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

The trick to process rocess validation is defined in the Validation often requires a joint effort —

P
validation, these industry supplementary information section of planning and expertise from several groups. Units
experts argue, is to
the Federal Register as “a QA function or departments that help create the validation
understand that it is a
that helps ensure drug product quality package often include process development,
process that stretches
through the whole product by providing documented evidence that engineering, manufacturing, quality assurance,
life cycle. Some secrets the manufacturing process consistently and quality control. Some of the critical process
of success: Take a team does what it purports to do” (1). It has validation activities are described in the “Critical
approach; focus on the also been defined as the act of “establishing Activities box” (2–4).
timing of the various documented evidence that provides a high degree A big dilemma during process validation
stages of validation; avoid of assurance that a specific process will consis- development is how much to do and when to do
some common mistakes
tently produce a product meeting its predeter- it. Although it is essential to be thinking of vali-
(see page 20); and build
mined specifications and quality attributes” (2). dation as early as possible, the validation process
your documentation
as you go. Process validation is required in license often changes as the product goes from phase 1 to
submissions for all products regulated by CBER phase 3 and beyond. Creating the package too
or CDER and usually is a subject of intense early — doing too much too soon — can mean
scrutiny and lots of activities that culminate in an redoing a lot of the work. Insights into those
acceptable validation package. FDA expects that “how much” and “when” questions are shared
each step of a CGMP manufacturing process here by some process experts from our industry.
must be controlled to maximize the probability
that the finished product meets all quality and JOSEPH F. NOFERI AND EDWARD R. ARLING
design specifications. Quality Assurance, Pharmacia Corporation
Validation has become one of the pharmaceutical
industry’s most recognized and discussed
subjects. It is a critical success factor in product
approval and ongoing commercialization.
Defined, it is the “act of establishing documented
evidence that provides a high degree of assurance
that a specific process will consistently produce a
product meeting its predetermined specifications
and quality attributes.” Despite the simplistic defi-
nition, validation is subject to variable
interpretations both by industry and by regulators.
Approach and philosophy. Our industry’s
approach to validation is often flawed. We tend
to think deductively — indeed, companies train
us to think deductively. We order our thoughts
and approach problem resolution by listing
options and selecting the optimum course, priori-
tizing issues and selecting actions. Deductive
thinking identifies a solution and looks for all the
problems that might be solved by the solution.
Inductive thinking results in an analytical rather
18 BioPharm OCTOBER 2002
than a systemic approach to problem resolution. Validation is about control — that is, adequate
Analysis focuses on structure; it reveals how controls supporting and surrounding the process
components work. Synthesis focuses on function; — without which, the validation will fail. For bio-
it reveals why components operate as they do. pharmaceuticals, the more frequently misvalidated
Systems thinking is synthesis, putting things activities include maximum cell age, impurity
together. Analysis is taking them apart. The two profiles, column resin life, and viral clearance.
approaches are complementary: Analysis yields Critical success factors. Success requires leader-
knowledge; synthesis yields understanding (5). ship and management with a focus on quality
Process validation requirements. Validation is a across all functional areas. Preparing and planning
dynamic process. It is expected to follow a can never be overemphasized. The small costs in-
timeline stretching from initial design through curred up front pale when compared with reper-
ongoing commercial operation — the product life forming the entire exercise. Validation should be
cycle. Typical expectations are that design qualifi- written as if FDA were the customer. The agency
cation (DQ), installation qualification (IQ), and is the reviewer. And of course, corrective actions
operational qualification (OQ) should be to solve any validation problems need to be im-
nearly complete early in the development process, plemented before a preapproval inspection.
at phase 1 and 2. Process qualification (PQ) should Critical Activities
be complete at end of phase 3. But everything GAIL SOFER Some of the critical process
evolves with the process, and that process is Regulatory Services, BioReliance Corporation
validation activities include
usually a moving target until registration. Regulations mandate compliance with Good Manu-
the following (2–4).
Controls must be applied to all manufacturing facturing Practices (GMP) for the manufacture of
clinical trial materials (6,7). Validation is one com- • Create a validation master
steps, critical starting materials, components, and
plan (VMP) that shows
the master cell bank. Those controls need to ponent of the GMPs, but it is not feasible to com-
“when” and “what” activities
increase as the process develops toward final plete validation before a process is fully developed.
will be performed.
isolation and purification. Those controls must Instead, it is important to understand what is re-
quired for phase 1, 2, and 3 clinical trials. Under- • Develop a strategy that
cover all process steps identified as critical —
allows revisiting the VMP as
those steps that can affect the quality and purity of standing the requirements means that the needs of
the process “changes” from
the final product. The subcomponents that affect the patient and the expectations of the regulatory
a phase 1 process to a
the process include equipment, facilities and utili- agencies are considered. From both patient and reg- phase 3.
ties, systems, computers, cleaning, analytical ulatory perspectives, safety is critical. An under-
• Identify components that
method transfers, and sterilization among others. standing of the risks associated with product source,
will take place at small and
A frequent problem is the failure to address the manufacturing methods, and the product itself is es-
at large scale.
life cycle of the system as a whole — most valida- sential for making decisions about what has to be
validated for the various clinical trial phases. • Know the implications of
tion efforts focus on individual subcomponents of
CGMP on raw material
the process and stop after three commercial runs. Phase 1. Validation “must-haves” for the earliest
usage, facility maintenance,
Process validation requirements for active clinical trials include those related to safety. Annex
documentation
pharmaceutical ingredients (API) differ from 13 to the EU GMP Guide: Manufacture of Investi- requirements, utilities,
those for finished dosage forms (drug products). gational Medicinal Products states that “Validation equipment cleaning, and
The standards vary with the type of API, the of the sterilization process is no different than for personnel training.
range of specifications, and “other factors.” For licensed product. Virus clearance, where relevant,
• Define “critical” process
drug products, the regulators expect validation of and removal of other impurities of biological origin parameters.
all manufacturing steps: cleaning, weighing, should be no less rigorous than for licensed product
• Identify process parameters
measuring, mixing, blending, filling, packaging, and should, therefore, be validated” (7). The
as “critical” and “noncritical.”
and labeling. For an API, the expectation is that product must also be shown to be stable during the
time it is at the clinic, and that can require valida- • Determine the proven
all critical processing steps determined to affect
tion or, at least, qualification of stability-indicating acceptable range (PAR) and
the quality and purity of the API be validated.
the normal operating range
Frequent problems. Regulatory expectations at assays. The catch is that validation of a process re-
(NOR) for each critical
each functional stage often exceed the best efforts quires validated assays. Before phase 1, sponsors’
process parameter.
of those charged with translating management assays are often still in the research unit, performed
by only one or, at best, a few people. • Demonstrate that critical
directives into direct action. The “Validation
parameters can be
Steps Often Missed” sidebar illustrates each Sterility and mycoplasma testing. Sterility and
monitored and controlled
functional stage and the problems inspectors mycoplasma testing are crucial to the safety of a
during manufacturing runs.
frequently find there. It is not intended to be biotechnological product produced in mammalian
or insect cells. Sterility assays are essential not —Courtesy of Anurag Rathore
all-inclusive and should not be construed as
definitive. only for aseptic processing validation but also for
establishing product stability. Sterility assays are
BioPharm OCTOBER 2002 19
 also used to determine the acceptability of to be shipped, then shipping conditions must also
unprocessed bulk for further processing. These be validated to demonstrate no loss of bacteria,
assays must be validated according to the latest fungi, or mycoplasma.
regulatory requirements, such as 21 CFR 610.12 Viral clearance validation is usually contracted
(8), United States and European pharmacopoeia out: The assays are validated by a testing com-
standards, or FDA Points to Consider. pany and each sponsor’s test article is evaluated
Sterility and mycoplasma assays can’t be for interference and cytotoxicity. But validation
considered validated without stasis testing to of process scale-down is often overlooked in the
ensure that the test article doesn’t interfere with rush to get into the clinic.
the assays, which can cause false negatives that Scale-down. Time and money are wasted if
can lead to adverse patient reactions that can virus clearance evaluation studies are performed
potentially terminate clinical trials. If samples are without a validated scale-down model of the
process. Validation requires that the sponsor
Validation Steps Often Missed understands critical process and control parame-
ters, uses qualified equipment and validated
Regulatory expectations at each functional stage often exceed the best efforts of
those charged with translating management directives into direct action. The
assays, and follows a protocol defining the study
following list of items that frequently cause problems if missed is not intended to and the expected outcome. A sponsor must also
be all-inclusive and should not be construed as definitive. understand what each unit of operation does and
At the design qualification (DQ) stage, elements often missed include:
how output is measured. Unfortunately, at this
early stage of development, many of the assays
• Adequate description of the equipment’s intended use
that enable that understanding are not validated.
• Clear specifications for all critical design parameters At worst, a sponsor might lack understanding of
• Setting design parameters that allow future flexibility (for example, the what a unit of operation accomplishes.
process will likely change, but the equipment may not) Take, for example, an immobilized protein A
• Specifications that take CGMPs into account. column used to purify a monoclonal antibody
At the installation qualification (IQ) stage, elements often missed include: (MAb). Initially, product yield might be deter-
• A list of all equipment that, when operating, has the potential to affect product mined using high-performance liquid chromatog-
quality or process performance raphy (HPLC) and a total protein assay. But what
• As-built drawings and specifications for all purchased equipment, new or used about activity and impurities associated with the
eluted antibody? Without an impurity profile or
• Verification that all purchased equipment and its installation meets the
original intent (functional specifications and design parameters), including
without knowing how much of a MAb’s biologi-
applicable building, electrical, plumbing, and other such codes cal activity has been retained, the control parame-
ters (such as flow rate and pH) cannot be estab-
• Preventive maintenance plans and schedules for all such equipment.
lished with certainty. A sponsor in this situation
At the operational qualification (OQ) stage, elements often missed include:
should go back and understand both purity and
• Process operating parameters for each module, including those designated impurity profiles. The assays used to determine
as critical those profiles are unlikely to be validated at this
• An OQ protocol designed to demonstrate that the equipment used in each stage, but they must be “qualified.” Typically, a
module operates as intended throughout each process operating parameter reference standard is run along with each assay to
range ensure the assay is working according to protocol.
• Task reports describing the successful execution of each OQ protocol Which viruses? Another issue in viral clearance
• A list identifying each module (step, unit of operation, or stage) of the studies is how many and which viruses to use in the
process. first studies so that clinical trials can begin. The
At the performance qualification (PQ) stage, elements often missed include: opinion of regulatory agencies varies. If a sponsor
• A fully defined process, including identifying critical processes and their intends to begin clinical trials worldwide, it is
acceptable operating parameter range (traceable to development reports or essential to understand the latest regional concerns
small-scale supporting studies) and defining potential adverse consequences related to viral safety so that validation of virus
• Completed product specifications clearance will stand up to regulatory scrutiny.
Other impurities of biological origin. Developing a
• Scientific rationale or basis for criteria — usually exists but is poorly
documented validated assay that demonstrates the removal of
impurities (such as host cell proteins) for phase 1
• IQ and OQ steps completed and reports written, reviewed, and approved
is seldom possible. There is one exception. By
• Operating personnel trained and qualified
using a parental cell line and similar culture con-
• Change control procedures in place. ditions to produce multiple products, a generic
—Courtesy of Joseph F. Noferi assay can be validated and qualified for each new
product. However in other situations, the host cell
20 BioPharm OCTOBER 2002
 larger virus panels are used, mass balance analyses
5.075 are attempted, and duplicate runs are tested. For
5.050 A
UCL
5.0554 chromatography steps that claim to remove
B viruses, end-of-resin lifetime studies are performed
Figure 1. Control chart showing 5.025
C to demonstrate consistent virus clearance.
that operator skill, 5.000 C
Avg
5.0089
Preventing surprise. One of the most common
manufacturing equipment, and B FDA form 483 observations is the lack of process
written instructions are sufficient 4.975 A
validation. Planning ahead during early develop-
LVL
4.9624
to adjust oxidation pH to the 4.950 ment can prevent unpleasant surprises, such as
1 2 3 4 5 6 7 8 9
same set-point from lot to lot. specifications that are so tight they cannot be met,
Lot Number
—Courtesy of Peter Watler analytical assays that can be performed only by one
operator and so are not validatable, or a process
protein population is unlikely to be consistent that cannot be scaled up or down without redesign.
until the scale and conditions of the final culture Surprises like that result in processes that cannot be
have been established. validated and often multiple failed batches.
Enzyme-linked immunosorbent assay (ELISA) I find that for new processes and products, the
kits on the market may be sufficient for clinical greatest validation problems arise when insuffi-
trial material, but companies are still required by cient resources and insufficient time are budgeted
today’s regulations to develop their own assays for understanding and optimizing production. For
for licensed product. The development of those biological and biotechnological products, phase 1
assays is time-consuming — they can take more validation activities are carried out to ensure safety
than a year. Planning ahead is essential. In addi- and the manufacturing process should be control-
tion to host cell proteins, other impurities for lable. Phase 2 manufacturing processes should be
which validated assays may be required include well-controlled and better understood with vali-
any toxic or potentially immunogenic substances. dated assays that demonstrate that control. Phase 3
Phase 2. During phase 1 and 2 clinical trials, should be very well controlled and full process
the process is generally improved. Both upstream validation should take place during this phase.
and downstream processes change, analytical
methods are usually transferred to QC during this PETER WATLER
stage, and potency assays validated. No specific Pilot Plant Engineering, Amgen Inc.
process validation activity is required at this Process validation demonstrates the consistency
stage, unless the changes made to improve the of multiple batches at full-scale. The validation
process have the potential to affect the results of shows that the process is operated in a consistent
previously performed validation studies for steril- manner and that contaminants are reproducibly
ity, virus clearance, and specific impurity reduced to acceptable levels. That is accomplished
removal. However, while the process is being by monitoring those parameters that demonstrate
optimized, assay validation efforts should consistent operation of the process, consistent
continue so that process validation at pilot or full formation of the product, and consistent removal
scale for licensure can take place during phase 3. of the contaminants. Processes can be validated
Phase 3. Sometime before phase 3, the process through two distinct metrics: operational parame-
is, hopefully, finalized to avoid bridging studies in ters (inputs) and performance parameters (out-
the clinic. Assays not yet validated must now be puts). Operational parameters are process control
validated so that process validation required for set-points for variables such as flow rate, tempera-
biologics licensure can be completed during phase ture, and concentration. These parameters define
3. This is the phase of heavy-duty validation and the process recipe and are used to demonstrate
requires equipment qualification, then process val- that the facility, equipment, and staff can execute
idation during three or more consecutive batches. the process consistently. Performance parameters
Cleaning validation and lifetime studies for chro- reflect the outcome of a given step and indicate
matography columns are important validation that the process gave the desired result.
elements. Clearance studies to remove host cell Process validation is an investment in future pro-
proteins, DNA, viruses, and other impurities may duction because it sets the bar by which the process
eliminate lot release testing. Some clearance stud- will be judged at future inspections. Process valida-
ies are performed at a smaller scale than manufac- tion is the point at which the science of the process
turing, so the small-scale model must be validated. can be explained to regulatory agencies. Following
Once the process is finalized, viral clearance stud- validation, regulatory agencies are looking for con-
ies are carried out again. At this stage, however, sistent process operation within the ranges and
22 BioPharm OCTOBER 2002
 meeting the criteria specified in the validation. Sci- Operational control parameters are input variables
entific arguments explaining operation and perfor- with set-points or ranges specified in the manufac-
mance outside of specified ranges are likely to fall turing procedure to define how a process is
upon deaf ears. Therefore it is essential to generate executed. Control charts can be used to demon-
bench and pilot scale process data that establish key strate that the set-points are consistently achieved
parameters and their acceptance criteria before within the specified ranges. Typical operational
process validation in the GMP facility. parameters include pH, raw material quantities, re-
When to validate. Process validation is action times, flow rates, temperature, and pressure.
expensive, involving copious sampling, extensive For example, the control chart in Figure 1 shows
analysis, and detailed documentation. Because of that operator skill, manufacturing equipment, and
the complexity and cost, process validation is best written instructions are sufficient to adjust oxida-
performed during phase 3 trials following the deci- tion pH to the same set point from lot-to-lot.
sion to file a Biologics License Application (BLA). Performance parameters are output variables that
This “delayed” strategy offers several advantages. reflect the outcome of a given step, indicating that
The process is better understood, which means that the process performed as expected. Validation
key parameters and acceptance criteria can be bet- should demonstrate that the process is capable of
ter specified. In addition, the commercial process is consistently removing three classes of contami-
in place at this point, and validation does not have nants: process-related, host cell-related, and prod-
to contend with process changes, adjustments, and uct-related. Process-related contaminants are
inexperience that can lead to deviations from the reagents required by the process (for example,
validation protocol. By the time products enter the guanidine, glycerol, and antifoam). Host cell-
regulatory review phase, the likelihood of product related impurities are derived from the organism
success increases to 90%, meaning that there is less used to generate the product (for example, nucleic
risk that process validation will have been for acids, CHO proteins, and endotoxins). Product-
naught. Then too, although process validation is a related contaminants are variants and isoforms of
required component of a BLA submission, it is not the target protein (for example, oxidized, deami-
required for clinical trials. dated, aggregated, and clipped forms). The valida-
What to validate. Demonstrating process consis- tion should also demonstrate at which step the
tency requires multiple, full-scale batches. Three contaminant is removed and at what point in the
to five consecutive purification runs from three process it meets acceptance criteria. Table 1
consecutive fermentation lots should generate suf- shows that E. coli proteins were reduced by the
ficient consistency data. Process validation can be
filtration step, significantly reduced at cation ex-
made more efficient and more consistent by using
change 1, and further reduced to below the
templates that identify key input and output para-
meters for a unit operation. For example, consis- acceptance criteria by cation exchange 2.
tent operation of a tangential-flow filtration step Process consistency is also shown by monitor-
can be shown by monitoring cross-flow rate, ing yield and product concentration at each step.
transmembrane pressure (TMP), temperature, re- With sufficient advance planning, earlier process
tentate tank volume, and diafiltration volume. data, and careful execution, the data from the
Consistent performance of this step can be shown validation should demonstrate that the manufac-
Table 1. Validation results by monitoring excipient removal, protein contam- turing procedure consistently and effectively
demonstrating that E. coli inant removal, product concentration, product re- yields product that meets specifications. It will
proteins were reduced by the covery, pH, and conductivity. Such parameters also demonstrate that the process can be operated
filtration step, significantly at demonstrate that the step achieved its purpose in consistently and that the process, product, and
cation exchange 1, and further the process. These validation templates for unit
host cell contaminants are reduced to acceptable
reduced to below acceptance operation can be customized to address any
levels.
criteria by cation exchange 2. unique specifics of the process or product.
—Courtesy of Peter Watler

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

24 BioPharm OCTOBER 2002

 RHONA O’LEARY variability. Such runs may include running the
Process Sciences, Genentech, Inc. process at the extremes of its operating ranges,
Manufacturing processes for biologicals are testing different feedstocks, using different lots of
continually optimized as the product moves through resins, testing cell culture components or peptone
various stages of development, from the preclinical lots, among others. Sometimes we perform these
stage through the investigational new drug (IND) runs at the 400 liter scale, more often at lab scale.
filing to the common technical document (CTD) fil- The timing of these scaled-down runs also
ing and on to approval. The early-stage product may provides the scientist with the ability to incorpo-
have a basic and somewhat unpolished manufactur- rate last minute changes that result from data
ing process. By midstage the process is optimized gathered during these runs to further optimize
and becomes more robust. By the end of develop- and improve the robustness of the process.
ment, the final manufacturing process is fully char- It is sometimes easy to overlook the impor-
acterized, validated, and qualified. tance of a complete evaluation of the robustness
Development validation activities. At Genentech, of the process for both cell culture (or fermenta-
our goal is to have the manufacturing process tion) and recovery before finalizing the process.
completely finalized before the manufacturing Although it is hoped that the manufacturing
campaign that supplies the phase 3 clinical trials. process will run at its optimum operating condi-
That means that the critical process parameters tions on a routine basis, knowing how the process
have been identified, the operating ranges have works under conditions beyond its optimum set
been set, and the process is deemed robust points (such as flowrate, pH, and temperature
enough to tolerate the manufacturing environ- excursions) is important. If the process is running
ment and produce marketable product. close to the edge of failure, the time to discover
Phase 1 process validation activities. Validation and correct that is before finalizing the process.
evolves along with development of the process, The studies performed at this stage require fully
and different types and levels of validation are developed analytical assays, and those assays are
appropriate at the various developmental stages. examined to ensure the process meets the accep-
At the time of IND filing, the only validation tance criteria set in the Certificate of Analysis.
requirement is a demonstration that the process Although this phase of development is not
removes retroviral particles. If modifications are considered part of the official validation process,
made as the process moves through development, if well documented, it can become the foundation
the process for removing retroviruses must be on which the process validation program is based.
reassessed. Complete process validation is not Final process with product in phase 3 trials. Genen-
required until the CTD is filed. tech finalizes the manufacturing process for the
Phase 2 and 3 process validation activities. Valida- start of the phase 3 clinical trials — that is, the ma-
tion is performed after the process is fully devel- terial that is made in the manufacturing campaign
oped and finalized. Process validation can be a to supply the phase 3 trials is from the fully devel-
combination of manufacturing-scale and scaled- oped, fully characterized, and finalized manufac-
down studies. The qualification runs are performed turing process. No process changes are expected
on validated equipment, at full manufacturing beyond this stage. At this point, the analytical as-
scale, in the facility in which the product is to be says are also finalized and are used to release the
routinely manufactured. The validation studies material to the clinic. Once that product is manu-
performed at manufacturing scale typically include factured, some of that material can be used to sup-
the validation of the process to clear impurities port process validation. Validation of the analyti-
(host cell proteins and DNA) and small molecules cal methods and the definition of the control
and the cleaning of resins and membranes. system begin at this stage.
Scaled-down studies typically include the vali- Generic process validation activities. Some of the
dation of resin lifetimes, in-process hold times, validation studies we perform are common across
buffer stability, virus validation, harvest criteria, many of our processes: viral validation, resin and
filter extractables, resin leachables, and cell age. membrane sanitization and storage, buffer stability,
Genentech has found that a combination of filter extractables, and resin leachables. Rather than
manufacturing-scale and scaled-down studies continuing to repeat these studies time and time
provides the best overall process validation plan. again, we developed matrices to support a wide
Before phase 3, evaluate robustness. Before the variety of conditions — conditions that typically
process is finalized several scaled-down runs are reflect our most common operating parameters.
performed to evaluate process robustness. These We try to assess the worst-case situation, after
runs provide an adequate history of the process’ which all other conditions are well covered by
26 BioPharm OCTOBER 2002
 those conditions. For example, by testing the ability be taken, distinct commonalities underlie all the
of a particular resin matrix (such as Sepharose) to approaches described. The validation approaches
be sanitized in the presence of a protein mixture, discussed here seek to overcome challenges often
we can use that study to support the same sanitizing encountered during development of a validation
agent for all other similar protein mixtures, on sim- program for biotechnology products. An accept-
ilar resin matrices. If we did not adopt such an ap- able validation program addresses the following
proach, it is likely that the huge workload currently difficulties encountered by others:
required for process validation would be doubled. • Lack of overall strategy
The ability to develop such large matrices for our • Failure to consult early with the regulatory
various processes is reflective of the large number authorities
of products that have been validated in Genentech • Inadequate product definition
• Failure to follow CGMP regulations
in recent years. A smaller company might not have
• Poorly defined cell bank genealogy
the luxury of developing the necessary database of
• Inadequate analytical procedures
information to support a matrix design, but it could • Too many process changes after validation
use the worse-case scenario to cover not just a sin- activities have been completed
gle product, but also many future products. • Not enough process characterization or
Process validation is a combination of efforts. The qualification
importance of documentation is also frequently • Commencing process validation too early
overlooked. Throughout all phases of develop- • Inappropriate acceptance criteria.
ment, we try to capture all development work and
all decisions in development reports. These VALIDATION PERSPECTIVES
reports, in turn, support any questions that the We hope that this article is useful to those in the
regulatory agencies may have about how a biopharmaceutical industry who are currently in-
process was developed or why certain decisions volved in validation activities and especially use-
were made. These reports are stored in a database ful to those who are planning to start validation
for easy access. They also serve as a resource into activities in the near future.
the development history if an employee leaves This article series presents opinions and
and as guidance for new employees. viewpoints of some of the industry experts on issues
The biostatisticians are also on hand to assist that are routinely faced in process development and
in the design of characterization studies and to manufacturing of biopharmaceuticals.
evaluate process robustness, frequently using
fractional factorial designs. Sometimes these REFERENCES
(1) FDA, “Current Good Manufacturing Practices;
results are incorporated into resin lifetime Proposed Amendment of Certain Requirements for
studies. Quality assurance, analytical chemistry, Finished Products: Supplementary Information,”
Federal Register 61(87), p. 20104 (3 May 1996).
and process development groups work closely Available at www.fda.gov/cber/genadmin/cgmp.pdf.
together to ensure that all appropriate studies are (2) Sofer G. and Hagel L., “Validation,” Handbook of
Process Chromatography: A Guide to Optimization,
performed, that the data are audited for accuracy, Scale-up, and Validation (Academic Press,
and that reports are generated and approved. New York, 1997), pp. 119–243.
Developing a manufacturing process that will (3) Martin-Moe S. et al., “Validation of Critical Process
Input Parameters in the Production of Protein
ultimately receive regulatory approval and yield a Pharmaceutical Products: A Strategy for Validating
marketable product requires a development plan New Processes or Revalidating Existing Processes,”
PDA J. Pharm. Sci. Technol. 54(4), 315–219
that follows regulatory requirements. The devel- (July/August 2000).
opment plan must have goals that are met during (4) Seely R.J. et al., “Defining Critical Variables in Well-
Characterized Biotechnology Processes,” BioPharm
development before it moves onto the next stage. 12(4), 33–36 (April 1999).
In addition, we have found that having standard- (5) Noferi J.F., Arling E.R., and Worden D.E., “Beyond
Corresponding author QSIT,” BioPharm 15(4), 40–44, 67 (April 2002).
ized study designs, protocols, and reports have
Anurag S. Rathore, GG3K, (6) Center for Drug Evaluation and Research, Guideline on
700 Chesterfield Parkway North,
led to successful validation programs. The the Preparation of Investigational New Drug Products
program described has undergone four FDA (Human and Animal) (FDA, Rockville, MD, March
Chesterfield, MO 63017 1991). Available at www.fda.gov/cder/guidance/
[email protected]; preapproval inspections since 1998. old042fn.pdf.
Joseph F. Noferi, Esq. and (7) European Commission, “Annex 13 Manufacture of
Investigational Medicinal Products” Medicinal Products
Edward R. Arling are also at MEETING THE CHALLENGE for Human and Veterinary Use: Good Manufacturing
Pharmacia Corporation, In this article, experts from different companies in Practices, Vol. 4, Draft 1, F2/AN D(2001)
www.pharmacia.com. our industry have outlined the strategy that they (Enterprise Directorate-General, Brussels, Belgium,
Gail Sofer is at BioReliance, November 2001).
use to arrive at a validated process that can be (8) “General Biological Products Standards: Sterility,”
Peter Watler is at Amgen Inc., and Code of Federal Regulations: Food and Drugs,
Rhona O’Leary is at Genentech, Inc.
submitted to regulatory agencies. Although it is
Title 21, Part 610.12 (U.S. Printing Office, Washington,
interesting to see the different approaches that can DC, revised 1 April 2001). BP
28 BioPharm OCTOBER 2002