CLIN. CHEM.

40/11,

Interference
Martin

H.

1996-2005

with

Clinical

and

Kroll

(1994)

Ronald

Laboratory

Analyses

J. Elin

Interference

were

with assays

by endogenous
and exogenous
substances
for clinical analytes
is a common
problem
in
laboratory
medicine.
For this review, we defined interference as the effect of a substance
present in the sample
that alters the correct
value of the result,
usually
expressed
as concentration
or actMty,
for an analyte.
There are four major endogenous
compounds
that consistently
interfere with laboratory
results: hemoglobin,
bilirubin,
lipids, and paraproteins.
The major exogenous
sources
of interference
are drugs prescribed
for the pa-

(blood

tient; and there are several excellent compendia

of the
effect of drugs on clinical laboratory tests. We recommend
determining
whether the interference
is dependent
or independent
of the analyte for the assay. Further,
we propose an approach
to the identification
and resolution
of an
interference
problem
for the clinical laboratory
and make

recommendations

to manufacturers.

Terms: variation, source

bin/bilirubin/Iipemia/paraproteinemia
Indexing

The

rate

goal

and

analysis,

enous
fluids
an

of the
precise
have

and
may

analyte.

clinical

laboratory

Body

a complex

the

and

fluids,

is to produce

the

undefined

compounds

the

metabolites/hemoglo-

accurate

interference

material

matrix.

present

of

determination
may

also

be found

control
materials
and survey
specimens
for proficiency
testing.
The importance
of interference
on clinical
laboratory
analyses
may be estimated
by frequency
and impact
on
patient
care. The frequency
of interference
with clinical
laboratory
analyses
is difficult
to determine.
A study
in
1971 assessed
the interference
by drugs
on laboratory
tests
in 100
outpatients
(1). The
percentage
of tests
affected
by interference
was 7% when
the patient
took
one drug,
16.7% when the patient
took
two drugs,
66.7%
when
the patient
took
three
or four drugs,
and 100%
when
the patient
took five drugs.
The average
number
of drugs administered
to the patient
was 1.76. An analysis of the affected
tests
showed
that about
one-third
Clinical
Chemistry
Service,
Clinical
Pathology
Department,
Bldg.
10, Room 2C-306,
Warren
G. Magnuson
Clinical
Center,
National
Institutes
of Health,
9000 Rockville
Pike, Bethesda,
MD
20892.
Present
address:
Department
of Pathology,
The Johns Hopkins
School of Medicine,
Clinical
Chemistry
Division,
600 N. Wolfe St.,
Meyer,
B-125, Baltimore,
MD 21287.
for correspondence.
Fax 301-402-1612.
Received
June 28, 1994; accepted
August
31, 1994.

CLINICAL

CHEMISTRY,

tests.

Vol. 40, No. 11,

1994

one-fourth

were

coagulation),
Since

and

1971,

hematology

the

there

tests

remainder
were
been a great
in

has

in the number
of drugs
available
to physicians
for
patient
care and in information
about the effect of these
drugs
on clinical
laboratory
tests.
A crude
index of this
is the number
of pages
in the three
editions
of the drug
ifie

by Young
et al. (2-5);
the first edition
in 1973 had
the second
edition
in 1975 had 432, the third
edition
in 1990 had 937, and the 1991 supplement
to the

262 pages,

third
edition
had 360. However,
the most
important
aspect
of interference
with laboratory
tests
is the detriment to the care of the patient
and added cost for health
care. To the best of our knowledge,
there are no data to
address
these
important
questions.
Nonetheless,
it is
important
for the laboratorian
to be sensitized
to the
interference
phenomena
and to know
how to identify
and to resolve
interference
problems.
In this review,
we

our definition

tion and
resources

of interference,

recommend
resolution
available

and evaluate

classify

an approach

to the

of interference
problems,
to resolve
interference

identifica-

and outline
problems.

Definition

Endog-

in body

with

1996

chemistry

and

crease

give

accufor

tests,

counts

interferences,

results.

(or) exogenous
interfere
with
Further,

of/drug

urinalysis

two common
definitions
applied
to laboratory

There are at least

interference
when
first

definition,

atic

error

nent

interference

does

measuring

exclude

the

by

itself

effect

of

absence
of the analyte
as well
ences of hemoglobin,
bilirubin,
common
definition,
the effect
step

in

lytic

activity

the

i.e.,

substances

determination

of the
other

The
system-

by a sample
compoproduce
a signal
in the
(6, 7), we find problematic
because
it

not

system
to

is the

caused

of measurement

which

seems

Analytical

of the term
analysis.

of the

analyte
than

reagents
or detection
system
Thus,
the primary
difference

(8),
the

the

interferent

in

the

as the common
interferand lipemia.
The second
of a substance
upon
any
concentration

includes

analyte

of the
between

or cata-

nonspecificity,
reacting

with

these

the

method.

analytical

two

defini-

tions
is the exclusion
of nonspecificity
in the first and
inclusion
in the second.
Further,
the first definition
is
more
commonly
used in Europe
and the second
more
commonly
in the US.
We prefer the second
definition
but have modified
it to
expand
the concept
of interference.
Our definition
of
interference
for this review
is: the effect of a substance
present
in the sample
that alters the correct
value
of the
result,
usually
expressed
as concentration
or activity,
for an analyte.
When
interference
is present,
we be-

Lieve it is important
Ices not depend

to determine
on the analyte,

snalyte-dependent
?nce, respectively
?nce

and
(9-11).

is encompassed

lefinition

but

Evaluation
The

step

is to

interference

mon detection
)f the

by

patients

by

our

the

first.

interferinterfer-

definition

and

the

second

of Interference

establish
with

systems

it depends
or
referred
to as

analyte-independent
Analyte-independent

and Classification

first

possible

not

whether
herein

a detection
laboratory

are the delta

system

tests.

check,

Two

for
corn-

a comparison

current

result
with a previous
one, and
letection
of abnormal
blanks
or unusual
reaction
rates
by the analyzer
(12).
Further,
many
times
clinicians
all the laboratory
with concerns
about inconsistent
resuits. We suggest
that
one consider
these
calls
an opporbunity
to investigate
a problem;
this is not only good
practice
of laboratory
medicine,
but also good public
relations.

Several

experimental

options

are available

to resolve

luestions
of interference.
The best means
to establish
snd quantify
an interference
with a given
method
is to
:letermine
the analyte
by another
method
that
would
riot demonstrate
the interference.
A Reference
Method
is appropriate
for this analysis,
provided
the putative
interferent
has no effect
on the Reference
Method.
It is
riot often
evident
that one has such a method,
nor do

ative

interferences

with

added

of the final result

original
result
(before
supplementation).
For a given
method,
the interference
for each analyte
is shown
as a function
of the
interferent.
The scale typically
ranges
from 0% to 200%.
An envelope
of 10% around
the 100% line indicates
a
zone of noninterference.
Analyses
whose
final results
deviate
no more than
3% from the original
result
are
deemed
to be free from interference
and are represented
by an asterisk
on the graph.
Glick
et al. summarized
their results
by ranking
analyzers
according
to the freedom from interferences
(14). This method
for summathing
interference
data is effective
for identification
of
interference
among
instruments
and
methods.
However, the results
are limited
to endogenous
interference
assessed
at one analyte
concentration
or activity.
Another
way
of classifying
interference
is by the
source
of the interferent.
The aberrant
concentration
or
activity
of the substance
in the sample
may arise from
endogenous
sources,
may be exogenous
but inherent
in
the sample,
or may be a compound
added
at the time of
specimen
collection.
Endogenous

as the

interferent

percentage

divided

by

the

Interferents

by atomic
absorption
is unaffected
by citrate,
unlike
most compleximetric
methods
for calcium,
but most
labDratories
do not have
the atomic
absorption
method
readily
available.
A second
approach
is to analyze
the sample
with
a
different
method.
If the results
for the two methods
are

There
are four major
sources
of endogenous
interferents for patients
samples:
hemolysis,
iipemia,
bilirubinemia,
and paraproteinemia.
Hemolysis.
In vivo or in vitro hemoiysis
produces
an
increase
in the hemoglobin
concentration
of serum
or
plasma.
The major cause of hemolysis
occurs
in vitro at
the time of specimen
collection
or transport.
Serum does
not look hemolyzed
until the concentration
of hemoglobin reaches
0.2 g/L (16); jaundiced
serum
has the potential to hide even greater
concentrations
of hemoglobin

different,

(17).

most

laboratories

employ

them.

For

example,

calcium

this suggests
that an interferent
is present.
If
are similar
for the two methods,
either
no
interferent
is present
or the interference
by the interrerent
is essentially
the same
for the two methods.
The
greater
the
dissimilarity
between
the methodological
principle
of the two methods,
the greater
the likelihood
that the interference
will be detected.
For example,
an
interferent
with the Jaff#{233}
reaction
(alkaline
picrate)
for

the results

the determination
of creatiine
is unlikely
to interfere
with the enzymatic
reaction
using
creatinine
amidohyirolase.
A third
approach
is to add serially
higher
concentrations of the putative
interferent
to aliquots
of the same
material.
This
approach
determines
the presence
and
diegree of interference
with the given
method.
One may
snalyze
the results
by linear
regression
analysis.
If the
slope differs
significantly
from zero as determined
with
the Students
t-test,
interference
is present.
A simple
way to determine
the t-value
is to divide
the slope by the
standard
error
of the
slope.
P <0.05 indicates
interfernce is present,
and the magnitude
of the interference
is
given by the slope.
The above
method
is used to compare
common
groups
f analytes
or methods.
Glick et al. have published
exbensive
reviews
of the effects
of endogenous
interferents
n many
analyzers
(13-15).
These
graphs
show the rel-

Hemolysis

and

hemoglobin

have

a direct

effect

on

several
tests.
Hemoglobin
at a concentration
of 5 g/L
inhibits
serum
lipase
activity
by 50% (18); hemoglobin
also
causes
a significant
negative
interference
with
methods
for bilirubin
(19, 20). Of course,
the cellular
components
of erythrocytes
may in fact raise
the analyte concentration
as a contaminant,
given
that the intracellular
concentration
of lactate
dehydrogenase
is
160-fold
greater,
potassium
22-fold
greater,
and magnesium
3-fold
greater
than
in plasma
(20,
21).
This
increases
the concentration
of lactate
dehydrogenase
by
670 U/L, potassium
by 3 mmol/L,
and serum
magnesium
concentration
by 0.05 mmol/L
for each
1 g/L of
hemoglobin
(20-22).
However,
these
elevations
in arialytes are false only if the hemolysis
occurs in vitro. With
in vivo hemolysis,
such as can occur
with paroxysmal
nocturnal
hemoglobinuria,
these
elevations
are present
in vivo
(23).
Other
analytes,
such
as aspartate
aminotransferase,
alanine
aminotransferase,
creatine
kinase
(CK),
and
iron,
may
also be affected
(24).
The
degree
of interference
depends
on the methodology
and
individual
instrumentation.
Sonntag
has
published
guidelines
on when
not to report
a result,
based
on the
hemoglobin
concentration
(25).
In vitro
hemolysis
cccurs with
mechanical
destruction,
freezing,
hyperosCLINICAL

CHEMISTRY,

Vol. 40, No. 11, 1994

1997

motic
shock,
detergents,
exhaustion
of glucose
sample,
and increased
fragility
due to inherited

in the

diseases

(26).

One way to define

the problem
of hemolysis
is to
estimate
the hemoglobin
concentration
on all samples.
Many
modern
instruments
detect
the hemoglobin
of a
sample
blank.
Sonntag
and Glick studied
the efficacy
of
the Hitachi
serum index for hemoiysis
and found that it
correlated
well with the hemoglobin
concentration
(27).
The value
of these
serum
indices
is that one can detect
and flag hidden
hemolysis.
The causes
of interference
due to hemolysis
are not
always
apparent,
but some studies
have been done. Hemoglobin
begins
to absorb
around
340 nm and thus
affects
methods
based on the absorbance
properties
of
NADH
or NADPH
(28). When
hemoglobin
becomes
oxidized
to form methemoglobin,
the absorbance
at 340
nm decreases
(28). Other constituents
besides
hemoglobin cause effects
when
hemolysis
occurs.
Hemolysis
affects the determination
of CK activity
through
the release of adenylate
kinase,
which
catalyzes
the following
reaction:
2 ADP
CK

catalyzes

the

Creatine

phosphate

ATP

-*

following

AMP

reaction
+

ADP

Further,
adenyiate
kinase
electrophoretic
determination
nyiate
kinase
migrates
cathodal

-*

(29):
creatine

ATP

may
interfere
with
the
of CK isoenzymes.
Adoto CK-MM
and may be

spuriously
interpreted
as macro
CK-2 (30, 31). Hemolysis
causes
both
positive
and negative
interferences
with bilirubin
and theophylline,
as determined
with the
Hitachi
717 and Boehringer-Mannheim
reagents
(32).
Finally,
substitutes
for erythrocytes
may cause
interferences.
Cross-linked
hemogiobins
are acellular
and interfere with
common
methods
for alanine
and aspartate
aminotransferases,
bilirubin,
and alkaline
phosphatase
(33).
Lipemia.
Lipemia
is the result
of circulating
chyiomicrons, which
are large lipid particles.
They cause interferences
by turbidity
or light
scattering
and volume
displacement.
Light scattering
raises
the absorbance
of
the blank
and thereby
reduces
the operating
scale
for
colorimetric
methods.
Likewise,
light
scattering
can
have profound
effects
on methods
based on light scattering, such as nephelometry.
Volume
displacement
affects
most methods
by displacing
water,
thereby
giving
rise
to such problems
as pseudo-hyponatremia.
Methods
are
not affected
by lipemia
when the determination
is made
in the aqueous
fraction,
such as methods
involving
ionselective
electrodes.
The reagents
for many
enzyme
methods
contain
detergents
that clear the lipoproteins
and alter
the absorbance
rate.
For example,
lipemia
may have a spurious
effect on the determination
of triglycerides,
since
enzymatic
methods
employing
lipase
clear the turbidity
from the lipemia
during
the reaction
(34). Lipemia
interferes
with determinations
of glucose,
phosphorus,
total
biirubin,
uric acid, and total
protein
1998

CLINICAL

CHEMISTRY,

Vol.40,

No. 11, 1994

by the Beckman
Synchron
CX5 (35). The interference
caused
by lipemia
has been difficult
to quantify
with a
measurable
entity
(27). In general,
the three
predominant interferences
are electrolyte
exclusion,
partitioning of nonpolar
analytes
(chromatographic
and immunoassays),
and light scattering
(36).
Bilirithin.
Elevated
concentrations
of bilirubin
are another
source
of endogenous
interference.
Part
of the
interference
arises from the spectral
properties
of bilirubin,
another
part
from
bilirubins
ability
to react
chemically
with reagents.
Bilirubin
reacts
with peroxidase-catalyzed
reactions,
as are used
in the detection
systems
for glucose,
cholesterol,
and uric acid (32). Bilirubin
causes
a negative
interference
with
the Jaff#{233}
method
for creatinine,
as measured
with the Paramax
(Baxter),
the
Dimension
(DuPont),
or the
Chem-1
(Bayer-Technicon)
analyzers
(37,
38).
Also,
bilirubin
causes
a negative
interference
with
the
enzymatic
method
for creatinine
that is based on creatinine,
creatinase,
and sarcosine
oxidase
and detects
H2O2 with peroxidase
(39). Bilirubin
interferes
with determinations
of
uric acid, cholesterol,
and triglycerides
that use peroxidase-coupled
reactions
(40). Part
of the interference
is
chemical,
thought
to be caused
by the destruction
by
bilirubin
of a reaction
intermediate,
which
is partially
corrected
by adding
ferrocyanide
to the reagents
(40).
Another
part
is spectral,
related
to the strong
absorbance
of bilirubin
(conjugated
and unconjugated)
between
400 and 540 nm (40). Some of the analyzers,
such
as the Kodak
Ektachem,
avoid
the problem
by using
a
chromogen
that reacts
with hydrogen
peroxide
but not
with bilirubin
(41). Bilirubin
has the ability
to react
with electrophilic
reagents
such as diazotized
sulfanilic
acid, but it may also react
with
intermediates
in the
reaction
catalyzed
by hydrogen
peroxide
and thus yield
a negative
edge,
the

(42). To the best of our knowlof bilirubin

in serum
or plasma
between
conjugated
or unconjugated
is not a significant
factor
for interference.
However,
urine
would
contain
only
conjugated
bilirubin
(monoand diglucuronide);
unconjugated
and, thus,

interference
proportion

and delta bilirubin

are bound
to albumin
are excluded
by the glomerular
membrane.
Paraproteinemia.
Elevated
protein
concentrations
cause interferences
in many
assays.
Paraproteins
circulate as the result
of multiple
myeloma
or similar
diseases.
The concentrations
of immunoglobulin
(Ig) are
greatly
increased,
not only for a particular
class,
but
also for a particular
idiotype.
Sometimes
the viscosity
of
serum
is greatly
increased,
which
interferes
with proper
volume
sampling.
Sonnenblick
et al. reported
that paraproteins
interfered
with
phosphate
determination
by
the ferrous
sulfate/ammonium
molybdate
method
(43).
Paraprotein
interference
is method
specific;
that is, they
interfere
with
urea
as determined
with
o-phthalaldehyde but not with the change
in conductivity
as urea is
converted
to ammonia
and bicarbonate
by urease
(44).
However,
others
have
reported
a positive
interference
by paraprotein
with this latter
method
(45). Immuneglobulins
may inadvertently
bind other analytes
in serum. For example,
IgG may complex
with CK-BB
and

macro-CK
1, which
has a longer
circulating
halflife than CK and a concomitant
elevated
serum activity
(45). Immunoglobulins
can also form macro
complexes
with lactate
dehydrogenase
and amylase
(45). Further,
Hirano
et al. reported
an IgG that
inhibited
lactate
dehydrogenase
activity
and thus caused
a negative
interference
(46).
Immunoglobulin
may cause
negative
interferences
by complexing
with thyroxine
measured
by RJA but may also cause a positive
interference
when
determined
by a double-antibody
or solid-phase
assay
(47). IgG and f3-endorphin
may show structural
homologies,
and high
concentrations
of IgG interfere
with
immunoassays
for -endorphin
(48).
Others.
Other
endogenous
substances,
especially
in
abnormal
concentrations,
may also cause interferences.
The chloride
electrode
from Boehringer-Mannheim
is
sensitive
to bicarbonates.
Thus,
when
the bicarbonate
concentration
is low (<15 mmol/L),
chloride
is underestimated,
but when the bicarbonate
concentration
is high
(>40
mmoliL),
chloride
is overestimated
(49).
Ketone
bodies
and protein
interfere
with the Jaff#{233}method
for
creatinine,
but these interferences
are variable,
depending on the method
used (50-52).

form

Exogenous

Interferents

Exogenous

interferents

may

effect

a positive

or nega-

tive bias in the result.

A positive
interference
increases
the value
of the result
to greater
than the true concentration
of the analyte;
a negative
interference
acts in the
opposite
direction.
For example,
cephalothin
and cefoxitin cause a positive
interference
with the Jaff#{233}method
for creatinine
(53, 54), but ascorbic
acid causes
a negative interference
with glucose
oxidase
methods
for glucose (55). The type of interference
that will arise,
positive or negative,
is not readily
obvious
from the nature
of the compound
or material.
Understanding
how and
why such interferences
occur
requires
knowledge
of the
chemistry
of the exogenous
interferents
and the analytical reaction.
Here, we present
an approach
to analyzing
resolving
with
suitable
interferents
and

tions,

additives,

problems
with
exogenous
interferents
examples.
We have
grouped
exogenous
into the following
four categories:
medica-

materials,
and matrix
effects.
medications
interfere
with the determination
of analytes.
Any medication
given
a patients
by any route (e.g., intravenous,
oral, subcutaneous)
could
interfere
with
analytical
methods.
Most
medications
are low-molecular-mass
(<1 kDa)
organic
compounds
with functional
groups
that run the gamut
from nonpolar
to ionic. Medications,
being
designed
to
be biologically
active
and given
in high pharmacological
doses, have a high probability
of reacting
with analytes
or reagents.
Metabolites
of drugs
may cause
interfer-

Medications.

test
Many

ences
and are as important
to consider
as the parent
drug.
Many
drugs
are chemically
converted
to morepolar forms,
especially
those that are lipid-soluble
(56).
Most of the transformations
take
place
in the hepatic
microsomes
and consist
of the basic reactions
of oxidation; reduction;
ester and amide
hydrolysis;
conjugation
with
glucuronide,
glycine,
or sulfate;
and
acetylation

(56).
tions
more

The oxidation,
reduction,
and conjugation
reacproduce
metabolites
that have the potential
to be
reactive
than the parent
compound.
Two different

outcomes
metabolite
but react

can occur
with
regards
to interference:
the
may have
altered
pharmaceutical
function
in the analytical
system
similar
to the parent

compound,

or the parent
compound
does not react in the
system
but the metabolite
does.
Oxidation,
with
the formation
of hydroxy
(alcohol)
and aldehyde
groups,
often
makes
drugs
more reactive
in chemical
systems,
as does the addition
of an acetyl
group.
Phenytom and its major
metabolite,
5-(p-hydroxyphenyl)-5phenylhydantoin,
create
false-positive
reactions
in a
method
for barbiturates
(57).
The hydroxylation
and
demethylation
metabolites
of cyclosporin
G show
5-100%
of the activity
of the parent
drug, depending
on
whether
they are measured
by RIA, fluorescence
polarization,
or enzyme
immunoassay
(58).
Furosemide
by
itself does not interfere
with the Jaff#{233}
method
for creatmine,
but furosemide
given
in high doses
leads
to formation
of a metabolite
that causes
a negative
interference (59). The interferences
caused
by the metabolites
of drugs may be more difficult
to detect
than those of the
parent
compounds,
because
the concentrations
of the
metabolites
are more unpredictable
and the metabolites
are often
unknown.
Analysts
typically
need
another
method
based
on a principle
different
from the original
one to detect
interference
by a metabolite.
Lastly,
fillers
and additives
in the formulation
of the medication
may
be a source
of interference.
Endogenous
interferents
may be administered
externally,
as a medication.
As a replacement
for a blood
transfusion,
cross-linked
hemoglobins
are being
tested
as a source
of oxygen-carrier
replacement.
These
hemoglobin
compounds
are given
intravenously
and
are not
analytical

separated
from serum
by centrifugation
effect is similar
to in vivo hemolysis.
preparations
have
been
shown
to
with

(33). Thus,
the
These
hemoglobin
cause
interference

alkaline

phosphatase,
alanine
aminotransferase,
aminotransferase,
total
bilirubin,
calcium,
and phosphorus
by several
methods
and with activated
partial
thromboplastin
time and prothrombin
time by
one method
(33, 60, 61).
Additives.
Materials
are added to the blood collection
tube for anticoagulation,
for inhibition
of glycolysis,
and
to seal the stopper.
All of the common
additives
for
aspartate

anticoagulation
(heparin,
EDTA,
citrate,
and oxalate)
have been documented
to interfere
with methodologies
for many
analytes
(4, 5). Fluoride
and
(or) iodoacetate,
frequently
added
as antiglycolytic
agents,
also may interfere
with methodologies
for a variety
of analytes
(4,
5). Fluoride
inhibits
the enzyme
systems
involved
in
glycolysis,
and iodoacetate
inhibits
CK activity.
Lastly,
the stopper
in some evacuated
collection
tubes
is sealed
with a silicone
coating.
The ensuing
contamination
of
the specimen
with silicone
has been shown
to interfere
with the determination
of ionized
magnesium
with an
ion-selective
electrode
(62).
Test materials.
Materials
used with
instruments
in
the laboratory
may contain
added substances
that cause
CLINICAL

CHEMISTRY,

Vol. 40, No. 11, 1994

1999

interference

with

analysis.

Control,

calibration,

and

proficiency
survey
materials
typically
are made from a
defibrinated
plasma
base. Naturally
occurring
analytes
may be removed
from the plasma
base and analytes
of
interest
added.
Thus, these
materials
may contain
very
high
concentrations
of analytes
such as bilirubin
and
glucose
as well as multiple
drugs. These exogenous
compounds
commonly
interfere
with methods,
but the effect
may go unnoticed
for calibrators
and control
materials,
because
the exogenous
compounds
are in constant
concentration
throughout
the lot. This situation
is more
problematic
with proficiency
survey
material
that clinical laboratories
use to compare
their analytical
results
with those of other laboratories.
When
two methods
for
a given
analyte
have
different
responses
to the same
interferent,
they
will
produce
different
results,
even
when
they
might
have
given
identical
results
in the
absence
of the interferent.
For example,
the therapeutic
drug
monitoring
proficiency
survey
material
distributed
by the College
of American
Pathologists
contains
multiple
drugs,
including
lithium.
The
survey
results
showed
that one ion-selective
electrode
for the determination
of lithium
gave consistently
higher
results
than
other electrodes
with this survey
material
(63). Analysis of the proficiency
survey
data showed
a relationship
between
the determined
lithium
concentration
and the
quiidine
concentration
(63).
Thus,
in this case,
the
polypharmacy
of the survey
material
identified
an interferent
(quinidine)
with one ion-selective
electrode
for
lithium.
Salicylate
added to proficiency
survey
material
produces
a falsely
elevated
value
for theophylline
(63).
Lastly,
ascorbic
acid effects
a negative
interference
with
the glucose
oxidase
method,
but not the hexokinase
method,
for glucose
(64, 65).
Thus,
glucose
oxidase
methods
show a low bias vs hexokinase
methods
when
proficiency
survey
materials
contain
ascorbic
acid. The
above
examples
emphasize
that
the results
of proficiency
surveys
are a fertile
tification
of interference

source
among

for the selective

idenmethods
for a given

analyte.
Matrix
effects.
The types
above will usually
be ascribed
times
the processing
of the
change
in results
that cannot

of interference
described
to a specific
cause.
Somesample
may
introduce
a
be ascribed
to a specific

chemical
constituent.
These
changes
are usually
referred to as matrix
effects.
Thus, a matrix
effect is one in
which
the
preparation
of the
material
introduces
changes
in the solution
that alter
the values
obtained
with the assay.
This problem
occurs when the specimen
has undergone
special
processing.
The change
in the
value
for the analyte
is due to unknown
or poorly
characterized
factors
in the sample
(66). These
factors
have
been ascribed
to the source
of the solution
(e.g., equine,
bovine,
human),
preservatives
(sodium
azide or antibiotics),
stabilizers
(dithiothreitol
or acetylcysteine),
or
physical
changes
such as lyophilization
or freezing
(66).
Enzyme
activities
as well as binding
characteristics
of
ions are sensitive
to changes
in the solute-solvent
interactions.
Processing
and lyophilizing
serum
pools affect the result
for cholesterol
for certain
analyzers
and
2000

CLINICAL

CHEMISTRY,

Vol. 40, No. 11, 1994

have interfered
with the standardization
process
for this
analyte
(67-69).
Thus,
matrix
effects
are not a problem
with fresh human
serum
but may occur in control,
calibration,
and proficiency
survey
materials.

Mechanistic

Perspective

The number
of agents,
both endogenous
and exogenous, that interfere
with the determination
of analytes
is great.
The problem
is compounded
by the myriad
reagents
and instruments
used for each method.
Besides
compilations
of data in files and computer
programs,
an
ability
to predict
possible
and probable
interferences
is
needed.
To predict
interferences,
one needs
a way to
understand
them.
A typical
way to understand
interactions
in chemistry
is through
elucidating
the pathway
taken
by the reactions.
Such pathways
are known
as
mechanisms.
Mechanisms
may be global
(generic)
or specific,
analytical
or chemical.
A global
mechanism
describes
the
overall
pathways
that are shared
by many interferents
and methods.
Specific
mechanisms
describe
pathways
for an individual
substance
or class of substances
with a
specific
type
of reagents.
Analytical
mechanisms
describe
how
an interferent
exhibits
its behavior.
The
term chemical
mechanisms
describes
the chemical
reactions
in detail,
with use of principles
of organic,
inorganic,
biochemical,
and physical
chemistry.
The global
mechanisms
of interference
are, for the
most part,
general
analytical
mechanisms.
We divide
them
into whether
the interference
pathway
involves
the analyte
or products
of the analyte,
or whether
the
interference
pathway
excludes
the analyte
or its products-that
is, analyte-dependent
and
analyte-independent interferences,
respectively.
One can distinguish
between
analyte-dependent
and
analyte-independent
interference
by determining
the
analyte
concentration
at four or more concentrations
of
analyte
and of interferent
and
then performing
multivariate
regression
analysis
(9-11).
When
an interferent
causes
a positive
interference
in the analyte-dependent
mechanism,
it indicates
a nonspecificity
of the method.
Knowledge
of the overall
mechanism
is important,
because
performance
of interference
studies
at only one
concentration
of analyte
for analyte-dependent
interferences may result
in underestimation
of the interference
in clinically
relevant
concentrations
of analyte
and interferent.
Hemoglobin
interferes
with the measurement
of conjugated
bilirubin,
but with a positive
interference
for bilirubin
concentrations
<32.5
molJL
(19 mg/L),
a
negative
interference
for >32.5
mol/L,
and no interference at 32.5 jmol/L
(10). For a pool with a conjugated
bilirubin
concentration
of 32.5 mol/L,
one might
believe there was no interference,
a false conclusion.
Different
investigators
using
different
analyte
concentrations
could
be led
to different
conclusions
and
to
disagreements.
These
investigators
would
be technically
correct
for their narrow
range
of conjugated
bilirubin
concentration
but would
fail
to understand
the
totality
of the problem.
It is always
best to determine
the general

mechanism

for interference

before

postulat-

ng specific
mechanisms
or attempting
corrections.
Inerferences
can become
quite complex,
with more than
ne interferent
acting
at a time.
For example,
both bilrubin
and hemoglobin
simultaneously
interfere
with
m enzymatic
method
for creatinine,
yielding
a negative
)las for hemoglobin
and a positive
bias for bilirubin
70).

A mechanism
describes
a pathway
from an initial
to a final state.
With interferences,
a mechanism
lescribes
the pathway
from the initial
concentrations
of
Lnalyte,
reagents,
and interferent
to a determined
value
br the analyte,
including
the bias for the interferent.
:nterference
mechanisms
are useful
in predicting
the
legree of interference
and in suggesting
ways of circumrenting
the interference.
Establishing
the interference
nechanism
involves
an understanding
of the chemical
eactions
in the method.
Elucidation
of the mechanism
)athway
usually
requires
the knowledge
of whether
the
nethod
is an endpoint
(equilibrium)
or rate (kinetic)
eaction;
however,
the pathway
for individual
steps
in
;he pathway
often requires
examination
of the kinetics
)f the reactions
(71).
Knowledge
of the chemical
reaction
or reactions
inrolved in the determination
of the analyte
is useful
in
tate

leveloping
a structure-function
relationship.
The struc;ure-function
relationship
is a common
theme
used in
)harmacology
to understand
and predict
pharmacologcal action
of compounds
and potential
uses of new medcations.
It can also be useful
in understanding
interfernces.
As an example,
consider
the interference
of cepha
mtibiotics
with the Jaffe
method
for creatinine.
Cepha
mtibiotics
form a colored
complex
with
picrate
that
Lbsorbs
in the same
region
as the creatiine-picrate
omplex
(72).
Although
the rate
for the reaction
of
)icrate
with cefoxitin
is -5% of that for creatinine,
the
nolar
absorptivity
and
the concentration
of cefoxitin
re usually
high enough
to significantly
bias the readngs
for the creatinine
assay.
Certain
aspects
of this
nterference
are unclear.
Cephalothin,
a cephalosporinype antibiotic
structurally
similar
to cefoxitin,
interres
positively
with some methods
for creatinine,
negthvely
with
others
(73,
74).
At first
glance,
there
ippears
to be no common
ground
between
the interfernce caused
by cephalothin
and that caused
by cefoxitin.
n fact, cephalothin
reacts
with picrate
similarly
to the
-ay cefoxitin
does (absorbing
light
with a similar
ab;orbance
band),
but more
quickly;
however,
once the
naximum
absorbance
has been reached,
after -45 s, the
tbsorbance
decreases
(75). The interference
is positive
vhen the readings
for the creatinine
assays
are made
luring
the period
of increasing
absorbance
for the cephilothin-picrate
reaction,
and
negative
when
read durng the period
of decreasing
absorbances.
In terms
of structure,
it is not apparent
why some
ephalosporin
antibiotics
interfere
and others
do not
76). If not all compounds
of the same class react, then it
s difficult
to propose
a general
mechanism.
Further
tudy
of a large group
of cephalosporin-like
antibiotics
ias shown
that all react with picrate
to form a complex
hat absorbs
light
with
a peak
similar
to that of the

creatinine-picrate
complex;
interference
depends
on
whether
the reaction
rate is fast enough
to be comparable with that of creatinine
and picrate
(77).
Studies
of mechanism
showed
that the ketone
groups
on creatinine
and the cephalosporin-like
antibiotics
react with the picrate
(77,
78). This reaction
is accentuated when
the ketone
group
is adjacent
to a nitrogen.
Thus
two different
reaction
rates
may be observed-a
fast reaction
for a straight-chain
ketone
and
a slow
reaction
(at around
the same
rate as creatinine)
for a
ketone
member
in a fourto six-member
ring
(78).
Cefoxitin
demonstrates
both
fast
and
slow
reaction
rates,
due to the straight-chain
ketone
and its lactam
ring, respectively;
this is illustrated
by the loss of the
slow reaction
when aged solutions
of cefoxitin
that, presumably,
have
lost their
lactam
rings
are used
(79).
This knowledge
of structural
components
allows
one to
predict
which
compounds
have a potential
to interfere
with the Jaff#{233}
reaction
for creatinine.
The enzymatic
methods
for creatinine
are not immune
to interferences
either.
In the sarcosine
oxidase
method,
creatinine
is converted
to creatine,
the creatine
to sarcosine
and urea; the sarcosine
is oxidized
to glycine, formaldehyde,
and hydrogen
peroxide
by sarcosine
oxidase,
and the hydrogen
peroxidase
is determined
by
reaction
with a leuco dye (80). Spindler
et al. reported
a
positive
interference
for creatinine,
as determined
with
the Ektachem,
in patients
receiving
lidocaine
(81). The
liver metabolizes
lidocaine
into 4-hydroxy-2,6-xylidine
and N-ethylglycine,
the latter being
almost
identical
to
sarcosine
(N-methylglycine),
differing
by only
one
methyl
group
(82). Because
the Ektachem
does not dilute the sample,
the N-ethylglycine
may be present
in
higher
concentration
in the reaction
than
that found
with traditional
wet chemistries,
enough
for the crossreactivity
with sarcosine
oxidase
to form a positive
interference
(83-85).
Other
mechanisms
can be based
on physical
principles. Turbid
samples,
such as are found
with
lipemia,
scatter
light and give a significant
baseline
absorbance
(34, 36). This background
absorbance
presents
a problem when there is no sample
blank
or when the turbidity changes
during
the reaction,
as in the enzymatic
determination
of triglycerides
(34). Only a bichromatic
blank
(not a reagent
or serum
blank)
appropriately
corrects the changing
interference
from turbidity
for the
enzymatic
determination
of triglycerides
(34).
Other
mechanisms
are more specific,
such as the similarities
of
reaction
of Br
and Cl
in both colorimetric
and chloride-selective
electrode
methods
because
of the similarity of the atomic
structure
of these
ions (86, 87).
Immunoassays
involve
a special
subset
of macromolecular
binding
interactions.
Immunoassays
include
as
reagent
at least
one antibody.
For convenience,
some
authors
have
divided
the
different
assays
into
six
groups:
classical,
i.e., labeled
analyte
and
limited
concentration
of antibody;
labeled
antibody
with a limited
concentration
of reagent;
direct
detection
of immune
complexes
without
a labeled
reagent
(precipitation,
nephelometric,
and
turbidimetric
immunoassays);
all
CLINICAL

CHEMISTRY,

Vol.40,

No. 11,

1994

2001

in excess,
with one reagent
labeled,
e.g., twosite sandwich
assays;
quantification
of specific
antibodies
with
diluted
test serum
added
to immobilized
excess
antigen;
and
separation-free
(homogeneous)
immunoassays
with either
an increase
or decrease
in label
activity
after
binding
by the analyte
(88).
Cross-reactivity
between
the analyte
and
other constituents
in the sample
represents
the most recognized
form
of interference
in immunoassays.
Unlike
most
chemical
reactions,
which
involve
a change
in covalent
bonds,
the antibody-antigen
interaction
involves
weak
intermolecular
forces
such as hydrogen
bonds and van
reagents

der Waal
interactions.
Antibody-antigen
interactions
(structure)
of the antigen
rather

depend
on the shape
than the chemical
re-

activity.
The plethora
of shapes
assumable
by antibody
variable
regions
gives rise to the vast array of antibodies capable
of binding
with different
biological
constituents
while
still
maintaining
specificity.
Because
the
binding
is noncovalent,
the ligand
bound
by the antibody is unchanged
in most cases.
The variable
regions
in antibodies
are small
enough
to recognize
relatively
small
regions
of their ligands;
thus, small differences
in
shape
between
ligands
have little
effect on the affinity
between
the antibody
and an altered
ligand.
This property allows
one to label
the ligand
with a radioactive
isotope,
such as 1251.
These properties
also represent
one of the weaknesses
of immunoassays,
because
often the differences
between
the analyte
and other chemical
constituents
are small.
One immunoassay
for cortisol
actually
showed
a higher
affinity
for hydrocortisone
sodium
succinate
or acetate
(>700%
cross-reactivity)
and fludrocortisone
acetate
(320%
cross-reactivity)
than
for cortisol
(89).
The sodium
succinate
ester
version
of hydrocortisone
differs
from cortisol
only in that sodium
succinate
is covalently
bonded
through
an ester linkage
with the alcohol
group
off the C21 carbon.
Compared
with cortisol,
the succinate
ester
linkage
increases
the water
solubility
but leaves
the biological
activity
unaltered.
Hydrocortisone
acetate is similar
in structure
to cortisol
except
for having
an ester linked
to the C21 alcohol
group.
Fludrocortisone
is similar
to cortisol
except
that a fluorine
is covalently
bonded
to the C9 carbon.
Fludrocortisone
shows
greater
anti-inflammatory
activity
and markedly
more sodiumretention
potential
than
cortisol.
The antibody
in this
assay
(89) had greater
affinity
for these
three
drugs
than for cortisol,
and their
presence
yielded
a positive
interference.
The structures,
and therefore
the shapes,
of the analyte and interferent
do not have to be closely
identical
for cross-reactivity
to occur. Cyproheptadine,
an antihistamine
(H1-antagonist),
in concentrations
>400
g/L
causes
a positive
interference
with a screen
for serum
tricycic
antidepressants
(90). Cyproheptadine
and the
tricycic
antidepressants
both
contain
a seven-membered
ring with two phenyl-groups
attached,
one opposite from the other,
but the functional
groups
on the
antidepressants
differ from those of cyproheptadine.
The
drugs
show markedly
different
biological
and chemical
2002

CLINICAL

CHEMISTRY,

Vol. 40, No. 11,1994

activities,
but display
structural
similarities
from th
perspective
of the antibody.
With immunoassays,
exogenous
and endogenous
co
pounds
can interfere
with
the detection
system.
Th
Abbott
TDx
uses fluorescence-labeled
antigen
and flu
rescent
polarization.
Ophthalmologists
occasionally
ad
minister
sodium
fluorescein
in fundoscopic
examin
tions; the resulting
concentration
attained
in blood ca
be sufficient
to result
in high background
fluorescenc
with low background
polarization
in assays
for thyrox
ine and triiodothyronine-uptake
(91). In such instances
the analyzer
either
reports
falsely
elevated
values
a
fails to report
any result
(91, 92).
Patients
sera often contain
antibodies
against
mous
immunoglobulins
(referred
to as heterophile
antibodies
such as are used in two-site
immunoassays
(93). Suck
anti-murine
antibodies
can result
in falsely
elevata
values
for (e.g.) thyrotropin
(94). In the two-site
immu
noassay,
the interfering
anti-murine
antibody
binds
t
the immobilized
and labeled
antibodies,
creating
a corn
plex between
the two; the net effect is to bind the sign#{225}
antibody
to the capture
antibody
(95). When the concen
tration
of heterophile
antibodies
is high in relationshil
to the reagents
antibodies,
the interfering
antibod3
forms
complexes
with
only one reagent
antibody
at
time,
and
falsely
positive
responses
are not generated
but the assay
response
may be decreased
(93, 94). Thi
heterophile
antibodies
do not bind to the analyte
bind.
ing site but rather
to either
the Fc or F(ab)2
fragments
in either
case, the binding
affinity
of the reagent
anti
bodies
can be reduced,
affecting
RIAs as well as two-sit
immunoassays
(93).
When
specific
antibodies
are given
to a patient
foi
therapy,
the antibodies
may compete
with immunoas
says for the analyte.
The best example
of this is admin.
istration
of antigen-binding
fragments
(Fab)
derivec
from specific
anti-digoxin
antibodies
(Digibind)
to treat
digoxin
intoxication.
The ability
of Fab to interfere
ir
immunoassays
has been well-documented
(96). The Fal
compete
with
the antibodies
of the assay
system
foi
digoxin,
disturbing
the
competition
of antigen
anc
tracer
for antibody
binding
sites. Using
immunoassay
to measure
significant

digoxin
in samples
containing
Fab will yield
variability
in the results
(97, 98). The freE
digoxin
concentration
is needed
for the clinical
care ol
the patient
and may be achieved
by ultrafiltration
of thE
serum
or with selected
immunoassays
for digoxin
(99).

Information

Sources

There
are several
sources
of information
to assist
thE
laboratorian
with the resolution
of an interference
problem. A starting
point
is the manufacturers
packagE
insert,
in which
the description
of the method
usually
includes
an analysis
of probable
interfering
substances,
The endogenous
interferents
hemoglobin,
biirubin,
and
lipemia
are routinely
evaluated
for interference
by
manufacturers.
In addition,
the manufacturer
usually
evaluates
several
other
compounds
that have been
reported
to interfere
with the assay
or that have a proba.
bility
for interference
with the manufacturers
method-

ology.
Further,
many
manufacturers
provide
a toll-free
telephone
number
so that the laboratorian
can discuss
the problem
with a technical
consultant.
Thus, the manufacturer
of the analytical
test is usually
an excellent
source
of information
about
interference.
Several
publications
have
organized
the vast literature related
to interference
with laboratory
tests.
In the
US, the publications
by Young
et al. (2-5)
are available
in most
laboratories:
three
editions
of the Effects
of
Drugs
in Clinical
Laboratory
Tests,
and a supplement
to
the third
edition.
The information
in this drug
file is
organized
by laboratory
test and drugs,
and
an index
lists all laboratory
tests and drugs.
The information
for
a given
test is delineated
by the type of sample
(serum,
plasma,
urine,
etc.) and type of effect [in vivo (physiological)
or methodologic
(analytical)].
A short statement
identifies
each interference
and cites a reference
to the
literature.
Faced with a questionable
result,
laboratory
personnel
can query
the drug
ifie for possible
interferences
for the test.
If, on the other
hand,
a patient
is
known
to be taking
a particular
medication,
the specific
tests
affected
by that medication
are
identified
in the
publication.
Other
drug
files contain
different
information.
Sonntags book (100) not only gives information
about falsely
increased
or decreased
values
with interference,
it also
presents
information
for cases where
no interference
is
found in the therapeutic
range
but is found in the toxic
range,
or for cases
where
interference
is found
within
the therapeutic
range.
This book also breaks
down the
interference
studies
into the individual
methods,
which
is very helpful
in light of the general
rule of drug interference
that one method
is affected
while
another
is not.
The book by Tryding
and Roos (101)
is arranged
in an
easy-to-read
format
that clearly
demarcates
biological
from analytical
effects;
their
information
is also available on disc for IBM or IBM-compatible
systems.
Salwayss
drug
file (102) separates
the tests by physiological
entity:
bone
and
joint
diseases,
cardiovascular
diseases,
glucose
and
diabetes,
hormones,
kidney-function tests,
liver-function
tests,
therapeutic
drug
monitoring,
and thyroid-function
tests.
Through
the use of
tables,
the references
relating
to the presence
or absence
of biological
or analytical
effects
can easily
be found.
If
interferences
are present,
their positive
or negative
nature is shown.
For each effect, the specimen,
dose, duration,
subjects,
description,
and
summary
of the study
used to provide
the information
are presented.
Siest and
Galteau
describe
the analytical
interference
of 21 common chemistry
analytes
on the basis
of method
and
pharmacological
effects
(103); they also include
the metabolic
background
of the analyte
and the physiological
and pathological
variations
in the reference
interval.
Using
interferographs,
Glick
et al. have
depicted
the
effect
of common
endogenous
interference
on clinical
chemistry
instruments
(13-15).
Lastly,
the National
Committee
for Clinical
Laboratory
Standards
has published
a proposed
guideline
on interference
testing
in
clinical
chemistry
(EP7-P)
in 1986 (8), although
thus far
it has not progressed
to the tentative
stage.
This

guideline
statistical

presents
analysis

protocols

of the

for interference
data, and several

testing
and
appendices

contain
recommendations
of concentrations
to use for
evaluating
many
drugs
for interference,
preparation
of
test pools, and guidelines
for assays
based
on separation
techniques
and immunochemical
principles.
Access
to
one or more of the above
references
should
greatly
facilitate
the resolution
of interference
problems.

Approaches

to Laboratory

Test Interference

Problems

Recommendations
to laboratory
personnel.
Identification and resolution
of an interferent
with laboratory
test
results
are usually
handled
by laboratory
personnel.
The process
usually
starts
when
the laboratorys
monitoring
system
indicates
a problem
or the
patients
physician
calls
the laboratory
because
of an apparent
unusual
laboratory
result.
Interference
due to an endogenous
compound
is usually
detected
by laboratory
personnel
and would
appear
as part of the report.
A drug
taken
by the patient
is frequently
the exogenous
compound interfering
with a laboratory
test result.
Initially,
it is important
to review
a drug
file and to discuss
the
matter
of drug interference
with the patients
physician.
Stopping
the medication
may be possible.
Second,
laboratory
personnel
should
determine
the analyte
by a
different
methodology,
cant difference
between

if available.
If there
is a signifithe two methods,
drug interfer-

ence is likely.
Third,
one can contact
the pharmacy
to
identify
specimens
from other patients
taking
the same
medication,
and ask the pharmacy
for a small
quantity
of the medication.
By adding
the drug to an aliquot
of an
appropriate
specimen
(one that is free of the drug being
studied
and without
lipemia,
hemolysis,
hyperbilirubinemia, etc.) and assaying,
one can determine
whether
the
analyte
result
changes.
This study,
if the results
are
positive,
provides
an answer
that can then
be refined
with more structured
studies,
as discussed
above
(8, 10,
11). If they are negative,
drug
metabolites
may be the
cause of the interference
with the laboratory
test. Thus,
the failure
to document
interference
with
this simple
procedure
does not rule
out this particular
drug
as the
cause of the interference.
Finally,
contact
the manufacturer
of the instrument
or supplier
of the reagents
for
the assay
and ask for the most recent
information
from
other users
about drug interference
with the assay.
Recommendations
to manufacturers.
Manufacturers
of instruments
and analytical
methods
are required
to
include
detailed
and specific
information
on interference
in their package
inserts,
their method
sheets,
and their
510K submission
to the Food and Drug Administration.
The information
should
include
the basic design
of the
interference
study,
and the interference
should
be reported as unit of analyte
per unit of interferent,
with an
indication
of the uncertainty
of the measure.
Reporting
the interference
in this manner
allows
a laboratory
scientist to interpret
the value of the analyte
in light of the
degree
of interference.
For example,
what if the creatinine
is elevated
and
biirubin
is present
in sufficient
concentrations
to cause
a negative
interference?
One
could
stifi derive
useful
information
if one knew
the
CLINICAL

CHEMISTRY,

Vol. 40, No. 11, 1994

2003

effect of the bilirubin

and could calculate
the true
of creatinine.
Manufacturers
must
document
interference,

value
espe-

cially
as users
raise
suspicions.
Further
study is necessary
to characterize
fully
the
interferences
and,
if
possible,
to work out the mechanism.
During
test development,
manufacturers
should
strive to minimize
interferences.
Just
as important,
manufacturers
must
improve
detection
of interferents
and provide
advice
and
support
for the resolution
of customers
interference
problems.
References
1. Munzenberger
P, Emmanuel
S. The incidence
of drug-diagnostic test interferences
in outpatients.
Am J Hosp Pharm
1971;28:
786-91.
2. Young DS, Thomas DW, Friedman
RB, Pestaner
LC. Effects of
drugs on clinical
laboratory
tests. Clin Chem 1972;18:1041-303.
3. Young
DS, Pestaner
LC, Gibberman
V. Effects
of drugs on
clinical
laboratory
tests. Clin Chem 1975;21:1D-432D.
4. Young
DS. Effects of drugs on clinical
laboratory
tests, 3rd ed.
Washington,
DC: AACC
Press,
1990:937p.
5. Young DS. Effects of drugs on clinical
laboratory
tests, 3rd ed.,
1991 Supplement.
Washington,
DC: AACC Press, 1991:360p.
6. IFCC provisional
recommendation
on quality
control in clinical
chemistry.
J Clin Chem Clin Biochem
1976;14:270.
7. Buttner
J, Borth R, Boutwell
JH, Broughton
PMG, Bowyer Re.
IFCC approved recommendation
(1978) on quality
control in clinical chemistry.
Part I. General
principles
and terminology.
J Clin
Chem Clin Biochem
1979;18:69--77.
8. Powers DM, Boyd JC, Glick MR, Kotschi
ML, Leteffier
G,
Miller
WG, et al. Interference
testing
in clinical
chemistry;
proposed guidelines.
NCCLS
Document
EP7-P.
Villanova,
PA: National
Committee
for Clinical
Laboratory
Standards,
1986.
9. Kroll MH, Chesler
R. Rationale
for using multiple
regression
analysis
with complex
interferences.
Eur J Clin Chem Clin Biochem 1992;30:415-24.
10. Kroll MH. Analyte-dependent
interference
and multi-interference. Dtsch (las Klin Chem Mitt 1991;22:13-20.
11. Kroll MH, Ruddel
M, Blank
DW, Elm
RJ. A model
for
assessing
interference.
Clin Chem 1987;33:1121-3.
12. Ladenson
JH. Patients
as their own controls:
use of the
computer
to identify
laboratory
error.
Clin Chem 1975;21:164853.
13. Glick MR. Ryder KW, Jackson
SA: Graphical
comparisons
of
interferences
in clinical
chemistry
instrumentation.
Cm
Chem
1986;32:470-5.
14. Glick MR, Ryder KW. Analytical
systems
ranked
by freedom
from interferences.
Clin Chem 1987;33:1453-8.
15. Glick MR, Ryder
KW. Interferographs:
a users guide
to
interferences
in clinical
chemistry
instruments.
Indianapolis:
Science Enterprises
Inc., 1987.
16. Behrendt
H. Chemistry
of erythrocytes.
Springfield,
IL:
Charles
C Thomas,
1957.
17. Caraway
WT. Chemical
and diagnostic
specificity
of laboratory tests. Am J Clin Pathol
1962;37:445-64.
18. Henry
RJ, Sobel
C, Berkman
S. On the determination
of
pancreatic
lipase in serum.
Clin Chem 1957;3:77-89.
19. McGann
CJ, Carter RE. The effect of hemolysis
on the Van der
Bergh reaction
for serum bilirubin.
J Pediatr
1960;57:199-203.
20. Meites
5, Hogg CK. Direct spectrophotometiy
of total serum
bilirubin
in the newborn.
Clin Chem 1960;6:421-8.
21. Eke RJ, Hosseini
JM. Magnesium
content
of mononuclear
blood cells. Cliii Chem 1985;31:377-80.
22. Brydon NG, Robert LB. The effect of hemolysis
on the determination
of plasma
constituents.
Clin Chim Acta 1972;41:435-8.
23. Blank
DW, Kroll
MH, Ruddel
ME, Eke RJ. Hemoglobin
interference
from in vivo hemolysis.
Cliii Chem 1985;31:1566-9.
24. Laessig
RH, Hassemer
DJ, Paskey
TA, Schwartz
TH. The
effect of 0.1 and 1.0 percent
erythrocytes
and hemolysis
on serum
chemistry
values.
Am J Clin Pathol
1976;66:639-44.
25. Sonntag
0. Haemolysis
as an interference
factor in clinical
chemistry.
J Clin Chem Cliii Biochem
1986;24:127-39.
2004

CLINICAL

CHEMISTRY,

Vol. 40, No. 11,

1994

26. Guder
in clinical

WG. Haemolysis
chemistry

as an influence
and interference
factor
[Editorial].
J Clin Chem
Clin Biochem

1986;24:124-6.
27. Sonntag
0, Glick MR. Serum-Index
und Interferograin-ein
neuer Weg zur Prufung
und Darstellung
von Interferengen
durch
Serumchromogene.
Lab Med 1989;13:77-82.
28. Fonseca-Wolheim
FD. Haemoglobin
interference
in the bichromatic
spectrophotometry
of NAD(P)H
at 340/380nm.
Eur J
Clin Chem Clin Biochem
1993;31:595-601.
29. Greenson
JK, Farber
SJ, Dubin SB. The effect of hemolysis
on
creatine
kinase determination.
Arch Pathol
Lab Med 1989;113:
184-5.
30. Lott JA, Stang ,JM. Serum
enzymes
and isoenzymes
in the
diagnosis
and differential
diagnosis
of myocardial
ischemia
and
necrosis.
Clin Chem 1980;26:1241-50.
31. Lee KN, Csako G, Bernhardt
P, Elm RJ. Relevance
of macro
creatine kunase type 1 and type 2 isoenzymes
to laboratory
and
clinical
data. Clin Chem 1994;40: 1278-83.
32. Perlstein
MT, Thibert
RJ, Watkins
R, Zak B. Spectrophotometric study of bilirubin
and hemoglobin
interactions
in several
hydrogen
peroxide
generating
procedures
[Abstract].
Clin Chem
1977;23:1133.
33. Christenson
RH, Gregory
LC, Dick S, Mackenzie
C, Trump
BF. Hemoglobin
based blood substitutes:
interferences
with routine chemical
tests [Abstract].
Clin Chem 1993;39:1129.
34. Sampson
M, Ruddel ME, Elm RJ. Effect of specimen
turbidity
and glycerol
concentration
on nine enzymatic
methods
for triglyceride determination.
Clin Chem 1994;40:221-6.
35. Randall
AG, Garcia-Webb
P, Beilby JP. Interference
by haemolysis,
icterus
and lipidemia
in assays on the Beckman
Synchron
CX5 and methods for correction.
Am J Clin Biochem
1990;27:34552.
36. Creer MH, Ladenson
J. Analytical
error due to lipemia.
Lab
Med 1983;14:351-5.
37. Jay DW, Provask
D. Characterization
and mathematical
correction
of hemolysis
interference
in selected
Hitachi
717 assays.
Cim Chem 1993;39:1804-10.
38. Blijenberg
BG, Liesting
EC, Zwang
L. Creatinine
and automatic analyzers
in relation
to icteric specimens.
Eur J Cim Chem
Biochem 1992;30:779-84.
39. Crocker
H, Shephard
MDS, White
GH. Evaluation
of an
enzymatic
method
for determining
creatinine
in plasma.
J Clin
Pathol
1988;41:576-81.
40. Spain MA, Wu AHB. Bilirubin
interference
with determination of uric acid, cholesterol,
and triglycerides
in commercial
peroxidase-coupled
assays,
and the effect of ferrocyanide.
Cke
Chem 1986;32:518-21.
41. Sonntag
0, Schumann
G. Evalvierung
der Ektachem-singleslide Methode
zur Bestimmung
der Creatinin-konzentration
im
Serum.
Lab Med 1989;13:351-6.
42. Witte DL, Brown LF, Feld RD. Effects of bilirubin
on detection
of hydrogen
peroxide
by use of peroxidase.
Cm
Chem
1978;24:
1778-82.
43. Sonnenblick
M, Eylath
U, Brisk
R, Eldad C, Hershko
C.
Paraproteun
interference
with colorimetry
of phosphate
in serum
of some patients
with multiple
myeloma.
Cim Chem
1986;32:
1537-9.
44. Pierce
GF, Garrett NC, Koenig
J, Lichti DA, Chan K-M.
Interference
by monoclonal
proteins
in the o-phthalaldehyde
method
for blood urea nitrogen.
Cke Chem Acta 1986;154:233-6.
45. Pudek MR, Nanji Ak Antibody
interference
with biochemical
tests and its clinical
significance.
Cim Biochem
1983;16:275-80.
46. Hirano
T, Matsuzaki
H, Miura
M, Kojima
E, Tamura
N,
Sekine T. An immunoglobulin
G inhibiting
lactate dehydrogenase
activity.
Clin Chim Acta 1986;159:17-25.
47. Ginsberg
J, Segal D, Ehrilich
RM, Walfish
PG. Inappropriate
T3 and T4 radioimmunoassay
levels
secondary
to circulating
thyroid hormone
antibodies.
Cke Endocrinol
1978;8: 133-S.
48. Hashimoto
T, Miyabo
5, Nishiby
M, Matsubara
F, Migita
S.
Interference
of immunoglobulins
in the radioimmunoassay
of
human
p-endorphin.
J Cliii Chem Cliii Biochem
1990;28:937-41.
49. Faulkner
AM, Peake MJ. Bicarbonate
interference
with Hitachi chloride
electrodes.
Ann Cliii Biochem
1991;28:107-8.
50. Spencer
K. Analytical
reviews
in clinical
biochemistry:
the
estimation
of creatunine.
Ann Cliii Biochem
1986;23:1-25.

51. Watkins
PJ. The effect of ketone bodies on the determination
f creatinine.
Clin Chim Acta 1967;18:191-6.
52. Larsen
K. Creatinune
assay by a reaction
kinetic
principle.
Cliii Clam Acta 1971;32:81-5.
53. Sash AJ, Koch TR, Drusano
GL. Cefoxitun falsely
elevates
creatinine
levels. J Am Med Assoc 1982;247:205-6.
54. Watkins
RE, Feldkamp
CS, Thibert
RJ, Zak B. Interesting
interferences
in a direct serum creatinune
reaction.
Microchem
J
1976;21:370-84.
55. Siest G, Dawkins
SJ, Galteau
MM. Drug effects on laboratory
tests. J Pharm Biomed
Anal 1983;1:247-57.
56. Benet
LZ, Sheiner
LB. Pharmacokunetics:
the dynamics
of
irug
absorption,
distribution
and elimination.
In: Gilman
AG,
Goodman
LS, Rail TW, Murad F, eds. Goodman
and Gilmans
the
pharmacological
basis
of therapeutics,
New York: Macmillan,
1985:13-22.
57. Siff KS, Finkler
AE. False-positive
barbiturate
test in urine
awing
to phenytoin
and 5-(p-hydroxyphenyl)-5-phenylhydantoun
[Tech Brief]. Cim Chem 1988;34: 1359-40.
58. Yatscoff
RW, Langman
U, LeGatt
DF. Cross reactivities
of
cyclosporin
G (NVa2 cyclosporin)
and metabolites
in cyclosporun
A
immunoassays.
Cke Chem 1993;39:1089-92.
59. Murphy JL, Hurt TL, Griswold
WR, Peterson
BM, Rodarte A,
Krouse
HF. Interference
with creatinune
concentration
measurement by high dose furosemide
infusion.
Crit Care Med 1989;17:
889-90.
60. Leissing
N, Mattia-Goldberg
C, Oskeroba
D. Modification
of
clinical chemistry
methods
to overcome
interferences
from diaspirin crosslinked
hemoglobin
(DCLHb)
[Abstract].
Cliii Chem 1993;

39:1144.
Si. Leissing
N, Mattia-Goldberg
C, Blama
D. Modification
of
hematology
methods
to overcome
interferences
from diaspirin
crosslinked
hemoglobin
(DCLHb)
[Abstract].
Cliii Chem 1993;39:
1142.
62. Sachs C, Marsoner
HJ, Ritter
C, Kindermas
C, Ghahramani
M. Ionized
magnesium
measurement
in serum
[Abstract].
Cliii
Chem 1993;39:1176.
63. Witte
DL. Matrix
effects
in therapeutic
drug
monitoring
surveys.
Arch Pathol Lab Med 1993;117:373-80.
64. Lott JA, Turner
K. Evaluation
of Trinders
glucose
oxidase
method
for measuring
glucose
in serum
and urine.
Cke Chem
1975;21:1754-60.
65. Siest G, Appel W, Blijenberg
GB, et al. Drug interference
in
clinical
chemistry:
studies
on ascorbic
acid. J Cliii Chem Cke
Biochem
1978;16:103-1O.
66. Rej R. Accurate
enzyme
activity
measurements.
Arch Pathol
Lab Med 1993;117:352-64.
67. Kroll
MH, Chesler
R, Eke RJ. Effect of lyophuization
on
results
of five enzymatic
methods
for cholesterol.
Cim Chem
1989;35:1523-6.
68. Miller WG. Matrix effects in the measurement
and standardization
of lipids and apolipoproteins.
Curr Opin Lipidol
1992;3:
361-4.
69. Ross JW, Myers
GK, Gilmore
BF, Cooper
GR, Naito
HR,
Eckfeldt
J. Matrix
effects
and accuracy
of cholesterol
analysis.
Arch Pathol Lab Med 1993;117:393-400.
70. Schlebush
H, Gruenn
V, Scheider
C. Simultaneous
interference of bilirubun
and hemoglobin
in creatunine
determinations
[Abstract].
Cliii Chem 1993;39:1144.
71. Jencks
WP. Catalysis
in chemistry
and enzymology.
New
York: Dover, 1969:555-90.
72. Kroll MH, Eke RJ. Mechanism
of cefoxitin
and cephalothin
interference
with the Jaff#{233}method
for creatinune.
cm Chem
1983;29:2044-8.
73. Nichols
EM. Cephalothin
interferes
in automated
assays
of
plasma creatmine
[Letter].
cm Chem 1981;27:1953-4.
74. Eke 1W, Chesler
RA. Interference
with the determination
of
creatunine
by cepha antibiotics
[Abstract].
Clin Chem
1982;28:
1580.
75. Kroll MH, Nealon
L, Vogel MA, Eke 1W. How certain drugs
interfere
negatively
with the Jaff#{233}
reaction
for creatinine.
Cke
Chem 1985;31:306-8.
76. Kroll
MH, Koch TR, Drusano
GL, Warren
JW. Lack of
interference
with
creatinune
assays
by four cephalosporin-like
antibiotics.
Am J Clin Pathol
1984;82:214-6.

77. Kroll MH, Hagengruber

C, Elm 1W. Reaction
of picrate
with
creatinine
and cepha antibiotics.
Clin Chem 1984;30:1664-6.
78. KroU MH, Roach NA, Poe B, Eke 1W. Mechanism
of interference with the Jaff#{233}
reaction
for creatunine.
Clin Chem
1987;33:
1129-32.
79. Kroll MH. Some observations
on the reaction
mechanism
of
cefoxitin
and cephalothin
with picrate.
Microchem
J 1990;42:
241-9.
80. Kodak Ektachem
Clinical
Chemistry
Slides.
Rate methodologies summary,
MP2-75.
Rochester,
NY: Eastman
Kodak Co., 1986.
81. Spindler
SA, Anderson
BD, Kindwall
KE. Lidocaune
interference with Ektachem
analyzer
determinations
of serum creatizune
concentration.
Cliii Pharm
1989;8:659-63.
82. Ritchie
JM, Greene
NM. Local anesthetics.
In: Gilman
AG,
Goodman
LS, Rail TW, Murad F, eds. Goodman
and Gilmans
the
pharmacological
basis of therapeutics,
7th ad. New York: Macmillan,
1985:310.
83. Sena SF, Syed D, Romeo
R, Krzymowski
GA, McComb
RB.
Lidocaine
metabolite
and creatinune
measurements
in the Ektachem
700: steps to minimize
its impact
on patient
care. Clin
Chem 1988;34:2144-8.
84. Couttie
K, Earle J, Coakley
J. Evaluation
of single-slide
creatinine
method
on the Kodak
Ektachem
700 shows
positive
interference
from lignocaine
metabolites.
Clin Chem
1987;33:
1674.
85. Gosney
K, Adachi-Kirkland
J, Schiller
HS. Evaluation
of
lidocaine
interference
in the Kodak Ektachem
700 analyzer
singleslide method
for creatinine.
Cliii Chem 1987;33:2311.
86. Eke 1W, Robertson
EA, Johnson
E. Bromide interferes
with
determination
of chloride
by each of four methods.
Clin Chem
1981;27:778-9.
87. Emancipator
K, Kroll M. Bromide
interference:
is less really
better?
Cliii Chem 1990;36:1470-3.
88. Gosling
JP. A decade of development
in immunoassay
methodology.
Cim Chem 1990;36:1408-27.
89. Berthod
C, Roy F. Enormous
cross-reactivity
of hydrocortisone
hemusuccmate
in the RAINEN
RIA kit for cortisol determination
[Techl Brief]. Cke Chem 1987;34: 1988.
90. Wians FH Jr, Norton JT, Wirebaugh
SR. False-positive
serum
tricycic
antidepressant
screen
with cyproheptadine
[Letter].
Cliii
Chem 1993;39:1355-6.
91. Bowie U, Kirkpatrick
PB, Dohnal
JC. Thyroid-function
testing with the TDx: interference
from endogenous
fluorophore
(Tech
Brief]. Cliii Chem 1987;33:1467.
92. Bloom
JN, Herman
DC, Eke
1W, Sliva CA, Ruddel
ME,
Nussenblatt
RB, Palestine
AG. Intravenous
fluorescein
interference with clinical
laboratory
tests.
Am J Opthalmol
1989;108:
375-9.
93. Boscato
LM, Stuart
MC. Heterophilic
antibodies:
a problem for
all immunoassays.
Cke Chem 1988;34:27-33.
94. Zweig MH, Csako
G, Benson
CC, Weintraub
BD, Kahn BB.
Interference
by anti-immunoglobuhn
antibodies
in immunoradiometric assays of thyrotropin
involving
mouse monoclonal
antibodies. Cke Chem 1987;33:840-4.
95. Boscato
LM, Stuart
MC. Incidence
and specificity
of interference in two-site immunoassays.
Cke Chem 1986;32:1491-5.
96. Hursting
MJ, Raisys
VA, Opheim
KE. Drug-specific
Fab
therapy
in drug overdose.
Arch Pathol Lab Med 1987;111:693-7.
97. Rainey
PM. Digoxun
immune
Fab (DigibandR)
markedly
affects the results
of immunoassays
for digoxin
[Abstract].
Cke
Chem 1989;35:1165.
98. Lehrer
M, Catinella
M. Observations
of a renal patient treated
with Digiband;
comparison
of digoxun levels by Abbott
TDx and
Opus [Abstract].
Cke Chem 1993;39: 1240.
99. Bahar
I, Case A, Farrell
K, Ziegler W, Inbar S, Metzmann
E.
The effect of Digibund on Opus digoxin assay [Abstract].
Cke Chem
1991;37: 1016.
100. Sonntag
0. Arzneimittel-Interferenzen.
Stuttgart:
George
Thieme
Verlag,
1985.
101. Trydung
N, Roos KA. Drug interferences
and drug effects in
clinical
chemistry,
5th ad. Stockholm:
Apoteksbolaget,
1989.
102. Salway JG. Drug-test
interactions
handbook.
London: Chapman and Hall Medical,
1990.
103. Siest G, Galteau
MM. Drug effects on laboratory
test results:
analytical
interferences
and pharmacological
effects.
Littleton,
MA: PSG Publishing,
1988.

CLINICAL

CHEMISTRY,

Vol. 40, No. 11,

1994

2005