.
The XRD represented by the green lines is that of
Neusilin US2, and it gives no peaks indicating its amorphous
nature. The other two XRD patterns shown by blue line and
Fig. 4. XRD patterns of a Aceclofenac, b Neusilin, c Mixture at 0 h, and d Mixture at 20 h
611
compared to spectra of pure drug without milling. The only

difference observed is reduction in particle size from volume
mean diameter (VMD) 144 to 18.2 m. So, it is conrmed
that only ball milling does not produce amorphous form of
aceclofenac, but it is neusilin which is playing a signicant
role to achieve conversion of drug from crystalline to
amorphous via formation of hydrogen bonding.
In vitro Drug Dissolution Studies
Fig. 5. Comparison of dissolution prole of pure drug and mixture

collected at 5 and 20 h interval
red line are of co ground mixtures of aceclofenac neusilin at

0 h and aceclofenac neusilin at 20 h intervals, respectively. It
is observed that the intensity of the peak reduces in case of
blue lines (mixture at 0 h interval) indicating that Neusilin
US2 interferes with the aceclofenac probably by forming
hydrogen bond and thus reduces the crystallinity of the drug.
Further, the XRD pattern of the mixture collected at 20
h interval, indicated by red lines, indicates complete amorphiza
tion of the drug because XRD pattern of red lines does not show
any peaks. Hence, XRD data reveals conversion of aceclofenac
from crystalline form to its amorphous form.
Aceclofenac was ball milled alone to check the effect of
ball milling on the drug properties, and the results show that
there is no change in XRD, FTIR, and DSC spectra
In vitro drug dissolution studies were carried out for 8 h

on three different powder samples, i.e., aceclofenac alone
(pure drug), mixture of Neusilin US2, and aceclofenac
collected at the end of 5 h (Mix5 h) and mixture of Neusilin
US2 and aceclofenac collected at the end of 20 h (Mix20 h).
The dissolution study was carried out to obtain the
cumulative percent drug dissolution for 8 h. Figure 5 shows
the graph of percent dissolution versus time.
The results of in vitro dissolution studies carried out of
aceclofenac pure, a mixture of neusilin and aceclofenac
collected at 5 h interval and at the end of 20 h co grinding
are shown in the Fig. 5.
From the graph, it is evident that pure aceclofenac shows
the cumulative percent drug dissolution of 92% at the end of
8 h. While co ground mixture of Neusilin US2 and aceclofe
nac, collected at the end of 5 h interval and at the end of 20 h
interval, show 103% drug dissolution within 3 h. Hence co
ground mixtures give faster dissolution rates compared to
crystalline aceclofenac (pure drug), which can be due to dual
effect produced during milling, i.e., conversion of crystalline
form of aceclofenac to amorphous form during co grinding as
well as particles size reduction of aceclofenac during co
grinding, as reduction in particles size will provide larger
surface area for the drug to dissolve in dissolution media.
Further, when comparing the dissolution prole of both
the mixtures, the mixture collected at 20 h interval showed
initial drug dissolution slightly faster than the mixture collected
Fig. 6. XRD patterns of a Aceclofenac, b Neusilin, c Mixture at 20 h, and d Mixture after four weeks
storage at 40C and 75% RH
612
Fig. 7. FTIR spectra of a Aceclofenac, b Neusilin, c Mixture at 20 h, and d Mixture after four weeks
storage at 40C and 75% RH
at 5 h interval. So, it indicates that upon further co grinding

after 5 h, complete amorphization of aceclofenac occurs.
Stability Study of Co-Ground Mixture of Aceclofenac and
Neusilin US2
The physical stability of the resulting amorphous state
of the drug after co grinding with Neusilin was monitored by
XRD and DSC. The FTIR was also taken to identify the
mechanism of interaction. The samples were stored in vials
at 40C and 75% RH for 4 weeks. The data collected from
the 4 week stability samples were compared with those
of pure aceclofenac and co ground mixture collected at
20 h interval.
The XRD pattern of pure aceclofenac gives the charac
teristics peaks at 18.32, 19.23., 22.07, 24.30 and 25.76 and
31.97 at 2 . The peak at 25.76 at 2 was used to compare

the XRD pattern of fresh mixture collected at 20 h and
mixture after 4 weeks of storage. The XRD patterns of both
the mixtures shows no peaks indicating complete amorphiza
tion (Fig. 6). This indicates no reversion of drug form its
amorphous state to its crystalline state during its storage at
40C and 75% RH for 4 weeks
Further, FTIR data also conrms stability of H bond
formed between aceclofenac and neusilin formed during
milling. From FTIR of the mixture stored for 4 weeks, it
can be seen that broadening of peak 3,469.63 cm 1 still
persists as shown in Fig. 7.
Further, supporting evidence for stability of amorphiza
tion state was obtained from DSC analysis of the mixture
stored for 4 weeks (Fig. 8). No peak was observed at melting
point peak of aceclofenac at 153.683C in this sample.
Fig. 8. DSC analysis of the mixture stored for 4 weeks at 40C and 75% RH

So, from all the three analysis, it is conrmed that
aceclofenac retains its amorphous state after 4 week stability
period, and no reversion to its crystalline state occurs.
Formation of the amorphous state of aceclofenac is
feasible by ball milling with Neusilin. The resulting amor
phous states of the drug appear to be physically stable during
storage at 40C and 75% RH for up to 4 weeks.
It was shown by Tong et al. that stronger electrostatic
interactions between the carboxylate group of indomethacin
and counterions, such as sodium and potassium, can increase
the glass transition temperature, Tg, of the amorphous salts,
resulting in higher physical stability of the salt in comparison
with the acid at a particular storage temperature. The Tg of
the salt of indomethacin with monovalent lithium (139C)
(46) was much higher than that of the melt quenched acid
(44C) (47), and it was suggested that the Tg might be even
higher for salts with divalent magnesium (present in Neusilin),
thereby improving the physical stability of the amorphous salt
of indomethacin. A similar hypothesis can be applied for
aceclofenac that the Tg of the amorphous neusilin bound drug
may very well be higher than the melting point of the drug
thereby improving the physical stability of the co ground
mixture. The physical stabilization of aceclofenac co ground
mixture can also be attributed to its restricted molecular
mobility due to mechanochemical reaction on milling with
neusilin and formation of bridging bonds as suggested by
Watanable et al. (48).
CONCLUSION
The present study shows the utility of a novel porous
carrier such as neusilin to enhance the solubility and dissolu
tion of poorly water soluble drug aceclofenac. A simple
method of co grinding was adopted to convert crystalline form
of drug to amorphous form. The results indicate that neusilin
enhances the dissolution of aceclofenac and also provides
physical stability by preventing reversion of drug from
crystalline state to amorphous state after co grinding. This
approach can be further extended for dissolution enhancement
of other BCS class II drugs.
ACKNOWLEDGEMENTS
The authors are thankful to Gangwal Chemical, Mumbai
and Alembic Ltd. Vadodara for providing free gift samples of
Neusilin US2 and aceclofenac, respectively. The authors are
also thankful to Mr. Vipul Vaghela for providing valuable
suggestions during these studies and SICART for providing
the analytical service for this work.
REFERENCES
1. Gupta MK. Factors inuencing the release of poorly water
soluble drugs from solid dispersion granules during storage,
PhD Dissertation, 2002; University of Connecticut.
2. Bogner RH, Bahl D. Amorphization of Indomethacin by Co
grinding with Neusilin US2: Amorphization kinetics, Physical
stability and mechanism. Pharm Res. 2006;23(10):2317 25.
3. Hancock BC, Parks M. What is the true solubility advantage for
amorphous pharmaceuticals. Pharm Res. 2000;17:397 404.
613
4. Gupta MK, Goldman D, Bogner RH, Tseng YC. Enhanced drug

dissolution and bulk properties of solid dispersions granulated
with a surface adsorbent. Pharm Technol. 2001;6:563 72.
5. Gupta MK, Goldman D, Bogner RH, Tseng YC. Mechanism for
further enhancement in drug dissolution from solid dispersions
granules upon storage. Pharm Dev Technol. 2002;7:103 12.
6. Gupta MK, Vanwert A, Bogner RH. Formation of physically
stable amorphous drugs by milling with Neusilin. J Pharm Sci.
2003;92:502 17.
7. Watanabe T, Wakiyama N, Usui F, Ikeda M, Isobe T, Senna M.
Stability of amorphous indomethacin compounded with silica.
Int J Pharm. 2001;226:81 91.
8. Watanabe T, Wakiyama N, Hasegawa S, Kusai A, Senna M.
Prediction of apparent equilibrium solubility of indomethacin
compounded with silica by C13 solid state NMR. Int J Pharm.
2002;248:123 9.
9. Fujii M, Okada H, Shibata Y, Teramachi H, Kondoh M,
Watanabe Y. Preparation, characterization, and tabletting of a
solid dispersion of indomethacin with crospovidone. Int J Pharm.
2005;293:145 53.
10. Watanabe T, Wakiyama N, Hasegawa S, Kusai A, Senna M.
Comparison of polyvinylpyrrolidone and silica nanoparticles as
carriers for indomethacin in a solid state dispersion. Int J Pharm.
2003;250:283 6.
11. Otsuka M, Matsumoto T, Kaneniwa N. Effect of environmental
temperature on polymorphic solid state transformation of indo
methacin during grinding. Chem Pharm Bull. 1986;34:1784 93.
12. Ali AS, Yamamoto K, El sayed AM, Habib FS, Nakai Y.
Molecular behavior of ufenamic acid in physical and ground
mixtures with orite. Chem Pharm Bull. 1992;40:1289 94.
13. Nakai Y, Fukuoka E, Nakajima S, Iida Y. Effect of grinding on
physical and chemical properties of crystalline medicinals with
microcrystalline cellulose. II. Retention of volatile medicinals in
ground mixture. Chem Pharm Bull. 1978;26:2983 89.
14. Kaneniwa N, Ikekawa A, Sumi M. A decrease in crystallinity of
Amobarbital by mechanical treatment in the presence of the
diluents. Chem Pharm Bull. 1978;26:2734 43.
15. Kinoshita M, Baba K, Nagayasu A, Yamabe K, Shimooka T.
Improvement of solubility and oral bioavailability of a poorly
water soluble drug, TAS 301, by its melt adsorption on a porous
calcium silicate. J Pharm Sci. 1997;91:362 70.
16. Friedrich H, Nada A, Roland B. Solid state and dissolution rate
characterization of co ground mixtures of nifedipine and hydro
philic carriers. Drug Dev Ind Pharm. 2005;31(8):719 28.
17. Chuang IS, Maciel GE. Probing hydrogen bonding and the local
environment of silanols on silica surfaces via nuclear spin cross
polarization dynamics. Phys Chem B. 1996;101:401 6.
18. Chuang IS, Maciel GE. A detailed model of local structure and
silanol hydrogen bonding of silica gel surfaces. J Phys Chem B.
1997;101:3052 64.
19. Fuji C. Company Literature on Neusilin, Fuji Chemical Industry,
Toyama, Japan; 1997.
20. Zawilla NH, Mohammad MAA, El Kousy NM, El Moghazy Aly
SM. Determination of aceclofenac in bulk and pharmaceutical
formulations. J Pharm Biomed Anal. 2002;27:243 51.
21. The British Pharmacopoeia 1999:34.
22. Buckton G, Choularton A, Beezer AE, Chaltham SM. The effect
of the comminution technique on the surface energy of a powder.
Int J Pharm. 1988;47:121 8.
23. Kitamura S, Miyamae A, Koda S, Morimoto Y. Effect of
grinding on the solid state stability of cexime trihydrate. Int J
Pharm. 1989;56:125 34.
24. Otsuka M, Ofus T, Matsuda Y. Effect of environmental humidity
on the transformation pathway of carbamezepine polymorphic
modications during grinding. Colloids Surf B. 1999;13:263 73.
25. Ferraz HG, Carpentieri LN, Watanabe SP. Dissolution prole
evaluation of solid pharmaceutical forms containing chloram
phenicol marketed in Brazil. Braz Arch Biol Technol. 2007;50
(1):57 65.
26. Feng T, Pinal R, Carvajal MT. Process induced disorder in
crystalline materials: differentiating defective crystals from the
amorphous form of griseofulvin. J Pharm Sci. 2008;97(8):3207 21.
27. Lian Y. Amorphous pharmaceutical solids: preparation, charac
terization and stabilization. Adv Drug Del Rev. 2001;48:27 42.
614
28. Gusseme D, Neves C, Willart JF, Rameau A, Descamps M.
Ordering and disordering of molecular solids upon mechanical
milling: the case of fananserine. J Pharm Sci. 2008;97(8):5000 12.
29. Brand B. Polymorphic transitions of cimetidine during manufac
ture of solid dosageforms. Int J Pharm. 1996;140:195 206.
30. Desprez S, Descamps M. Transformations of glassy indometha
cin induced by ball milling. J Non Cryst Solids. 2006;352:4480 5.
31. Descamps M, Willart JF, Dudognon E, Caron V. Transformation
of pharmaceutical compounds upon milling and comilling: the
role of Tg. J Pharm Sci. 2007:96:1398 407.
32. Caron V, Willart JF, Danede F, Descamps M. The implication of
the glass transition in the formation of trehalose/mannitol
molecular alloys by ball milling. Solid State Commun. 2007;
144:288 92.
33. Tsukushi I, Yamamuro O, Matsuo T. Solid state amorphization
of organic molecular crystals using a vibrating mill. Solid State
Commun. 1995;94:1013 18.
34. Lee B, Jung H. Enhanced bioavailability of poorly water soluble
aceclofenac using PEG based solid dispersion in rats, beagle dogs
and human subjects. AAPS Annual Meeting, New Orleans, LA,
USA. Pharm Sci Supplement. 1999;1(4):S614 614 November 14 18.
35. Kim T, Shin J, Lee B. Enhanced dissolution and bioavailability of
poorly water soluble aceclofenac using solid dispersion system.
Denver, Colorado, USA: AAPS, Annual Meeting; 2001 October
21 25.
36. Patel AR, Joshi VY. Evaluation of SLS: APG mixed surfactant
systems as carrier for solid dispersion. AAPS PharmSciTech.
2008;9(2):583 90.
37. Mutalik S, Anju P, Manoj K, Usha AN. Enhancement of dissolution
rate and bioavailability of aceclofenac: a chitosan based solvent
change approach. Int J Pharm. 2008;28:350(1 2):279 90.
38. Usha AN, Mutalik S, Reddy MS, Ranjith AK, Kushtagi P,
Udupa N. Preparation and, in vitro, preclinical and clinical
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
studies of aceclofenac spherical agglomerates. Eur J Pharm

Biopharm. 2008;70(2):674 83.
Dahiya S, Pathak K. Physicochemical characterization and
dissolution enhancement of aceclofenac hydroxypropyl beta
cyclodextrin binary systems. PDA J Pharm Sci Technol. 2006;60
(6):378 88.
Konno T, Kinuno K, Kataoka K. Physical and chemical changes
of medicinals in mixtures with adsorbents in the solid state. I.
Effect of vapor pressure of the medicinals on changes in
crystalline properties. Chem Pharm Bull. 1986;34:301 7.
Kim KH, Frank MJ, Henderson NL. Application of differential
scanning calorimetery to the study of solid drug dispersions. J
Pharm Sci. 1985;74:283 9.
Oguchi T. Improved dissolution of naproxen from solid disper
sions with porous additives. Yakuzaigaku 1997;57:168 73.
Sharma S, Sher P, Badve S, Pawar AP. Adsorption of meloxicam
on porous calcium silicate: characterization and tablet formula
tion. AAPS PharmSciTech. 2005;6(4):Article 76
Byrne RS, Deasy PB. Use of commercial porous ceramic particles
for sustained drug delivery. Int J Pharm. 2002;246:61 73.
Watanabe T, Hasegawa S, Wakiyama N, Usui F, Kusai A. Solid
state radical recombination and charge transfer across the
boundary between indomethacin and silica under mechanical
stress. J Solid State Chem. 2002;164:27 33.
Tong P, Taylor LS, Zogra G. Inuence of alkali metal counter
ions on the glass transition temperature of amorphous indo
methacin salts. Pharm Res. 2002;19:629 54.
Tong P, Zogra G. Solid state characteristics of amorphous
sodium indomethacin relative to its free acid. Pharm Res.
1999;16:1186 92.
Watanabe T, Ohno I, Wakiyama N, Kusai A, Senna M. Stabiliza
tion of amorphous indomethacin by co grinding in a ternary
mixture. Int J Pharm. 2002;241:103 11.

DOI: 10.1208/s12249 009 9240 3
Research Article
Investigations on the Physical Structure and the Mechanism of Drug Release
from an Enteric Matrix Microspheres with a Near-Zero-Order Release Kinetics
Using SEM and Quantitative FTIR
Wasfy M. Obeidat,1,4,5 Safwan M. Obeidat,2 and Nizar M. Alzoubi3
Abstract. The objectives of this study were to evaluate the physical structure and the release mechanisms
of theophylline microspheres made of Eudragit S 100 polymer as an enteric polymer, combined with a
nonerodible polymer, Eudragit RL 100. In the preparation process, polymer combinations (1:1) were
dissolved in an organic solvent mixture composed of acetone and methanol at a specic ratio containing a
theoretical drug loading of approximately 15%. Two microsphere formulations (LS1 and LS2) were
prepared at two different total polymer concentrations (10% in LS1 and 12.7% in LS2). Dissolution
studies were carried out using US Pharmacopeia Dissolution Apparatus II in an acidic medium for 8 h
and in an acidic medium (2 h) followed by a slightly basic buffered medium for 10 h. Both LS1 and LS2
microsphere formulations produced particles that were spherical in shape and had very narrow size
distributions with one size fraction comprising 70 80% of the yield. Scanning electron microscopy and
quantitative Fourier transform infrared were used for microsphere physical structure evaluation. Except
for the absence of drug crystals, photomicrographs of both LS microspheres after dissolution in pH 1.2
and 7.2 buffer solutions were similar to those before dissolution. Dissolution results indicated the ability
of LS microspheres to minimize drug release during the acid stage. However, in the slightly basic medium
that followed the acidic stage, the drug release was sustained and controlled in its kinetics and data tted
to Peppas equation indicated a case II transport suggesting that the drug release is mainly through
swelling/erosion mechanism.
KEY WORDS: enteric; FTIR; microspheres; narrow size; SEM.
INTRODUCTION
Theophylline, a methylxanthine alkaloid that is used in
the treatment of asthma as a bronchodilator, has a narrow
therapeutic index in the range of 5 20 g/ml and has
therefore received a considerable amount of attention in oral
sustained release formulations (1 4). The development of a
dosage form such as a microsphere is believed to enhance
tolerance and to control the release of theophylline to achieve
a safe therapeutic concentration in the blood. Controlled
release multiple unit dosage forms, such as microspheres,
have advantages over single unit ones (5 7). The emulsion
solvent evaporation method is one of these methods and has
been extensively studied to prepare such microspheres (8).
1
College of Pharmacy, University of Sharjah, P.O. Box 27272

Sharjah, UAE.
2
College of Science, Philadelphia University, P.O. Box 1 Amman,
11392, Jordan.
3
College of Pharmacy, Applied Science University, P.O. Box 926296
Amman, 11931, Jordan.
4
Jordan University of Science and Technology, P.O. Box 3030 Irbid,
22110, Jordan.
5
To whom correspondence should be addressed. (e mail: wobeidat@
sharjah.ac.ae; Obeidatw@just.edu.jo)
When using the emulsion solvent evaporation method

for preparing microspheres from binary polymers, one can
achieve either matrix structures where both polymers may be
homogeneously or heterogeneously dispersed in each other
or otherwise one polymer will engulf the other polymer
resulting in a double wall spherical matrix structure. Of
course, systems in between can also be expected (9).
In this study matrix, microspheres made of polymer
mixtures are considered for purposes of achieving a con
trolled release of theophylline. Matrix microspheres are very
rugged and can be prepared for many drugs (10). Micro
sphere characteristics are greatly affected by processing and
formulation variables (2,11,12).
Eudragit S 100 is an anionic copolymer based on
methacrylic acid and methacrylate. It is a polyacrylic resin
that has been suggested to be used in microencapsulation for
controlled release applications due to its unique solubility
prole (3,13,14). The free carboxylic acid groups make the
polymer pH sensitive, being soluble at pH 6 7.5 (15).
In a previous work (4), because microspheres composed
of an enteric or pH sensitive hydrophilic polymer (Eudragit S
100) and a nonerodible polymer hydrophobic (CAB 551
0.01) could not be produced as usual using a common solvent
(s) due to the difculty of selecting a solvent or mixture of
solvents to dissolve both polymers, Eudragit S 100 was
615
616
dispersed (not dissolved) in CAB 551 0.01/acetonitrile poly
mer solution to produce microspheres for purposes of
achieving controlled drug release. Although the microspheres
could modify drug release rates compared to release rates
from CAB 551 0.01 microspheres alone, high temperature
(65C up to 4 h) was required which would be undesirable for
heat sensitive drugs. In addition, the microspheres possessed
wrinkled nonsmooth surfaces.
The rst aim of this study was to prepare and evaluate
theophylline microspheres made of Eudragit S 100 polymer
as a pH sensitive enteric hydrophilic polymer combined with
a nonerodible hydrophilic polymer (other than CAB 551
0.01) such as Eudragit RL 100 using suitable common organic
solvent(s). After reviewing the prior art, we could not nd
such combinations of these specic polymers. Thus, this work
provides solutions for producing microspheres from such
combination and elucidates the composition of such micro
spheres. Therefore, an essential second aim of this work was
to study the distribution of polymers within the microspheres
and the microscopic shape of the entire microspheres using
scanning electron microscopy and Fourier transform infrared
(FTIR) spectroscopy.
EXPERIMENTAL
Materials
Materials used are: Eudragit S 100 (Rhm), Eudragit RL
100 (Rhm), theophylline (Sigma Aldrich Co.), magnesium
stearate (CDH, Ltd., Bombay, New Delhi, India), heptane
(GFS Chemicals, Inc., Columbus, OH, USA), mineral oil
(Ruger Chemical Co. Inc., Irvington, NJ, USA), acetone
(Carlo Erba Reactifs SDS), methanol (Carlo Erba Reactifs
SDS), potassium phosphate tribasic, and hydrochloric acid
50% and sodium hydroxide 50% w/w solution (J. T. Baker
Inc., Phillipsburg, NJ, USA).
Instruments
Instruments used are: stirrer (Lab. Stirrer DLH, VELP
Scientica, Europe), US Pharmacopeia (USP) Dissolution
Apparatus II (ERWEKA, DT D6, Fed. Rep. of Germany),
UV visible spectrophotometer CE (Cintra 5, GBC Scientic
Equipment, Bausch & Lomb, Rochester, NY, USA),
SCHOTT pH meter (CG 843, Germany), standard sieves
series, differential scanning calorimeter (DSC 50 Shimadzu,
Japan), scanning electron microscope, and FTIR spectropho
tometer (Shimadzu 8400S, Japan) with KBr pellets.
Preparation of Microspheres
Two microsphere formulations (LS1 and LS2) were
prepared using Eudragit S 100 and Eudragit RL 100 polymer
combinations at a 1:1 ratio and with two different total
polymer concentrations (10% in LS1 and 12.7% in LS2). In
the preparation process, both Eudragit S 100 and Eudragit
RL 100 polymer combinations were dissolved in a solvent
mixture composed of acetone and methanol (4.5:1). Then, the
model drug, manually micronized theophylline, was dispersed
in the polymer solution phase at a theoretical loading of 15%
before it was emulsied with ve times its volume of mineral
Obeidat, Obeidat and Alzoubi

oil containing 1% magnesium stearate. While incubated in a
water bath at 321C, the dispersion system was continuously
stirred for 2 3 h at an agitation speed adjusted for each
formulation to yield an approximately similar particle size
distribution. The stirrer consisted of two propellers on a single
shaft. After the formation of microspheres and the evaporation
of solvent, the microspheres were separated from the oil phase,
washed with n heptane, and dried at 50C. Blank LS1 and LS2
microspheres (without drug) were also prepared using the
same method and conditions described above.
Particle Size Distribution
The size distributions were evaluated by sieve analysis
using a set of standard USP sieves from 125 to 825 m. The
microspheres were placed at the topmost sieve (825 m) and
tapped by hand. The weight of microspheres retained on each
individual sieve was recorded.
Drug Loading
Drug content analysis was performed by the following
procedure. An accurately weighed sample of microspheres
was placed in a 50 ml volumetric ask and a solvent mixture
composed of methanol and phosphate buffer (pH 7.2) at 1:1
ratio was added to dissolve the polymers and the drug. After
the proper dilution and ltration using micropore lter
(0.45 m), the drug concentration was determined spectro
photometrically at 274 nm. At the specied wavelengths, no
spectrophotometric interferences were observed for blank
microspheres (without drug) dissolved in the same solvent
mixture.
In Vitro Dissolution Analysis
In vitro dissolution studies were carried out on micro
sphere samples under two different procedures: (1) an
accurately weighed sample of microspheres (100 mg) was
suspended in 900 ml acidic medium (pH 1.2 0.2) and
dissolution was followed to determine the effect of the enteric
polymer on drug release and (2) a two stage dissolution
procedure where the rst stage is at pH 1.20.2 for 2 h
followed by pH 7.2 media for 10 h or until the drug was depleted
from the microspheres. For the two stage dissolution, an
accurately weighed sample of microspheres (100 mg) was
suspended in the dissolution medium consisting of 500 ml of
0.1 N (pH 1.20.2) hydrochloric acid without enzymes and
dissolution was performed for the rst 2 h. At the end of the 2 h,
400 ml of 0.1 M tribasic sodium phosphate was added to all
dissolution vessels; the pH was adjusted to 7.2 (0.2) and the
dissolution was continued for ten more hours for a total of 12 h.
All the dissolution studies were carried out on the
microspheres at 37C (0.5C) and 100 rpm with USP
Dissolution Apparatus II. Aliquots of dissolution uid were
withdrawn at specied time intervals to assay the released
drug spectrophotometrically at 271 nm in both dissolution
experiments. Three batches of each formulation were pre
pared and evaluated. Three samples were tested from each
batch and each graphical data point was an average of
dissolution data from all three batches. Corrections were
made for the removal of samples.
Investigations on Physical Structure and Mechanism of Drug Release
617
Table I. Composition, Agitation Intensities, and Microsphere Size Fractions for the Two Microsphere Formulations (LS1 and LS2)
Formulation
Weight of
ERL 100 (g)
Weight of
ES 100 (g)
Weight of
theophylline (g)
Agitation
intensity (rpm)
LS1
0.75
0.75
0.265
1,450
LS2
0.90
0.90
0.320
2,000

Scanning electron microscope was utilized to study the
shape of the whole microspheres in terms of size uniformity,
surface smoothness and roughness before and after dissolu
tion in the two dissolution media. In addition, cut or cleaved
microspheres were evaluated. Micrographs were taken for
microspheres of different sizes of both microsphere prepara
tions (LS1 and LS2). Blank microspheres (without drug) were
also viewed before and after dissolution in phosphate buffer
(pH 7.20.2). Microspheres were prepared and mounted on a
conductive carbon adhesive attached to aluminum stubs.
Microspheres were coated with gold (Bio Rad, Polaron
division E6100). Gold was evaporated onto the specimen to
a thickness of approximately 15 nm. The sample images were
digitally captured using the FEI Quanta 200 scanning electron
microscope.
Quantitative FTIR
To investigate the way the two polymers (Eudragit S 100
and Eudragit RL 100) mix or are deposited in the micro
spheres during production, four main groups of samples were
examined by FTIR. Each group contains 12 replicates of the
same type of sample. Group 1 contains pure Eudragit S 100
raw material; group 2 contains pure Eudragit RL 100; group 3
contains 1:1 mixture of both polymers before their subjection
to the slightly basic buffer dissolution and group 4 contains
the same mixture in group 3 but after buffer dissolution. Each
sample in the four groups was obtained by grinding LS1 or
LS2 blank microspheres and then scanned by FTIR from 850
to 300 cm 1. The spectrum for each sample was recorded
after background correction. The data of each spectrum were
saved in an Excel le to form a column matrix. Then, the data
from all samples from the entire four groups were
concatenated using MATLAB 7.0.4 in one Excel le to form
a two dimensional data matrix with dimensions of (481,800).
Each column represents the FTIR data collected for one
sample. The columns in the matrix were arranged randomly.
Principal component analysis (PCA) was applied to the
resulting data matrix which was mean centered prior to
PCA application.
Microsphere size
fraction (m)
300
425
300
425
425
600
425
600
Percent from the yield

25.83.2
74.25.7
19.87.1
80.26.6
preparations and microsphere size fractions are shown in

Table I for the two microsphere formulations (LS1 and LS2).
Drug Loading
Generally, the efciency of drug loading varied from
88% (5) for 425 600 m particles to 92% (3) for 300 425
m particles in LS1 and LS2 preparations. No signicant
differences in loadings were observed between LS1 and LS2
microsphere formulations.
Blank Microspheres. Blank LS1 and LS2 microsphere
formulations were viewed after production as whole
microspheres as well as cleaved ones and compared with
micrographs after dissolution in the phosphate buffer solution
(pH 7.20.2). Both LS1 and LS2 microsphere formulations
showed basically the same characteristics, so the results of
scanning electron microscopy presented in this section are for
LS1 preparation only.
The shape and size uniformity of blank LS1 micro
spheres are shown in Fig. 1. Figure 2a, b shows the surface of
LS1 microspheres before and after buffer dissolution, respec
tively. Microspheres were similar in terms of surface uniformity,
smoothness, and homogeneity. In addition, photomicrographs

Size Distribution
At the conditions used in this work, both LS1 and LS2
microsphere formulations were produced in very narrow size
distributions with one size fraction comprising 70 80% of the
yield. Composition and agitation intensities for microsphere
Fig. 1. Scanning electron micrographs of blank LS1 microsphere

preparation from a 425 600 m sieve fraction just after production
showing the shape and size uniformity
618
Fig. 2. a Scanning electron micrographs of the surface of blank LS1 microsphere preparation from a 425 600 m sieve
fraction before buffer dissolution. b Scanning electron micrographs of the surface of blank LS1 microsphere preparation
from a 425 600 m sieve fraction after buffer dissolution
Drug Containing Microspheres. LS1 and LS2 micro

sphere formulations were viewed after production as whole
microspheres as well as cut or cleaved ones and compared

with micrographs after dissolution in an acidic medium
(pH 1.2 0.2) and in the slightly basic buffer solution
(pH 7.20.2) as well. Figure 4 shows the shape and size
uniformity of LS1 microspheres containing theophylline after
production. In addition, Fig. 5a, b shows the surfaces of LS1
and LS2 microspheres containing theophylline, respectively,
where drug crystals are present on the surfaces of both
microsphere formulations with a relatively less concentration
in LS2 microspheres compared to LS1. In both acid and
slightly basic buffer dissolution tests, LS1 and LS2 micro
sphere formulations showed basically the same character
istics, so the results of scanning electron microscopy
presented in this section are for LS1 preparation only.
Fig. 3. Cross sectional view of a cleaved LS1 blank microsphere

preparation from a 425 600 m sieve fraction after buffer dissolution
Fig. 4. Scanning electron micrographs of drug containing LS1

microsphere preparation from a 300 425 m sieve fraction just after
production showing the shape and size uniformity
of microspheres after buffer dissolution did not show evidence

of nonuniform distribution of Eudragit S 100 (such as grooves or
cavities on the surface) that would be seen if it was not evenly
distributed in the matrix. For further investigation, cut or
cleaved microspheres before and after buffer dissolution were
also viewed. The microspheres were similar in shape with the
absence of clusters of any of the polymers within the micro
spheres before dissolution and the absence of empty spaces or
cavities within the microspheres after dissolution. Figure 3
shows a cross sectional view of a cleaved LS1 blank microsphere
after dissolution.
619
Fig. 5. a Scanning electron micrographs of drug containing LS1 microsphere preparation from a 300 425 m sieve fraction
just after production showing drug crystals on the surface. b Scanning electron micrographs of drug containing LS2
microsphere preparation from a 300 425 m sieve fraction just after production showing less drug crystals on the surface
Figure 6a, b shows the shape of LS1 microspheres after acid

and buffer dissolutions, respectively. After acid and buffer
dissolution tests, theophylline dissolved out of the microsphere
surfaces. As for blank ones, drug containing LS1 and LS2
microspheres were similar in terms of surface uniformity,
smoothness, and homogeneity. In addition, similar to blank
microspheres in buffer medium, photomicrographs of drug
loaded microspheres after buffer dissolution did not show
evidence of nonuniform distribution of the Eudragit S 100 in
the microspheres (such as grooves or cavitations on the
surface) that the Eudragit S 100 would leave behind when it
dissolves out. Cleaved microspheres were also viewed before
and after buffer dissolution tests. Except for the presence of
drug crystals before dissolution, photomicrographs were simi
lar in shape with the absence of clusters of any of the polymers
within the microspheres before dissolution and the absence of

empty spaces or cavitations within the microspheres after
dissolution. Figure 7 shows a cross sectional view of a cleaved
LS1 blank microsphere after dissolution.
Dissolution Characteristics of LS1 and LS2 Microspheres
In this study, the two microsphere formulations LS1 and
LS2 were fabricated using equal proportion of Eudragit S 100
and Eudragit RL 100 polymers, with LS2 microspheres
containing higher total polymer concentrations/viscosities.
Two different size fractions of microspheres (300 425
and 425 600 m) from both LS1 and LS2 formulations, each
containing about 15.35% theoretical drug loadings of theoph
ylline, were subjected to dissolution at 37C (0.5C) and 100
Fig. 6. a Scanning electron micrographs of the surface of drug containing LS1 microsphere preparation from a 425 600 m
sieve fraction after acid dissolution. Surface drug crystals dissolved during dissolution. b Scanning electron micrographs of
the surface of drug containing LS1 microsphere preparation from a 425 600 m sieve fraction after buffer dissolution
620
Fig. 7. Cross sectional view of a cleaved LS1 microsphere prepara

tion from a 300 425 m sieve fraction after buffer dissolution
rpm with USP Dissolution Apparatus II as described in the

EXPERIMENTAL section for in vitro dissolution.
Figure 8 shows the dissolution prole of the two size
fraction microspheres of LS1 and LS2 in an acidic medium
(pH 1.20.2). It is apparent that the cumulative amounts of
drug released at the end of this stage (8 h) were about 30%,
while it was less than 18% during the rst 2 h. These results
indicated the ability of LS microspheres to minimize the drug
release during the acid stage, thus having the potential to be
used as enteric oral dosage form. Although amount of drug
released is insufcient to draw a conclusion of the drug
release model, a nearly constant drug release rate from
microsphere formulations in this acidic stage can be seen
which could be due to the presence of the drug in a relatively

more localized way in the core of the microspheres. And
neither Eudragit S 100 nor Eudragit RL 100 polymers
dissolved in the acidic medium (pH 1.20.2); however, they
swelled to various degrees allowing the creation of pores and
channels from which theophylline leached out (16,17).
A representative dissolution proles of the same size
fraction microspheres (300 425 and 425 600 m) from LS1
and LS2 preparations in the two stage media (pH 1.20.2) for
2 h followed by a slightly basic buffer medium (pH 7.20.2)
for 10 h or until the depletion of drug release from
microspheres is shown in Fig. 9.
When the pH of the dissolution medium was increased
to the slightly basic buffer medium (pH 7.20.2) after
subjecting the microspheres to dissolution for 2 h in the
acidic medium, the cumulative amount of drug released
increased substantially with the progressing times of
dissolution. Since theophylline has similar solubility in
acidic and basic media (18), it was suggested that the
increase of drug release rates in the slightly basic medium
in Fig. 9 was due to the increase of the porosity of the
microspheres due to the enhanced solubility of Eudragit S
100 (17) in this medium and its subsequent gradual
removal from the microsphere matrix structure. Addition
ally, Eudragit RL 100 swelled, allowing further increase in
drug release rates. The combination of the swelling of
Eudragit RL 100 and the erosion (dissolution) of Eudragit
S 100 within the microspheres resulted in a nearly constant
drug release rate during the dissolution in the pH 7.2
buffer medium. Models such as Higuchi model, rst order
model, and the power law equation of Ritger and Peppas
for the drug release were tted to the data obtained. The
100
90
80
70
LS1
LS1
LS2
LS2
percent release
40
percent release
50
300-425 M
425-600 M
300-425 M
425-600 M
60
50
40
30
30
LS1 300-425 M
LS1 425-600 M
20
20
LS2 300-425 M
10
10
LS2 425-600 M
0
0
10
12
time hr
0
0
10
time hr
Fig. 8. Release proles of theophylline in an acidic medium (pH 1.2
0.2) at 37C from different size fractions of microspheres from LS1
and LS2 formulations. Error bars represent the standard deviation
Fig. 9. Release proles of theophylline in the two stage media;

(pH 1.20.2) for 2 h followed by a slightly basic buffer medium
(pH 7.20.2) at 37C from different size fractions microspheres from
preparations LS1 and LS2. Error bars represent the standard
deviation
621
Table II. Values of n Exponents (Slopes), k (Intercepts), and Regression Coefcients Obtained from Fitting Peppas Equation to Drug
Release Data in the Slightly Basic Buffer Medium
Microsphere
formulation
LS1
LS1
LS2
LS2
Microsphere size
fraction, m
300
425
300
425
425
600
425
600
n exponent (slope)
in Peppas equation
Intercept k in
Peppas equation
Regression
coefcient
0.8607
0.8536
0.8721
0.8712
0.3734
0.4748
0.4915
0.6262
0.9882
0.9954
0.9959
0.9928
best t was to Ritger and Peppas model (the power law)

(19).
Mt =M1 k t n
Where Mt/M is the fraction of drug released at time t; k is
the coefcient constant which accounts for the structural
and geometrical properties of the matrix, and n is the
diffusional exponent indicative of the mechanism of drug
release. The resulting n values were found to be close to
0.85 (as shown in Table II) indicating a case II transport
(near zero order drug release kinetics) suggesting that the
drug release mechanism could be through swelling/polymer
relaxation/erosion with insignicant or negligible contribu
tion of Fickian diffusion in the drug transport process. At
the end of dissolution, the release rate decreased with time
because of the increase in the diffusion path length of the
drug. The change in diffusion path length was not only
because of the gradual depletion of the drug from the
matrix but also because of the moving boundaries such as
swelling and erosion.
This explanation of the mechanism of drug release was
not based only on the results of best t to the suggested
models but also on the results we have from other studies
such as the quantitative FTIR and the scanning electron
microscopy studies. In these studies, the two polymers were
found to be molecularly dispersed in the microsphere matrix
structure. Eudragit S 100 and Eudragit RL 100 polymers swell
in the acidic medium (16,17) before being exposed to the
buffer medium where the swelling then becomes accompa
nied with the major portion of drug being released. In
addition, Eudragit S 100 is pH sensitive that dissolves
(erodes) after swelling. Thus, these results and explanations
would support our results of the proposed mechanism of drug
release.
It was evident from the results of the dissolution
proles in the two stage media that the drug release was
controlled in both kinetics and duration. These properties of
such a dosage form are of high importance. The use of
controlled release technology in the formulation of pharma
ceutical products has become increasingly important during
the last few years since they have made signicant potential
to enhance clinical efcacy and reduce total treatment costs,
thereby providing economic value compared to conventional
immediate release dosage forms, even when the initial
acquisition costs are higher (20). In this study, both LS1
and LS2 microsphere formulations with the size fractions
300 600 m would have the potential to provide a controlled
and sustained drug release. In addition, both microsphere
formulation possessed enteric properties in the acidic medium
(pH 1.20.2) where only small percentage of the drug was

released. This property can be utilized in oral controlled
release dosage forms to protect drugs from the low pH of the
stomach and/or to protect the gastrointestinal tract from the
irritating effects of drugs.
Effect of Microsphere Size and Polymer Concentration/
Viscosity on the Dissolution Profile
The dissolution proles of LS1 and LS2 microspheres in
the acidic medium (pH 1.20.2) and in the slightly basic
buffer medium (pH 7.20.2) in Figs. 8 and 9, respectively,
showed that the drug release rate increased as the micro
sphere size decreased for both microsphere formulations.
Also, it was evident that the increase in drug release rate with
decreasing particle size was more pronounced for micro
spheres of low polymer concentration or viscosity (LS1
microspheres) in both of the dissolution media.
These observations could be attributed to the fact that an
increase in organic phase viscosity would hold the drug rmly
inside the microsphere so that the diffusion of the drug is
slowed (2).
Also, an increase of the viscosity of the polymer solution
phase could decrease the porosity of the microspheres. It was
apparent from the dissolution proles in Figs. 8 and 9 that the
initial release was decreased substantially as the polymer solution
viscosity was increased. The effect of viscosity on the initial
release could be due to the holding and the prevention of drug
particle migration to the surfaces during the solvent evaporation
Fig. 10. Two dimensional PCA score plot. Each point represents a
single spectrum acquired for each replicate from each sample
622
(migration) and during the spinning centrifugal forces created
by shearing procedure throughout the microsphere preparation.
This could be clearly seen by comparing the scanning electron
micrographs for LS1 (low concentration/viscosity) in Fig. 5a and
LS2 (high concentration/viscosity) in Fig. 5b.
Quantitative FTIR Studies
It was thought from the results of the dissolution studies
in the acidic medium (pH 1.20.2), where microspheres
exhibited minimal drug release, that double wall micro
spheres were produced wherein Eudragit S 100 polymer
engulfs Eudragit RL 100 polymer or at least dominates the
surface of the microsphere as a separate layer. To investigate
this thought, a series of experiments including surface tension
measurements, solubilities of both polymers in the common
solvent mixture, and rheological studies were conducted.
Unfortunately, except for the rheological properties, all other
results were similar for both polymer solutions and were
insufcient to draw a conclusion about the location of
polymers within the microspheres. In addition, as shown
earlier, the scanning electron microscopy studies did not reveal
the presence of polymer layers or double wall microspheres
before dissolution in the slightly basic medium; neither did it
reveal the presence of empty cavities after the dissolution.
To investigate the way the two polymers (Eudragit S 100
and Eudragit RL 100) do mix and whether a layer of Eudragit
S 100 could exist on the surface of the microspheres, samples
were examined by the FTIR as described in the EXPERI
MENTAL section. Among the samples tested were the LS
microspheres before and after dissolution in the slightly basic
buffer medium. Thus, in this study, scanning electron
microscopy and quantitative FTIR were used because of
their availability with the useful software to analyze the data,
convenience in conducting the experiments, and the high
precision results they can provide.
PCA is a multidimensional data analysis tool for investi
gating differences and similarities among samples spectra
through recognizing the pattern in a data set. PCA relies upon
extracting the eigenvectors for the covariance or correlation
matrix of the original data matrix containing the measured
variables. The rst principal component (PC) which accounts
for as much of the variability in the data as possible has the
same direction as the eigenvector associated with the largest
eigenvalue. And each succeeding component accounts for as
much of the remaining variability as possible (21).
In this study, the rst two components are identied as
PC1 and PC2. They are mutually orthogonal. The greater the
separation in this two dimensional principle component space
among the four types of samples, the greater is the statistical
difference between sample spectra (22).
The resulted PCA model accounts for more than 92% of
the total variation in the data using the rst two PCs as shown
in Fig. 10. From this model, it could be seen that pure
Eudragit S 100 polymer (diamond) and pure Eudragit RL 100
polymer (triangles) cluster in different spaces of the PCA
model. The distance between the center of the two clusters
means that they have different spectral signatures. In the
same model, it could be seen that polymer mixtures of LS
microspheres before the dissolution (lled squares) lie
between the two clusters of the pure Eudragit S 100 and

Eudragit RL 100 which means that the LS microspheres were
composed of the pure Eudragit S 100 and pure Eudragit RL 100
in a comparable ratio. To investigate the composition of the same
mixture of LS microspheres after dissolution in the slightly basic
buffer, the spectral data of the mixture were also included in the
same PCA model (empty squares). Data from the polymer
mixture of LS microspheres after dissolution clustered at the
same space of the mixture before the dissolution (lled squares)
as shown in Fig. 10. This indicated an almost similar spectral
signature of the LS microspheres before and after dissolution
and hence similar composition. Thus, upon prolonged dissolu
tion in the pH 7.2 buffer medium (48 h), it seems that Eudragit S
100 slowly but incompletely dissolved out of the microspheres. If
Eudragit S 100 was dominating the surface of the microspheres,
it would come out faster leaving only Eudragit RL 100 polymer
and the spectral data would be close to pure Eudragit RL 100
cluster. This was supported by the gravimetric studies (data not
shown) where only a few percentage of the microsphere weight
was lost with the prolonged dissolution. These results suggested
that Eudragit S 100 could be molecularly dispersed and
distributed all through the microspheres including the surfaces
that gave rise to their enteric properties.
CONCLUSIONS
Microspheres of narrow size distributions were produced
with one size fraction comprising 70 80% of the yield. LS
microspheres seemed to be composed of a homogeneous
molecular dispersion of Eudragit S 100 and Eudragit RL 100.
This was indicated to some degree by the quantitative FTIR
studies. In addition, the absence of double walled or layered
microspheres before dissolution and the absence of cavities
after dissolution in pH 7.2 buffer medium was conrmed by
scanning electron microscopy studies.
ACKNOWLEDGEMENT
The author would like to thank Professor James Price at
UGA for his remarks.
REFERENCES
1. Stithit S, Chen W, Price JC. Development and characterization
of buoyant theophylline microspheres with near zero order
release kinetics. J Microencapsul 1998;15:725 37.
2. Obeidat WM, Price JC. Viscosity of polymer solution phase and
other factors controlling the dissolution of theophylline micro
spheres prepared by the emulsion solvent evaporation method. J
Microencapsul 2003;20:57 65.
3. Obeidat WM, Price JC. Preparation and evaluation of Eudragit
S100 microspheres as pH sensitive release preparations for
piroxicam and theophylline using the emulsion solvent evapo
ration method. J Microencapsul 2006;23:195 202.
4. Obeidat WM, Obaidat IM. Effect of the dispersion of Eudragit
S100 powder on the properties of cellulose acetate butyrate
microspheres containing theophylline made by the emulsion
solvent evaporation method. J Microencapsul 2007;24:263 73.
5. Li SP, Kowalski CR, Feld KM, Grim WM. Recent advances in
microencapsulation technology and equipment. Drug Dev Ind
Pharm 1988;14:353 76.
6. Conti B, Giunchedi P, Conte U. Cellulose microparticles in drug
delivery. STP Pharm Sci 1997;7:331 42.

7. Hosny EA, El Mahrouk GM, Gouda MW. Formulation and in
vitro and in vivo availability of diclofenac sodium enteric coated
beads. Drug Dev Ind Pharm 1998;24:7661 6.
8. ODonnell PB, McGinity JW. Preparation of microspheres by
the solvent evaporation technique. Adv Drug Deliv Rev
1997;28:25 42.
9. Pekarek KJ, Jacob JS, Mathiowitz E. Double walled polymer
microspheres for controlled drug release. Nature 1994;367:258 60.
10. Sayed HAM, Price JC. Tablet properties and dissolution
characteristics of compressed cellulose acetate butyrate micro
capsules containing succinyl sulfathiazole. Drug Dev Ind Pharm
1986;12:577 87.
11. Shukla AJ, Price JC. Effect of drug loading and molecular weight
of cellulose acetate propionate on the release characteristics of
theophylline microspheres. Pharm Res 1991;8:1396 400.
12. Obeidat WM, Price JC. Preparation and in vitro evaluation of
propylthiouracil microspheres made of Eudragit RL 100 and
cellulose acetate butyrate polymers using the emulsion solvent
evaporation method. J Microencapsul 2005;22:281 9.
13. Squillante E, Morshed G, Bagchi S, Mehta KA. Microencapsu
lation of galactosidase with Eudragit L 100. J Microencapsul
2003;20:153 67.
14. Rodriguez M, Vila Jato JL, Torres D. Design of a new
multiparticulate system for potential site specic and controlled
15.
16.
17.
18.
19.
20.
21.
22.
623
drug delivery to the colonic region. J Control Release

1998;55:67 77.
Peppas NA, Wood KM, Blanchette JO. Hydrogels for oral
delivery for oral delivery of therapeutic proteins. Expert Opinion
Biological Therapy 2004;4:881 7.
Bodmeier R, Chen H. Preparation and characterization of
microspheres containing the anti inammatory agents, indo
methacin, ibuprofen and ketoprofen. J Control Release
1989;10:167 75.
Rhm Pharma GmbH. Eudragit: technical literature. Darmstadt:
Rhm Pharma GmbH; 2005.
Kimberley AL, Sanna T, Paul JS, James EP. Solubility and
permeability determination of anhydrous theophylline with
application to the biopharmaceutics classication system. Mid
land: The Dow Chemical Company Literature; 2002.
Ritger PL, Peppas NA. A simple equation for description of
solute release. II. Fickian and anomalous release from swellable
devices. J Control Release 1987;5:37 42.
Saks SR, Gardner LB. The pharmacoeconomic value of con
trolled release dosage forms. J Control Release 1997;48:237 42.
Brereton RG. Chemometrics: data analysis for the laboratory
and chemical plant. Hoboken: Wiley; 2005.
Obeidat S, Rayson G, Bai B, Anderson D. A multi source
portable LED spectrouorometer. Appl Spectrosc 2008;62:327 32.

DOI: 10.1208/s12249 009 9244 z
Research Article
Fabrication of Triple-Layer Matrix Tablets of Venlafaxine Hydrochloride
Using Xanthan Gum
Mukesh C. Gohel1,2 and Shital H. Bariya1
Abstract. The objective of present investigation was to develop venlafaxine hydrochloride layered tablets
for obtaining sustained drug release. The tablets containing venlafaxine hydrochloride 150 mg were
prepared by wet granulation technique using xanthan gum in the middle layer and barrier layers. The
granules and tablets were characterized. The in vitro drug dissolution study was conducted in distilled
water. The tablets containing two lower strengths were also developed using the same percentage
composition of the middle layer. Kinetics of drug release was studied. The optimized batches were tested
for water uptake study. Radar diagrams are provided to compare the performance of formulated tablets
with the reference products, Effexor XR capsules. The granules ready for compression exhibited good
ow and compressibility when xanthan gum was used in the intragranular and extragranular fractions.
Monolayer tablets failed to give the release pattern similar to that of the reference product. The drug
release was best explained by Weibull model. A unied Weibull equation was evolved to express drug
release from the formulated tablets. Lactose facilitated drug release from barrier layers. Substantial water
uptake and gelling of xanthan gum appears to be responsible for sustained drug release. The present
study underlines the importance of formulation factors in achieving same drug release pattern from three
strengths of venlafaxine hydrochloride tablets.
KEY WORDS: layered tablet; radar diagram; venlafaxine hydrochloride; weibull model; xanthan gum.
INTRODUCTION
An active pharmaceutical ingredient is uniformly distrib
uted within a polymer matrix in hydrophilic matrix system (1).
The drug release is extended over a much greater time from a
matrix system as compared to immediate release dosage forms.
In the recent years, preparation of matrix tablets has been
demonstrated with the publication of numerous patents and
research papers and their utilization in new products. The
widespread and successful use of such polymeric systems could
be attributed to their ease of manufacturing, relatively low cost,
high biocompatibility, favorable in vivo performance, and
versatility in controlling the release of drugs with a wide range
of physicochemical properties (2,3).
Xanthan gum is a hydrophilic polymer, secreted from
Xanthomonas campestris (a Gram negative, yellow pig
mented bacterium) (4). It is used for the fabrication of
matrices with uniform drug release characteristics (5 11).
Xanthan gum is the only bacterial polysaccharide produced
industrially on a large scale. It is a natural carbohydrate
commercially produced by fermenting glucose with the
appropriate microorganisms. Xanthan gum contains glucose
1
Department of Pharmaceutics and Pharmaceutical Technology,

L. M. College of Pharmacy, P.O. Box 4011, Navarangapur, Ahmedabad,
380 009, Gujarat, India.
2
To whom correspondence should be addressed. (e mail: mukeshgo
hel@hotmail.com)
37%, mannose 43.4%, glucuronic acid 19.5%, acetate 4.5%,

and pyruvate 4.4%. Xanthan gum swells in gastric uid to
produce a highly viscous layer around the tablet through
which the drug can slowly diffuse. This property makes
xanthan gum a useful ingredient for controlled release and
sustained release (SR) applications. Its compatibility with a
wide variety of ingredients makes it particularly effective in
these applications. Xanthan gum has been evaluated as a
hydrophilic matrix for different model drugs including
theophylline (12), cephalexin (13), prednisolone (14), indo
methacin (15), and diclofenac sodium (16).
Venlafaxine imparts its antidepressant effects by inhibit
ing the neuronal uptake of norepinephrine, serotonin, and to
a lesser extent, dopamine (17). The short biological half life
(52 h) and the fast clearance make the drug suitable
candidate for the development of once a day formulation.
Furthermore, it is an antidepressant, and so it is required to
be taken for quite a long period. The recommended dose of
venlafaxine hydrochloride is 75 to 450 mg/day. The use of
extended release formulation is associated with less nausea
and dizziness at the initiation of therapy (18). Effexor XR
capsules containing coated pellets were used as reference
product. The major advantages of multiplayer approach over
the coating method are higher productivity, shorter process
ing time, and minimum variation between and within batches.
US patents 7090867 and 6607751 and patent application
20030091634 covered the use of cellulose ether along with
microbial polysaccharide for the development of modied
624
Fabrication of Venlafaxine Hydrochloride-Layered Tablets

release tablet of venlafaxine hydrochloride. US patents
6274171 and 6403120 employed coating of spheroids (pellets)
with water insoluble excipients like ethyl cellulose. An effort
was made in the present investigation to develop functional
dosage forms of venlafaxine hydrochloride using a hybrid
technique of direct compression and wet granulation. The
objective of the present study was to obtain drug release
prole similar to that of the reference product using a simpler
method.
Venlafaxine hydrochloride Dv; 0:12:79; Dv; 0:5
10:04; and Dv; 0:943:98 was received as a gift sample
from Cadila Healthcare Ltd., Ankleshwar. Xanthan gum was
obtained from Alok International, Mumbai. Microcrystalline
cellulose (Avicel PH 101, Avicel PH 102) and lactose mono
hydrate (Pharmatose DCL 21) were received from Colorcon
Asia Pvt. Ltd., Goa. Magnesium stearate was purchased from
Laser Chemicals, Ahmedabad. Effexor XR Capsules (Wyeth
Pharmaceuticals Inc.) containing 150, 75, and 37.5 mg of
venlafaxine HCl with expiry dates 04/2011, 01/2011, and 04/
2011, respectively, were used as reference products.
Preparation and Evaluation of Venlafaxine Hydrochloride
Tablets
Modied release tablets of venlafaxine hydrochloride
were prepared by hybrid wet granulation technique. Xanthan
gum was used as a matrix forming material while Avicel PH
101 was used as a granulation facilitator and compression aid.
The drug, intragranular fraction of xanthan gum (25% of
total xanthan gum), and Avicel PH 101 were blended and
granulated with water using rapid mixer granulator (Saral
Engineering Company, Mumbai, India). The wet mass was
dried in a tray dryer at 605C temperature until loss on
drying was below 3%. The partially dried blend was sieved
through mesh #24. The granules were blended using conta
blender (Saral Engineering Company, Mumbai, India) for
10 min with the extragranular fraction of xanthan gum (batch
A1 to A3) and Avicel PH 102 (batch A3). The formulation of
monolayer tablets of venlafaxine hydrochloride (A1, A2, and
A3) is shown in Table I. The blends were lubricated with
magnesium stearate. The granules ready for compression
625
were evaluated for angle of repose, Carrs index, and
Hausner ratio. The batch size was 1,000 tablets for batches
A1 to A3. The tablets were prepared by compressing the
lubricated blend using 16 station rotary tablet press. Triple
layer tablets were prepared by putting drug free barrier
layers on either side of the middle layer. The granules of
middle layer were prepared as described above. Table I
depicts the composition of batches A4 to A9. The core was
made only of intragranular composition and the drug free
barrier layer consisted of xanthan gum, Avicel PH 102 or
pharmatose DCL 11, and magnesium stearate (extragranular
composition, Table I). Each layer was sequentially lled in
die cavity. Finally, the compression force was applied. The
Rimek triple layer tablet compression machine (Karnavati
Engineering Ltd., Mehsana, India) was used for compression
of layered tablets. The batch size was 2,000 tablets for batches
A4 to A9. The monolayer and triple layer tablets were
examined for uniformity of weight, thickness, crushing
strength, friability, and in vitro drug dissolution. The thickness
and crushing strength were measured on hardness tester (Dr.
Schleuniger Pharmatron AG, Switzerland). Friability was
measured using Roche type friabilitor (Electrolab, Mumbai)
by rotating the tablets for 4 min for 100 rotation.
Modied release tablets of venlafaxine hydrochloride with
lower strengths were also formulated. From industrial point of
view, it is always preferable to go for scale up scale down for
different strengths. The composition of granules for middle
layer and barrier layers were kept same as that of batch A8,
but the weight of barrier layer was changed to obtain release
prole similar to that of reference product. Table II shows
composition of tablets of lower strengths. The optimized
formulation in each category was compared using similarity
factor (f2 value) (19). The amount of drug released from the
three strengths were compared using surface area of the
tablets. The tablet was assumed to be cylinder in shape; hence,
the surface area was calculated using the following equation:
Surface area 2rh 2r2
Where r is radius of the tablet in centimeters and h is thickness

of the tablet in centimeters. Round punches with 10 mm
diameter were used for preparing batches A1 to A8. The
batches B1 to B4 and C1, C2 were prepared using punches
with 7.8 mm diameter. The thickness of the venlafaxine
Table I. Composition of Venlafaxine Hydrochloride Tablets

Batch code
Ingredients (mg)
A1
A2
A3
A4
A5
A6
A7
A8
A9
169.8
42.5
53.0
169.8
63.7
53.0
169.8
42.5
53.0
169.8
42.5
53.0
169.8
42.5
53.0
169.8
22.5
53.0
169.8
42.5
53.0
169.8
42.5
53.0
169.8
42.5
53.0
127.4
191.0
127.4
50.0
127.4
50.0
127.4
127.4
85.0
75.0
65.0
7.5
7.5
7.5
7.5
50.0
7.5
50.4
7.5
50.2
7.5
60.2
7.5
70.2
7.5
Intragranular composition
Venlafaxine hydrochlorideb
Xanthan gum
Avicel PH 101
Extragranular compositiona
Xanthan gum
Avicel PH 102
Pharmatose DCL 11
Magnesium stearate
a
b
For batches A4 to A9, intragranular composition and extragranular composition represent core and barrier layer, respectively
169.8 mg venlafaxine hydrochloride is equivalent to 150 mg of venlafaxine base
626
Gohel and Bariya

Table II. Composition of Venlafaxine Hydrochloride Tablets
Batch code
Ingredients (mg)
Intragranular composition
Venlafaxine hydrochloride
Xanthan gum
Avicel PH 101
Extragranular composition
Xanthan gum
Pharmatose DCL 11
Magnesium stearate
B1
B2
B3
B4
C1
C2
84.9
21.3
26.5
84.9
21.3
26.5
84.9
21.3
26.5
84.9
21.3
26.5
42.5
10.1
13.8
42.5
10.1
13.8
75.0
60.0
7.5
60.0
48.0
6.0
50.0
40.0
5.0
37.5
30.0
3.8
45.0
36.0
4.5
37.5
30.0
3.8
hydrochloride tablets were 0.60, 0.45, and 0.39 cm, respectively,

for batches A8, B3, and C1. Hence, the surface area of batches
A8, B3, and C1 was 3.45, 2.11, and 1.91 cm2, respectively. The
drug release in milligrams per square centimeter was calculated
and converted in percent value (Table III).
In Vitro Dissolution Studies
In vitro drug release study (n=3) was carried out in USP
apparatus I (Electrolab TDT 06 T, Mumbai, India) in 900 mL
of distilled water at 370.5C. Five milliliter samples were
pulled at predetermined times. The drug solution was
replaced with equal volume of distilled water. The samples
were diluted with water and analyzed at 226 nm using UV
visible spectrophotometer (Shimadzu 1700, Japan). The dis
solution study was also performed for reference products
(Effexor XR Capsules 150, 75, and 37.5 mg). The optimized
batch was also investigated for drug dissolution in distilled
water containing 10% ethanol.
Drug-Release Kinetics
In order to investigate the kinetics of drug release from
matrix tablets, the data of in vitro drug release were tted to
different models (20 24). The program was developed using
FORTRAN language for zero order, rst order, Higuchi,
Hixson Crowell, Korsmeyer Peppas, and Weibull models. The
F value was employed to select the most appropriate kinetic
model.
Water-Uptake Study
The swelling behavior of batches A8, B3, and C1 was
studied (25). The tablets (n=3) were kept in beaker contain
ing 100 mL distilled water at 372C. At selected time
points, the tablets were withdrawn, wiped with tissue paper,
and weighed. The percent water uptake by the tablet was
calculated using the following formula:
h
Percentage water uptake 100 Wt
W0=
W0
Where Wt was weight of tablet at time t and W0 was initial

weight of the tablet.
The Radar Graphs
In the dissolution study, higher or lower % drug release
than a target is permitted up to a certain limit. Shah et al.
proposed that the maximum difference can be 10% (f2 =50) for
establishing similarity (26). The reference products dissolution
data were used as ideal release pattern. The ideal % drug
release will get a score of 5 on a scale of 0 (ideal 10%) to 10
(ideal +10%). The lower and high permissible % of drug
release will get a score of 0 and 10, respectively. The score of
optimized batches (A8, B3, and C1) were calculated at each
dissolution time point.
Score 5 f%Tt
%Rt =2g
Table III. Comparison of Drug Release Prole Per Unit Surface Area of All Strengths
Drug release per unit surface area (mg/cm2)
Percent drug release
Time (h)
A8
B3
C1
A8
B3
C1
1
2
4
6
8
10
16
24
5.7
9.9
19.1
26.0
32.5
37.3
40.0
42.9
5.2
7.6
15.4
21.4
25.6
28.8
32.0
36.0
2.6
4.8
8.9
12.0
14.5
16.2
17.5
19.1
13.1
22.8
44.0
59.8
74.9
85.8
92.1
98.7
14.5
21.3
43.3
62.2
72.8
81.0
90.0
101.1
13.5
24.2
45.4
61.2
74.0
82.5
89.0
93.7

Where %Tt is percentage of drug released from test batch
while % Rt is percentage of drug released from reference
product at the same time.

Xanthan gum has ability to take up water when it comes
in contact with aqueous environment. The processing of
xanthan gum becomes difcult due to its sticky nature on
wetting, especially when it is used in higher amount. The
concept of adding a part of xanthan gum intragranularly and
a part extragranularly was adopted to achieve ease in
processibility and exibility in achieving the desired drug
release prole. Use of xanthan gum both intra and extra
granularly permits addition of higher amount of xanthan gum
compared with the classical wet granulation technique. The
concept can be adopted only for those excipients that show
good ow and compression behavior like xanthan gum. The
drug release pattern can be tailored by adjusting the
proportion of intra and extragranular fraction. The process
can be considered as a hybrid process between wet granula
tion and direct compression.
The reference product, Effexor XR capsules contained
coated pellets. The dissolution of the reference product was
performed in distilled water and was targeted for the formulated
venlafaxine hydrochloride matrix tablets. The reason for
choosing distilled water as a dissolution medium is that the
Food and Drug Administration (FDA) endorses the use of it as
a dissolution medium for the generic version of venlafaxine
hydrochloride. The reference product exhibited sustained drug
release. The result shown in Fig. 1 reveals that the drug release
was less than 25% in rst 2 h. The pellets of reference product
might be coated with water insoluble coating agent.
The granules of batches (A1 to A9) showed good ow
and compressibility as the value of the angle of repose, Carrs
index, and Hausner ratio were in the range of 22 to 26, 15%
to 18%, and 1.17 to 1.22, respectively. The developed tablets
fullled the requirements of crushing strength (>6 kp) and
friability (<1%). The problems of weight variation and
content variation were not observed. Moreover, high speed
tablet press can be used for manufacturing of tablets. It is well
known that preparation of pellets and subsequent coating
requires expertise and time.
Fig. 1. Comparative release proles of monolayer matrix tablets
627
The batches A1, A2, and A3 were compressed as mono
layer matrix tablets while batches A4 to A9 were compressed
as three layer matrix tablets. In the batches A1 and A2,
xanthan gum to venlafaxine hydrochloride ratio was 1:1 and
1.5:1, respectively. The data for in vitro dissolution show that
with increase in the amount of xanthan gum, the drug release
was decreased (Fig. 1). Neither batch A1 nor batch A2
showed release prole similar to that of the reference
product. The two most important challenges in the develop
ment of matrix tablets are slow drug release in the earlier phase
and complete drug release in the terminal phase with a fairly
uniform drug release in between. The batch A1 showed faster
drug release until 4 h and comparable drug release with the
reference product thereafter while batch A2 showed com
parable drug release up to 2 h and slow drug release thereafter.
The high aqueous solubility of venlafaxine hydrochloride and
high gel viscosity appears to be responsible for the behavior of
batches A1 and A2. Thus, it can be concluded that by varying
the amount of xanthan gum in uncoated monolayer tablets,
the required release prole could not be achieved. Our
objective was to avoid the use of time consuming two stage
procedures, i.e., compression and subsequent solvent coating or
pelletization and coating for the development of sustained
release venlafaxine HCl formulation.
The multilayered matrix system overcomes inherent
disadvantages of nonlinearity associated with diffusion con
trolled matrix devices by providing adequate drug release
rate with time (27). Few researchers developed multilayer
tablets for modulating release of active pharmaceutical
ingredient (API) from hydrophilic polymeric system (17,28
31). The use of xanthan gum has not been explored in layered
tablets. Geomatrix technology was used to reduce the
active surface area to engineer the API release at the initial
time points. Directly compressible microcrystalline cellulose
(Avicel PH 102) was added to the xanthan gum to augment
compressibility and increase weight of barrier layer. It is very
important to remember that the middle layer and barrier
layers should maintain their integrity in the layered tablets
during manufacturing, storage, and drug release study. The
excipients were selected considering the stated objectives.
The middle layer was formulated using 25% of the total
xanthan gum present in the formulation to prevent quick drug
release. The gelled particles of xanthan gum provide the
required hindrance to drug release. Figure 2 represents the
comparative release prole of monolayer matrix tablet (A3)
and triple layer matrix tablet (A4) of same composition. The
release rate is reduced in batch A4 compared to batch A3.
The probable reason could be availability of limited surface
area in batch A4. However, the drug release from batch A4
was slower than the release shown by the reference.
The release rate of API can be increased by incorpora
tion of soluble pore forming material in the barrier layers
(Batch A5), by reducing the percent of polymer in core layer
(Batch A6), or by reducing the percent of polymer in the
barrier layers (Batch A7). Figure 3 shows that incorporation
of water soluble excipients such as lactose monohydrate
(Pharmatose DCL 11) facilitated the API release rate after
2 h as desired. Batch A6 showed higher drug release than the
required release due to quick tablet erosion. Batch A7
showed drug release prole very close to the reference
product. Batches A8 and A9 were prepared to ne tune the
628
Gohel and Bariya
Fig. 2. Comparative release proles of monolayer matrix tablets

(Batch A3) with three layer matrix tablets (Batch A4)
Fig. 4. Comparative release proles of three layer matrix tablets of

venlafaxine hydrochloride 84.9 mg
drug release (Fig. 3). The batches A7, A8, and A9 showed
similarity factor f2 values of 62, 74, and 70, respectively.
The FDA has recently enforced the testing of modied
release dosage forms in dissolution media containing ethanol.
The FDA mentioned that the potentially fatal interaction of a
modied release system might be observed on consumption
with alcohol which resulted in impairment of the formulation
and dose dumping (32). Hence, effect of ethanol on release of
venlafaxine hydrochloride was studied. Ten percent concen
tration of ethanol typical of those found in alcoholic beverages
was included in dissolution medium (distilled water). The
dissolution study of batch A8 was performed using same
dissolution conditions with and without ethanol. Batch A8
showed similarity factor (f2) 92 with and without ethanol. The
matrix of xanthan gum will not collapse in presence of alcohol
since it is insoluble in alcohol (33). Thus, we can conclude that
the developed formulation is robust and is safe to take with
alcohol. Batch A8 was selected for development of tablets with
other strengths, i.e., 75 and 37.5 mg.
The Effexor XR capsules are available in three
strengths, 150, 75, and 37.5 mg. Respective strengths were
used as reference for development of venlafaxine hydrochlo
ride tablets. Four batches of venlafaxine hydrochloride with
75 mg drug content and two batches of venlafaxine hydro
chloride with 37.5 mg drug content were prepared and
evaluated (Table II). The comparative release proles are
shown in Figs. 4 and 5. The batch B1 showed slower drug

release as the thickness of barrier layer was higher compared
to that of batch A8. The batches B1, B2, B3, B4, C1, and C2
showed f2 values of 60, 70, 83, 51, 89, and 79, respectively.
Hence, the tablets of batches B3 and C1 were selected and
evaluated for swelling behavior.
It is well known that the drug release prole from the SR
tablet changes with the surface area of tablet. To correlate the
drug release rate from tablets of different strength, the
surface area of optimized batches was calculated followed
by normalization of drug release in milligrams per unit area.
For batches A8, B3, and C1, the complete release corre
sponded to 43.43, 35.61, and 19.63 mg/cm2. The drug release
was expressed in terms of percentage considering the three
computed values. The results shown in Table III reveals that
the percentage of drug release from the three optimized
batches, using calculated normalized amount of drug release
(mg/cm2), was almost identical. The percentage composition
of core layer of the three batches was identical. However, the
barrier layer composition was different. Formulation
development time can be shortened in industry simply by
focusing on barrier layer composition.
The goodness of t test was used to determine the
mechanism of drug release. The in vitro dissolution data of
batches A7, A8, and A9 were tted to different mathematical
models using software developed in FORTRAN language.
Fig. 3. Comparative release proles of three layer matrix tablets of

venlafaxine hydrochloride
Fig. 5. Comparative release proles of three layer matrix tablets

venlafaxine hydrochloride 42.5 mg
Fig. 6. Water uptake of optimized batches of venlafaxine hydrochlo

ride tablets of all strength
Weibull model showed best t. The dissolution data of

Effexor XR capsule was also subjected to model tting. The
Fishers ratio (F) was 20 for the Weibull model. The next
objective was to develop unied equation to correlates drug
release from different batches within the area of interest with
time. Weibull equation is given below (34):
M1

t =
Where M is the cumulative amount of drug released at time t,

is slope parameter (slope), and is the scale parameter
(intercept). The term was replaced by term Td (time
necessary to dissolve 63.2% of drug) using the relationship
=(Td). The unied Weibull model was:
h
i
t=T
d
e
M1
As is equal to equation of slope, the Eq. 5 can be

written as shown below:
h
M1
t=T
i0 033* X0 475
d
The linear equation of slope (r=0.99) was evolved by the

method reported by Kirilmaz using the dissolution data of
batches A7 to A9. The values of Td for batches A7, A8, and
A9 were 6.54, 6.41, and 6.13, respectively. The evolved model
629
was validated by comparing calculated and predicted release
proles. The calculated percent drug release and experimen
tal percent drug release of the optimized batch A8 can be
considered as comparable since f2 is equal to 88. Thus, by
using the unied weibull equation (Eq. 6), we can modulate
the drug release pattern. This investigation demonstrates that
the release of venlafaxine hydrochloride can be modied by
changing the amount of xanthan gum and using the triple
layer concept. The optimized batches B3 of venlafaxine
hydrochloride 75 mg and C1 of venlafaxine hydrochloride
37.5 mg also followed the Weibull model and the calculated F
values were 9 and 6, respectively.
Figure 6 shows the average value of water uptake of the
optimum batches (A8, B3, and C1). The study showed that all
three batches showed almost identical and substantial water
uptake. The water taken up by the tablet is responsible for
gelling of xanthan gum.
The radar diagrams of batches A8, B3, and C1 are shown
in Fig. 7. The dissolution pull times are shown on the
periphery of radar diagrams. The outer surface of radar
graphs shows highest score (10) while the centre shows lowest
score (0). Ideally, all the data points should fall on score line
of 5, i.e., in the middle of radar diagram. The radar diagrams
of batches A8, B3, and C1 show that the formulated batches
and reference products show almost similar dissolution at all
the time points. The sums of absolute value of difference
between reference and test at all time points were 8.7, 8.9,
and 5.9, respectively, for batches A8, B3, and C1. The low
values of computed difference quantitatively show the
difference.
CONCLUSION
The drug release rate was found to be dependent on the
percentage of xanthan gum, pore forming agent like pharma
tose DCL 11, and surface area of the formulation exposed to
the dissolution medium. The optimized formulation showed
media independent drug release in distilled water and in 10%
aqueous ethyl alcohol solution. The drug release was
explained by Weibull model. The use of unied Weibull
model is demonstrated to investigate the inuence of minor
changes in the formulation. A drug release prole similar to
that of the reference product (Effexor XR Capsule) was
achieved by adopting systemic formulation approach. The use
of radar diagram is demonstrated.
Fig. 7. Comparison of release proles of optimized batches of all strengths of venlafaxine

hydrochloride with respective reference product using radar graphs
630
REFERENCES
1. Lin SY, Lin TL. Different types of direct compressible excipients
affecting the release behavior of theophylline controlled release
tablets containing eudragit resins. Drug Dev Ind Pharm
1993;19:1613 21.
2. Durig T, Fassihi R. Guar based monolithic matrix systems: effect
of ionizable and non ionizable substances and excipients on gel
dynamics and release kinetics. J Control Rel 2002;80:45 56.
3. Peppas NA, Gurny R, Doelker E, Buri P. Modelling of drug
diffusion through swellable polymeric systems. J Membr Sci
1980;7:241 53.
4. Lachke A. Xanthan A Versatile Gum. Resonance 2004;9:25 33.
5. Talukdar MM, Mooter VD, Augustijns P, Maga TT, Verbeke N,
Kinget R. In vitro evaluation of xanthan gum as a potential
excipient for oral controlled release matrix tablet formulation.
Int J Pharm 1998;169:105 13.
6. Talukdar MM, Vercammen JP. Evaluation of xanthan gum as a
hydrophillic matrix for controlled release dosage forms. Drug
Dev Ind Pharm 1993;19:1037 46.
7. Cox PJ, Khan KA, Munday DL, Sujja areevath J. Development
and evaluation of a multiple unit oral sustained release dosage
form for S(+) ibuprofen: preparation and release kinetics. Int J
Pharm 1999;193:73 84.
8. Billa N, Yuen KH. Formulation variables affecting drug release
from xanthan gum matrices at laboratory scale and pilot scale.
AAPS PharmSciTech 2000;1:35 42.
9. Munday DL, Cox PJ. Compressed xanthan and karaya gum
matrices: hydration, erosion and drug release mechanisms. Int J
Pharm 2000;203:179 92.
10. Tobyn MJ, Staniforth JN, Baichwal AR, McCall TW. Prediction
of physical properties of a novel polysaccharide controlled
release system. Int J Pharm 1996;128:113 22.
11. Sujja areevath J, Munday DL, Cox PJ, Khan KA. Relationship
between swelling, erosion and drug release in hydrophilic natural
gum mini matrix formulations. Eur J Pharm Sci 1998;6:207 17.
12. Lu MF, Woodward L, Borodkin S. Xanthan Gum and Alginate
Based Controlled Release Theophylline Formulations. Drug
Dev Ind Pharm 1991;17:1987 2004.
13. Dhopeshwarkar V, OKeeffe JC, Zatz JL, Deete R, Horton M.
Development of an Oral Sustained Release Antibiotic Matrix
Tablet Using In Vitro In Vivo Correlations. Drug Dev Ind Pharm
1994;20:1851 67.
14. Watanabe K, Yakou S, Takayama K, Machida Y, Nagai T.
Factors affecting prednisolone release from hydrogels prepared
with water soluble dietary bers, xanthan and locust bean gums.
Chem Pharm Bull 1992;40:459 62.
15. Watanabe K, Yakou S, Takayama K, Machida Y, Isowa K, Nagai
T. Investigation on rectal absorption of indomethacin from
sustained release hydrogel suppositories prepared with water
soluble dietary bers, xanthan gum and locust bean gum. Biol
Pharm Bull 1993;16:391 4.
16. Yeole PG, Galgatte UC, Babla IB, Nakhat PD. Design and
evaluation of Xanthan gum based sustained release Matrix
tablets of Diclofenac sodium. Indian J Pharm Sci 2006;68:185 9.
Gohel and Bariya

17. Nicholas TW, Jones L, Chamberlain JC. A possible case of
venlafaxine induced Stevens Johnson syndrome. J Clin Psychia
try 2004;65:1431 2.
18. Olver JS, Burrows GD, Norman TR. The treatment of depres
sion with different formulations of venlafaxine: a comparative
analysis. Hum Psychopharmacol 2004;19:9 16.
19. Gohel MC, Patel TP, Bariya SH. Studies in preparation and
evaluation of pH independent sustained release matrix tablets of
verapamil HCl using directly compressible eudragits. Pharm Dev
Technol 2003;8:323 33.
20. Costa P. An alternative method to the evaluation of similarity
factor in dissolution testing. Int J Pharm 2001;220:77 83.
21. Higuchi T. Mechanisms of sustained action medication. Theoret
matrices. J Pharm Sci 1963;52:1145 9.
22. Ertan G, Karasulu HY, Karasulu E, Ege MA, Kose T, Guneri T.
A new in vitro/in vivo kinetic correlation method for nitro
furantoin matrix tablet formulations. Drug Dev Ind Pharm
2000;26:737 43.
23. Langenbucher F. Linearization of dissolution rate curve by the
Weibull distribution. J Pharm Pharmacol 1972;24:979 81.
24. Gohel MC, Parikh RK, Padshala MN, Sarvaiya KG, Jena DG.
Formulation and optimization of directly compressible isoniazid
modied release matrix tablet. Indian J Pharm Sci 2007;69:640 5.
25. Roy DS, Rohera BD. Comparative evaluation of rate of
hydration and matrix erosion of HEC and HPC and study of
drug release from their matrices. Eur J Pharm Sci 2002;16:193 9.
26. Shah VP, Tsong Y, Sathe P, Liu JP. In vitro dissolution prole
comparison statistics and analysis of the similarity factor, f2.
Pharm Res 1998;15:889 96.
27. Shajahan A, Poddar SS. A exible technology for modied release
of drugs: multi layered tablets. J control Rel 2004;97:393 405.
28. Conte U, Maggi L. Modulation of the dissolution proles from
Geomatrix multi layer matrix tablets containing drugs of
different solubility. Biomaterials 1996;17:889 96.
29. Maggi L, Bruni R, Conte U. High molecular weight polyethylene
oxides (PEOs) as an alternative to HPMC in controlled release
dosage forms. Int J Pharm 2000;195:229 38.
30. Siahi MR, Jalali MB, Monajjemzadeh F, Ghaffari F, Azarmi S.
Design and evaluation of 1 and 3 layer matrices of verapamil
hydrochloride for sustaining its release. AAPS PharmSciTech
2005;6(4):EE626 EE32. doi:10.1208/pt060477.
31. El Nabarawi MA. Modulation of tenoxicam release from hydro
philic matrix: Modulator membrane versus rate controlling
membrane. Chem Pharm Bull 2005;53:1083 7.
32. FDA Alert for Healthcare Professionals (July 2005): Hydro
morphone Hydrochloride Extended Release Capsules (mar
keted as Palladone). http://www.fda.gov/cder/drug/InfoSheets/
HCP/hydromorphoneHCP.pdf (accessed 08/03/08), part of http://
www.fda.gov (accessed 08/03/08).
33. Rowe RC, Sheskey PJ, Weller PJ. Xanthan gum. In: Rowe RC,
Sheskey PJ, Weller PJ, editors. Handbook of pharmaceutical
excipients, 4th edn. London: Pharmaceutical; 2003. p. 691 3.
34. Dokoumetzidis A, Papadopoulou V, Macheras P. Analysis of
dissolution data using modied versions of noyes whitney
equation and the Weibull function. Pharm Res 2006;23:256 61.

DOI: 10.1208/s12249 009 9246 x
Research Article
Preparation and In Vitro Evaluation of Solid Dispersions of Total Flavones
of Hippophae rhamnoides L.
Yan Xie,1 Guowen Li,1,2 Xiurong Yuan,1 Zhenzhen Cai,1 and Rong Rong1
Abstract. The purpose of this study was to enhance the dissolution of total avones of Hippophae
rhamnoides L. (TFH) by solid dispersions consisting of the drug and a polymeric carrier, poloxamer 188
(PXM). The solvent evaporation method was used to prepare solid dispersions. A 32 full factorial design
approach was used for optimization wherein the amount of solvent (X1) and the drug to polymer ratio
(X2) were selected as independent variables and the percentage of TFH dissolved in 10 min (Q10) was
selected as the dependent variable. Multiple linear regression analysis revealed that a suitable level of X1
and X2 was required for obtaining higher dissolution of TFH from PXM solid dispersions. Solid
dispersions were characterized by differential scanning calorimetry, X ray diffraction, Fourier transform
infrared spectroscopy, scanning electron microscopy, and dissolution tests. Characterization studies
revealed that solid dispersion of TFH PXM showed enhancement of TFH dissolution due to the
conversion of TFH into a less crystalline and/or amorphous form. In conclusion, dissolution enhancement
of TFH was obtained by preparing its solid dispersions in PXM using solvent method.
KEY WORDS: poloxamer 188; solid dispersions; solvent method; total avones of Hippophae
rhamnoides L.
INTRODUCTION
Total avones of Hippophae rhamnoides L. (TFH) are
extracted from a Chinese herbal medicine, sea buckthorn (1). It
is reported that in the high performance liquid chromatograms
of TFH, 12 compounds have been identied, such as quercetin
3 O glucoside, isorhamnetin 3 O rutinoside, quercetin, kaemp
ferol, isorhamnetin, etc. (2). With their major constituents
including quercetin, isorhamnetin, and kaempferol (3,4),
TFH have been demonstrated with most of the bioactive
properties of sea buckthorn. Animal and human studies suggest
that sea buckthorn avonoids may have antioxidant, anti
ulcerogenic, and hepatoprotective actions, which also can
scavenge free radicals, lower blood viscosity, lower blood
pressure, enhance cardiac function, and suppress platelet aggre
gation (5 7). TFH have very low and erratic oral bioavailability
due to their poor water solubility. Therefore, it is important to
introduce effective methods to enhance their dissolution, hence
their bioavailability.
Solid dispersion (SD) is dened as the dispersion of one
or more active ingredients in inert carriers at solid state
prepared by fusion, solvent, or solvent fusion methods (8 10).
This system provides the possibility of reducing the particle
Shanghai University of Traditional Chinese Medicine, Science and

Technology Center, Shanghai, 201203, Peoples Republic of China.
2
To whom correspondence should be addressed. (e mail: lgwshutcm@
163.com
size of drugs to nearly a molecular level, to transform the

drug from the crystalline to the amorphous state, and/or to
locally increase the saturation solubility (8). Therefore,
preparing solid dispersions of a poorly soluble drug with
water soluble polymers is a promising method for improving
the dissolution characteristics and bioavailability of the drug
(11). In recent years, the binary and ternary solid dispersions
were prepared to enhance the dissolution of poorly soluble
drugs to improve oral absorption of these drugs (12 14).
Many water soluble carriers have been employed for prepar
ing solid dispersions, such as polyethylene glycols, polyvinyl
pyrrolidone (15), mannitol, hydroxypropyl methylcellulose
(16), poloxamer (17), etc. In the present paper, poloxamer
188 (PXM) was used. It is a nonionic block copolymer
composed of two hydrophilic polyoxyethylene chains con
nected by a hydrophobic polyoxypropylene chain. It has been
used by researchers to increase the aqueous solubility of
poorly water soluble drugs (18 20).
Designing drug delivery formulations with the mini
mum number of trials is very crucial for pharmaceutical
scientists (21). Factorial design is an efcient method of
nding the relative signicance of number of variables and
their interaction based on the response or outcome of the
study. Optimization procedure involving factorial designs
and analysis of response surfaces is powerful, efcient, and
also proven to be a more systematic tool. It has been widely
used in the development of various dosage formulations
(22 24).
In current study, TFH were selected as a model drug and
poloxamer 188 as a polymeric carrier to prepare solid
631
Xie et al.
632
preparation, and c is the weight of PXM taken for solid

dispersion preparation.
dispersions to enhance the dissolution of TFH. Solid disper

sions of PXM TFH were prepared using solvent evaporation
method and studied systematically using an optimization
technique. Differential scanning calorimetry (DSC), Fourier
transform infrared spectroscopy (FTIR), X ray diffraction
(XRD), scanning electron microscopy (SEM), and dissolution
tests were employed to characterize the prepared solid
dispersions.
Experimental Design
To study all the possible combinations, three level full
factorial design (32) was constructed and conducted in a fully
randomized order. Percent of drug dissolved in 10 min (Q10)
and percent yield of SD were selected as dependent variables
and studied at three different levels of the selected
independent variables (factors). Details of the selected
independent variables of the 32 full factorial design are
shown in Table I. The range of a factor was chosen in order
to adequately measure its effect on the response variables.
This type of design was selected as it provides sufcient
degrees of freedom to resolve the main effects as well as the
factor interactions. A statistical model incorporating
interactive and polynomial terms is used to evaluate the
response.

Materials
Total avones of H. rhamnoides L. powder was obtained
from Sichuan Medco Pharmaceutical (Sichuan, China); polox
amer 188 was received as gift samples from BASF, Germany.
Tween 80 and ethanol were obtained from Sinopharm Chemical
Reagent Co. Ltd. (Shanghai, China).
Preparation of Solid Dispersions
Y b0 b1 X1 b2 X2 b12 X1 X2 b11 X12 b22 X22 ;
The solid dispersions of THF PXM were prepared by

the solvent evaporation method. Briey, poloxamer 188 was
dissolved in ethanol (80%) under stirring, until a clear
solution was obtained, followed by TFH addition and stirring
for 45 min. The solvent was recovered by evaporation at 60C
to 70C under vacuum (Rotavapor, Heidolph, Germany) and
used for the preparation of the next SD batch. The recovered
solvent was used for the next batch. The resultant solid
dispersions were stored in a desiccator at room temperature
for 24 h before pulverization and sieving. The yield of solid
dispersion was calculated according to the equation shown
below:

Yield

a
100;
bc
where, Y (Q10, the percentage of TFH dissolved in 10 min) is

the dependent variable; b0 is the arithmetic mean response of
the nine runs, and bi is the estimated coefcient for the factor
Xi. Here, X1 represents the amount of solvent used for SD
preparation and X2 is the drug to polymer mass ratio. The
main effects (X1 and X2) represent the average result of
changing one factor at a time from its low to high value. The
interaction terms (X1X2) show how the response changes
when two factors are simultaneously changed. The
polynomial terms (X12 and X22 ) are included to investigate
nonlinearity. The composition of the factorial design batches
SD1 to SD9 are shown in Table I. Two way analysis of
variance (ANOVA) with post hoc Tukeys test has been
applied for data comparison in the factorial design results.
Multiple linear regression analysis was used to nd out the
control factors that affects signicantly on response variables.
The statistical evaluation of the factorial design results was
where a is the weight of the solid dispersion sifted through a

#120 sieve; b is the weight of THF taken for solid dispersion
Table I. Composition of Factorial Design Batches (x SD , n 3)

Variable levels in coded form
Batch code
X1a
X2b
% Q10 SD
% YieldSD
SD1
SD2
SD3
SD4
SD5
SD6
SD7
SD8
SD9
1
1
1
0
0
0
+1
+1
+1
1
0
+1
1
0
+1
1
0
+1
68.983.28
80.251.18
85.031.97
70.543.20
95.140.95
92.061.53
71.282.15
87.331.65
88.982.17
810.78
861.23
840.94
851.53
880.64
850.95
891.72
810.89
891.16
Values represent the meanSD of three experiments

X1 the amount of solvent, X2 the drug to polymer mass ratio, Q10 the percentage of TFH dissolved in 10 min
a
Actual values: 1 150 mL; 0 200 mL; +1 250 mL
b
Actual values:1 1:2 (1.5:3 g); 0 1:4 (1.5:6 g); +1 1:6 (1.5:9 g)
Preparation and In Vitro Evaluation of Solid Dispersions
633
Table II. Factors in Preliminary Experiments and the Results of One Way ANOVA (x SD , n 3)
Level
% Q10 SD
P value
Fcrit
1:2
1:3
1:4
1:5
1:6
25 min
45 min
60 min
60%
80%
100%
150 mL
200 mL
250 mL
300 mL
84.282.92
72.940.11
86.691.09
92.211.23
87.344.24
89.350.83
92.211.23
92.670.97
89.141.64
87.762.97
92.211.23
88.752.02
83.170.56
88.112.85
92.211.23
90.410.77
3.04E08
111.07
3.48
0.055
4.88
5.14
0.056
4.85
5.14
2.42E05
27.04
3.48
Factor
The drug to polymer mass ratio
Stirring time
Alcohol content
Solvent amount
350 mL
Values represent the meanSD of three experiments. The drug to polymer mass ratio: 1:2 means 1.5:3 g, 1:3 means 1.5:4.5 g, 1:4 means 1.5:6 g,
1:5 means 1.5:7.5 g, 1:6 means 1.5: 9 g; solvent amount means 1:4 (1.5:6 g) in 150, 200, 250, 300, 350, respectively. The statistical evaluation of
the results in preliminary experiments was performed using SPSS 13.0
Characterization
generated from the standard curve (r2 =0.9999). Our previous

tests have conrmed that there was no change in the max of
TFH despite the presence of poloxamer 188 dissolved in the
dissolution medium. All tests were performed in triplicates.
In Vitro Dissolution Studies
Dissolution tests were performed with a dissolution

apparatus (RCZ 8A, Precise Apparatus of Tianjin University
Co., Ltd., China) using the paddle method according to US
Pharmacopeia XXIX (25). Samples of original TFH, physical
mixtures of TFH and polymeric carrier, and various solid
dispersions equivalent to 40 mg of TFH were added to 900
mL deionized water with 0.5% Tween 80 to satisfy the sink
condition. Paddle rotation speed was set at 100 rpm and the
temperature was maintained at 370.5C. At predetermined
intervals (5, 10, 20, 30, 45, 60, 90, and 120 min), 5 mL sample
was withdrawn from each vessel, ltered with a 0.45 m
membrane lter. The absorbance of each sample was
analyzed spectrophotometrically at 374 nm (model UV 1601
UV Visible spectrophotometer, Shimadzu). The same volume
of fresh medium was replaced after sampling. The percentage
of TFH dissolved was calculated using a regression equation
A Nicolet Nexus FTIR 670 spectrometer (Thermo Fisher

Scientic, Inc., Waltham, MA, USA) was used for FTIR
analysis. The samples were ground and mixed thoroughly
with KBr, an infrared transparent matrix. Sample disks were
prepared by compressing the mixtures. The scans were
executed from 450 to 4,400 cm 1.
performed using SPSS 13.0 software (http://www.ts.vcu.edu/

faq/software).

DSC was performed by a Perkin Elmer DSC7 differen
tial scanning calorimeter with a Pyris Series Workstation
(Perkin Elmer, Waltham, MA, USA). The accurately
weighed sample was placed in an aluminum pan and an
empty aluminum pan was used as reference. The heating rate
of the DSC run was 10C/min starting from 40C to 280C.
Table III. Results of Regression Analysis

Coefcient response Q10
FM
RM
b0
b1
b2
b11
b22
b12
91.310
87.573
2.222
2.222
9.212
9.212
0.413*
5.605*
8.095
8.095
FM full model, RM reduced model

*P 0.05, response is insignicant at this level
Xie et al.
634
Table IV. Calculations for Testing the Model in Portions
Regression
df
SS
MS
FM
RM
Error
FM
RM
5
3
733.314
669.802
146.663
223.267
0.970
0.927
0.045
0.014
9.689
10.249
3
5
45.411
108.924
15.137
21.785
df degree of freedom, SS sum of squares, MS mean of squares, R

regression coefcient, FM full model, RM reduced model
Liquid nitrogen was pumped at 20 mL/min through the

system for cooling purpose.
X Ray Diffraction Studies

Vacuum grease was applied over a glass slide to stick the
sample. About 100 mg of sample was sprinkled over it to
make a layer having a thickness of 0.5 mm. All the
experiments were performed on an X ray diffraction (XPert
PRO, PANalytical, The Netherlands) having a sensitivity of
0.1 mg. The samples were exposed to CuK radiation under
40 kV and 40 mA over the 2 range from 0 to 40 in
increments of 0.5/min every 0.033.
Scanning Electron Microscopy Analysis

SEM was used to study the surface morphology of TFH,
poloxamer 188, and the solid dispersion of TFH with
poloxamer 188. The samples were mounted on a brass stage
using adhesive carbon tape and placed in a low humidity
chamber prior to analysis. Samples were coated with gold
palladium, and microscopy was performed using a PHILIPS
environmental scanning electron microscope (XL 30 ESEM,
PHILIPS Inc.) operating at an excitation voltage of 20 kV.
TFH was found to be 23.37% and 50.88% dissolved in

5 min and in 2 h, respectively, explaining the need for
improvement of its dissolution characteristics. Therefore, a
solid dispersion technique using poloxamer 188 was
employed for dissolution enhancement of TFH in the present
investigation.
Preliminary investigations of the process parameters
revealed that factors X1 (solvent amount) and X2 (the drug
to polymer mass ratio) highly inuenced the rate of in vitro
dissolution (Table II) and, hence, were used for subsequent
systematic studies. The Q10 for the nine batches (SD1 to SD9)
showed a wide variation from 68.98 to 95.14 (Table I). The
data clearly suggested that X1 and X2 strongly inuence the
Q10. All the batches of factorial design exhibited yield greater
than 80% (Table I).
Data comparison in Table I was performed by two way
ANOVA with post hoc Tukeys test since they are normally
distributed and the variances of the different treatments are
equal. The results revealed that both X1 and X2 and the
interaction term X1 X2 strongly inuenced Q10 (P<0.001).
When factors X1 and X2 are considered separately, a
statistically signicant difference was observed between the
levels used (P<0.001), with level 2 (i.e., coded value 0, Table I)
performing better for both factors. Therefore, level 2 was
selected for further investigation.
Multivariable linear regression was performed to evalu
ate the relationship obtained between the response and
independent variables. The tted equation relating the
response Q10 to the transformed factors is shown in Table III.
It is obvious from the results that X2 (the drug to polymer
mass ratio) showed the largest positive effect, whereas the
term X1 (the amount of solvent) showed insignicant positive
effect on Q10. The quadratic term of X2 presented slight
negative effect and the quadratic term of X1 exhibited slight
positive effect on percentage of drug dissolved. However,
when both X1 and X2 were changed simultaneously, a
statistically signicant negative effect on Q10 was observed
(P<0.05). This may be probably attributed to the different
Fig. 1. Dissolution proles of optimized formulation, physical mixture, and pure drug (n 3)
Fig. 2. FITR spectra of TFH, poloxamer 188, and FITR spectra of solid dispersion
(SD5)
635
Xie et al.
636
Fig. 3. DSC curves of TFH, poloxamer 188, and solid dispersion (SD5)
Fig. 4. X ray diffraction spectra of TFH, poloxamer 188, physical mixture of TFH and
poloxamer 188, solid dispersion (SD5)
637
Xie et al.
638
Fig. 5. SEM of THF a, poloxamer 188 b, and TFH poloxamer 188 SDs (SD5) particles c
effect exhibited separately by X1 and X2 on Q10. Accordingly,

the interaction between these two factors needs to be further
investigated in future work.
The signicance levels of the coefcients b11 and b22
were found to be P=0.846 and 0.134, respectively, so they
were omitted from the full model to generate a reduced
model. The results of the regression analysis are shown in
Table III. The coefcients b0, b1, b2, and b12 were found to be
signicant at P<0.05; hence, they were retained in the
reduced model. The reduced model was tested in portions
to determine whether the coefcients b11 and b22 signicantly
affect the prediction of Q10. The results of the model testing
are shown in Table IV. The critical value of F for =0.05 is
equal to 9.28 (df=3, 3). Since the calculated value (F=9.689)
in FM and the calculated value (F=10.249) in RM are more
than the critical value (F=9.28), it may be concluded that the
interaction terms b11 and b22 did not contribute signicantly
to the prediction of Q10 and should be omitted from the full
model to generate the reduced model.
Batch SD5 (1:4 ratio) and SD6 (1:6 ratio) exhibited the
largest Q10 values, i.e., 95.14% and 92.06%, respectively.
Considering lower drug to polymer mass ratio and higher
percentage drug dissolved simultaneously, batch SD5 may be
a promising formulation batch for dissolution enhancement of
TFH. Hence, this batch was selected for subsequent physico
chemical characterization.
The dissolution proles of optimized formulation (1:4

ratio, 200 mL), physical mixture (1:4 ratio), and pure drug are
shown in Fig. 1. It is clear from the gure that the dissolution
rate of pure drug and physical mixture is very low as
compared to the optimized formulation. It is also observed
from the dissolution prole of optimized formulation that the
total quantity of the drug present in the solid dispersion gets
dissolved within 20 min.
The solid dispersion of best batch SD5 was evaluated for
physical characterization via FTIR, DSC, XRD, and SEM.
TFH and pure poloxamer 188 were also run as control. The
samples used for the study were freshly prepared (48 h in
advance) and preserved in desiccator before use.
The FTIR spectrum of TFH, poloxamer 188, and solid
dispersion are shown in Fig. 2. The characteristic peaks of
TFH can be found at 2,926, 1,654, and 1,617 cm 1 due to
stretching of C H, C=O, and C=C groups. The poloxamer
188 exhibits characteristic peaks at 3,447, 2,886, and
1,112 cm 1 due to stretching of O H, C H, and C O
groups. The peak at 1,654 and 1,617 cm 1 of the C=O and
C=C is the important characteristics of TFH. The
characteristic carbonyl stretching band of TFH was
inconspicuous in FTIR spectra of solid dispersion.
Furthermore, the peak at 2,926 cm 1 of C H in TFH is
shifted to a lower frequency of 2,877 cm 1 in the solid
dispersion. The lowering of shift to 49 cm 1 and the

disappearance of the peak at 1,654 and 1,617 cm 1 in the solid
dispersion could be attributed to the Van der Waals
interactions and/or intermolecular hydrogen bonding
between the drug and polymers, which may result in
dissolution enhancement of TFH.
Figure 3 shows the DSC curve of TFH, poloxamer 188,
and solid dispersion. The TFH, poloxamer 188, and solid
dispersion show endothermic peaks at 245.500C, 54.666C,
and 51.500C, respectively. The endothermic peak
corresponding to melting of TFH was absent in the DSC
thermogram of solid dispersion which could be attributable to
the transformation of TFH into amorphous form in the solid
dispersion or just simply because of lack of crystalline
structure of TFH in the solid dispersion.
XRD patterns of TFH, poloxamer 188, and solid
dispersion of TFH and poloxamer 188 are shown in Fig. 4.
For TFH, very sharp characteristic peaks at diffraction angle
(2) of 9.02, 9.78, 13.35, and 25.82 were observed
indicating the presence of crystalline lattice of the drug,
whereas, for pure poloxamer 188, the characteristic peaks
were observed at 19.17 and 23.33. For the prepared solid
dispersion, however, characteristic peaks for the crystalline
drug disappeared and only exhibited the characteristic peaks
of those of poloxamer 188 indicating the formation of
amorphous drug within crystalline polymer matrix. It is well
known that, at amorphous state, drug possesses higher Gibbs
energy than its more stable crystalline state exhibiting a higher
dissolution rate. This can be the reason why the dissolution
rates of total avones in solid dispersions of TFH and
poloxamer 188 were higher than those of drug themselves.
The scanning electron micrographs of TFH, poloxamer
188, and TFH poloxamer 188 SDs revealed an unformed
sheet with particles size less than 5 m for TFH (Fig. 5a),
spherical particles with smooth surface for poloxamer 188
(Fig. 5b), and a smooth irregular shaped mixed mass for SDs
(Fig. 5c). The particle size of SDs is much bigger than TFH
but similar to that of poloxamer 188 and the surface of SDs is
slightly rougher compared with poloxamer 188, suggesting
that the surface morphology of SDs is much similar to the
surface morphology of poloxamer 188.
CONCLUSIONS
The results of the experimental study conrmed that the
factors X1, X2 and X1 X2 signicantly inuence the dissolu
tion rate of TFH. The application of experimental design
techniques for optimization of formulation helps in reaching
the optimum point in the shortest time with minimum efforts.
Characterization studies revealed that solid dispersion of
TFH poloxamer 188 showed enhancement of TFH dissolu
tion probably due to the conversion of TFH into a less
crystalline and/or an amorphous form. In conclusion, dissolu
tion enhancement of TFH was obtained by preparing its solid
dispersions in PXM using solvent method.
ACKNOWLEDGMENT
This work was supported by the Shanghai City College
Scientic Research Fund for Choosing and Cultivating
Excellent Youth Teacher of China (szy 07034). The authors
are grateful to Vice professor Yue Su for her instruction on
639
the statistics analysis, Ms. Fuyuan Ye for her FTIR and X ray
recording, and Mr. Jun Li for his DSC recording.
REFERENCES
1. Lachman J, Pivec V, Hubacek J, Rehakova V. Flavonoid
substances in the fruit s of sea buckthorn (Hippophae rham
noides). Sci Agric Bohem 1985;3:169 82.
2. Chen C, Zhang H, Xiao W, Yong ZP, Bai N. High performance
liquid chromatographic ngerprint analysis for different origins
of sea buckthorn berries. J Chromatogr A 2007;1154:250 9.
3. Zhang Q, Cui H. Simultaneous determination of quercetin,
kaempferol, and isorhamnetin in phytopharmaceuticals of
Hippophae rhamnoides L. by high performance liquid chroma
tography with chemiluminescence detection. J Sep Sci 2005;28:
1171 8.
4. Zu Y, Li C, Fu Y, Zhao C. Simultaneous determination of
catechin, rutin, quercetin kaempferol and isorhamnetin in the
extract of sea buckthorn(Hippophae rhamnoides L.) leaves by
RP HPLC with DAD. J Pharm Biomed Anal 2006;41:714 9.
5. Suomela JP, Ahotupa M, Yang B, Vasankari T, Kallio H.
Absorption of avonols derived from sea buckthorn (Hippophae
rhamnoides L.) and their effect on emerging risk factors for
cardiovascular disease in humans. J Agric Food Chem.
2006;54:7364 9.
6. Pang X, Zhao J, Zhang W, Zhuang X, Wang J, Xu R, et al.
Antihypertensive effect of total avones extracted from seed
residues of Hippophae rhamnoides L. in sucrose fed rats. J
Ethnopharmacol. 2008;117:325 31.
7. Cheng J, Kondo K, Suzuki Y, Ikeda Y, Meng X, Umemura K.
Inhibitory effects of total avones of Hippophae rhamnoides L.
on thrombosis in mouse femoral artery and in vitro platelet
aggregation. Life Sci. 2003;72:2263 71.
8. Damian F, Blaton N, Naesens L, et al. Physicochemical
characterization of solid dispersions of the antiviral agent UC
781 with polyethylene glycol 6000 and Gelucire 44/14. Eur J
Pharm Sci. 2000;10:311 22.
9. Chiou WL, Riegelman S. Pharmaceutical applications of solid
dispersion systems. J Pharm Sci. 1971;60:1281 302.
10. Ford JL. The current status of solid dispersions. Pharm Acta
Helv. 1986;61:69 88.
11. Sethia S, Squillante E. Solid dispersion of carbamazepine in PVP
K30 by conventional solvent evaporation and supercritical
methods. Int J Pharm. 2004;272:1 10.
12. Liu L, Wang X. Improved dissolution of oleanolic acid with
ternary solid dispersions. AAPS PharmSci Tech. 2007;8(4):E1 5.
13. Cirri M, Maestrelli F, Corti G, Mura P. Fast dissolving tablets of
glyburide based on ternary solid dispersions with PEG 6000 and
surfactants. Drug Deliv. 2007;14:247 55.
14. Kwon SH, Kim SY, Ha KW, Kang MJ, et al. Pharmaceutical
evaluation of genistein loaded pluronic micelles for oral delivery.
Arch Pharm Res. 2007;30:1138 43.
15. Abdul Fattah AM, Bhargava HN. Preparation and in vitro
evaluation of solid dispersions of halofantrine. Int J Pharm.
2002;235:17 33.
16. Zheng X, Yang R, Tang X, Zheng L. Part I: characterization of
solid dispersions of nimodipine prepared by hot melt extrusion.
17. Chen Y, Zhang GGZ, Neilly J, Marsh K, Mawhinney D, Sanzgiri
YD. Enhancing the bioavailability of ABT 963 using solid
dispersion containing Pluronic F 68. Int J Pharm. 2004;286:69 80.
18. Passerini N, Albertini B, Gonzalez Rodriguez ML, Cavallari C,
Rodriguez L. Preparation and characterization of ibuprofen
poloxamer 188 granules obtained by melt granulation. Eur J
Pharm Sci. 2002;15:71 8.
19. Kwon GS. Polymeric micelles for delivery of poorly water
soluble compounds. Crit Rev Ther Drug Carrier Syst.
2003;20:357 403.
20. Yong CS, Yang CH, Rhee JD, et al. Enhanced rectal bioavail
ability of ibuprofen in rats by poloxamer 188 and menthol. Int J
Pharm. 2004;269:169 76.
640
21. Hamed E, Sakr A. Application of multiple response optimization
technique to extended release formulations design. J Control
Rel. 2001;73:329 38.
22. Rhee Y S, Chang S Y, Park C W, et al. Optimization of ibuprofen
gel formulations using experimental design technique for
enhanced transdermal penetration. Int J Pharm. 2008;364:14
20.
23. Patel VF, Patel NM. Statistical evaluation of inuence of
viscosity and content of polymer on dipyridamole release from
Xie et al.
oating matrix tablets: a technical note. AAPS PharmSciTech.
2007;8(Sup 3):E1 5.
24. Huang YT, Tsai TR, Cheng CJ, et al. Formulation design of
an HPMC based sustained tablet for pyridostigmine bromide
as a highly hygroscopic model drug and its in vivo/in vitro
dissolution properties. Drug Dev Ind Pharm. 2007;33:
1183 91.
25. USP. The United States Pharmacopeia (USP) XXIX. NF XXIV.
Rockville: US Pharmacopeial Convention; 2006. p. 2673.

DOI: 10.1208/s12249 009 9259 5
Research Article
Preparation and Evaluation of Differently Sulfonated StyreneDivinylbenzene
Cross-linked Copolymer Cationic Exchange Resins as Novel Carriers
for Drug Delivery
Prasert Akkaramongkolporn,1 Tanasait Ngawhirunpat,1,2 and Praneet Opanasopit1
Received 22 December 2008; accepted 23 April 2009; published online 19 May 2009
Abstract. The differently sulfonated styrene divinylbenzene cross linked copolymer cationic exchange
resins were prepared by oil in water polymerization and varied degrees of sulfonation. Several
characteristics of the obtained resins were evaluated, i.e., Fourier transform infrared spectra, the ion
exchange capacity, microscopic morphology, size, and swelling. The resin characteristics were altered in
relation to the degree of sulfonation, proving that differently sulfonated resins could be prepared. The
behavior of chlorpheniramine (CPM) loading and in vitro release in the USP simulated gastric (SGF) and
intestinal uids (SIF) of the obtained resins were also evaluated. The CPM loaded in the resinates (drug
loaded resins) increased with the increasing degree of sulfonic group and hence the drug binding site in
the employed resins. The CPM release was lower from the resins with the higher degree of sulfonic group
due to the increase in the diffusive path depth. The CPM release was obviously lower in SGF than SIF
because CPM, a weak base drug, ionized to a greater extent in SGF and then preferred binding with
rather than releasing from the resins. In conclusion, the differently sulfonated resins could be utilized as
novel carriers for drug delivery.
KEY WORDS: cationic exchange resin; chlorpheniramine maleate; different sulfonation; drug loading;
drug release; oil in water polymerization.
INTRODUCTION
Ion exchange resins are swellable cross linked copoly
mers that can reversibly interchange counterions. The resins
are organized into two main types depending upon the charge
of the counterions with which they exchange. The cationic
exchange resin contains the negatively ionizable group such
as a sulfonic group, which is capable of interchanging the
positively charged or cationic counterion. The anionic ex
change resin interchanges the negatively charged or anionic
counterion due to the existence of the positively ionizable
group such as a quaternary ammonium group (1,2).
The preparation of these two resins is quite similar,
consisting of two stages (2). First, the cross linked copolymer
bead is synthesized, to which the ionizable or ion exchangeable
group is added later. The cross linked copolymer between
styrene and divinylbenzene is commonly tailored in both resin
types. The spherical bead can be obtained using oil in water
emulsion polymerization. In this method, the monomer mixture
containing styrene, divinylbenzene, and benzoyl peroxide (as a
Department of Pharmaceutical Technology, Faculty of Pharmacy,

Silpakorn University, Nakhon Pathom, 73000, Thailand.
2
To whom correspondence should be addressed. (e mail: tanasait@
email.pharm.su.ac.th)
catalyst) is gradually added into a well stirred aqueous phase of

a stabilizing agent. Then, the polymerization is begun after the
drops of the monomer mixture are formed. Having received the
bead, the negatively ionizable group (i.e., sulfonic group) is
simply added by a treatment with concentrated sulfuric acid,
dubbed sulfonation. The addition of the positively ionizable
group (i.e., quaternary ammonium group), however, is some
what complex, requiring multiple steps. Up until present time,
when preparing a resin, the ion exchangeable group is always
added until the maximum capacity of the resin is reached.
In pharmaceutics, the resins have diversied applica
tions, the primary among which is as carriers for drug
delivery. A drug that ionizes into either a positively or
negatively charged molecule can act as an incoming counter
ion, replacing the counterion and electrically interacting with
the oppositely charged ionized group of the resins. This drug
and resin combination is interchangeably referred to as the
drug resin complex or resinate (1). The loaded drug will
substantially release from the resinate on exposure to a like
charge ion (another counterion) present in the gastrointesti
nal tract. The drug release can be tuned to a desired rate by
selecting suitable cross linked resins (3), entrapping (4), or
coating the resinate with suitable polymers (5). However, the
above application employs only the commercial resins with
the ion exchangeable site fully lled. To the best of our
knowledge, the characteristic and behavior of different
partially sulfonated resins in delivering a drug has never
been presented.
641
642
Akkaramongkolporn, Ngawhirunpat and Opanasopit

Table I. Assignment, IEC, and Yield Weight of Differently Sulfonated Resins
Sulfonation time (min)
15
28
35
40
90
210
Assigned code
IEC (meq/g)a
Resin weight (g)b
% Weight increasec
R/0
0.0000.000
3.000
0
R/15
0.0000.000
3.115
3.8
R/28
0.4800.001
3.303
10.1
R/35
2.0570.007
4.178
39.3
R/40
3.5860.002
5.159
72.0
R/90
4.1380.046
6.220
107.3
R/210
4.1440.012
6.410
113.7
Mean SD from triplicate measurements

After drying at 50C for 6 h
c
100 resin weight bead weight=bead weight , where bead weight
b
As a novel approach, this work was aimed at preparing

and characterizing the differently sulfonated resins. The
behavior of the obtained resins in terms of drug (chlorphenir
amine maleate) loading and in vitro release were evaluated
and discussed. Chlorpheniramine maleate was used as a
representative of a low dose drug. The different partially
sulfonated resins had lower ion exchange than the usual
resins with the full ion exchange capacity; thus, they were
feasible for use in delivering the low dose drug.

Materials
Styrene (Sigma Aldrich, Germany), divinylbenzene (Sig
ma Aldrich), polyvinyl alcohol (MW 85,000 124,000, 87 89%
hydrolyzed, Sigma Aldrich), benzoyl peroxide (Sigma
Aldrich), concentrated sulfuric acid (Mallinckrodt Baker
Inc., USA), chlorpheniramine maleate (Green Chemical
Company Ltd., Japan), dichloromethane (Ajax Finechem,
Australia), sodium chloride (Ajax Finechem), potassium
chloride (Ajax Finechem), sodium hydroxide (Ajax Fine
chem), potassium dihydrogen orthophosphate (Ajax Fine
chem), and concentrated hydrochloric acid (Mallinckrodt
Baker Inc.) were purchased from various suppliers and used
as received. Deionized water prepared by a water purier
(Barnstead/Thermolyne D 4745, USA) was used entirely in
this work.
3.000 g
Methods
Preparation of Differently Sulfonated Resins
The cross linked copolymer bead was prepared by oil in
water emulsion polymerization in a 2 l Erlenmeyer ask
tted with a mechanical agitator (IKA RW20, Germany) in a
temperature controllable oil bath (IKA Werke). The aqueous
phase (1.5 l) of a 0.5% (w/v) polyvinyl alcohol solution was
added to the ask, and the temperature was raised to around
65C. Under a xed stirring (400 rpm), the monomer mixture
containing styrene (75 ml), divinylbenzene (3 ml), and
benzoyl peroxide (3 g) was gradually poured into the aqueous
phase. Then, the temperature was raised to around 85C and
maintained at that temperature until the polymerization was
terminated (4 h). Thereafter, the bead was washed several
times with deionized water (totally 2 l) and methanol (totally
400 ml) and then sieved. The fraction in the range of 74
149 m (100 200 mesh) was collected, dried at 50C for 6 h in
a hot air oven, kept in a tightly closed container, and used for
sulfonation. This bead fraction (25.7 g) approximately corre
sponded to 60% of the whole beads (42.5 g) and 33% of the
employed monomers (78 g), respectively.
Prior to sulfonation, the dried copolymer bead (3 g) was
swollen by contacting with dichloromethane (12 ml) for
30 min. The swollen bead was then sulfonated with a xed
volume (30 ml) of concentrated sulfuric acid (H2SO4) in the
oil bath maintained at 70C. The slurry was periodically
shaken during the sulfonation. The degree of sulfonation was
Fig. 1. Fourior transform infrared spectra of differently sulfonated resins
StyreneDivinylbenzene Cross-linked Copolymer Exchange Resins

regulated by varying the reaction times from 0 (no sulfona
tion) to 210 min. The sulfonated resin was ltered, washed
with excess deionized water until neutral pH, and nally dried
at 50C for 6 h. The nal resin was kept in a tightly closed
container until further investigation.
Characterizations of Copolymer Bead and Differently
Sulfonated Resins
Fourier Transform Infrared Spectroscopy. The resins that
were previously dried at 100C until reaching a constant weight
were crushed and pressed into KBr pellets. The infrared (IR)
spectra of the obtained pellets were recorded over the range
400 4,000 cm 1 by a Fourier transform infrared spectro
photometer (Nicolet Magma IR system 750, USA).
643
method. In this process, 0.5 g of resin was placed into a 1.0%

(w/v) CPM solution in water (100 ml). The preliminary test
showed that this selected drug solution could provide maximum
drug loading for the resins. The mixture was allowed to come to
equilibrium for drug exchange (24 h) at 35C under constant
agitation (BioSan Environmental shaker incubator ES 20, Lat
via). A preliminary study proved that 24 h was sufcient for
achieving equilibrium. Thereafter, the obtained resinate was
washed with deionized water to remove the unloaded drug. The
resinate was dried overnight in a hot air oven at 50C and then
stored in a tightly closed container. The drug content was deter
mined by eluting 50 mg of each resinate with a 1 N KCl solution
Ion Exchange Capacity. The ion exchange capacity of

resins was determined by the salt splitting titration (2). An
accurate amount (0.1 g) of the resins was weighed and added
into a 125 ml Erlenmeyer ask containing a 2 N sodium
chloride solution (25 ml). The slurry was swirled periodically
and left for 3 h to allow for the displacement of H+ from the
resins. Thereafter, the slurry was titrated slowly with a 0.1 N
standardized sodium hydroxide (NaOH) solution using
phenolphthalein as the indicator. The ion exchange capacity
of the resins (IEC, meq/g) was calculated from:
IEC
cv
w
where c is the standardized concentration (N), v is the volume

(ml) at an endpoint of the NaOH solution, and w is the
weight (g) of determined resins.
Microscopic Morphology. The dried resins (the dry state)
and those put into a drop of excess deionized water (the wet
state) were photographed by a microscope (Olympus, Japan)
equipped with a digital camera (AnMo AM 423X DinoEye,
Taiwan). The dried resins were also viewed by a scanning
electron microscope (CamScan MX 2000, UK). Before viewing,
the samples were xed on stubs and sputter coated with gold in
a vacuum evaporator (Cressington Sputter Coater 108, UK).
Size and Swelling. The resins dried at 50C for 6 h and
those suspended in deionized water for 3 h were photo
graphed with the digital microscope. Ferets diameter of 200
particles on the images was randomly measured with a
calibrated digital micrometer of the image analysis program
(AnMo DinoCapture version 2.3.0.0, Taiwan). Then, the
average diameter of the dried (ddry) and swollen (dswell)
resins was determined, and the swelling (%) of the resins was
calculated using the following equation (6):
Swelling
dswell ddry
100:
ddry
Behavior of Differently Sulfonated Resins as Drug Carriers

CPM Loading. Chlorpheniramine (CPM) loading into the
differently sulfonated resins was carried out by the batch
Fig. 2. Photomicrographs of differently sulfonated resins
644
Fig. 3. Scanning electron micrographs of differently sulfonated resins
(200 ml) and then calculated in % w/w as the amount of

drug=amount of resinate 100 (7). The eluted drug was
assayed by a UV spectrophotometer (Perkin Elmer Lambda
2, Germany) at 261 nm.
CPM Release. In vitro CPM release was investigated in
the USP simulated gastric and intestinal uids (450 ml) by a
USP release testing apparatus I (Prolabo Dissolutest, France)
(8). Each resinate prepared from the differently sulfonated
resins was weighed to obtain an equivalent of 24 mg CPM
and then added into the release vessel. The rotation and
temperature were set at 50 rpm and 371C, respectively,
throughout testing. At the predetermined times, small
portions (5 ml) of the medium were withdrawn through a
lter and assayed by the UV spectroscopic method. The same
volume of fresh medium was returned to maintain the volume
entirely constant. The release testing was conducted in
triplicate.
Fig. 4. Scanning electron micrograph showing fractures found in R/35
Fig. 5. Average diameter in dry (square) and wet (triangle) states and
swelling (circle) of differently sulfonated resins

Preparation and Characterizations of Differently
Sulfonated Resins
The copolymer beads were prepared by oil in water
emulsion polymerization and then sulfonated for varied
reaction times to transform into the differently sulfonated
resins, the codes of which are displayed in Table I. To
evaluate the success of this process, the IR spectra and the
ion exchange capacity of the obtained resins were deter
mined. The IR spectra of the differently sulfonated resins
were shown in Fig. 1. There were signicant new peaks
around 1,126 1,217 and 3,450 cm 1 in the IR spectra of the
copolymer beads after sulfonation or, in the other words, the
differently sulfonated resins. These peaks were attributed to
the stretching vibrations of the S=O and O H of the sulfonic
group ( SO3H), respectively (9,10). The peaks obviously
appeared more prominently in the IR spectra of the resins
treated with the longer sulfonation periods, demonstrating
the higher degree of sulfonic group introduced into the resins.
In addition, the IR spectra of the obtained resins resembled
those of identical or commercial sulfonated styrene
divinylbenzene (DVB) copolymer resins reported elsewhere
(9,11). This evidence indicated that the differently sulfonated
styrene DVB cross linked copolymer resins could be
successfully prepared.
To determine the ion exchange function of the intro
duced sulfonic group, the IEC of the prepared resins as well
as the copolymer beads were determined (Table I). The
645
copolymer beads showed no ion exchange property (IEC =

0 meq/g) due to the absence of sulfonic group. In contrast, the
ion exchange property and hence IEC were found in the
sulfonated resins. The IEC increased as the resins were
treated with the longer sulfonation periods. This evidence
conrmed the successful addition of varied degrees of
sulfonic group into the resins. The sulfonic group was the
ion exchangeable site of the resins. In an ionic solution (e.g.,
NaCl as an ionic solution used in the determination of IEC),
the sulfonic group ionized and interchanged its counterion
(i.e., H+) with another cationic counterion (i.e., Na+). The
greater degree of sulfonic group added in the resins, the greater
interchange of the counterions, thus providing the higher
IEC.
Also, it was found that the weight and hence the percent
weight increase of the obtained resins in relation to the
employed copolymer bead was increased as the sulfonation
time was increased (Table I). This was due to the more
sulfonic group introduced into the resins. However, this work
could not provide the percent yield of the resins in relation to
the used reactants because the actually reacted amount of
sulfuric acid, which was added in excess for sulfonation, was
not determined.
The photomicrographs of the differently sulfonated
resins in the dry and wet states are shown in Fig. 2. The
gure depicted, especially from the wet state, the progress of
the sulfonation, proceeding from the outside into the center
of the copolymer beads. The beads without sulfonation (R/0
in Fig. 2) were seen to be totally unchanged between the dry
and wet states. In contrast, the outer region, which was
already reacted, of the different partially sulfonated resins (R/
15, /28, /35, and /40 in Fig. 2) signicantly swelled, which
clearly distinguished it from the remaining unreacted core.
The virgin cores were of irregular shapes, demonstrating the
uneven rates of sulfonation proceeding toward the bead
center in each direction. Nevertheless, the irregular shaped
cores became smaller as the sulfonation further progressed
and eventually disappeared in the completely sulfonated
resins (R/90 and /210 in Fig. 2). The fully sulfonated resins
swelled evenly throughout the beads and then returned to a
uniformly spherical shape. The swelling occurrence helped
conrm the success of the introduction of sulfonic group into
the resins. The embedded sulfonic group had strong hydro
philicity and afnity for water, enabling it to transform the
rigid cross linked copolymer into water swellable but insolu
ble gelled beads (9).
The scanning electron micrographs of the resins sup
ported the nding obtained from the photomicroscopic study
presented above (Fig. 3). The surfaces and shapes of the
different partially sulfonated resins were wavy and irregular
Table II. CPM Loading and Resinate Formulation Obtained from Differently Sulfonated Resins
Resin
R/0
R/28
R/35
R/40
R/90
R/210
CPM loaded in resinate (% w/w)

Amount of resinate formulation (mg)a
Amount of employed resin (mg)b
0
0
0
12.8
187.5
163.5
37.6
63.8
39.8
48.9
49.1
25.1
55.0
43.6
19.6
56.7
42.3
18.3
a
b
Containing equivalently 24 mg CPM

The amount of resinate formulation minus 24 mg CPM
646

regions were likely to be stretched by a greater force or, in
other words, had a greater internal tension than the adjacent
or other regions, thus making them prone to fracture (12). On
the other hand, the fully sulfonated resin beads (R/90 and /
210) swelled evenly, keeping the internal tension small
enough that no fractures occurred. Additionally, no fractures
were found in R/28 and R/0 owing to the very low and no
swelling, respectively. In fact, the fractures and breakages
could be observed even in commercial resins if treated
improperly during use; nevertheless, they did not signicantly
inuence on the ion exchange property of the resins (9).
The average diameters in the dry and wet states as well
as the swelling of the resins are illustrated in Fig. 5. In the dry
state, the average diameters of the differently sulfonated
resins were comparable (100 115 m), whereas those in the
wet state progressively increased from 103 to 211 m in
relation to the increased degree of sulfonic group introduced
in the resins. As described above, the sulfonic group brought
about the increased hydration, thus resulting in the larger wet
size and hence the higher swelling of the resins.
Behavior of Differently Sulfonated Resins as Drug Carriers
Fig. 6. CPM release from various resinate formulations prepared

from (triangle) R/28, (circle) /35, (square) /40, (diamond) /90, and
(multiplication sign) /210, respectively, as determined in a SGF and b
SIF
due to the shrinkage of the outer sulfonated region super

imposed upon the rigid irregular unreacted core. They
reverted to uniformly smooth and spherical shapes in the
completely sulfonated resins. However, according to the
scanning electron microscopy study, fractures were observed
on the hollow surface of moderately sulfonated resins, i.e., R/
35 (Fig. 4) and /40. The fully (R/90 and /210) and poorly (R/
28) sulfonated resins had no such fractures. This evidence
might indicate that the fractures did not occur during the
sulfonation step but rather during the post treatment, i.e., the
washing step where the swelling of resins occurred. During
swelling, the hollows in R/35 and /40 were the point of highest
swelling (Fig. 2 and 3). Because of uneven swelling, these
After the resin was placed in the drug solution, CPM

dissociated into a cationic molecule would exchange with the
resin counterion (H+) and then form the resinate. The CPM
loading (% w/w) in the resinates prepared from the
differently sulfonated resins are presented in Table II. As
expected, the drug loading was greater in the resins treated
with the longer sulfonation times and hence the higher degree
of sulfonic group. This is due to the fact that the sulfonic
group is the ion exchangeable or drug binding site of the
resins. As it increased, the drug loading in the resinates
accordingly increased.
According to the aforementioned trend, the different
partially sulfonated resins had the lower ion exchange and
hence the drug loading than the resins with the full ion
exchange capacity. Therefore, to deliver the same amount of
CPM (24 mg), greater amounts of resins were necessary, thus
requiring higher amounts of resinate formulations than the
resins with the full ion exchange capacity (Table II). This
point could be considered to be an advantage of the partially
sulfonated resins for use in delivering especially low dose
drugs. The increased amount of the resinate formulations
may facilitate dispensing and weighing. Additionally, it might
Fig. 7. Ionization of CPM in a SGF (pH 1.2) and b SIF (pH 7.5)

reduce the use of llers when the resinate is further prepared
in certain dosage forms, e.g., capsules.
The in vitro release of CPM from various resinates in
SGF and SIF is illustrated in Fig. 6. In both media, a similar
pattern was observed of the drug release being slower from
the resins with the higher degree of sulfonic group. This
behavior may be attributed to several causes. The rst cause
resulted from the sulfonation process, which began at the
outer shell and proceeded toward the resin center (Fig. 2).
Therefore, the resins with the higher amount of sulfonic
group would have a wider diffusive path depth (or length) for
drug passage. It should be noted that the wet size of the resins
with the higher sulfonic group was larger because of the
greater swelling (Fig. 2), simultaneously expanding the
diffusive path depth. In terms of distances, the drug located
at the deeper site near the resin center would require more
time to diffuse out. Moreover, it was less releasable as
compared with that located at the region near the resin
surface (3), which would be described in more detail below.
These effects thus led the resins with the higher degree of
sulfonation to provide a slower rate of drug release. The other
reason might be linked to the amount and hence the surface
area of the resinate formulation exposed to the release
medium. As presented in Table II, the resins with the higher
sulfonic group were correspondingly employed in lower
amounts than those with the smaller sulfonic group, offering,
qualitatively speaking, lower numbers of resinate particles in
the obtained formulation. As the number of resinate particles
in the formulation was decreased, the surface area exposed to
the release media would decrease, thus providing a slower
release rate.
The CPM release in both SGF and SIF was not complete
(Fig. 6) because it was driven by the ion exchange process
toward equilibrium (13). From surveying existing literature, it
was found that the equilibrium release of a drug was
apparently not constant but instead was likely to vary with
various factors, namely, the cross linkage and particle size of
resin (3,13,14), the amount of loaded drug (11), the type of
release medium (13), and the sampling procedure of release
testing (15). Nevertheless, the effect of the degree of resin
sulfonation on equilibrium release has never been mentioned.
As seen in Fig. 6, the equilibrium release from the resinates
seemed to decrease with increasing degree of sulfonic group
in the employed resins. This behavior could potentially be
explained by the heterogeneous nature of the cross linked
copolymer matrix. Irwin et al. (3) reported that not all ion
exchangeable sites (sulfonic group) were in the same
accessible and releasable locations within a resin. The sulfonic
group located at the deeper site near the center was less
accessible and less releasable than that at the outer shell of
the resin, offering a limitation for diffusion and release of the
loaded drug. This effect was reportedly more pronounced
with an increase in the cross linkage and the particle size of
the resin. As with this case, the resins with the higher degree
of sulfonic group had both larger diffusive path depth and
larger wet size (Figs. 2 and 5). Therefore, the extent of the
drug loaded in the less releasable site was likely to be higher
in the resins with the higher degree of sulfonic group, thus
providing the lower equilibrium release.
In existing literature, the type of release medium always
affects the drug release from resinates. In the present case,
647
the CPM release determined in SGF was obviously lower

than that in SIF (Fig. 6), although the total cation concentra
tion in SGF (104.2 mN) was higher than that in SIF
(87.0 mN), which was similar to the previous work (13). This
might indicate that the difference in the release observed was
primarily caused by the ionization of the drug rather than the
total cation concentration in the release media (SGF and
SIF). For CPM (Fig. 7), two pKa values (9.2 and 4.0) were
reported, corresponding to the ionization of the tertiary
amino group and the pyridine ring, respectively (16). In SIF
(pH 7.5), only the tertiary amino group of the drug ionized. In
contrast, not only the tertiary amino group but also the
pyridine ring of the drug was able to ionize in SGF (pH 1.2).
Therefore, the drug in SGF ionized and then preferred to
bind with the resin to a greater extent than that in SIF, thus
allowing a smaller amount of the drug to be available for
release.
CONCLUSION
In this study, the differently sulfonated resins were
successfully prepared. The resin characteristics, CPM loading,
and in vitro release were clearly affected by the degree of
sulfonic group in the prepared resins. In addition, the release
also depended on the ionization of the drug in the release
medium. The different partially sulfonated resins are novel
carriers for drug delivery and can be used for application in a
controlled drug delivery system. In spite of experiencing
surface fracture, the resins could be utilized as carriers,
especially for delivering low dose drugs, which provided the
increased amount of the resinate formulation. However,
further experiments should be performed to eliminate the
surface fracture of the resins.
ACKNOWLEDGMENTS
The authors wish to thank the Thailand Research Fund
and the Commission on Higher Education, Ministry of
Education, Thailand for its funding of this work
(MRG5280242), the Department of Pharmaceutical Technol
ogy, Faculty of Pharmacy, Silpakorn University, Nakhon
Pathom, Thailand for its instrument support, and also the
Faculty of Science, Silpakorn University for its assistance in
the SEM study.
REFERENCES
1. Borodkin S. Ion exchange resins and sustained release. In:
Swarbick J, Boylan JC, editors. Encyclopedia of pharmaceutical
technology, vol. 8. New York: Marcel Dekker; 1993. p. 203 16.
2. Harland CE. Ion exchange: theory and practice. UK: Royal
Society of Chemistry; 1994. p. 21 32, 73.
3. Irwin WJ, Belaid KA, Alpar HO. Drug delivery by ion
exchange, part III: interaction of ester prodrug of propranolol
with cationic exchange resins. Drug Dev Ind Pharm. 1987;13:
2047 66.
4. Sriwongjanya M, Bodmeier R. Entrapment of drug loaded ion
exchange particles within polymeric microparticles. Int J Pharm.
1997;158:29 38.
5. Zhang ZY, Ping QN, Xiao B. Microencapsulation and charac
terization of tramadol resin complexes. J Control Release.
2000;66:107 13.
648
6. Halder A, Sa B. Entrapment efcacy and release characteristics
of polyethyleneimine treated or untreated calcium alginate
beads loaded with propranolol resin complex. Int J Pharm.
2005;302:84 94.
7. Cuna M, Jato JLV, Torres D. Controlled release liquid emulsions
based on ion exchange particles entrapped within acrylic micro
capsules. Int J Pharm. 2000;199:151 8.
8. USP 29. The United States Pharmacopeial Convention, Rock
ville, MD; 2006. p. 3174.
9. Oliveira AJB, Aguiar AP, Aguiar MRMP, Maria LCS. How to
maintain the morphology of styrene divinylbenzene copolymer
beads during the sulfonation reaction. Mater Lett. 2005;59:1089 94.
10. Liu H, Zhang S, Nie S, Zhao X, Sun X, Yang X, et al.
Preparation and characterization of a novel pH sensitive ion
exchange resin. Chem Pharm Bull. 2005;53:631 3.
11. Akkaramongkolporn P, Yonemochi E, Terada K. Molecular state
of chlorpheniramine in resinates. Chem Pharm Bull. 2000;48:
231 4.

12. Coutinho FMB, Souza RR, Gomes AS. Synthesis, characteriza
tion and evaluation of sulfonic resins as catalysts. Eur Polym J.
2004;40:1525 32.
13. Pongjanyakul T, Prakongpan S, Rungsardthong U, Chancham P,
Priprem A. Characteristics and in vitro release of dextromethor
phan resinates. Powder Technol. 2005;152:100 6.
14. Burke GM, Mendes RW, Jambhekar SS. Investigation of the
applicability of ion exchange resins as a sustained release drug
delivery system for propranolol hydrochloride. Drug Dev Ind
Pharm. 1986;12:713 32.
15. Pongjanyakul T. Effect of sampling procedures of release testing
on drug release and scale up production feasibility of multiple
unit dextromethorphan resinate tablets: a technical note. AAPS
PharmSciTech. 2007;8:E117.
16. Villoslada IM, Gonzalez F, Rivas BL, Shibue T, Nishide H.
Tuning the pKa of the antihistamine drug chlorpheniramine
maleate by supramolecular interactions with water soluble
polymers. Polymer. 2007;48:799 804.

DOI: 10.1208/s12249 009 9248 8
Research Article
Assessment of Feasibility of Maillard Reaction between Baclofen and Lactose
by Liquid Chromatography and Tandem Mass Spectrometry, Application to Pre
Formulation Studies
Farnaz Monajjemzadeh,1 Davoud Hassanzadeh,1,4 Hadi Valizadeh,1,5,7 Mohammad R. Siahi-Shadbad,1,5
Javid Shahbazi Mojarrad,2,6 Thomas Robertson,3 and Michael S. Roberts3
Abstract. The aim of this study was to determine any possible, baclofen lactose Maillard reaction
products. Granules and tablets of baclofen and lactose were prepared and maintained in heat ovens for a
certain time period. The effects of lactose type, addition of magnesium stearate, and water were
monitored. Heated lactose and baclofen were analyzed using reverse phase HPLC. Liquid chromatog
raphy tandem mass spectroscopy revealed nominal mass values consistent with baclofen lactose, early
stage Maillard reaction condensation products (ESMRP). Multiple reaction monitoring confirmed the
presence of ESMRP as well. FTIR analysis proved the formation of imine bond. The results indicated
that baclofen undergoes a Maillard type reaction with lactose.
KEY WORDS: baclofen; lactose; LC MS/MS; Maillard reaction; solid state incompatibility.
(3) Formation of heterocyclic nitrogen compounds.

Browning occurs at this stage.
INTRODUCTION
Maillard reaction is named after Louis Maillard, who
reported over 95 years ago that some amines and reducing
carbohydrates react to produce brown pigments (1). The
reaction is actually a series of complex reactions between
reducing sugars such as D glucose and free amino groups of
amino acids, peptides, or proteins (2,3). The mechanism of
the Maillard reaction is very complicated; however, it is
generally divided into three stages (4):
(1) The first stage involves the sugar amine condensa
tion and the Amadori rearrangement. The reaction
steps have been well defined and no browning occurs
at this stage
(2) The second stage involves sugar dehydration and
fragmentation, and amino acid degradation via the
Strecker reaction especially at high temperatures.
1
Department of Pharmaceutics, Tabriz University of Medical Scien

ces, Tabriz, Iran.
2
Department of Medicinal Chemistry, Tabriz University of Medical
Sciences, Tabriz, Iran.
3
Therapeutics Research Unit and Southern Clinical Division, Uni
versity of Queensland, Princess Alexandra Hospital, Brisbane,
Queensland 4102, Australia.
4
Drug applied research center, Tabriz University of Medical Sciences,
Tabriz, Iran.
5
Research center for Pharmaceutical nanotechnology, Tabriz University
of Medical Sciences, Tabriz, Iran.
6
Biotechnology Research Center, Tabriz University of Medical
Sciences, Tabriz, Iran.
7
To whom correspondence should be addressed. (e mail: valizadeh@
tbzmed.ac.ir)
Generally, the first product of this reaction is a simple

glycosylamine (5), which readily undergoes the Amadori
rearrangement to produce 1 amino 1 deoxy 2 ketoses (6).
The later stages are complex and variable, depending on the
reaction conditions, and involve further dehydration and
fission of the initial reaction products (7). Melanoidines are
the final products of the reaction and are actually brown
nitrogen containing polymeric substances that decompose
with difficulty (8,9). Browning is usually measured spectro
photometrically (at 490 nm) and expressed in absorbance
units (8).
Maillard reaction has been extensively studied and
reviewed, especially in the food and nutrition related litera
ture and it is a well documented process for the degradation
of lactose during the heating of milk (3,10 14). Nowadays,
mass spectrometry (MS) has been adopted by many research
ers attempting to reveal the details of Maillard chemistry
(9,15 18). Whereas traditional mass analysis would provide
the fragment ions resulting from all compounds contained
within a sample, tandem mass spectrometry provides infor
mation regarding the fragment ions that originate from a
single molecular ion (9). Mass spectroscopy along with a
separation technique such as HPLC results in an extremely
powerful analytical technique (9). Fay et al. demonstrated the
benefits of tandem MS in their study of a pentose glycine
reaction system (19).
Among pharmaceutical excipients, lactose, a reducing
sugar, is widely used as a filler in tablet and capsule for
mulations (20). A survey of the Physicians Desk Reference
database shows that there are many pharmaceutical formula
649
Monajjemzadeh et al.
650
tions where amino compounds and lactose are both present
(21,22). In formulations, solid state reactions of the drug
substance are of great interest, especially in cases where the
drug substance is intrinsically and chemically reactive or
unstable (23). Recently, the possible reaction of amine groups
of drug entities with carbonyl groups of common tablet
excipients, such as lactose, starch, and cellulose has gained a
pharmaceutical interest (7,18,22 26). Baclofen is an amine
containing skeletal muscle relaxant, which is used to relieve
the signs and symptoms of spasticity due to spinal cord injury
or multiple sclerosis (27). It is a zwitterion containing both
carboxylic and amine moiety in its structure, which makes it a
good candidate to react with a reducing sugar such as lactose.
Solid phase chemical reaction is dependent on physical
contact between solid components. It has been shown that
extensive solid state reactions can occur after tableting of
pharmaceutical formulations which is due to increased
contact between reactants.
Although the chemistry of the Maillard reaction is well
known, detailed data using mass spectroscopy in pharmaceu
tical technology are limited and needs more examples with
detailed mechanism of the reaction to show that the reducing
sugars should be avoided in formulating of amine containing
drugs in the pharmaceutical industry.
Recently, Cutrignelli et al. studied the comparative
effects of some hydrophilic excipients on the rate of
gabapentin and baclofen lactamization in lyophilized formu
lations. They suggested that a possible gabapentin/lactose
(not baclofen/lactose) Maillard type interaction results in a
moderate increase in lactam formation (28). This study was
generally about the rate and extent of baclofen and gaba
pentin lactamization. To the best of our knowledge, evalua
tion of the Maillard reaction products of baclofen and lactose
has not been investigated yet. In the present study, different
mixtures of the drug and excipient have been studied by
HPLC, FTIR, visible spectrophotometry, and tandem mass
spectroscopy to determine the possible early stage Maillard
reaction condensation products and final stage brown color.
The effect of lactose type, temperature conditions, and the
presence of magnesium stearate has been studied as well.
The other novelty of this study is that not only the
HPLC, mass analysis, and FTIR is done to prove the possible
Maillard reaction between the drug and excipients but also
visible spectrophotometry as a less sophisticated method of
evaluating the Maillard reaction which was widely used in the
past, has been investigated simultaneously to show that still,
with some considerations, spectroscopic assessment of the
observed browning phenomena is a remarkable tool to
evaluate this type of reaction in pharmaceutical sciences.
This study also introduces a new stability indicating HPLC
method with regard to baclofen/lactose Maillard reaction
products, using internal standard.

Materials
Baclofen (R,S Amino 3 (4 chlorophenyl) butyric acid)
was obtained from Pfizer groups, Istanbul, Turkey. Lactose
monohydrate (Pharma grade 200 Mesh) and anhydrous
lactose provided from DMV Chemical Co, Netherlands. All

chemicals were of HPLC or analytical grade obtained from
LAB scan analytical science, Ireland. Commercial tablets of
baclofen (Zahravi, Kimidaroo, and Clofen) were purchased
directly from Iran and Australia.
Methods
Analytical Methods
HPLC. The HPLC system consisted of a SCL 10A XL
auto injector, SCL 10A VP system controller, LC 10AT liquid
chromatograph and a SPD M10AVP, UV VIS, photo diode
array (PDA) detector and a FRC 10A fraction collector, all
from Shimadzu (Kyoto, Japan).
Samples were injected onto a Hypersil C18 BDS column
(100 mm, 4.60 mm, 5 m; Phenomenex, Torrance, CA)
maintained at ambient temperature. Mobile phase was 13%
acetonitrile in phosphate buffer 25 mM and pH was adjusted
to 9.3, using sodium hydroxide. Flow rate was 1 mL/min with
detection at 220 nm. Data were analyzed with Class VP
software (version, 6.14 SP1). A solution of 3 methyl salicylic
acid (0.5 mg/mL in mobile phase) was used as the internal
standard. Internal standard solution (10 L) was added to
each experimental sample (100 L). The analytical method
was validated with respect to parameters such as linearity,
intermediate precision, accuracy, and selectivity (29,30).
LC MS/MS. The LC system consisted of a SIL 10AD
VP auto injector, SCL 10A VP system controller, LC
10ADVP liquid chromatograph and a DGU 12A degasser,
all from Shimadzu (Kyoto, Japan).
Samples were introduced into the mass spectrometer
through a C18 Gemini column (25200 mm, Phenomenex)
eluted at a flow rate of 0.2 mL/min, at ambient temperature.
Elution was performed, with 97% solvent A (0.1% formic
acid in water) and 3% solvent B (0.1% formic acid, 90%
acetonitrile, 5% methanol, and 5% water). Mass spectromet
ric detection was performed with an Applied Biosystems
MDS Sciex (Ontario, Canada) API 2000 triple quadrupole
mass spectrometer equipped with an electrospray ionization
(ESI) interface in the positive ion mode. The tandem mass
spectrometer was operated at unit resolution in the multiple
reactions monitoring mode (MRM), monitoring the transition
of the protonated molecular ions to the product ions. Q1 was
used from 150 600 amu in a mass resolving mode to select
the parent ion. The ion source temperature was maintained at
350C. The ionspray voltage was set at 5,500 V. The curtain
gas (CUR; nitrogen) was set at 15 and the collision gas
(CAD) at 7. The collision energy (CE), declustering potential
(DP), focusing potential (FP) and entrance potential (EP),
were set at 40, 75, 200, and 8 V, respectively. This system was
set to the multiple reaction monitoring (MRM) mode, that is,
selecting precursor ions, dissociating them and finally analyz
ing the product ions reaching the high selectivity and
sensitivity of this mode for mass analysis and detection. In
the MRM mode, data acquisition and processing were
accomplished using the Applied Biosystems Analyst version
1.4.1 software.
Baclofen and Lactose Maillard Reaction Products
651
Spectrophotometer. As mentioned before, brown poly

mers are formed at the final stages of the Maillard reaction
(4,13,31). In order to determine the intensity of brown color
samples were monitored spectrophotometrically using a plate
reader (Spectra Count Packard, Meriden, USA) in a Nunc
96 well plate with 490 nm filter and the optical density (OD)
was recorded.
reversed phase chromatography and LC MS/MS. The intensity

of the brown color was also measured at 490 nm. Different
samples of baclofen (solid state solution, pH=9.2), aqueous
mixture of baclofen and lactose (pH=9.2) and commercial
tablets were heated in order to yield degradation products. All
solid and liquid samples were heated in 90C ovens and 60C
water bath for 24 and 72 h, respectively.
FTIR. Infrared spectra were recorded on a Bomem MB

100 spectrometer (Bomem, Quebec, Canada), in the range of
400 4,000 cm 1, using KBr disks. Resolution was 4 cm 1,
37 scans/min. The spectrum was a mean of ten consecutive
scans on the same sample. Processing of the FTIR data was
performed using GRAMS/32 version 3.04 (Galactic Industries
Corporation, Salem, NH).
Tablet preparation. Baclofen and lactose either anhy

drous or monohydrate were mixed in 1:1, 1:5, and 1:10 w/w ratios.
Binary mixtures were subjected to dry granulation by slug
preparation technique. Granules were directly compressed
to tablets and were kept in silica gel containing desiccators
for 10 days. Dry samples were kept in well closed contain
ers. The containers were placed inside silica gel chambers
and maintained at 25, 40, 50, and 60C in ovens for a 6
month period. The relative humidity in the chambers was
kept at almost zero which was confirmed by a hygrometer
(Lutron HD 3008 Hygrometer, Taiwan). Each sample was
dissolved in mobile phase and analyzed using HPLC DAD
(diode array detection), LC MS/MS, and FTIR. The
intensity of the brown color was also measured spectro
photometrically (at 490 nm).
Formulation Methods
Preparation of baclofen lactose adduct mixture. Baclofen
(0.5 g) and lactose monohydrate (2.5 g) were dissolved in
50 mL of United States Pharmacopoeia (USP) borate buffer
(0.1 M, pH=9.2) with the aid of stirring and ultrasound (30).
The ionic strength of the solution was adjusted to 25 mM by
sodium chloride. Triethylamine was added in an equimolar
ratio with baclofen to aid solubility. The clear solution was
then refluxed at 60C using a water bath (Contherm Scientific
Ltd, New Zealand) for 12 h and dried overnight at the same
temperature in an open Pyrex beaker using a heat oven.
The dried mixture is referred to as the adduct mixture. Adduct
mixtures were dissolved in mobile phase to get 1 mg/mL
concentration with respect to the baclofen and was subjected to
Commercial tablets and screening tests. The Serajuddin

(32) method with minor modifications was employed to
monitor the probable solid state interaction of drug with
excipient in an accelerated manner. Briefly, drug powders
with 20% added water was kept in well closed 4 mL HPLC
grade glass vials at 95C for 12 h. The total weight of the
drug:excipient blend in a vial was kept at 200 mg for each of
the three brand formulations. Controls were done using the
Table I. Composition and Assay Results of Screening Samples

Composition
Samplesa
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
a
b
c
Baclofen
+b
+
+
+
+
+
+
+
+
+
Brand
Brand
Brand
Brand
Brand
Brand
Mg stearate
1
1
2
2
3
3
+
+
+
+
Physical mixtures of the contents

Presence
Absence
Lactose
anhydrous
+
+
+
+
+
+
Assay
lactose
monohydrate
+
+
+
+
+
+
water
Baclofen (%)
Unknown 1
83.96
1.34
92.21
1.15
82.75
1.21
90.77
2.11
89.39
76.24
91.79
0.49
80.09
0.55
77.42
0.83
0.08
1.03
0.11
0.24
0.49
0.45
0.58
0.51
0
0
0.73
0.08
0.46
0.18
0.31
1.18
OD
0
0.191
0.016
0.212
0.015
0.192
0.016
0.233
0
0
0.01
0.43
0.00
0.41
0.07
0.50
0
0
0
0
652
Table II. Intermediate Precision and Accuracy of the HPLC Method
Actual
concentration
(g/ml)
62.5
125
250
500
Measured concentration
(mg/ml), RSD (%)
Intra day
Inter day
0.0614, 1
0.126, 2
0.248, 1.4
0.49, 0.9
0.616,
0.124,
0.258,
0.512,
0.6
1.9
2.1
1.2
Accuracy (%)
Intra day
Inter day
98.24
100.8
99.2
98.0
98.56
99.2
103.2
102.4
same procedure without lactose. The composition of the

screening samples is listed in Table I. Samples were analyzed
using HPLC to calculate the amount of the remaining drug.
The brown color was measured spectrophotometrically at
490 nm.
Screening tests were conducted to examine the effect of
lactose type (monohydrated or anhydrous), addition of water
and magnesium stearate as a widely used lubricant. Mixtures
were prepared according to Table I and kept at 95C for 12 h.
Samples were analyzed to determine the amount of remaining
baclofen, possible Maillard reaction adduct/s and brown color
Fig. 1. HPLC chromatogram of a baclofen, b adducts mixture (heated baclofen and lactose), c heated baclofen in aqueous media, d intact
brand 3. (A) baclofen, (B) internal standard, (C) unknown l, (D) unknown 2, (E) unknown 3, (F) unknown 4. Note: in a, b, and c flow rate is 1
but in d flow rate 2
653
formation according to specified methods. Each experiment

was done in duplicate.
The presence of lactose in selected formulations was
examined according to the British Pharmacopoeia (BP), that
is heating of a mixture of lactose (equivalent to 0.25 mg) with
added ammonia (5 mL) and water (5 mL). Development of
red color confirms the presence of lactose in the formulations
(33). Twenty tablets of three different brands were finely
powdered and assayed according to the United States
Pharmacopoeia (USP) and then were kept at 95C for 24 h
in dry conditions.
RESULTS AND DISSCUSION

Analytical Methods
HPLC
HPLC method validation. The standard solutions for
linearity test were prepared five times at different concentra
tion levels. Peak area ratios of baclofen to internal standard
were calculated and plotted versus respective concentrations
and linear regression analysis performed. The constructed
calibration curve was linear over the concentration range of
7.8 1,000 g/mL. Correlation coefficient was found to be
more than 0.996 with relative standard deviation (RSD)
values ranging from 0.16 3.03% within the concentration
ranges studied. Repeatability of measurements of peak area
was carried out using seven replicates of the same concentra
tion (62.5 g/mL). The RSD was found to be 1.2%.
The intra and inter day precision of the method was
carried out at four different concentrations (62.5, 125, 250,
and 500 g/mL). The low RSD values of within day and day
to day variations revealed that the proposed method is
precise (Table II).
Limit of detection (LOD) and limit of quantification
(LOQ) were determined based on signal to noise ratios using
an analytical response of three and ten times the background
noise, respectively (30). The LOD and LOQ were found to be
0.9 and 3 g/mL, respectively. Selectivity of the method was
tested using heated samples of baclofen with or without
Table III. Some Standard Chromatographic Parameters

Peak name
Solvent front
Internal standard
a b
c b
d b
e b
f b
c d
d a
a e
a
Retention time
0.10
4.10
2.30
1.60
2.00
2.70
11.20
Compared to internal standard

Peak c compared to peak d
c
Peak d compared to peak a
d
Peak a compared to peak e
b
Selectivity factora
Resolutionb
2.38a
5.17a
3.10a
1.82a
3.29a
1.67b
1.30c
1.31d
3.64a
2.70a
3.23a
5.86a
6.17a
1.07b
0.55c
0.67d
lactose. Chromatograms are presented in Fig. 1. Some useful

standard chromatographic parameters have been calculated
and reported in Table III.
HPLC analysis of the adduct mixtures. Each adduct
mixture was dissolved in mobile phase to produce a solution
with a nominal baclofen concentration of 500 g/mL. In
comparison to initial solution HPLC analysis of the baclofen
lactose adduct mixture revealed extra peaks (labeled as c, d,
e, and f) in addition to baclofen and the internal standard
(Fig. 1a and b). These extra peaks were named as unknown
1 4. Baclofen elutes at 2.3 min, whereas the internal standard,
elutes at 4.1 min.
Heated baclofen showed two peaks (labeled as e and f)
and named unknown 3 and unknown 4 respectively (Fig. 1c).
Two major unknowns eluted before baclofen indicated as
unknown 1 and unknown 2 in Fig. 1b. These species are more
polar than baclofen but have UV spectra very similar to it,
with maxima near 220 nm. Lactose samples either anhydrous
or hydrous that were heated alone showed no extra peaks in
the same chromatographic conditions compared to that
obtained with the mixture.
Mass Spectroscopy
The same solution injected to HPLC was examined by
LC MS. A mass spectrometer compatible (salt free) gradient
method was developed to give similar separation to the
HPLC method. Mass spectra (MS2) are presented in Fig. 2a
f. The full scan positive ion electrospray product ion mass
spectra showed that the precursor ions of baclofen and
unknown (1 4) were the protonated molecule, [M+H]+, of
m/z 214.7, 538.3, 376.2, 242.4, and 196.2, respectively (Fig. 2a
e). After collision induced dissociation, the characteristic ions
in the product ion mass spectrum were at m/z 151.2, 358.2,
358, 186.3, and 151.1, respectively. Proposed structures for
unknown 1 4 are shown in Figs. 3 and 4, respectively. The
nominal molecular mass of unknown 1 is consistent with the
baclofen lactose condensation product formed by the
elimination of a single molecule of water from the parent
compounds (Figs. 2b and 3a). According to Fig. 2c, unknown
2 indicates another condensation product of 376 amu which is
related to baclofen galactose or baclofen glucose adducts
(Fig. 3b). Galactose and glucose may be produced from lactose
hydrolysis in aqueous solutions and high temperatures.
According to BP (33) there are two major impurities in
baclofen powder, one of them is an oxopentanoic acid deriva
tive: (3R,S) 5 amino 3 (4 chlorophenyl) 5 oxopentanoic acid
(CAS: 1141 23 7) and the other one contains a lactam ring and
is named (4R,S) 4 (4 chlorophenyl) pyrrolidin 2 one (CAS:
22518 27 0; Fig. 4). The nominal mass values of these
impurities are 195.6 and 241.7, respectively. It should be
noted that unknown 3 and 4 shown in Fig. 1 (eluting after
baclofen) have shown masses of 242.4 and 196.2 amu,
respectively (Fig. 2d and e). This result corresponds to
protonated molecule, [M+H]+ of baclofen major impurities.
The MRM mode was set for baclofen (214.7/151.2),
unknown 1 (538.3/358.2) and unknown 2 (376.2/358). MRM
chromatogram shown in Fig. 2f indicates the presence of 538
and 376 masses before baclofen in LC MS/MS system as well.
Manually collected HPLC fractions of unknown 1, unknown
654
Fig. 2. Positive ion mode electrospray mass spectrum of a baclofen, b unknown 1, c unknown 2, d unknown 3, e unknown 4, f MRM
chromatogram of three pairs; (A) baclofen, 214.1/151.2 amu, (B) unknown 1, 538.3/358.3 amu, and (C) unknown 2, 376.2/358.2 amu
2 were correlated by LC MS/MS analysis. The unknown 1

and unknown 2 from Fig. 1 also correlated to the LC MS/
MS chromatograms by similar relative peak areas.
section under Analysis of Prepared Granules and Tablets and

Screening tests and Commercial Tablets subsections.
FTIR Spectroscopy
Spectrophotometry
Calculated absorbances of the samples have been presented
in Table IV and are discussed in the Formulation Methods
As mentioned before, the first step of the Maillard

reaction leads to the formation of an imine known as a
Schiffs base. In Fig. 5, proposed changes in infra red
655
Fig. 3. Proposed Maillard reaction products for baclofen and lactose/glucose/galactose
absorption due to Maillard reaction is presented (34,35). The

C=N stretching band appears at 1,630 1,650 cm 1 in the
infrared spectra of imine containing compounds (35,36).
Wnorowski et al. monitored carbonyl amine reaction
between pyruvic acid and amino alcohols by FTIR
spectroscopy and reported that the resultant Schiffs base
absorbs infrared light at about 1,647 cm 1 (35). Later at 2006,
Namli et al. followed the reaction process between benzalde
hyde or salicylaldehyde and or 2 pyridinecarboxaldehyde with
aniline and showed that resulting imine, shows a band in FTIR
spectra at about 1,630 cm 1 (36). Other scientists have reported
imine formation using FTIR spectroscopy with similar
absorption bands (37,38). In the baclofen FTIR spectrum, the
absorption band at about 1618 cm 1 is consistent with baclofen
N H bending vibration (Fig. 6b). Adduct mixtures showed
another absorption band around 1,648 cm 1 which is related to
C=N stretching (Fig. 6a). The possibility that this imine is
converted into its isomeric enamine form during the Maillard
reaction has been proposed by Holtermand and has been
referred to as a transamination reaction (35,39). The absorption
pattern of baclofen lactose 1:5 w/w after mixing and incubation
at room temperature or 60C for 6 months is shown in Fig. 6c, d,
and e, respectively. Clearly the N=H bending vibration and C=N
stretching absorption have appeared at 1,616 and 1,630 cm 1,
respectively. In the Fig. 6e an extra band at 1,666 cm 1 could be
ascribed to the transamination process described in Fig. 5.
solid degradation products were calculated based on the

assumption that the HPLC detectors response factors for
baclofen and the degradation products are the same. This
kind of assumption has been previously made by Abdoh et al.
in the assessment of amlodipine besylate and excipients
interaction in solid dosage form (7).
Visual changes in heated tablets are presented in Fig. 7.
The granules and the tablets were analyzed to determine the
amount of remaining baclofen and unknown 1. Intensity of
brown color was also measured. Results are presented in
Table IV.
A closer look at Table IV reveals that in granules after
6 months incubation at different temperature conditions, the
amount of remaining baclofen in the preparations containing
anhydrous lactose is always more than the hydrated form
which is in contrast with tablets. Although the first stage of
Maillard reaction involves the removal of water molecule, but
according to Serajuddin (32) and Abdoh (7) the presence of
Formulation Methods
Analysis of Prepared Granules and Tablets
Because the degradation products are not available as
reference materials or not known, their percentages in the
Fig. 4. Baclofen major impurities according to BP
67 06
0 39
0 016
78 23
0 24
0 016
80 1
1 64
0 015
74 01
0 28
0 014
82 84
0 37
0 014
83 9
1 42
0 013
53 68
0 23
0 013
75 88
0 36
0 015
87 49
2 05
0 014
45 22
0 41
0 008
72 15
0 38
0 007
84 40
1 79
0 009
89 72
0 35
0 002
90 22
0 50
0 003
90 26
2 20
0 007
92 23
0 25
0 000
95 87
0 48
0 005
92 55
2 19
0 002
51 00
0 24
0 002
94 76
0 55
0 003
96 43
1 83
0 003
50 11
0 38
0 001
90 06
0 50
0 007
92 05
2 11
0 006
90 10
0 33
0 002
90 23
0 52
0 009
92 76
1 67
0 005
91 99
0 22
0 000
92 37
0 40
0 007
94 00
2 33
0 005
79 08
0 27
0 003
98 68
0 55
0 004
97 98
2 29
0 003
59 71
0 28
0 001
96 35
0 54
0 003
94 24
2 21
0 004
89 38
0 32
0 001
97 02
0 62
0 002
95 89
2 26
0 002
10d
5c
M lactose monohydrate, A lactose anhydrated

a
Baclofen lactose 1 1 w/w
b
RSD for all the experiments is between 0 33 1%
c
d
93 98
0 25
0 014
98 38
0 67
0 002
99 65
2 39
0 003
A
M
A
M
A
M
A
M
A
80 63
0 26
0 002
98 18
0 53
0 003
97 36
2 09
0 002
1
A
M
A
M
A
M
Tablet
Granule
Tablet
Granule
Tablet
M
60 19
0 19
0 001
97 96
0 38
0 002
96 98
2 12
0 001
Remaining Baclofen (%)
Unknown-1(%)
OD
Unknown-1(%)
OD
Unknown-1(%)
OD
b
a
Tablet
Granule
60C
50C
Granule
Red color appearance in the lactose identification test,

confirmed the presence of lactose in the brand formulations
studied. The relatively high temperature of 95C was chosen
for expediency; the results are for comparison within the
experiment, not for prediction of overall rates of decompo
sition of drug mixtures under normal storage conditions, as
have been previously utilized by Wirth et al. in the study of
fluoxetine incompatibility with lactose (26). Results of the
screening test are presented in Table I. As it was expected
(7,32), the presence of water promoted the degradation
reactions including the Maillard reaction as the brown color
develops more significantly in water containing mixtures
(samples 2, 4, 6, and 8).
In fast screening samples, the results with anhydrous and
monohydrate lactose were similar. These results could be
explained by the fact that high temperature forces the
reaction to the end point but the long term studies which
were conducted at lower temperatures are at an intermediate
stage of reaction progress. Therefore, it can be assumed that
the rate of reaction with anhydrous and monohydrated
lactose may be different, but at the end of the reaction, the
amount of baclofen degradation is almost the same.
Wirth et al. showed that the addition of magnesium
stearate catalyzed the formylation process between fluoxetine
and lactose. They concluded that this may be due to localized
changes in pH rather than changes in physical mobility within
the solids due to the lubricant (26). In another investigation,
Abdoh et al. demonstrated that catalytic effect of magnesium
in the Maillard reaction of amlodipine besylate and lactose is
attributed to local pH changes (7). According to Table I,
except sample 4, all other magnesium stearate containing
samples (3, 7, and 8), showed less baclofen loss in comparison
to those prepared without magnesium stearate (1, 5, and 6).
This may be due to a lubricant effect of magnesium stearate
rather than its basic nature which prevents the effective
contact between drug and excipient particles. However, more
study is needed to establish the difference seen here
compared to previous reports (7,26).
The amount of unknown 1 (ESMRP) simply indicates
the occurrence of a reaction and cannot be considered as an
indicator of reaction promotion because the intermediate
40C
Screening Tests and Commercial Tablets
25C
some water promotes the Maillard reaction. This is in

accordance with our observation, that the baclofen remaining
in heated granules of monohydrated lactose is less than in
granules prepared with anhydrous lactose. But in tablets, the
trend is reversed. In granules, the physical contact between
the chemical reactants is more limited than in tablets which
have been compressed. Recent findings of Buisignies et al.
indicates that lactose pseudopolymorphism seems to affect
the compressibility more than anomerization or partial
amorphization (40,41). It can be concluded that tablets
prepared with anhydrous lactose will be harder than those
prepared with monohydrate lactose. Stronger compaction of
anhydrous lactose leads to better particle interlocking and
enabling effective physical contact between the drug and
excipient. Thus, the most important factor affecting the
Maillard reaction in prepared tablets and granules are
physical contact and moisture content, respectively.
Table IV. Percentage of Unknown-1 and Remaining Baclofen and the Absorbance at 490 nm (optical density) in Tablets and Granules after Incubation at Different Temperatures for 6 months
656
Fig. 5. Maillard reaction with FTIR peak changes. Adopted from Yates et al. with minor modifications (35)
Fig. 6. FTIR spectra of (A) adduct mixture, (B) baclofen, (C) baclofen:lactose
monohydrate (1:5 w/w) after mixing, (D) sample C after 6 months at 25C, (E) Sample
C after 6 months at 60C
Fig. 7. Visual changes in prepared tablets after heating at 30, 40, 50, and 60C for 6 months (left to right)
657
658
component is consumed during the progression of the
Maillard reaction with changes to other intermediate. These
intermediates are all then changed to brown colored mela
noidines (8). Brown color development is almost similar in
the presence and absence of magnesium stearate containing
samples. It can be concluded that addition of magnesium
stearate has not influenced the Maillard reaction progress
significantly.
The HPLC chromatograms for intact brands were similar
and the chromatogram for brand 3 is shown in Fig. 1d. The
peak representing unknown 1 appeared at about 1 min.
According to Table I, in the absence of lactose (sample
9), baclofen was degraded up to 10% while in tablets in the
dry state, the loss of drug was more than 20% (sample 13 and
15) except brand 1 (sample11). In wet conditions, the loss of
drug content in brands is higher than the heated baclofen
alone in wet conditions. In both the Maillard and carameliza
tion reactions, highly UV absorbing colorless compounds are
formed at intermediate stages, whereas in the Maillard
reaction, brown polymers are formed at the final stages
(4,13,31). OD results showed a visible dark brown color
development in wet conditions. Carbohydrates may undergo
a caramelization reaction at very high temperatures. The
brown color relates to brown non nitrogenous water soluble
polymers named caramel colors (42,43). Results from samples
17 20 ruled out this type of reaction as no browning occurred.
CONCLUSION
According to results, the introduced HPLC method
which uses a new internal standard is selective, linear,
repeatable, accurate, and has intermediate precision. Thus,
it can be used as a stability indicating method.
Early stage Maillard reaction product (ESMRP) which
were characterized with tandem mass spectrometry showed
that baclofen and lactose (either monohydrate or anhydrous)
undergo the Maillard reaction. Condensation products of
metoclopramide, amlodipine, and hydrochlorothiazide have
been detected using mass spectrometry in a similar way by
other scientists (7,18,25). It can be concluded that the most
important factor affecting the Maillard reaction in prepared
tablets and granules are physical contact and moisture
content, respectively. FTIR analysis confirmed the formation
of imine in the heated mixtures. Similar imine bonds have
been reported in Maillard type reactions but it has never
been used in pharmaceutical evaluation of the Maillard
reaction (37,38). Developing the brown color development
is a key factor of the Maillard reaction occurrence and is also
in accordance with the modern techniques used in this study.
HPLC analysis showed the ESMRP or unknown 1 peak
in binary/tertiary mixtures of drug and lactose in the presence
of magnesium stearate and also in solid state formulations
(granules and tablets) along with the brands tested. These
findings indicate an incompatibility which takes place as a result
of the Maillard reaction between a reducing sugar such as
lactose and an amine containing compound such as baclofen.
There are some reports of mutagenicity and carcinoge
nicity of Maillard reaction products on human cells in the
nutritional literature (44), but the safety of the compounds
that are generated during the Maillard reaction of baclofen
and lactose remain to be investigated. It can be assumed that
this type of reaction reduces drug potency and may cause

some adverse effects relevant to drug safety. Therefore, it is
advisable that lactose should be avoided in the formulation of
baclofen.
ACKNOWLEDGMENTS
The authors are grateful to Tabriz University of Medical
Sciences and University of Queensland for providing financial
assistance. We wish to thank Dr. Nokhodchi for his good
explanation of the tableting process. We are grateful to
Zahravi Co. Tabriz, Iran for their interest and supply of
baclofen and lactose.
REFERENCES
1. Maillard L. Action Des acides amins sur les sucres. Formation
des melanoidins par voie methodique. Compt Rend. 1912;154.
2. Ledl F, Schleicher E. New aspects of the maillard reaction in foods
and in the human body. Angew Chem Int 1990;29(6):565 94.
3. Yaylayan VA, Huyghues Despointes A. Chemistry of Amadori
rearrangement products: analysis, synthesis, kinetics, reactions,
and spectroscopic properties. Crit Rev Food Sci Nutr 1994;34
(4):321 69.
4. Mauron J. The Maillard reaction in food; a critical review from
the nutritional standpoint. Prog Food Nutr Sci 1981;5(1 6):5 35.
5. Paulsen H, Pflughaupt KW. Glycosylamines. In: Horton D,
Pigman W, editors. The carbobydrates, chemistry and biochem
istry. New York: Academic; 1980.
6. Hodge JE. The amadori rearrangement. Adv Carbobydr Chem
1955;10:169 205.
7. Abdoh A, Al Omari MM, Badwan AA, Jaber AM. Amlodipine
besylate excipients interaction in solid dosage form. Pharm Dev
Technol 2004;9(1):15 24.
8. Martins SIFS, Van Boekel MAJS. Melanoidins extinction
coefficient in the glucose/glycine Maillard reaction. Food Chem
2003;83:135 42.
9. Fayle SE, Gerrard JA. The Maillard reaction. RSC food analysis
monographs. Cambridge: Royal Society of Chemistry; 2002.
10. Van Boekel MAJS, Berg HE. Kinetics of the early Maillard reaction
during heating of milk. In: Labuza TP, Reineccius GA, Monnier V,
OBrien J, Baynes J, editors. Maillard reactions in chemistry, food,
and health. Royal Soc Chem, London 1994. p. 170 175.
11. Schneider M, Klotzsche M, Werzinger C, Hegele J, Waibel R,
Pischetsrieder M. Reaction of folic acid with reducing sugars and
sugar degradation products. J Agric Food Chem 2002;50(6):1647 51.
12. Ellis GP. The Maillard reaction. Adv Carbohydr Chem
1959;14:63 134.
13. Hodge JE. Dehydrated foods, chemistry of browning reactions in
model systems. J Agric Food Chem 1953;1:928 43.
14. Rychlik M, Mayr A. Quantitation of N2 [1 (1 carboxy)ethyl]folic
acid, a nonenzymatic glycation product of folic acid, in fortified
foods and model cookies by a stable isotope dilution assay. J
Agric Food Chem 2005;53(13):5116 24.
15. Dai Z, Nemet I, Shen W, Monnier VM. Isolation, purification
and characterization of histidino threosidine, a novel Maillard
reaction protein crosslink from threose, lysine and histidine.
Arch Biochem Biophys 2007;463(1):78 88.
16. Davidek T, Kraehenbuehl K, Devaud S, Robert F, Blank I.
Analysis of Amadori compounds by high performance cation
exchange chromatography coupled to tandem mass spectrome
try. Anal Chem 2005;77(1):140 7.
17. Roscic M, Versluis C, Kleinnijenhuis AJ, Horvat S, Heck AJ.
The early glycation products of the Maillard reaction: mass
spectrometric characterization of novel imidazolidinones derived
from an opioid pentapeptide and glucose. Rapid Commun Mass
Spectrom 2001;15(12):1022 9.
18. Harmon PA, Yin W, Bowen WE, Tyrrell RJ, Reed RA. Liquid
chromatography mass spectrometry and proton nuclear magnetic
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
resonance characterization of trace level condensation products

formed between lactose and the amine containing diuretic
hydrochlorothiazide. J Pharm Sci 2000;89(7):920 9.
Fay LB, Blank I, Cerny C. Flavour science: recent developments.
In: Taylor AJ, Mottram DS, editors. Cambridge: Royal Society
of Chemistry, Information Services; 1996.
Rowe RC, Sheskey PJ, Weller PJ. Handbook of pharmaceutical
excipients. 4th ed. London, Washington, DC.: Pharmaceutical
Press; American Pharmaceutical Association; 2003.
Physicians desk reference. Montvale, N.J.: Medical Economics
Co.; 2004.
Qiu Z, Stowell JG, Morris KR, Byrn SR, Pinal R. Kinetic study
of the Maillard reaction between metoclopramide hydrochloride
and lactose. Int J Pharm 2005;303(1 2):20 30.
Byrn SR, Xu W, Newman AW. Chemical reactivity in solid state
pharmaceuticals: formulation implications. Adv Drug Deliv Rev
2001;48(1):115 36.
Kumar V, Banker GS. Maillard rection and drug stability. In:
Labuza TP, Reineccius GA, Monnier V, OBrien J, Baynes J,
editors. Maillard reactions in chemistry, food, and health. Cam
bridge: Royal Society of Chemistry; 1994.
Qiu Z, Stowell JG, Cao W, Morris KR, Byrn SR, Carvajal MT.
Effect of milling and compression on the solid state Maillard
reaction. J Pharm Sci 2005;94(11):2568 80.
Wirth DD, Baertschi SW, Johnson RA, Maple SR, Miller MS,
Hallenbeck DK, et al. Maillard reaction of lactose and fluoxetine
hydrochloride, a secondary amine. J Pharm Sci 1998;87(1):31 9.
Martindale: The Complete Drug Reference. 35 ed., Pharmaceu
tical Press. 2006.
Cutrignelli A, Denora N, Lopedota A, Trapani A, Laquintana V,
Latrofa A, et al. Comparative effects of some hydrophilic
excipients on the rate of gabapentin and baclofen lactamization
in lyophilized formulations. Int J Pharm 2007;332(1 2):98 106.
Miller JC. Statistics for analytical chemistry. 22. Chichester:
Horwood; 1984.
United States Pharmacopeia, USP 30 Nf 25. United States
Pharmacopeial Convention: Rockville, Md. 2006.
Ajandouz EH, Tchiakpe LS, Dalle Ore F, Benajiba A, Puig
server A. Effects of pH on caramelization and maillard reaction
kinetics in fructose lysine model systems. J Food Sci 2001;66
(7):926 31.
659
32. Serajuddin AT, Thakur AB, Ghoshal RN, Fakes MG, Ranadive
SA, Morris KR, et al. Selection of solid dosage form composition
through drug excipient compatibility testing. J Pharm Sci 1999;88
(7):696 704.
33. British pharmacopoeia. British Pharmacopoeia Commission:
London; 2006.
34. Yates EA, Jones MO, Clarke CE, Powell AK, Johnson SR,
Porch A, et al. Microwave enhanced reaction of carbohydrates
with aminoderivatised labels and glass surfaces. J Mater Chem
2003;13:2061 63.
35. Wnorowski A, Yaylayan VA. Monitoring carbonyl amine reac
tion between pyruvic acid and alpha amino alcohols by FTIR
spectroscopy a possible route to Amadori products. J Agric
Food Chem 2003;51(22):6537 43.
36. Namli H, Turhan O. Background defining during the imine
formation reaction in FT IR liquid cell. Spectrochim Acta A Mol
Biomol Spectrosc 2006;64(1):93 100.
37. Saini A, Carlin CM, Patterson HH. Confirmation of the presence
of imine bonds in thermally cured polyimides. J Polymer Sci
Polymer Chem 1993;31:2751 8.
38. Nazr H, Arc C, Emregul KC, Atakol O. A crystallographic
and spectroscopic study on the imine amine tautomerism of
2 hydroxyaldimine compounds. Z Kristallogr 2006;221:699
704.
39. Holtermand A. The browning reaction. Die Strke 1966;18:319 28.
40. Busignies V, Tchoreloff P, Leclerc B, Besnard M, Couarraze G.
Compaction of crystallographic forms of pharmaceutical granular
lactoses. I. Compressibility. Eur J Pharm Biopharm 2004;58
(3):569 76.
41. Busignies V, Tchoreloff P, Leclerc B, Hersen C, Keller G,
Couarraze G. Compaction of crystallographic forms of pharma
ceutical granular lactoses. II. Compacts mechanical properties.
Eur J Pharm Biopharm 2004;58(3):577 86.
42. Kamuf W, Nixon A, Parker O, Barnum GC. Overview of
caramel colors. Cereal Foods World 2003;48(2):64 9.
43. Yaylayan VA, Kaminsky E. Isolation and structural analysis of
Maillard polymers: caramel and melanoidin formation in glycine/
glucose model system. Food Chem 1998;63(1):25 31.
44. Cotterill JV, Chaudhry MQ, Matthews W, Watkins RW. In silico
assessment of toxicity of heat generated food contaminants.
Food Chem Toxicol 2008;46(6):1905 18.

DOI: 10.1208/s12249 009 9249 7
Research Article
Enhancement of Oral Bioavailability of Cilostazol by Forming its Inclusion
Complexes
Samir G. Patel1,3 and Sadhana J. Rajput2
Abstract. The study was designed to investigate the effect of cyclodextrins (CDs) on the solubility,
dissolution rate, and bioavailability of cilostazol by forming inclusion complexes. Natural CDs like CD,
CD, and the hydrophilic CD derivatives, DM CD and HP CD, were used to prepare inclusion
complexes with cilostazol. Phase solubility study was carried out and the stability constants were
calculated assuming a 1:1 stoichiometry. Solid cilostazol complexes were prepared by coprecipitation and
kneading methods and compared with physical mixtures of cilostazol and cyclodextrins. Prepared
inclusion complexes were characterized by Fourier transform infrared spectroscopy, differential scanning
calorimetry (DSC), and X ray diffraction (XRD) studies. In vitro dissolution study was performed using
phosphate buffer pH 6.4, distilled water, and HCl buffer pH 1.2 as dissolution medium. The optimized
inclusion complex was studied for its bioavailability in rabbit and the results were compared with those of
pure cilostazol and Pletoz 50. Phase solubility study showed dramatic improvement in the solubility of
drug by formation of complexes, which was further increased by pH adjustment. The dissolution rate
of cilostazol was markedly augmented by the complexation with DM CD. DSC and XRD curves
showed sharp endothermic peaks indicating the reduction in the microcrystallinity of cilostazol. Selected
inclusion complex was also stable at ambient temperature up to 6 months. The in vivo study revealed that
DM CD increased the bioavailability of cilostazol with low variability in the absorption. Among all
cilostazol cyclodextrins complexes, cilostazol DM CD inclusion complex (1:3) prepared by coprecipi
tation method showed 1.53 fold and 4.11 fold increase in absorption along with 2.1 fold and 2.97 fold
increase in dissolution rate in comparison with Pletoz 50 and pure cilostazol, respectively.
KEY WORDS: bioavailability; cilostazol CD inclusion complex; dissolution; solubility; stability study.
INTRODUCTION
Cilostazol,{6 [4 (1 cyclohexyl 1H tetrazol 5 yl)butoxy] 3,4
dihydro 2(1H) quinolinone}(Fig. 1) (1), is a cyclic adenosine
monophosphate (cAMP) phosphodiesterase III inhibitor,
inhibiting phosphodiesterase activity and suppressing cAMP
degradation with a resultant increase in cAMP in platelets and
blood vessels, leading to inhibition of platelet aggregation and
vasodilation. Cilostazol is slightly soluble in methanol, ethanol,
and practically insoluble in water, 0.1 N HCl and 0.1 N NaOH.
The reported Log P value is 3.048 (http://www.chemspider.
com/Search.aspx; accessed August 15 2008). The pharmacoki
netic parameters of cilostazol following oral administration are
generally highly variable. Peak plasma concentration has been
shown to be 1.2 g/mL after a single oral dose of 100 mg,
generally obtained 3 to 4 h after oral administration (1).
Cilostazol absorption in the gastrointestinal tract is slow,
variable, and incomplete. A high fat meal increased absorp
1
Pharmaceutics Department, Ramanbhai Patel College of Pharmacy,

Education Campus Changa, Changa. Dist.: Aanand, Gujarat, India.
2
Pharmaceutical Quality Assurance Laboratory, Pharmacy Depart
ment, The M. S. University of Baroda, Vadodara, Gujarat, India.
3
To whom correspondence should be addressed. (e mail: samirpatel
1602@gmail.com)
tion, with approximately 90% increase in Cmax and a 25%

increase in area under curve (AUC) (1). The absolute
bioavailability of cilostazol is not known and relative bio
availability is unpredictable. After oral administration, ap
proximately 56% cilostazol and its metabolites (19%) are
eliminated through urine (74%) and the remaining is
excreted in feces (20%). The apparent elimination half life
of cilostazol and its active metabolite is 11 13 h in adults with
normal renal functions (1). The aim of the study was to
develop the inclusion complexes of cilostazol to enhance the
solubility and oral bioavailability.
Cyclodextrins (CDs) are cyclic oligosaccharide consisting
of at least six (1 4) linked glucopyranose units. , , and
CD consist of six, seven, and eight glucopyranose units,
respectively (2). These are often depicted as hollow truncated
cones with exterior hydrophilic surface and an electron rich
hydrophobic interior surface. Exterior hydrophilic surface is
favorable for enhancement of absorption rate through the
gastrointestinal tract and the hydrophobic cavity generally
provides a favorable environment for hydrophobic molecules
or parts of a molecule thus improving the solubility of
hydrophobic compounds in aqueous solutions (3,4). The
solubilization abilities of CDs have been attributed to
the formation of inclusion complexes between CDs and the
guest molecules. Generally, this complexation involves the
660
Enhancement of Oral Bioavailability of Cilostazol
Fig. 1. Chemical structure of cilostazol
inclusion of the guest molecule in the cavity of the host

molecule, such as CD, with no covalent bonding (5).
However, practical use of natural CDs as cilostazol carriers
is restricted by their lower aqueous solubility. Several
hydrophilic CD derivatives have been used to solve the
problem (i.e., the methylated, hydropropylated, and sulfobu
tyl ether CD derivatives, etc.) (6 8). All these derivatized
CDs offer better solubility, higher aqueous stability, and
increased bioavailability and have less undesirable side effects
compared to natural CDs (9 12).
661
The phase solubility of cilostazol was conducted accord
ing to Higuchi and Connors (14). An excess amount of
cilostazol (50 mg) was added to 5 mL of water or aqueous
solutions of CDs and its derivatives (10 50 mol/L) individu
ally in 10 mL stoppered glass tubes. The tubes were shaken
for 24 h at 50 cycles per minute in a water bath at 370.5C.
At equilibrium after 2 days, aliquots were withdrawn, ltered
(0.45 m cellulose nitrate lters), and suitably diluted.
Concentration of cilostazol was determined spectrophotomet
rically at 257.0 nm. The phase solubility study was further
carried out in HCl buffer pH 1.2 (13) and phosphate buffer
pH 6.8 (13).
A plot of total molar concentration of the cilostazol
against the total molar concentration of CDs gave phase
solubility diagrams from where the apparent solubility
constant, KC, was calculated for all the pH values using their
regression lines to the following equation.
Stability constant KC
Slope
S 1 Slope
Where So is the intrinsic solubility of the cilostazol

studied under the conditions (15 17).
Preparation of Inclusion Complexes
Materials
The CDs used for the preparation of inclusion complexes

were CD, CD, HP CD, and DM CD. Cilostazol
CDs inclusion complexes were prepared in 1:1, 1:2, and 1:3
molar ratios by using two different methods: (1) kneading
method and (2) coprecipitation method and compared with
physical mixtures of cilostazol CDs (18).
Cilostazol was generously provided by Cadila Pharma

ceutical Laboratory, Ahmedabad, India, as a gift sample.
CD was purchased from Hi media Laboratories Pvt. Ltd.
Mumbai. Hydroxypropyl cyclodextrin (HP CD) was
obtained as a gift sample from Sun Pharma Advance
Research Company, Vadodara, India. CD and dimethyl
cyclodextrin (DM CD) were procured as gift samples from
Roquette Pharma, USA. High performance liquid chroma
tography (HPLC) grade methanol and acetonitrile were
purchased from SD Fine chemicals (Maharashtra, India).
All other chemicals used were of analytical grade and used as
received without further purication. Triple distilled water
was used throughout the study. The HCl buffer pH 1.2 and
phosphate buffer pH 6.8 were prepared as per IP 1996
procedure (13).
Phase Solubility Study
The analytical method based on UV spectrophotometry
was developed before starting the phase solubility study. The
calibration curve was prepared by measuring the absorbance
of standard methanolic solutions of cilostazol in the concen
tration range 5 50 g/mL at 257.00 nm. The method was
validated as per the International Conference on Harmoni
zation (ICH) guidelines. The coefcient of determination
value (R2) for the calibration curve was 0.9998. The intraday
and interday (3 days, n=3) accuracy was 99.58 100.32% and
99.80 100.4%, respectively. The intraday and interday
(3 days, n=3) precision was expressed as relative standard
deviation and were in range of 1.39 1.98% and 1.37 1.97%,
respectively. The limit of detection (LOD) and limit of
quantication (LOQ) of the developed method were 0.231
and 0.770 g/mL, respectively.
(a) Physical mixture

The physical mixtures were prepared by mixing pulver
ized powder of cilostazol with CD, CD, HP CD, and
DM CD in different cilostazol to CD ratios like 1:1, 1:2, and
1:3. The mixture was then passed through sieve mesh # 100
and stored in a dessicator until further evaluation.
(b) Kneading method
A paste containing one part of cilostazol and three parts
of any one CD like, CD, CD, HP CD and DM CD,
was introduced in to a kneading machine with progressive
addition of the cilostazol. The paste was mixed for 1 h. The
viscosity of the mixture was increased indicating the forma
tion of the complex. Auxiliary step involved washing the
paste with solvent and water followed by ltration. Finally the
paste was dried in an oven at 45C until dry. The paste was
ground to get a ne powder and passed through sieve mesh
#100. All the prepared inclusion complexes were stored in
dessicator until further evaluation.
(c) Coprecipitation method
In this method, an organic solution of the cilostazol was
poured under agitation into an aqueous solution of CD with
contentious stirring. A solid inclusion complex was obtained
either spontaneously or after evaporation of excess solvent.
After the precipitation step, the inclusion complex was
thoroughly washed with solvent and water, ltered, and dried
to get a pure inclusion complex. Finally, dried complex was
662
Patel and Rajput
passed through sieve mesh # 100 and stored in dessicator until

further evaluation.
Inclusion Efficiency Study
All inclusion complexes of cilostazol and their physical
mixtures (25 mg) were separately taken in 25 mL volumetric
asks. Ten milliliters of methanol were added to it, mixed
thoroughly, and sonicated for 30 min at ambient temperature.
The volume was made up to mark with methanol. An aliquot
from the each solution was suitably diluted with methanol to
get the nal concentration of 10 g/mL of cilostazol and
spectrophotometrically assayed for cilostazol content at
257.00 nm. Inclusion efciency was calculated using the
formula:
Inclusion efficiency estimated % cilostazol content=
theoretical % cilostazol content 100
Physicochemical Characterization
1. IR spectroscopy study
The infrared (IR) spectra were obtained using a Shi
madzu Fourier transform infrared (FTIR) 8400S spectropho
tometer with IR solution software (Shimadzu). The samples
were prepared by grinding a small amount of the dried
sample and the corresponding amount of potassium bromide
(2/98 w/w) in an agate mortar. Data were collected over a
spectral region from 4,000 to 650 cm 1 with resolution of
4 cm 1 and 100 scans.
scanning rate employed was 1 min 1 over the 0 100

diffraction angle (2) range. The divergence and receiving
slits were chosen in order to ensure a high resolution mode
for the crystalline phases (20).
Dissolution rate studies were performed in a solution
containing HCl buffer pH 1.2, distilled water, or phosphate
buffer pH 6.4, 900 mL, at 370.2C, using US Pharmacopeia
(USP) XXIII (18,21,22) apparatus (Electrolab, India Pro
grammable tablet dissolution test apparatus USP XXI/XXII,
TDT 06P) with a basket rotating (Apparatus I) at 50 rpm.
The best results in terms of higher percent drug release,
dissolution pattern, and dissolution efciency were obtained
in phosphate buffer pH 6.4. Hence, phosphate buffer pH 6.4
was used as dissolution medium for further studies. Physical
mixtures and inclusion complexes, each containing 50 mg of
cilostazol, were lled into empty hard gelatin capsule shell
and subjected to dissolution study. Samples (5 mL) were
withdrawn at 10, 20, 30, 40, 50, 60, 75, 90, 105, and 120 min,
ltered through Whatman lter paper no. 41, and assayed
spectrophotometrically for cilostazol content at 257.8 nm. The
experiments were carried out in triplicate and the maximum
mean cumulative percent drug released along with standard
deviation (SD) of pure cilostazol, marketed formulation
(tablet, Pletoz 50), physical mixtures, and inclusion complexes
were identied. The percent dissolution efciency (DE120 min.)
values based on the dissolution data along with SD were
calculated as per Khan and Rhodes (23) method and time
taken for 50% cilostazol dissolved (T50% value) was identied
from the dissolution proles. The release kinetics of diffusion
was also studied by calculating the regression coefcient for
zero order, rst order, and Higuchis equations.
2. Differential scanning calorimetry

Differential scanning calorimetry (DSC) analysis was
performed for pure cilostazol, pure CDs and its derivatives,
physical mixtures of cilostazol CDs, and inclusion complexes
using a Shimadzu DSC (Model DT 60) instrument. Samples
(0.5 1.0 mg) were weighed on Shimadzu Libror AEG 220
electronic balance and transferred to the aluminum boat,
crimped and hermetically sealed by using crimping machine.
An empty aluminum boat was used as reference. Samples
were heated in sealed aluminum boat at a rate of 10C/min in
a 0 400C temperature range under nitrogen stream. The
analysis was repeated after cooling down to temperature in
order to estimate the DSC baseline. The instrument was
calibrated using indium (melting point, 156.61C; enthalpy of
fusion, 28.71 J/g) (18,19).
3. X ray powder diffraction study
X ray diffraction (XRD) pattern of pure cilostazol, pure
CDs and its derivatives, physical mixtures of cilostazol CDs,
and inclusion complexes was collected with a PW1701
diffractometer at 30 mA and 40 kV using CuK radiation
with a graphite monochromator on the diffracted beam. The
powdered samples (a fraction of 100 180 m sieved powder)
were deposited on an adhesive support and placed in the
diffractometer. XRD patterns were recorded in step scan
mode in the range 3260 with step size 0.06. The
Stability Study
Approximately 5 g of the formulation was lled in USP
type III glass vial and sealed using VP6 crimp on spray pump
tted with 10 m actuator. The chemical stability was studied
as per ICH guideline (photo stability testing for new drug
substances and products Q1A (R2)).The stress stability was
conducted at 602C in an incubator and the accelerated
stability was studied at 302C/655% relative humidity
(RH) and 402/755% RH as per ICH guidelines. The
duration of stability was 6 months and samples were
withdrawn at predetermined time intervals, after 1, 2, 4, and
6 months. Withdrawn samples were tested for their in vitro
dissolution study and results were compared with in vitro
dissolution proles of the test product before and after
stability study as per the SUPAC IR guidelines which assure
similarity in the product performance and bioequivalence.
The similarity factor f2 was calculated using the reported
equation (24,25) in which initial dissolution data were
considered as reference values.

f2 50LOG
1X
t
n
nRt
1
Tt
2 0:5

100
Where Rt and Tt are the percentage of cilostazol

dissolved for reference and test samples at each time point.

An f2 value between 50 and 100 suggests the dissolution
proles can be considered as similar.
Comparative Bioavailability Study
The pharmacokinetic study was performed using New
Zealand rabbits weighing 1.0 to 1.5 kg. All experiments and
protocol for this study were approved by the Institutional
Animal Ethics Committee, The M. S. University of Baroda,
and are in accordance with the committee for the purpose of
control and supervision of the experiments on Animal,
Ministry of Social Justice and Empowerment, Government
of India. The dose for the rabbit was calculated to be 4.66 mg/
kg of rabbit weight (26). The study was performed in parallel
experiments. The animals were fasted overnight prior to the
experiment but had free access to water. The conventional
tablet(Platoz 50) powder equivalent to 7.0 mg of cilostazol,
cilostazol DM CD inclusion complex in the ratio of 1:3
(equivalent to 7.0 mg of cilostazol), and pure cilostazol
(7.0 mg) were taken and mixed with 1 mL of 0.5% of sodium
carboxymethyl cellulose solution. Prepared suspensions were
administered orally to the rabbits. There were four groups
containing eight rabbits in each group, two rabbits for
placebo, two rabbits for pure cilostazol, two rabbits for
conventional tablet, and two rabbits for cilostazol DM
CD inclusion complex. The blood samples (approximately
400 500 L) from rabbits were collected from the marginal
ear vein using heparinized needle (20 24 size) at 1, 2, 4, 6, 12,
18, and 24 h after oral administration. The heparinized blood
samples were immediately transferred to centrifuge tubes
(5 mL) and centrifuged at 20,000 rpm at 0C for 15 min.
Supernatant layer of plasma was separated into another
centrifuge tube and stored at 20C for further experimental
procedure.
HPLC analysis for Estimation of Cilostazol in Plasma
Drug concentration from plasma was determined by
reverse phase HPLC using isocratic pump connected to
manual rheodyne injector with 20 L xed loop and a
Phenomenex Luna column (C18 ODS 250 mm4.6 id, 5 ,
Torrance, USA) preceded with ODS guard column (10
5 mm id). The mobile phase was acetonitrile: water (60:40)
with the ow rate of 0.4 mL/min. Chromatograms were
recorded by ultraviolet (UV) detection at a xed wavelength
663
of 254.00 nm. Exactly 0.200 mL of thawed plasma was taken
in 5 mL polypropylene centrifuge tube with at caps by
adding 0.4 mL of methyl tertiary butyl ether, and the tube was
vortexed for 30 s at high speed. The sample was shaken on a
rotating shaker (50 rotations per minute) for 3 min. Finally,
the sample was centrifuged for 20 min at 2,000 rpm at 4C
and the clear methyl tert butyl ether layer was transferred
with a calibrated pipette to a disposable glass tube and
evaporated under a gentle stream of nitrogen. The dried
residue was taken up with 25 L of mobile phase, and 20 L
of this mixture was injected in the HPLC system. The
linearity of this method was in the range of 0.1 20 g/mL
having coefcient of determination value (R2 =0.9989). The
average extraction efciency of cilostazol in plasma was
83.42% to 95.13%. The LOD and LOQ of the developed
method were 0.0248 and 0.0845 g/mL, respectively.
Pharmacokinetic Data Analysis
After oral administration of inclusion complex, marketed
formulation, and pure cilostazol, the plasma samples were
analyzed by HPLC for their cilostazol content. A curve of
cumulative drug absorbed vs. time over 0 to 24 h was plotted
to calculate the AUC. The pharmacokinetic data were
calculated using Microsoft Excel work sheet.
The phase solubility proles for the complex formation
between cilostazol and CDs in three aqueous solutions (HCl
buffer pH 1.2, distilled water, and phosphate buffer pH 6.8) at
37C are shown in Fig. 2a c. Initial studies indicated that
cilostazol is chemically stable in water for at least 7 days at
37C. The extremely low solubility of cilostazol (0.101
0.004 g/mL in water at 37C) was linearly increased in a
concentration dependent manner with the increase in CD
concentration. Such linear curves are referred to as AL type
curves in solubility diagrams (27).The regression coefcient
values were (R2) >0.99 in the specied concentration range
(14). These linear cilostazol CDs curves suggested the
formation of 1:1 (mol/mol) cilostazol CD inclusion
complexes at different pH values. The calculated stability
constant values are shown in Table I. These results indicated
Fig. 2. a Phase solubility study of cilostazol in HCl buffer at pH 1.2 at 257.2 nm. Key (empty triangles) DM CD; (empty squares) HP CD;
(filled diamonds) CD; (empty circles) CD. All values are represented as meanSD (n 3). b Phase solubility study of cilostazol in water at
257.00 nm. Key (filled triangles) DM CD; (filled squares) HP CD; (filled diamonds) CD; (empty circles) CD. All values are represented as
meanSD (n 3). c Phase solubility study of cilostazol in phosphate buffer at pH 6.8 at 257.8 nm. Key (empty triangles) DM CD; (filled
squares) HP CD; (filled diamonds) CD; (empty circles) CD. All values are represented as meanSD (n 3)
664
Patel and Rajput
Table I. Apparent Inclusion Complex Stability Constant (Kc) of Cilostazol with Different CDs at 37C and Types of Phase Solubility Curve for
Cilostazol with Various CDs at Different pH Solutions
Cilostazol
Solutions
Solubility (g/mL)
K 1:1 [M 1]SDa
R2
Type of curve
HCl buffer pH 1.2

Water
Phosphate buffer
pH 6.8
HCl buffer pH 1.2
Water
Phosphate buffer
pH 6.8
HCl buffer pH 1.2
Water
Phosphate buffer
pH 6.8
HCl buffer pH 1.2
Water
Phosphate buffer
pH 6.8
114.456
87.195
147.748
105.3677
79.895
138.80411
0.9984
0.9942
0.9945
AL
AL
AL
80.1203
66.262
98.645
68.346
55.094
89.778
0.9976
0.995
0.9925
AL
AL
AL
125.303
108.581
181.328
121.71311
101.8113
177.59115
0.9972
0.9972
0.9957
AL
AL
AL
146.046
217.467
393.524
146.10617
214.399720
390.07135
0.9948
0.999
0.9952
AL
AL
AL
CDs
CD
CD
HP CD
DM CD
All values are represented as meanSD (n 3)

Table II. Inclusion Efciency Data for Cilostazol CD Inclusion Complexes and Physical Mixtures
Inclusion complexes and physical mixtures of cilostazol CDs (w/w)
Cilostazol CD (w/w)
Coprecipitation method
Kneading method
Physical mixture
Cilostazol CD (w/w)
Kneading method
Physical mixture
Cilostazol HP CD (w/w)
Kneading method
Physical mixture
Cilostazol DM CD (w/w)
Kneading method
Physical mixture
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
The results are mean valuesSD derived from three different experimental batches
Inclusion efciency(%)SDa
%RSDa
75.22.12
87.42.23
98.51.50
65.362.31
76.961.87
85.61.53
55.51.50
61.60.8
68.51.2
69.20.92
80.40.83
99.42.06
71.411.34
79.731.66
89.11.87
49.21.53
58.90.85
67.51.72
67.62.32
83.941.66
98.90.81
61.491.82
74.382.03
88.01.92
57.501.23
79.11.35
84.71.92
77.21.12
89.41.33
99.641.86
80.462.19
84.371.93
89.131.36
68.70.50
79.30.45
88.90.2
2.16
2.29
1.23
1.88
1.75
1.39
1.51
1.05
1.3
1.36
1.89
2.23
1.58
1.63
2.12
1.81
1.55
0.83
1.31
2.39
1.69
1.43
2.12
1.88
1.91
1.55
1.23
1.96
1.89
1.73
1.87
1.56
1.67
2.51
1.65
2.39
665
Fig. 3. DSC thermograms of [A] pure cilostazol, [B] pure CD, [C] pure CD, [D] pure HP CD, [E] pure DM CD,
[F] cilostazol CD physical mixture (1:3), [G] cilostazol CD inclusion complex by coprecipitation method (1:3), [H]
cilostazol CD inclusion complex by kneading method (1:3), [I] cilostazol CD physical mixture (1:3), [J] cilostazol CD
inclusion complex by coprecipitation (1:3), [K] cilostazol CD inclusion complex by kneading method (1:3), [L] cilostazol
HP CD physical mixture (1:3), [M] cilostazol HP CD inclusion complex by coprecipitation method (1:3), [N] cilostazol
HP CD inclusion complex by kneading method (1:3), [O] cilostazol DM CD physical mixture (1:3), [P] cilostazol DM
CD inclusion complex by coprecipitation method (1:3), [Q] cilostazol DM CD inclusion complex by kneading method
(1:3)
that cilostazol CD complexes (1:1 molar ratio) were

sufciently stable in phosphate buffer pH 6.8. The solubility
of cilostazol was highest (393.524 g/mL) in phosphate
buffer pH 6.8 with stability constant value of 390.07135 M 1.
Inclusion Efficiency Study
The results of inclusion efciency study are shown in
Table II. The data indicate that the percent inclusion
efciency of 1:3 cilostazol CD inclusion complexes prepared
by coprecipitation method was more than 98.5% 1.50,
whereas other inclusion complexes prepared by kneading
method and physical mixtures have values in the range of
49.2%1.53 to 89.4%1.33 suggesting that cilostazol was
uniformly distributed in all 1:3 inclusion complex whereas the
inclusion complexes and physical mixtures prepared in other
ratios did not show satisfactory drug incorporation.
Physicochemical Characterization
FTIR Spectroscopy
Pure cilostazol was characterized by aromatic and
aliphatic C H stretching peaks at 2,867 to 3,315 cm 1, C=N
stretching of tetrazole at 1,757 cm 1, N H stretching of
quinolinone at 3,315 cm 1, N=N stretching of tetrazole at
1,687 cm 1, aliphatic C=O stretching band at 1,822.61 cm 1,
and aromatic C=C stretching band at 1,506 cm 1. The FTIR
spectra of CD, CD, HP CD, and DM CD showed
intense bands at 3,465 3,247 cm 1 corresponding to
absorption by hydrogen bonded OH groups. The bonds that
appeared at 3,000 2,800 cm 1 were assigned to stretching
vibration of the bonds in CH and CH2 groups. In the
physical mixtures of cilostazol CDs, the spectra were the
superimpositions of those of the pure compounds with
666
Patel and Rajput
Fig. 4. X ray diffraction patterns of [A] pure CD, [B] pure CD, [C] pure HP CD, [D] pure DM CD, [E] pure
cilostazol, [F] cilostazol CD physical mixture (1:3), [G] cilostazol CD inclusion complex (1:3), [H] cilostazol CD
physical mixture (1:3), [I] cilostazol CD inclusion complex (1:3), [J] cilostazol HP CD physical mixture (1:3), [K]
cilostazol HP CD inclusion complex (1:3), [L] cilostazol DM CD physical mixture (1:3), [M] cilostazol DM CD
inclusion complex (1:3)
attenuation of the cilostazol peaks. However, the spectra of

inclusion compounds showed rightward shifts of the band
corresponding to hydrogen bonded OH groups (from
3,352.05 to 3,315 cm 1 for the cilostazol CD complex,
from 3,380 to 3,340 cm 1 for the cilostazol CD complex,
from 3,380.98 to 3,313 cm 1 for the cilostazol HP CD
complex, and from 3,465.84 to 3,461.99 cm 1 for the
cilostazol DM CD complex). These results suggested that
some of the existing bond formed between the OH groups on

the narrow side of CD molecules might be disturbed after the
formation of inclusion complexes.
Evidence for the interaction between cilostazol and CDs
can be identied by DSC study and obtained thermograms
Table III. Maximum Percent Cumulative Drug Release of all Physical Mixtures, Inclusion Complexes, Pletoz 50, and Pure Cilostazol
Maximum percent cumulative drug releaseSDa
Method
Physical mixture
Kneading method
Cilostazol CDs ratios
Cilostazol CD
Cilostazol CD
Cilostazol HP CD
Cilostazol DM CD
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
45.351.59
56.671.93
69.321.75
47.831.47
63.761.65
73.281.93
69.321.75
68.881.56
76.491.09
51.982.62
56.091.55
71.131.24
63.991.04
66.42.33
83.721.24
67.871.82
78.951.64
87.082.02
61.551.45
66.561.59
72.342.14
55.761.85
70.52.67
79.891.24
78.320.830
83.721.24
95.202.31
53.220.13
69.760.25
78.621.78
61.962.85
73.611.33
84.861.23
78.212.56
88.552.14
98.162.24
PLETOZ 50 46.562.1; pure cilostazol 32.980.91

The results are mean valuesSD derived from three different experimental batches
667
the data, it can be concluded that coprecipitation method was
the most suitable method for the formation of cilostazol CD
inclusion complexes.
X ray powder diffraction study
Fig. 5. Cumulative percent drug released from cilostazol DM CD

inclusion complex (1:3) by coprecipitation method for stability study
at different time intervals. The results are mean valuesSD derived
from three different experimental batches
are shown in Fig. 3. Thermogram of pure cilostazol shows a

characteristic endothermic peak at 165.13C corresponding to
its melting point. The DSC thermogram of CD has large
broad endothermic peak between 130.00C and 160.00C
(148.59C) that was related to the loss of hydration water of
the starting material (20,28,29). CD decomposes at about
300.00C, so there was no trace of melting peak of CD in
the chosen temperature range, while any endothermic peak
was not observed in CD, HP CD, and DM CD. The
cilostazol endothermic peak and the CD dehydration peaks
though rudimentary were observed in the thermogram of the
physical mixture of cilostazol with CD suggesting the
partial complexation. This may be due to the right proportion
of CD and cilostazol and the appropriate cavity size of
CD. Cilostazol endothermic peak between 162C and 165C
was observed in all physical mixtures of CD, HP CD, and
DM CD suggesting partial interaction between cilostazol
and CDs in these systems. DSC spectra of inclusion complex
prepared by coprecipitation method showed complete disap
pearance of cilostazol endothermic peak in all 1:3 ratios of
cilostazol CDs, while partial inclusion complex formation was
observed from the DSC thermograms of all cilostazol CDs
inclusion complexes prepared by kneading method. From all
Powder XRD study was used to measure the crystallinity

of the formed inclusion complexes. The peak position (angle
of diffraction) is an identication tool of a crystal structure,
whereas the number of peaks is a measure of sample
crystallinity in a diffractogram. The formation of an amor
phous state proves that the drug was dispersed in a molecular
state with CD. All the XRD patterns of cilostazol CDs
inclusion complexes and physical mixtures are shown in
Fig. 4. The powder X ray diffraction pattern of pure cilostazol
exhibited a series of intense peaks at 2 value of 12.67, 12.98,
15.35, 15.76, 17.98, 18.71, 19.52, 22.19, and 22.58 which were
indicative of their crystallinity. However, the patterns of
CD, CD, and DM CD are all crystalline in nature while
only HP CD was amorphous. In physical mixtures of
cilostazol CDs, most of the principal peaks of cilostazol and
CDs are present and the curves are approximate superimpo
sition of the pattern of the raw materials. This indicated
partial interaction between the pure components in physical
mixtures.
On the other hand, inclusion complexes of cilostazol
HP CD and cilostazol DM CD in the ratio of 1:3,
prepared by coprecipitation method, have completely dif
fused diffraction patterns. These results suggested the forma
tion of amorphous inclusion complexes, whereas the
diffraction patterns of inclusion complexes of cilostazol
CD and cilostazol CD in the ratio of 1:3 were completely
different from that of pure cilostazol, CD, CD, and its
physical mixtures, suggesting the formation of partial inclu
sion complexes. Hence, it was concluded that HP CD and
DM CD are the suitable excipients for the preparation of
cilostazol inclusion complexes.
In vitro dissolution study of the drug in aqueous solution
is the rate limiting step for the absorption of poorly water
soluble drugs. In the phase solubility study, it was suggested
that the phosphate buffer pH 6.4 was the most suitable
dissolution medium as it showed maximum drug release.
Dissolution study in phosphate buffer pH 6.4 was carried out
for all cilostazol CD physical mixtures and inclusion com
plexes prepared in different ratios (1:1, 1:2, and 1:3) by using
coprecipitation and kneading methods. The dissolution pro
les of cilostazol CDs (inclusion complex and physical
mixtures) are compared with that of marketed formulation
and pure drug. The maximum mean cumulative percent drug
releasedSD for 120 min for all the ratios are shown in Table
Table IV. Similarity Factor (f2) and Students T Test Values for the Stability Prole of Cilostazol DM CD Inclusion Complex (1:3) by
Coprecipitation Method for 6 Months
Batch
f2
T stat
T cri
Reference batch compared with after 6 months of storage batch
66.85
0.5990
2.0859
668
Patel and Rajput

Stability Study
Fig. 6. Comparison of pharmacokinetic proles of pure cilostazol,

conventional tablet, and cilostazol DM CD inclusion complex
III. It is evident from the data that optimized cilostazol CD

inclusion complexes prepared by coprecipitation method
showed better drug release than other inclusion complexes,
physical mixtures, marketed formulation (Pletoz 50), and
drug solution. Cilostazol DM CD inclusion complex (1:3ra
tios) showed a 2.10 fold substantially higher drug release
compared to Pletoz 50 and 2.97 fold higher drug release
compared to cilostazol solution. The graphs of percent drug
dissolved vs. time were found to be nonlinear, suggesting that
the dissolution pattern did not follow zero order kinetics.
However, the correlation coefcients indicated that Higuchis
model was found to be the best t curve compared with zero
order and rst order kinetic for all the tested formulations.
The values of DE120 min and T50 study for cilostazol CDs
physical mixtures, inclusion complexes, cilostazol, and Pletoz
50 showed that inclusion complex of cilostazol DM CD in
the ratio of 1:3, prepared by coprecipitation method, was
having the highest dissolution efciency (89.59%) and lowest
time taken for 50% dissolution (T50 value 11.81 min) and
indicated that the inclusion complex of cilostazol DM CD
(1:3) prepared by coprecipitation method was having a rapid
and higher dissolution rate among all inclusion complexes
and physical mixtures.
Cilostazol DM CD inclusion complex having a ratio of

1:3 prepared by coprecipitation method showed satisfactory
in vitro dissolution results; hence, it was evaluated for
chemical stability study. The study was carried out as per
ICH Q1A (R2) and SUPAC IR guidelines and the withdrawn
samples up to 6 months were characterized by in vitro
dissolution study. Obtained data are represented in Fig. 5,
which shows that the change of cumulative percentage of
drug release after 6 months was from 2.0% to 5.0%. The
dissolution curve also showed similar changes in dissolution
patterns indicating that cilostazol was stable in the inclusion
complex. The results of Students t test and similarity factor
(f2) are mentioned in Table IV showing insignicant differ
ence between the dissolution proles. The calculated f2 values
for the batches are higher than 50. It can be concluded from
the data that there is insignicant change in the formulated
product on storage.
Comparative Bioavailability Study

Cilostazol DM CD inclusion complex (1:3) prepared
by coprecipitation method showed satisfactory results in all
preliminary studies; hence, it was selected for in vivo
absorption studies. The plasma concentration vs. time prole
after 24 h is shown in Fig. 6 and the pharmacokinetic
parameters are mentioned in Table V. The results show that
the Cmax of cilostazol DM CD inclusion complex is 2.23
fold higher than marketed formulation (Pletoz 50) and 4.11
fold higher than pure cilostazol. The value of T1/2 was also
decreased in cilostazol DM CD inclusion complex (7.46 h)
compared with Pletoz 50 (11.83 h) and pure cilostazol
(12.68 h). Relative bioavailability of cilostazol inclusion
complex was 1.53 times higher than marketed formulation.
A sharp reduction in T1/2 value indicated faster in vivo
absorption compared to Pletoz 50 and pure cilostazol. The
enhancement of oral absorption rate may be attributed to (1)
complete cilostazol incorporation into CD, (2) improved
dissolution because of the hydrophilic exterior surface of
Table V. Different Pharmacokinetic Parameters of Cilostazol in Rabbits after Oral Administration of 4.66 mg/kg of Cilostazol
Parameters #
Pure cilostazol
Marketed formulation (Platoz 50)
Cilostazol DM CD inclusion complex (1:3)
Cmax (ng/ml)
tmax (h)
ka (h 1)
kel (h 1)
T1/2 (h)
Vd (L)
AUC0 (ng h ml 1)
MRT (h)
Cl (ml h 1)
TCR (L/h)
Relative bioavailabilitya
1,150.96101*
3.901.32
0.770.006
0.050.002
12.681.84*
1.25730.03
21,007.191543
9.621.32
0.070.002
0.070.004
2,126.48185*
4.000.93
0.630.003
0.060.0032
11.831.56*
0.73050.05
34,197.551383
9.511.56
0.040.001
0.040.004
4,734.88252*
3.500.66
0.710.001
0.090.0022
7.461.58*
0.30200.007
52,625.591482
8.572.01
0.030.0023
0.030.0015
1.53
MeanSD, n 6
AUC area under the curve, MTR mean residence time, Cl clearance, TCR total clearance rate
a
Relative bioavailability (AUCInc. Comp/AUCTAB)(dose TAB/doseInc.Comp)
*p<0.05 (ANOVA test)

CD, and (3) the stability of inclusion complex in the
gastrointestinal tract.
CONCLUSION
Cilostazol formed its inclusion complexes in either its
neutral or charged form with DM CD and HP CD. CD
having larger cavity size was not suitable for the formation of
inclusion complex with cilostazol. The interaction between
cilostazol and DM CD was stronger than HP CD and
neutral CDs. Data obtained from the IR, DSC, and X ray
studies showed that it is possible to form inclusion complex
between cilostazol and CDs and the coprecipitation technique
produced amorphous complexes. DM CD improved the
oral bioavailability with low variability in vivo as evidenced
by the faster dissolution rate of cilostazol. A shorter T50% of
dissolution is found for the formulation prepared by copre
cipitation method. All the obtained data suggested that the
cilostazol DM CD inclusion complex (1:3) prepared by
coprecipitation method may have grater utility in the fast
dissolving dosage forms with possible enhancement of oral
bioavailability.
ACKNOWLEDGEMENT
The authors are thankful to The Maharaja Sayajirao
University of Baroda for providing nancial support to one
of the author (S.G.P.). The authors are also thankful to
Roquette Pharma, U.S.A., and Sun Pharma Advance
Research Center, Vadodara, India for their generous gift of
CD, DM CD, and HP CD and to Cadila Pharmaceut
icals, Ahmedabad, India for providing the gift sample of
cilostazol. Our sincere thanks to Food and Drug Laboratory,
Vadodara, for providing necessary facilities for animal study
in the laboratory.
REFERENCES
1. Medical EconomicsPhysicians. Physicians desk reference. 59th
ed. Stamford: Thomson; 2005. p. 2564 6.
2. Larsen KL. Large cyclodextrins. J Inclusion Phenom Macro
Chem. 2002;43(1 2):1 13.
3. Szejtli J, Osa T. Comprehensive supramolecular chemistry:
cyclodextrins, vol 3. Oxford: Pergamon; 1996.
4. Uekama K, Hirayama F, Irie T. Cyclodextrin drug carrier
systems. Chem Rev. 1998;98(5):2045 76.
5. Mosher G, Thompson D. Complexation and cyclodextrins.
encyclopedia of pharmaceutical technology, Vol. 19. New York:
Marcel and Dekker; 1999. p. 49 88.
6. Homans SW. A molecular mechanical force eld for the
conformational analysis of oligosaccharides: comparison of the
theoretical and crystal structure of Man1 3Man1 4GlcNAc.
Biochemistry 1990;29(39):9110 8.
7. Zuo Z, Kwon G, Stevenson B, Diakur J, Wiebe LI. Flutamide
hydroxypropyl cyclodextrin complex: formulation, physical
characterization, and absorption studies using the Caco 2 in vitro
model. J Pharm Pharmceut Sci. 2000;3(2):220 7.
669
8. Dollo G, Corre PL, Chollet M, Chevanne F, Bertault M, Burgot JL,
et al. Improvement in solubility and dissolution rate of 1,2 dithiole
3 thiones upon complexation with cyclodextrin and its hydrox
ypropyl and sulfobutyl ether 7 derivatives. J Pharm Sci.
1999;88:889 95.
9. Duchene D. New trends in cyclodextrins and derivatives. Paris:
Editios de Sante; 1992. p. 351 407.
10. Fromming KH, Szejtli J. Cyclodextrins in pharmacy (Topics in
inclusion science). Dordrecht: Kulwer Academic; 1994.
11. Thompson DO. Cyclodextrin enabling excipients: their present
and future use in pharmaceuticals. Crit Rev Ther Drug Carrier
Syst. 1997;14:1 104.
12. Masson M, Loftsson T, Masson G, Stefansson E. Cyclodextrin as
permeation enhancer: some theoretical evaluations and in vitro
testing. J Control Release 1999;59(1):107 18.
13. Govt. of India, Ministry of Health and Family Welfare.
Pharmacopoeia of India, vol I. 4th ed. New Delhi: Controller
of Publication; 1996. p. A 144 7.
14. Higuchi T, Connors KA. Phase solubility techniques. Adv Anal
Chem Inst. 1965;4:117 212.
15. Attama AA, Ndibe ON, Nnamani PO. Studies on diclofenac
cyclodextrin inclusion complexes. J Pharm Res. 2004;3:47 9.
16. Liu X, Lin HS, Thenmozhiyal JC, Chan SY, Ho PC. Inclusion of
acitretin cyclodextrins: phase solubility, photostability, and phys
icochemical characterization. J Pharm Sci. 2003;92(12):2449 57.
17. Hidetoshi A, Kiyokazu Y, Kouzou M, Tetsumi I, Fumitoshi H,
Kaneto U. Comparative studies of the enhancing effect of
cyclodextrins on the solubility and oral bioavailability of
tacrolimus rats. J Pharm Sci. 2001;90(6):690 701.
18. Rawat S, Jain SK. Rocoxib beta cyclodextrin inclusion complex
for solubility enhancement. Pharmazie. 2003;58(9):639 41.
19. Manca ML, Zaru M, Ennas G, Valenti D, Sinico C, Loy G,
Fadda AM. Diclofenac cyclodextrin binary systems: physical
characterization and in vitro dissolution and diffusion studies.
AAPS Pharm Sci Tech. 2005;6(3):E464 72.
20. Castro Hermida JA, Gomez Couso H, Ares Mazas ME, Gonza
lez Bedia MM, Castaneda Cancio N, Otero Espinar FJ, et al.
Anticryptosporidal activity of furan derivative G1 and its
inclusion complex with beta cyclodextrin. J Pharm Sci. 2004;93
(1):197 206.
21. Nalluri BN, Chowdary KPR, Murthy KVR, Satyanarayana V,
Hayman AR, Becket G. Inclusion complexation and dissolution
properties of nimesulide and meloxicam hydroxypropyl cyclo
dextrin binary systems. J Inclusion Sci Macro Chem. 2005;53(1
2):103 10.
22. United State Pharmacopoeia. Dissolution study, XXIII, NF
XVIII. Rockville: USP Convention; 1995. p. 1791 9.
23. Khan KA, Rhodes CT. Concept of dissolution efciency. J
Pharm Pharmacol. 1975;27(1):48 9.
24. Moore JW, Flanner HH. Mathematical comparison of dissolution
proles. Pharm Technol. 1996;20(6):64 74.
25. Shah VP, Tsong Y, Sathe P, Liu JP. In vitro dissolution prole
comparison statistics and analysis of the similarity factor (f2).
Pharm Res. 1998;15(6):889 96.
26. Gibaldi M, Perrier D. Pharmacokinetics (Drugs and the phar
maceutical sciences), 2nd edn, vol. 15. New York: Marcel
Dekker, 1982, revised and expanded.
27. Challa R, Ahuja A, Ali J, Khar RK. Cyclodextrins in drug delivery:
an updated review. AAPS Pharm Sci Tech. 2005;06(02):E329 57.
28. Wan Liang L, Qiang Z, Li Z, Hua W, Rong Yu L, Li Feng Z, et
al. Antipyretic, analgesic and anti inammatory activities of
ketoprofen cyclodextrin inclusion complexes in animals. Bio
Pharm Bull. 2004;27(10):1515 20.
29. Ali H, Marzouqi A, Shehatta I, Jobe B, Dowaidar A. Phase
solubility and inclusion complex of itraconazole with cyclo
dextrin using supercritical carbon dioxide. J Pharm Sci. 2006;95
(2):292 304.

DOI: 10.1208/s12249 009 9253 y
Research Article
ChitosanChondroitin Sulfate Based Matrix Tablets for Colon Specific Delivery
of Indomethacin
Jitendra R. Amrutkar1 and Surendra G. Gattani1,2
Abstract. The different approaches for targeting orally administered drugs to the colon include coating
with pH dependent polymers, design of time release dosage forms, and the utilization of carriers that are
degraded exclusively by colonic bacteria. The aim of the present study was to develop a single unit, site
specic drug formulation allowing targeted drug release in the colon. Matrix tablets were prepared by
wet granulation using cross linked chitosan (ChI) and chondroitin sulfate (ChS) polysaccharides as
binder and carrier. ChS was used to form polyelectrolyte complexes (PEC) with ChI, and its potential as
a colon targeted drug carrier was investigated. Indomethacin was used as a model drug. The ChI and ChS
PEC was characterized by Fourier transform infrared spectroscopy (FTIR), differential scanning
calorimetry (DSC) and powder X ray diffraction studies (XRD). The matrix tablets were tested in vitro
for their suitability as colon specic drug delivery systems. FTIR demonstrated that the PEC forms
through an electrostatic interaction between the protonated amine (NH3+) group of ChI with the free
carboxylate (COO ) group and sulfate (SO42 ) group of ChS. DSC and XRD indicated that the PEC has
different thermal characteristics from ChI or ChS. The dissolution data demonstrates that the dissolution
rate of the tablet is dependent upon the concentration of polysaccharide used as binder and matrix and
time of cross linking. The study conrmed that selective delivery of indomethacin to the colon can be
achieved using cross linked ChI and ChS polysaccharides.
KEY WORDS: chitosan; chondroitin sulfate; colonic delivery; cross linking; indomethacin.
INTRODUCTION
Colonic drug delivery has gained increased importance
not just for the delivery of drugs for the treatment of local
diseases of colon but also for its potential for the delivery of
proteins and peptides (1). Over the last few years, different
approaches have been reported in order to achieve specic
colonic drug delivery. Most of the previous literature reports
on colonic targeting have focused on the development of a
colonic delivery system based on time and pH dependent
delivery systems as well as systems that utilize bacteria, which
colonizes the colon or enzymes produced by these bacteria to
affect drug release (2,3). The poor site specicity problem
occurs with time release dosage form due to large variation in
gastric emptying time (4) and passage across the ileocecal
junction. In addition, poor site specicity of pH dependent
system was very well established due to large variation in pH
of the gastrointestinal tract (GIT) (5,6). Biodegradable
systems formulated using natural polysaccharides are increas
ingly being developed (7,8). Use of naturally occurring
polysaccharides is attracting attention for drug targeting to
the colon, since these polymers of monosaccharides are found
Department of Pharmaceutics and Q.A, R. C. Patel College of

Pharmacy, Shirpur 425 405, Dhule, Maharashtra, India.
2
To whom correspondence should be addressed. (e mail: sggattani@
rediffmail.com)
in abundance, inexpensive, and available in variety of

structures with varied properties (9). They can be easily
modied chemically and biochemically and are non toxic,
hydrophilic, gel forming, as well as biodegradable in nature.
Conventionally, various polysaccharides are used in the tablet
formulations to retard drug release. These have been used
either as matrices or as a coating material. For matrices,
generally a high concentration of polymer is required.
Alternatively, these can be used as binders in tablets (10).
Thus, varying the polysaccharides and their concentration
affects drug release from the prepared tablet. Based on the
above assumption, two different polysaccharides, namely,
chitosan (ChI) and a chondroitin sulfate (ChS) were selected
for the present study. ChI ChS complex erodes slowly in
phosphate buffer at pH values higher than 6.5, and this
behavior leads to suppression of the initial drug release in the
upper segments of GIT and controls release in the colon
where pH value is in the range of 6.5 7.0 (11).
In the present investigation, for the rst time, authors
have formulated matrix tablets for colonic delivery using
novel cross linked ChI ChS PEC. Indomethacin was select
ed as a model drug because it has good indication for
colonic delivery (12,13). Indomethacin is widely used as
non steroidal anti inammatory drug. In vitro drug release
studies were carried out in simulated colonic uid with and
without rat cecal content, also by using Bacteroides ovatus
culture due to its known polysaccharide degradable activity
(14).
670
16 125
25
215
F12
671
75
21 5
21 5
78 375
ChIChS Based Matrix Tablets for Colon Specific Delivery of Indomethacin
F11
10 75
25
215
5 375
25
215
16 125
25
215
75
21 5
21 5
89 125
F10
F9
Methods
75
16 125
16 125
89 125
Indomethacin was a generous gift from Micro Lab

Limited (Bangalore) India. Chitosan (degree of deacetylation
85%) was obtained from Central Institute of Fisheries, Kochi,
India. Chondroitin sulfate was a gift from Markson Pharma.,
Goa, India. Culture of B. ovatus MTCC 3298 was made
available from microbial type culture collection, Chandigarh,
India. All other ingredients used were of analytical grade.
75
21 5
21 5
83 75
Materials
F8
10 75
25
215
5 375
25
215
16 125
25
215
10 75
25
215
5 375
25
215
25
215
25
215
25
215
75
5 375
5 375
121 375
F7
F6
75
16 125
16 125
89 125
75
10 75
10 75
105 25
F5
F4
75
5 375
5 375
121 375
75
16 125
16 125
89 125
16 125
F3
F2
Characterization of Tablets
The properties of the compressed matrix tablet, such as
hardness, friability, weight variation, and content uniformity,
were determined using reported procedure (16). Briey,
hardness was determined using Monsanto hardness tester
Sr
Tablets weighing 215 mg containing 75 mg of indometh

acin were compressed on a ten station rotary machine
(Karnavati Engg., India) using a 7 mm round, concave, plain
die punch.
NoName of Ingredient
Preparation of Tablets
ChI chitosan, ChS chondroitin sulfate
where TD is the tap density and BD is the bulk density
F1
Indomethacin
ChI
ChS
MCC (Emocel)
Ethyl Cellulose (10%
Alcoholic Solution)
06
Starch paste (10%)
07
PVP K30
08
Magnesium stearate
Total weight of tablet
TD BD
100
TD
01
02
03
04
05
CI
75
10 75
10 75
105 2
10 75
Before compression, granules were evaluated for their

characteristic parameters. Angle of repose was determined by
funnel method. Bulk density and tapped density were
determined by cylinder method, and Carrs index (CI) was
calculated using the following equation (15).
75
5 375
5 375
121 375
5 375
Characterization of Granules
Formulation codes
All the powdered ingredients were weighed, mixed, and

granulated with the binder paste prepared as above. This
mixture was thoroughly blended manually and passed
through a sieve with a nominal aperture of 1 mm. The
granules were prepared and dried in an oven at a tempera
ture between 30C and 40C for 4 h. The dried granules were
screened, mixed with lubricants, and stored at room temper
ature for tableting.
Quantity/tablet (mg)
Formulation of Granules
Table I. Formulation Composition of Matrix Tablets of Indomethacin
Binder paste of the polysaccharides in various concen

trations were prepared by mixing the weighed amount of ChI
in a 1% v/v solution of acetic acid in distilled water with
magnetic stirrer for 0.5 h so as to enable the ChI to swell. To
this mixture, weighed amount of ChS was added slowly. The
mixing was carried out for different time intervals such as 12,
18, and 24 h. The cross linked paste was used as a binder for
the powder mixture during wet granulation (Table I).
75
10 75
10 75
105 25
Preparation of Binder Paste
672
(Campbell Electronics, Mumbai, India), and friability was
determined using Roche friability testing apparatus (F.
Hoffmann La Roche Ltd, Basel, Switzerland). Drug content
studies were carried out to evaluate the amount of drug
present in the prepared tablet.
Determination of Drug Content
Randomly, 20 tablets were selected from each formula
tion and tested for its drug content. The tablets were nely
powdered, and a quantity of powder equivalent to 75 mg of
indomethacin was accurately weighed and transferred to
100 ml volumetric asks containing approximately 50 ml of
methanol. The asks were shaken in a shaking water bath at
37C for 3 h to solubilize the drug. The volume was made up
with methanol and mixed thoroughly. The solutions were
ltered through membrane lter paper and analyzed for the
content of indomethacin using UV spectrophotometer (1700,
Shimadzu, Japan) at 318 nm.
The electrostatic interaction study of polymers was done
by Fourier transform infrared spectroscopy (FTIR) spectro
scopic analysis. ChI, ChS, and cross linked ChI ChS complex
were mixed separately with IR grade KBr in the ratio of
100:1, and corresponding pellets were prepared by applying
10 metric ton of pressure in hydraulic press. The pellets were
scanned over a wave range of 4,000 400 cm 1 using FTIR
spectrophotometer (8400S, Shimadzu, Japan).
Differential scanning calorimetry (DSC) of ChI, ChS,
and cross linked ChI ChS complex were performed by using
Mettler Toledo DSC 822e instrument. The cell had a nitrogen
purge owing at approximately 40 cm3/min. Sample (5 mg)
was scanned in aluminum pan over a temperature range
between 75C to 300C at a scanning rate of 5C/min. An
indium pan served as reference, and all scans were performed
in triplicate. The instrument was calibrated before sample
analysis, using an indium standard.
Powder X Ray Diffraction Studies
Powder X ray diffraction (XRD) pattern of polymers
were obtained using Philips diffractometer (PW3710, Almelo,
The Netherland) and Cu K line as a source of radiation, which
was operated at the voltage 40 kV and the current 30 mA. All
samples were measured in the 2 angle range between 0 and
80 with a scanning rate of 1/min and a step size of 0.02.
Amrutkar and Gattani

08L, Electrolab, Mumbai, India) at 75 rpm and a temperature
of 370.5C. Initial drug release studies were conducted in
900 ml of 0.1 N HCl for 2 h. Then, dissolution media was
replaced with phosphate buffer pH 7.4 (900 ml) for 3 h as
average small intestine transit time is 3 h and nally in
phosphate buffer pH 6.8 up to 24 h. Samples were withdrawn
after regular intervals of time and replaced with the same
media. The drug release was analyzed spectrophotometrically
(UV 1700, Shimadzu, Japan) at a wavelength of 318 nm.
Preparation of Rat Cecal Medium
Male Wistar rats weighing 110 125 g maintained on
normal diet were used. Rat cecal content medium at 2% and
4% w/v level was obtained after 7 days of enzyme induction
with 1 ml of 2% w/v ChI and ChS dispersion provides the best
condition for assessing the susceptibility of ChI and ChS to
colonic bacterial degradation (17). Forty ve minutes before
the commencement of drug release studies, six rats were
killed by spinal traction, the abdomens were opened, and
cecum was traced, ligated at both ends, dissected, and
immediately transferred into phosphate buffer pH 6.8 previ
ously bubbled with sterile nitrogen. The entire procedure
mentioned above was carried out under sterile nitrogen in
order to maintain anaerobic conditions.
The susceptibility of ChI and ChS in matrix tablets to
enzymatic action of colonic bacteria was assessed by continu
ing the drug release studies in 100 ml of phosphate buffer pH
6.8 containing 2% and 4% w/v rat cecal content (18). Drug
release studies were carried out in USP XXIII dissolution
apparatus I (75 rpm, 370.5C) with slight modication. In
this modication, beaker (capacity 250 ml) containing 100 ml
of dissolution medium was immersed in water containing
1,000 ml vessel, which was, in turn, to water bath apparatus.
The tablets were placed in the basket and immersed in the
dissolution medium containing rat cecal contents. The exper
iment was carried out with continuous nitrogen supply into
medium to simulate anaerobic environment of the cecum.
The drug release studies were carried out for 24 h (usual
colonic transit time is 20 30 h), and 1 ml test samples were
withdrawn at 1 h interval without prelter and replaced with
1 ml of fresh phosphate buffer pH 6.8 bubbled with nitrogen.
Withdrawn samples were diluted with 1 ml of methanol to
ensure the solubility of nely suspended drug particles
released due to digestion of polysaccharides by cecal
enzymes. The volume was made up to 10 ml with phosphate
buffer pH 6.8 and centrifuged, and the supernatant was
ltered through bacteria proof lter and ltrate analyzed for
indomethacin at 318 nm (18) using UV spectrophotometer
(1700, Shimadzu, Japan).
Dissolution Study Using Culture of Bacteroides ovatus
Drug Release Studies

The ability of the prepared tablets to retard drug release
in the physiological environment of the stomach, small
intestine, and the colon was assessed by conducting drug
release studies in simulated stomach, small intestine, and
colonic uid, respectively. The changing pH media, method I,
USP XXIII, for delayed release tablets was used (10).
Dissolution test was conducted in USP I apparatus (TDT
In addition, polysaccharide prepared matrix tablet with

the potential for site specic delivery to the colon has been
evaluated using culture of B. ovatus. The culture was
cultivated anaerobically and maintained on nutrient dissolu
tion media (19).
After a 5 h dissolution study, the phosphate buffer pH
7.4 was replaced with nutrient medium containing culture of
B. ovatus (109 CFU/ml). All of the material was then

transferred to a sterile jar of dissolution apparatus. Sterile
nitrogen was bubbled through vessels to maintain anaerobic
condition. Samples (1 ml; after each interval) were withdrawn
periodically and diluted with phosphate buffer pH 6.8. Each
sample was subsequently centrifuged at 2,000 rpm for 5 min; the
absorbance of resulting supernatant was measured at 318 nm.
Kinetic Analysis of Dissolution Data
To study the mechanism of drug release from the matrix
tablets, the release data were tted to zero order, rst order,
and Higuchi equations (20). The dissolution data was also
tted to the well known exponential equation (Korsmeyer
equation), which is often used to describe the drug release
behavior from polymeric systems (21).
Mt
log k n log t
M1
where Mt is the amount of drug release at time t; M is the

amount of drug release after innite time; k is a release rate
constant incorporating structural and geometric character
istics of the tablet; and n is the diffusional exponent indicative
of the mechanism of drug release.
The cumulative percent of indomethacin released from
cross linked ChI ChS matrix tablets (n=3) in the dissolution
medium at 24 h with and without rat cecal contents (control
study) was compared, and the statistical signicance was
tested by using Students t test. A value of P<0.05 was
considered statistically signicant.
RESULTS
Characterization of Granules
The granules for matrix tablet were prepared according
to the composition given in Table I and characterized with
respect to angle of repose, bulk density, tapped density, Carrs
673
index, and total drug content (Table II). Physical character

ization of granules was found to be in acceptable range (15).
Characterization of Tablets
Weight variation, friability, hardness, and content unifor
mity (98 101%) falls within the limits of Indian Pharmaco
poeia (16).
The IR spectra of ChI (A), ChS (B) and cross linked
ChI ChS complex (C), respectively, are shown in Fig. 1. IR
spectra of ChI (A) conrm the presence of OH and N H
stretching vibration at 3,435 cm 1 in which the OH stretching
vibration are overlapped by N H stretching; the absorption of
C H stretching of methyl or methylene group of ChI is at
2,920 cm 1; the peak at 1,648 cm 1 corresponds to the amide
bonds; the peak at 1,599 cm 1 corresponds to the symmetrical
stretch vibration of amino group; the peak at 1,384 cm 1
corresponds to stretching vibrations of C=N bond; the stretch
vibrations of C O are found at 1,079 and 1,030 cm 1.
The IR spectra of ChS (B) conrm the presence of OH
and N H stretching vibration at 3,436 cm 1 in which the OH
stretching vibration is overlapped by N H stretching; the
absorption of C H stretching of methyl or methylene group is
at 2,923 cm 1; the stretch vibrations of C O are found at
1,039 cm 1; the peak at 1,636 cm 1 corresponds to the amide
bands; the bands at 1,413 and 1,380 cm 1 are due to the
coupling of the C O stretch vibration and O H variable angle
vibration and indicates the existence of the free carboxyl
group; the peak at 1,253 cm 1 corresponds to the stretching
vibrations of S=O bond (SO42 ), which is the characteristic
absorption peak of ChS.
In the IR spectra of cross linked ChI ChS complex (C),
the peak at 1,648 cm 1 (CONH2) in the spectrum of ChI shifts
to 1,642 cm 1, whereas the peak of amino group disappeared,
which indicates cross linking of amino group of ChI with ChS.
The peak at 1,253 cm 1 (S=O) turns weaker and shifts to
1,250 cm 1, which indicates cross linking of sulfate group of
ChS with ChI. The coupling peak of the C O stretch
vibrations and O H variable angle vibration turns weaker
Table II. Characterization of Granules and Matrix Tablets of Indomethacin

Formulation Code
Parameters
Granules
Angle of repose (deg)
Bulk Density (g/mL)
Tap Density (g/mL)
Carrs Index
Total drug content (%)
Tablets
Weight variation (%)
Friability (%)
Hardness (kg/cm2)
Content uniformity (%)
F7
F8
F9
F10
F11
F12
281.1
0.68
0.78
13.00.5
98.03.6
250.8
0.57
0.65
12.30.5
98.53.9
270.9
0.46
0.53
13.21.1
99.03.7
220.6
0.53
0.61
13.10.6
98.24.1
281.2
0.43
0.50
14.01.2
98.73.8
270.9
0.41
0.48
14.61.1
99.13.7
4.0
0.24
5.50.1
98.34.3
3.0
0.32
5.70.2
98.74.0
2.0
0.11
6.10.4
99.00.5
3.0
0.22
5.70.3
98.44.3
2.0
0.19
5.60.4
98.84.0
4.0
0.17
6.30.2
99.33.5
*All values represent meanSD (n 3)
674
Fig. 1. FTIR spectras of a ChI, b ChS, and c cross linked ChI ChS.
ChI chitosan, ChS chondroitin sulfate, ChI ChS chitosan chondroitin
sulfate polyelectrolyte complex
and divide into two peaks (the peak at 1,413 cm 1 divide into
1,411 and 1,402 cm 1, the peak at 1,380 cm 1 divide into 1,383
and 1,378 cm 1), which indicates that one part of free
carboxyl group reacted with ChI.
It is evident from Fig. 2a that ChI powder exhibited one
endothermic and exothermic transition each at 52.90C and
253.11C, respectively. The DSC thermogram of ChS
revealed one endotherm at 94.24C and one exotherm at
211.89C in Fig. 2b. However, cross linked ChI ChS exhibited
a large endotherm at 75.30C followed by another endotherm
at 185.32C (Fig. 2c).
Powder X-Ray Diffraction Studies
XRD patterns of ChI, ChS, and cross linked ChI ChS is
shown in Fig. 3. The characteristic peaks appeared in the
XRD of ChI (Fig. 3a) and ChS (Fig. 3b) at different angles.
Fig. 3. Powder X ray diffraction patterns of a ChI, b ChS, and c

cross linked ChI ChS. ChI chitosan, ChS chondroitin sulfate, ChI
ChS chitosan chondroitin sulfate polyelectrolyte complex
ChI and ChS is crystalline in the solid state, but its diffracto
gram exhibits well dened peaks at 2=9.4 and 19.1 for ChI
(Fig. 3a) and 46.2 for ChS (Fig. 3b). The complexes (Fig. 3c)
give diffractograms showing only the amorphous part, and the
intensity of two peaks mentioned above, which is character
istic of ChI and ChS, is much lower than the peaks in
individual XRD pattern of ChI and ChS.
In Vitro Drug Release Studies
Fig. 2. DSC thermograms of a ChI, b ChS, and c cross linked ChI

ChS. ChI Chitosan, ChS chondroitin sulfate, ChI ChS chitosan
chondroitin sulfate polyelectrolyte complex
Formulations F1 to F6 were disintegrated within 45 min

in 0.1 N HCl and hence were not studied further. Studies
using cross linked ChI and ChS complex as a binder showed
that, at a concentration of 5% (F7) the drug release in the
initial 5 h was not more than 10%. It indicates that cross
linked polymers are capable of preventing drug from being
released completely in the physiological environment of
stomach and small intestine. Increasing the concentration of
ChI and ChS complex in the formulation up to 10% in F8 and
up to 15% in F9 further retards the drug release as compared
to F7. At the end of 5 h, the drug release was found to be 6%
in F8 and to 4% in F9.
Effects of cross linking time of ChI and ChS on the
percentage drug release at different time periods from the
formulation F7 to F12 without rat cecal content are shown in
Fig. 4. Results indicate that, as the degree of cross linking
time of ChI and ChS increases, drug release from matrix
tablet decreases. Drug released after 24 h cross linking from
formulation F7 to F12 was found to 57.12%, 55.1%, 45.64%,
54.0%, 52.12%, and 51.9%, respectively, in the absence of rat
675
Fig. 4. Effect of cross linking time of ChI and ChS on the percentage of drug released at different time periods from matrix tablets from the
formulations F7 to F12 without rat cecal content (a 12 h cross linking; b 18 h cross linking, and c 24 h cross linking). ChI chitosan, ChS
chondroitin sulfate. All results were calculated as mean3SD
study with culture of B. ovatus provides the best alternative

method to dissolution study with 4% w/v rat cecal content.
cecal contents (control study). Hence, F9 formulation was

selected for further study, as it shows only 45.64% release
after 24 h. On the other hand, in the presence of 2% and 4%
w/v rat cecal content, drug release was found to be 69.84%
and 99.5%, respectively, from F9 formulation (Fig. 5).
Pure colonic bacteria B. ovatus selected for its known
polysaccharide degradable activity (14). The results of
dissolution experiment using B. ovatus MTCC 3298 showed
that drug is rapidly released after predetermined lag time
from matrix tablet, as shown in Fig. 6. Result of dissolution
Drug release data of F9 formulation in presence of rat

cecal content showed good t into the rst order equation
(r2 =0.90), partially tted in zero order equation (r2 =0.88)
and into the Higuchi equation (r2 =0.85) and show high
linearity with Korsmeyer equation r2 =0.91. Value of release
Fig. 5. Percent cumulative release from optimized (F9) formulation

with 2% and 4% w/v rat cecal content. All results were calculated as
mean3SD
Fig. 6. Percent cumulative release from optimized (F9) formulation

in presence of culture of bacteroides ovatus (109 CFU/ml); all results
were calculated as mean3SD
Kinetic Analysis of Dissolution Data
676
exponent n determined for F9 in the presence of rat cecal
content is 0.72, indicating combined effect of diffusion and
predominately erosion mechanisms for controlled drug
release.
The cumulative percent of indomethacin released from
cross linked ChI ChS matrix tablets (n=3) in the dissolution
medium at 24 h with and without rat cecal contents (control
study) were compared statistically using Students t test. An
increase in the amount of indomethacin released was found
signicant (p<0.05) from F9 with rat cecal contents when
compared to the in vitro release without rat cecal contents.
DISCUSSION
Angle of repose was less than 30 for all the batches of
granules, indicating satisfactory ow behavior. Other param
eters for granules were also determined and found to be in
acceptable range (15). The weight variation and friability was
less than 4% and 0.4%, respectively. Good drug content
uniformity was found among different batches of the tablets
(Table II). FTIR behavior reects the interaction between the
amino group of ChI with the sulfate and free carboxyl group
of ChS. PEC was obtained after the interaction between the ChI
and ChS. From DSC studies, the endothermic peak (189C) can
be assigned to the interaction between ChI and ChS. It was
observed that the XRD of cross linked ChI ChS PEC shows
absence of characteristic peaks of ChI and ChS, and intensity of
peaks in cross linked ChI ChS PEC was also reduced.
The purpose of colon targeted drug delivery system is
not only to prevent the drug from being release in the
physiological environment of the stomach and intestine but
also to release the drug in the colon after enzymatic
degradation of polysaccharides from matrix tablets by colonic
bacteria. Hence, the ability of the polymers used in the
formulations to retain the integrity of tablet in upper GI tract
was assessed by conducting drug release studies in 0.1 N HCl
for 2 h and phosphate buffer pH 7.4 for 3 h (condition
mimicking mouth to the colon transit).
The decrease in release rate on increasing the concen
tration of ChI and ChS can be explained on the basis that, as
the concentration of binder in the system is increased,
hardness, porosity, and capillary sizes are reduced. This
reduces the wicking of water into the tablet, which reduces
the disintegration and dissolution processes. These tablets
formulated using ChI did not show any swelling in basic
environments, so drug release due to swelling and polymer
erosion was minimized. This explains why the rate of drug
release was not as high as in the case of other swellable gums.
The presence of rat cecal contents or B. ovatus culture in
the dissolution medium resulted in improved drug release at
different periods when compared to control, indicating that
polysaccharides are degraded by colonic bacterial species. It
appears that several contaminants, grown in the dissolution
medium, might have broken down the matrix tablets. In fact,
a bioreactor (with a relevant culture medium) was used by
earlier workers to test the colon specic delivery systems
(14). However, due to several limitations, alternative meth

ods, like rat cecal content medium or a dissolution medium
containing a specic enzyme, were developed.
Thus, systems formulated using ChI as a binder have
been found to protect majority of drug release during the
usual upper GIT transit time of 5 h. However, there have
been a number of reports where ChI has been found to be
digested by the microora of the colon (8,22,23,25). On
similar lines, once these matrix tablets reaches the colon, ChI
and ChS shall be broken down by the microora of the colon,
and the total amount of drug shall be released from the
dosage form. These tablets can also tolerate variation in
upper GIT transit time, since the rate of drug release before
arrival into the colon remains retarded. From the results, it
seems that the complex, i.e., cross linked ChI ChS is more
susceptible for bacterial degradation and may release the
drug completely in the colon when compared with drug
release in the absence of microbial ora.
However, it has been reported that when water solubility
of the drug is low, as is in the case of indomethacin, the
possibility of release by diffusion is practically zero, and
release takes place by surface erosion (24,25). The kinetic
data obtained from in vitro drug release studies in presence of
rat cecal content indicates that the drug release was followed
by mixed order transport, i.e., release of the drug from tablets
was by more than one mechanism. The P value found to be
less than 0.05 denotes signicant difference at 24 h in the
amount of indomethacin released from F9 with rat cecal
contents when compared to the in vitro release without rat
cecal contents.
CONCLUSION
Based on the results of the present investigation, it can
be concluded that polyelectrolyte complex of ChI ChS might
assist in improving drug delivery to the colon.
ACKNOWLEDGMENTS
Authors are thankful to the Principal, R.C. Patel College
of Pharmacy for providing facilities to carry out the research
work. The authors acknowledges Micro Lab Limited, Banga
lore, India, Central Institute of Fisheries, Kochi, and Markson
Pharma., Goa, India, for providing gift samples of indometh
acin, chitosan, and chondroitin sulfate, respectively.
REFERENCES
1. Gupta K, Beckert TE, Price JC. A novel pH and time based
multi unit potential colonic drug delivery system. I. Develop
ment. Int J Pharm. 2001;213:83 91.
2. Rodriguez M, Vila Jato JL, Torres D. Design of a new multi
particulate system for potential site specic and controlled drug
delivery to the colonic region. J Control Release. 1998;55:67 77.
3. Khan MZ, Prebeg Z, Kurjakovic N. A pH dependent colon
targeted oral drug delivery system using methacrylic acid
copolymers. I. Manipulation of drug release using Eudragit
L100 55 and Eudragit S100 combinations. J Control Release.
1999;58:215 22.
4. Davis SS, Hardy JG, Taylor MJ, Stockwell A, Whalley DR,
Wilson CG. The in vivo evaluation of an osmotic device (Osmet)
using gscintigraphy. J Pharm Pharmacol. 1984;36:740 2.

5. Ashford M, Fell JT, Attwood D, Sharma HL, Woodhead PJ. An
in vivo investigation into the suitability of pH dependent
polymers for colonic targeting. Int J Pharm. 1993;95:193 9.
6. Evans DF, Pye G, Bramely R, Clark AG, Dysoa TJ. Measure
ment of gastrointestinal pH proles in normal ambulant human
subjects. Gut. 1988;29:1035 41.
7. Sinha VR, Kumria R. Colonic drug delivery: prodrug approach.
Pharm Res. 2001;18:557 64.
8. Sinha VR, Kumria R. Polysaccharides for colon specic drug
delivery. Int J Pharm. 2001;224:19 38.
9. Hovgaard L, Brondsted H. Current applications of polysacchar
ides in colon targeting. Crit Rev Ther Drug Carrier Syst.
1996;13:185 223.
10. Sinha VR, Kumria R. Binders for Colon specic drug delivery:
an in vitro evaluation. Int J Pharm. 2002;249:23 31.
11. Tapia C, Escobar Z, Costa E, Hagar J, Valenzuela F, Basualto
C, et al. Comparative studies on polyelectrolyte complex and
mixtures of chitosan carageenan as prolonged diltiazem clo
rhydrate release systems. Eur J Pharm Biopharm. 2004;57:
65 75.
12. Hull MA, Gardner H, Hawcroft G. Activity of the non steroidal
anti inammatory drug indomethacin against colorectal cancer.
Cancer Treat Rev. 2003;29:309 20.
13. Kapitanovic S, Cacev T, Antica M, Kralj M, Cavric G, Pavelic K, et
al. Effect of indomethacin on E cadherin and catenin expression
in HT 29 colon cancer cells. Exp Mol Pathol. 2006;80:91 6.
14. Chourasia MK, Jain SK. Pharmaceutical approaches to colon
targeted drug delivery systems. J Pharm Sci. 2003;6(1):33 66.
15. Aulton ME, Wells TI, editors. . The science of dosage form
design. London, UK: Churchill Livingstone; 1988.
16. Government of India Ministry of Health and Family Welfare. The
pharmacopoeia of India. Delhi, India: Controller of Publication; 2007.
677
17. Krishnaiah YSR, Bhaskar Reddy PR, Satyanarayana V, Karthi

keyan RS. Studies on the development of oral colon targeted
drug delivery systems for metronidazole in the treatment of
amoebiasis. Int J Pharm. 2002;236:43 55.
18. Krishnaiah SR, Satyanarayana S, Rama Prasad YV, Narasimha
Rao S. Evaluation of gaur gum as a compression coat for drug
targeting to colon. Int J Pharm. 1998;171:137 46.
19. Macfarlane GT, Hay S, Macfarlane S, Gibson GR. Effect of
different carbohydrates on growth, polysaccharidase and glyco
sidase production by Bacteroides ovatus, in batch and continuous
culture. J Appl Bacteriol. 1990;68(2):179 87.
20. Higuchi T. Mechanism of sustained action medication: theoreti
cal analysis of rate of release of solid drugs dispersed in solid
matrices. J Pharm Sci. 1963;52:1145 9.
21. Korsmeyer RW, Gurny R, Docler E, Buri P, Peppas NA.
Mechanism of solute release from porous hydrophilic polymers.
Int J Pharm. 1983;15:25 35.
22. Tozaki H, Komoike J, Tada C, Maruyama T, Terabe A, Suzuki T,
et al. Chitosan capsules for colon specic drug delivery:
improvement of insulin absorption from the rat colon. J Pharm
Sci. 1997;86:1016 21.
23. Tozaki H, Fujita T, Odoriba T, Terabe A, Suzuki T, Tanaka C, et
al. Colon specic delivery of R68070, a new thromboxane
synthase inhibitor, using chitosan capsules: therapeutic effect
against 2, 4, 6, trinitrobenzene sulphonic acid induced ulcerative
colitis in rats. Life Sci. 1999;64:1155 62.
24. Lapidus H, Lordi NG. Some factors affecting the release of a
water soluble drug from a compressed hydrophilic matrix. J
Pharm Sci. 1966;55:840 3.
25. Nigalaye AG, Adusumilli P, Bolto S. Investigation of prolonged
drug release from matrix formulations of chitosan. Drug Dev Ind
Pharm. 1990;16:449 67.

DOI: 10.1208/s12249 009 9258 6
Research Article
Effect of EDTA and Methionine on Preventing Loss of Viscosity
of Cellulose-Based Topical Gel
Junyan A. Ji,1,2 Erika Ingham,1 and John Y. Wang1
Abstract. Methylcellulose and hydroxypropylmethylcellulose (hypromellose) are used in topical
formulations of a protein to form a viscous hydrogel. Five lots of hypromellose raw material were made
into 3% gel; all showed viscosity loss after sterilization by autoclave. EDTA (edetate disodium)
minimized the viscosity loss caused by autoclaving in the presence of up to 100 ppm H2O2. These results
suggest that EDTA may prevent loss of viscosity of the hydrogel when peroxide is present. H2O2 at low
levels (2 50 ppm) caused signicant viscosity loss over time at either 40C or 5C in 3% methylcellulose
or hypromellose gel. EDTA slowed the rate of viscosity loss during storage under stress by H2O2 but did
not completely prevent the loss. Methionine was effective in completely preventing gel viscosity loss
during storage in the presence of up to 50 ppm H2O2. On the basis of these results, it is recommended
that methionine be added to the protein topical formulation as a stabilizer against viscosity loss.
KEY WORDS: autoclave; EDTA; HPMC; hydrogel; hypromellose; methionine; methylcellulose; protein;
viscosity.
INTRODUCTION
Growth factors such as broblast growth factors (acidic
and basic), epidermal growth factor, vascular endothelial
growth factor, platelet derived growth factor (Regranex
Gel), etc. have been investigated as topically applied
therapeutics for treatment of diabetic skin ulcer or wound
healing. For this kind of application, the dosage form is often
a hydrogel. These topical formulations may contain either
methylcellulose USP/NF (Methocel A4M) or hydroxypro
pylmethylcellulose USP/NF (hypromellose, Methocel E4M)
as the gelling agent. Methylcellulose and hypromellose share
the same cellulose backbone but differ in the substitution of
certain hydroxyl (HO ) groups by methoxyl (CH3O ) in the
case of methylcellulose and hydroxypropyl (CH3CH(OH)
CH2O ) plus methoxyl groups in the case of hypromellose.
In pharmaceutical preparations, cellulose derivatives are
widely used to impart viscosity to topical, ophthalmic, and
vaginal formulations. The stability of the viscosity of these
cellulose based hydrogels at various pH levels and temper
atures has been studied (1 3). In addition to acid hydrolysis
and breakdown by high temperatures, oxidative degradation
has also been suspected to cause loss of gel viscosity (4).
Viscosity loss in hydroxyethylcellulose (HEC) thickened latex
Late Stage Pharmaceutical and Processing Development, Process

Research and Development, Genentech, Inc., 1 DNA Way, South
San Francisco, California 94080, USA.
2
To whom correspondence should be addressed. (e mail: jji@gene.
com)
ABBREVIATIONS: EDTA (disodium salt), Ethylenediamine
tetraacetate; HEC, Hydroxyethylcellulose; HPC, Hydroxypropylcellulose.
paint has been attributed to chemical oxidants through

measurement of the oxidation reduction potential (5). In
addition, it has been reported that the thermal degradation
temperatures and activation energies for cellulose ethers such
as methylcellulose, ethylcellulose, sodium carboxymethylcel
lulose, HEC, and hypromellose are usually lower in air than
in nitrogen (6).
When a hydrogel is sterilized, the gel is exposed to high
temperatures, which may cause viscosity loss. Chu and Doyle
(7) reported signicant viscosity loss of the sodium carboxy
methylcellulose gel used for compounding platelet derived
growth factor (in becaplermin [Regranex Gel]) when the
autoclave tank was ushed with oxygen, whereas viscosity
was well maintained when the tank was ushed with nitrogen.
Although sterility is not claimed for Regranex, it is nonethe
less produced with autoclave sterilization to minimize bio
burden. Autoclave sterilization of cellulose derivatives is not
limited to topical products applied to open wounds. Numer
ous ophthalmic products contain cellulose derivatives to
increase viscosity, and they too are autoclave sterilized.
Recognizing the role that oxidation plays in the stability
of cellulose, Dahl and coworkers (8) studied the effect of
spiked hydrogen peroxide (H2O2) at 20 and 200 ppm on an
HEC based gel and observed a rapid reduction in gel
viscosity. They also reported that butylated hydroxyanisole
had a marginal protective effect, as an anti oxidant, on gel
viscosity.
Recently, hydroxypropylcellulose (HPC) from different
lots was found to contain substantial amounts of oxidant,
equivalent to 50 890 nmol hydroperoxide per gram of HPC,
with signicant lot to lot and manufacturer to manufacturer
variations (9). In the case of HPC, peroxide can be extracted
into the aqueous medium without an interfering gel phase,
678
Preventing Loss of Viscosity of Cellulose-Based Topical Gel
679
and the peroxide level can be determined. However, the exact

level of peroxide in methylcellulose or hypromellose cannot
be determined by the same method because once the
cellulose derivatives are placed in water, they are hydrated
and become a viscous liquid or gel that renders the assay
inoperable. Because these cellulose derivatives are made in a
similar manner, one may expect that the peroxide level in
HPC may be extended to methylcellulose or hypromellose.
Hydroperoxide at 890 nmol/g would result in 0.89 ppm of the
peroxide in a gel when 3% of the cellulosic gelling agent is
used in formulation.
There is further evidence of the presence of oxidant in a
cellulose based preparation and of its possible effect on
protein stability. Nguyen (10) reported in the early 1990s that
relaxin was oxidized much faster in methylcellulose gel than
in aqueous solution. Therefore, the concern related to the
presence of oxidant in cellulose derivatives is not just limited
to viscosity loss, but also to the degradation of any oxidation
sensitive drug substance in the formulation.
In this work, we investigated accelerated oxidative
damage to cellulose gels by spiking in various levels of
H2O2 and monitoring viscosity loss. Levels of H2O2 from 2 to
20 or 100 ppm were used to induce an oxidative reaction.
These levels are higher than those described in the afore
mentioned report on HPC (9) for example, 0.96 ppm
peroxide when made into a 3% gel. The evaluation of higher
levels, up to 20 100 ppm of H2O2, can be justied because
autoclaving can generate additional amounts of reactive
oxidizing species and because other excipients such as
polysorbate or polyethylene glycol may also introduce
additional peroxides. The study described in this report
examined the effect of EDTA on viscosity loss during
autoclave sterilization of hypromellose gel in the presence
of H2O2 and the effect of EDTA and methionine, alone or in
combination, on viscosity loss during long term storage
(6 months) in the presence of H2O2.
not. Sodium succinate (5 mM, pH 5.0) buffer was weighed

into 500 mL Kimax bottles and heated to 70C. Hypromellose
powder was added to make a 4.7% (w/w) gel. The exact
amount of powder added was determined by weight to ensure
that any variation was not due to differences in powder
transfer. The powder was dispersed into the buffer to form
slurry by swirling the bottle. One bottle of each lot of gel to
be tested was then autoclaved for 45 min at 121C and 15 psi.
After the autoclave cycle, the bottles were shaken to disperse
the precipitated hypromellose and then weighed. Buffer was
added to all bottles (autoclaved and not autoclaved) to make
3% gels (based on powder weight). The bottles were then
placed on a bottle roller at 2 8C until the appearance of the
gel was homogeneous.

Materials
Methylcellulose USP/NF (Methocel A4M) and hypro
mellose USP/NF (Methocel E4M, 4 Lots TK, UF, UI, and
TH) were obtained from Dow (Midland, MI). Hypromellose
lot OS (the fth lot) was prepared and tested by an oversea
supplier. Sodium succinate 6 hydrate and succinic acid were
purchased from JT Baker (Phillipsburg, NJ). H2O2 (30%
solution) was purchased from VWR International (West
Chester, PA) and EDTA and methionine from Sigma Aldrich
(St. Louis, MO).
Methods
Preparation of Gels
For safety, all autoclaved bottles were tted with a cap
having a ltered vent.
Gels from various lots of hypromellose. Two hydrogels
were prepared from each of four lots of hypromellose
powder. One gel of each lot was autoclaved and one was
Addition of H2O2 and EDTA to gels before autoclave

heat treatment. One stock gel for hypromellose and one for
methylcellulose were prepared. Sodium succinate (5 mM, pH
5.0) buffer was heated to 70C. Gel powder was then added
to make a 4.7% gel. The suspension was mixed with a
Heidolph RZR 2021 propeller mixer until it formed a
uniform hydrogel. The gel was then allowed to sit at room
temperature until all the bubbles had risen out of it. Ten
grams of gel was then weighed into 100 mL Kimax bottles.
Aliquots of 1,000 ppm EDTA stock solution, 3,000 ppm
peroxide stock solution, or both were added to each gel to
achieve the desired nal concentrations. The gels were
autoclaved for 20 min at 121C and 15 psi. The bottles were
shaken after the autoclave cycle to disperse the precipitated
methylcellulose or hypromellose and then placed on bottle
rollers at 2 8C. To study the effect of autoclaving, these
samples were tested without further storage.
Preparation of Gels for Storage Studies. One stock gel for
hypromellose and one for methylcellulose were prepared for
each study. Sodium succinate (5 mM, pH 5.0) buffer was
heated to 70C. Gel powder was then added to make a 4.7%
gel. The suspension was mixed with a Heidolph RZR 2021
propeller mixer until it formed a uniform hydrogel. The gels
were then transferred to Kimax bottles and autoclaved for
45 min at 121C and 15 psi. The bottles were shaken after the
autoclave cycle to disperse the precipitated methylcellulose or
hypromellose and then placed on bottle rollers at 2 8C.
Different formulations of each gel type were made by
transferring and weighing approximately 13 g of 4.7% gel into
50 mL polypropylene tubes. Buffer alone, buffer with
640 ppm EDTA, buffer with 5 mg/mL methionine (not used
prior to autoclaving), or buffer with EDTA and methionine
was added for a nal concentration of 3% gel, with 400 ppm
EDTA and 1.8 mg/mL methionine in the appropriate gels. To
these gels, 3,000 ppm stock H2O2 was added to a 0, 2, 5, 20, or
50 ppm concentration. The gels were then mixed at 2 8C on
a bottle roller. After mixing, 1 g of each gel was dispensed
into a 5 cc West glass vial with a serum stopper. The vials
were stored at 5C and 40C for designated period of time.
Viscosity measurements. Viscosity was determined using
a Paar Physica UDS 200 cone and plate rheometer at a
constant shear rate of 120 s 1 (about 20 rpm) using a 25 mm
cone, temperature controlled to 25C. A positive displacement
pipette was used to put approximately 200 L of gel on the
680
Ji, Ingham and Wang
plate; excess was removed once the cone was lowered into
measuring position. The average value of all readings at half
minute intervals over a span of 10 min was taken as the
viscosity measurement. Typically, three independent
measurements were taken for each sample. Each reported
viscosity value was the average of three measurements for each
sample. In general, without added H2O2, the viscosity of
methylcellulose and hypromellose gels was about 2,800 cP
under the measurement conditions (shear rate, 120 s 1).
Fig. 2. Percentage of viscosity remaining in the absence and presence

of EDTA after autoclaving. H2O2 at 1 ppm (white bars), 10 ppm (gray
bars), and 100 ppm (black bars) caused the gel viscosity loss. MC
methylcellulose, HPMC hypromellose
Viscosity Loss Due to the Autoclave Sterilization Process

Viscosity loss due to the autoclave sterilization process
has previously been observed with sodium carboxymethylcel
lulose (7). We chose to study uncharged cellulose derivatives
such as methylcellulose and hypromellose because of con
cerns about potential interactions between charged cellulose
derivatives and protein. To evaluate the susceptibility of
hypromellose to autoclave sterilization, four lots of domestic
hypromellose raw material (lots TK, UF, UI, and UH) and
one lot from oversea were made into 3% gel, and the gel
viscosity before and after autoclave sterilization was assessed.
All ve gels showed viscosity loss to some extent (Fig. 1), and
the gel made from oversea hypromellose lost the most
viscosity after autoclave sterilization.
Effect of EDTA on Autoclave-Induced Viscosity Loss
in the Presence of Peroxide
benet of EDTA was more prominent at higher concentra

tions of H2O2 than at lower concentrations.
Effect of Hydrogen Peroxide Concentration
and Temperature on Gel-Viscosity Loss during Storage
In this study, the effect of different levels of H2O2 on gel
viscosity during storage at 5C and 40C was studied.
Figure 3a and b shows that H2O2, at concentrations at or
above 5 ppm, caused a rapid, substantial decrease in the
viscosity of hypromellose gel. The higher the concentration of
H2O2, the greater was the loss of gel viscosity. At 5C
(Fig. 3a, where the scale is shown in weeks), the percentage of
remaining viscosity at week 8 was 70%, 45%, 40%, and 40%
for the gels containing 2, 5, 20, and 50 ppm H2 O 2 ,
respectively. The decrease at 40C was much faster than that
at 5C. At 40C (Fig. 3b, where the scale is shown in days),
An attempt to use butylated hydroxyanisole as an anti

oxidant in HEC gel to prevent viscosity loss was not
successful (8). There are many anti oxidants that can be used
to inhibit the oxidation of pharmaceuticals. However, major
ity of these anti oxidants are prone to oxidation during
autoclaving at high temperature and high pressure and thus
are unsuitable for any preparation need be autoclaved.
EDTA and certain analogues, because of their simple
chemical structure and stability at elevated temperature, are
the exceptions. For this reason, viscosity stabilization by
EDTA was evaluated. In the absence of EDTA, the loss of
viscosity increased with increasing amounts of H2O2 for both
methylcellulose and hypromellose gels (Fig. 2). As shown in
the same gure, when EDTA was added at 400 ppm to gels
containing H2O2, the viscosity loss was minimized. The
Fig. 1. Viscosity of 3% hypromellose gels before (dotted column) and

after (grid column) autoclave sterilization. Each gel was made from a
different lot of hypromellose powder (TK, UF, UI, TH, or OS)
Fig. 3. Percentage of viscosity remaining relative to time zero in 3%

hypromellose gel in the presence of various concentrations of H2O2 at
5C (a) and 40C (b). Diamonds 0 ppm H2O2, squares 2 ppm H2O2,
triangles 5 ppm H2O2, multiplication signs 20 ppm H2O2, asterisks
50 ppm H2O2
681
the percentage of remaining viscosity on day 3 compared to

that at time zero was 70%, 40%, 15%, and 10% for the gels
containing 2, 5, 20, and 50 ppm H2O2, respectively. The
decrease in viscosity at 40C reached a plateau after 3 days,
while the viscosity of the gels at 5C with 20 and 50 ppm
H2O2 continued to decrease. The viscosity loss of the gels
with 2 and 5 ppm H2O2 at 5C reached a plateau after
12 weeks. The explanation for this observation is that,
concurrent with the loss of viscosity, the H2O2 concentration
in the gel also decreased. Although peroxide was not
measured, one may expect that when the gel was spiked with
a large amount of H 2O 2, it took more time for the
concentration of H2O2 to decrease to a very low level at
5C than it did at 40C. After 8 weeks, there was still a fair
amount of H2O2 residue in the gel at 5C, and the viscosity of
the gel continued to decrease.
Figure 4a and b shows the percentage viscosity remaining
in the methylcellulose gels in the presence of different levels
of H2O2 at 5C and 40C. Phenomena very similar to the
hypromellose gels were observed. The slight increase in
viscosity observed in the sample containing 0 ppm H2O2 is
experimental error; the percent error of the viscosity mea
surement is about 10%.
Both hypromellose and methylcellulose contain a poly
meric backbone of cellulose, a natural carbohydrate that
contains a basic repeating structure of anhydroglucose units.
Gel viscosity loss is due mainly to the cleavage of the
glycosidic linkage of the cellulose ether by hydroxyl radicals
(11). One of the mechanisms proposed is that the C H bonds
are rst oxidized, in a stepwise process, to form hydroperox

ide ( C OOH). Decomposition of the hydroperoxide formed
at the acetal C H can cleave the main polymer chain and
cause a signicant decrease in gel viscosity (8,12). More
recently, using 1, 4 anhydrocellobitol as the model compound,
it was found that the hydroxyl radicals can simply cleave any
randomly encountered glycosidic linkages in cellulose and
that the main degradation mechanism is primarily substitution
reactions of hydroxyl radicals at the anomeric carbon (13).
Fig. 4. Percentage of the viscosity remaining relative to time zero in

3% methylcellulose gel in the presence of various concentrations of
H2O2 at 5C (a) and 40C (b). Diamonds 0 ppm H2O2, squares 2 ppm
H2O2, triangles 5 ppm H2O2, multiplication signs 20 ppm H2O2,
asterisks 50 ppm H2O2
Effect of EDTA and Methionine on Gel-Viscosity Loss

during Storage
Given that EDTA exhibited stabilization of viscosity loss
caused by the autoclave sterilization process (Fig. 2), it was
hypothesized that EDTA may also stabilize viscosity loss
during storage in the presence of added peroxide. In addition
to EDTA, a useful anti oxidant, methionine, was considered
since it can be added to a hydrogel after autoclaving.
Therefore, the effects of EDTA and methionine on the
stability of gel viscosity during storage were investigated.
The experiment was carried out at 40C for 4 weeks
(Fig. 5) and at 5C (Fig. 6) for 6 months. At 40C and without
added EDTA or methionine, gel viscosity decreased signi
cantly in the presence of 2, 5, 20, or 50 ppm H2O2, as
discussed previously and shown in Fig. 3. Figure 5a shows that
EDTA slowed the rate of viscosity loss (Fig. 3b). However,
EDTA alone could not completely prevent viscosity loss
during storage at 40C. When methionine was added to the
gel at a concentration of 1.8 mg/mL, it was sufcient to
prevent viscosity loss during storage at 40C even in the
presence of 50 ppm H2O2 (Fig. 5b). Figure 5c shows that
when both EDTA and methionine were present in the gel, gel
viscosity was also maintained. Because methionine alone
provided excellent protection, the results shown in Fig. 5c do
not allow the effect of adding EDTA in this case to be
discerned.
Figure 6a c shows the stability of gel viscosity during
storage at 5C in formulations containing EDTA, methionine,
or both, respectively, in the presence of different levels of
H2O2. In results similar to those obtained at 40C, methionine
at 5C effectively prevented the gel viscosity loss caused by
H2O2. However, the data in Fig. 6a indicate that EDTA alone
prevented viscosity loss at 5C. Unlike the observations at
40C (Fig. 5a), the protection at 5C was signicant when a
comparison is made between the absence of EDTA (Fig. 3a)
and the presence of EDTA (Fig. 6a). Since EDTA appears to
have a protective effect, it is possible that a free radical
reaction is occurring. Because EDTA sequesters and inacti
vates the trace metals that catalyze free radical reactions,
EDTA is considered a stabilizer against such reactions. It is
possible that at 5C the generation of free radicals is slow,
allowing EDTA the chance to exert its sequestering effect on
the catalytic metals and thus minimize the quantity of free
radicals generated. Consequently, loss of viscosity for gels
stored at 5C can be minimized by the addition of EDTA. In
contrast, at 40C, free radical generation is much faster and
sequestration of metal by EDTA cannot reduce the formation
of free radicals, and thus EDTA exhibits a minimal effect.
The use of H2O2 to stress a gel made of cellulose
derivatives is well founded because H2O2 has been found in
682
Ji, Ingham and Wang

CONCLUSIONS
Hypromellose gels made from ve lots of raw material
were found to show viscosity loss after autoclave sterilization.
Detectable viscosity loss occurred in the presence of various
levels of H2O2 (e.g., 1 100 ppm) in hypromellose and
methylcellulose gels. EDTA minimized the viscosity loss in
the presence of H2O2 during autoclaving.
During gel storage at 40C, EDTA slowed the rate of
viscosity loss but did not completely prevent it. At 5C,
EDTA showed better protection. Methionine was effective in
completely circumventing the loss of gel viscosity during
storage at 5C or 40C when the gel was stressed by 2 50 ppm
H2O2. Based on this nding, methionine may be considered
for use as a gel viscosity stabilizer in cellulose containing
formulations.

hypromellose gel in the presence of 400 ppm EDTA (a), 1.8 mg/mL
methionine (b), and both 400 ppm EDTA and 1.8 mg/mL methionine
(c) and various concentrations of H2O2 at 40C for 4 weeks.
Diamonds 0 ppm H2O2, squares 2 ppm H2O2, triangles 5 ppm H2O2,
multiplication signs 20 ppm H2O2, asterisks 50 ppm H2O2
various sources of cellulose raw material (9). The appropriate

level of peroxide to use in a stressed study is debatable. In
this study, a wide range of peroxide concentrations, from 2 to
50 ppm, was used. According to a previous report (9), 2 ppm
may be present in cellulose raw material when it is made into
a 3% gel. Since autoclaving may generate additional free
radicals and peroxide, 50 ppm may be the likely level for a gel
at the initiation of long term storage. Therefore, 2 50 ppm
could be an appropriate and realistic range for evaluation.
The results and recommendations made from the study
should lead to a robust and stable formulation.
Figures 5 and 6 show methionine to be an effective
stabilizer against viscosity loss of a hypromellose gel stressed
by H2O2. EDTA exhibits a marginal stabilization effect under
storage conditions. Based on these results, methionine is a
recommended component for a cellulose containing formula
tion when stored after autoclave sterilization.

hypromellose gel in the presence of (a) 400 ppm EDTA, (b) 1.8 mg/
mL methionine, and (c) both 400 ppm EDTA and 1.8 mg/mL
methionine and various concentrations of H2O2 at 5C for 6 months.
Diamonds 0 ppm H2O2, squares 2 ppm H2O2, triangles 5 ppm H2O2,
multiplication signs 20 ppm H2O2, asterisks 50 ppm H2O2

ACKNOWLEDGMENTS
The authors wish to thank Dr. Sherry Martin Moe for
her encouragement of this work and Ms. Linda Khym for
editing the manuscript.
REFERENCES
1. Huikari A. Effect of heat sterilization on the viscosity of
methylcellulose solutions. Acta Pharm Fennica. 1986;95(1):9 17.
2. Huikari A, Karlsson A. Viscosity stability of methylcellulose
solutions at different pH and temperature. Acta Pharm Fennica.
1989;98:231 8.
3. Ferdous AJ. Viscosity and stability studies of hydroxylpropyl
methylcellulose polymer solutions. Pak J Pharm Sci. 1992;5:115 9.
4. Ha AN, Martin CF, Firestone BA. Viscosity degradation effects of
aqueous carbomer dispersions during heat sterilization and
oxygen exposure. Pharm Res. 1999;S1(4):2867.
5. Vaughn L, Irwin M, Williams M. Viscosity loss in hydroxyethyl
cellulose thickened latex paints caused by chemical oxidants
methods of detection. J Coatings Technol. 1980;52:71 4.
6. Li XG, Huang MR, Bai H. Thermal decomposition of cellulose
ethers. J. Appl Polym Sci. 1999;73:2927 36.
683
7. Chu PI, Doyle D. Development and evaluation of a laboratory
scale apparatus to simulate the scale up of a sterile semisolid and
effects of manufacturing parameters on product viscosity. Pharm
Dev Technol. 1999;4:553 9.
8. Dahl T, He G, Samuels G. Effect of hydrogen peroxide on the
viscosity of a hydroxyethylcellulose based gel. Pharm Res. 1998;
15:1137 40.
9. Wasylaschuk WR, Harmon PA, Wagner G, Harman AB,
Templeton AC, Xu H, et al. Evaluation of hydroperoxides in
common pharmaceutical excipients. J Pharm Sci. 2007;96:106 16.
10. Nguyen TH. Oxidation degradation of protein pharmaceuticals.
In: Cleland JL, Langer R, editors. Formulation and delivery of
proteins and peptides. ACS Symposium no. 567. Washington,
DC: American Chemical Society; 1994. p. 59 71.
11. McBurney LF. Degradation of cellulose. In: Ott E, Bikales NM,
Segal L, editors. Cellulose and cellulose derivatives: part I. New
York: Wiley Interscience; 1954. p. 99.
12. Kuramshin EM, Kulak LG, Nazarov MN, Zlotsky SS, Rakh
mankulov DL. Oxidation of cyclic acetals as a preparative
method of diol monoester production. J Prakt Chem. 1989;331:
591 9.
13. Guay DF, Cole BJW, Fort RC Jr, Hausman MC, Genco J.
Mechanisms of oxidative degradation of carbohydrates during
oxygen delignication. III. Reaction of photochemically gener
ated hydroxyl radicals with 1,5 anhydrocellulobitol and cellulose.
J Pulp Paper Sci. 2002;28(7):217 26.

DOI: 10.1208/s12249 009 9256 8
Research Article
Fabrication of Modified Transport Fluconazole Transdermal Spray Containing
Ethyl Cellulose and Eudragit RS100 as Film Formers
Mukesh C. Gohel1,2 and Stavan A. Nagori1
Abstract. The present investigation was undertaken to fabricate modied transport uconazole
transdermal spray using ethyl cellulose and Eudragit RS100 as lm forming polymers. Eudragit
RS100 (X1) and ethyl cellulose (X2) were selected as independent variables in 32 full factorial design,
whereas drug transport in rst hour (Y1) and the time required for 50% drug transport (Y2) were selected
as dependent variables. Eutectic blend of camphor and menthol was used as permeation enhancer cum
solvent for lm forming polymers. The pH, viscosity, volume of solution delivered upon each actuation,
spray angle, ex in vivo physical evaluation and in vitro drug transport of the formulated products were
evaluated. The optimized batch B16 containing 5.25% w/w ethyl cellulose and 10.6% w/w Eudragit
RS100 was formulated by overlapping the contour plots of Y1 and Y2. The pH, viscosity, volume of solution
sprayed upon each actuation and spray angle of the batch B16 was 6.3, 52.9 cPs, 0.24 ml and 82.6
respectively. The lm of optimized batch was exible and dermal adhesive. The responses Y1 and Y2 of
batch B16 were 7.91 g/ml and 347 min respectively. The kinetics of drug transport was best explained by
the Korsmeyer and Peppas model. The eutectic mixture consisting of equal parts of camphor and menthol
showed improved drug permeation through shed snake skin. Short term stability study demonstrated
insignicant changes in performance characteristics.
KEYWORDS: uconazole; modied transport; transdermal spray; factorial design and short term
stability study.
INTRODUCTION
Tinea infections are caused by three species of fungi
collectively known as dermatophytes. Tinea infections are
named considering the affected area of human body, i.e.,
Tinea corporis (general skin), Tinea cruris (groin), and
Tinea pedis (feet). Topical therapy is generally successful
unless the infection covers an extensive area or is resistant
to initial therapy. In these cases, systemic therapy may be
required.
Fluconazole, a novel triazole antifungal drug, is used in
the treatment of supercial and systemic fungal (Tinea)
Department of Pharmaceutics and Pharmaceutical Technology, L. M.

College of Pharmacy, P.O. Box 4011, Navrangpura, Ahmedabad,
Gujarat 380 009, India.
2
To whom correspondence should be addressed. (e mail: mukeshgohel@
hotmail.com)
ABBREVIATIONS: f2, similarity factor; MIC, minimum inhibitory
concentration (g/ml); PEG 400, polyethylene glycol; X1, concentration
of Eudragit RS100 (% w/w); X2, concentration of ethyl cellulose (%
w/w); Y1, the amount of drug transported (g) per milliliter at the end
of rst hour; Y2, the time (min) required to transport 50% of the drug
(t50%); US FDA, US Department of Health and Human Services Food
and Drug Administration.
infection. The drug has moderate (8 mg/ml) solubility in

water (1,2). Like other imidazole and triazole class antifun
gals, uconazole inhibits the fungal cytochrome P450 enzyme
14 demethylase (3). Fluconazole is primarily fungistatic,
however may be fungicidal against certain organisms in a
dose dependent manner. Ayub et al. investigated in vitro skin
penetration and permeation of uconazole from emulsions
containing propylene glycol and isopropyl myristate as
penetration enhancers (4). El Laithy and El Shaboury
evaluated inuence of vehicle on the release and perme
ation of uconazole dissolved in jojoba oil (5). Rivera et al.
prepared uconazole loaded poly(D,L lactic co glycolic) acid
microspheres by spray drying process (6). Kavitha et al.
formulated and characterized topical drug delivery systems
of uconazole in form of ointment, cream and gel using
water soluble and water insoluble base in absence of
penetration enhancer (7).
The aim of the present research work was to develop
patient friendly modied transport transdermal spray of
uconazole using ethyl cellulose and Eudragit RS100 as
lm forming polymers. A 32 full factorial design was
employed for optimization. The formulation contained
eutectic mixture of menthol and camphor (dermal
penetration enhancers) that readily partition into the
stratum corneum (8 14). Eutectic mixture of camphor and
menthol also acted as a solvent for lm forming polymers,
684
Fluconazole Transdermal Spray
685
imparted cooling effect to the skin and possessed antifungal

activity (15,16).
weight of undissolved solid was recorded. The ltrate was

carefully observed for clarity.
Preliminary Studies
Materials
Film formers (ethyl cellulose and Eudragit RS100) and

plasticizer (polyethylene glycol) were sequentially dissolved
in the eutectic mixture consisting of equal proportion of
camphor and menthol. Fluconazole was separately dissolved
in vehicle blend consisting of 80 parts of alcohol (80% v/v)
and 20 parts of acetone. The solution of ethyl cellulose/
Eudragit RS100 in eutectic mixture was gradually added to
the solution of uconazole and mixed for 15 min at 80
100 rpm. The resulting solution was lled in a rellable
container containing plastic dip tube of 75 mm length and
1.2 mm internal diameter. The aperture size of the tube was
0.3 mm. Table I displays the composition of the formulated
batches (B1 B6). The sprays were analyzed for pH, viscosity,
volume of solution delivered upon each actuation, spray
angle, ex in vivo physical characteristics and in vitro drug
transport. The results of drug transport are expressed in
Fig. 1.
Fluconazole USP and ethyl cellulose (Ethocel standard

10FP premium) were received as gift samples from Zydus
Cadila (Ahmedabad, India). Eudragit RS100 was received
as a gift sample from Degussa Pvt. Ltd. (Mumbai, India).
Camphor and menthol were purchased from Gem Corpora
tion (Ahmedabad, India) and Shreeji Pharma International
(Ahmedabad, India), respectively. Polyethylene glycol (PEG
400) and acetone were purchased from S. D. Fine Chemical
Pvt. Ltd. (Ahmedabad, India). Alcohol I.P was purchased
from Baroda Chemicals Industries Ltd. (Baroda, India). The
other chemicals and reagents were of analytical grade.
Method
Determination of Solubility of Fluconazole and Polymer
in Various Solvents
Factorial Design
The solubility of uconazole was determined in a
mixture containing 80 parts of alcohol (80% v/v) and 20 parts
of acetone. The drug solubility was also determined in a
eutectic mixture of camphor and menthol (1:1). The solubility
of ethyl cellulose and Eudragit RS100 was determined in
the eutectic mixture. An excess amount of sample was added
to 35 ml of solvent and stirred at 100 rpm on a magnetic
stirrer (Remi Electronics, Ahmedabad, India) for 30 min at
room temperature (352C) in a closed vessel. The mixtures
were then ltered through 0.22 m millipore lters and the
A 32 full factorial design was used for optimization of

formulated product. The concentration of Eudragit RS100
(X1) and ethyl cellulose (X2) were selected as independent
variables, whereas the amount of drug transported in 1 h/ml
(Y1) and the time required to transport 50% of the drug
(t50%, Y2) were selected as dependent variables. Tables II and
III show the composition, design layout for the optimization
study and the responses. The method for preparation and
evaluation of the formulated batches (B7 B16) was similar to
Table I. Composition and Evaluation of Fluconazole Sprays

Batch code
Ingredient (% w/w)
Fluconazole
Ethyl cellulose
Eudragit RS100
PEG 400
Eutectic blend
Alcohol and acetone blend (q.s.)
Tests
Viscosity (a2.7 cPs)
Volume of solution delivered upon each actuation (b0.03 ml)
Spray angle (c0.4)
Y1a
Y2b
Ex in vivo lm formation timec (d35 sec)
Feeling of warmth and subsequent cooling sensationc (e1 min)
Appearance of the lmc
Dermal adhesion and exibility of the lmc
Water washabilityc
a
b
c
B1
0.5
3
B2
0.5
5
0.25
0.25
10
10
100
100
Average results (n 3)
15.0
26.4
0.32
0.30
78.0
79.1
21.07
17.07
94
115
130
160
12
14
+
+
++
++
++
++
Y1 is the amount of drug transported (g) per ml at the end of rst hour
Y2 is the time (min) required to transport 50% of the drug (t50%)
Measured for placebo batches replacing uconazole with blend of alcohol and acetone
B3
0.5
7.5
B4
B5
B6
0.5
0.5
0.5
0.25
10
100
5
0.25
10
100
10
0.25
10
100
15
0.25
10
100
43.5
0.26
81.0
12.55
196
210
15
++
++
++
7.6
0.39
76.1
23.44
120
110
12
+
++
+++
20.2
0.31
78.3
14.12
160
152
14
+
+++
+++
51.6
0.24
82.2
11.73
210
230
16
++
+++
++
686
Gohel and Nagori

the initial weight of the formulation before actuation, and Dn
is the density of the formulation.
3. Spray angle
The method of impingement of spray on a piece of paper
was used for the study. Sudan red (10 mg) was dissolved in
formulation to facilitate visualization. The sprays were
actuated in horizontal direction onto a white paper mounted
at a distance of 15 cm from the nozzle. The radius of the
circle, formed on the paper, was recorded in triplicate from
different directions. Spray angle () was calculated by eq. 2.
Fig. 1. In vitro drug transport from batches B1 B6 through nylon mem

), B5 (
), B6 ( )
brane; B1 ( ), B2 ( ), B3 ( ), B4 (
Spray angle tan1 l=r
where l is the distance of paper from the nozzle, and r is the

average radius of the circle.
that described under preliminary studies. The results are
expressed in Fig. 2 and Tables II and III.
Evaluations
1. Viscosity
The viscosity of the solutions was measured at 251C
using Brookeld viscometer (digital viscometer model DV II+,
Stoughton, MA, USA). The ULA spindle was rotated at
1 rpm.
2. Volume of solution delivered upon each actuation
The volume of solution delivered upon each actuation
was calculated using eq. 1.
AL Wt
Wo =Dn
where AL is the volume of solution delivered upon each

actuation, Wt is weight of formulation after actuation, Wo is
4. Ex in vivo physical evaluation

Placebo batches (B1P B16P) were actuated on the left
hand palms of three healthy human volunteers of 21 27 years
in age, four times every 10 s from a distance of 15 cm. The
purpose of the study was fully explained and the volunteers
gave written consent. The departmental review board ap
proved the study. The time required for lm formation,
appearance of the lm, dermal adhesion and exibility of the
lm, feeling of warmth and subsequent cooling sensation,
irritation potential and water washability were recorded. The
appearance of the lm was graded as shiny and transparent (+)
or shiny and translucent (++) or dull and opaque (+++). Dermal
adhesion, exibility and water washability of the lm were
graded as poor (+), moderate (++) or good (+++). To check
dermal adhesion of the lm, the palms of each volunteer were
rotated for 10 min in anti clock wise direction with occasional
opening and closing of palm, after 8 min of actuation of the
spray. During the study (720 min), the nature of the lm was
carefully evaluated for any fracture, separation or removal.
After 720 min, water washability of the lm was checked.
Table II. Composition and Evaluation of Fluconazole Sprays Based on 32 Full Factorial Design
Batch Code
Ingredient (% w/w)
Fluconazole
Ethyl cellulose
Eudragit RS100
PEG 400
Eutectic blend
Alcohol and acetone blend (q.s.)
Tests
Volume of solution delivered upon each actuation (b0.03 ml)
Spray angle (c0.4o)
Ex in vivo lm formation timea (d35 sec)
Feeling of warmth and subsequent cooling sensationa (e1 min)
Appearance of the lma
Dermal adhesion and exibility of the lma
Water washabilitya
a
B7
B8
0.5
0.5
7.5
5
15
15
0.25
0.25
10
10
100
100
Average results
128.3
92.8
0.14
0.18
91.5
88.2
532
435
19
18
+++
++
+++
+++
+
+
B9
0.5
3
15
0.25
10
100
(n 3)
70.8
0.2
83.8
390
17
++
+++
++
Measured for placebo batches replacing uconazole with blend of alcohol and acetone
B10
B11
B12
B13
B14
B15
B16
0.5
7.5
10
0.25
10
100
0.5
5
10
0.25
10
100
0.5
3
10
0.25
10
100
0.5
7.5
5
0.25
10
100
0.5
5
5
0.25
10
100
0.5
3
5
0.25
10
100
0.5
5.25
10.6
0.25
10
100
69.7
0.21
84.2
372
18
++
+++
++
48.9
0.26
81.8
328
16
++
+++
++
38.1
0.31
80.6
287
15
+
+++
++
52.5
0.26
82.6
340
17
++
+++
++
34.0
0.3
79.8
282
13
+
++
++
23.7
0.33
78.7
249
13
+
++
++
52.9
0.24
82.6
335
17
++
+++
++
687
from the local snake habitat (Sundarvan, Ahmedabad, India).

The composition of batches B16 and B17 was similar except
Real Values Transformed Values Dependent Variables that batch B17 was formulated without using the eutectic
mixture.
Table III. Design Layout for 32 Factorial Design
Batch Code
X1
X2
B7
B8
B9
B10
B11
B12
B13
B14
B15
B16*
15
15
15
10
10
10
5
5
5
10.6
7.5
5
3
7.5
5
3
7.5
5
3
5.25
X1
X2
Y1
Y2
1
1
1
0
0
0
1
1
1
0.12
1
0
1
1
0
1
1
0
1
0.1
4.53
6.15
10.7
5.73
8.49
12.85
10.56
16.15
21.7
7.91
660
540
292
466
320
165
240
165
107
347
B16* is the point batch, X1 is the concentration of Eudragit RS100 (%

w/w), X2 is the concentration of ethyl cellulose (% w/w), Y1 is the
amount of drug transported (g) per milliliter at the end of rst hour, Y2
is the time (min) required to transport 50% of the drug (t50%)
5. In vitro drug transport

A nylon membrane with pore rating equal to 0.22 m
was mounted in a Franz cell. The surface area of membrane
available for drug transport, was 4.9 cm2. One milliliter of the
drug formulation and 88 ml of solution (pH 7.4, 100 rpm)
containing 0.9% w/v sodium chloride and 1% w/v sodium
lauryl sulfate were lled in donor and receptor compartments,
respectively. Throughout the experiment, the temperature of
media present in the receptor compartment was maintained
at 322C using a hot water jacket (372C). Sodium lauryl
sulfate was used to provide sink condition. Aliquots of 10 ml
samples were withdrawn at different time intervals from the
receptor compartment, and uconazole was estimated
spectrophotometrically at 264 nm (5). Equal amount of
fresh dissolution medium was replaced after each
withdrawal. Figures 1 and 2 show the drug transport
proles. The amount of drug transported in 1 h/ml (Y1) and
the time required to transport 50% of the drug (Y2) were
found. The UV spectrum of uconazole was observed for
uconazole formulation excipient interaction. The in vitro
drug transport from the optimized batch B16 and batch B17
was also determined using shed snake skin as a membrane in
Franz diffusion cell (Fig. 4). The shed snake skin was procured
Fig. 2. In vitro drug transport from batches B7 B16 A; B7 ( ), B8

), B11 (
), B12 ( ), B13 ( ),
( ), B9 ( ), B10 (
), B16 ( ), B16 A ( )
B14 ( ), B15 (
Kinetics of Drug Transport

The method of Bamba et al. was adopted to ascertain
kinetics of drug transport from the formulated batches B1
B16 (17). In vitro drug transport data of batches B1 B16 were
analyzed by zero order, rst order, Higuchi, Hixson Crowell,
Korsmeyer Peppas and Weibull models (18 23). A FOR
TRAN software, developed in house, was used. The least
value of sum of square of residuals (SSR) and Fishers ratio
(F) were used to select the most appropriate kinetic model.
Similarity Factor
The similarity factor (f2) was calculated by comparing
test and reference in vitro drug transport prole using eq. 3
(24).
9
8"
#0:5
=
<
n
1 X
3
Rt Tt 2
100
f2 50 log 1
;
:
n n 1
where n is the number of pull points, Rt and Tt are percentage
drug transported from reference and test products, respec
tively, at time t.
Criteria for Optimized Batch
The lm formation time was arbitrarily xed as less than
8 min. The lm should be mucoadhesive and exible in
nature to allow day to day activity yet should be water
washable. Viscosity of the formulations should be less than
80 cPs. The selected limits for drug transport were: (1) Y1:
Fig. 3. Overlapped contour plots; dotted line Y1, dashes and dots Y2
688
Fig. 4. In vitro drug transport from batches B16 and B17 through
shed snake skin; B16 (closed diamonds), B17 (open squares )
amount of drug transported in 1 h/ml should be equal to the

minimum inhibitory concentration of uconazole (minimum
inhibitory concentration (MIC)=8 g/ml)5%, and (2) Y2:
time required to transport 50% of the drug (t50%) should be
equal to 360 min5%.
Unpaired t test with equal variance was used to nd any
statistically signicant difference in the in vitro drug transport
prole between batches B16 and B17 through shed snake
skin at 5% level of signicance.
Stability Study
Optimized formulation (batch B16) was stored for 2 months
at 252C away from light. At the end of second month, the
formulation was subjected to various tests like volume of
solution delivered upon each actuation, pH, viscosity, solution
delivered upon each actuation, spray angle, ex in vivo physical
characteristics and in vitro drug transport. The procedure
employed for the study was identical to that described above.
The results are shown in Table IV.
A eutectic mixture is a mixture of two or more solids
which has lower melting temperature than any of its
constituents. Various substances such as ibuprofen, menthol,
chloral hydrate, beta naphthol, lidocaine, and prilocaine form
eutectic mixtures. The primary criterion for eutectic forma
tion is the mutual solubility of the components in the liquid.
Camphor and menthol also forms a hydrophobic eutectic
mixture. Camphor and menthol are powerful penetration
enhancers (8 14). Camphor and menthol cause leaching of
the lipids present in the skin and thus cause subsequent pore
formation (25). In the present study, attempt was made to use
eutectic mixture of camphor and menthol as multifunctional
excipient (solvent for the lm formers, powerful permeation
enhancer and antifungal agent).
Preliminary Batches Containing Ethyl Cellulose
The solubility of uconazole in a mixture consisting of 80
parts of alcohol (80% v/v) and 20 parts of acetone and the 1:1
Gohel and Nagori

eutectic mixture of camphor and menthol was 150 and less
than 2 mg/ml, respectively. Acetone and primary alcohols are
preferred solvents in concentration of 40 65% in formulation
of topical solutions (26). As per US Department of Health
and Human Services Food and Drug Administration (US
FDA) inactive ingredient guide, limit of alcohol and acetone
in topical solutions is more than 83% and 12%, respectively
(27). Acetone was used along with alcohol in the vehicle
blend to facilitate faster lm formation (<8 min) on skin.
Ethyl cellulose, a lm former, was soluble (>100 mg/ml) in the
eutectic mixture. Hence, uconazole was dissolved in vehicle
blend of alcohol and acetone, whereas ethyl cellulose was
dissolved in eutectic mixture of camphor and menthol.
The pH of formulated batches (B1 B3) ranged from 5 to
7. The pH of human skin is in between pH 5.5 and 6.5 (28).
Hence, the pH adjustment was unnecessary. Table I shows
that the viscosity of the formulated batches containing ethyl
cellulose as lm former (B1 B3) ranged from 15 to 43.5 cPs.
The viscosity of the formulations increased with ethyl
cellulose concentration. Formulated batches (B1 B3) showed
good sprayability. The volume of solution delivered upon
each actuation and spray angle ranged from 0.26 to 0.32 ml
and 78 to 81, respectively. The stated parameters were
correlated with polymer concentration and viscosity of the
formulation. Ex in vivo lm formation of placebo batches
(B1P B3P) passed the desired criteria of lm formation,
which varied from 130 to 210 s after actuation. The lm
turned from transparent to translucent on increase in
viscosity. Dermal adhesion and exibility of lms were
moderate. Feeling of warmth and subsequent cooling sensa
tion were perceived after application of spray (around
12 min) because of camphor and menthol in the formulation.
None of the placebo formulations resulted in irritation,
rashes and itching in any of the volunteers. Water washability
of the placebo batches (B1P B3P) was moderate.
The formulation excipients did not show absorbance at
264 nm. The UV spectrum remained unchanged during in
vitro drug transport study, indicating stability of uconazole
during study. The excipients or membrane constituents did
not show absorbance at 264 nm. The MIC of uconazole
against Candida species is 8 g/ml (29). The problem of dose
Table IV. Results of 2 Months Stability Study of Optimized Batch

B16
Tests
Volume of solution delivered upon
each actuation (b0.06 ml)
Spray angle (c0.4)
Ex in vivo lm formation timea (d30 s)
Feeling of warmth and subsequent cooling
sensationa (e0.5 min)
Appearance of the lma
Dermal adhesion and exibility of the lma
Water washabilitya
a
Average results
(n 3)
54.7
0.24
83.4
341
16
++
+++
++
Measured for placebo batches replacing uconazole with blend of

alcohol and acetone

dumping and lack of sustained drug transport was seen with
formulated batches (B1 B3) with values of Y1 greater than
12 g/ml and Y2 less than 200 min (Table I and Fig. 1).
Considering the results of viscosity, the concentration of ethyl
cellulose was not increased beyond 7.5% w/w.
Preliminary Batches Containing Eudragit RS100
Eudragit RS100 was soluble (>120 mg/ml) in the
eutectic mixture. Hence, Eudragit RS100 was dissolved in
eutectic mixture of camphor and menthol. The pH of
formulated batches (B4 B6) ranged from 5 to 7. Table I
shows that the viscosity of batches B4 B6 increased with
polymer concentration and ranged from 7.6 to 51.6 cPs. The
formulated batches containing Eudragit RS100 as a lm
former (B4 B6) showed good sprayability. The volume of
solution delivered upon each actuation and spray angle
ranged from 0.24 to 0.39 ml and 76 to 82, respectively.
The polymer concentration and viscosity were found to affect
the stated two parameters. Ex in vivo lm formation of
placebo batches (B4P B6P) passed the desired criteria of lm
formation. The appearance of lm changed from shiny
transparent to translucent on increase in viscosity. Dermal
adhesion and exibility of lms were good except batch B4P
containing lower concentration of Eudragit RS100. Eudra
git RS100 is a widely used polymer in formulation of
mucoadhesive dosage forms (30). Feeling of warmth and
subsequent cooling sensation were perceived after application
of spray (around 12 min). None of the placebo formulations
resulted in irritation, rashes and itching in any of the
volunteers. Water washability of the placebo batches (B4P
B6P) was good except batch B6P containing higher concen
tration of Eudragit RS100. Concentration of Eudragit
RS100 was not increased beyond 15% w/w, considering the
results of water washability. The formulated batches contain
ing Eudragit RS100 as a lm former (B4 B6) failed to meet
the desired in vitro drug transport criteria with values of Y1
greater than 11.75 g/ml and Y2 less than 215 min (Fig. 1).
Considering the results of preliminary study, a combination of
ethyl cellulose and Eudragit RS100 was tried to achieve
desired in vitro drug transport.
Factorial Design
A two factor, three level full factorial design was used
for formula optimization. The levels of independent variables
were determined from the results of preliminary batches. For
Eudragit RS100, the low, medium and high levels were 5%,
10% and 15%, respectively. Ethyl cellulose was used at low
(3%), medium (5%) and high (7.5%) levels. The pH of
formulated batches (B7 B15) ranged from 5 to 7. Table II
shows that viscosity, volume of solution delivered upon each
actuation and spray angle of formulated batches (B7 B15)
ranged from 23.7 to 128.3 cPs, 0.14 to 0.33 ml and 78.7 to
91.5, respectively. Viscosity of the liquids increased with total
increase in polymeric concentration. All placebo batches
(B7P B15P) passed ex in vivo lm formation criteria
(<8 min) except batch B7P containing high level of Eudragit
RS100 and ethyl cellulose. The appearance of lm changed
from shiny transparent to dull and opaque on increase in
viscosity. Dermal adhesion and exibility of lms were good
689
except batches B14P and B15P. Feeling of warmth and
subsequent cooling sensation were perceived after application
of spray (around 13 min). None of the placebo formulations
resulted in irritation, rashes and itching in any of the volunteers.
Water washability of the placebo batches (B7P B15P) was
moderate except batches B7P and B8P, which showed poor water
washability because of presence of higher level of Eudragit
RS100 along with moderate to low level of ethyl cellulose.
The values of Y1 (amount of drug transported at the end of
rst hour) and Y2 (time required to transport 50% of the drug)
for formulated batches (B7 B15) varied from 4.53 to 21.7 g/ml
and 107 to 660 min, respectively (Table III and Fig. 2). Hence, it
can be concluded that selected independent variables exhibit
signicant inuence on the values of Y1 and Y2. Further
optimization was carried out by evolving mathematical models
using linear regression analysis. Equation 4 shows the relation
ship between the amount of drug transported at the end of rst
hour (Y1) and the independent variables.
Y1 9:02
4:5X1
4:07X2 1:24X1 X2 2:6X12
(Multiple R=0.994, p<0.05)

Equation 5 shows the relationship between the time
required to transport 50% of the drug (Y 2) and the
independent variables.
Y2 328:33 163:33X1 133:66X2 58:75X1 X2
(Multiple R=0.992, p<0.05)

Contour plots of Y1 and Y2 were overlapped to locate
the region of acceptability. The critical observation of the
contour plot (Fig. 3) reveals that if X1 and X2 is varied from
0.5 to 0.12 and 0.1 to 1, respectively, the responses Y1 and Y2
are close to the target values of 8 g/ml and 360 min,
respectively. This limit may be considered for ne tuning of
the formulation. Three formulations (A: X1 = 0.1; X2 =0.4, B:
X1 = 0.45; X2 = 1; and C = batch B16: X1 = 0.12; X2 = 0.1)
theoretically satisfy the values of Y1 and Y2. A check point
batch (B16) was prepared. The experimental values of Y1 and
Y2 for batch B16 were 7.91 g/ml and 347 min, respectively
(Table III). Batch B16 showed good sprayability. Ex in vivo
lm formation of batch B16 passed the desired criteria of lm
formation (<8 min). The appearance of the lm was shiny
translucent. Dermal adhesion and exibility of the lm was
good, whereas water washability was moderate. Thus, looking
to overall results of sprayability, water washability and in vitro
drug transport, batch B16 was ranked as one of the optimized
batches and taken for further study.
A spray formulation (batch B16 A, Fig. 2) was prepared
to evaluate the effect of acetone on in vitro drug transport by
replacing acetone with alcohol (80% v/v ethanol) in batch
B16. The Y1 and Y2 of batch B16 A was 16 g/ml and
190 min, respectively. The probable reason for faster drug
transport is delayed ex in vivo lm formation (720 s). The
similarity factor (f2 =38.5) calculated using in vitro drug
transport data of batches B16 and B16 A as reference and
test prole, respectively, indicated signicant difference in
drug transport between both the batches (24). It is therefore
concluded that acetone is indispensable in getting desired
drug transport prole.
690
Gohel and Nagori
The in vitro drug transport data were analyzed for

determining kinetics of drug transport. Model tting was
done using an in house computer program, developed by the
authors. Zero order, rst order, Higuchi, Hixson Crowell,
Korsmeyer Peppas and Weibull models were tested. Kors
meyer Peppas model showed the least of SSR and therefore,
it was used for further data analysis. The slope and intercept
values were 0.706, 0.799, 0.737 and 2.031, 2.413, 2.161,
respectively for the formulations B9, B10, and B11. Equa
tions 6 and 7 for slope and intercept, respectively, were
evolved using linear regression analysis taking slope or
intercept as dependent variable and ratio of X1 and X2 in
the formulation as independent variable. The values of
multiple R were greater than 0.85 in both the equations
indicating good correlation. The unied equation (eq. 8) was
evolved by combining the equations of slope and intercept.
Equation of slope 0:81
Equation of intercept
0:02 ratio of X1 and X2
2:43 0:08 ratio of X1 and X2

7
log f equation of slope log t equation of intercept

8
Where f and t are fraction of drug transported and time,
respectively. Model validation was done by comparing the
values of observed and predicted drug transport for three
formulations (batches B9 B11) as well as for a check point
batch B16 containing 10.6% Eudragit RS100 and 5.25%
ethyl cellulose. Similarity factor (f2) was calculated consider
ing calculated and observed drug transport data as reference
and test prole, respectively (24). Batches B9 B11 and B16
showed insignicant difference in observed and predicted
drug transport with similarity factor (f2) greater than 64 in
each case.
The effect of eutectic mixture on drug permeation was
checked using shed snake skin as a biological membrane
(31,32) in Franz diffusion cell. The UV spectrum remained
unchanged during in vitro drug transport study, indicating
stability of uconazole during the analysis. Figure 4 shows
that the amount of drug transported and present in the
receptor compartment per milliliter at the end of rst hour,
Y1 was 8.18 and 2.78 g/ml for batches B16 and B17,
respectively. The Y2 values were 367 and >720 min, respec
tively. Incomplete drug transport (less than 45%) was seen
from the batch B17 formulated without using eutectic
mixture. The probable reason for this difference could be
presence of camphor and menthol in batch B16. It is reported
that camphor and menthol works as penetration enhancer (8).
Camphor and menthol causes leaching of the lipids present in
the skin and thus causes pore formation (24). The difference
in drug transport between batches B16 and B17 through shed
snake skin was found to be signicant at 5% level with
tcalculated 2:7 > tcritical twotail 2:05 . Hence, it can be con
cluded that eutectic mixture of camphor and menthol
signicantly improved the drug permeation. Batch B16
showed similar in vitro drug transport through nylon (refer
ence prole) and biological membrane (test prole) with

similarity factor (f2) of 86. Probable reason for this behavior
could be high concentration (10%) of eutectic mixture in the
formulation.
Stability Study
Short term stability study of the optimized batch B16 was
carried out for 2 months at 252C. Table IV shows that pH,
viscosity, volume of solution delivered upon actuation, spray
angle and ex in vivo physical characteristics of optimized
batch B16 remained unchanged during the study. Unpaired t
test with equal variance indicated insignicant difference in
the in vitro drug transport from the optimized batch B16 at p=
5% with tcalculated 0:07 < tcritical onetail 1:71 . The amount of
drug transported at the end of 1 h (Y1) and time required to
transport 50% of the drug (Y2, t50%) of optimized batch (B16) at
the end of stability study were 7.78 g/ml and 364 min,
respectively.
CONCLUSION
Modied transport transdermal spray of uconazole was
developed using 32 full factorial design. The optimized batch
B16 containing 10.6% Eudragit RS100 and 5.25% ethyl
cellulose yielded mucoadhesive and exible lm on the
human skin. The in vitro drug transport at the end of rst
hour was equal to the minimum inhibitory concentration (8
0.4 g/ml) with t50% of (36018 min). The eutectic mixture
showed increased penetration of the drug through shed snake
skin. Batch B16 passed short term stability study, carried out
at 252C, with no change in performance characteristics of
the product.
ACKNOWLEDGMENTS
We are grateful to Indian Council of Medical Research
(ICMR) for providing senior research fellowship (SRF) to
one of the authors for this project. We are thankful to Zydus
Cadila (India) and Degussa Pvt. Ltd. (India) for kindly
providing the gift samples.
REFERENCES
1. Susan B. Merck index. 12th ed. NJ: Merck & Co. Inc.; 1996.
2. United States Pharmacopoeia, XXIX, NF XIV, The United States
Pharmacopoeial Convention Inc.; 2006. pp. 911 3.
3. Laura K, Stephanie DT, Christopher JP, Charles LC, David MS.
Assessing pregnancy risks of azole antifungals using a high
throughput aromatase inhibition assay. Endocr Res. 2002;28:129
40.
4. Ayub AC, Gomes AD, Lima M, Vianna Soares CD, Ferreira LA.
Topical delivery of uconazole: In vitro skin penetration and
permeation using emulsions as dosage forms. Drug Dev Ind
Pharm. 2007;33:273 80.
5. El laithy HM, El Shaboury KMF. The development of cutina
lipogels and gel microemulsion for topical administration of
uconazole. AAPSPharmSciTech. 2002;3:77 85.
6. Rivera PA, Martinez Oharriz MC, Rubio M, Irache JM, Espuelas
S. Fluconazole encapsulation in PLGA microspheres by spray
drying. J Microencapsulation. 2004;21:203 11.

7. Kavitha K, Sivaramkrishnan M, Nalini CN, Nappinnai M.
Formulation and evaluation of topical drug delivery system of
uconazole. Indian Drugs. 2003;40:720 3.
8. Khan M. Bioavailability of vitamin B12 using a small volume
nebulizer opthalmic drug delivery system. Clin Exp Ophthalmol.
2005;33:402 7.
9. Ho H, Chen L, Lin H, Sheu M. Penetration enhancement by
menthol combined with a solubilization effect in a mixed solvent
system. J Control Release. 1998;51:301 11.
10. Flashner Barak M, Lerner I, Rosenberger V, Moldavski N,
inventors. Menthol solutions of drugs. Patent WO/2004/073686.
September 2, 2004.
11. Aqil M, Ahad A, Sultana Y, Ali A. Status of terpenes as skin
penetration enhancers. Drug Discov Today. 2007;12:1061 7.
12. Rhee Y, Choi J, Park E, Chi S. Transdermal delivery of
ketoprofen using microemulsions. Int J Pharm. 2001;228:161 70.
13. Jain R, Aqil M, Ahad A, Aliand A, Khar RK. Basil oil is a
promising skin penetration enhancer for transdermal delivery of
labetolol hydrochloride. Drug Dev Ind Pharm. 2008;34:384 9.
14. Williams AC, Barry BW. Terpenes and the lipid protein partitioning
theory of skin penetration enhancement. Pharm Res. 1991;8:17 24.
15. Pattnaik S, Subramanyam VR, Bapaji M, Kole CR. Antibacterial
and antifungal activity of aromatic constituents of essential oils.
Microbios. 1997;89:39 46.
16. Nrnberg E. Welche galenischen Grundlagen werden heute fr
die. Hautbehandlung eingesetzt? Hautarzt. 1978;29:61 7.
17. Bamba M, Puisievx F, Marty JP, Carstensen JT. Release
mechanism in gel forming sustained release formulation. Int J
Pharm. 1979;2:307 15.
18. Gibaldi M, Feldman S. Establishment of sink conditions in
dissolution rate determinations: theoretical considerations and
application to nondisintegrating dosage forms. J Pharm Sci.
1967;56:1238 42.
19. Korsmeyer R, Gurny R, Doelker E, Buri P, Peppas N.
Int J Pharm. 1983;15:25 35.
691
20. Higuchi T. Rate of release of medicaments from ointment bases
containing drugs in suspension. J Pharm Sci. 1961;50:874 5.
21. Higuchi T. Mechanism of sustained action medication. J Pharm
Sci. 1963;52:1145 9.
22. Higuchi T. Analysis of data on the medicament release from
ointments. J Pharm Sci. 1962;51:802 4.
23. Langenbucher F. Linearization of dissolution rate curve by
Weibull distribution. J Pharm Pharmacol. 1972;24:979 89.
24. U.S. Department of Health and Human Services, Food and Drug
Administration, Center for Drug Evaluation and Research.
Guidance for Industry: Dissolution Testing of Immediate Re
lease Solid Dosage Forms, August 1997.
25. Chew N, Reed B, Morgan T, Finnin B, inventors. Topical
delivery of antifungal agents. US Patent US20040081684. April
29, 2004.
26. Warren R, Reed J, inventors. Topical tanning compositions.
European patent EP19870306848. February 8, 1989.
27. http://www.accessdata.fda.gov/scripts/cder/iig/getiigWEB.cfm.
Accessed on February 1, 2009.
28. Nadakatti S, Naik V. Detergent bar and process of manufacture,
inventors. US patent US2006287206. December 21, 2006.
29. Pfaller MA, Diekema DJ, Sheehan DJ. Interpretive breakpoints
for uconazole and Candida revisited: A blueprint for the future
of antifungal susceptibility testing. Clin Microbiol Rev.
2006;19:435 47.
30. Adriana T, Valentino L, Nunzio D, Angela L, Annalisa C,
Massimo F, et al. Eudragit RS 100 microparticles containing 2
hydroxypropyl cyclodextrin and glutathione: Physicochemical
characterization, drug release and transport studies. E J Pharm
Sci. 2007;30:64 74.
31. Li GL, Geest RVD, Chanet L, Zanten EV, Danhof M, Bouwstra
JA. In vitro iontophoresis of R apomorphine across human
stratum corneum: structure transport relationship of penetration
enhancement. J Control Release. 2002;84:49 57.
32. Higuchi T, inventor. Method for in vitro determination of
transdermal absorption US Patent US4771004. August 29, 1986.

DOI: 10.1208/s12249 009 9255 9
Research Article
Dental Mold: A Novel Formulation to Treat Common Dental Disorders
Soma Ghosh,1 Gopa Roy,1 and Biswajit Mukherjee1,2
Abstract. Oral administration of antibiotics to treat dental problems mostly yields slow actions due to
slow onset and hepatic rst pass. Again, commonly used dental paints are generally washed out by
saliva within few hours of application. To overcome the challenges, polymeric molds to be placed on an
affected tooth (during carries and gum problems) were prepared and evaluated in vitro for sustained drug
release for prolonged local action. Here, amoxicillin trihydrate and lidocaine hydrochloride were used as
model drugs. Dental molds were prepared using corn zein, carbopol 934 P, gum karaya powder, and
poloxamer 407 by mixing and solvent evaporation technique. Different physicochemical evaluation
studies such as tooth adhesion test, surface pH, swelling index, and drug distribution pattern were carried
out. Percentage swelling varied from 56% to 93%. Average tooth adhesion strength and mean initial
surface pH of the formulations were 50 g and 6.5, respectively. As assessed by scanning electron
microscopy, drug distribution was uniform throughout the matrix. Cumulative percentage release of
lidocaine hydrochloride and amoxicillin trihydrate in simulated saliva were 98% and 50%, respectively.
In vitro drug release studies revealed the sustained release patterns of the drugs in simulated saliva at
least for 24 h. The stability study shows that the drugs were stable in the formulations following the
conditions as per ICH guideline. The formulation is a novel approach to deliver the drug(s) for a
prolonged period for local action upon its application on an affected tooth.
KEY WORDS: amoxicillin trihydrate; dental molds; lidocaine hydrochloride; tooth adhesion test.
INTRODUCTION
The most familiar symptom of a tooth problem is
toothache. The severity of the pain depends on the level of
which part of the tooth is affected and how deeply the decay
extends (1). Tooth decay is a destruction of the tooth enamel
(2). Each tooth has a coating of enamel, which protects the
underlying dentine containing soft tissue and nerves. When
foods are frequently left on the teeth, bacteria that live in the
mouth thrive on these foods and produce acids over a period
of time; these acids destroy tooth enamel, resulting in tooth
decay (3). To treat dental problems such as pain due to dental
caries, periodonitis, gingivitis, and other gum infections, pain
killers along with antibiotics and some dental paints are the
commonly prescribed drugs by dentists as initial mode of
treatment (4). But the common side effects of most of the
painkillers are hyperacidity and gastric irritation upon oral
administration. On the other hand, most antibiotics due to
Department of Pharmaceutical Technology, Jadavpur University,

Kolkata, 700 032, West Bengal, India.
2
To whom correspondence should be addressed. (e mail: biswajit55@
yahoo.com)
NOTATIONS: cm, centimeter; cp, centipoise; C, degree Celsius; FTIR,
Fourier transform infrared; g, gram; h, hour; KV, kilo volt; , lambda;
g/mL, microgram/mililiter; mL, mililiter; M 1 cm 1, molar 1centimeter 1;
nm, nanometer; rpm, rotation per minute; SEM, scanning electron
microscopy; UV, ultraviolet.
slow onset of action and hepatic rst pass effect fail to

produce prompt and prolong actions (5). Moreover, most of
the dental formulations are washed out by saliva within a few
hours of application. Hypothesis of the study is that if a soft
moldable gummy material containing analgesic as well as
antibiotic drugs is attached to an offending tooth and
sustained drug release occurs from it, a prolonged local
action of the drugs is achieved. The polymeric mold should
have an appropriate adhesiveness so that it may be easily
xed on the affected tooth and can be removed easily
whenever necessary. Amoxicillin trihydrate, a lactam
antibiotic, inhibits the cross linkage between the linear
peptidoglycan polymer chains of the cell wall of Gram
positive and Gram negative bacteria (6). The drug has been
used to treat dental microbial plaques and periodontal
diseases as bacterial infection (7,8). Lidocaine hydrochloride,
an anesthetic agent, is available as dermal formulation,
intravenous injection, intravenous infusion, nasal spray, oral
gel, and topical gel (9). Lidocaine has been widely used as
anesthetic agent for dental scaling, root planning, pain
sensitivity, and early wound healing following nonsurgical
periodontal therapy (10,11).
Sustained release delivery systems allocate extended
drug action to treat dental and periodontal diseases compared
to the conventional dosage forms. Better patient compliance
in terms of application frequency, better relief for longer
period of time, and reduction in dose of drug leads to
overcome the adverse reactions due to higher dose to achieve
the same effectiveness when given orally. Besides faster local
692
693
Table I. Drug Polymer Composition of Denticap Formulations

Ingredients (ratio by weight)a
Formulation
Denticap L
Denticap A
Denticap AL
Weight of drug (milligram)
Corn zein, Carbopol 934 P, Gum karaya, and Poloxamer 407

(8:8:4:1; ethyl cellulose was used for coating only)
Lidocaine hydrochloride 50 mg
Amoxicillin trihydrate 70 mg
70 mg for amoxicillin trihydrate and 50 mg for
lidocaine hydrochloride
a
Total weight of the polymer blend in each formulation 262.5 mg
L lidocaine hydrochloride, A amoxicillin trihydrate, AL amoxicillin trihydrate lidocaine hydrochloride
action as compared to slow onset of action by oral route and

avoidance of hepatic rst pass effect are important advan
tages of this formulation. Search of sustained release devices
is a relatively new area in dentistry. Many researches are in
progress to develop the similar formulations. Dental gels (12),
mucoadhesive tablets (13), lms (14), injectable semisolid
systems (15), inserts (16), and sponges (17) are some of the
sustain release drug delivery approaches for the treatment of
periodontal diseases. However, they are mostly on experi
mental level.
Therefore, an effort was made here to develop and
evaluate in vitro a dental mold denticap (since it is cap like,
we name it here as denticap) containing lidocaine hydrochlo
ride and amoxicillin trihydrate as model drugs for sustained
local action.

Materials
Carbopol 934 P (Corel Pharma Chem, Ahmedabad,
India), gum karaya powder # 150 (Nutriroma, Hyderabad,
India), lidocaine hydrochloride (Heer Pharmaceuticals Pvt.
Ltd., Mumbai, India), and amoxicillin trihydrate (Unimerk
Remedies, Birganj, Nepal) were obtained as gift samples.
Corn zein (Sigma, St. Louis, MO, USA), disodium hydrogen
Fig. 1. Outline diagram of a denticap on an affected tooth
orthophosphate anhydrous, potassium dihydrogen phosphate,

sodium chloride (Process Chemical Industries, Kolkata,
India), ethyl cellulose (S.D Fine Chemical Ltd., Boisar,
India), poloxamer 407 (Sciencelab.com, Houston, TX,
USA), and absolute ethanol (Changsu Yangyuan Chemical,
Jiangsu, China) were purchased.
Development of Formulation
Initially several polymers mixed at different ratios (by
weight) were screened for developing dental molds. The best
formulation based on physicochemical properties and drug
release has been reported here. Moldable denticaps were
formulated with combination of polymers such as corn zein,
carbopol 934 P, gum karaya, poloxamer 407, and ethyl
cellulose (Table I). Here, corn zein was used to form drug
matrix for sustained release of drugs (18). Carbopol 934 P
and gum karaya were used for achieving adhesiveness in the
formulations since they are known to possess mucoadhesive
property (19,20). Poloxamer 407 was a wetting agent (21).
Ethyl cellulose was used for coating purpose (22).
Drugs and polymers were mixed together in ethanol
(95%), using homogenizer to form a paste. The mixture was
poured in a cap shaped ethanol proof plastic molds (internal
diameter 1 cm, height 1 cm) and was then subjected to
evaporation of solvent at 37C for 2 h. Finally, the formula
tions were coated from all the sides except the side for drug
release on the affected tooth (Fig. 1). Prior to coating, the
Fig. 2. Mucoadhesive strength test assembly
694
Ghosh, Roy, and Mukherjee
Fig. 3. FTIR spectra of a amoxicillin trihydrate; b lidocaine

hydrochloride; c a mixture of amoxicillin trihydrate and lidocaine
hydrochloride in a ratio of 1:1; d a mixture of polymers corn zein,
carbopol 934 P, gum karaya, poloxamer 407, and ethyl cellulose; e a
mixture of amoxicillin trihydrate and lidocaine hydrochloride with
those polymers
formulations were removed from the plastic molds. The

coating was done using 5% ethanolic solution of ethyl
cellulose following the method available (23) keeping drug
release side covered.
Drug-Excipient Interaction Study Using FTIR Spectroscopy

Drug excipient interaction study is a preformulation
experiment (24) to conrm whether the excipients of a
formulation interact chemically among themselves as well as
with drug(s) present in the formulation. FTIR spectroscopy
was used here for the purpose. Pure lidocaine hydrochloride
and amoxicillin trihydrate separately; mixture of them;
mixture of lidocaine hydrochloride and amoxicillin trihydrate
with polymeric combinations; and polymeric combinations
alone were mixed separately with IR grade KBr in the ratio
1:100 and corresponding pellets were prepared by applying
5.5 metric ton of pressure in a hydraulic press. The pellets
were scanned over a wave number range of 4,000 400 cm 1 in
Magna IR 750 Series II (Nicolet, USA) FTIR spectroscope.
Tooth Adhesion Test

Teeth of goats collected from a slaughterhouse were
xed on a plaster of Paris (CaSO4, 0.5H2O) base to prepare
tooth model. Teeth were cleaned with distilled water before
xing. The individual experimental formulation was attached
(at the surface which had not been coated with ethyl
cellulose) to the tooth model after being made wet with
simulated saliva (pH 6.8, viscosity 0.740 cp at 370.5C) for
2 min. A physical balance with two circular pans, hanged from
a rod which was balanced with a fulcrum on a stand (Fig. 2),
was used as a modied tooth adhesion test assembly (25).
Lower surface of a pan was xed to the denticap attached to
the tooth model as described above, by both side adhesive
tape. Weights were given on the other pan continuously until
the denticap was detached from the tooth model. Simulated
saliva was prepared by dissolving 2.38 g Na2HPO4, 0.19 g
KH2PO4, and 8 g of NaCl in a liter of distilled water (26).
695
Surface pH
The denticaps were incubated in a Petri dish in simulated
saliva, pH 6.8 at 370.5C for 2 h. Then, the surface pH was
determined by touching the electrode of a pH meter
(Toshniwal Instruments, Ajmer, India) in the excess simulat
ed saliva present at the surface of the denticaps (29).
The drug distribution patterns of denticaps (before and
after drug release) were studied using scanning electron
microscope (JSM 6700F, JEOL, Tokyo, Japan). Experimental
samples were cut and mounted onto stubs, and then platinum
was sputtered under vacuum. They were visualized at an
acceleration voltage of 5 KV.
In Vitro Drug-Release Study
Release of lidocaine hydrochloride amoxicillin trihy
drate from denticaps was carried out in a USP apparatus I.
In this apparatus, denticaps were placed inside the basket and
the drug release occurs from only one side of the formulation,
which remains open towards the reservoir containing 100 mL
of simulated saliva, pH 6.8 as a dissolution media at 370.5
C, and 50 rpm (16). The ow rate of saliva was maintained by
replacing the dissolution medium from time to time at a
predetermined time intervals (30). Amoxicillin trihydrate and
lidocaine hydrochloride were assayed by simultaneous equa
tion UV method (31) at their max, respectively, using Cary 50
UV VIS spectrophotometer (Varian, Palo Alto, CA, USA).
Both the drugs obeyed linearity at the concentration of 10
100 g/mL, and the correlation coefcient (R2) was less than
1 in both the cases (for lidocaine hydrochloride, 0.9996 and
for amoxicillin trihydrate, 0.9994). The calculated molar
absorptivities (E1%, 1 cm) of lidocaine hydrochloride were
Percent Swelling
The original weights of the denticaps (Wo) were deter
mined and allowed to swell on a Petri dish in simulated saliva,
pH 6.8 at 370.5C. At predetermined time intervals (1 5 h),
increase in weights (wet weight) was reported after removal
of excess saliva with lter paper. When the weight became
constant (Wt), percent swelling was calculated in terms of
water uptake (27,28).
Wt W0 100
Percent swelling
W0
Fig. 4. Percent swelling in terms of water absorption capacity of

denticaps. Data show meanSD (n 3)
696
Table II. Surface pH and Tooth Adhesive Strength of Denticaps (Containing both Lidocaine Hydrochloride and Amoxicillin Trihydrate)
Stored at Different Temperature and Humidity Conditions
Surface pHSD (n 3)
Tooth adhesive strength (g)SD (n 3)
Storage condition
Initial
1 months
3 months
Initial
1 months
3 months
302C/60% RH
452C/75% RH
2 8C
6.50.75
6.970.43
7.30.36
6.690.19
6.90.58
6.950.44
6.770.32
504.56
252.21
103.75
302.25
204.89
297.88 and 460.36 M 1 cm 1 at 273 and 263 nm, respectively,

and for amoxicillin trihydrate, the values were 1,048.62 and
796.95 M 1 cm 1 at 273 and 263 nm, respectively.
Dental mold was taken in 100 mL simulated saliva, pH
6.8 in a volumetric ask, and the mixture was stirred for 48 h
at room temperature using a magnetic stirrer (Remi Equip
ments, Mumbai, India). The drug content analysis for
amoxicillin trihydrate and lidocaine hydrochloride was done
by simultaneous equation UV method as described above.
Initially, the time of analysis of the method was
standardized by taking formulation with measured amount
of drug in simulated saliva and determination of amount of
drug released with the duration. It was found that 100%
release of both the drugs was achieved in 48 h. Therefore, the
time of drug content analysis was chosen up to 48 h.
Temperature-Dependent Stability Study
Denticaps containing the drugs were stored at elevated
temperatures and relative humidity (30 2C/60% RH,
452 C/75% RH) in a stability analysis chamber (Darwin
Chambers Company, St. Louis, MO, USA) over a period of
3 months as per ICH guideline (32). Denticaps stored at
2 8C were used as control. The surface pH, drug content, tooth
adhesion strength, in vitro release proles, and FTIR data of the
samples kept for 1 and 3 months for stability analysis were
compared with those of the control formulations (33).
amounts of amoxicillin trihydrate and lidocaine hydrochloride

in each formulation were 70 and 50 mg, respectively.
Here, drug excipient interaction was studied using FTIR
spectroscopy. Fig. 3a e shows the FTIR spectra of amoxicillin
trihydrate; lidocaine hydrochloride; a mixture of amoxicillin
trihydrate and lidocaine hydrochloride; a mixture of polymers
corn zein, carbopol 934 P, gum karaya, poloxamer 407, and
ethyl cellulose; and a mixture of amoxicillin trihydrate and
lidocaine hydrochloride with those polymers, respectively. The
FTIR spectra of pure amoxicillin trihydrate (Fig. 3a) and
lidocaine hydrochloride (Fig. 3b) shows that all the character
istic peaks of amoxicillin trihydrate and lidocaine hydrochlo
ride are present (34,35). Again, the FTIR spectra of the
combination of both drugs (Fig. 3c) show the presence of their
characteristic peaks. This indicates that the drugs do not
interact chemically in their physical mixture. Fig. 3d provides
FTIR spectra of the excipients alone. When the drugs were
mixed with the excipients, variation of some peaks were
noticed (Fig. 3e). When Fig. 3c and d were compared, between
3,200 and 2,800 cm 1 and between 1,800 and 1,000 cm 1 wave
numbers, variations at transmission spectroscopy data were
noted. Alkenyl (>C=C<; 3,020 3,100 cm 1), amide (>NH;
1,000 1,250 cm 1) ketonyl (>C=O; 1,710 1,720 cm 1), phenolic
( OH; 970 1,250 cm 1) stretches are mainly responsible
functional groups for those regions. This suggests the
formation of weak to medium intensity bonds due to the
forces such as Vander Waal force, dipole moment, and
Statistics
Data were assessed by one way ANOVA followed by
Tukey HSD Test using Vassar Stats software (USA). P<0.01
has been considered as statistical signicance.
RESULTS
Several combinations of various polymers were screened
to obtain the required physicochemical properties and in vitro
drug release pattern. The denticap reported here consisted of
corn zein, carbopol 934 P, gum karaya, and poloxamer 407 at
a ratio of 8:8:4:1 (total amount of the polymer blend was
262.5 mg). The polymeric ratio 8:8:4:1 was selected through
random screening procedure and selection of polymers, and
their ratios in mixture were established experimentally and
the best formulation has been reported in the study. The
Fig. 5. SEM photograph of amoxicillin trihydrate lidocaine hydro

chloride denticap before drug release
Fig. 6. SEM photograph of amoxicillin trihydrate lidocaine hydro

chloride denticap after drug release
electrostatic force between drugs, and the excipients in those

regions since no major shifting of peaks was noted.
The percent swelling results were expressed in terms of
percentage water uptake at 37C (36). Figure 4 shows the
swelling pattern of denticaps. The water uptake of the
denticaps can be arranged as follows: blank (only polymers) <
amoxicillin trihydrate denticap < amoxicillin trihydrate
lidocaine hydrochloride denticap < lidocaine hydrochloride
denticap. The results show that the percentage swelling of
the various denticaps varied from 56% to 93%.
Table II represents the tooth adhesive strength of the
experimental denticaps containing both amoxicillin trihydrate
and lidocaine hydrochloride. The average initial tooth
adhesive strength (measured in g) was found to be 50 g.
Average mucoadhesive strength required for mucoadhesion
was reported to be 30 g (37), whereas the mean adhesive
strength obtained in our formulation was 50 g. This was
advantageous because the surface of mucous layer of buccal
cavity is uneven compared to tooth surface; hence, more
adhesion strength is required to attach the formulation on
tooth surface (38,39). Moreover, tooth adhesive strength was
found to be good enough to hold the formulations to the
tooth, and again, the adhesiveness in our formulations was
such that the formulations were easily removable from the
tooth with a little effort.
697
The surface pH obtained (Table II) in this study were

within the limits and showed hardly any variation from time
to time which omits the chances of irritation in the mucosal
membrane upon application.
Figures 5, 6 show the distribution of amoxicillin trihy
drate and lidocaine hydrochloride in the matrix before and
after the release, respectively. Drug particles were distributed
throughout the matrix, and particle sizes were about 10 nm.
After drug release study, most of the drug particles were
found to diffuse out of the formulation. Further, the matrix
(background) as shown in Fig. 6 was found to be different
from that of Fig. 5. Upon the exposure of the matrix to the
simulated saliva during drug release study, penetration of
liquid into the formulation eventually caused the matrix to
swell as supported by the results of Fig. 4. This resulted in
differences in appearance in polymer matrices in the scanning
electron microscopy (SEM) microphotographs between Fig. 5
(without any appearance of swelling) and Fig. 6 (appearance
of swelling).
The drug content analysis of denticaps containing both
amoxicillin trihydrate and lidocaine hydrochloride shows that
the initial recoveries of amoxicillin trihydrate and lidocaine
hydrochloride were 98.85% and 99.89%, respectively
(Table III). The data were reproducible.
Release patterns of lidocaine hydrochloride and amoxi
cillin trihydrate from the prepared denticaps were studied in
simulated saliva (pH 6.8). Figure 7 shows the release of
amoxicillin trihydrate and lidocaine hydrochloride from the
denticap for 24 h. The cumulative percentage release of
lidocaine hydrochloride and amoxicillin trihydrate were about
98.18% and 50.46%, respectively, over a time period of 24 h.
To investigate the drug release kinetic pattern, drug
release data were checked using Zero Order, First Order,
Higuchi, Korsmeyer Peppas, and Hixson Crowell kinetic
models. R2 values for the studied kinetics have been
included as Table IV. The release of lidocaine hydrochloride
from the denticap showed to follow apparent zero order
kinetics (Table IV). Likewise, amoxicillin trihydrate release
was found to obey apparent zero order kinetics, too.
Tables II, III represent the various data of stability test
samples stored at different temperature and humidity con
ditions as per ICH guideline (32). There was no predominant
variation in surface pH values of the stored formulations.
However, 2% to 3% degradation of both the drugs were
noticed when the samples had been stored at 452C/75%
RH. Adhesive strength was found to be reduced during
storage. This could be due to the loss of moisture from the
Table III. Percentage Drug content of Denticaps (Containing Both Lidocaine Hydrochloride and Amoxicillin Trihydrate) Stored at Different
Temperature and Humidity Conditions
% Drug contentSD (n 3)
Storage condition
Initial
1 months
3 months
302C/60% RH
Lidocaine hydrochloride
99.893.53 Amoxicillin
trihydrate 98.850.84
Lidocaine hydrochloride 99.516.39

Amoxicillin trihydrate 99.015.11

452C/75% RH
2 8C
698

Fig. 8. FTIR spectra a polymer mixture stored at 2 8C; b polymer
mixture along with amoxicillin trihydrate and lidocaine hydrochloride
stored at 2 8C; c polymer mixture stored at 302C/60% RH; d
polymer mixture along with amoxicillin trihydrate and lidocaine
hydrochloride stored at 302C/60% RH; e polymer mixture stored
at 452C/75% RH; f polymer mixture along with amoxicillin
trihydrate and lidocaine hydrochloride stored at 452C/75% RH
interactions as compared to control samples were noted in the

formulations stored at the other mentioned conditions.
DISCUSSION
Fig. 7. In vitro release of lidocaine hydrochloride and amoxicillin

trihydrate from the denticap having both amoxicillin trihydrate and
lidocaine hydrochloride in simulated saliva (pH 6.8). Data shows
mean (n 6)SD. Values were signicantly different as accessed by
one way ANOVA followed by Tukey HSD test (p<0.01)
formulation during storage. Figure 8a f shows the FTIR

spectra of polymer mixture stored at 2 8C; polymer mixture
along with amoxicillin trihydrate and lidocaine hydrochloride
stored at 2 8C; polymer mixture stored at 302C/60% RH;
polymer mixture along with amoxicillin trihydrate and
lidocaine hydrochloride stored at 302C/60% RH; polymer
mixture stored at 452C/75% RH; and polymer mixture
along with amoxicillin trihydrate and lidocaine hydrochloride
stored at 452C/75% RH, respectively. Figure 8f (denticaps
with drugs stored at 452C/75% RH for 3 months) appeared
to have some variations in peaks, in particular, between 1,500
and 1,000 cm 1 and between 800 and 400 cm 1. Wave
numbers between 1,250 and 1,000 cm 1 is the stretching
vibration zone of C N and between 1,250 and 970 cm 1 is the
stretching vibration zone of C O. Wave numbers between
1,360 and 1,350 cm 1 and between 1,390 and 1,370 cm 1 are
the bending vibration zone of CH2 and CH3 deformation
and OH bending. Further, the wave numbers between
900 and 660 cm 1 are the bending vibration zone of C H
bending and ring puckering, OH bending (out of plane),
C H deformation, and NH2 and N H wagging due to shifts
on H bonding. Since N H, OH, NH2, O=, C H, N H,
N=, CH2 groups are present in drugs and excipients, there
might be some physical interactions between drugs and
excipients in those regions at high temperature and humidity
condition (452C/75% RH) which differs the spectra of Fig. 8f
from the rest (Fig. 8a e). No different drug excipient
A cap like, soft gummy dental mold (denticap) contain

ing amoxicillin trihydrate and lidocaine hydrochloride was
developed so that the drugs might release from the formula
tion for a prolonged period upon its application on an
affected tooth. The doses of the drugs were selected as per
the earlier reports (14,40).
One of the important preformulation studies to develop
a new dosage form is drug excipient interaction. Among the
various methodologies available to evaluate drug excipient
interaction, the FTIR spectroscopy is an easy and simple
option (41). FTIR spectroscopy shows interaction between
molecules at the level of functional groups (42). The FTIR
spectra suggests that there may be physical interactions due
to formation of weak to medium intensity bonds since no
major shifting of peaks was noted (36). Blend of polymers is
known to change the rate and pathway of diffusion of drug
molecules by varying entanglement in polymeric network
(43). Thus, the physical interactions might be helpful in
sustaining the release of drug molecules from the experimen
tal formulations.
Moisture uptake by polymers and subsequent swelling in
polymer matrix are common phenomenon in case of most of
the available polymers (44). Swelling depends on the polymer
concentration and the moisture absorption capacity of the
polymers (45). Flexibility of polymer chain from individual
polymer is important for interpenetration and entanglement,
and in presence of water molecules, polymers become cross
linked and the mobility of individual polymer chain decreases
and swelling occurs (46). In the present study, drug polymer
matrix swelled more (40%) as compared to the polymeric
matrix without drug (Fig. 4). Incorporation of drug shows
more swelling of the matrix, and this may be because of the
hygroscopic nature of lidocaine hydrochloride (47) and
presence of water of crystallization in amoxicillin trihydrate.
This might provide more water molecules deep inside of the
Table IV. Drug Release Kinetics of Denticaps
Higuchi kinetics
Korsmeyer et al.
kinetics
Hixon Crowell
kinetics
Formulation
Zero order kinetics
First order kinetics
Combined denticap
(lidocaine hydrochloride release)
Combined denticap
(amoxicillin trihydrate release)
y 1.8678x+53.509
R2 0.9796
y 0.7728x+32.133
R2 0.9455
y 0.0935x+2.0739 y 22.359x1.5838 y 0.7573x+1.1241 y 0.1348x+0.3062

R2 0.9323
R2 0.8486
R2 0.8244
R2 0.9445
y 0.0068x+1.8526 y 7.594x+16.249 y 0.3239x+1.3066 y 0.0237x+0.4462
R2 0.9156
R2 0.8649
R2 0.8997
R2 0.7762
699
700
polymeric matrix as compared to the blank formulation
(without drug), where seepage of water molecules from the
surface to the core might take more time, which ultimately
caused less swelling at the same time points except at the rst
hour where the value was similar to that of denticaps
containing both drugs (reason is unknown). Further presence
of drug in the polymeric network structure might make more
pathways for the seepage of water into the core of the
polymers faster as compared to the polymer matrix without
drug. The highest percentage of moisture uptake was
obtained for lidocaine hydrochloride denticap. The reduced
swelling ability of amoxicillin trihydrate denticap may be
because of less water afnity of amoxicillin trihydrate than
lidocaine hydrochloride (47,48) and presence of amoxicillin
trihydrate in the matrix.
The tackiness of the formulations was determined by
tooth adhesion test. Tooth adhesive strength was found to be
good enough to hold the formulations to the tooth and again,
the tackiness was such that the denticaps were easily
removable from the tooth with a little effort.
SEM photographs show that drug particles were very
small (approximately 10 nm) in size and were distributed
uniformly throughout the matrix. Following drug release
study, most of both drugs were found to diffuse. However,
presence of amoxicillin trihydrate in drug matrix at 24 h after
drug release was predominantly more than that of lidocaine
hydrochloride. This was further supported by studying the
drug release patterns from the formulations. Cumulative
amount of lidocaine was found to be twice more than that
of amoxicillin at 24 h in drug release study. This could be
because of the higher aqueous solubility of lidocaine hydro
chloride as compared to amoxicillin trihydrate. However, the
study suggests that the formulations would release the drugs
for an extended period of time in the oral cavity and the
drugs (amoxicillin trihydrate and lidocaine hydrochloride)
would be available for permeation through gums and gingival
mucosa. Available reports regarding the mucosal permeation
of lidocaine hydrochloride (49) and amoxicillin trihydrate
(50) suggest that the drugs would provide local action, and
there may be possibility of systemic drug action form the
formulations also. However, the present work was intended
for local action only, and no study was performed here to
evaluate systemic drug absorption from the formulation.
When drug release patterns were investigated, data
suggest that initially, faster diffusion of the drug molecules
from the surface of the denticaps showed rst order drug
release. But the drug release patterns varied in 2 h for
amoxicillin trihydrate and in 5 h in case of lidocaine
hydrochloride. It was noted that drug release gradually
changed from concentration dependent rst order release
kinetics to concentration independent zero order kinetic
pattern. With the time, the swelling of the polymers varied
the entanglement of polymeric pathways to control the
diffusion of the drugs from the formulations (51). Again,
amoxicillin trihydrate was found to release much slower than
lidocaine hydrochloride with the duration. This could be due
to the higher aqueous solubility of lidocaine hydrochloride
than that of amoxicillin trihydrate (47,48).
Results of the stability study show that drug content and
surface pH were almost the same as those of the control
formulations, whereas the tooth adhesive strength decreased

with time and elevated temperature conditions. The stability
study was carried out following ICH guideline for stability
testing of new drug substances and products (32) in a
container closure system. Since the study was conducted in a
container closure system, the product could have lost a percent
age of moisture in the container closure system during its
storage for 3 months at 2 8C temperature. This possibly
resulted in a decrease of tooth adhesive strength value as
compared to the fresh formulations (i.e., before storage). The
release patterns (not presented here) for both amoxicillin
trihydrate and lidocaine hydrochloride also followed the same
pattern as the control denticaps. FTIR spectra of all the samples
which have undergone the stability study had similarity with
those of the control ones, and there was hardly any variation
among the samples stored at different temperature and
humidity conditions except that the samples stored at 452C/
75% RH for 3 months where some variations in the spectra
were noticed. Therefore, it is obvious from the results that the
denticaps were stable both physically and chemically at those
elevated temperature and humidity conditions other than 452
C/75% RH. Our data regarding the stability of the drugs are
also supported by the reported ndings (52,53).
CONCLUSION
Denticap is a novel approach to local drug delivery for
a prolonged period by applying it on the affected tooth. More
patient compliance over the conventional dosage forms is the
expected outcome of such approach. In the future, in root
canal treatment and in different gum therapy, this type of
dosage form may be very effective. Therefore, denticap may
open a new era in dentistry in near endeavor.
ACKNOWLEDGEMENTS
This work was supported by grants from the Council of
Scientic and Industrial Research (Grant No. 27(0149)/05/
EMR II) and the Indian Council of Medical Research (Grant
No. 45/55/2004 PHA/BMS).
REFERENCES
1. Feiglin B. Aspects of dentinal and pulpal pain. Diagnosing dental
pain. Ann R Australas Coll Dent Surg. 1994;12:143 52.
2. Carvalho RM, Tay FR, Giannini M, Pashley DH. Effects of pre
and post bonding hydration on bond strength to dentin. J Adhes
Dent. 2004;6:13 7.
3. Marthaler TM. Changes in dental caries 1953 2003. Caries Res.
2004;38:173 81.
4. Vyas SP, Sihorkar V, Mishra V. Controlled and targeted drug
delivery strategies towards intraperiodontal pocket diseases. J
Clin Pharm Ther. 2000;25:21 42.
5. Letendre L, Scott M, Dobson G, Hidalgo I, Aungst B.
Evaluating barriers to bioavailability in vivo: validation of a
technique for separately assessing gastrointestinal absorption
and hepatic extraction. Pharm Res. 2004;21:1457 62.
6. Orosz SE, Jones MP, Cox SK, Zagaya NK, Frazier DL.
Pharmacokinetics of amoxicillin plus clavulanic acid in blue
fronted Amazon parrots (Amazona aestiva aestiva). J Avian Med
Surg. 2000;14:107 12.

7. Bascones Martinez A, Fiquero Ruiz E. Periodontal diseases as
bacterial infection. Med Oral Patol Oral Cir Bucal. 2004;9:92 100.
8. Akincibay H, Orsal SO, Sengn D, Tzm TF. Systemic admin
istration of doxycycline versus metronidazole plus amoxicillin in
the treatment of localized aggressive periodontitis: a clinical and
microbiologic study. Quintessence Int. 2008;39:e33 9.
9. Khaliq W, Alam S, Puri N. Topical lidocaine hydrochloride for
the treatment of postherpetic neuralgia. Cochrane Database Syst
Rev. 2007;18:CD004846.
10. Stoltenberg JL, Osborn JB, Carlson JF, Hodges JS, Michalowicz
BS. A preliminary study of intra pocket topical versus injected
anaesthetic for scaling and root planing. J Clin Periodontol.
2007;34:892 6.
11. Kasaj A, Heib A, Willershausen B. Effectiveness of a topical
salve (Dynexan) on pain sensitivity and early wound healing
following nonsurgical periodontal therapy. Eur J Med Res.
2007;12:196 9.
12. Stoltze K, Stellfeld M. Systemic absorption of metronidazole
after application of a metronidazole 25% dental gel. J Clin
Periodontol. 1992;19:693 7.
13. Perioli L, Ambrogi V, Rubini D, et al. Novel mucoadhesive
buccal formulation containing metronidazole for the treatment of
periodontal disease. J Control Release. 2004;95:521 33.
14. Ahuja A, Ali J, Rahman S. Biodegradable periodontal intra
pocket device containing metronidazole and amoxycillin: formu
lation and characterisation. Pharmazie. 2006;61:25 9.
15. Schwach Abdellaoui K, Vivien Castioni N, Gurny R. Local
delivery of antimicrobial agents for the treatment of periodontal
diseases. Eur J Pharm Biopharm. 2000;50:83 99.
16. Barat R, Srinatha A, Pandit JK, Ridhurkar D, Balasubramaniam
J, Mittal N, et al. Niridazole biodegradable inserts for local long
term treatment of perodontis: possible new life for an orphan
drug. Drug Deliv. 2006;13:365 73.
17. Ang CY, Samsudin AR, Karima AM, Nizam A. Locally
produced bovine bone sponge as a haemostatic agent. Med J
Malaysia. 2004;59:149 50.
18. Liu X, Sun Q, Wang H, Zhang L, Wang JY. Microspheres of corn
protein, zein, for an ivermectin drug delivery system. Biomate
rials. 2005;26:109 15.
19. Patel VM, Prajapati BG, Patel MM. Effect of hydrophilic
polymers on buccoadhesive Eudragit patches of propranolol
hydrochloride using factorial design. AAPS PharmSciTech.
2007;8:45.
20. Collys KD, Roma de Sousa AC, Smeyers Verbeke J. Soluble
denture adhesives: pH and sodium content. Eur J Prosthodont
Restor Dent. 1997;5:63 7.
21. Wong SM, Kellaway IW, Murdan S. Enhancement of the
dissolution rate and oral absorption of a poorly water soluble
drug by formation of surfactant containing microparticles. Int J
Pharm. 2006;317:61 8.
22. Bussemer T, Bodmeier R. Formulation parameters affecting the
performance of coated gelatin capsules with pulsatile release
proles. Int J Pharm. 2003;267:59 68.
23. Oh E, Luner PE. Surface free energy of ethylcellulose lms and
the inuence of plasticizers. Int J Pharm. 1999;188:203 19.
24. Bruni G, Amici L, Berbenni V, Marini A, Orlandi A. Drug
excipient compatibility studies. Search of interaction indicators. J
Thermal Anal Cal. 2002;68:561 73.
25. Gupta A, Garg S, Khar RK. Measurement of bioadhesive buccal
tablets: design of an in vitro assembly. Indian Drugs. 1993;30:
152 5.
26. Yvonne TFT, Kok KP, Al Hanbali O. Effect of carbopol and
polyvinylpyrrolidone on the mechanical, rheological and release
properties of bioadhesive polyethylene glycol gels. AAPS Pharm
Sci Tech. 2000;1(3):24.
27. Semalty A, Bhojwani M, Bhatt GK, Gupta GD, Shrivastav AK.
Design and evaluation of mucoadhesive buccal lms of diltiazem
hydrochloride. Indian J Pharm Sci. 2005;67:548 52.
28. Vishnu MP, Bhupendra GP, Harsha VP, Karshanbhi MP.
Mucoadhesive bilayer tablets of propranolol hydrochloride.
AAPS Pharm Sci Tech. 2007;8(3):77.
29. Nafee NA, Ismail FA, Boraie NA, Mortada LM. Mucoadhesive
buccal patches of miconazole nitrate:in vitro/in vivo performance
and effect of ageing. Int J Pharm. 2003;264:1 14.
701
30. Colin D. Salivary ow patterns and the health of hard and soft
oral tissues. J Am Dent Asso. 2008;139:18 24.
31. Devarajan S, Lakshmi S. Simulataneous spectrophotometric
determination of valdecoxib and tizanidine in tablets. Indian J
Pharm Sci. 2006;68:240 2.
32. International Conference on Harmonisation of Technical
Requirements for Registration of Pharmaceuticals for Human
Use. Stability testing of new drug substances and products Q1A
(R2). Current Step 4 version dated 6 February 2003.
33. Garg M, Dutta T, Jain NK. Stability study of stavudine loaded
O palmitoyl anchored carbohydrate coated lyposomes. AAPS
PharmSciTech. 2007;8:38.
34. Brittain HG. Solid state uorescence of the trihydrate phases of
ampicillin and amoxicillin. AAPS PharmSciTech. 2005;6:E444 8.
35. Abu Huwaij R, Assaf S, Salem M, Sallam A. Mucoadhesive
dosage form of lidocaine hydrochloride: I. Mucoadhesive and
physicochemical characterization. Drug Dev Ind Pharm. 2007;33:
855 64.
36. Mukherjee B, Das S, Patra B, Layek B. Nefopam containing
transdermal matrix based on pressure sensitive adhesive poly
mers. Pharm Tech (USA). 2006;30:146 63.
37. Singh B, Ahuja N. Development of controlled release buccoad
hesive hydrophilic matrices of diltiazem hydrochloride: optimi
zation of bioadhesion, dissolution, and diffusion parameters.
38. Badran D, Soutar DS, Robertson AG, Payne AP, McDonald
SW, Scothorne RJ. Scanning electron microscopy of the
surface morphology of supercial cells of buccal mucosa is
unlikely to be useful in monitoring radiotherapy. Clin Anat.
2005;7:34 41.
39. Hamada S, Torii M, Kotani S, Tsuchitani Y. Adherence of
Streptococcus sanguis clinical isolates to smooth surfaces and
interactions of the isolates with Streptococcus mutans glucosyl
transferase. Infect Immun. 1981;32:364 72.
40. Leopold A, Wilson S, Weaver JS, Moursi AM. Pharmacokinetics
of lidocaine delivered from a transmucosal patch in children.
Anesth Prog. 2002;49:82 7.
41. McCauley JA, Brittain HG. In: physical characterizations of
pharmaceutical solids. New York: Marcel Dekker; 1995.
42. Ktting C, Gerwert K. Monitoring protein ligand interactions by
time resolved FTIR difference spectroscopy. Methods Mol Biol.
2005;305:261 86.
43. Singh SK, Fan LT. A generalized model for swelling controlled
release systems. Biotechnol Prog. 1986;2:145 56.
44. Bagyalakshmi J, Vamsikrishna RP, Manavalan R, Ravi TK,
Manna PK. Formulation development and in vitro and in vivo
evaluation of membrane moderated transdermal systems of
ampicillin sodium in ethanol: pH 4.7 buffer solvent system.
AAPS PharmSciTech. 2007;8:7.
45. Moustane RI, Margulis EB, Sibgatullina LF, Kemenova VA,
Van den Mooter G. Comparative evaluation of interpolyelec
trolyte complexes of chitosan with Eudragit L100 and Eudragit
L100 55 as potential carriers for oral controlled drug delivery.
Eur J Pharm Biopharm. 2008;70:215 25.
46. Narasimha B, Peppas NA. The role of modeling studies in the
development of future control release devices. In: Park K, editor.
Controlled drug delivery: challenges and strategies. Washington
D.C.: American Chemical Society; 1997. p. 529 557.
47. Gulsen D, Chauhan A. Effect of water content on transparency,
swelling, lidocaine diffusion in p HEMA gels. J Membr Sci.
2006;269:35 48.
48. Song M, Li N, Sun S, Tiedt LR, Liebenberg W, de Villiers MM.
Effect of viscosity and concentration of wall former, emulsier
and pore inducer on the properties of Amoxicillin microcapsules
prepared by emulsion solvent evaporation. Farmaco. 2005;60:
261 7.
49. Perry DA, Gansky SA, Loomer PM. Effectiveness of a trans
mucosal Lidocaine delivery and root planning. J Clin Periodon
tal. 2005;32:590 4.
50. Ahuja A, Rahman S, Ali J, Chaudhry R. Effect of dental lms
containing amoxicillin and metronidazole on periodontal patho
gens: microbiological response. Pharmazie. 2003;58:716 20.
51. Mukherjee B, Mahapatra S, Gupta R, Patra B, Tiwari A, Arora
P. A comparison between povidone ethyl cellulose and povi
702
done eudragit transdermal dexamethasone matrix patches based
on in vitro skin permeation. Eur J Pharm Biopharm. 2005;59:
475 83.
52. Kayumba PC, Risha PG, Shewiyo D, Msami A, Masuki G,
Ameye D, et al. The quality of essential antimicrobial and
antimalarial drugs marketed in Rwanda and Tanzania: inuence

of tropical storage conditions on in vitro dissolution. J Clin
Pharm Ther. 2004;29:331 8.
53. Abdelmageed R, Labyad N, Watson DG, Pournamdari M,
Cable CG, Stanley E. Evaluation of the stability of morphine
sulphate in combination with InstillagelR. J Clin Pharm Ther.
2008;33:263 71.

AAPS PharmSciTech Volume 10, Number 2 6.09a PDF

Încărcat de

Informații document

Descriere originală:

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

AAPS PharmSciTech Volume 10, Number 2 6.09a PDF

Încărcat de

Drepturi de autor:

Formate disponibile

AAPS PharmSciTech, Vol. 10, No.

2, June 2009 (# 2009)

Division of Pharmaceutics, College of Pharmacy, The University of

Examples of drug instability due to the Maillard

1530-9932/09/0200-0317/0 # 2009 American Association of Pharmaceutical Scientists

Gokhale, Kearney, and Kirsch

Fig. 1. Structure of kynurenine

Glycosylation of Aromatic Amines I: Characterization of Reaction Products and Kinetic Scheme

METHODS AND MATERIALS

equation relating the absorbance to the fraction and absorp

Gokhale, Kearney, and Kirsch

desired pH and subjected to a 1D proton and a 2D HMQC

where obs is the observed chemical shift, acid is the chemical

For separation and LC MS analysis of reaction products

Reaction products were identied using UV, NMR, and

Glycosylation of Aromatic Amines I: Characterization of Reaction Products and Kinetic Scheme

Total concentration M103

experiments were also performed on products I and II setting

reactions were carried out in the pH range of 1 6. The

Kinetics of Kynurenine and Reducing Monosaccharides

The apparent Ka value of 0.01445 (pKa =1.84) at 40C for

Reaction mixtures of 1.2 mM kynurenine and 0.50 M

Table II. Experimental Conditions of Reactions of 1 mM Kynurenine

Gokhale, Kearney, and Kirsch

chemical shift (ppm)

chemical shift (ppm)

Isolation and Characterization of Reaction Products

Fig. 4. Representative chromatogram of reaction mixture of kynur

Reactions of kynurenine with galactose, mannose, allose,

Glycosylation of Aromatic Amines I: Characterization of Reaction Products and Kinetic Scheme

conditions by maintaining constant pH, temperature, and

Table III. Chemical Shift Assignments in ppm for Kynurenine

Chemical shift (ppm)

Side chain region

determined for the products I and II, respectively, by decoupling

Reactions of kynurenine in the presence of excess

aromatic amine proton

Gokhale, Kearney, and Kirsch

Side chain region

observed rate constant (kobs) for the overall loss of kynurenine

estimated to be 1.84 at 40C using UV spectroscopy.

Table IV. Chemical Shift Assignments for the Glycosylamines Based

Determination of pKa Values

Fig. 6. Typical concentration area time proles of reaction of 1.2 mM

Glycosylation of Aromatic Amines I: Characterization of Reaction Products and Kinetic Scheme

Half life (h)

Percent of initial concentration at equilibrium

water and deuterium oxide can be described by a linear

pKH $pK a b pKH

where a=0.332 and b=0.040 (38). Substituting the pKa value

The chemical shift assignments for the protons on the

Gokhale, Kearney, and Kirsch

The decoupling experiment was performed by setting the

half life (hr)

1. Decoupling of either of those protons affected the

total acyclic content (%)

overlapping each other, one from the protons in the side

sugar Conc (M)

The reaction of kynurenine and glucose reversibly

Glycosylation of Aromatic Amines I: Characterization of Reaction Products and Kinetic Scheme

8. A. Sianipar, J. E. Parkin, and V. B. Sunderland. The reaction of

Gokhale, Kearney, and Kirsch