Validation of API Manufacturing

This article Reprinted from The Official Journal of ISPE

presents a PHARMACEUTICAL ENGINEERING® March/April 2005, Vol. 25 No. 2

procedure for

Procedure for the Validation of

the validation of
biological API
manufacturing
processes. It
details each
Biological Active Pharmaceutical
step of the
procedure for
Ingredients (APIs) Manufacturing
biological
processes
Processes
applications, by Josée Ethier
and covers the
regulatory
requirements for
biological APIs.

T
Introduction • processes for which the product can not be
he validation of biological Active Phar- fully characterized and/or verified
maceutical Ingredients (APIs) manu- • processes for which the complete product
facturing processes is more complex characterization is very expensive or impos-
compared to standard chemical APIs, sible to perform (in this case, product quality
due to the lack of both product and process is assessed using process validation)
characterization. The following definition of • processes for which the product quality is
Process Validation can be found in ICH Guide- critical and for which a slight variation in its
line Q7A:1 “Process validation is the documented composition may result in severe reactions
evidence that the process, operated within es- in the patient
tablished parameters, can perform effectively
and reproducibly to produce an intermediate or Validation of processes involving micro-organ-
API meeting its pre-defined specifications and isms probably present the highest difficulty
quality attributes.” degree due to the important variability of living
The objective of process validation is to: organisms. Variations in the behavior and pro-
ductivity level of the micro-organism used may
• demonstrate process stability and reliabil- result in differences in the composition of the
ity culture media at the end of the production,
• evaluate the impact of a variation in a criti- which can, in turn, impact the purification
cal process parameter on product quality process. Processes involving living micro-or-
ganisms are more sensitive to operating condi-
Not all manufacturing processes require vali- tions and may show larger variability from one
dation. The examples mentioned hereafter usu- batch to another. Therefore, biological processes
ally require process validation: are more difficult to validate, and require more
considerations. In addition, as
Table A. Recommended stated by Kirrstetter,2 the raw ma-
Recommended Procedure for the Validation of API Manufacturing
procedure for the
Processes terials used in biological APIs
validation of API
manufacturing processes may re-
manufacturing • Formation of a Validation Committee
processes. • Description of the process manufacturing steps, utilities, services and sult in microbiological contamina-
equipment tion, and more stringent controls of
• Critical Analysis equipment, utilities, and services
• Identification of critical process parameters and validation requirements
• Elaboration of a Validation Master Plan (VMP) are required to minimize the risk of
• Redaction of process validation protocols contamination. Slight variations of
• Protocols execution the manufacturing conditions may
• Reports
be observed when undesired un-

©Copyright ISPE 2005 MARCH/APRIL 2005 PHARMACEUTICAL ENGINEERING 1

Validation of API Manufacturing
characterized substances are found in the final product. and quality characteristics. Process validation was originally
Biological APIs are difficult to characterize and, are influ- not required for clinical trial material manufacturing due to
enced by the numerous parameters involved in their manu- the numerous modifications brought to the procedures dur-
facturing process. ing process fine tuning and due to the limited number of
As the biotechnology sector faces important growth with batches produced at this stage.
the increasing number of expiry patents,3 the number of The Annex 13 of European current Good Manufacturing
biological API manufacturing processes to validate will in- Practices4 has been modified in 2003 to require process
crease in the next few years, enhancing the importance of a validation for clinical material manufacturing.
structured procedure for process validation. The Canadian regulatory agency has decided to adhere to
the new European directive.5 FDA still maintains its position
Process Validation Requirements and requires process validation for commercial batches only.
Process validation is generally required prior to product However, the US FDA appropriated control and monitoring
commercialization to demonstrate that the process consis- of the critical process parameters and demonstrated trace-
tently results in a product having the required specifications ability for clinical lots.
Process validation will now be required earlier, prior to the
completion of process development. The validation procedure
presented here constitutes a systematic approach that also
can help in process development and optimization, and allow
the initiation of process validation while its development is
still in progress.

Main Validation Steps

The proposed validation procedure for biological API manu-
facturing processes is described here. The procedure also can
be applied to standard (non-biological) APIs.
Process validation may follow Installation, Operation,
and Performance Qualification of equipment, utilities, and
services, as a process can be developed and implemented in
an existing premise using qualified equipment. However, the
procedure presented here is more general and can be applied
to new processes implemented in new premises using new
services and equipment. As mentioned above, the proposed
procedure can be initiated in early development stages and
help in the identification of critical process parameters and
improve process understanding. It also ensures that an
appropriate distance has been taken to visualize the process
as a whole and that equipment, utilities, services, process
environment, and parameters impact have been properly
considered and evaluated.
The main validation steps are described below and sum-
marized in Table A.

Formation of a Validation Committee

First, a validation committee should be formed and include
people from the following departments: Quality Assurance,
Engineering, Production, and Development (R&D). Repre-
sentatives from additional departments could be included if
required (Laboratories, Logistics, and Regulatory Affairs, as
an example), depending on the nature of the process to be
validated and the organizational structure. This committee
is responsible for the management and the execution of
process validation and to ensure compliance with regulatory
authorities. The committee should meet on a regular basis as
long as process validation is not completed and be involved
earlier during process development.
The modified Annex 13 to the European cGMPs4 requires the
Figure 1. Typical flowchart for a cell culture process (and
functional blocks). presence of a Qualified Person (QP) that plays a similar role.

2 PHARMACEUTICAL ENGINEERING MARCH/APRIL 2005 ©Copyright ISPE 2005

 Validation of API Manufacturing
Process Description Critical Analysis Results
Once the committee is formed, a complete and detailed
description of the manufacturing process can be established • Identification of the critical process parameters that must be monitored
and controlled.
and should include the following steps: • Identification of validation requirements: IQ, OQ, PQ, analytical methods,
cleaning, operation, inactivation, process
1. A Process Flow Diagram (PFD) should be prepared and • Identification of the required characteristics: raw materials, final
product, in-process samples, etc.
functional blocks should be defined, each block having a • Identification of “markers” for process comprehension and follow-up.
clear and distinct function. Figure 1 illustrates a typical
flow chart of a cell culture process and the corresponding Table B. Description of the critical analysis results.
suggested functional blocks. Typical unit operations in-
volved in a biological API manufacturing process are: probability of detection (D), the possible occurrence of the
fermentation, inoculation, cell harvest, filtration, cen- problem (O), and the gravity of the resulting consequence (G)
trifugation, diafiltration, chromatography, formulation, are evaluated (refer to the Failure Mode and Effect Analysis
filling, freeze drying, sterilization. Transfer steps should procedure).6 The critical parameter (C= D*O*G) can then be
be included in the PFD. calculated, and represents a quantitative measure of the
critical of each possible source of non-conformity. Appropri-
2. The equipment, utilities, and systems required can be ated solutions can be proposed for each possible source
listed for each functional block. The corresponding speci- identified to reduce the critical (C) to an acceptable level,
fications and validation status (if any) should be specified. either by improving the detectability of the problem (and
reduce factor D) and/or reducing the possible occurrence of
3. The In-Process Controls (IPC) with the corresponding the problem (and reduce factor O). The calculation of the
tolerances should be identified. Critical Factor (C) can be performed again, considering the
solutions proposed.
4. Product final specifications with the corresponding toler- Most of the time, the solutions proposed will allow the
ances should be identified. identification of critical process parameters and the defini-
tion of validation requirements: perform cleaning or design
5. The analytical methods, required instruments, and final validation, add an in-process control, control the environ-
specifications with tolerances should be specified. The ment (by having a validated HVAC system as an example),
corresponding validation status (if any) should be speci- etc. Table C illustrates an example of a part of a Critical
fied. Analysis (inoculation functional block). Four possible sources
of nonconformities have been identified to illustrate the
6. The process parameters which impact product quality procedure (not exhaustive):
should be identified.
• the inoculum cell concentration could be out of specifica-
7. Process inlets and outlets should be identified and illus- tion
trated using a cause and effect diagram (“Fish Bone,” see • the inoculum could be contaminated
Figure 2). The “Fish Bone” illustrates the impact of a • the transfer of the inoculum to the bioreactor could be
variation in process parameters on product specifications deficient
and helps in process understanding. • the bioreactor cleaning could be deficient

Critical Analysis The impact of each source is then identified and evaluated.
The validation of a Drug Product (DP) manufacturing process An out of specification inoculum cell concentration would
requires the qualification of each manufacturing step whereas result in unusual growth kinetics that could be detected using
the validation of an API manufacturing process requires the an in-process control prior to bioreactor inoculation. Accord-
qualification of the critical manufacturing steps only. ingly, the possibility of detection (D) would be rated at 3, the
To help in the identification of validation requirements occurrence (O) at 2, and the gravity (G) at 3 since the
and critical process parameters (and complete the lists pre- productivity level would be affected. The resulting critical
pared during the previous step) a Critical Analysis (or Risk factor (C) would then be 18 (3*2*3). The proposed corrective
Analysis) needs to be performed - Table B. Most product action is the implementation of optical density verification
nonconformities result from either errors performed during prior to inoculation. The probability of detection (D) would
manufacturing or from variations in process parameters or then be rated at 1, the occurrence (O) at 1, resulting in a
immediate environment. critical factor (C) of 3 (1*1*3).
A Critical Analysis consists of the identification of those The final critical factor evaluation (rightmost column in
possible sources of errors and process variations that could Table C) allows the identification of most critical process
result in product non-conformity. Once the sources are iden- parameters with a series of corrective actions that are re-
tified, their impact on product quality and/or process safety quired to keep the risks at a minimal level and ensure product
(including environment and operators) are evaluated. The consistency and quality, in addition to personnel safety and

©Copyright ISPE 2005 MARCH/APRIL 2005 PHARMACEUTICAL ENGINEERING 3

Validation of API Manufacturing

Figure 2. Example of a cause and effect diagram (“Fish Bone.”)

environmental considerations. The corrective actions could ing and labelling.

be an in-process control, a standard operating procedure, a
validation, a verification, personnel training, or any other Identification of Critical Process Parameters
action required to reduce the occurrence of the problem and/ and Validation Requirements
or improve its detectability. The Critical Analysis is a very efficient tool that allows the
Generally, the following process steps should be identified identification of critical process parameters and the identifi-
as critical for biological processes and be included in the cation of validation requirements. Any possible source of non
Critical Analysis: equipment and instrument cleaning, raw conformity is analyzed and solutions to control variability are
material characterization (media, components, and cells), proposed until the critical factor (C) is reduced to its minimal
weighing, solution, and media preparation, inoculation prepa- value. The acceptable operation range for each critical pro-
ration, bioreaction (pH, dissolved oxygen, stirring,... ), har- cess parameter also can be defined. The risk to forget an
vesting, refolding (if required), purification, formulation (when important parameter or to omit a required qualification is
applicable), filling, freeze-drying (when applicable), packag- reduced to its minimum, and benefits are reflected on the

Manufacturing Step Source of Non-Conformity Impact on Quality/Safety D O G C Corrective Action D O G C

Out of specification cell Unusual growth kinetics 3 2 3 18 - In-process control of optical 1 1 3 3
concentration of the density before inoculation
inoculum
Contamination while Presence of contaminant in 3 2 4 24 - SOP for inoculation 1 1 4 4
inoculation the bioreactor - Installation of a laminar flow
hood
- SIP of inlet ports to be
Bioreactor developed
Inoculation - Monitor optical density during
growth phase
Forget to open inlet port Loss of inoculum 1 2 3 6 - SOP for inoculation 1 1 3 3
Bioreactor cleaning deficient Presence of contaminant in 3 2 4 24 - Cleaning validation required 1 1 4 4
the bioreactor - TOC/swab prior to bioreactor
inoculation
LEGEND D = probability of detection O = possible occurrence of the problem G = gravity of the resulting consequence C = critical factor
Table C. Example of a critical analysis.

4 PHARMACEUTICAL ENGINEERING MARCH/APRIL 2005 ©Copyright ISPE 2005

 Validation of API Manufacturing
whole validation team, for which the validation/development List of Biological Process Validation Pre-Requisites
objectives are clear even before the redaction of the Valida-
tion Master Plan. Some corrective actions could involve • Defined raw materials
• Defined equipment, utilities and services
process development and/or engineering design. • Defined process parameters and acceptance range
Some “markers” can be chosen among the critical process • Standard Operating Procedures (SOPs)
parameters identified, based on their fair representation of • Process documentation
• Critical analysis (risk analysis)
the process condition, to limit in-process control related costs • Validation Master Plan
during process validation and commercial manufacturing • IQ/OQ/PQ for equipment, utilities and services with closed deviations and
(methods development and validation, as well as time and nonconformities
• Validated analytical methods
consumables). The optimal value of each marker should be • Qualified instruments
identified and accompanied with the corresponding analyti- • Maintenance and calibration programs
cal method, sampling procedure, specification and tolerance, • Change control program
• Identification of critical process parameters and markers, with their
and a reference standard (if required). Markers are used respective acceptance range
mainly for not-well characterized biologics (mainly complex • Bioassays defined and qualified
vaccines, blood products, viral vectors, and cell therapies)7 to
Table D. List of the information required prior to process validation.
follow the product quality and consistency.
Of all unit operations, cell culture is the processing step • a description of all manufacturing steps and in-process
resulting in most important variations. Living micro-organ- controls, including packaging operations and acceptance
isms can show slight differences in growth characteristics limits for each process parameter
and can release proteins and debris in the culture media that • the final product specifications
affect the growth curve. The Critical Analysis is therefore of
major importance for biological processes validation. The following verifications should be included for a biological
Bioassays also are critical and should be qualified since API manufacturing process:
they allow the determination of the product’s tertiary struc-
ture and activity.7 Bioreaction
As mentioned above, bioreaction is the most variable manu-
Validation Master Plan facturing step as micro-organisms are involved.7 The process
Now that the critical process parameters, critical manufac- duration, temperature, pH, conductivity, nutrients, and prod-
turing steps, markers, and validation requirements are iden- uct concentration should be characterized. The establish-
tified, the preparation of a Validation Master Plan should ment and maintenance of the Working Cell Bank (WCB) is
constitute the normal following step of the process validation not covered by the ICH Q7A cGMP. However, it is recom-
procedure. mended to perform the characterization of the Master and
For biological processes, cell stability and purity (includ- Working Cell Bank (MCB/WCB) to identify the history of the
ing viral clearance) also need to be validated. The stability of organism (media used, storage conditions, pressure factor
the genetically modified micro-organism must be demon- used, etc.) and demonstrate the absence of virus and myco-
strated, and the maximum cell division number identified. plasma.
This can determine the longest continuous cell culture that Growth kinetics, product formation rate, yields, cell den-
could be performed before mutation or transformation oc- sity, stirring conditions, and optical density needs to be
curs. Viral clearance is usually performed by inactivation characterized for successful process scale-up and validation.
(using pH, solvents and/or detergents, or heat) or removal The demonstration that the bioreactor sterility can be main-
(using filtration and/or chromatography). tained during normal operation of the bioreactor (sampling,
For yeast and bacterial cultures, viral inactivation is not addition of antibiotics, etc.) should be included in the perfor-
required since these micro-organisms are usually not in mance qualification.
contact with viruses nor TSE (Transmissible Spongiform Cleaning procedures are critical for bioreactors as they
Encephalopathies). must eliminate any risk of cross contamination. It is strongly
recommended to dedicate bioreactors to bacteria, yeast, or
Process Validation Protocol Redaction mammalian cell culture to avoid contamination.
The structure of process validation protocol should be the Continuous operation of a bioreactor (including perfusion)
same as for standard pharmaceutical processes. However, for is often preferred to batch and fed-batch operation for mam-
biological processes, particular verifications are required to malian cell culture due to higher product yields obtained.
ensure process consistency and reproducibility. However, validation is more difficult to perform on continu-
The process validation protocols should include: ous systems since the production can last for months (com-
pared to days/weeks for batch and fed-batch cultures). The
• a general decisional process flow chart cell line stability can be difficult to demonstrate. Cell muta-
• a list of the equipment (both critical and ancillary) and tion can occur and can be difficult to detect, resulting in
instruments used, including their identification number complex validation. It may be difficult to determine when a
and calibration state (if required) lot begins and finishes, and the production of three distinct

©Copyright ISPE 2005 MARCH/APRIL 2005 PHARMACEUTICAL ENGINEERING 5

Validation of API Manufacturing
batches can take a year to perform. In addition, if a contami- Purification efficiency is closely related to chromatogra-
nation occurs, it will be detected only few days later, contami- phy operation parameters such as ionic strength, pH and flow
nating part of the lot. rate of the elution solution, the column diameter, the bed
Finally, it is important to characterize product stability for height, and both the impurity and protein concentration of
the holding period between bioreaction and harvest. The the inlet solution to be purified. The efficiency is also depen-
maximum holding duration and storage conditions need to be dant on packing quality. The evaluation of the Height Equiva-
specified. lent to a Theoretical Plate (HETP) should be performed
following each column packing to demonstrate the absence of
Harvest and Recovery channeling and assess packing quality. Finally, the non-
Most unit operations used in harvest and recovery steps are specific binding of the protein to the chromatography resin
centrifugation, microfiltration (tangential, diafiltration), dis- should be quantified to confirm resin quality and safely
integration, and refolding (for inclusion bodies). At the end of process the target protein using this media.
the recovery step, the product is most of the time inactivated For ultrafiltration operations, the maximum flow rate and
using sterile filtration to eliminate any risk of contamination membrane pore size should be characterized and an operat-
of the equipment used for downstream processing (mainly ing range specified. The non-specific binding of the protein on
chromatography media) and to contain viable micro-organ- the membrane should be quantified as for chromatography
isms into the production room. resins.
The efficiency of each filtration step should be character- The use of disposable filtration units eases the validation
ized and verified. The filter integrity should be demonstrated of filtration steps and reduces cleaning validation efforts.
before and after filter use using standard methods. Filter
integrity also should be demonstrated before and after the Validation Execution
sterile filtration step. Sterile filtration qualification should ICH1 requires three consecutives runs for prospective and
include the demonstration of absence of viable particles concurrent process validation and from 10 to 30 consecutive
following filtration. runs for retrospective process validation. More runs may be
The acceptance range of each critical parameter needs to required for complex processes. For prospective and concur-
be specified for each unit operation involved in harvest and rent validation, three different lots of raw materials should
recovery operations. Refolding step should be defined and its be used. Each run needs to be completely independent from
consistency demonstrated. the others: the inoculum and culture media should be fresh
The endotoxin and protein concentration shall be charac- and the equipment cleaned using cleaning procedures in
terized at the end of the recovery step. And as previously place between each run.
specified, the hold period between recovery and purification The application of the validation procedure described here
needs to be characterized (maximal duration and storage ensures that all the information required to perform process
conditions) to ensure product integrity and stability. validation is available - Table D.
Product characterization can be performed using the speci-
Downstream Processing fied analytical methods. Data needs to be compiled and
Unit operations usually involved in downstream processing statistically analyzed.
are chromatography (gel filtration, ion exchange, affinity, Finally, a validation report comparing the product speci-
hydrophobic interaction), extraction, and ultrafiltration. Five fications with product characterization should be prepared.
to 10 purification steps are normally required to reach the Deviations and nonconformities should be summarized in
required product purity level. this report.
Chromatography requires several verifications to demon- A change control procedure is required to follow and
strate process consistency and reproducibility and to guar- control any modification performed on the process and/or its
anty product quality. Chromatography media is very difficult utilities and services. Each modification should be evaluated
to clean, and is often used for different applications. Cleaning commonly by the Validation Committee or by the Qualified
validation is critical for chromatography columns and media. Person, depending on the structure of the society. Process re-
The Total Organic Carbon (TOC) of the final rinsing water validation may be required when a major modification is
should be measured and be kept below the specification to performed on an equipment, a service, or an utility, on the
demonstrate the absence of residues of cleaning agents and premise itself or on the manufacturing procedure. Any devia-
confirm the absence of resin leaching. The gel lifetime also tions of the markers identified should result in a process
should be determined. A procedure is required for column investigation to quickly identify the source of the problem
sanitization and needs to be rigorously followed by the per- and reduce the risk of getting out-of-specifications product.
sonnel.
Chromatography resins should be considered as a raw Conclusion
material and therefore requires full characterization as well The proposed procedure may seem time-consuming and heavy
as acceptance specifications. Resins properties can vary con- to implement. However, considerable reduction of both devel-
siderably from one batch to another and the acceptance opment and validation steps duration results from the use of
specification range should consider such possible variability. this procedure. Qualification can be initiated prior to the

6 PHARMACEUTICAL ENGINEERING MARCH/APRIL 2005 ©Copyright ISPE 2005

 Validation of API Manufacturing
completion of process development and accelerate product
marketing.

References
1. ICH Harmonized Tripartite Guideline Q7A: GMP Guide
for Active Pharmaceutical Ingredient, November 2000.
2. Kirrstetter, R., “Requirements for Process Validation of
Biotech Active Pharmaceutical Ingredients (APIs),” Phar-
maceutical Engineering, Volume 22, No. 3, May/June
2002.
3. ABN-AMRO, “Generic Biologics: The Next Frontier,” June
2001.
4. EU Guide to GMP (Volume 4), Annex 13: Manufacture of
Investigational Medicinal Products, July 2002.
5. Health Canada: Annex 2 to the Current Edition of the
Good Manufacturing Practices Guidelines: Manufacture
of Drugs Used in Clinical Trials, April 2004.
6. FMEA information centre (Failure Mode and Effect Analy-
sis), [Link]
7. Nash, Robert A. and Wachter, Alfred H., “Pharmaceutical
Process Validation,” Third Edition, Marcel Dekker, Inc.,
New York, 2003.
8. EU Guide to GMP, Annex 15: Qualification and Valida-
tion, July 2001.
9. ICH Harmonized Tripartite Guideline Q6B: Specifica-
tions: Tests Procedures and Acceptance Criteria for
Biotechnical/Biological Products, March 1999.

Acknowledgement
The author would like to thank Validapro Inc. for their
support in the preparation of this article. Any questions
regarding the content of this article may be addressed to
Ethier or Validapro Inc. ([Link])

About the Author

Josée Ethier is Project Manager at ProMetic
BioSciences Inc. She has a chemical engi-
neering degree from École Polytechnique de
Montréal (1992) (biotechnology specialty) and
a Masters degree in chemical engineering
processes also from École Polytechnique de
Montréal (1996). She is involved in biological
and chemical process design and validation
as well as cleanroom design. She has more than 13 years of
experience in the biotechnology and pharmaceutical industry
in the design, process development, User Requirement Speci-
fications (URS) development, project management, and criti-
cal analysis. Ethier has filed two patent applications and is
member of the Quebec Order of Engineers. She can be
contacted by e-mail: [Link]@[Link] or by telephone:
1-514/341-2115.
ProMetic BioSciencesInc., 8168 Montview, Mont-Royal,
Quebec H4P 2L7, Canada. [Link].

©Copyright ISPE 2005 MARCH/APRIL 2005 PHARMACEUTICAL ENGINEERING 7