Sunteți pe pagina 1din 13

Process

Validation
How Much to Do and When to Do It
Anurag S. Rathore, Joseph F. Noferi, Edward R. Arling,
Gail Sofer, Peter Watler, and Rhona O’Leary
2002;15(10):18–20,22,24,26,28

The trick to process validation, these industry experts argue, is to


understand that it is a process that stretches through the whole product
life cycle. Some secrets of success: Take a team approach; focus on the
timing of the various stages of validation; avoid some common mistakes;

P
and build your documentation as you go.

rocess validation is defined in the controlled to maximize the probability that


supplementary information sec- the finished product meets all quality and
tion of the Federal Register as ­design specifications.
“a QA function that helps ensure Validation often requires a joint effort
drug product quality by provid- — ­planning and expertise from several
ing documented evidence that groups. Units or departments that help cre-
the manufacturing process con- ate the validation package often include
sistently does what it purports to do” (1). It process development, ­engineering, manu-
has also been defined as the act of “estab- facturing, quality assurance, and quality
lishing documented evidence that provides control. Some of the critical process vali-
a high degree of assurance that a specific dation activities are described in the “Criti-
process will consistently produce a product cal ­Activities box” (2–4).
meeting its predetermined specifications A big dilemma during process valida-
and quality attributes” (2). tion development is how much to do and
Process validation is required in license when to do it. Although it is essential to
submissions for all products regulated by be thinking of validation as early as pos-
CBER or CDER and usually is a subject sible, the validation process often changes
of intense scrutiny and lots of activities as the product goes from phase 1 to phase
that culminate in an acceptable validation 3 and beyond. Creating the package too
package. FDA expects that each step of early — doing too much too soon — can
a CGMP manufacturing process must be mean redoing a lot of the work. Insights

142 | BioPharm International


into those “how much” and “when” ques-
tions are shared here by some process ex-
perts from our industry.

Joseph F. Noferi
and Edward R. Arling
Quality Assurance, Pharmacia Corporation
Validation has become one of the pharma-
ceutical industry’s most recognized and
discussed subjects. It is a critical success
Critical Activities factor in product approval and ongoing
commercialization. Defined, it is the “act
Some of the critical process validation
of establishing documented evidence that
activities include the following (2–4).
provides a high degree of assurance that a
• Create a validation master plan specific process will consistently produce a
(VMP) that shows “when” and “what” product meeting its predetermined specifi-
activities will be performed.
cations and quality attributes.” Despite the
• Develop a strategy that allows simplistic definition, validation is subject
revisiting the VMP as the process to variable interpretations both by industry
“changes” from a phase 1 process to and by regulators.
a phase 3. Approach and philosophy. Our industry’s
• Identify components that will take approach to validation is often flawed. We
place at small and at large scale. tend to think deductively — indeed, com-
• Know the implications of CGMP on panies train us to think deductively. We
raw material usage, facility mainte- order our thoughts and approach problem
nance, documentation requirements, resolution by listing options and selecting
utilities, equipment cleaning, and per- the optimum course, prioritizing issues and
sonnel training. selecting actions. Deductive thinking identi-
• Define “critical” process parameters. fies a solution and looks for all the prob-
• Identify process parameters as “criti- lems that might be solved by the solution.
cal” and “noncritical.” ­Inductive thinking results in an analytical
rather than a systemic approach to problem
• Determine the proven acceptable
resolution. Analysis focuses on structure; it
range (PAR) and the normal operating
range (NOR) for each critical process
reveals how components work. Synthesis
parameter. focuses on function; it reveals why compo-
nents operate as they do. Systems thinking
• Demonstrate that critical parameters
is synthesis, putting things together. Analy-
can be monitored and controlled dur-
sis is taking them apart. The two approaches
ing manufacturing runs.
are complementary: Analysis yields knowl-
—Courtesy of Anurag Rathore edge; synthesis yields understanding (5).

Elements of Biopharmaceutical Production Series | 143


Process validation requirements. Valida- manufacturing steps: cleaning, weighing,
tion is a ­dynamic process. It is expected to measuring, mixing, blending, filling, pack-
follow a timeline stretching from initial de- aging, and labeling. For an API, the expec-
sign through ­ongoing commercial operation tation is that all critical processing steps
— the product life cycle. Typical expecta- determined to affect the quality and purity
tions are that design qualification (DQ), in- of the API be validated.
stallation qualification (IQ), and ­operational Frequent problems. Regulatory expecta-
qualification (OQ) should be nearly com- tions at each functional stage often ex-
plete early in the development process, at ceed the best efforts of those charged with
phase 1 and 2. Process qualification (PQ) translating management directives into di-
should be complete at end of phase 3. But rect action. The “Validation Steps Often
everything evolves with the process, and Missed” sidebar illustrates each functional
that process is ­usually a moving target until stage and the problems inspectors fre-
registration. quently find there. It is not intended to be
Controls must be applied to all manu- all-inclusive and should not be construed
facturing steps, critical starting materials, as definitive.
components, and the master cell bank. Validation is about control — that is, ad-
Those controls need to increase as the pro- equate controls supporting and surrounding
cess develops toward final isolation and the process — without which, the validation
purification. Those controls must cover all will fail. For biopharmaceuticals, the more
process steps identified as critical — those frequently mis­validated ­activities include
steps that can affect the quality and purity maximum cell age, impurity ­profiles, col-
of the final product. The subcomponents umn resin life, and viral clearance.
that affect the process include equipment, Critical success factors. Success requires
facilities and utilities, systems, comput- leadership and management with a focus
ers, cleaning, analytical method transfers, on quality across all functional areas. Pre-
and sterilization among others. A ­frequent paring and planning can never be over-
problem is the failure to address the life emphasized. The small costs incurred up
cycle of the system as a whole — most front pale when compared with reperform-
validation efforts focus on individual sub- ing the entire exercise. Validation should
components of the process and stop after be written as if FDA were the customer.
three commercial runs. The agency is the ­reviewer. And of course,
Process validation requirements for ac- corrective actions to solve any validation
tive pharmaceutical ingredients (API) dif- problems need to be implemented before a
fer from those for finished dosage forms preapproval ­inspection.
(drug products). The standards vary with
the type of API, the range of specifica- Gail Sofer
tions, and “other factors.” For drug prod- Regulatory Services, BioReliance Corporation
ucts, the regulators expect validation of all Regulations mandate compliance with Good

144 | BioPharm International


Validation Steps Often Missed
Regulatory expectations at each functional stage often exceed the best efforts of those charged with
translating management directives into direct action. The following list of items that frequently cause
problems if missed is not intended to be all-inclusive and should not be construed as definitive.
At the design qualification (DQ) stage, elements often missed include:
• Adequate description of the equipment’s intended use
• Clear specifications for all critical design parameters
• Setting design parameters that allow future flexibility (for example, the process will likely
change, but the equipment may not)
• Specifications that take CGMPs into account.
At the installation qualification (IQ) stage, elements often missed include:
• A list of all equipment that, when operating, has the potential to affect product quality or pro-
cess performance
• As-built drawings and specifications for all purchased equipment, new or used
• Verification that all purchased equipment and its installation meets the original intent (functional
specifications and design parameters), including applicable building, electrical, plumbing, and
other such codes
• Preventive maintenance plans and schedules for all such equipment.
At the operational qualification (OQ) stage, elements often missed include:
• Process operating parameters for each module, including those designated as critical
• An OQ protocol designed to demonstrate that the equipment used in each module operates as
intended throughout each process operating parameter range
• Task reports describing the successful execution of each OQ protocol
• A list identifying each module (step, unit of operation, or stage) of the process.
At the performance qualification (PQ) stage, elements often missed include:
• A fully defined process, including identifying critical processes and their acceptable operating
parameter range (traceable to development reports or small-scale supporting studies) and
defining potential adverse consequences
• Completed product specifications
• Scientific rationale or basis for criteria — usually exists but is poorly documented
• IQ and OQ steps completed and reports written, reviewed, and approved
• Operating personnel trained and qualified
• Change control procedures in place.
—Courtesy of Joseph F. Noferi

Elements of Biopharmaceutical Production Series | 145


BP OCT 01
RATHORE FIG. 1
FILENAME: Rathore F1

evant, and removal of other impurities of


5.075
biological origin should be no less rigorous
5.050 A
UCL5.0554
than for licensed product and should, there-
5.025
B
fore, be validated” (7). The ­product must
C
Avg5.0089 also be shown to be stable during the time
5.000 C
B it is at the clinic, and that can require valida-
4.975 A
LVL4.9624
tion or, at least, qualification of stability-in-
4.950 dicating assays. The catch is that validation
1 2 3 4 5 6 7 8 9
Lot Number of a process requires validated ­assays. Be-
fore phase 1, sponsors’ assays are often still
Figure 1. Control chart showing that in the research unit, performed by only one
operator skill, manufacturing equipment, or, at best, a few people.
and written instructions are sufficient to Sterility and mycoplasma testing. Sterility
adjust oxidation pH to the same set-point and mycoplasma testing are crucial to the
from lot to lot. safety of a biotechnological product pro-
—Courtesy of Peter Watler duced in mammalian or insect cells. Steril-
ity assays are essential not only for aseptic
Manufacturing Practices (GMP) for the man- processing validation but also for estab-
ufacture of clinical trial materials (6,7). Vali- lishing product stability. Sterility assays
dation is one component of the GMPs, but it are also used to determine the acceptability
is not feasible to complete validation before of unprocessed bulk for further processing.
a process is fully developed. Instead, it is im- These assays must be validated accord-
portant to understand what is required for ing to the latest regulatory requirements,
phase 1, 2, and 3 clinical ­trials. Understand- such as 21 CFR 610.12 (8), United States
ing the requirements means that the needs of and European pharmacopoeia standards, or
the patient and the expectations of the regu- FDA Points to Consider.
latory agencies are considered. From both Sterility and mycoplasma assays can’t
patient and regulatory perspectives, safety be ­considered validated without stasis test-
is critical. An understanding of the risks as- ing to ­ensure that the test article doesn’t
sociated with product source, manufacturing interfere with the assays, which can cause
methods, and the product itself is essential false negatives that can lead to adverse
for making ­decisions about what has to be patient reactions that can ­potentially ter-
validated for the various clinical trial phases. minate clinical trials. If samples are to be
Phase 1. Validation “must-haves” for the shipped, then shipping conditions must
earliest clinical trials include those related also be validated to demonstrate no loss of
to safety. Annex 13 to the EU GMP Guide: bacteria, fungi, or mycoplasma.
Manufacture of Investigational Medicinal Viral clearance validation is usually con-
Products states that “Validation of the ster- tracted out: The assays are validated by a
ilization process is no different than for li- testing company and each sponsor’s test
censed product. Virus clearance, where rel- article is evaluated for interference and cy-

146 | BioPharm International


totoxicity. But validation of process scale- might be determined using high-perfor-
down is often overlooked in the rush to get mance liquid chromatography (HPLC) and
into the clinic. a total protein assay. But what about activ-
Scale-down. Time and money are wasted ity and impurities associated with the eluted
if virus clearance evaluation studies are antibody? Without an impurity profile or
performed without a validated scale-down without knowing how much of a MAb’s
model of the process. Validation requires biological activity has been retained, the
that the sponsor ­understands critical pro- control parameters (such as flow rate and
cess and control parameters, uses qualified pH) cannot be established with certainty. A
equipment and validated assays, and fol- sponsor in this situation should go back and
lows a protocol defining the study and the understand both purity and impurity pro-
expected outcome. A sponsor must also files. The assays used to determine those
understand what each unit of operation profiles are unlikely to be validated at this
does and how output is measured. Unfortu- stage, but they must be “qualified.” Typi-
nately, at this early stage of development, cally, a reference standard is run along with
many of the assays that enable that un- each assay to ensure the assay is working
derstanding are not validated. At worst, a according to protocol.
sponsor might lack understanding of what Which viruses? Another issue in viral
a unit of operation accomplishes. clearance studies is how many and which
Take, for example, an immobilized pro- viruses to use in the first studies so that
tein A column used to purify a monoclonal clinical trials can begin. The opinion of
antibody (MAb). Initially, product yield regulatory agencies varies. If a sponsor in-
tends to begin clinical trials worldwide, it
is ­essential to understand the latest regional
Table 1. Validation results demonstrat- concerns related to viral safety so that vali-
ing that E. coli proteins were reduced by dation of virus clearance will stand up to
the filtration step, significantly at cation regulatory scrutiny.
exchange 1, and further reduced to below Other impurities of biological origin. Devel-
acceptance criteria by cation exchange 2. oping a validated assay that demonstrates
—Courtesy of Peter Watler the removal of impurities (such as host cell

Purification Lot Number (% w/w)

Process Stream 23004 23006 23007 23008 23010 Average

Oxidation 6.39 4.60 13.4 6.78 11.8 8.60


Filtration 4.13 3.26 2.42 4.83 7.31 3.58
Cation exchange 1 0.915 0.494 0.64 0.532 1.18 0.65
Cation exchange 2 <0.009 <0.003 <0.004 <0.003 <0.011 <0.006
Purified bulk <0.007 <0.014 <0.003 <0.009 <0.023 <0.011

Elements of Biopharmaceutical Production Series | 147


p­ roteins) for phase 1 is seldom possible. bridging studies in the clinic. Assays not
There is one exception. By using a parental yet validated must now be validated so that
cell line and similar culture conditions to process validation required for ­biologics
produce multiple products, a generic assay ­licensure can be completed during phase
can be validated and qualified for each new 3. This is the phase of heavy-duty valida-
product. However in other situations, the tion and ­requires equipment qualification,
host cell protein population is unlikely to then process validation during three or more
be consistent until the scale and conditions consecutive batches. Cleaning validation
of the final culture have been established. and lifetime studies for chromatography
Enzyme-linked immunosorbent assay columns are important validation ­elements.
(ELISA) kits on the market may be suffi- Clearance studies to ­remove host cell pro-
cient for clinical trial material, but compa- teins, DNA, viruses, and other impurities
nies are still required by today’s regulations may eliminate lot release testing. Some
to develop their own assays for licensed clearance studies are performed at a smaller
product. The development of those assays scale than manufacturing, so the small-scale
is time-consuming — they can take more model must be validated. Once the process
than a year. Planning ahead is essential. is finalized, viral clearance studies are car-
In addition to host cell proteins, other im- ried out again. At this stage, however, larger
purities for which validated assays may be virus panels are used, mass balance analyses
required include any toxic or potentially are attempted, and duplicate runs are tested.
immunogenic substances. For chromatography steps that claim to re-
Phase 2. During phase 1 and 2 clinical move viruses, end-of-resin lifetime stud-
­trials, the process is generally improved. ies are ­performed to demonstrate consistent
Both upstream and downstream processes virus clearance.
change, analytical methods are usually Preventing surprise. One of the most com-
transferred to QC during this stage, and mon FDA form 483 observations is the lack
potency assays validated. No specific pro- of process validation. Planning ahead during
cess validation activity is required at this early development can prevent unpleasant
stage, unless the changes made to improve surprises, such as specifications that are so
the process have the potential to affect the tight they cannot be met, analytical assays
results of ­previously performed valida- that can be performed only by one opera-
tion studies for sterility, virus clearance, tor and so are not validatable, or a process
and specific impurity ­removal. However, that cannot be scaled up or down without
while the process is being ­optimized, assay ­redesign. Surprises like that result in pro-
validation efforts should ­continue so that cesses that cannot be validated and often
process validation at pilot or full scale for multiple failed batches.
licensure can take place during phase 3. I find that for new processes and prod-
Phase 3. Sometime before phase 3, the ucts, the greatest validation problems arise
process is, hopefully, finalized to avoid when insufficient resources and insufficient

148 | BioPharm International


time are budgeted for ­understanding and op- inspections. Process validation is the point
timizing production. For biological and bio- at which the ­science of the process can be
technological products, phase 1 validation ­explained to regulatory agencies. Following
activities are carried out to ensure safety validation, regulatory agencies are looking for
and the manufacturing process should be consistent process operation within the ranges
controllable. Phase 2 manufacturing pro- and meeting the criteria specified in the
cesses should be well-controlled and better validation. Scientific arguments explaining
understood with validated ­assays that dem- operation and performance outside of speci-
onstrate that ­control. Phase 3 should be very fied ranges are likely to fall upon deaf ears.
well controlled and full process validation Therefore it is essential to generate bench and
should take place ­during this phase. pilot scale process data that establish key pa-
rameters and their acceptance ­criteria before
Peter Watler process validation in the GMP facility.
Pilot Plant Engineering, Amgen Inc. When to validate. Process validation is
Process validation demonstrates the consis- ­e xpensive, involving copious sampling,
tency of multiple batches at full-scale. The ­extensive analysis, and detailed documenta-
validation shows that the process is operated tion. Because of the complexity and cost,
in a consistent manner and that contami- process validation is best performed dur-
nants are reproducibly ­reduced to accept- ing phase 3 trials following the decision to
able levels. That is accomplished by moni- file a Biologics License Application (BLA).
toring those parameters that demonstrate This “delayed” strategy offers several ad-
consistent ­operation of the process, consis- vantages. The process is better understood,
tent ­formation of the product, and consistent which means that key parameters and ac-
removal of the contaminants. Processes can ceptance criteria can be better specified. In
be validated through two distinct metrics: ­addition, the commercial process is in place
operational parameters (inputs) and perfor- at this point, and validation does not have to
mance parameters (outputs). Operational contend with process changes, adjustments,
parameters are process control set-points and inexperience that can lead to deviations
for variables such as flow rate, temperature, from the validation protocol. By the time
and concentration. These parameters define products enter the regulatory review phase,
the process recipe and are used to demon- the likelihood of product success increases
strate that the facility, equipment, and staff to 90%, meaning that there is less risk that
can ­execute the process consistently. Per- process validation will have been for naught.
formance parameters reflect the outcome Then too, a­lthough process validation is a
of a given step and indicate that the process required component of a BLA submission, it
gave the desired result. is not required for clinical trials.
Process validation is an investment in What to validate. Demonstrating process
future production because it sets the bar by consistency requires multiple, full-scale
which the process will be judged at future batches. Three to five consecutive puri-
fication runs from three consecutive fer-

Elements of Biopharmaceutical Production Series | 149


mentation lots should generate sufficient agents required by the process (for ­example,
consistency data. Process validation can be guanidine, glycerol, and antifoam). Host
made more efficient and more consistent cell-­related impurities are derived from the
by using templates that identify key input organism used to generate the product (for
and output parameters for a unit operation. ­example, nucleic acids, CHO proteins, and
For example, consistent operation of a tan- endotoxins). Product-related contaminants
gential-flow filtration step can be shown by
are variants and isoforms of the target pro-
monitoring cross-flow rate, transmembrane
tein (for example, ­oxidized, deamidated, ag-
pressure (TMP), temperature, retentate tank
volume, and diafiltration volume. Consis- gregated, and clipped forms). The validation
tent performance of this step can be shown should also demonstrate at which step the
by monitoring excipient removal, protein contaminant is ­removed and at what point
contaminant removal, product concentra- in the process it meets acceptance criteria.
tion, product recovery, pH, and conduc- Table 1 shows that E. coli proteins were
tivity. Such parameters demonstrate that reduced by the filtration step, ­significantly
the step achieved its purpose in the pro- reduced at cation exchange 1, and further
cess. These validation templates for unit reduced to below the ­acceptance criteria by
operation can be customized to address any cation exchange 2.
unique specifics of the process or product. Process consistency is also shown by
Operational control parameters are input
monitoring yield and product concentra-
variables with set-points or ranges speci- tion at each step. With sufficient advance
fied in the manufacturing procedure to planning, earlier process data, and care-
define how a process is ­executed. Control ful execution, the data from the validation
charts can be used to demonstrate that the should demonstrate that the manufactur-
set-points are consistently achieved within ing procedure consistently and effectively
the specified ranges. Typical operational yields product that meets specifications. It
­parameters include pH, raw material quanti- will also demonstrate that the process can
ties, reaction times, flow rates, temperature, be operated consistently and that the pro-
and pressure. For example, the control chart cess, product, and host cell contaminants
in Figure 1 shows that operator skill, manu- are reduced to acceptable levels.
facturing equipment, and written instruc-
tions are sufficient to ­adjust oxidation pH to Rhona O’Leary
the same set point from lot-to-lot. Process Sciences, Genentech, Inc.
Performance parameters are output vari- Manufacturing processes for biologicals are
ables that reflect the outcome of a given continually optimized as the product moves
step, indicating that the process performed through various stages of development,
as expected. Validation should demonstrate from the preclinical stage through the in-
that the process is capable of consistently vestigational new drug (IND) filing to the
removing three classes of contaminants: pro- common technical document (CTD) filing
cess-related, host cell-related, and product- and on to approval. The early-stage product
related. Process-related contaminants are re- may have a basic and somewhat unpolished

150 | BioPharm International


manufacturing process. By midstage the teins and DNA) and small molecules and
process is optimized and becomes more ro- the cleaning of resins and membranes.
bust. By the end of development, the final Scaled-down studies typically include
manufacturing process is fully character- the validation of resin lifetimes, in-process
ized, validated, and qualified. hold times, buffer stability, virus validation,
Development validation activities. At Ge- harvest criteria, filter extractables, resin
nentech, our goal is to have the manufac- leachables, and cell age. Genentech has
turing process completely finalized before found that a combination of manufacturing-
the manufacturing campaign that supplies scale and scaled-down studies provides the
the phase 3 clinical trials. That means best overall process validation plan.
that the critical process parameters have Before phase 3, evaluate robustness. Be-
been identified, the operating ranges have fore the process is finalized several scaled-
been set, and the process is deemed robust down runs are performed to evaluate
enough to tolerate the manufacturing envi- process robustness. These runs provide
ronment and produce marketable product. an adequate history of the process’ vari-
Phase 1 process validation activities. Vali- ability. Such runs may include running the
dation evolves along with development of process at the extremes of its operating
the process, and different types and levels ranges, testing different feedstocks, using
of validation are appropriate at the various different lots of resins, testing cell culture
developmental stages. At the time of IND components or peptone lots, among oth-
filing, the only validation ­requirement is a ers. Sometimes we perform these runs at
demonstration that the process removes ret- the 400 liter scale, more often at lab scale.
roviral particles. If modifications are made The timing of these scaled-down runs also
as the process moves through development, provides the scientist with the ability to
the process for removing retroviruses must incorporate last minute changes that result
be reassessed. Complete process validation from data gathered during these runs to
is not required until the CTD is filed. further optimize and improve the robust-
Phase 2 and 3 process validation activities. ness of the process.
Validation is performed after the process It is sometimes easy to overlook the im-
is fully developed and finalized. Process portance of a complete evaluation of the
validation can be a combination of manu- robustness of the process for both cell cul-
facturing-scale and scaled-down studies. ture (or fermentation) and recovery before
The qualification runs are ­performed on finalizing the process. Although it is hoped
validated equipment, at full manufacturing that the manufacturing process will run at
scale, in the facility in which the ­product its optimum operating conditions on a rou-
is to be routinely manufactured. The vali- tine basis, knowing how the process works
dation studies performed at manufacturing under conditions beyond its optimum set
scale typically include the validation of the points (such as flowrate, pH, and tempera-
process to clear impurities (host cell pro- ture excursions) is important. If the pro-

Elements of Biopharmaceutical Production Series | 151


cess is running close to the edge of fail- We try to assess the worst-case situation,
ure, the time to discover and correct that is after which all other conditions are well cov-
before finalizing the process. The studies ered by those conditions. For example, by
performed at this stage require fully devel- testing the ability of a particular resin ma-
oped analytical assays, and those assays are trix (such as Sepharose) to be sanitized in
examined to ensure the process meets the the presence of a protein mixture, we can
acceptance criteria set in the Certificate of use that study to support the same sanitizing
Analysis. Although this phase of develop- agent for all other similar ­protein mixtures,
ment is not ­considered part of the official on similar resin matrices. If we did not adopt
validation process, if well documented, it such an approach, it is likely that the huge
can become the foundation on which the workload currently required for process
process validation program is based. validation would be doubled. The ability to
Final process with product in phase 3 trials. ­develop such large matrices for our various
Genentech finalizes the manufacturing pro- processes is reflective of the large number
cess for the start of the phase 3 clinical tri- of products that have been validated in Ge-
als — that is, the material that is made in nentech in recent years. A smaller company
the manufacturing campaign to supply the might not have the luxury of developing the
phase 3 trials is from the fully developed, necessary database of information to support
fully characterized, and finalized manufac- a matrix design, but it could use the worse-
turing process. No process changes are ex- case scenario to cover not just a single prod-
pected beyond this stage. At this point, the uct, but also many future ­products.
analytical assays are also finalized and are Process validation is a combination of ef-
used to release the material to the clinic. forts. The importance of documentation is
Once that product is manufactured, some also frequently overlooked. Throughout all
of that material can be used to support pro- phases of development, we try to capture
cess validation. Validation of the analytical all development work and all decisions
methods and the definition of the control in development reports. These reports, in
system begin at this stage. turn, support any questions that the regula-
Generic process validation activities. Some tory agencies may have about how a pro-
of the validation studies we perform are cess was developed or why certain deci-
common across many of our processes: sions were made. These reports are stored
viral validation, resin and membrane sani- in a database for easy access. They also
tization and storage, buffer ­stability, filter serve as a resource into the development
extractables, and resin leachables. Rather history if an employee leaves and as guid-
than continuing to repeat these studies time ance for new employees.
and time again, we developed matrices to The biostatisticians are also on hand
support a wide ­variety of conditions — to assist in the design of characterization
­conditions that typically reflect our most studies and to evaluate process robustness,
common operating parameters. frequently using fractional factorial de-

152 | BioPharm International


signs. Sometimes these results are incor- a­ pproaches discussed here seek to over-
porated into resin lifetime studies. Quality come challenges often encountered dur-
assurance, analytical chemistry, and pro- ing development of a validation program
cess development groups work closely to- for biotechnology ­products. An acceptable
gether to ensure that all appropriate studies validation program ­addresses the following
are performed, that the data are audited for difficulties encountered by others:
accuracy, and that reports are generated • Lack of overall strategy
and approved. • Failure to consult early with the regula-
Developing a manufacturing process tory authorities
that will ultimately receive regulatory • Inadequate product definition
• Failure to follow CGMP regulations
approval and yield a marketable product
• Poorly defined cell bank genealogy
requires a development plan that follows
• Inadequate analytical procedures
regulatory requirements. The development • Too many process changes after valida-
plan must have goals that are met during tion activities have been completed
development before it moves onto the next • Not enough process characterization or
stage. In addition, we have found that hav- qualification
ing standardized study designs, protocols, • Commencing process validation too early
and reports have led to successful valida- • Inappropriate acceptance criteria.
tion programs. The program described has
undergone four FDA preapproval inspec- Validation Perspectives
tions since 1998. We hope that this article is useful to those
in the biopharmaceutical industry who are
M e e t i n g t h e C h a ll e n g e currently involved in validation activities
In this article, experts from different com- and especially useful to those who are
panies in our industry have outlined the planning to start validation activities in the
strategy that they use to arrive at a vali- near future.
dated process that can be submitted to reg- This article series presents opinions and
ulatory agencies. Although it is interesting viewpoints of some of the industry experts
to see the different approaches that can be on ­issues that are routinely faced in pro-
taken, distinct commonalities underlie all cess ­development and manufacturing of
the approaches described. The validation biopharmaceuticals.

Elements of Biopharmaceutical Production Series | 153


References
(1) FDA, “Current Good Manufacturing Practices; (5) Noferi J.F., Arling E.R., and Worden D.E.,
Proposed Amendment of Certain Requirements “Beyond QSIT,” BioPharm 15(4), 40–44, 67
for Finished Products: Supplementary (April 2002).
Information,” Federal Register 61(87), p. (6) Center for Drug Evaluation and Research,
20104 (3 May 1996). Available at www.fda. Guideline on the Preparation of Investigational
gov/cber/genadmin/cgmp.pdf. New Drug Products (Human and Animal)
(2) Sofer G. and Hagel L., “Validation,” Handbook (FDA, Rockville, MD, March 1991). Available
of Process Chromatography: A Guide to at www.fda.gov/cder/guidance/ old042fn.pdf.
Optimization, Scale-up, and Validation (7) European Commission, “Annex 13
(Academic Press, Manufacture of Investigational Medicinal
New York, 1997), pp. 119–243. Products” Medicinal Products for Human
(3) Martin-Moe S. et al., “Validation of Critical and Veterinary Use: Good Manufacturing
Process Input Parameters in the Production of Practices, Vol. 4, Draft 1, F2/AN D(2001)
Protein Pharmaceutical Products: A Strategy (Enterprise Directorate-General, Brussels,
for Validating New Processes or Revalidating Belgium, November 2001).
Existing Processes,” PDA J. Pharm. Sci. (8) “General Biological Products Standards:
Technol. 54(4), 315–219 (July/August 2000). Sterility,” Code of Federal Regulations: Food
(4) Seely R.J. et al., “Defining Critical Variables in and Drugs,
Well-Characterized Biotechnology Processes,” Title 21, Part 610.12 (U.S. Printing Office,
BioPharm 12(4), 33–36 (April 1999). Washington, DC, revised 1 April 2001). BP

Originally published in the October 2002 issue of BioPharm International.

154 | BioPharm International