Technical Report No.

60
Process Validation:
A Lifecycle Approach

Paradigm Change in
Manufacturing OperationsSM

Bethesda Towers
4350 East West Highway
Suite 200
Bethesda, MD 20814 USA
Tel: 1 (301) 656-5900
PDA Task Force on Technical Report No. 60: Process Validation: A Lifecycle Approach
Authors

Scott Bozzone, Ph.D., Chair, Pfizer, Inc. Raj Jani, Baxter Healthcare Corporation
Harold S. Baseman, Co-Chair, Valsource, LLC Peter F. Levy, PL Consulting, LLC
Vincent Anicetti, Parenteral Drug Association, Keck Michael Long, PhD Concordia Valsource, LLC
Graduate Institute
John McShane, Roche-Genentech, Inc.
John A. Bennan, Ph.D., ComplianceNet, Inc.
Victor G. Maqueda, Sr., Consultant
Michael N. Blackton, Imclone Systems, Inc.
José Luis Ortega, Pharma Mar S.A. Sociedad Unipersonal
Vijay Chiruvolu, Ph.D., MBA, Amgen, Inc.
Elizabeth Plaza, Pharma-Bio Serv, Inc.
Rebecca A. Devine, Ph.D., Consultant to the Bio-
pharmaceutical Industry Praveen Prasanna, Ph.D., Shire Human Genetic
Therapies, Inc.
Stephen Duffy, Covidien, LLC
David Reifsnyder, Roche-Genentech, Inc.
Panna L. Dutta, Ph.D., The Medicines Company
Kurtis Epp, BioTechLogic, Inc. Markus Schneider, Ph.D., Novartis Pharma AG

Igor Gorsky, Shire Pharmaceuticals, Inc. Iolanda Teodor, Baxter Healthcare Corporation
Norbert Hentschel, Boehringer Ingelheim Pharma Mark Varney, Abbott Laboratories
GmbH & Co., KG Alpaslan Yaman, Ph.D., Biotech, Pharma and Device
Pedro Hernandez, Ph.D., PHPD, LLC Consulting, LLC
Irwin Hirsh, Novo Nordisk A/S Wendy Zwolenski-Lambert, Abbott Laboratories

This technical report was developed as part of PDA’s Paradigm Change in Manufacturing Operations (PCMO)
project. The content and views expressed in this Technical Report are the result of a consensus achieved by the
members of the authorizing Task Force, and are not necessarily the views of the organizations they represent.
 Process Validation: A
Lifecycle Approach
Technical Report No. 60

ISBN: 978-0-939459-51-3
© 2013 Parenteral Drug Association, Inc.
All rights reserved.

Bethesda Towers
4350 East West Highway
Suite 200
Bethesda, MD 20814 USA
Tel: 1 (301) 656-5900
Fax: 1 (301) 986-0296
E-mail: [email protected]
Web site: www.pda.org
Paradigm Change in Manufacturing Operations (PCMOSM)
PDA launched the project activities related to the PCMOSM program in December 2008 to help imple-
ment the scientific application of the ICH Q8, Q9 and Q10 series. The PDA Board of Directors ap-
proved this program in cooperation with the Regulatory Affairs and Quality Advisory Board, and the
Biotechnology Advisory Board and Science Advisory Board of PDA.

Although there are a number of acceptable pathways to address this concept, the PCMO program fol-
lows and covers the drug product lifecycle, employing the strategic theme of process robustness with-
in the framework of the manufacturing operations. This project focuses on Pharmaceutical Quality
Systems as an enabler of Quality Risk Management and Knowledge Management.

Using the Parenteral Drug Association’s (PDA) membership expertise, the goal of the Paradigm
Change in Manufacturing Operations Project is to drive the establishment of ‘best practice’ docu-
ments and /or training events in order to assist pharmaceutical manufacturers of Investigational
Medicinal Products (IMPs) and commercial products in implementing the ICH guidelines on Phar-
maceutical Development (ICH Q8, Q11), Quality Risk Management (ICH Q9) and Pharmaceutical
Quality Systems (ICH Q10).

The PCMO program facilitates communication among the experts from industry, university and regula-
tors as well as experts from the respective ICH Expert Working Groups and Implementation Working
Group. PCMO task force members also contribute to PDA conferences and workshops on the subject.

PCMO follows the product lifecycle concept and has the following strategic intent:
• Enable an innovative environment for continual improvement of products and systems
• Integrate science and technology into manufacturing practice
• Enhance manufacturing process robustness, risk based decision making and knowledge management
• Foster communication among industry and regulatory authorities

The Product Lifecycle

Pharmaceutical Technology Commercial Product

Development Transfer Manufacturing Discontinuation

For more information, including the PCMOSM Dossier, and to get involved, go to
www.pda.org/pcmo
 Table of Contents

1.0 INTRODUCTION...................................................1 4.3.2.4 PPQ Using Bracketing, Matrix,

1.1 Purpose and Scope......................................... 1 and Family Approaches ..................... 36
1.2 Background.................................................... 1 4.3.2.5 Bracketing Approach.......................... 36
4.3.2.6 Matrix Approach ............................... 36
2.0 GLOSSARY OF TERMS.........................................6 4.3.2.7 Family (Grouping) Approach............... 37
4.3.2.8 Process Analytical Technology .......... 38
2.1 Acronyms....................................................... 9
4.3.2.9 Sampling Strategy.............................. 39
3.0 BUILDING AND CAPTURING PROCESS 4.3.2.10 Setting PPQ Acceptance Criteria..... 39
KNOWLEDGE (STAGE 1 — PROCESS DESIGN).10 4.4 PPQ Protocol ................................................ 40
4.5 PPQ Report................................................... 42
3.1 Deliverables from Stage 1
4.6 Transition to Continued Process Verification.43
Process Validation........................................ 12
3.2 Quality Target Product Profile (QTPP)............ 12
5.0 CONTINUED PROCESS VERIFICATION
3.3 Critical Quality Attributes.............................. 13 (STAGE 3) ...........................................................44
3.4 Define the Manufacturing Process ............... 14
3.5 Analytical Methods....................................... 20 5.1 Establishing a Monitoring Program............... 44
3.6 Risk Assessment and Parameter Criticality 5.1.1 Purpose and Strategy.............................. 44
Designation................................................... 20 5.1.2 Documenting the CPV Program............... 44
3.7 Process Characterization.............................. 23 5.1.3 Legacy Products and Continued Process
3.8 Product Characterization Testing Plan........... 23 Verification............................................... 46
3.9 Control Strategy............................................ 24 5.1.4 Demonstrating Continued Process
3.10 Clinical Manufacturing Experience – Batch Verification............................................... 47
Records and Production Data....................... 25 5.1.5 CPV Monitoring Plan................................ 48
3.11 Process Design Report.................................. 26 5.1.6 Data Analysis and Trending..................... 48
3.12 Process Validation Master Plan.................... 26 5.2 Incorporation of Feedback from
3.13 Stage 1 Manufacturing and Technology CPV Monitoring............................................. 49
Considerations.............................................. 26 5.2.1 Quality Systems and CPV........................ 49
5.3 CPV Data Review and Reporting................... 50
4.0 PROCESS QUALIFICATION (STAGE 2)................28
6.0 PROCESS VALIDATION ENABLING SYSTEMS
4.1 Strategies for System Design and
AND TECHNOLOGY.............................................51
Qualification.................................................. 28
4.1.1 Engineering and Design........................... 29 6.1 Application of Risk Management ................... 51
4.1.1.1 Risk Assessment................................ 29 6.1.1 Risk Management in Stage 1 –
4.1.2 Installation............................................... 29 Process Design........................................ 52
4.1.3 Qualification Plan..................................... 29 6.1.2 Risk Management in Stage 2 –
4.1.3.1 Test Functions and Process Qualification............................... 53
Acceptance Criteria............................ 30 6.1.3 Risk Management in Stage 3 –
4.1.4 Maintaining Systems in a Continued Process Verification................ 54
State of Control ...................................... 30 6.1.4 Raw Material Risk Management
4.2 Process Performance Qualification .............. 31 Considerations ........................................ 54
4.2.1 PPQ Readiness........................................ 31 6.2 Statistical Analysis Tools ............................. 55
4.3 Design Strategy for Process Performance 6.2.1 Design of Experiments (DoE)................... 57
Qualification (PPQ)........................................ 33 6.2.2 Statistical Process Control and Process
4.3.1 Use of Prior Knowledge and Stage 1 Capability................................................. 59
Data to Support PPQ................................ 33 6.2.2.1 Statistical Process Control Charts...... 60
4.3.2 PPQ Study Design ................................... 34 6.2.2.1.1 Factors to Consider in Designing a
4.3.2.1 Number of Batches............................. 35 Control Chart................................... 62
4.3.2.2 PPQ at Normal Operating Conditions.. 35 6.2.2.1.2 Types of Control Charts................... 62
4.3.2.3 PPQ Using Individual Unit Operation 6.2.2.1.3 Process Capability........................... 62
Studies............................................... 36 6.2.3 Statistical Acceptance Sampling.............. 64
 6.2.4 Number of Lots for Stage 2 Process p×100% lot failure rate.......................... 81
Performance Qualification (PPQ).............. 66 8.1.3 Within and Between Lot Normal Tolerance
6.3 Process Analytical Technology (PAT)............ 66 Intervals................................................. 82
6.3.1 Selection of PAT System.......................... 67 8.1.4 Statistical Process Control Charts ......... 82
6.3.2 Process Validation Considerations 8.1.5 Ppk , Cpk Process Capability Metrics........ 83
During the PAT System Design Stage...... 69 8.1.6 Assure the Lot Conformance
6.3.2.1 Risk Assessment................................ 69 Rate is Above an Acceptable Rate
6.3.2.2 In-Process Application and With Specified Confidence..................... 84
Method Development......................... 69 8.1.7 Wald Sequential Probability Ratio.......... 84
6.3.3 Process Qualification 8.1.8 Narrow Limit Gauging............................ 85
Considerations for PAT............................. 69 8.1.9 Demonstrate Between-Lot Variation is
6.3.4 Continued Process Verification Less Than Within-Lot Variation (Anova) 8. 5
Considerations for PAT............................. 70 8.1.10 Sample Size .......................................... 86
6.4 Technology Transfer ..................................... 70 8.1.11 Demonstrate the Between-Lot Standard
6.5 Knowledge Management.............................. 73 Deviation σb ≤ Acceptable Value X........ 86
8.1.12 Demonstrating equivalence
7.0 EXAMPLES.........................................................75 between lots.......................................... 86
7.1 Large Molecule (Biotech).............................. 75 8.2 Appendix 2: Types of Control Charts............. 87
7.2 Small Molecule (Parenteral).......................... 77 8.2.1 Control Charts for Variables Data............. 87
8.2.2 Control Charts for Attributes Data........... 88
8.0 APPENDICES.......................................................81 8.2.3 Performance of Control Charts:
8.1 Appendix 1: Statistical Methods for Average Run Length (ARL)...................... 88
Determining the Number of Lots................... 81
8.1.1 Average Run Length (ARL) to detect a 9.0 REFERENCES.......................................................89
FIGURES AND TABLES INDEX

Figure 1.2-1 Applicability of ICH Q8 (R2) through Q11 Figure 6.2.2-1 Process in Classical Statistical Control;
Relative to the FDA Stage Approach to Common Cause Variation only............ 59
Process Validation................................ 3 Figure 6.2.2-2 Process Not in Statistical
Figure 1.2-2 Common Timing of Process Validation Control -Special Cause Variation......... 60
Enablers and Deliverables to Validation Figure 6.2.2-3 A Process with Both Within-lot and
Stage Activities.................................... 5 Between-lot Variation......................... 60
Figure 3.0-1 Overall Sequence of Process Figure 6.2.2.1-1 Xbar/S Control Chart for Fill Weight,
Validation Activities ........................... 11 n=5 per group.............................. 61
Figure 3.4-1 Example Process Diagram for a Figure 6.2.2.1.3-1 Process Capability Statistics Cp
Tangential Flow Filtration Step........... 15 and Cpk........................................... 63
Table 3.4-1 Example Process Parameter Table Figure 6.2.2.1.3-2 Examples of Process Capability
for a Tangential Flow Filtration Step... 16 Statistics Cp and Cpk...................... 63
Figure 3.6-1 Decision Tree for Designating Table 6.2.2.1.3-2 Relationship Between Capability and
Parameter Criticality........................... 22 % or Per Million Nonconforming.... 64
Figure 4.1-1 Typical System Qualification Sequence... 28 Figure 6.2.3-1 Example of an Operating Characteristic
Table 4.1.3-1 Qualification Information..................... 30 Curve.................................................. 65
Figure 4.3.1-1 Relationship of Prior Knowledge to the Table 6.3.3-1 Examples of PAT Tools and Their
Amount of PPQ Data Required............ 33 Application......................................... 68
Table 4.3.2.6-1 Illustration of a Matrix Approach for Table 6.4-1 Technology Transfer Activities
Filling Process PPQ............................. 37 Throughout Product Lifecycle............. 71
Table 4.4-1 Example of PPQ Acceptance Figure 6.4-1 Distribution of Technology Transfer
Criteria Table...................................... 42 Activities throughout the Product
Lifecycle............................................. 73
Figure 5.1.2-1 Development of a Continued Process
Verification Plan.................................. 45 Table 7.1-1 Stage 1: Process Design..................... 75
Figure 5.1.2-2 CPV Plan within Validation Table 7.1-2 Stage 2: Process Qualification
Documentation System...................... 45 (Continued)......................................... 76
Table 7.1-3 Stage 3: Continued Process Verification.......77
Figure 5.1.3-1 CPV Plan Determination for
Legacy Products................................. 47 Table 7.2-1 Stage 1: Process Design..................... 77
Figure 5.2.1-1 Body of Knowledge and Maintenance Table 7.2-2 Stage 2: Process Qualification ........... 79
of Process Control.............................. 49 Table 7.2-3 Stage 3: Continued Process Verification.... 80
Figure 6.1-1 Quality Risk Management: A Lifecycle Table 8.1.2-1 Expected Between-Lot Variation
Tool for Process Development and Coverage in nL Lots............................. 81
Validation............................................ 52
Table 8.1.6-1 Number of lots to demonstrate
Figure 6.1-2 Product Attribute Criticality Risk confidence for lot conformance rate... 84
Assessment Example......................... 53
Figure 8.1.7-1 Wald’s Sequential Probability Ratio
Table 6.1-1 Risk-Based Qualification Planning....... 53 Example.............................................. 85
Table 6.1.2 Severity Rating and Sampling Table 8.1.9-1 Effect of between-lot variation on the
Requirements..................................... 54 total process variance........................ 85
Table 6.2-1 Statistical Methods and the Typical Table 8.1.10-1 Sample Size to estimate a standard
Stages at Which They Are Used........ 56 deviation to within ±X% of true value..... 86
 1.0 Introduction

1.1 Purpose and Scope

This Technical Report (TR) is intended to provide practical guidance on the implementation of a
lifecycle approach to pharmaceutical process validation (PV). It contains information that enables
manufacturers to implement globally-compliant PV programs consistent with the principles of re-
cent lifecycle-based PV guidance documents and current expectations for Pharmaceutical Quality
Systems (1-4). In pharmaceutical manufacturing, “process validation” is the collection and evaluation
of data -from the process design stage through commercial production that establishes scientific evi-
dence that a process is capable of consistently delivering quality product (3). The U.S. FDA and EMA
consider PV a requirement in both general and specific terms in current Good Manufacturing Prac-
tice (cGMP) guidelines and an essential element in the assurance of drug quality (2,3,5).

The PV lifecycle concept links product and process development, the qualification of the commercial
manufacturing processes, and maintenance of the commercial production process in a coordinated
effort (3). When based on sound process understanding and used with quality risk management prin-
ciples, the lifecycle approach allows manufacturers to use continuous process verification (enhanced
approach) in addition to, or instead of, traditional PV (1,2,6).

The information in this TR applies to the manufacturing processes for drug substances and drug
products, including:
• Pharmaceuticals, sterile and non-sterile
• Biotechnological/biological products, including vaccines
• Active Pharmaceutical Ingredients (APIs)
• Radiopharmaceuticals
• Veterinary drugs
• Drug constituents of combination products (e.g., a combination drug and medical device)

This report is prepared for global use and applies to new and existing (i.e., legacy) commercial manu-
facturing processes. Its scope does not include manufacturing processes for:
• Medical devices
• Dietary supplements
• Medicated feed
• Human tissues

Although these product categories are outside the scope of this TR, its recommendations are based
on modern quality concepts, ICH Quality Guidelines, and recent regulatory authority guidance docu-
ments. As such, it may be a useful reference in the development of PV lifecycle approaches for other
product categories. The validation of ancillary supporting operations used in pharmaceutical manu-
facturing processes is not discussed in the report. Many PDA TRs already provide specific guidance for
such procedures; for example, cleaning, aseptic process simulation, moist heat sterilization and dry
heat sterilization (7-10).

1.2 Background
The lifecycle concept includes all phases in the life of a product from initial development through
commercial production and product discontinuation (4,11). The use of a lifecycle approach to phar-
maceutical product quality is widely thought to facilitate innovation and continual improvement as
well as strengthen the link between pharmaceutical development and manufacturing (ICH Q10). The
lifecycle philosophy is fundamental in the ICH guidance documents for Pharmaceutical Develop-

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 1

 ment (ICH Q8 (R2)), Quality Risk Management (ICH Q9) (12), Pharmaceutical Quality Systems (ICH
Q10), and Development and Manufacture of Drug Substances (ICH Q11). The principles they contain
provide the product lifecycle framework and quality system enablers that have been used in recent
pharmaceutical process validation guidance documents. A central concept in these documents is that
PV is not a one-time event, but rather, an activity that spans the product lifecycle, linking process de-
velopment, validation of the commercial manufacturing process, and its maintenance during routine
commercial production.

The ICH Q8 (R2) guidance document for pharmaceutical development defines procedures for link-
ing product and process development planning to the final commercial process control strategy and
quality system. It describes an enhanced scientific and risk-based approach to product and process
development that emphasizes statistical analysis, formal experimental design, and the incorporation
of knowledge gained from similar products and processes. Manufacturing capabilities and the quality
system must be integrated into the process development plan to ensure effective and compliant com-
mercial operations. The functionality and limitations of commercial manufacturing equipment are a
primary consideration in the process design.

The ICH Quality Risk Management guidance document (ICH Q9) describes the use of a risk-based
approach to pharmaceutical development and manufacturing quality. These approaches identify and
prioritize those process parameters and product quality attributes with the greatest potential to af-
fect product quality. Specific guidance on the application of the ICH Q9 concepts can be found in
PDA Technical Report 54: Implementation of Quality Risk Management for Pharmaceutical and Biotechnology
Manufacturing Operations and PDA Technical Report 59: Utilization of Statistical Methods for Production
and Business Processes (13,14). The FDA process validation guidance document stresses a risk-based
approach to develop criteria and process performance indicators, and improve the design and execu-
tion of other validation-related activities, such as developing confidence levels and sampling plans (3).

Both the FDA and EMA process validation guidance documents aim to integrate PV activities into the
pharmaceutical quality system. To achieve the goals outlined in ICH Q10, it is essential to integrate
the process design stage into the quality system. Throughout the development effort, product and
process development input and alignment from the Quality Unit are required to ensure compatibility
with the quality system. Key considerations in product and process design include the commercial
control strategy and use of modern quality risk management procedures. Quality and Regulatory
organizational components should be part of the cross-functional product team from the beginning
of the process validation study design. Their participation is essential to ensure that the study design
is compatible with the firm’s quality system, and that submissions will meet regulatory agency ex-
pectations.

The Quality Unit should provide appropriate oversight and approval of process validation studies re-
quired under GMPs. Although not all process validation activities are performed under GMPs (for
example, some Stage 1 – Process Design studies) (4), it is wise to include the Quality and Regulatory
representatives on the cross-functional team. The degree and type of documentation required varies
during the validation lifecycle, but documentation is an important element of all stages of process
validation. Documentation requirements are greatest during the process qualification and verification
stages. Studies during these stages should conform to GMPs and be approved by the Quality Unit.

The Process Validation Master Plan (PVMP) should describe the rationale, overall validation strategy,
and list of specific studies. It should reside within the firm’s quality documentation system (15). A suc-
cessful validation program is one that is initiated early in the product lifecycle and is not completed
until the process or product reaches the end of that lifecycle. A comprehensive corporate policy that

2 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

defines the expectations and commitment to process validation lifecycle principles is the foundation
of a successful validation program. This policy should define the quality management philosophy,
components of validation, periodic review or requalification time frames, documentation require-
ments (including a process validation master plan), validation protocols and reports, and responsibili-
ties of key stakeholders within the organization (16).

This TR follows the principles and general recommendations presented in current regulatory process
validation guidance documents. Of particular note, is that the TR uses the traditional/nontraditional
(enhanced) process validation terminology employed by EMA (1). In this context, nontraditional or
enhanced process validation may use Continuous Process Verification as an alternative approach to
traditional PV. In the enhanced approach, manufacturing process performance is continuously moni-
tored and evaluated. It is a science and risk-based real-time approach to verify and demonstrate that
a process operates within specified parameters and consistently produces material that meets quality
and process performance requirements.

The FDA three-stage process validation lifecycle nomenclature (Stage 1-Process Design, Stage 2– Pro-
cess Qualification, and Stage 3–Continued Process Verification) is used in this TR. Implementation of
these stages is discussed in detail in Sections 3–5. It should be noted that Continued Process Verifica-
tion and Continuous Process Verification are distinct terms and have different meanings. Continuous
Process Verification refers to validating manufacturing processes that utilize advanced manufacturing
and analytical technologies (e.g., PAT systems). FDA uses the term Continued Process Verification
generally to mean those activities which maintain the process in a state of control and encompasses
all manufacturing scenarios, i.e., traditional manufacturing, manufacturing employing advanced tech-
nologies of any kind or any combination thereof.

These are defined in Section 2.0 and are also discussed later in this TR. Figure 1.2-1 shows the re-
lationship between the relevant ICH guidance documents and the FDA stage approach to process
validation across the product lifecycle.
Figure 1.2-1 Applicability of ICH Q8 (R2) through Q11 Relative to the FDA Stage Approach to Process Validation

Lifecycle
Development – ICH Q8(R2)
Need to make this read ICH Q8(R2)
Quality Systems / Knowledge Management – ICH Q10
Development and Manufacture of Drug Substances – ICH Q11

Stage 1 Stage 2 Stage 3

Processing Design/ Systems Qualification/ Continued Process
Characterization Commercial Process Verification
Performance Qualification (PPQ)

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 3

 This TR is based on the experiences and knowledge of the Task Force on Process Validation: A Lifecy-
cle Approach. It represents a cross-section of industry professionals covering products within its scope
(e.g., large and small molecules, and sterile and non-sterile products), and presents approaches to best
practices that are scientifically sound, good business practices, and designed to meet current regulatory
expectations. The document does not include isolated responses to individual inspection or review is-
sues. These are most often case-by-case requirements particular to specific organizational needs.

The intent of this TR is not to establish mandatory standards, but rather to be a single-source over-
view that complements existing regulatory authority guidance documents. References throughout
the document provide greater detail on various topics. It is always advisable to consult with the ap-
propriate regulatory authorities for agreement on the strategies employed for product development
and lifecycle management strategies.

Figure 1.2-2 illustrates the progression of typical process validation enablers or deliverables relative
to validation activities that are conducted throughout the product lifecycle. The figure represents
stages and validation studies as single “point in time” events. However, in practice, the exact timing
of product development activities or validation studies may vary with the specific product develop-
ment strategy. For example, the enablers for Stage 1 process validation activities will be much less
extensive for a production formulation change than for development of a new molecular entity. Thus,
the figure presents an overall sequence of activities and their approximate correlation to the stages of
process validation.

4 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Figure 1.2-2 Common Timing of Process Validation Enablers and Deliverables to Validation Stage Activities

Process Validation Stages

Process Validation Enablers and Deliverables Product Lifecycle Validation Stage Activities
Quality Target Product Profile (initial) Early Drug Substance Process Development
Quality Attributes Evaluation (initial)
Initial Formulation and DP Process Development
Clinical Process Description
Clinical Production Master Batch Records Development of PAT and/ or Analytical
methodologies
Reference Standard(s)
IND APPLICATION Risk Assessment for Robustness Studies

Stage 1
Quality Attributes Evaluation (updated)
Initiate Formal Stability Studies
Quality Target Product Profile (updated)
Clinical Manufacturing
Process Validation Master Plan
Process Parameter – Categorization Qualify Manufacturing Equipment & Facility
Process Parameter – Acceptable Ranges
Continued Assay and Process Development

Process Equipment Qualification Protocols and Reports

Develop and Qualify Scaled — down Models

Process Characterization Studies — Clearance,

Commercial Process Description Robustness, and Other Qualification Studies
Stage 2

Defined Process Control Strategy (Design Space Established, if applicable)

Risk Assessment for Commercial Manufacturing

Commercial Production Master Batch Records Assay Qualification / Validation
Process Performance Qualification
Protocols and Reports Assess Risk (Process + Equipment + Operation)
Process Performance Qualification Technical Implementation of Process Control Strategies
Summaries for Filing / Inspection
REGULATORY LICENSE APPLICATION
Manufacture PPQ Batches
Stage 3

Review and Update:

Risk Assessments PRE-APPROVAL INSPECTION
Process Monitoring Review / Periodic Report REGULATORY APPROVAL
Implement Continued Verification Program
Commercial Manufacturing & Distribution
Completion of Lifetime Validation Studies Lifecycle Management within Quality System

Tools used throughout the lifecycle (e.g., risk management, statistical analysis, Process Analytical Tech-
nology [PAT], technology transfer, documentation, and knowledge management) are described in Sec-
tion 6.0. Examples of the lifecycle approach for a large and small molecule are described in Section 7.0.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 5

 2.0 Glossary of Terms

Terminology usage may differ by company, at individual companies and some terms may be subject
to change over time. Those terms used in a validation program should be clearly defined, docu-
mented, and well-understood. Terminology definitions that are widely recognized by the industry
should be considered when establishing internal definitions. These can be found in regulatory guid-
ance documents. Definitions of company-specific terminology should also be included in the valida-
tion documents to provide clarity and context. This Technical Report uses the terms below, which are
accompanied by their definitions, synonyms, and references where applicable:

Active Pharmaceutical Ingredient (API; Quality Attribute

Equivalent to Drug Substance for large A molecular or product characteristic that is
molecules) selected for its ability to indicate the quality of
Any substance or mixture of substances intended the product. Collectively, the quality attributes
to be used in the manufacture of a drug (medici- define identity, purity, potency and stability of
nal) product and that, when used in the produc- the product, and safety with respect to adven-
tion of drug, becomes an active ingredient of the titious agents. Specifications measure a select-
drug product. Such substances are intended to ed subset of the quality attributes (18).
furnish pharmacological activity or other direct Attribute
effect in the diagnosis, cure, mitigation, treat- A physical, chemical, or microbiological prop-
ment, or prevention of disease or to affect the erty or characteristic of an input or output ma-
structure and function of the body (17). terial (19).
Active Pharmaceutical Ingredient (API) Continued Process Verification (CPV)
Starting Material Assuring that during routine production the pro-
A raw material, intermediate, or an API that is cess remains in a state of control (3).
used in the production of an API and that is in-
corporated as a significant structural fragment Continuous Process Verification
into the structure of the API. An API Starting An alternative approach to process validation
Material can be an article of commerce, a mate- in which manufacturing process performance is
continuously monitored and evaluated (11).
rial purchased from one or more suppliers under
contract or commercial agreement, or produced Continuum of Criticality
in-house. API Starting Materials normally have As Used for Parameters
defined chemical properties and structures (17). A non-discrete scale where parameters or attri-
butes are evaluated relative to their impact on
Attributes drug substance and drug product quality (3).
Critical Quality Attribute (CQA)
A physical, chemical, biological or microbio- As Used for Attributes
logical property or characteristic that should be Following comprehensive assessments of sci-
within an appropriate limit, range, or distribu- entific evidence and risk, quality attributes
tion to ensure the desired product quality (11). are ranked according to the degree of criti-
cality. The continuum, as opposed to binary
Process Performance Attribute (Synonym classifications of Critical and Non-Critical, is
– Process Performance Parameter) thought to “more accurately reflect complex-
An output variable or outcome that cannot be ity of structure-function relationships and the
directly controlled, but is an indicator that the reality that there is some uncertainty around
process performed as expected (15). attribute classification” (20).

6 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Control Strategy Intermediate (or In-Process Material)
A planned set of controls, derived from current A material produced during the steps of the pro-
product and process understanding, which en- cessing of an API that undergo further molecu-
sures process performance and product quality. lar change or purification before it becomes an
The controls can include parameters and attri- API. Intermediates may or may not be isolated
butes related to drug substance and drug product (17).
materials and components, facility and equipment
operating conditions, in-process controls, finished Lifecycle
product specifications, and the associated meth- All phases in the life of a product, from the initial
ods and frequency of monitoring and control (4). development through marketing until the prod-
uct’s discontinuation (11).
Design Space
The multidimensional combination and interac- Normal Operating Range (NOR)
tion of input variables (e.g., material attributes) A defined range, within (or equal to) the Proven
and process parameters that have been demon- Acceptable Range, specified in the manufactur-
strated to provide assurance of quality. Working ing instructions as the target and range at which
within the design space is not considered a change. a process parameter is controlled, while pro-
Movement out of the design space is considered ducing unit operation material or final product
to be a change, and would normally initiate a meeting release criteria and CQAs (23).
regulatory post-approval change process. Design
space is proposed by the applicant and is subject Parameters
to regulatory assessment and approval (11). Critical Process Parameter (CPP; Synonym
– Critical Operational Parameter)
Drug Product (DP) A process parameter whose variability has an
The dosage form in the final immediate packag- impact on a critical quality attribute and there-
ing intended for marketing (17). fore should be monitored or controlled to en-
sure the process produces the desired quality
Drug Substance (DS; Equivalent to Active
Pharmaceutical Ingredient for small (11).
molecules) Key Process Parameter (KPP; Synonym –
The material which is subsequently formulated Key Operational Parameter)
with excipients to produce the drug product. It An input process parameter that should be
can be composed of the desired product, prod- carefully controlled within a narrow range
uct-related substances, and product- and process- and is essential for process performance. A
related impurities. It may also contain excipients
key process parameter does not affect product
including other components such as buffers (21).
quality attributes. If the acceptable range is
Formal Experimental Design (Synonym – exceeded, it may affect the process (e.g. yield,
Design of Experiments) duration) but not product quality (15).
A structured, organized method for determin-
Non-Key Process Parameter (Non-
ing the relationship between factors affecting a
process and the output of that process (11).
KPP; Synonym – Non-key Operational
Parameter)
Good Engineering Practice (GEP) An input parameter that has been demonstrat-
Those established engineering methods and ed to be easily controlled or has a wide accept-
standards that are applied throughout the life- able limit. Non-key operational parameters
cycle to deliver appropriate and cost-effective may have an impact on quality or process per-
solutions (22). formance if acceptable limits are exceeded (15).

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 7

 Process Parameter (Synonym – Process Validation
Operational Parameter) US FDA
An input variable or condition of the manufac- The collection and evaluation of data from
turing process that can be directly controlled the process design stage to commercial pro-
in the process. Typically, these parameters are duction,, which establishes scientific evidence
physical or chemical (e.g. temperature, process that a process is capable of consistently deliv-
time, column flow rate, column wash volume, ering quality products (3)..
reagent concentration, or buffer pH) (15).
EMA
Platform Manufacturing The documented evidence that the process,
Development of a production strategy for a new operated within established parameters, can
drug starting from manufacturing processes simi- perform effectively and reproducibly to pro-
lar to those used to manufacture other drugs of duce a medicinal product meeting its predeter-
the same type (the production for which there mined specifications and quality attributes (2).
already exists considerable experience) (6).
Process Validation Master Plan (Synonym
Process Analytical Technology (PAT) – Validation Master Plan)
A system for designing, analyzing, and control- A document that defines the process validation
ling manufacturing through timely measure- scope and rationale and that contains the list of
ments (i.e., during processing) of critical quality process validation studies to be performed (15).
and performance attributes of raw and in-pro-
cess materials and processes with the goal of en- Proven Acceptable Range (PAR)
suring final product quality (11). A characterized range of a process parameter
for which operation within this range, while
Process Performance Qualification (PPQ) keeping other parameters constant, will result
The second element of the Process Qualifica- in producing a material meeting relevant quality
tion. It includes a combination of the actual criteria (11).
facility, utilities, equipment, and the trained
personnel with the commercial manufacturing Quality
process, control procedures, and components to The suitability of either a drug substance or
produce commercial batches. A successful PPQ drug product for its intended use. This term in-
will confirm the process design and demonstrate cludes such attributes as the identity, strength
that the commercial manufacturing process per- and purity (24).
forms as expected. Batches prepared are also Quality by Design (QbD)
called Conformance batches or PPQ batches (3). A systematic approach to development that be-
Process Qualification gins with predefined objectives and emphasizes
Confirming that the manufacturing process, as product and process understanding and process
designed, is capable of reproducible commercial control, based on sound science and quality risk
manufacturing (3). management (11).

Process Robustness Quality Target Product Profile (QTPP)

Ability of a process to tolerate variability of A prospective summary of the quality charac-
materials and changes of the process and equip- teristics of a drug product that ideally will be
ment without negative impact on quality (11). achieved to ensure the desired quality, taking
into account safety and efficacy of the drug
product (11).

8 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Target Product Profile (TPP) term that encompasses types of approaches to
A format for a summary of a drug development ensure the systems are fit for such as qualifica-
program described in terms of labeling concepts tion, commissioning and qualification, verifica-
to facilitate communication regarding a particu- tion, system validation, or other (26).
lar drug development program (25).
Worst Case
Validation Master Plan A set of conditions encompassing upper and
See Process Validation Master Plan. lower processing limits and circumstances, in-
Verification cluding those within standard operating proce-
A systematic approach to verify that manufactur- dures, that pose the greatest chance of process
ing systems, acting alone or in combination, are or product failure (when compared to ideal con-
fit for intended use, have been properly installed, ditions). Such conditions do not necessarily in-
and are operating correctly. This is an umbrella duce product or process failure (8).

2.1 Acronyms
API — Active Pharmaceutical Ingredient KPP — Key Process Parameter
CMA — Critical Material Attribute NOR — Normal Operating Range
CPP — Critical Process Parameter PAR — Proven Acceptable Range
CPV — Continued Process Verification PAT — Process Analytical Technology
CQA — Critical Quality Attribute PPQ — Process Performance Qualification
DoE — Design of Experiments PTT — Product Technical Team
DP — Drug Product PVMP — Process Validation Master Plan
DS — Drug Substance QbD — Quality by Design
FMEA — Failure Mode Effects Analysis QTPP — Quality Target Product Profile
HCP — Host Cell Protein TPP — Target Product Profile
ICH — International Conference Harmonization TT — Technology Transfer

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 9

 3.0 Building and Capturing Process Knowledge
(Stage 1 — Process Design)
This section focuses on approaches used during development to implement robust manufacturing
processes. It addresses the first Stage of process validation in which process and product knowledge
are explored to establish the control strategy. Risk assessment and management are used to focus the
development effort. Process and product knowledge evolve through the course of the pharmaceutical
development program. Designing a comprehensive and efficient program for a lifecycle approach to
process validation compels thoughtful planning very early in development. Early planning facilitates
appropriate data gathering in Stage 1, with the objective of enhancing the effectiveness and success
of Stage 2 Commercial Process Qualification. It also establishes a foundation for continued process
verification in Stage 3.

Sources of knowledge available prior to (and that may be used during) Stage 1 of the Process Valida-
tion lifecycle, include:
• Previous experience with similar processes (e.g., platform processes)
• Product and process understanding (from clinical and pre-clinical activities)
• Analytical characterization
• Published literature
• Engineering studies/batches
• Clinical manufacturing
• Process development and characterization studies

The following sections outline the Stage 1 outputs from a general lifecycle approach to Process Valida-
tion, as depicted in Figure 3.0-1.

10 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Figure 3.0-1 Overall Sequence of Process Validation Activities

Establish Target Product Profile and Quality

Target Product Profile – QTPP
(Living Document)
Process
Development
Identify Critical Quality Attributes (CQAs)

Define the Manufacturing Process

Stage 1

Process Perform Quality Risk Assessment Initial

Characterization Categorization of Parameters
Scale Up, Tech Transfer,
and Set Design Space
Boundaries (Optional)

Perform Proces Characterization Experiments

(eg. DOE, multivariate, univariate)

Final Categorization of Parameters Based on

Criticality and Establish Control Strategy

Implement Process Control Strategy

Commercial
Stage 2 Process Facilities, Utilities, and Equipment Qualification
Qualification

Process Performance Qualification

Stage 3 Continued Process Verification

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 11

 3.1 Deliverables from Stage 1 Process Validation
The list below summarizes the information needed to transition from Stage 1 (Process Design) to
Stage 2 (Performance Qualification) in the Process Validation Lifecycle. The sections in this section
discuss these deliverables in more detail and provide references for additional information.
• Quality Target Product Profile (QTPP) — This is done at the initiation of Stage 1
• Critical Quality Attributes (CQAs) with corresponding Criticality Risk Assessment and desired confidence
• Manufacturing Process Design
• Process description showing process inputs, outputs, yields, in-process tests and controls, and process
parameters (set points and ranges) for each unit operation
• Process solution formulas, raw materials, specifications
• Batch Records and production data from laboratory or pilot scale production

• Analytical Methods (for product, intermediates, and raw materials)

• Quality Risk Assessment
• Initial risk-based categorization of parameters prior to process characterization

• Criticality and Risk Assessments

• Identification of Process Parameters with corresponding criticality and risk analysis

• Process Characterization
• Process Characterization Plan and Protocols
• Study Data Reports

• Process Control Strategy

• Release Specifications
• In-Process Controls and Limits
• Process Parameter set points and ranges
• Routine Monitoring requirements (including in-process sampling and testing)
• Storage and time limitations for intermediates, process solutions, and process steps
• Raw Material/Component Specifications
• Design Space (if applicable)

• Process Analytical Technology applications and algorithms (if PAT is used)

• Product Characterization Testing Plan (i.e., tests not included in the product Release Test panel)
• Manufacturing Technology – assessment of production equipment capability and compatibility with process
requirements (may be covered in Stage 2a)
• Scale-up/Scale-down Approach (Evaluation/Qualification of Laboratory Models)
• Development Documentation
• Process Design Report

• Process Validation Master Plan

3.2 Quality Target Product Profile (QTPP)

The aim of pharmaceutical development is to design a quality product with a manufacturing process
that consistently delivers the intended performance of the drug product. Pharmaceutical develop-
ment begins with the establishment of pre-defined objectives. These are described in the Quality Tar-

12 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

get Product Profile (QTPP). The QTPP is defined at the initiation of Stage 1 and is referenced through-
out the product lifecycle.

The QTPP captures all relevant quality requirements for the drug product. Consequently, it is peri-
odically updated to incorporate any new data that may be generated during pharmaceutical develop-
ment. However, the QTPP should not depart from the core targets established in the drug product
Target Product Profile (TPP).

Note: TPP is used as a tool that facilitates sponsor-regulator interactions and communication. Con-
sequently, the TPP contains such information as Drug Indications and Use; Dosage and Administra-
tion; Dosage Forms and Strengths; Contraindications; Warnings and Precautions; Adverse Reac-
tions; Drug Interactions; Abuse and Dependence; and Overdose that are not covered under the
scope of this document (25).

It addresses relevant characteristics that include:

• Intended use in the clinical setting (e.g., dosage form and strength, route of administration, delivery systems,
container and closure system).
• Drug substance quality attributes appropriate to the drug product dosage form being developed (e.g., physical,
chemical, and biological properties).
• Drug product quality attributes appropriate for the intended marketed product (e.g., purity/impurities, stability,
sterility, physical, and chemical properties)
• Therapeutic moiety release or delivery, and attributes affecting pharmacokinetic characteristics (e.g., dissolution,
aerodynamic performance) appropriate to the drug product.
• Excipient and component quality attributes, drug-excipient compatibility, and drug-container compatibility that
affect the process ability, stability, or biological effect of the drug product.

The QTPP summarizes the quality attributes of the product that ensure safety and efficacy. It pro-
vides a starting point for assessing the criticality of product quality attributes.

3.3 Critical Quality Attributes

A Critical Quality Attribute (CQA) is a physical, chemical, biological, or microbiological property or
characteristic that should be within an appropriate limit, range, or distribution to ensure the desired
product quality. CQAs can be associated with drug substances, drug products, excipients, intermediates
(in-process materials), and container/closure components. At an early stage of process development,
the information available on product attributes may be limited. For this reason, the first set of CQAs
may come from prior knowledge obtained during early development and/or from similar products
rather than from extensive product characterization. The degree of criticality assigned to quality attri-
butes is derived using risk-based tools and the potential impact of the attributes on safety and efficacy.
Following comprehensive assessments of scientific evidence and risk, quality attributes are ranked ac-
cording to the degree of criticality, which may be a continuum that more accurately reflects the com-
plexity of structure-function relationships and varying levels of uncertainty around attribute classifica-
tion. Attributes not assigned as CQAs should also be considered in the development of the process.

CQAs are not synonymous with specifications. In addition, there is not necessarily a one-to-one rela-
tionship between CQAs and specifications. Specifications are a list of tests, references to analytical pro-
cedures, and appropriate acceptance criteria that are numerical limits, ranges, or other criteria for the
tests described. Several product attributes identified as CQAs may be detected by a single test method,
and therefore, built into a single test specification (e.g., API solubility, hardness, porosity are CQAs
evaluated using a single test: dissolution). Some CQAs may not be included in the specifications if they

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 13

 are very well-controlled and consistently achieved within the process (e.g., viral clearance is not tested
for every batch), while some attributes not considered critical may be included in the specifications.

The identification of potential CQAs is an ongoing activity initiated early in product development. It
makes use of general knowledge about the product and its application, as well as available clinical and
non-clinical data. CQAs are subject to change in the early stages of product development, and thus re-
quire a quality risk management approach that evolves as knowledge about the product and process is
generated (for discussion, see Section 6.1 “Application of Risk Management”). CQAs for commercial
products should be defined prior to initiation of Stage 2 activities.

3.4 Define the Manufacturing Process

A manufacturing process is designed to consistently provide a product that will meet its required qual-
ity attributes. As the process is being defined during development, a process description is a tool that
is used to assist in execution of risk assessments and in the development of the control strategy. The
manufacturing process is described as a series of constituent unit operations in a process description,
block diagram, or process flow diagram that describes each unit operation. Each unit operation in the
manufacturing process should be depicted with a similar level of detail. The following information
should be included in the description of each:
• Process requirements, including raw materials, scale, and order of operations
• Set points and ranges for the process parameters
• Identification and quantity of all material flows (additions, wastes, product streams)
• Testing, sampling, and in-process controls
• Hold times and hold conditions for product and additional solutions
• Estimated step yields and durations
• Sizing for equipment, including such items as chromatography columns and filtration units.
• Specific identification (manufacturer, part number) for manufacturing (e.g., filters) and product components
(e.g., vials, stoppers)
• Other information necessary to successfully reproduce the process

A process diagram for a single unit operation is presented as an example in Figure 3.3-1 and a sample
description table is provided in Table 3.4-1. The evolution of process knowledge and understanding
is reflected in clinical batch records; these are an important source of information for defining the
manufacturing process in the process description. Data collected from clinical trial material manu-
facture may be useful to determine process capabilities, set specifications, design PPQ protocols and
acceptance criteria, evaluate laboratory models, and transfer processes. Strategies and fundamentals
of knowledge management are discussed further in Section 6.5, Knowledge Management.

Process descriptions are documented in reports and may be incorporated into the Technology Trans-
fer (TT) Package for the product. The process may change during Stage 1 due to increases in material
demand (i.e., process and analytical development, clinical needs), improved product understanding
that leads to changes to CQAs, or improved process understanding that results in addition, elimina-
tion or adjustments of unit operations. Documentation should capture these changes and the sup-
porting justifications. This information should be archived in the Knowledge Management System.

Development and documentation of the commercial manufacturing process in Development Re-

ports should precede formal process characterization studies. Increased knowledge gained during
process characterization may require additional changes to the process description. All changes to
the process should be approved through change control procedures as defined by the Quality System.

14 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Figure 3.4-1 Example Process Diagram for a Tangential Flow Filtration Step

Regulator CIP in

AIR Diafiltration Buffer Conductivity

m S/cm pH Ultraviolet
Filter 0.2 Micron pH, cond., conc., diavolumes

Filter 0.2 Micron

Sample
Flow L/min
JACKET TANK
Drain
Sample Flow Control
Average Flux. LMH
Conductivity Product
m S/cm Conc., g/L PIT PIT
13 PSIG 14 PSIG

pH
UFDF
JACKET membrane
Agitation, Temperature, C FEED PERMEATE
rpm
Area, m2
Pump PSH
Flow 10 Size, kD

Recirculation rate, PIT PSIG PIT PSIG

DRAIN L/min 11 12

To product pool tank DRAIN

g/L, L, % yield

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 15

 Table 3.4-1 Example Process Parameter Table for a Tangential Flow Filtration Step
Process Variable Attributes
Process Variable Proven Rationale
Parameter Product or
Set Point Acceptable Expected Range
Designation Process Attribute
Range
General
Membrane Area 2 m2
Molecular Wt Cut-Off 30 kDa
Membrane Polymer Polysulfone
Pre-Use Cleaning & Flushing
Cleaning Solution: Concentration 0.4 to 0.6 N NaOH Non-Key Low risk of product or process impact
Adequate recirculation is needed to insure
Recirculation rate 10 L/min 8 to 12 L/min Non-Key proper cleaning, but acceptable results are
achieved over a wide range.
Low risk of product or process impact over
Transmembrane Pressure (TMP) 10 PSI 5 to 15 PSI Non-Key
a wide range.
May impact cleaning effectiveness if far out
of range. Procedural controls in place such
Temperature 30 °C 25 to 35 °C Non-Key
that the risk of running outside the range is
unlikely.
Wide range. Directly controlled to prevent run-
Time 60 min 60 to 90 min Non-Key
ning outside of the validated range.
WFI Flush volume 20 L/m2 ≥ 20 L/m2 Non-Key Wide range. Directly controlled.
Pre-Use Qualification
If test pressure is incorrect, the test result
Integrity Test Pressure 15 PSIG 15 – 18 PSIG Critical
is invalid.
Water Flux TMP 10 PSIG 8 – 12 PSIG Non-Key Water flux can be corrected for actual pressure.
Water flux can be corrected for actual tem-
Water Flux Temperature 20 °C 18 to 22 °C Non-Key
perature
Verification of filter integrity is crucial to en-
Manufacturer’s sure process effectiveness. Filter integrity
Filter Integrity Pass Critical
Specifications testing is an output of the prequalification,
but an input to processing the feed stream.

16 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Process Variable Attributes
Process Variable Proven Rationale
Parameter Product or
Set Point Acceptable Expected Range
Designation Process Attribute
Range
System Priming
All process buffer specifications categorized
as critical even though procedural controls
are in place to prevent release of non-con-
Buffer conductivity & pH Solution Acceptance Criteria Critical
forming buffers to production. Buffers outside
of the established ranges may impact product
quality during processing.
Unlikely to affect product or process. Directly
Buffer volume 35 L 25 to 50 L Non-Key
controllable.
Unlikely to affect product or process. Directly
Recirculation rate 8 L/min 4 to 12 L/min Non-Key
controllable.
Unlikely to affect product or process. Directly
Transmembrane Pressure (TMP) 12 PSI 10 to 15 PSI Non-Key
controllable.
Unlikely to affect product or process. Directly
Temperature 20 °C 15 to 25 °C Non-Key
controllable.
Process – Initial Concentration Step
Initial product concentrations and volumes
(total mass) may be critical due to relationship
Initial Product Total Mass 1225 g 900 to 1600 g Critical
with system volume constraints and ability to
reach DF and final concentration targets.
In some circumstances initial product con-
centrations and volumes (total mass) may
be critical – may be related to system vol-
Initial Product Volume 75 L 50 to 100 L Key ume constraints and ability to reach DF and
final concentration targets. In other situa-
tions, the initial volume may be Key (affect
process time and or yield).

17 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Process Variable Attributes
Process Variable Proven Rationale
Parameter Product or
Set Point Acceptable Expected Range
Designation Process Attribute
Range
Crossflow rate can impact flux; however,
only processing time is impacted unless
rate is excessively low (causing significant
Recirculation rate 8 L/min 6 to 10 L/min Key
membrane polarization of protein) Depends
on impact of TMP on flux. In many cases
this parameter can be categorized as “Key”.
Minor impact on flux unless operated ex-
cessively high or low (outside of PAR). At
Transmembrane Pressure (TMP) 12 PSI 10 to 15 PSI Key
low values TMP may have a significant im-
pact on flux.
Temperature 20 °C 15 to 25 °C Non-Key Minor impact on flux (approx 2% per degree)
Output of process conditions including TMP, Process
Process Flux (average) recirculation rate, product concentration. May Performance 20 to 30 LMH
be used to track batch-to-batch consistency Attribute
Output of the initial concentration stage / Critical Quality
Product Concentration 30 to 40 g/L
Input to diafiltration. Attribute
Volume must be in range validated for proper
volume control within the system during DF
Product (Retentate) Volume 35 L 30 to 40 L Critical and within equipment/tankage constraints for
total volume of DF buffer needed deliver re-
quired volumetric exchange during diafiltration.
Process – Diafiltration Step (constant volume)
Diafiltration buffer directly impacts the for-
Diafiltration Buffer pH and
Solution Acceptance Criteria Critical mulation of the final bulk drug substance
Conductivity
and ultimately drug product.
Crossflow rate can impact flux; however,
processing time is impacted if rate is exces-
Recirculation rate 8 L/min 6 to 10 L/min Key
sively low (outside of PAR) causing signifi-
cant membrane polarization of protein.
Minor impact on flux unless operated ex-
Transmembrane Pressure (TMP) 12 PSI 10 to 15 PSI Key cessively high or low. Operating outside of
PAR will impact process time.

18 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Process Variable Attributes
Process Variable Proven Rationale
Parameter Product or
Set Point Acceptable Expected Range
Designation Process Attribute
Range
System Volume During Potential to under-diafilter if variability or un-
35 L 30 to 40 L Critical
Diafiltration certainty in this parameter.
Extent of buffer exchange is dependent on
Number of Diavolumes 7 7 to 10 Critical
number of Diavolumes.
Output of process conditions including TMP,
Process Perfor-
Process Flux (average) recirculation rate, product concentration. May 25 to 30 LMH
mance Attribute
be used to track batch-to-batch consistency
Retentate pH and Conductivity at Critical Quality
Direct impact to product quality To Specification
end of step Attribute
Process – Final Concentration & Product Recovery
Chase Buffer pH and Conductivity Per solution specification Critical Direct impact to product quality
More likely to significantly affect flux at higher
Recirculation rate 8 L/min 6 to 10 L/min Key
product concentrations
Transmembrane Pressure (TMP) 10 PSI 5 to 15 PSI Key Impacts flux
Minor effect on flux. Assume no effect on
Temperature 20 °C 15 to 25 °C Non-Key
product quality over fairly wide range.
Process Perfor-
Process Flux (average) 15 to 20 LMH
mance Attribute
Chase Buffer Volume Determined by in-process measurement Procedural controls.
Must be in range to facilitate next process
Product Concentration after step. If final step in drug substance manu- Critical Quality
To Specification
Recovery & Chase facture, must be consistent with require- Attribute
ments for formulating drug product.
System Cleaning & Storage
Directly controllable and unlikely to affect
Cleaning Solution 0.4 to 0.6 N NaOH Non-Key
product or process
Adequate recirculation is needed to insure
Recirculation rate 10 L/min 8 to 12 L/min Non-Key
proper cleaning, but range is wide.
No impact to cleaning effectiveness over a
Transmembrane Pressure (TMP) 10 PSI 5 to 15 PSI Non-Key
wide range.

19 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Process Variable Attributes
Process Variable Proven Rationale
Parameter Product or
Set Point Acceptable Expected Range
Designation Process Attribute
Range
May impact cleaning effectiveness if far out
of range. Procedural controls in place such
Temperature 30 °C 25 to 35 °C Non-Key
that the risk of running outside the range is
unlikely.
Wide range, directly controlled to prevent
Time 60 min 60 to 90 min Non-Key
running outside of the validated range.
Directly controllable. Unlikely to affect prod-
Storage Solution Normality 0.09 to 0.10 N NaOH Non-Key
uct or process

3.5 Analytical Methods

Analyses of raw materials, in-process samples, drug substance, and drug product are important aspects of the Control Strategy (Section 3.8) and process character-
ization studies. Analytical methods used for such studies should be appropriate for their intended use, scientifically sound, reliable, and reproducible. Strategies for
qualification/ validation of the analytical methods used during development have been published, and provide approaches for evaluating tests used at this stage of
the lifecycle (27). Guidance on expectations for the analytical methods is also outlined in the FDA Guidance on Process Validation (3). Information on the analytical
methods used during process characterization studies should be included in the Process Characterization Plan, and documented in the study reports. Qualification
of the methods should also be documented. Since process characterization studies may be performed in development laboratories, instruments must be adequately
calibrated and maintained.

3.6 Risk Assessment and Parameter Criticality Designation

Risk assessment plays an important role in the development of a commercial control strategy. Risk assessments are performed by interdisciplinary teams at several
points during stage 1 of the lifecycle, and serve a number of purposes. (See Section 6.1 – Application of Risk Management) Risk assessment tools provide a struc-
tured means for documenting data and rationale associated with the risk assessment outcome, and becomes part of the documented process development history.

As shown in Figure 3.0-1, the initial identification of critical quality attributes is followed by a quality risk assessment in stage 1. The initial quality risk assessment
is a cause and effect type of analysis to identify process input parameters where variability is likely to have the greatest impact to product quality or process perfor-
mance. This assessment is based primarily on prior knowledge or early development work, and the outcome of this assessment provides the foundation for process
characterization studies that follow.

Understanding the impact of process parameter variability and applying the appropriate controls is a fundamental element in development of the commercial con-
trol strategy. ICH Q8 (R2) defines a Critical Process Parameter (CPP) as, “one with variability that has an impact on a CQA, and therefore, should be monitored or controlled
to ensure that the process produces the desired quality (3).

20 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Process parameters may be further categorized based on impact to the process. In certain circum-
stances, process performance is controlled and monitored as an additional means of ensuring a consis-
tent state of control. Process parameters that have been shown experimentally to impact process per-
formance may be classified as key process parameters (KPP). KPPs may impact process performance
attributes (such antibody titer in cell culture processes or yield in downstream purification), but do
not impact critical product quality attributes (15). In some processes, identifying and appropriately
controlling KPPs is useful since process performance measures may be an important means of dem-
onstrating intra-batch consistency.

Beyond the generally recognized definition of a critical process parameter from of ICH Q8 (R2),
however, process parameter designations are not standardized and approaches may vary. For this
reason, definitions for parameter designations must be clearly documented and understood within
the organization. Definitions should remain consistent throughout the process validation lifecycle.

Figure 3.6-1 provides an example of a decision tree developed to guide the assignment of parameter
designations in conjunction with the quality risk assessments. The decision tree facilitates categori-
zation of process parameters as critical, key, or non-key (see definitions). Decision making tools can
facilitate common understanding among participants, and have the advantage increases consistency
in the decision making process as well as consistent documentation of rationales as part of the risk
assessment process.

The decision tree can be used for risk assessments both before and after the supporting data from
process characterizations studies are available.
• Parameter or Attribute: Process variables can be outputs from one unit operation and inputs to another. For
a given unit operation, each variable is initially established as a parameter or an attribute on the basis of direct
controllability
Yes — Directly controllable process input parameters can theoretically contribute to process variability.
No — Process outputs that are not directly controllable are attributes that are monitored and may be indicative of
process performance or product quality.

• Process Parameters: Potential impact to critical quality attributes.

Yes — If impact is suspected, or if data show that variability in a parameter could impact a CQA, the parameter
is designated as a CPP. Although a parameter may be initially classified as a CPP, data from robustness studies
conducted during process characterization may show that CQAs are not impacted despite exaggerated variations
in the parameter. In these cases, the second risk assessment serves to change the assessment to non-CPP.

No — Parameter is a non-CPP and is further evaluated

• Non-CPP: Potential to impact process performance or consistency if run outside of defined range.
Yes — Parameter designated a KPP
No — Parameter has little impact to the process over a wide range. Parameter is designated a non-KPP.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 21

 Figure 3.6-1 Decision Tree for Designating Parameter Criticality

Control Process Output

Process Variable Can the Variable be NO Process Performance Attribute
or
Controlled? Product Quality Attribute

YES

Process Input
Process Parameter

Product Quality Impact

Potential Impact to Critical
YES Quality Attributes? NO

Potential CPP Non-CPP

NO

Risk Assessment Risk Assessment

Product Quality Risk Process Performance Risk
Process Control Strategy and Risk of Process Control Strategy and Risk of
Impact to Critical Quality Attributes Impact to Process Performance
Severity Severity
Likelihood Likelihood
Detectability Detectability

YES NO

YES

CPP KPP Non-KPP

Low Risk of Impact to
Potential to Impact Critical Potential to Impact Process
Process Performance or
Quality Attribute Performance or Consistency
Product Quality Attributes

22 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

3.7 Process Characterization
Process characterization is a set of documented studies in which operational parameters are pur-
posely varied to determine their effect on product quality attributes and process performance. The
approach uses the knowledge and information from the risk assessments to determine a set of process
characterization studies to examine proposed ranges and interactions for process parameters. The re-
sulting information is used to define the PPQ ranges and acceptance criteria. It can also be used to set
the final parameter ranges and can be used to develop a Design Space if using an enhanced approach,
i.e., incorporating advanced analytical and/or manufacturing control technologies, to process de-
velopment. Experiments can be designed to examine proposed ranges and explore ones wider than
those that will normally be used in operation. An element of process characterization may include
multivariate designed experiments to define process design space. While univariate approaches are
appropriate for some variables to establish a proven acceptable range (PAR), multivariate studies ac-
count for interactions between process parameters/material attributes (1).

Since Studies designed to characterize the process and setting acceptable ranges for process param-
eters are usually performed at laboratory scale. The ability of laboratory-scale studies to predict pro-
cess performance is desirable. When a laboratory scale model is used in development, the adequacy
of the model should be verified and justified. When there are differences between actual and expected
performance, laboratory models and model predictions should be appropriately modified. In that the
conclusions drawn from the studies are applied directly to the commercial-scale process, qualification
of laboratory-scale models is essential. Qualification of the scaled- down models should confirm that
they perform in a manner that is representative of the full-scale process. This is shown by comparing
operational parameters and inputs and outputs, including product quality attributes.

Scaled- down models for chromatography steps for protein products can be qualified by performing mul-
tiple runs with input parameters at set points and comparing the results to the full-scale unit operation.
Parameters evaluated should include those that affect process consistency, such as step yields, elution
profile, elution volume, and/or retention time. These should then be combined with those that represent
product quality, such as pool purity and levels of process-related and host cell-related impurities.

Pilot-scale models of small molecules that are representative of the commercial manufacturing pro-
cess may be used for supportive PPQ data. In solid and liquid oral dosage forms, 10% of the commer-
cial batch size and/or 100,000 units have been considered a representative scale (1). Scale-up effects for
certain processes, such as mixing freely soluble substances, tablet compression, or liquid filling may
be well-known. Batch sizes at 10% of bulk size or run times of 100,000 dosage units provide a suf-
ficient duration to determine a degree of control and process characterization, while uncovering any
preliminary major problems. Full-scale confirmation/evaluation may be carried out when small-scale
studies are used to support PPQ.

For scale-down studies, the raw materials, component attributes, equipment, and process parameters
should be comparable and indicative of the process intended for the commercial product.

3.8 Product Characterization Testing Plan

Some product characteristics may not be tested as part of the routine release test panel. Examples
of such product characteristics include residual DNA levels for biotechnology products (when DNA
clearance has been established at a level that clearly exceeds safety requirements) or final product
porosity for solid oral dosage products (when dissolution testing is performed). In addition to release

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 23

 specifications, Stage 1 deliverables should include other tests on the DS, DP, or critical intermediates
that are needed in order to claim a comprehensive understanding of the product and process.

3.9 Control Strategy

Establishing an effective and appropriate process control strategy is one of the most important out-
comes of pharmaceutical development in Stage 1. An appropriate control strategy is based on knowl-
edge and experience gained in Stage 1 and its effectiveness will dictate the extent to which a manufac-
turing process remains in a state of control. As with the other aspects of stage, 1 discussed above, the
development of an effective process control strategy is an iterative process. It starts, early in develop-
ment and evolves as process and product knowledge increase. A robust control strategy encompasses
all elements of individual unit operations in the process. All product quality attributes and process
parameters, regardless of whether they are classified as critical, are included in a complete process
control strategy which includes the following elements:

Raw Material Controls

The ability to manage the quality of the inputs (raw materials and components) to assure a consistent
output is an essential aspect of a process control strategy. Inputs should be categorized based on their
potential risk for introducing variability or contaminants into the product and/or process. Product
variability may include changes to CQAs, whereas process variability may include inconsistencies in
yield, reaction kinetics, filterability, or other non-product, quality-related effects. For many raw ma-
terials used in the manufacturing process, selection of appropriate grades (based on purity, chemical
and physical characteristics, and/or microbial specifications, such as endotoxin) may be an adequate
level of control. For higher risk raw materials, understanding the contribution to product and process
variability may be essential to establishing specifications for those materials. Once the relationships
are understood, appropriate risk reduction steps can be made part of the control strategy (see Section
6.1.4).

In-Process and Release Specifications

In-process and product specifications may be related to product safety and efficacy or may assure
product consistency. Confirmed failure to meet a product specification (in-process or product) dis-
qualifies material from clinical or commercial use. Guidance on setting specifications is provided in
ICH guidance documents Q6a and Q6b.

In-Process Controls
In-Process Controls (IPCs) are inputs to the process and are checks performed during production to
monitor and, if appropriate, to adjust the process, and/or to ensure that the intermediates or product
conform to specifications or other defined quality criteria.

Performance Parameters
Performance parameters (e.g., tablet/capsule disintegration; harvest or peak growth cell densities/
viability) are process outputs that cannot be directly controlled but are indicators that the process has
performed as expected.

Process Parameter Set Points and Ranges

Knowledge of the effects of process parameter variability on the output of each Unit Operation and
on the final product evolves during Process Development and Process Characterization (Section 3.7).
This information, along with process equipment capability (Section 4.1), is used to establish param-

24 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

eter set points and ranges (including ranges for alarms and deviations). It may also be used to assess
the severity of process deviations caused by parameter excursions. Parameter ranges may be desig-
nated as normal operating ranges (NORs), or where proven by supportive data, as proven acceptable
ranges (PARs).

Process Monitoring (Data Review, Sampling, Testing)

Process monitoring includes measurement data (e.g., flow rates, temperatures, volumes, pH), in-
process sampling plans, and appropriate analytical assays. Data collection and analysis begins in Stage
1 and are integral parts of Stage 2, Process Performance Qualification. The data collection effort
eventually evolves into the continued process monitoring program described for Stage 3, Continued
Process Verification (see Section 5.0, “Continued Process Verification, Stage 3”).

Processing and Hold Times

Hold conditions and times are an essential part of the process control strategy for all process interme-
diates (or in-process materials), drug substance, bulk drug product, and prepared solutions. Studies
should be performed to support these limits. Time limits for processing steps should also be part of
the control strategy.

Process Analytical Technology (PAT)

Process Analytical Technology (PAT) is one approach to implement the Control Strategy (28). Using
PAT, CQAs are monitored in real-time (using on-line or at-line analytics), and results are used to adjust
CPPs during production to decrease product variability (CQAs) or achieve consistent CQAs at desired
ranges with low variability.

PAT uses product and process knowledge as well as equipment automation and analytical instrumen-
tation technologies. Successful application of PAT requires a thoroughly characterized process (Sec-
tion 3.7) in which the relationship between CPPs and CQAs is explored using mathematical models,
such as multivariate analysis. Application of this understanding to the Control Strategy (Section 3.9)
also affects the design and qualification of the instrumentation and control systems in the manufac-
turing process.

To support implementation of PAT, Stage 1 deliverables must describe the CQA monitoring scheme
and the algorithm for adjusting CPPs based on the process response. Qualification of the equipment,
measurement system, and process (Stage 2) must demonstrate the capability to adjust CPPs accord-
ing to the established algorithm and confirm that these adjustments result in acceptable and predict-
able outputs. Therefore, PAT-based control methods need to be qualified (29).

3.10 Clinical Manufacturing Experience – Batch Records and

Production Data
During Stage 1, clinical batches are used in clinical trials to support product approval. The data may
be used along with formal process characterization data to support the establishment of manufactur-
ing process parameters and the process control strategy. These data also comprise the beginning of
the process monitoring that will continue after PPQ. Early-batch data may not include all controls
implemented in the final commercial process, but the information is still valuable for evaluating the
process performance. If used to support ranges and limits, clinical batch data should be included in
the final process design report that will justify the process and the control strategy. The final batch
records should be generated at the end of Stage 1. They will support the finalized commercial process
and serve as a prelude to Stage 2.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 25

 In some cases, it may be appropriate to use data from clinical batches to support the PPQ Stage 2. The
rationale for this approach should be documented and included in the Process Validation Master Plan.

3.11 Process Design Report

The process design report is also a Stage 1 output. As a living document that describes in detail the
intended commercial process, it may have various titles in internal procedures. Stage 1 study data are
used to support this document and to justify the ranges, and process control strategy. Additional data
and process knowledge are gained gathered as the manufacturing process changes, and are incorpo-
rated during Stages 2 and 3. The process design report, this document should be updated to include
this new information. This comprehensive document includes:
• Reference to CQAs and supporting risk assessments
• Process flow diagrams
• Process description tables
• Inputs (in-process controls)
• Outputs (in-process tests and limits, in-process specifications)

• Process parameters and ranges

• Classification of parameters for risk of impact to CQAs and process performance
• Design space, as appropriate
• Justification and data supporting all parameter ranges (e.g., characterization data, development studies, clinical
manufacturing history)

3.12 Process Validation Master Plan

A process validation master plan may be initiated during Stage 1 to prepare for Stage 2 activities. It
should outline the validation strategy and supporting rationale, and typically includes:
• Process characterization plan
• Description of the manufacturing process and control strategy
• Functions and responsibilities
• PQ or PPQ plan:
• PPQ strategy (e.g., single unit operations or a combination of unit operations, bracketing, family, or matrix
approaches) and a list of individual protocols; applicable ancillary studies, (e.g., mixing, media preparation, in-
process pool hold time, resin lifetime)
• List of equipment and facilities to be used
• List of analytical methods and their status
• Sampling plan
• List of protocols to be executed under the plan

• Proposed timeline and schedule of deliverables

• Procedures for handling deviations and revisions
• Continued Process Verification plan

3.13 Stage 1 Manufacturing and Technology Considerations

The capability of the production equipment and procedures has a significant influence on the abil-
ity to maintain process parameters within pre-set limits. The measurement and control capability
of the process equipment is one of the subjects of Stage 2, Process Qualification, and can be found

26 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

in Section 4.1 Equipment qualification exercises should confirm the suitability of equipment for its
intended use.

Compatibility of the process streams with the equipment and materials that they contact (e.g., poly-
meric membranes, elastomers, disposable bags, and other plastic parts) is necessary to ensures prod-
uct safety and efficacy. Product contact materials as well as extractables and leachables need to be
evaluated for compatibility. This work should begin in Stage 1, may include studies that require long
lead times, and should be completed in conjunction with Stage 2.

Compatibility of the process streams with equipment surfaces is a measure of their reactivity, absorp-
tion, and stability when in contact during manufacturing. Compatibility tests should demonstrate
that the material properties of the equipment surfaces are not altered by contact with the solutions
or other product-related materials. In addition, the contact materials should not alter the process
solutions or materials (either by adsorption of product components or excessive leaching that could
adulterate the product).

Extractables are components of a material (e.g., a product contact surface that is used in drug manu-
facture or storage) that are recovered by use of an exaggerated force (solvent, time, temperature).
Leachables are contact material components from process equipment or storage containers that mi-
grate into the product under normal conditions of use.
The identity and quantity of leachables from polymeric wetted components (plastic storage contain-
ers, filters, primary packaging materials, gaskets and O-rings) used in drug manufacture, storage,
and packaging must be documented to assure that the product is not adulterated. A combination of
literature reviews, risk assessments, and laboratory studies can be used to address leachables. Various
approaches to determine the extent of testing and identification of leachable species, and the setting
of acceptable levels, have been published.
• “Safety Thresholds and Best Practices for Extractables and Leachables in Orally Inhaled and Nasal Drug Products” (30)
• “Evaluation of Extractables from Product-Contact Surfaces” (31)
• “Application of Quality by Design (QbD) Principles to Extractables/Leachables Assessment: Establishing a Design Space for
Terminally Sterilized Aqueous Drug Products Stored in a Plastic Packaging System” (32)
• “Leachables Evaluation for Bulk Drug Substances” (33)

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 27

 4.0 Process Qualification (Stage 2)

Process Qualification (PQ) during Stage 2 demonstrates that the process works as intended and yields
reproducible commercial product. It should be completed before release of commercial product lots,
and covers the following elements:
1. Design and qualification of the facility, equipment, and utilities (this should be completed prior to
qualification of the process)
2. Process Performance Qualification (PPQ), which demonstrates control of variability and the ability to
produce product that meets predetermined quality attributes

4.1 Strategies for System Design and Qualification

Facilities, equipment, utilities, and instruments (collectively referred to as systems) used in the manu-
facturing process should be suitable and capable for their intended process use, and their performance
during the operation should be reliable. Systems that affect product quality should be qualified to re-
duce the equipment performance as a process variable. The review and qualification of these systems
should be performed according to a pre-defined project plan. System qualification should precede
Stage 2 PPQ activities. Qualification studies should be completed, reviewed, and approved, with all
deviations addressed, prior to the start of PPQ studies.

The following section provides considerations for preparation and performance of system qualifica-
tion. More information on approaches to planning and performing system qualification may be found
in several sources (26,34,35). Figure 4.1-1 presents a typical sequence of activities that support the
system qualification effort.
Figure 4.1-1 Typical System Qualification Sequence

Transfer process design requirements and CPP

information from Stage 1 Process Design

Perform design qualification/review to confirm system

Engineering Design Perform system risk
design is aligned with process requirements
or impact assessment
to determine which systems
and system components are
Construction, Perform factory acceptance testing and site acceptance crtical to product quality
Installation Fabrication, and
Instalation testing to confirm system meets design specification

Perform Engineering and Commissioning Studies to

Start Up Commissioning Develop Qualification
Confirm systems are in reliable working order
Protocol, including test
functions and acceptance
criteria, based on process
System Evaluate information obtained during commissioning to requirements and
Qualification/ Testing system capabilities
Verification determine what can be levraged during qualification

Perform system qualification studies to ensure that

systems are capable of meeting process requirements

Evaluate qualification results and release system for

Process Performance Qualification

28 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

4.1.1 Engineering and Design
Facility, equipment, and utilities should be designed to meet process requirements. The design of the
facility and commissioning of the equipment and utilities should assure the capability of operating
as required for routine manufacturing. These activities and all commissioning-related tasks should be
conducted according to Good Engineering Practices (GEP), and recorded according to Good Docu-
mentation Practices (GDP), with oversight by the Quality Unit. Risk-based approaches may be used
to assure adequate controls and verification.

System design should be based on process parameters, control strategies, and performance require-
ments developed or identified during Stage 1 Process Design. This information is transferred to those
designing engineering requirements for facility and manufacturing systems. Design qualification in-
volves a review of the system design to assure that it is aligned with process control strategy and
performance requirements.

In situations where the process is being transferred to an established facility with qualified equipment,
a risk assessment should be performed to identify any equipment control gaps. These can be ad-
dressed through equipment modifications (which may require requalification) or through operational
controls.

4.1.1.1 Risk Assessment

Risk assessments determine which systems and system components have an impact on the establish-
ment and maintenance of process parameters and conditions that affect product quality. This infor-
mation helps develop system qualification plans, protocols, test functions, and acceptance criteria.
The process steps and systems that affect product quality, the mode of effects, and the correlation
between system performance and control of process variables should be understood. For more infor-
mation on risk assessments, see Section 6.1.

4.1.2 Installation
Upon completion, system testing and inspection should be used to verify that the systems have been
fabricated, constructed, and installed to engineering, and process specifications. The information
from this verification should be accurate, reliable, and useful. If so, then information from these ac-
tivities may be leveraged or used to support qualification testing.

The start-up and commissioning of these systems should confirm that they are in good working order
and operate as designed. Engineering studies can provide confidence that the systems will perform
under process conditions. Adjustments to the systems to achieve the specified level of performance
and operation may be needed. Information on modifications or adjustments should be documented
and transferred to the team preparing the qualification plans and protocols.

4.1.3 Qualification Plan

The qualification plan may be developed at any time once the process requirements and correlation
to process systems are understood. Early development of the qualification plans may provide valu-
able guidance to the design, installation, and commissioning efforts. However, to capture any changes
that result from start-up and commissioning, it may be prudent to complete the qualification plans
and protocols after all information from the commissioning has been transferred. This approach also
enables a better understanding of the type and amount of information that can be leveraged from
pre-qualification activities. This approach means that Stage 2 activities may be underway during and
prior to completion of all Stage 1 activities.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 29

 Table 4.1.3-1 presents examples of some of the information that can be used to help develop test
functions and acceptance criteria.
Table 4.1.3-1 Qualification Information
Information Purpose
Control Capability Assessment for each controlled parameter indicating control band
widths under test conditions that are relevant to the process that is be-
ing qualified. Analysis conducted in conjunction with process control
requirements and may impact process parameter designations.
Capacity Range of operating capacities for each unit operation or step, demon-
strating consistency with proposed process.
Detection Capability Assessment of monitoring instruments and sampling points showing
accuracy of instrument outputs and sampling limitations. (To be used
in performing Risk Assessment and establishing the monitoring strat-
egy for the process.)
Alarms and Interlocks Check for consistency with process requirements and safety concerns.
Process Stream/Product Listing of all contact materials for all equipment used to process prod-
Contact Materials ucts or materials added to the product stream. Used for compatibility
assessment and leachables/extractables analyses.
Maintenance and Documentation of required equipment service and calibration of all in-
Calibration Programs strumentation.

4.1.3.1 Test Functions and Acceptance Criteria

System qualification tests or studies should be based on knowledge gained from previous activities,
including Stage 1 Process Design, and engineering studies. Test functions should be based on good
scientific and engineering principles designed to demonstrate and assure that anticipated operating
parameters will be met throughout the manufacturing process in a consistent and predictable man-
ner. Acceptance criteria should be based on sound scientific rationale; the criteria should be useful,
attainable, and where appropriate, quantifiable.

If sufficient process understanding is not available, or the scale-up effect is unknown, existing knowl-
edge may be used during design and commissioning to define user requirements.

Formal system operating and maintenance procedures or instructions should be in place prior to the
execution of test functions. Operators and those conducting studies should be trained in the opera-
tion of the systems and conduct of the tests. These should be conducted under GMP conditions and
documented according to GDPs. All measuring and test instruments should be calibrated and trace-
able to appropriate standards.

Deviations in the execution of qualification testing should be documented, investigated, and ad-
dressed. Conclusions should be based on the suitability and capability of the system to meet the pro-
cess requirements. When necessary, systems may be modified and studies repeated.

4.1.4 Maintaining Systems in a State of Control

Qualification studies ensure that the manufacturing systems, as designed and operated, are in a state
of control. For the process to remain valid and controlled, the systems must be maintained in a state
similar to that demonstrated during qualification.

30 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Periodic assessment and evaluation of the system to determine its control status is important. The
assessment should include a review of information that indicates or supports assurance of control.
This information may include, but is not limited to, such items as:
• Calibration records • Work orders
• Preventative and corrective maintenance records • Monitoring results and trends
• Equipment logs
• Non-conformance reports and deviations
• Training records
• Standard Operating Procedures • Failure investigations
• Change requests • Re-qualification studies

Periodic assessment of systems may lead to additional qualification-related activities or testing. In addition to
periodic assessment, event-driven assessments and re-qualifications may arise from process-related changes,
out-of-specification results and trends, and investigations. The System assessments and the events that trig-
ger event-driven assessments should be recorded in a formal procedure that also addresses the mechanism
for deciding when re-qualification is warranted, the criteria for doing so, and those responsible. It is recom-
mended that Subject matter experts and the quality unit should also be involved in these decisions.

4.2 Process Performance Qualification

Process performance qualification marks the transition from development and clinical manufacturing
to routine commercial production. Process Performance Qualification (PPQ) demonstrates the valid-
ity of the process design and the suitability of the process control strategy at the commercial manu-
facturing scale. PPQ provides confidence that the systems of monitoring, control, and procedures in
routine manufacturing are capable of detecting and compensating for potential sources of process
variability over the product lifecycle.

The number of “successful” batches executed during the PPQ study should not be viewed as the
primary objective of a PPQ campaign. While successful runs of commercial-scale batches can indi-
cate overall operational proficiency and sound process design, these batches should also be viewed
as a means to obtain information and data needed to demonstrate that the process control strategy
is effective. The type and amount of information should be based on understanding of the process,
the impact of process variables on product quality, and the process control strategy developed during
Stage 1 Process Design. As appropriate, other prior knowledge should be used as well. The number
of batches needed to acquire this information and data, may be based in part on a statistically sound
sampling plan that supports the desired confidence level. It may also be influenced by the approach
selected to demonstrate that the batch-to-batch variability of CQAs is acceptable.

This section will discuss design strategy for the PPQ, recommended content for the protocol and
report, and the transition to Stage 3 of the process validation lifecycle.

4.2.1 PPQ Readiness

The transition from Stage 1 to Stage 2 of the process validation lifecycle is not strictly sequential.
Completion of some Stage 1 activities may overlap with those from Stage 2. Likewise, some prepara-
tive Stage 2 activities will be initiated in parallel with those from later Stage 1 activities. Components
of Stage 1 PPQ activities (as discussed in Section 3.1) include, among others: drafting of the Process
Validation Master Plan; initiation of the qualification of facilities, utilities, and equipment; drafting
of protocols for the PPQ studies; training of personnel; or drafting an initial CPV plan. Although
initiation of PPQ activities does not is not depend on completion of all Stage 1 activities, a readiness
assessment should be conducted to determine the timing of sufficient information and completion of

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 31

 activities to support moving forward with PPQ batch manufacture. The readiness assessment should
include deliverables from Stage 1 (as outlined previously in Section 3.1) and other elements:
Quality Target Product Profile — Initiated in at the start of Stage 1, but updated to reflect knowledge
obtained from Stage 1 prior to initiating PPQ.

Critical Quality Attributes with Criticality Assessment — CQAs are identified early in Stage 1. They
are confirmed to account for additional analytical characterization, clinical and/or non-clinical data
and information gathered during Stage 1. CPPs that impact QCAs are reviewed and updated based on
detectability and occurrence (11,36).

Commercial Manufacturing Process Description — This is started in Stage 1 and updated to

reflect the finalized commercial process supported by Stage 1 studies/data. These include elements
outlined in Section 3.4, and any changes resulting from the qualification of the facilities, utilities, and
equipment as outlined in Section 4.1.

Analytical Methods — Appropriately validated or suitably qualified methods should be identified and
their status documented. Methods for product release and stability should be fully validated accord-
ing to ICH requirements prior to initiating PPQ batch testing. Additional tests beyond normal release
testing used to support PPQ should be identified and suitably qualified/validated prior to being used
to test PPQ batches. The justification of the status for use in the PPQ studies (qualified and/or vali-
dated) should be fully documented for each analytical method.

Approved commercial batch records — Changes may be made to batch records during Stage 1 should
enhance, clarify, or optimize manufacturing instructions and/or to reflect knowledge gained during pro-
cess characterization. Batch records reflecting the final commercial process to be studied in PPQ should
be approved prior to PPQ batch execution.

Process Design Report — This report (as described in Section 3.11) is the repository for the process de-
sign justification, and includes parameter risk ranking, and ranges for the process that will undergo PPQ
study. The data summarized in this report will support the selection of the elements of the PPQ stud-
ies and proposed PPQ acceptance criteria. The process development summary should provide the link
between the detailed process description, risk assessments, control strategy description, characterization
reports, rationale for parameter designations, and clinical manufacturing history. It is a best practice for
this information to be finalized prior to PPQ study design since it provides the scientific support to justify
the PPQ acceptance criteria.

Process Validation Master Plan (PVMP) — Drafting of the process validation master should begin
in Stage 1 and be finalized prior to PPQ study initiation. Elements of the Process Validation Master
Plan are outlined in Section 3.12.

Quality System and Training — Qualified and trained personnel will be integral to the PPQ studies.
Detailed, documented training specific to the PPQ is recommended for functional groups directly
involved in the execution of the study. To minimize the risk of human error, personnel should un-
derstand their role in protocol execution to minimize the risk of human error. Quality Unit approval
of PPQ activities should be completed prior to PPQ study initiation, and all PPQ studies should be
conducted within the quality system.

Approved protocols for PPQ Studies — Protocols for each study should be approved and finalized
prior to initiation of PPQ studies. Design and content of process performance qualification protocols
is discussed in Section 4.4.

32 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

4.3 Design Strategy for Process Performance Qualification (PPQ)
4.3.1 Use of Prior Knowledge and Stage 1 Data to Support PPQ
In a lifecycle approach to process validation, sources of data and information outside of the PPQ
batches may be used to support a high degree of confidence in an ongoing state of process control.
Prior knowledge is that which has been gained from similar products and processes. It may come
from experience with a portfolio of similar molecules, where platform manufacturing strategies have
been developed using existing facilities and equipment (e.g., platform manufacturing processes for
monoclonal antibodies), or from similar process and unit operations. Leveraging the body of data
from similar products and processes may provide an additional level of confidence in the process con-
trol of a product and process that uses a similar control strategy and unit operations.

By contrast, first-in-class molecules and/or products manufactured in new facilities/equipment will

not have a similar depth of prior knowledge and data prior to development. In these instances, in-
creased emphasis on data gathering in Stage 1 may be applied to support PPQ readiness. To gather
sufficient data to demonstrate an acceptable level of confidence in the commercial manufacturing
process when little prior knowledge or Stage 1 data are available, the scope and extent of PPQ may be
greater. The rationale and scientific justification for the use of existing data (prior knowledge) to sup-
port the PPQ Stage should be documented in the process validation master plan. All prior knowledge
and Stage 1 data used in to support PPQ must be retrievable, traceable, verified, and generated using
good scientific practices.

Figure 4.3.1-1 illustrates the relationship of the amount of knowledge to Stage 1 and 2 activities.
Where there is greater prior knowledge or process design for a new product or process, PPQ studies
may be decreased. Less prior knowledge will require more Stage 1 and/or PPQ data.
Figure 4.3.1-1 Relationship of Prior Knowledge to the Amount of PPQ Data Required

Comprehensive Prior Knowledge May Support Requiring Fewer PPQ Runs

Prior Knowledge Process Design PPQ

Limited Prior Knowledge May Require More PPQ Runs

Prior Process Design PPQ

Knowledge

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 33

 Some examples of cases where prior knowledge may be useful to for PPQ include:
• Setting of acceptance criteria in PPQ studies: For example, bioburden and endotoxin in-process acceptance criteria
in cases where facility history and limits for other processes can be applied to similar processes that employ the
same facility and equipment. (Assumes the limits for the previous product are appropriate for the quality of the
new product.)
• Use of data from other product PPQ supportive studies: For example, in platform purification operations where
same or similar buffer formulations will be used in the same vessels, buffer hold studies already performed for a
different product can be used to support the PPQ for buffers used for the new product.
• Using prior experience on similar processes: In non-sterile solid and liquid dosage manufacturing, such as
granulating and film coating solution preparations, or bulk solution mixing and filling, prior experience with similar
solutions or filling equipment may be applied to justify the number of PPQ batches for those unit operations.
Past knowledge of common excipients, such as fillers, binders, disintegrants, lubricants, and preservatives in the
formulation and process is also an important factor.

Use of Stage 1 Data for PPQ

Processes and products for which there is little or no prior knowledge may require a greater emphasis
on Stage 1 and PPQ activities to demonstrate an acceptable level of confidence in the process control
strategy. Data from Stage 1 process characterization studies and clinical manufacturing are generally
used to support the establishment of the control strategy for new products, as discussed in Section
3.0. Stage 1 data may be used to support PPQ if sufficient scientific evidence for its use is available.
At a minimum, the studies claimed to support PPQ should represent the commercial manufacturing
scale (e.g., be scale independent) or derived from qualified small-scale model(s) proven to represent
the full-scale process. In some cases, data from clinical manufacturing batches may be used in con-
junction with that gathered during PPQ to increase the amount of data that can be used to achieve an
acceptable level of confidence in the process. Some examples of the use of Stage 1 data to support the
PPQ include (see Section 7.1 for details):
• Large molecule example
• Past experiences in clinical, and stability, and pilot batch manufacturing. Process evaluation batches help
determine the amount of PPQ data. For example, in an oral solid dosage form of multiple strengths, at least
8 Stage 1 batches with the same commercial formulation were from clinical supply manufacturing, stability,
pilot, process evaluation/design, and plant demonstration batches. The firm had extensive experience with the
components, equipment, and unit operations for the dosage form: wet granulation, fluid-bed drying, milling,
blending, compression, and film coating. These 8-plus batches played instrumental roles in the justification of
the number of PPQ batches.

In some cases, Stage 1 data that supports PPQ may be supported in some cases by adding stricter
testing for a defined number of batches to confirm the results obtained in the Stage 1 studies and the
PPQ batches. For example, small-scale column lifetime studies may be used to support column reuse
limits. These are then confirmed with a heightened level of impurity monitoring until the reuse pe-
riod has been reached at full scale.

4.3.2 PPQ Study Design

Process Performance Qualification is a means to demonstrate that all important elements of a process
unit operation are under the appropriate degree of control, and that all important variables and ele-
ments of the unit operation have been considered (facility, utilities, equipment, personnel, process,
control procedures, and components). During PPQ, critical process parameters and critical quality
attributes are monitored along with process performance parameters. Their evaluation is useful in
demonstrating consistency and can enhance confidence in the overall process control strategy when
included in the PPQ. All parameters and attributes intended for ongoing Continued Process Verifica-
tion in Stage 3 should be included in the PPQ.

34 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

4.3.2.1 Number of Batches
The PPQ should be viewed as a means to evaluate and confirm a sound process design, an effective
control strategy, and operational proficiency at commercial scale. The number of batches in the PPQ
study(ies) will be influenced by many factors such as
• the performance and acceptance criteria,
• the analyses to be performed and the type and amount of data necessary to perform those analyses.
• the level of process knowledge and understanding gained from Stage 1,
• the type and complexity of manufacturing technology employed in the various unit operations,
• knowledge from previous experience with similar well controlled processes
• the inherent/known variability of the process resulting from raw materials, age of the equipment, operator
experience,

Using risk-based approaches allows a balance between the number of batches studied and the risk of
the process. They can also be used in conjunction with objective approaches to determine the number
of batches to include.

Where practical, statistical methods are recommended to guide the determination of the number of
PPQ batches needed to achieve a desired level of statistical confidence (see Sections 6 and 8 on statisti-
cal approaches to determining the number of batches and sampling plans). However, this approach
alone may not always be feasible or meaningful. One such example where is PPQ studies of a protein
drug substance process with a limited number of clinical batches. This dearth of output could be due
to such factors as manufacturing scale or product indications (e.g., orphan drug) where infrequent
future manufacturing campaigns are to be performed. In addition to limitations on manufacturing
batch production, the nature of protein drug substance manufacturing makes increased sample sizes
of the process streams of limited usefulness to achieve a statistically-based sample size. When it is
not feasible or meaningful to use conventional statistical approaches, a practical, scientifically-based,
holistic approach may be more appropriate. In this case, the following factors may be used to support
the rationale for the number of PPQ batches selected:
• Prior knowledge and platform manufacturing information/data
• Risk analysis of the process to factor the level of risk into the batch number selection
• Increased reliance on Stage 1 data to support that the process is under control and to add to the data set
• Continuation of heightened sampling/testing plans during continued process verification until a sufficient
dataset to achieve statistical confidence has been accumulated.

When a combination of approaches and data are used, the rationale and justification should be clear-
ly documented in the process validation master plan. Also, references to all supporting source data
should be included.

4.3.2.2 PPQ at Normal Operating Conditions

PPQ studies are typically conducted in a manner that demonstrates a state of control under normal
operating conditions to assess process variability expected during routine production. Process char-
acterization (robustness) studies conducted during Stage 1 serve as the foundation for establishing
normal operating ranges, proven acceptable ranges, and design space, if appropriate. Effects of scale
should also be considered if scaled-down models are suitably qualified, well-planned, and executed.
Study data on robustness should support conducting commercial-scale PPQ under routine manufac-
turing conditions. Supplemental engineering studies at scale may be appropriate to evaluate extremes
of the normal operating range (e.g., line speed or compression speed). In most cases, available Stage

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 35

 1 data make it unnecessary to execute PPQ over the entire operating range during the commercial
manufacturing process. The process validation master plan should provide the justification for the
approach used and reference all source data.

4.3.2.3 PPQ Using Individual Unit Operation Studies

PPQ of a manufacturing process can be achieved by performing PPQ studies on each individual unit
operation (or related groups of operations). This approach calls for the writing of individual proto-
cols that outline the studies to be conducted on each unit operation. The overall objective is achieving
PPQ for the entire process. By emphasizing unit operations that have more variability, higher risk of
impact on CQAs, or more limited Stage 1 data available to support assurance of process, this strategy
may facilitate more flexibility in PPQ design. Protocols should define the testing performed and ac-
ceptance criteria for the output of the unit operation (intermediate). They and may also require that
the final drug substance or drug product meets all specifications and predefined acceptance criteria.

4.3.2.4 PPQ Using Bracketing, Matrix, and Family Approaches

Many operations involve similar or identical process operations or equipment. In these cases, designs
where grouping is used may be considered. Some process variables that might be amenable to ap-
proaches using bracketing, matrix, or family grouping PPQ include:
• Batch sizes • Various vial sizes and/or fill volumes of the same
drug product (e.g., smallest and largest vial size)
• Drug product dosage strength
• Filling line speeds (e.g., fastest and slowest line
• Identical equipment speed)
• Different size vessels, tanks, or similar • Product packaging (e.g., bottle heights or dosage
configurations of the same design and operating counts)
principle or in-kind equipment • Transport validation for biological products

4.3.2.5 Bracketing Approach

Bracketing qualifies processes that represent the extremes of process variables under the premise that
the extremes are fully representative of intermediate groups. The bracketing strategy is used when a
single process element can be varied while all other variables remain fixed.

Examples of cases where the use of bracketing approaches may be considered:

• Use of a common blend or solution, i.e., identical-strength tablets or those very closely related in composition
(e.g., for a tablet range made with different compression weights of a similar basic [common] granulation, or a
capsule range made by using different fill weights of the same basic composition into different size capsule shells).
• A blend concentration of 50 mg active ingredient /100 mg powder, could be compressed into a 100 mg active
(per 200 mg tablet weight), 200 mg active (400 mg tablet weight), and 300 mg active (600 mg tablet weight). The
same powder blend is common to the three tablet strengths.
• Capsules or liquid fills where common blends or solutions are used for filling into the final dosage form.

• Different container sizes or different fill volumes in the same container closure system.

The rationale for selection of representative groups and numbers of batches should be scientifically
justified, risk assessed, and outlined in the process validation master plan and PPQ protocols.

4.3.2.6 Matrix Approach

A matrix approach is appropriate for commercial manufacturing PPQ when configurations of the
same process and product have more than one variable. The approach is based on the assumption that
the batch configurations selected for inclusion in the PPQ fully represent processes for all combina-

36 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

tions. The rationale for the selection of combinations, and the number of batches representing each
combination, should be scientifically justified, risk assessed, and documented in the process validation
master plan and PPQ protocols. Some processes require a comprehensive PPQ. In those cases, it is
advisable select batches or lots from all combinations.

An example of a matrix design is shown for a PPQ of a filling process where manipulation of three
variables results in multiple drug product strengths. Variables in this example include:
• Fill Volume
• Bulk Drug Product Solution Concentration
• Final Drug Product Strength
Table 4.3.2.6-1 Illustration of a Matrix Approach for Filling Process PPQ

Fill Bulk Drug Solution Batch

Volume Concentration Final Drug Product
included in Rationale
Strength (mg)
(mL) (mg/mL) PPQ?

Lowest drug product strength;

Lowest Bulk Drug Product concen-
2.00 0.2 Yes*
0.10 tration;
Lowest fill volume
4.00 0.4 No Covered by matrix
4.00 0.6 No Covered by matrix
0.15
8.00 1.2 Yes* Highest drug conc. in fill solution
4.67 1.4 No Covered by matrix
6.00 1.8 No Covered by matrix
0.30 Highest drug conc. in final Drug
6.67 2.0 Yes* Product
Highest fill volume
*Based on the assumption that process variability is highest at these conditions

Rationale for selection of representative groups and number of batches should be scientifically justi-
fied and outlined in the process validation master plan and PPQ protocols.

4.3.2.7 Family (Grouping) Approach

A family approach is appropriate when multiple related but different entities can be grouped so that a
single one represents the common characteristics or worst case of each group. The rationale for fam-
ily groups and justification for the representative selection should be included in the validation master
plan and PPQ protocol. All variations in the formulation or method of manufacture should be de-
scribed and evaluated in detail. Two examples of the use of the family approach for PPQ are provided.

Equipment Family
Cell culture for biological product manufacturing can be performed in multiple trains using the same
equipment and process in each one. Use of a family or grouping approach may be valid for the PPQ
for the fermentation unit operations. This example shows how such an approach, which limits the
number of batches for the PPQ versus repeated multiple runs from each fermenter, could be used.
In this case, each equipment train was evaluated for similarity of the equipment (identical equip-
ment trains with duplicated equipment of the same model and manufacturer). Identical equipment
trains reduce the number of batches needed to show that the process is reliable in each one. In this

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 37

 case, there is ample prior knowledge on the performance of the process. Use of a reduced number
of batches in a family approach should take into consideration the amount of prior knowledge of
the process, the number and impact of the critical process parameters, and the ability to control the
parameters within the ranges. For a unit operation with no critical parameters, use of fewer batches
may be appropriate. In these cases, the approach should be clearly justified with reference to support-
ing data in the validation protocol.
Example 1
Equipment Family – PPQ Runs
Assessment Supporting Data
Production Bioreactor (Unit Operation)
#1 Compare: Bioreactor #1 – 3 batches Multiple small-scale
#2 • Physical design Bioreactor #2 – 1 batch process characterization
#3 • Design specs Bioreactor #3 – 1 batch runs available to support
• Materials of con- ranges
struction
• IQ
• Acceptance criteria
• Operating principles
• Process control
instruments and
software
Buffer Family Grouping Example
In assessing the stability of solutions and buffers to support commercial-scale bulk protein drug sub-
stance manufacturing, buffers and solutions of similar formulations and storage vessel types may be
assigned to family groupings. Following an analysis of concentrations, potential for interaction with
the vessel, susceptibility to contamination, and other appropriate factors, a “worst case” representa-
tive of each family grouping may be selected. The representative buffer is subjected to all testing
described in the protocol. Qualification of the family representative qualifies all other buffers or solu-
tions in the family grouping. This should be outlined in the rationale and in the process validation
master plan and protocols.

4.3.2.8 Process Analytical Technology

After developing a control strategy that incorporates PAT (Section 3.9, Section 6.3), process qualifi-
cation is performed to confirm that the monitoring, measurement, and process control or adjustment
systems are suitable, capable, accurate, and reliable. The key to effective PAT process control is the
reliable operation of instruments and equipment.

The use of PAT controls can provide an alternate approach to PPQ. If a PAT system is used to control
every commercial batch, then the PPQ stage will have a different focus. For example, if a powder
blending or solution mixing operation is controlled by a PAT system, such as NIR (near infrared) as-
says, the PPQ will involve demonstrating the control model and system and the process model works
as predicted in commercial manufacturing.

Qualification of the equipment, measurement system, and process must demonstrate the capability to
adjust CPPs according to the established algorithm and confirm that the adjustments result in accept-
able and predictable outputs. In other words, a PAT-based control method needs to be qualified (20).

38 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

4.3.2.9 Sampling Strategy
During the PPQ, increased sampling and analytical testing is expected to verify that the process is
under control, and to demonstrate consistency at intermediate steps, as well as in the final product.
Sampling plans for discrete units should include the statistical rationales that underlie the plans. (See
Section 6 and Appendix 8 for further information on statistically-based sampling plans.)

For processes or individual unit operations that yield a single homogenous pool of material, statisti-
cally based sampling plans may not be useful in ascertaining the level of intra-batch process variability.
For example, analysis of multiple samples from a homogeneous bulk solution or API provides infor-
mation on the variability of the analytical method only, not intra-batch variability of the process. In
these cases, extended characterization of intermediate pools and non-routine sampling performed at
certain points in the process and comparison of the data between batches can demonstrate process
control and reproducibility.

4.3.2.10 Setting PPQ Acceptance Criteria

The acceptance criteria for PPQ should be based on the body of data available from Stage 1, prior
knowledge, and equipment capabilities. The approach used to determine the acceptance criteria
should be outlined in the process validation master plan, and the justification of the individual ac-
ceptance criteria for each unit operation should be documented in the PPQ protocols. Statistical ap-
proaches should be used where appropriate, and each product and process variable should be evalu-
ated individually. Process justification documented in the Process Design Report (see Section 3.11)
provides the scientific basis and reference to the data supporting the acceptance criteria for process
parameter ranges, and product attributes. The rationale for PPQ acceptance criteria should be clearly
described. When sufficient data are available and statistical methods are used, the method(s) used and
the rationale for selection of that method should be described.

When establishing acceptance criteria for PPQ, the following considerations should be taken into account:
• Historical data / prior knowledge
• Preclinical, development, clinical, and pre-commercial batches
• Early analytical method suitability (if data is used from clinical lots)
• Amount of data available (level of process understanding)
• Sampling point in the process
• Compendial requirements can be met with high confidence

An overview of the factors considered for determining PPQ acceptance criteria should either be de-
scribed (or referenced, if included in a different document). Criteria for determining inter-and in-
tra-batch consistency should be defined. All parameters and attributes designated for tracking and
trending in Stage 3 Continued Process Verification should be included in PPQ acceptance criteria.
Acceptance criteria may include:
Incoming material — Meets designated criteria (may be raw material or the output of a preceding step).

Process Parameters — All process parameters are expected to remain within Normal Operating
Ranges; particular attention is focused on parameters with Critical or Key designations.
• Critical Process Parameters (CPP) with the potential to impact critical quality attributes
• Key Process Parameters (KPP) with the potential to impact process performance.

Attributes — All product quality and process performance attributes should meet pre-defined accep-
tance criteria and include statistical criteria where appropriate.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 39

 • Process performance attributes: may be impacted by KPPs (e.g., step yield or bioreactor titer) and demonstrate
process consistency between batches.
• Critical Quality Attributes: have the potential to impact safety or efficacy (e.g., impurities).
• Quality attributes: do not necessarily impact safety or efficacy, but can be used as a surrogate at certain process
steps to demonstrate process consistency (e.g., deamidation or oxidation that does not impact potency or
safety/immunogenicity)

4.4 PPQ Protocol

PPQ protocols are a documented plan for executing the PPQ studies. Protocols are reviewed and ap-
proved by cross-functional groups that include the quality unit. Protocols must be approved prior to
commencement of PPQ activities. PPQ protocols typically contain the following sections.

Introduction
The introduction should include a description of the process and/or specific unit operations under
qualification, including the intended purpose of the operations in the context of the overall manufac-
turing process. The introduction should provide an overview of the study(ies), and important back-
ground information.

Purpose and Scope

Describes the objective of the study and provides an overview of the study strategy, i.e., how it will be
performed, how data will be analyzed, and the expected outcome. Justifications or cross-referencing to
documents that contain justifications, such as the process validation master plan, should be included.

References
References to relevant documents related to the study should be included in the protocol:
• Development and/or Process Characterization Reports that provide supporting data for Operational Parameter
and Attribute ranges
• Process Design Report
• Process Validation Master Plan
• Commercial manufacturing batch records
• Related qualification documents (facilities, utilities, equipment, other PPQ studies)
• Analytical methods
• Specification documents
• Approved batch records

Equipment and Materials

A list of equipment, instrumentation, and materials necessary to perform the study should be in-
cluded. References to qualification of utilities and equipment should be provided as appropriate.

Responsibilities
A designation of various functional groups and their responsibilities as they relate to execution of the
study, and verification that appropriate training has been conducted for all contributors.

Description of Unit Operation/Process

The objective of PPQ is to provide confidence that all elements of unit operation/process are under
the appropriate degree of control. A comprehensive discussion of the control strategy similar to the
level of detail provided in the commercial manufacturing control strategy is appropriate to demon-
strate that all process elements have been considered. Although all elements are described, only a
subset of the process variables will comprise PPQ acceptance criteria. (See Acceptance Criteria.)

40 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Methodology
The step-by-step procedure needed to perform the study. This section clearly identifies the critical
and key process parameters under qualification and the methods by which the operation will be
monitored and recorded. A brief explanation of the relevance of these parameters and their potential
relationship to process performance and quality attributes is useful to further describe the PPQ strat-
egy. Documents containing the detailed rationale for critical and key parameter designations should
be referenced.

A discussion of the number of batches planned should be included, and the rationale should be stated.
The level of confidence expected at the conclusion of the PPQ study should be included as applicable.

Data Collection
Roles and responsibilities for various functional groups as they relate to collection and analysis of
PPQ data and documentation should be included. The list of process data to be collected and how it
will be analyzed should be stated.

Sampling Plan
A description of a defined prospective sampling plan and its Operating Characteristic Curve with de-
tails on the number of samples, frequency of sampling, and sampling points supported by statistical
justification, as applicable:
• Sampling points
• Number of samples and statistical basis for sampling, as appropriate
• Sample volume
• Non-routine sampling for extended characterization
• Sample storage requirements
• Analytical testing for each sample

See Section 6 and Appendix 8 for further information on statistically-based sampling plans.

Analytical Testing
The overall validation package includes the methods used for all analytical testing performed, from
assessment of raw materials to extended characterization of the drug product. A listing of all analyti-
cal methods used in each protocol and the validation or qualification status of each (and references to
source documents) should be included. Analytical method validation should also be included as part
of the process validation master plan.

Deviations
All potential deviations cannot be anticipated regardless of the level of characterization and knowl-
edge. A general framework for defining the boundaries of qualification is appropriate, for example:
• Out-of-specification or out-of-limits test results.
• Failure of a CPP to remain within normal operating range; a CPP is designated as such due to the potential
impact on a corresponding CQA. Failure to control may indicate overconfidence in an immature control
strategy. This would be grounds for protocol failure.
• Missed samples or samples held under incorrect storage conditions
• How individual batches or lots failing to meet validation acceptance criteria will impact the study.

Acceptance Criteria for PPQ

The objective of PPQ is to demonstrate that the commercial manufacturing process is in a state of
control, and the elements of the process control strategy provide confidence that a state of control

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 41

 will be maintained. The expectation for PPQ is that all process variables will remain within their des-
ignated ranges or meet acceptance criteria; subsets of these are used to define the PPQ acceptance
criteria. The protocol should clearly document the acceptance criteria to be met in order for the PPQ
to be considered successful. Acceptance criteria may be shown in tabular format in the protocol (see
the following example).
Table 4.4-1 Example of PPQ Acceptance Criteria Table
Process Parameter Designation Normal Operating Range
Parameter 1 CPP (x.xx - x.xx)
Parameter 2 CPP (x.xx - x.xx)
Parameter 3 KPP (x.xx - x.xx)
Parameter 4 KPP (x.xx - x.xx)
… … …
Attributes Acceptance Criteria
Recovery Process Performance (x.xx - x.xx)
Quality Attribute 1 Quality Attribute (x.xx - x.xx)
Quality Attribute 2 Quality Attribute (x.xx - x.xx)
Critical Quality Attribute 1 Critical Quality Attribute (x.xx - x.xx)
Critical Quality Attribute 2 Critical Quality Attribute (x.xx - x.xx)
Critical Quality Attribute 3 Critical Quality Attribute (x.xx - x.xx)
Critical Quality Attribute 4 Critical Quality Attribute (x.xx - x.xx)
… … …

4.5 PPQ Report

A report should be prepared for each study and will typically include the following sections:

Introduction
The introduction should include a concise description and outline of the unit operations or group of
unit operations that have been qualified. It should summarize the overall results of the study, provid-
ing background information and explanations as necessary.

Methods and Materials

A clear and concise summary of how the study was performed. It should identify how the objectives
of the study were accomplished using both methodology and references to appropriate procedures
and protocol requirements.

Deviations
A summary of the deviations and corresponding root causes, as well as a discussion of the potential
impact to the PPQ, should be included. Corrective actions resulting from deviations should be dis-
cussed. Their impact on the process, the PPQ, and on the affected batches should be provided.

Protocol Excursions
Protocol excursions and unexpected results should be included and fully described in the report. A ref-
erence to the root cause analysis should be provided if documented separately from the PPQ report.
Any corrective actions and their impact on PPQ should be outlined in the report.

Discussion: PPQ Results

This section should restate the key and critical process parameters and give the actual range of values

42 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

occurring during the PPQ. It should include how the data were collected as well as references for
analytical methods used.

Data summarized and compared with pre-defined acceptance criteria should be presented in tabular
or graphical format whenever possible, and data used from Stage 1 studies should be clearly identified.
A reference to the original study should be provided when data is used from outside the of PPQ is
used to augment the PPQ data set for statistical manipulation or other support The level of statistical
confidence achieved should be stated. If the desired level of statistical confidence was not achieved,
the reasons for this and follow up actions should be discussed.

The discussion should provide support for any study conclusions. The impact of ranges and devia-
tions should be discussed if they affect the study results. Risk assessment and any follow-up conclu-
sions, including corrective actions, should be stated.

Findings associated with batches or lots that fail to meet the acceptance criteria in the protocol should
be referenced in the final PPQ package; likewise, with any corrective measures taken in response to
the cause of failure
Conclusions
Conclusions as to whether data demonstrate that the process is in a state of control should be pro-
vided. Pass or fail results should be stated for each acceptance criteria and corresponding results.

When a unit operation approach is used, PPQ reports prepared for each unit operation study. A sum-
mary executive report that unifies all the study results to support the overall process PPQ should be
written.

4.6 Transition to Continued Process Verification

Following a successful PPQ, the CPV plan can be finalized and implemented. Any adjustments to
be made on the basis of the PPQ should be in place prior to manufacture of post-PPQ batches and
should be handled through the change control procedures. When appropriate, enhanced PPQ-level
sampling is recommended for a period of time following PPQ. However, this may not be necessary in
all cases. Further information is presented in Section 5.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 43

 5.0 Continued Process Verification (Stage 3)

5.1 Establishing a Monitoring Program

5.1.1 Purpose and Strategy
A program of Continued Process Verification (CPV) provides a means to ensure that processes re-
main in a state of control following the successful Process Qualification stage. The information and
data collected during Stages 1 and 2 set the stage for an effective control strategy in routine manu-
facturing and a meaningful CPV program. The understanding of functional relationships between
process inputs and corresponding outputs established in earlier stages is fundamental to the success
of the CPV program.

Continued monitoring of process variables enables adjustments to inputs covered in the scope of a
CPV plan. It compensates for process variability, to ensuring that outputs remain consistent. Since all
sources of potential variability may not be anticipated and defined in Stages 1 and 2, unanticipated
events or trends identified from continued process monitoring may indicate process control issues
and/or highlight opportunities for process improvement. Science and risk-based tools help achieve
high levels of process understanding during the development phase, and subsequent knowledge man-
agement across the product life stages, facilitates implementing continuous monitoring (see Sections
3.0 and 4.0).

5.1.2 Documenting the CPV Program

Planning for CPV begins during the establishment of the commercial-scale control strategy (Stage 1).

High-level quality system policies/documents outline how various departments interact and how
information is compiled and reviewed to ensure maintenance of the validated state. Under that policy
document as well as a process validation master plan, a product-specific CPV plan should include the
following elements:
• Roles and responsibilities of various functional groups
• Sampling and testing strategy
• Data analysis methods (e.g., Statistical Process Control methods)
• Acceptance criteria (where appropriate)
• Strategy for handling Out of Trend (OOT) and Out of Specification (OOS) results
• Mechanism for determining what process changes/trends require going back to Stage 1 and/or Stage 2
• Timing for reevaluation of the CPV testing plan

Figure 5.1.2-1 illustrates an example of the development of a CPV monitoring strategy throughout
the lifecycle stages. Ideally, the majority of the control strategy is established prior to Stage 2, when
PPQ is conducted. When adopting the concept of Continued Process Verification for legacy products,
the same general approach should be taken to document and execute the CPV program (see Section
5.1.3, Legacy Products and Continued Process Verification).

Because Stage 3 is part of the lifecycle validation approach (5.1.2-2), Continued Process Verification
should be governed by both an overarching quality system for validation practices and a process vali-
dation master plan. At a minimum, the process validation master plan should make high-level com-
mitments for both Process Design (Stage 1) and Continued Process Verification (Stage 3) in addition
to Process Qualification (Stage 2). The specifics of the CPV sampling/testing strategy may not be
finalized until completion of PPQ. Therefore, the process validation master plan may include general
commitments to the planned CPV strategy. These are then further clarified in a separate CPV Plan
referenced in the process validation master plan. It is still possible that a process validation master plan

44 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

can be considered complete at the end of Stage 2 (i.e., not left open-ended for the entire commercial
lifecycle) if the requirement that CPV activities, as required, are be initiated per the defined CPV Plan.
Figure 5.1.2-1 Development of a Continued Process Verification Plan
Stage 1

• Statistical methods
• Data to be trended/rationale
• Establish confidence in process based on
Draft Initial Plan small-scale models
• Frequency of reporting
Stage 2

• Revise commitment to # of batches

Make Adjustments to Plan
under CPV prior to reassessing accep-
Based on PPQ Learnings tance criteria
• Update statistical strategy based on
knowledge/confidence gained from PPQ
• Update frequency of data review based
on relevant statistical tools
Stage 3

Formalize Plan Prior to Start of

Post-PPQ Manufacturing

Periodic review to
assess state of control

Figure 5.1.2-2 CPV Plan within Validation Documentation System

Quality System (Governing Validation Practice)

Process Validation Master Plan

Process Equipment Cleaning

Facilities
Characterization Validation Validation PPQ Protocol(s) CPV Plan
Validation Protocol
Protocols Protocols Protocols

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 45

 5.1.3 Legacy Products and Continued Process Verification
Figure 5.1.3-1 outlines one possible approach to assessing what is necessary to apply the lifecycle ap-
proach to a legacy product. It may be the case that a legacy process is well-controlled and monitored,
and not much action is required. However, this decision should be based on an evaluation of the large
body of historical process and monitoring data and an assessment of process variability. In this ap-
proach, the historical data is used to determine the current state of control of the process. Measures
such as performance capability (Ppk) and other statistical approaches should be considered (see Sec-
tion 6.0) for assessment of the process. In addition to assessing process performance, the adequacy of
the set of parameters being used to monitor the performance of the process should also be evaluated.
Part of assessing the appropriateness of the current process control strategy is to provide a founda-
tion for determining what, if any, additional sampling/monitoring should be included during Con-
tinued Process Verification for the legacy product. A period of enhanced sampling will help generate
significant variability estimates that can provide the basis for establishing levels and frequency of
routine sampling and monitoring and should be considered. It is recommended that this ongoing
monitoring also be captured under a formal plan as outlined in Section 5.1.2, Documenting the CPV
Program.

In considering whether the sampling plans for legacy products are adequate, it may be determined
that a statistically-driven approach should be applied. However, the amount and type of data may also
lead to a decision that statistical justification of the sampling plan is unnecessary. This determination
should be part of the initial assessment of the historical data and monitoring approach. Although
statistically-derived models may not be required, the sampling plan should be scientifically sound and
representative of the process and each batch sampled.

46 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Figure 5.1.3-1 CPV Plan Determination for Legacy Products

Perform comprehensive review of Process Control Strategy

(PCS), historical production data trends, and events-based
data/information (deviations, complaints, etc)

Is the legacy Continue to monitor/trend on a

manufacturing process YES routine basis. Ensure events-
well controlled? * based data are integated.

NO

Use process knowledge, risk assessment, and/or historical

data to identify sources of process variability and/or PCS
deficiencies.

Can process variability Implement change and continue

be reduced via minor process YES collecting CPV data to confirm
change and/or addition of that variability is reduced.
process controls?

NO

Can process variability be

reduced via significant process Implement changes Resume CPV for
YES
changes which are supported Re-perform PPQ. improved process.
by existing data?

NO

Perform Process Design work required

to support process changes necessary
for ensuring process control.

* Is an appropriate Process Control Strategy (demonstrating understanding of the impact of process parameters on
CQAs) defined and does statistical of data show that variability is controlled?

5.1.4 Demonstrating Continued Process Verification

Two primary sources of data that need to be included in a Continued Process Verification (CPV) plan are:
1. Process parameters (i.e., process performance and product quality indicators)
2. Potential sources of variability that are not defined process parameters. Examples of such sources of data/
information may include:
a. Raw material quality
b. Redundant equipment and instrumentation comparability
c. Personnel impact on process (i.e., shift-to-shift consistency)

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 47

 Critical and key input parameters and the corresponding outputs related to process performance and
product quality attributes are established during Process Design (Stage 1) (see Figure 1.1-2). At the
commercial scale, Process Qualification (Stage 2) batches are produced to confirm that the process
operates as intended and to verify that the Process Control Strategy results in the consistent manufac-
ture of product that meets its predefined quality characteristics. The Process Control Strategy should
then also be used as the starting point for identifying the process data/information to be included in
a CPV plan.

5.1.5 CPV Monitoring Plan

Routine sampling will generate some data for the CPV Program, but non-routine sampling should
also be considered. The sampling/testing plan moving forward from Stage 2 into Stage 3 should be
a dynamic; it needs to be updated and reviewed periodically. An enhanced sampling plan (that may
include both off-line and on-line analyses) may be required to ensure that the appropriate data set is
collected. Since the PPQ Protocols already specify those process parameters and attributes (inputs
and outputs) that must be maintained within the specified ranges in order to make a product that
meets predefined quality attributes, the PPQ sampling plan is a logical foundation for the CPV sam-
pling plan. PPQ may provide sufficient assurance that certain parameters are well-controlled at the
commercial scale and do not need to be carried forward into a CPV plan. A biological process, for ex-
ample, requires sufficient clearance of a process residual (e.g., antifoam) or a process-related impurity
(e.g., DNA). These may be successfully demonstrated during PPQ batches, eliminating the need for
ongoing sampling and testing during CPV. In cases, where either historical data are limited or where
the data show a high degree of variability, testing and trending may be required post-Stage 2 to ensure
a high level of assurance that a particular impurity is well-controlled. This should be determined on a
case-by-case basis via risk assessment and/or statistical assessment of historical data.

The prospective CPV plan should provide specific instructions for analysis conducted to a limited
degree, and subsequently discontinued once a sufficient number of data points are accumulated to
determine process control. The number of batches sampled and the frequency of sampling within a
batch should be stated in a Stage 3 enhanced sampling plan. Depending on the data generated, sam-
ples collected and analyzed for information only (FIO) should have a designated end-point. A more
open-ended approach, where no specific number of batches is identified, could be used to address
data trends and results. A plan that describes an approach to reduce (step-down) or increase (step-up)
sampling and testing as a result of trending and results is also an option.

5.1.6 Data Analysis and Trending

The CPV plan should clearly state how the data collected will be analyzed. In some cases, it will be
compared to pre-defined acceptance criteria, especially for those data that are tightly controlled (e.g.,
a gradient elution slope for a critical column chromatography step). In other cases (e.g., unit opera-
tion yields), data may be statistically assessed to evaluate process trends. In such cases, the statistical
methods and rules used for continued process monitoring should be specified in the CPV plan. Con-
trol charts are commonly used to evaluate process control over time. They are appropriate for both
evaluating statistical process control and for detecting process trends. Under CPV, control charts are
generated and evaluated on a per batch basis (see Section 6.2, Statistical Analysis Tools and Appendix
8.1, Statistical Methods for Determining the Number of Lots for discussion of statistical data analysis).

Establish prospective criteria to ensure that the process is in a state of control. However companies
define it, an “out of control” result (e.g., Out-of-Trend, Out-of-Control, Out-of-Specification, outside
Action Limit) should trigger actions per the Quality System (e.g., investigation, impact assessment
to validated state). Specific actions will vary on a case-by-case basis, but the CPV plan should specify

48 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

what types of action. Section 6.0, Tools for the Process Validation Lifecycle, describes the tools avail-
able to address trending statistical trending and SPC, along with risk-based evaluations.

Section 5.1.4 covers, sources of process variability that may not be parameter-related (e.g., raw mate-
rials, personnel, and environment). As part of the overall CPV assessment, high-risk potential sourc-
es of variability should be risk-mitigated, and also assessed and demonstrated to be under control.
Trends in purity for a critical raw material, for example, may indicate subtle differences between
suppliers. Even seemingly innocuous changes by a supplier may lead to out-of-trend or out-of-speci-
fication events. These should be evaluated in light of overall process consistency and product quality.

5.2 Incorporation of Feedback from CPV Monitoring

5.2.1 Quality Systems and CPV
The best tools for continued confirmation and refinement of process control are the quality system
elements that provide feedback and objective measures of process control. The tools are based on
product and process understanding, and are enabled by procedures that monitor, measure, analyze,
and control the process performance (37). Once in commercial production, maintenance of the vali-
dated state requires an events-based system of review, in addition to process trending described in
Section 5.1, Establishing a Monitoring Program. Communication of review outcomes to the manu-
facturing, quality, and regulatory stakeholders to modify the control strategy (for improvement and/
or compliance reasons) is an iterative and essential part of the CPV. Feedback mechanisms can vary
between immediate (intra-batch or real-time), after each batch, or after a series of batches or a defined
time period. The CPV Plan should address when each of these mechanisms should be used.

Figure 5.2.1-1 depicts sources of data that contribute to continuous improvement of a manufactur-
ing process. While not intended to be an all-inclusive list, the figure shows typical categories of data
associated with product production and performance.
Figure 5.2.1-1 Body of Knowledge and Maintenance of Process Control

Process Trend
Analysis

Process Change
Periodic Review
Impact

Process Process Control Historical

Characterization
Strategy Production Data
Data

Extended
Deviation Review
Characterization
CAPA
Data

Product Complaints

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 49

 5.3 CPV Data Review and Reporting
The CPV plan needs to include a frequency of review of the information from data collection mecha-
nisms as well as Quality Systems. It should also identify circumstances for, and a process to allow for,
an immediate review based on significant issues identified with a process or product, and identify
the participants in the review. Per ICH Q10, this review must include senior management. They are
key stakeholders in the maintenance of an effective pharmaceutical quality system and advocates for
continual improvement.

The frequency of data review will depend heavily on risk. The period of review for various processes and
sub-processes is likely to vary greatly depending upon the levels of associated risk and the complexity of
control. The starting point for defining the review period will be the most recent process risk commu-
nication document. As more production data is generated, deeper process understanding is gained and
control is likely to be more easily demonstrated. Thus, the period or intensity of review may be reduced.

An annual commercial data compilation effort in preparation for Annual Product Review (APR) may
be sufficient. However, more frequent data reviews and comparisons to defined acceptance criteria
may help manufacturers be more proactive and less reactive. APR packages are necessary, as per regu-
latory guidelines. However, APR exercises are likely to become high-level reviews and summaries of
multiple, more frequent CPV data reviews. The APR will identify any gaps in the CPV data reviews
and will summarize long-term trends, but more frequent CPV data reviews should be performed by
the manufacturer at defined intervals.

Note: FDA 21 CFR 211.180(e) requires an evaluation at least annually. The periodicity of the review is
to be established by the manufacturer, but should be at least annually.

50 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 6.0 Process Validation Enabling Systems and Technology

This section presents tools and methods to assist in the planning and performance of the process validation
program. It includes sections on risk and knowledge management, statistical methodology, process analyti-
cal technology, and technology transfer. These tools can be used to identify, capture, and communicate
information needed for the design and assurance of process control. They facilitate informed decision-
making, prioritization of activities, and interpretation of results related to the process validation effort.

6.1 Application of Risk Management

This section addresses aspects of risk management specific to the process validation lifecycle ap-
proach. A detailed explanation of a Quality Risk Management program used to support the process
validation effort can be found in Section 5 of PDA Technical Report No. 54: How to Use Quality Risk Man-
agement as an Enabler (13). In addition, comprehensive lists of risk management tools can be found in
PDA Technical Report No. 54 and ICH Q9. For a comparison of risk management tools, see Technical
Report No. 54, Table 4.2-1: Comparison of Common Risk Management Tools.

The Quality Risk Management system is an “enabler” or “enabling system.” When correctly applied,
it adds supportive elements to the product lifecycle and other systems (e.g., the Pharmaceutical Qual-
ity System). The application of risk management principles and approaches is instrumental to effec-
tive decision-making in the Process Validation Lifecycle.

Management of variability is one example of applying risk management in the validation lifecycle.
The level of control required to manage variability is directly related to the level of risk that variability
imparts to the process and the product. The use of risk management to address variability requires
understanding of:
• The origin of the variability
• The potential range of the variability
• The impact of the variability on the process, product, and ultimately, the patient

Risk assessment should occur early in the lifecycle, be controlled appropriately, and effectively com-
municated. Risk Management increases product and process knowledge, which translates into greater
control of product and process variability, and a lower residual risk to patients.

The process validation lifecycle provides continued assurance that processes will manufacture prod-
uct in a predictable and consistent manner. Where decisions related to product quality or process per-
formance are made, risk can be assessed at several points throughout the process validation lifecycle.

Quality Risk Management applications throughout the process validation lifecycle include the follow-
ing (see Figure 6.1.1):
Stage 1 — Process Design
• Identification of product attributes that may affect quality and patient safety
• Criticality analysis of product quality attributes (CQA identification)
• Cause and Effect Analysis or Risk Ranking and Filtering, which link the process steps and parameters to process
performance or product quality attributes. These can be used to screen potential variables for future process
characterization (e.g., DoE) and testing.
• Preliminary Hazards Analysis (PHA) or early FMEA

Stage 1-2 — Transition from Process Design to Process Qualification

• Determining process control strategies that address the risk of failure for each process step
• Evaluation of residual risk remaining or created as a result of risk mitigation, process improvement, and process
knowledge

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 51

 Stage 2 — Process Qualification
• Determination of process steps and parameters to test in PPQ, including sampling plans and the confidence and
coverage they provide.
• Facility and equipment impact assessments to prioritize qualification efforts [may want to reference ISPE
Baseline Guide (34)]
• Determination of effective acceptance criteria for each test function
• Analytical test results and deviations

Stage 3 — Continued Process Verification

• Determination of parameters that should be monitored as well as how they should be sampled and analyzed
(e.g., sampling plans, confidence required and length of enhanced sampling).
• Evaluation of commercial manufacturing data to determine the best course for process improvement

Figure 6.1-1 depicts a quality risk management lifecycle tool for process development and validation (21).
Figure 6.1-1 Quality Risk Management: A Lifecycle Tool for Process Development and Validation

Process Validation Sequence

Stage 1 Stage 2 Stage 3

Based on product quality Is the process Are the variables When is confidence What is looked for
and patient safety known? known? achieved? and for how long?

1. Process 3. Process 4. Commercial 5. Monitoring

2. Process Design
Understanding Qualification Manufacture and Improvement

Risk Assessment
CQAs Parameters Support System Qualification Continuous Process Verification
Requirements Variables Commissioning Monitoring
Control Strategy IQ, OQ Reaction to Issues
DOE PQ Process Improvements
Statistical Sampling Plans

6.1.1 Risk Management in Stage 1 -Process Design

Conducting risk assessments during Stage 1 Process Design lays the groundwork for variables to be con-
trolled and monitored. It also determines the extent to which continued monitoring will ensure a state of
control during routine manufacturing. This begins with a criticality analysis: an initial definition of Prod-
uct Quality Attributes and an assessment of their relative importance. Inputs for the criticality analysis are:
• QTPP (quality target product profile)
• All relevant prior knowledge for the product being evaluated
Outputs from the criticality analysis are:
• Initial CQA list
• Initial relative severity listing of the CQAs

Criticality of product attributes is assessed along a continuum; i.e., it not a yes or no question. This is
accomplished by performing a risk assessment analysis that uses Severity and Uncertainty, rather than
the usual Severity and Occurrence. The process, which is iterative, is based on building product and

52 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

process knowledge. The level of severity assigned is based on the potential patient impact, while un-
certainty is based on how much information (product knowledge and clinical experience) is available
to determine the potential severity level for the specific attribute. Part of the output of this assessment
will be further scientific studies to reduce the amount of uncertainty for higher risk attributes (21).
(See Figure 6.1-2, Product Attribute Criticality Risk Assessment Example.)
Figure 6.1-2 Product Attribute Criticality Risk Assessment Example

Uncertainty

Low Medium High

(Large amount of in- (Some in-house (No/little in-house

house knowledge, large knowledge and knowledge, very limited
body of knowledge in scientific literature) information in scientific
literature) literature)

High (catastrophic
Critical Critical Critical
patient impact)
Severity

Medium (moderate
Potential Potential Potential
patient impact)

Low (marginal patient

Non-Critical Non-Critical Potential
impact)

6.1.2 Risk Management in Stage 2 -Process Qualification

Risk Management in Stage 2, the process qualification stage of the process validation lifecycle, is
much more tactical. Assessments assist in deciding where tests will be performed and at what level.
They are also used to fine-tune the control strategies drafted in the Process Design stage.

Risk management is commonly applied during the Facilities, Utilities, and Equipment Qualification phase
of Stage 2. Functional specifications are reviewed to help plan qualification activities. Higher-risk items
require a higher level of performance output, while lower-risk items can be satisfied by use of commis-
sioning activities with appropriate risk reviews and control. Risk assessment output ratings can be applied
against standard criteria to create the plan (see Table 6.1.1).
Table 6.1-1 Risk-Based Qualification Planning

Risk Assessment
Qualification Planning
Output Ratings

Testing to satisfy validation requirements will occur during qualification. Documentation

High
and sampling requirements are high.

A blend of Qualification and Commissioning activities can be used to satisfy validation

Medium requirements. Sampling requirements are moderate given appropriate controls and risk
reviews.

Testing to satisfy validation requirements can occur during commissioning phases.

Low
Appropriate controls and risk reviews should be in place.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 53

 Risk assessments performed during Stage 2 not only help prioritize qualification activities, but also
aid in the ongoing collection of knowledge and the planning of statistical sampling. Generally, three
factors - Severity, Occurrence, and Detection (also known as controls) - are evaluated to determine the
relative risk of specific failure modes. Each factor contributes to the validation plan in a different way.

Severity — Determines the level of testing required during Stage 2. The higher the severity rating for
a particular attribute, the higher the statistical confidence required (see Table 6.1-2).

Occurrence — The occurrence rating is tied directly to variation. High Occurrence rates may require
further testing or development to reduce variation and increase process knowledge. Testing at this
stage reduces additional and more costly testing during Stage 3. When the true occurrence rate is
unknown, additional development or engineering studies may be required. When testing is complete,
the occurrence ranking and overall risk rating for the failure mode can be updated with the new pro-
cess knowledge.

Detection (controls) — If the level of assessed controls is zero, the control strategy may need to be
updated or new controls created. Controls do not have to be technology-based. The HACCP system
is an example of a control, as are procedures and training.
Table 6.1.2 Severity Rating and Sampling Requirements

Risk Severity Rating Statistical and Sampling Example Confidence Level

Requirements Required

High +++ 99%

Med ++ 95%

Low + 90%

6.1.3 Risk Management in Stage 3 -Continued Process Verification

The Continued Process Verification stage is the longest segment of the process validation lifecycle. It
starts with an assessment of process capabilities and continues through a review of the output from
process characterization, PPQ, and historical data. The level of enhanced sampling that may be in
place when commercial manufacturing commences can be determined by a statistical review of the
PPQ data. The capabilities of the processes help determine the level of enhanced sampling for an
attribute and the length of time that sampling should continue at that level (see Section 6.2). The
statistical capability of the process is directly tied to the occurrence rating in the risk assessments. The
more robust a process, the lower the occurrence rate for a potential failure and the lower the overall
risk to the process. The level of risk can also determine the review period for certain product and
process attributes (14),

6.1.4 Raw Material Risk Management Considerations

Sources of variation should be understood, and where possible, mitigated for process validation to
succeed. In this context, using quality risk management to assess raw material quality and the po-
tential impact on the process is important (38). Risk identification through focused risk assessments
is the first step toward attaining the desired level of process control from both a risk-to-patient and
risk-to-business perspective. The assessment identifies risk in relation to the raw material, and how it
could impact the process and quality of product. The number and complexity of raw materials used

54 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

in pharmaceutical manufacturing is quite large, and all potential issues (e.g., fraud/counterfeiting)
should be addressed in the management of raw materials and components.

Risks-to-patient should also be addressed during commercial production. This can be done, through
a risk assessment process that builds on current understanding of risk and process knowledge, com-
bined with the Continuous Process Verification Program. QRM is a lifecycle process, with assess-
ments that occur throughout the lifecycle of the product.

Often subtle changes in raw materials can lead to significant and unforeseen variations in production.
The cause of a change in elution profile was lot-to-lot variation in particle size distribution in a chro-
matographic resin (39). Applications like Near Infrared (NIR) or even Nuclear Magnetic Resonance
(NMR) can be used to ensure that raw materials meet their specifications and CQAs. An important
risk mitigation strategy is for drug manufacturers to work with their suppliers so that each can under-
stand the other’s quality systems and demands.

6.2 Statistical Analysis Tools

Successful process validation depends on sound, scientific data and information. Table 6.2-1 illus-
trates where various statistical methods are most commonly used in the validation lifecycle process.
Three of the methods - Design of Experiments, Statistical Process Control, and Process Capability -
are described in more detail in the sections that follow. Additional information on statistical methods
can be found in PDA Technical Report 59: Utilization of Statistical Methods for Production and Business
Processes as well as Appendix 8.1 of this technical report (14).

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 55

 Table 6.2-1 Statistical Methods and the Typical Stages at Which They Are Used
Stage 1 Stage 2 Stage 3
Statistical Tool Process PQ CPV
Design
Descriptive Statistics – Mean, standard
X X X
deviation, etc.

Statistical Process Control Charts X X X

Statistical Power and Sample Size

X X X
Determination

Process Capability Study and Capability

X X X
Indices

Design of Experiments X

Measurement Systems Analysis (Gauge R&R) X

Robust Process Design / Tolerance Analysis /

X
Taguchi Methods

Multi-Vari Chart X

Regression and Correlation Analysis X

Analysis of Variance (ANOVA) X X X

Levene/Brown-Forsyth, Bartlett, Fmax Tests for

X X X
Variation

Hypothesis Tests / Confidence Intervals X X X

Pareto Analysis X X

Acceptance Sampling Plans X X

Normal and Nonparametric Tolerance Intervals X X

56 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

6.2.1 Design of Experiments (DoE)
The statistical design of experiments (DoE) is a powerful tool often used during Stage 1 Process De-
sign. Goals of DoE are to:
• Determine which process input parameters have a significant effect on the output quality attributes
• Help determine the “design space” levels of the input parameters that will produce acceptable output quality
attribute results
• Optimize the output of quality attributes, such as yield and acceptable levels of impurities
• Determine the levels of input parameters that will result in a robust process that reduces its sensitivity to
parameter variability

DoE differs from the classical approach to experimentation, where only one parameter is varied while
all others are held constant. This “one-factor-at-a-time” type of experimentation cannot determine
process parameter interactions, where the effect of one parameter on a quality attribute differs de-
pending on the level of the other parameters. The basic steps for the DoE approach are summarized
below:
1. Determine the input parameters and output quality attributes to study.
a. This is best done as part of a team approach to identify potential critical process parameters and
quality attributes; in many cases, the process may be well-understood and the parameters and
attributes for experimentation readily determined.
b. If there are a large number of input parameters, an initial screening design, such as a fractional
factorial or Plackett-Burman design, may be used (40). The purpose of a screening experiment
is to identify the critical parameters that have the most important statistical effect on the quality
attributes. Since screening designs do not always clearly identify interactions, the reduced number
of parameters identified by the screening experiment will be included in further experiments.
c. If the change is to an existing process, it is often valuable to construct a Multi-Vari chart
or SPC chart from current process data (41). A Multi-Vari chart can be used to identify if
the biggest sources of variation are within-batch variation, between-batch variation, or
positional variation (e.g., between fill heads on a multi-head filler). Variance components can
also be calculated from the data to determine the largest component of variance. Process
parameters that could be causing the largest sources of variation are then identified and
included in subsequent experiments.
For example, if within-batch variation appears to be the largest source of variation, then
charge-in of components done once at the beginning of the batch is not likely to be a key
contributor to this variation. Charge-in differences due to inadequate weighing, for example,
could cause between-batch variation rather than, within-batch variation. This simple but
powerful tool can sometimes discover important yet unsuspected critical parameters or
“lurking variables” that contribute to process variation, even if they are not initially on the
list of parameters.
The same data may also be used to create SPC charts to determine if the process is in statistical
control. Since a lack of statistical control will contribute to experimental error variation, it
will be more difficult to understand the results of an experiment if the process is not in
statistical control. Lack of statistical control may also mean that there are “lurking variables”
not on the list of process parameters that are contributing to process variation.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 57

 2. Conduct experiment(s) to determine which parameters have a significant main or interaction effect
on the quality attributes.
a. This will usually be a full factorial design for two to four parameters. A full 2-level factorial
design has a low (–) and high (+) level selected for each factor (parameter). At least one
experiment is run at each combination of the factor levels. For two factors, 22 = 4 combinations
exist; for three factors, 23 = 8 combinations exist; for four factors, 24 = 16 combinations exist.
Full factorial designs are seldom used for more than four factors since so many experiments
are required. Fractional factorial experiments, where only one-half or one-quarter of the
combinations are used, are often done for four to six parameters.
b. If possible, control runs at the nominal midpoints (0) between the low (–) and high (+)
levels of the factors should be included in the experimental design. Using control runs at
the beginning and the end of the factorial experiment, and ideally also during the factorial
experiment, will allow detection of any process drift during the experiments. Control runs at
the beginning and end of experiments that do not give similar results indicate the presence
of another uncontrolled variable. Replicate control runs at the nominal values also provide a
true estimate of inherent process variation (called experimental error). In addition, these can
serve as a basic check for a non-linear curvature effect between the parameters and quality
attributes.
c. If possible, the parameter effects on both the mean and variation of the quality attributes
should be determined. Some parameters may affect the mean only, variation only, or both.
This information can be used to minimize the variation while optimizing the mean, which
results in a robust process. Standard DoE approaches may be used for this as well as the
Taguchi method (42).
3. Optimize with response surface experiments and determine design space.
a. Occasionally, the science behind a process will be understood well enough to skip screening and
2-level factorial experiments and start with response surface experiments. If enough information
is learned from 2-level factorial studies, no additional experiments will be required and this step
can be skipped. However, it is often necessary to conduct more extensive experiments at three to
five levels for the parameters identified as most important from earlier factorial experiments.
The goal of response surface experiments is to develop an equation that accurately models the
relationship between the input parameters and output quality attributes. This equation is then
used to determine the design space region of the input parameters where the output quality
attributes will meet specifications
The most common response surface experimental designs are Box-Behnken, central composite,
3-level full factorial, and computer-generated D- and G-optimal designs (40). All of these have
experiments where at least three levels of the parameters are included in order to estimate
curvature (quadratic) effects. The Results are analyzed to determine regression equations to
model the process with such computer programs as Minitab, JMP, and SAS (41).
b. Another aspect of optimization is to develop a robust process. The regression equations already
developed can be used to locate input parameter settings that are “forgiving;” i.e., when the
process is run at these settings, variation in the input parameters will not result in unacceptable
variation in the quality attributes. The idea is to stay away from boundaries or areas in the
parameter design space where variation in the parameter will result in rapid quality deterioration.
This is accomplished by using the quadratic and interaction effects to minimize variation. The
Taguchi method of experimental design mentioned earlier uses a slightly different approach to
also develop robust processes.

58 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 c. The results may also be used to calculate the percent of total variation attributable to each
parameter. This is called a variance components analysis. The input parameters contributing the
most to the output quality attribute variation can be controlled the most tightly, made robust by
running the process at a particular level of the other parameters, or improved by a process design
change to reduce the impact of the parameter.
4. Confirm DoE results
Once the design space region for the input parameters that results in quality attributes that meeting
specifications is been determined, additional experiments can be used to confirm the expected DoE
results. This may consist of running a few experiments at various parameter combinations to verify
that the DoE equation adequately predicts the results. In some cases, where there is already good
confidence in the DoE results, Stage 2 PPQ results may be used. For further information on DoE, see
Montgomery (43) or Box, Hunter, and Hunter (44).

6.2.2 Statistical Process Control and Process Capability

Statistical Process Control (SPC) may be used to determine if a process is stable, predictable, and in
statistical control. Process Capability is used to determine if the process is capable of consistently
meeting specifications. A process is considered stable or “in statistical control” when only random
variation around a stable process mean is observed, i.e., only natural, common causes of variation are
present. Figure 6.2.2-1 illustrates a stable process that is in classical statistical control. Figure 6.2.2-2
shows a process that is not in statistical control and had a special cause of variation occur at lot 5.
Figure 6.2.2-1 Process in Classical Statistical Control; Common Cause Variation only

105
104
103
102
Total Variation
Within lot

101
100
99
98
97
96
95
Lot 1 Lot 2 Lot 3 Lot 4 Lot 5 Total Process Over Time

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 59

 Figure 6.2.2-2 Process Not in Statistical Control -Special Cause Variation

105 USL
104
103
102
101
Within lot

100
99
98
97
96
95 LSL
Lot 1 Lot 2 Lot 3 Lot 4 Lot 5 Total Process Over Time

A more complex form of a process that is also stable and in control is shown in Figure 6.2.2-3. This
pattern is typical of many processes where there is variation both within and between lots, but the
variation between lots is in control. One purpose of validation and CPV is to determine both within-
and between-lot variations.
Figure 6.2.2-3 A Process with Both Within-lot and Between-lot Variation
105
104
103
102
Total Variation
Within lot

Between lot

101
100
99
98
97
96
95
Lot 1 Lot 2 Lot 3 Lot 4 Lot 5 Total Process Over Time

6.2.2.1 Statistical Process Control Charts

Statistical process control charts are used to determine if a process is stable and in statistical control,
or if there are special causes of variation present in the process. The basic procedure to construct a
Statistical Process Control (SPC) chart to assess process stability is:
• Collect data from the process over time. Ideally, at least 20 subgroups should be collected, but preliminary
limits may be made with less data and updated as more data become available (40). Other references, such
as ASTM E2587 (45), have more detailed recommendations for the amount of data to collect initially. Plot
the summary statistics from each subgroup over time, such as mean (Xbar), standard deviation (S), percent
nonconforming, or individuals.
• Draw centerlines at the grand average of the statistic being plotted.
• Calculate the standard error of the plotted statistics and draw control limits at three standard errors on either
side of the centerlines. These limits are called “3-sigma” control limits.

60 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Values that fall outside the control limits indicate that special cause variation is likely present, and the
causes for these excursions should be investigated. In addition to a single value beyond the 3-sigma
limits, there are many other rules that may be used to check for process stability. Of these, the most
commonly used are (40,41):
• 8 in a row above or below the mean;
• 2 out of 3 beyond 2-sigma limits;
• 4 out of 5 beyond 1-sigma limits;
• 6 in a row increasing or decreasing.

Figure 6.2.2.1-1 shows an example of an Xbar/S chart for fill weight, where five vials from a single-
head filler were sampled every 15 minutes over a six hour production order or lot, for 24 samples.
Both the mean and standard deviation appear to be stable, with no values exceeding the 3-sigma con-
trol limits. The process appears to be stable and in a reasonable state of statistical control.
Figure 6.2.2.1-1 Xbar/S Control Chart for Fill Weight, n=5 per group

52.4 UCL=52.37
Sample Mean

52.2

52.0 X=52.04

51.8
LCL=51.70
1 3 5 7 9 11 13 15 17 19 21 23
Production Order
52.4 UCL=0.49
Sample StDev

52.2

52.0 S=0.24

51.8
LCL=0
1 3 5 7 9 11 13 15 17 19 21 23
Production Order
Control charts can be used during all three validation stages for within- or between-lot data. During
Stages 1 and 2, they can be used to determine if the process is stable and in control in order to com-
mence commercial production. Control charts are particularly useful during Stage 3 (CPV Stage).
Special causes of variation affect almost every process at some point. Control charts help identify
when such a special cause has occurred and when an investigation may be needed. As special causes
are identified and corrective actions taken, process variability is reduced and quality improved. Con-
trol charts are easy to construct and can be used by operators for ongoing process control. They also
create a common language for discussing process performance, and can prevent unnecessary adjust-
ments and investigations. They encourage staff to be responsible for monitoring and improving their
process, rather than just taking action when QC test results fail.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 61

 6.2.2.1.1 Factors to Consider in Designing a Control Chart
There are many factors to take into consideration when designing control charts, including:
• Characteristic(s) to chart
• Type of control chart to use
• Sample size and frequency of sampling
• How quickly the chart will detect a problem of a given magnitude
• Economic factors (costs of sampling and testing, costs associated with investigating out-of-control signals,
costs of allowing defective units to reach the customer)
• Production rate

6.2.2.1.2 Types of Control Charts

Control charts may be used for both variables and attributes data. Variables data are those that are
measured quantitatively, such as potency, weight, and pH. Attributes data are those obtained by count-
ing, such as number of rejected lots per month and percent of tablets rejected. For variables data, it
is important to control both the process mean and variation, and both should be charted. A change
in either indicates special causes acting on the process that should be investigated. For attributes data,
such as percent nonconforming units or number of cosmetic flaws in 100 glass vials, only a single
chart for the variable of interest might be kept. A separate chart for variation is not necessary because
the variation of attributes data is related to the mean value; for example, the number of cosmetic
flaws in 100 glass vials is usually modeled by the Poisson distribution, where the standard deviation is
the square root of the mean.

When possible, it is preferable to use variables data rather than attributes data. A measured value
contains more information than an attributes value, such as conforming/nonconforming. Control
charts for variables data have more statistical power and can use smaller sample sizes than attributes
data charts. Although the underlying theory for control charts assumes normally distributed and
uncorrelated data, control charts are robust and generally work well even when these assumptions
are not met (40). One exception is for attributes data with low values, which have a highly skewed
non-normal distribution. Bioburden monitoring is an example of a process with low attributes data
values, where many or most of the data are zeroes. Exact probability control limits use of the nega-
tive binomial, Poisson, or other suitable distribution that might be used to prevent too high of a false
alarm rate; see “Understanding Statistical Process Control, 2nd ed. (42). Additional information on control
charts is provided in Appendix 8.2, Types of Control Charts.

6.2.2.1.3 Process Capability

Statistical process control charts answer the question, “Is the process stable and consistent?” Process
capability statistics answer the question, “Is the process capable of meeting specifications?” Process
capability is the ability of a process to manufacture product that meets pre-defined requirements. It
can be assessed using a variety of tools, including histograms and process capability statistics. The two
most common process capability statistics, Cp and Cpk, are shown in Figure 6.2.2.1.3-1. Cp measures
the capability of a process to meet specifications if it is centered between the specification limits. Cpk
assesses if the process is actually meeting specifications when any lack of centering is considered.
Examples of normally distributed processes with various values of Cp and Cpk are shown in Figure
6.2.2.1.3-2.

62 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Figure 6.2.2.1.3-1 Process Capability Statistics Cp and Cpk

Figure 6.2.2.1.3-2 Examples of Process Capability Statistics Cp and Cpk

105 USL
104
103
102
101
100
99
98
97
96
95 LSL
Cp=2.0 Cp=2.0 Cp=1.0 Cp=1.33 Cp=1.33
Cpk=2.0 Cpk=1.0 Cpk=1.0 Cpk=1.33 Cpk=1.9

If the process is in statistical control, the standard deviation (s) used to calculate Cp and Cpk in Figure
6.2.2.1.3-1 is usually based on estimates derived from the control chart for the standard deviation or
range. These estimates of s will not include between-subgroup variation that may have occurred in
the mean. For an individuals chart where n=1 per subgroup, the standard deviation is usually based
on the moving range, which minimizes the effect of between-subgroup variation. If the standard
deviation is calculated by the familiar equation of all the data combined, this
estimate will include between-subgroup variation, such as between-lot variation, and the indices are
then called Pp and Ppk. If a process is in statistical control, there will be little difference between Cp and
Pp or between Cpk and Ppk. If a process is not in statistical control, it is difficult to determine process
capability because of the lack of process stability; see Figure 6.2.2-2. If a process is not in statistical
control, Pp and Ppk are preferred as they include variation due to lack of stability. However, this prac-
tice is somewhat controversial; see “Introduction to Statistical Quality Control, 6th ed.” (43)
Figure 6.2.2.1.3-2 shows the relationship between the process capability index Cpk and the probability
the process output will be out of specification. The table assumes the process is in statistical control,
normally distributed, and centered between the lower specification limits (LSL) and upper two-sided
specification limits (USL). If the process is not normally distributed, process capability methods for
non-normal distributions should be used.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 63

 Table 6.2.2.1.3-2 Relationship Between Capability and % or Per Million Nonconforming

USL – LSL ±2σ ±3σ ±4σ ±5σ ±6σ

Cpk 0.67 1.00 1.33 1.67 2.00
Nonconforming 4.6% 0.27% 63 ppm 0.6 ppm 2 ppb
% of specification used
150 100 75 60 50
(±3σ limits)

Acceptable values for Cpk depend on the criticality of the characteristic, but 1.0 and 1.33 are common-
ly selected minimum values. Six-sigma quality is usually defined as Cp≥ 2.0 and Cpk ≥ 1.5 for a normally
distributed process in statistical control. See Wheeler (40) or Montgomery (43) for more complete
treatments of SPC and process capability.

6.2.3 Statistical Acceptance Sampling

Statistical acceptance sampling is another commonly used statistical tool for validation. The general
principle is that the sampling used for validation should provide higher confidence than sampling used
during routine production. In validation, larger sample sizes, more replicates, and other such factors
are typically used. Commonly used acceptance sampling plans for validation to ensure that a high
percentage of individual units (e.g., tablets, vials) are conforming are:
• Single sampling for attributes data
• Double sampling for attributes data
• Variables sampling for quantitative data

Samples should be representative of the entire population being sampled. Random, stratified, and
periodic/systematic sampling are the most commonly used approaches. Targeted sampling to in-
clude suspected worst-case locations within the batch or process may be used when appropriate. For
example, samples from the very beginning and end of the batch may be selected to assure that these
potential trouble spots are included, while the rest of the required samples are randomly selected
from throughout the batch.

Reaching at least 90% confidence at the end of PPQ is desirable when using statistical acceptance
sampling for validation with little prior confidence. This means that the combined information from
the PPQ runs shows that there is at least 90% confidence that the validation performance level has
been met; 90% confidence is recommended as the minimum because it is the traditional confidence
associated with detecting unacceptable quality levels (called the Rejection Quality Level [RQL], Lot
Tolerance Percent Defective [LTPD], or Limiting Quality [LQ]) (46). Note that this use of the term
“confidence” is different than the traditional 95% confidence of acceptance associated with the Accep-
tance Quality Limit (AQL) in routine lot acceptance sampling. The AQL relates to the Type I error of
incorrectly rejecting an acceptable lot, while the 90% minimum confidence recommended here refers
to the Type II error of incorrectly accepting an unacceptable process.

Single sampling for attributes is the simplest type of sampling. For example, a sampling plan of n=388
units, accept on 1 nonconformance, reject on 2, would detect a 1% nonconformance rate with 90% con-
fidence. The statistical operating characteristic curve for this sampling plan is shown in Figure 6.2.3-1.

64 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Figure 6.2.3-1 Example of an Operating Characteristic Curve
OC Curve for n=388, a=1, r=2
1.0 0%
0.9 10%

Confidence= 1– pr(accept)
0.8 20%
Probability of Acceptance

0.7 30%
0.6 40%
0.5 50%
0.4 60%
0.3 70%
0.2 80%
0.1 90%
0.0 100%
0.0% 0.5% 1.0% 1.5%
% Nonconforming Units in Process

Double sampling plans for attributes may take a second set of samples depending on the results of
the first set. For example, the double sampling plan n1=250, a1=0, r1=2; n2=250, a2=1, r2=2 will
also detect a 1% nonconformance rate with 90% confidence. The values n1 and n2 are the stage 1 and
stage 2 sample sizes; a1 and a2 are the accept numbers; r1 and r2 are the reject numbers. If a1=0 non-
conformances are found in the first set of n1=250 samples, the sampling plan is passed. If exactly 1
nonconformance is found in the first sample of n1=250 units, an additional n2=250 units are sampled.
If the total number of non-conformances found in the combined 500 samples is no more than a2=1,
the sampling plan is passed. If the total number of nonconformances found in the combined 500
samples is r2=2 or greater, the sampling plan is failed. One advantage of double sampling plans is that
they often have lower false reject rates; i.e., good processes will not fail the sampling plan as often.

Several types of variables sampling plans may be used for validation, one of the most common being
the normal tolerance interval. For example, one normal tolerance interval sampling plan for two-
sided specifications is n=30, k=3.17. If the average ± 3.17 standard deviation is contained within the
specification limits, the sampling plan is passed. This plan also provides 90% confidence in detecting a
1% nonconformance rate. Variables sampling plans assume the data are normally distributed, and this
assumption should be confirmed with a suitable normality test. An advantage of variables sampling
plans is that they often are able to use much smaller sample sizes than attributes plans to provide the
same confidence.

Example: The validation will show with 90% confidence that the process averages ≤0.1% leaking contain-
ers after simulated shipping. This requires an attributes sampling plan of n=2300, accept=0, reject=1.
Three lots will be used for the Stage 2 PPQ, so n = 2300/3 = 767 containers per lot will be inspected for
leakage after simulated shipping. If no leakers are found in the combined n=2300 samples, the sampling
plan is passed.

ANSI/ASQ Z1.4 “Sampling Procedures and Tables for Inspection by Attributes” and ANSI/ASQ Z1.9 “Sam-
pling Procedures and Tables for Inspection by Variables” are commonly used sampling plans for routine
production (47,48). They should be used with care for validation, since they may not provide a high
enough level of confidence. For example, one Z1.4 tightened sampling plan for AQL 0.4% is n=315,

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 65

 a=2, r=3. If a validation lot has 2 nonconforming units in a sample of n=315, the validation lot would
pass the sampling plan. (However, note that 2/315 = 0.63% is substantially larger than the AQL of
0.4%.) Finding 0.63% nonconforming units in a sample does not provide high confidence that the pro-
cess is ≤0.4% nonconforming, if that was the goal of the PPQ. If Z1.4 and Z1.9 are used for validation,
the Operating Characteristic curves in the standards should be consulted to verify that the desired
confidence is achieved.

Not all sampling plans used to make accept/reject decisions are for percent nonconforming units. For
example, the USP test for content uniformity (of dosage units) is specified in terms of a two-stage
sampling plan given in USP. In this case, validation sampling should provide confidence that the USP
test can be passed with high confidence (49).

Example: The sampling plan will show with 95% confidence that the routine USP content uniformity (of
dosage units) test requirements can be met.

6.2.4 Number of Lots for Stage 2 Process Performance Qualification (PPQ)

The number of lots required for Stage 2 PPQ depends on the following:
• Prior information about the process available from Stage 1 Process Design or quality history from similar
processes. The more scientific evidence already available to establishes that the process is capable of consistently
delivering quality product, the fewer the number of PPQ lots required.
• Risk factors, including criticality of the product characteristics and extent of in-process quality control (e.g., PAT,
100% inspection)
• Type of data: attributes (pass/fail) or variables (quantitative)
• Statistical confidence desired
• Production rate (i.e., how often lots are produced). If only one commercial lot is produced per year, it will not
be feasible to require a PPQ with a large number of lots.

Depending on the prior information and/or risk involved, it may not be necessary to determine the
number of PPQ lots using statistical methods. The less information and confidence at the transition to
Stage 2 (PPQ), the more advisable it is to use statistical methods to help determine the number of PPQ
lots where feasible and meaningful. See the Appendix 8.1, Statistical Methods for Determining the
Number of Lots, for statistical approaches to determine the number of lots. Regardless of the number
selected and acceptance criteria used, the data collected during PPQ should be statistically analyzed to
help understand process stability, capability, and within (intra-) and between (inter-) lot variation.

Lots produced during Stage 1 under similar conditions as the PPQ lots may potentially be used to
reduce the number of lots required at PPQ. This can be done using Bayesian statistical methods or by
combining the Stage 1 data and Stage 2 PPQ results – if there are no significant differences in the data
(50). The criteria for combining Stage 1 data and PPQ data should be specified before the PPQ lots
are produced. These criteria would typically include such statistical comparisons as ANOVA (analy-
sis of variance) to compare lot means, Levene/Brown-Forsythe or Bartlett’s test to compare the lot
standard deviations, SPC charts, and equivalence tests to demonstrate that Stage 1 and PPQ data are
similar (51).

6.3 Process Analytical Technology (PAT)

PAT is a method of process control, where the product or in-process material quality attributes are
monitored and measured, and the process parameters and conditions are altered to maintain those
quality attributes. PAT can provide high levels of product quality assurance through the analysis of
material attributes, and the process adjustments. In that quality attributes do not vary outside of the

66 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

prescribed ranges, product and material quality is maintained (52).

PAT can provide an opportunity to enhance process analysis and process knowledge compared to
traditional tests. It can support process validation whether it is a parallel activity (concurrent with
process validation), reductive activity (reduces execution of existing tests), or replacement activity (al-
ternative to traditional testing). Effective use of PAT to provide process control relies on the selection
of correct quality attributes, process performance ranges, and methods for monitoring and reporting.
It also relies on the proper design, use, and validation of the PAT monitoring, measurement, and
control loop systems. The validation of the PAT system is based in part on the following principles:
1. Measurement of the correct product and in-process quality attributes
2. Accuracy and understanding of the correlation between these quality attributes and the process parameters
that will be adjusted
3. Reliability, suitability, capability, and accuracy of the monitoring, measurement, and process control loop or
adjustment systems
4. Acceptable performance of the PAT system throughout commercial manufacturing, including the ability to
identify opportunities for process improvement.

6.3.1 Selection of PAT System

PAT is an enabler to product and process understanding and an element of control strategy. Prior to
the selection of the PAT system, the product and manufacturing process must be developed and well-
understood. Selecting the right PAT system should be based on fitness for purpose, system rugged-
ness, and vendor customer service. Selection criteria should include, but are not limited, to, specificity,
sensitivity and accuracy, electronic integration requirements of information technology compatibil-
ity, data management, and communication. Table 6.3.3-1 provides a partial list of PAT systems, each
of which may provide information helpful to the understanding and validation of the respective drug
manufacturing processes.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 67

 Table 6.3.3-1 Examples of PAT Tools and Their Application

PAT Tools Process Application

Laser-based particle size analyzers crystallization, granulation, milling particle size, particle shape
chemical reactions reaction progress and completion
FT-Infra-Red
raw materials Identification
Nuclear Magnetic Resonance
chemical reactions reaction progress and completion
(NMR)
blending end point determination
Light induced fluorescence (LIF)
compression content uniformity, assay
blending, granulation end point determination
drying water content
Near Infra-red spectroscopy (NIR) compression content uniformity, assay
fermentation nutrient content
raw materials identification
blending end point determination
granulation water content, polymorphism
Raman spectroscopy compression content uniformity, assay
raw materials identification
lyophilization water content, polymorphism
Refractive Index (RI) blending or mixing end point determination
Turbidity blending or mixing end point determination
end point determination, water
Microwave blending, granulation
content
end point determination, water
Acoustic Absorption/Emission blending, granulation
content
end point determination, water
Effusivity blending, granulation
content
pH, Conductivity, Dissolved oxygen
reaction progress, end point
(DO), Oxidation-Reduction Potential fermentation
determination
(ORP)
Focused beam reflectance Formulation of suspensions and
measure particles and droplets
measurements (FBRM) emulsions
nutrient content,
Rapid High-Performance Liquid fermentation reaction progress,
Chromatography (Rapid HPLC) end point determination
chemical reactions reaction progress and completion

68 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

6.3.2 Process Validation Considerations During the
PAT System Design Stage
During PAT system design, information is developed to confirm that correct product and in-process
quality attributes are being measured, and that the correlation between these quality attributes and
the process parameters that will be adjusted is understood and accurate. During PAT system design,
an understanding of how process parameter changes affect product attributes is established. Process
monitoring and control systems are designed and linked to specific product attributes. Ranges of
acceptable process parameter variation are determined. PAT design efforts should include: risk as-
sessment, system feasibility and selection, in-process application development, and consideration of
regulatory requirements.

6.3.2.1 Risk Assessment

The Risk Assessment should identify product and in-process quality attributes that have an effect
on final product quality. The risk assessment should identify process steps and conditions that affect
these attributes and can be measured and adjusted to assure product quality. Quality attributes, and
corresponding process steps and conditions that are not monitored by the PAT system, may require
other means to assure or validate performance. Having PAT systems is expected to lower the risk
to product quality, by having additional controls, timely responses, increased detectability, increased
understanding, and information (e.g., identification, measurement, control of CQAs). These features
enable a more informed risk assessment decision. Tools for the assessment and evaluation of PAT
processes and systems are discussed in Section 6.1, as well as PDA TR 54, ICH Q9 and other publica-
tions (12,13,30).

6.3.2.2 In-Process Application and Method Development

The PAT methods for in-process product measurement and process adjustment should be selected
and validated for specificity, linearity, range, accuracy, precision, repeatability, robustness, detection
limit, and quantitation limit to ensure that the method is fit for purpose (13).

6.3.3 Process Qualification Considerations for PAT

The Process Qualification Stage is where information is developed to confirm that the monitoring,
measurement, and process control or adjustment systems are suitable, capable, accurate, and reliable.
One key to effective PAT process control is the reliable operation of instruments and equipment. For
implementation, an implementation and validation team should be assembled to categorize the vali-
dation requirements and propose acceptance criteria for each unit of operation, based on the applica-
tion or intended use of the PAT system and method. These requirements and criteria will ultimately
be included in a validation protocol and described in the validation report. The acceptance criteria
should be aligned with the expected specification, protocol requirements, development experience,
and manufacturing practice.

Function and operation of the equipment and instrumentation used in the PAT system should be
qualified to assure that it will monitor and control the process parameters accurately and reliably.
Equipment and instruments used during the process should be qualified to verify that they are suit-
able for in-process use, including compatibility with process materials and conditions, accuracy, sen-
sitivity, security, and reliability.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 69

 6.3.4 Continued Process Verification Considerations for PAT
The Continued Process Verification Stage is where information is obtained to confirm that the PAT
system performs at an acceptable level throughout commercial manufacturing. It also determines
where product and in-process quality attributes, or process parameters fall out of expected ranges;
those that do are identified, investigated for cause, and addressed.

By definition, PAT provides continuous process and product attribute verification. Stage 3 activities
should therefore focus on accuracy and reliability of control methods, possible process control im-
provements, and process variables missed during process development and qualification. Evaluation
of PAT and or in-process derived data should be part of the Quality System and review processes (11).
Where data trending shows excursions in anticipated monitoring results, analysis of the cause of the
excursion should be conducted to determine if changes to the control system are needed or opportu-
nities for process improvement can be identified.

When variables are found that are not being monitored adequately, changes to the monitoring meth-
ods may be needed. All changes should be evaluated for impact on the process and product attributes.
Changes should be evaluated and actions implemented to assure that residual risks do not adversely
affect process performance or product quality. These actions may include steps to qualify the changed
process and equipment.

6.4 Technology Transfer

For a lifecycle approach to process validation to be effective, all information that is available to sup-
port the understanding of the process, including that from other sites and similar processes, should
be considered. This information should be useful, accurate, and complete. The goal of technology
transfer (TT) activities is to communicate product and process knowledge between development and
manufacturing, and within or between manufacturing sites to achieve product realization. This infor-
mation forms the basis for the manufacturing process, control strategy, process validation approach,
and ongoing continual improvement (52). It also provides valuable insight into the development of
the process, including process variables, process performance, and process control strategies.

Technology transfer is successful if process understanding has increased, and there is documented
evidence that the recipient of the technology transfer can routinely reproduce the transferred prod-
uct, process, or method against a predefined set of specifications from the sender. Process understand-
ing and knowledge increase significantly during technology transfer, providing useful information
for process control strategy design and process validation. Technology transfer can occur at different
stages of the process validation lifecycle. If a new process is being transferred from research and de-
velopment to commercial manufacturing, the technology transfer may occur between Stages 1 and
2. However, if it occurs after a product has been launched and it is in the commercial manufactur-
ing phase, then transfer will occur during Stages 2 and 3. Refer to Table 6.4-1 below for distribution
of Technology Transfer Activities throughout the Product Lifecycle, which outlines the increasing
knowledge and process understanding with each technology transfer.

70 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Table 6.4-1 Technology Transfer Activities throughout Product Lifecycle

Process
Validation Activities Knowledge Development/Data Application
Lifecycle Stage
Stage 1 Process Design provides product and process develop- Development Report: Technology Transfer Batches manufactured during Stage 1
ment knowledge and data for technology transfer. • Development history, including criticality assessments are intended to establish comparability of product quality
between sites and, develop filing/market authorization data.
and DoE with sources of variation.
• Data and knowledge development from stability stud- Development Report summarizes activities from Stage 1.
ies and development batches
• Rationale for specifications and methods
• Critical Process Parameters (CPPs)
• Critical Material Attributes (CMAs)
• Critical Quality Attributes (CQAs)
• KPPs, PARs, NORs
• Manufacturing Process Description, Equipment Train
Stage 2 Most technology transfer activities in a product lifecycle Technology Transfer Strategy: Technology Transfer Batches manufactured during Stage
are carried out at Stage 2:
• Product and Process Description (as designed from 2 are intended to reproduce the manufacturing process,
• Development of Transfer Strategy Stage 1, and reported in the Development Report) including components and composition configurations at
the transfer site, and to conduct PPQ.
• Manufacturing of Commercial Scale PPQ Batches • Assessment of Site Change Requirements; e.g., Post-
• Site Equivalency Analysis (from receiving to sending unit) Approval and, Prior-Approval with rationale. Category Equivalency between sites is intended to compare equip-
under SUPAC guidelines, if applicable ment and facilities to assure that they are equivalent and
• Transfer and Validation of Analytical Methods qualified for commercial manufacturing
• Number of batches required to meet transfer require-
• Confirming CPPs at Commercial Scale. ments, including validation/PPQ strategy/Matrix Ap-
• Conducting Stability Studies at Commercial Scale under proach
Commercial Package configurations • Specifications and Methods Transfer Plan
• Confirming Risk Assessments, Criticality Analysis • Validation Plan
• Establish Sampling Plans and Statistical Methods at • Control Strategy
Commercial Scale
• Evaluation of Personnel Qualifications and Training

71 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Process
Validation Activities Knowledge Development/Data Application
Lifecycle Stage
Stage 3 Technology Transfer activities at Stage 3 are most likely Similar to activities in Stage 2, a Technology Transfer Transfer to a new location within a manufacturing site, to
carried out for products that have already been validated Strategy is recommended. The Strategy would include an alternate site of the company, or to a contract manufac-
and are on the market. These are known as post-approval data listed under Stage 2 of this Table. For products at turer. Filing requirements are defined by SUPAC, as these
changes under the SUPAC guidelines (for small mol- Stage 3, additional data and knowledge will be available. have different implications from the regulatory standpoint.
ecules), and apply to changes to alternate manufacturing It should be considered and evaluated prior to starting Validation requirements apply equally to any of the technol-
sites within a company or to contract manufacturers. technology transfer activities. ogy transfer scenarios.
At Stage 3, technology transfer activities may pose op-
portunities for process improvement at the receiving site
using historical control and quality systems data. Valu-
able data to evaluate include:
• Stage 2 Technology Transfer and Validation Reports
• Annual Product Reports, including Process Trending
and Process Capability
• History of Investigations, CAPA, Change Control, OOS,
Complaints Reports, Field Alerts, Stability Studies,
Yield Variations
• Executed Batch Records
• Sampling and Test Plans
• Analytical Data
• Conduct Gap Analysis at current vs. transfer site to
assess risks and variations, including:
o Manufacturing Equipment Train design and oper-
ating principle, as well as qualification status
o Confirmation of CPPs, equipment operating
ranges at new site
o Suppliers
o Personnel
o New Site state of compliance
Technology Transfer Strategy:
• Product and Process Description (as designed from
Stage 1,and reported in Development Report and Vali-
dation Reports)
• Assessment of Site Change Regulatory Require-
ments: Post-Approval, with rationale
• Number of batches required to meet transfer require-
ments, including validation/PPQ strategy/Matrix Approach
• Specifications and Methods Transfer Plan
• Validation Plan & Control Strategy

72 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Figure 6.4-1 Distribution of Technology Transfer Activities throughout the Product Lifecycle

Process Understanding

Knowledge

Information

Data

Scale Up, Commercial

Characterization Pilot of TT Stability Studies, PPQ TT Mfg.
Development Studios Optimization

Stage 1 Stage 2 Stage 3

TT: Technology Transfer

6.5 Knowledge Management

The effective and efficient capture and analysis of process-related information is essential to process
understanding and validation. Information that supports process validation should be identified, ana-
lyzed, communicated, maintained, and available. It is important to recognize that knowledge man-
agement is not just data collection. It involves a strategic, systemic, and methodical approach that
should include the acquisition of data at pivotal process steps, rigorous data analysis, easy access, and
controlled storage and dissemination of information about the product, process, and components. All
desired or necessary activities should be included, for example:
• Technology transfers
• Process understanding
• Product characterization

Knowledge management includes systems that capture review and feedback information in an effort
to ensure correct decisions were made, and identify where process improvements can be implemented.
Sources of knowledge include, but are not limited to:
• Prior knowledge (public domain or internally • Manufacturing experience
documented, such as similar processes)
• Risk assessments
• Pharmaceutical development studies
• Technology transfer activities • Continual improvement
• Process validation studies over the product lifecycle • Change management activities

The concept of sustainable and continually improved knowledge systems is essential to a lifecycle
process validation program. The flow of information from Stage 1 Process Design to Stage 2 Process

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 73

 Qualification and back from Stage 3 Continued Process Verification form the basis of the lifecycle
approach.

Knowledge management systems should be designed, installed, used, and maintained. They play a
pivotal role in finding problems and preventing process shifts by providing feedback for continuous
improvement efforts (4).

Appropriate information must be acquired, used, and archived. It should be accurate, timely, and use-
ful. Information should also be properly interpreted and effectively communicated.

Information or knowledge is gained in Stages 2 and 3 that can improve the process should be communi-
cated back to those responsible for process design and development. The information (including respon-
sible individuals, sampling plans, and justification) should be communicated via an appropriate tool.

Information needed to support the process validation effort should also be communicated to those
responsible for monitoring and providing feedback on commercial product manufacturing. A system
should be in place to provide feedback to those responsible for process design and development, to
confirm the accuracy of early process design assumptions, and to improve the process where possible.

When changes are made in Stages 2 and 3, they should be communicated to all affected parties in an
efficient, accurate, and timely manner. Formal Change Control procedures are a recommended and
required Quality System component (4).

Transparent interaction between teams collecting data, performing risk assessments, and transferring
information is essential to the process validation effort. Joint reviews between teams responsible for
process development, risk assessments, and data collection should be conducted throughout the life-
cycle of the process. These reviews enable the effective transfer of information from scale-up through
full-scale manufacturing batches, and help to ensure that the process operates in a reliable and predict-
able manner.

74 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 7.0 Examples

7.1 Large Molecule (Biotech)

An example of the three stages of process validation for a humanized IgG1 is provided in Table7.1-1 (Stage 1), Table 7.1-2 (Stage 2), and Table 7.1-3 (Stage 3).
Table 7.1-1 Stage 1: Process Design

Category Activities Outputs/ Deliverables Rationale/Examples

Process Development Establish TPP & QTPP Humanized IgG1; TPPs and QTPPs were established.
• Immunological indication; MOA (mechanism of action) requires both CDC (complement depended cytotoxicity) and ADCC (antibody dependent cell-mediated cytotox)
activity; IV administration at a fixed dosage;
• Liquid formulation with concentration at 20 mg/mL, iso-osmolar solution; material provided in a single use vial with a shelf life of at least 24 months at 2-8°C.
Identify Critical Quality • Presumptive CQAs (inherent attributes from the molecule that provide desired activity, purity, and safety) were identified based on prior knowledge. Deamidation, Aggregate, Host Cell Protein, Residual DNA, etc.
Attributes • Potential process parameters that impact the CQAs were identified for each unit operation based on platform information.
Define Manufacturing Prior knowledge, existing risk assessments for similar molecules, and early development data were used to define, unit operations: Seed train, bioreactor, harvest, Protein A,
Process viral inactivation, column purification 2, column purification 3, viral filtration, UFDF. In addition:
• Normal Operating Ranges identified
• Raw materials identified
• Cell line characterized to show free from adventitious agents.
• Master and working cell banks prepared and characterized.
• Analytical method development was started.
• Initial formulation development (liquid or frozen) was initiated. Due to ease of control, frozen was initially selected while the liquid formulation was being developed in parallel.
• A Process Design Summary Report was created with preliminary process information.
• Clinical Phase 1 & 2 manufacturing: Upstream Process Parameters:
o Material was produced for First-in-Human studies, in a GMP facility in a 2000L bioreactor facility, using a scaled-down version of the intended commercial process. • Viable cell density
Samples were put on stability to establish expiration times. • % Viability
o Material was produced for Phase 2 in a 2000L bioreactor process using the same GMP facility. Samples were taken and used for characterization studies in small-scale • Temperature
equipment (satellite studies) to define the eventual commercial process. Product was analyzed for the following (at a minimum):
• pH
• Appearance and identity
• Dissolved Oxygen
• Purity (IEC, SEC, CE SDS, endotoxin, bioburden, impurities
Downstream Process Parameters:
• Potency
o Initial product acceptance criteria based on targets were set from other molecules and early development studies. Stability studies were initiated using a subset of the • Protein load
release testing assays • Protein concentration
o Most of the analytical methods were qualified at this stage. • Elution buffer pH
• Clinical Phase 3 manufacturing was performed in a different 2000L bioreactor facility. Prior to the start of phase 3 material manufacture, some of the following activities • Viral inactivation pH
were performed • Diafiltration volumes
o Tech transfer process was conducted to transfer the process from the Phase 2 facility to a Phase 3 facility. • A team of scientists led the tech transfer effort by performing facility fit, generating
o Comparability study (DS & DP) protocols were generated technical reports, training of operators, and transferring of manufacturing process and
o Batch records were created associated scale-down models.
o Operator training was performed
o Primary containers were finalized
• After Phase 3 material manufacture, the Process Design Summary Report was updated (e.g., CQAs and CPPs, unit operations)
Quality Risk Assessment A modified FMEA was used to perform Quality Risk Assessment (QRA). A template created for similar products was used as a starting material with appropriate modifications.
Using the risk assessment process:
• Initial categorization of process parameters was performed
• Initial framework for control strategy was created based on high risks identified in the risk assessment

75 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Category Activities Outputs/ Deliverables Rationale/Examples
Process Characterization • Process characterization studies were designed based on prioritization developed from risks identified in the QRA. • Downstream process determined that acidic variants impacted biological activity.
• Statistical methods involving DoEs (screening designs to full factorial) were used to understand interactions of high-risk parameters and a design space developed wherever Placed tighter controls on in-process hold times to control level of acidic variants.
possible. Updated Quality Risk Assessment and the Control Strategy. Increased the concentra-
• Scale-down models were created and tested; some required qualification (e.g., virus clearance). In these cases, protocols were created and approved by Quality. tion of final bulk to save on storage capacity
• Based on characterization and small-scale model studies, operating ranges for process parameters were finalized.
• Acceptance ranges for performance parameters were established
Finalize CQAs and CPPs • Based on process characterization and scale-down model studies, the QRA was updated, which in several cases required re-scoring.
• In a cross-functional team, the CQAs and CPPs were reviewed and finalized. The final CQAs and CPPs were subject to approval by the Health Authorities wherever applicable.
• The control strategy was updated based on the understanding of CQAs, CPPs, process controls, and detection capabilities.
Documenting Process Design • The Process Design Summary Report was updated (CQAs, CPPs, unit operations, operating ranges, specifications, and acceptance criteria and controls).
• A commercial manufacturing was site was identified (12K bioreactor capacity), and a team of scientist and process engineers performed a facility fit analysis to identify any
gaps in equipment capabilities.
• Tech transfer process was initiated to the commercial site. A tech transfer risk assessment was performed to understand the high risks. Scale-down model process transfer
was also started in parallel.
• Around this time, the analytical method validation was completed.
Process Validation Master • Specific validation protocols were identified.
Plan • The process validation strategy and ancillary studies were described in the plan.
Process Qualification Equipment, Utilities, and • The facility fit assessment identified the requirement of a larger scale centrifuge.
Facility Qualification • Based on user requirements and design specifications, the new centrifuge was ordered. After FAT and SAT, the equipment was commissioned and qualified. To understand
the control required, a risk assessment was performed.

Table 7.1-2 Stage 2: Process Qualification (Continued)

Category Activities Outputs/ Deliverables Rationale/Examples

Process Performance Technology Transfer and • The transfer process used engineering runs to demonstrate that the process worked and to fine-tune the operation set points.
Qualification Engineering Runs • Two engineering runs were performed using GMP materials with draft batch production records. These runs enabled training on the new process for the operators.
• The Process Design Summary Report was updated with any changes to process parameters.
Process Performance • A checklist was used to ensure that all the processes and procedures were in place to start the PPQ process.
Qualification Readiness • PPQ protocols were drafted and approved.
Assessment • A sampling plan that described the sample points, number of samples, statistical justification, and analytical methods was created and approved.
• A Continued Process Verification plan was created to identify the parameters and attributes to be tested and monitored during PPQ and Stage 3 (Continued Process Verifica-
tion). Some of the elements included in the plan were justification of parameters, frequency of, statistical procedures used to determine state of control, and handling of
excursions.
• A qualitative decision tool was used to determine the number of PPQ runs. Some of the factors considered were:
o Process variability (e.g., novel and difficult scale-up unit operations, raw material variability, age of equipment and facility, level of commercial manufacturing experience
of operators, clinical manufacturing experience, robustness of control strategy).
• The tool suggested a range of 5-6 runs for the PPQ campaign.
• Discussions with the Health Authorities are helpful and generally a proposal is submitted for the number of runs.
• A similar approach was taken for DP PPQ campaign.
• Materials generated during the DP PPQ campaign will likely expire before approval. Depending upon company practices, one may perform one run at full scale and others
at reduced (approximately 10%) scale.
• Cleaning validation that is specific for the new process was performed concurrently with the PPQ runs.
• Product and process comparability was initiated with approved protocols.
• A comparability plan describes the actions to be taken in the event of significant process changes (including site change). The plan describes the testing program to be used
to demonstrate comparability between the Phase 3 and the commercial processes.
PPQ campaign • The qualification lots were scheduled in advance of the targeted submission date to allow for sufficient real-time stability data in the application. In general, Health Authorities require 6 months of real-time stability data at the time of
• PPQ campaign was conducted as per the protocols. submission.
• The PPQ was concluded to be successful after all the acceptance criteria were met. By meeting the statistically-derived acceptance criteria, the process was demonstrated • Any excursions were handled according to the established procedures.
to be in a preliminary state of control. The demonstration of state of control will continue into Stage 3. • Additional sampling is performed for all the runs in the event of an unforeseen inci-
• The PPQ reports were generated and approved. dent, which would have compromised the initial PPQ runs.
• The Process Design Summary Report was updated appropriately.

76 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Category Activities Outputs/ Deliverables Rationale/Examples
Stability • Three lots of DS and DP from the PPQ campaign were put into the stability program. • In addition to real-time testing and the designated storage temperature, stability at
• Multiple freeze and thaw cycles of the DS were also included in this study. accelerated conditions is performed per ICH guidelines.
• The stability program also includes a comprehensive study in which the DS is held at
its longest expiry and then used to prepare DP vials, which will are also held for the
entire expiry time.
• In addition to the primary stability data obtained during the PPQ runs, supportive
stability data acquired during clinical development is also used in the submission
package.

Table 7.1-3 Stage 3: Continued Process Verification

Category Activities Outputs/ Deliverables Rationale/Examples

Continued Process Process Monitoring • The CPV plan that was developed prior to start of the PPQ was submitted to the Health Authorities. • Preliminary control limits were established after 15 commercial batches (including
Verification • Testing and monitoring were performed during Stage 3 according to the CPV plan. PPQ batches) were manufactured.
• CPV data review was conducted as described in the CPV plan. • Final control limits were established after 30 commercial batches had been manu-
• The monitoring reports generated supplemented the Annual Product Review. factured.
• The CPV plan was used throughout the product lifecycle and helped to ensure that the process was in a state of control.
Product Technical Teams • Each commercial product had a Product Technical Team (PTT) that helped to oversee the process for the remainder of the product lifetime. • The PTT is cross-functional, including manufacturing, process development, analyti-
• The PTT was also responsible for reviewing data from multiple production sites to ensure consistent process performance and product quality. cal, quality, and statistics. The team is responsible for reviewing the processing data
that accumulates during commercial production.
• The PTT can recommend process changes and helps to ensure continuous improve-
ment.
Specification File • A manufacturing process specifications file was generated at the time of the license submission. The file is maintained throughout the product lifetime and is be updated to include in-
• The file was updated upon approval and contained the licensed parameters that had been agreed to by the agency. process and specification changes that might occur.

7.2 Small Molecule (Parenteral)

An example of the three stages of process validation for an organic, parenteral dosage form in Table 7.2-1 (Stage 1), Table 7.2-2 (Stage 2), and Table 7.2-3 (Stage 3).
Table 7.2-1 Stage 1: Process Design
Category Activities Outputs/ Deliverables Rationale/Examples
Process Development Establish TPP & QTPP Parenteral drug solution dosage form:, sterile formulation in three multiple strengths, intended to comply with the USP compendial requirements for injection. Target shelf life at The product development process had no clinical trials; therapeutic strength relied on
least 24 months at 25°C. bioequivalence. Thus, clinical manufacturing experience was minimal compared to a new
chemical entity.
Identify Critical Quality Active collaboration took place between R&D, development, formulators, and analytical scientists to identify potential CQAs and methods for detection. Experience with past
Attributes liquid dosage form manufacturing was vital in identifying CQAs.
Assays used to release product and test methods to release API were developed and verified at Stage 1 with the intention that they would be validated and transferred to the
manufacturing site to support PPQ.

77 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Category Activities Outputs/ Deliverables Rationale/Examples
Define Manufacturing Development was based on experience with previous and existing processes, excipients, and capabilities at the company’s current manufacturing sites. The lab scale formulation Samples from these pilot batches of 400 L were analyzed and tested to narrow down
Process batches were produced using identical primary packaging material. All raw materials were in the company’s GMP system. formulations based on compatibility and stability due to:
A DoE concluded that the DP was heat-sensitive, and therefore, would be manufactured aseptically and not terminally sterilized. Followed by lab feasibility / formulation batches, • Light sensitivity,
the pilot scale formulation stability batches (with at least three formulation pH levels) were prepared in an R&D Pilot Plant. The solution stability due to maximum temperature • Oxygen sensitivity, and
during compounding, filtration, filling and its impact on DP degradation rate and impurity profile was established. • Formulation pH,
At least two API supplier batches were considered. Intentions were to use standard and familiar unit operations and minimize the time to develop the process. The process for the • Container material incompatibility,
formulation studies performed at pilot scale (at least 10% of commercial scale) established knowledge on process variability, CPP, and CQAs. The process scale-up parameters, • Manufacturing material incompatibility,
manufacturing specification, analytical, and biological specifications were established through pilot scale runs. • Thermal stability,
• Color formation, and
• Any anticipated stability-limiting factors.
Solution temperature controls during mixing, filling, and storage were also followed as
controls.
• Scale-up models for unit operations were developed. R&D personnel provided justification for the models and documented the limitations through design and stage gate Tests of the Quality Attributes included: These studies established CPPs:
review processes. • Appearance and identity • RPM
• The Process Evaluation (PE) studies were initiated prior to manufacturing of the stability batch using the bracketing approach at the scale-up production GMP manufacturing • Purity test • Temperature
site. A total of two scale-up batches of highest DP concentration using active product ingredient from two different suppliers were manufactured. To establish and under-
• Solution pH • Dissolved oxygen
stand all CPP and CQA, both batches were produced at full scale. The study design was based on risk assessment accompanied by an extended in-process control program
defined in protocols and product- /batch-specific sampling plans. These studies established: • Osmolality • Mixing time
o drug dissolution profile • Dissolution profiles
o degradation over the manufacturing process • Process impurities
o solution filter compatibility • Particulate matter
o solution hold time • Microbiological attributes
o solution closure/container compatibility. • Sterility assurance levels
There were 10 sampling points throughout the PE batch of 2800 L during filling. These
• Extensive sampling and specification evaluations were conducted. Characterization and comparisons among the batches for both active ingredient and finished product were
performed. The data demonstrated that the DP met the finished product specification when produced using the worst case scenario. encompassed multiple (e.g., triplicate) samples at the beginning, middle, and end of the
process step. This approach is patterned after other heterogeneous system sampling
• Research personnel were primarily responsible for these batches, but manufacturing site personnel were also involved.
practices, such as the FDA’s bulk powder blend sampling schemes.
Quality Risk Assessments Formal risk assessments were performed during development. The scope was limited to the manufacturing risks of the product and processes. An approval of this document Risk Ranking and Filtering (RRF), which included severity and probability components,
indicated that the residual risks and associated risk scores with development DOE activities were acceptable to process with an entry into the stability design phase. Well after was used. This is a simpler tool to understand; it enabled focusing on the most important
Stage 2, other formal risk assessments were conducted during Stage 3 and during a long and successful commercial manufacturing phase. This included linking the worst case factors. Other tools used were FMEAs and Cause and Effect diagrams.
scenario for various operations, which aided in the development of the design space.
Process Characterization Unit operations were optimized to improve efficiency and robustness. Using experimental and scale-up studies, scientists were able to establish scale-up process parameters Qualified excipients were those that impacted CQAs, such as the ones that controlled pH
and perform evaluations prior to stability runs. and osmolality.
Improvements in the process included enhancing immediate dissolution through solution mixing process optimization, in-tank solution pH, and a dissolved oxygen monitoring
system. Process characterization or evaluation studies were designed using a DOE approach to minimize experiments. There were numerous research and scale-up / transfer
reports, along with qualification and process understanding reports. Qualification of the most critical excipients from a different vendor was performed on the full-scale drug product.
Finalize CQAs and CPPs • Issues with respect to API dissolution occurred during development and scale-up due to variations in raw material particle size. The scale-up process parameters for agitation CQAs were:
and solution temperature were modified and evaluated from model calculations. Manufacturing procedure specifications were modified and evaluated to confirm finished • Solution pH • Drug concentration (potency)
product CQA.
• Dissolved oxygen • Osmolality
• CQAs for the solution product were identified early in development, refined during Stage 1, and implemented as final specification in the manufacturing procedure. These gener-
• Drug dissolution and homogeneity • Microbiological attributes
ally inherent attributes from the molecule and formulation provided desired activity, purity, and safety. The review and approval of the CQAs was performed by a dedicated team
and documented in a formal report. The process parameters that impacted the CQAs were identified in PV protocols and their criticality determined from results of the PV studies. • Process (i.e., drug) impurities • Sterility assurance levels

Documenting Process Design • Analytical methods were not validated for PE/demonstration batches; however, they were validated and transferred from R&D to manufacturing sites prior to stability batch Analytical methods were dependable but not validated initially since:
production at a GMP site. Factors included specificity, forced degradation, precision, linearity, LOD/LOQ, accuracy, and robustness. • The knowledge-gathering phase with experimental batches early in the lifecycle were
• Scientists were encouraged to write technical reports that summarized different aspects of the process. In general, they focused on a single unit operation, describing carried out
changes and improvements. A technical review reference document was also prepared. It summarized all of the developmental reports covering methods, ranges, condi- • Draft specs were used and case changes made in the ranges
tions, and knowledge of the entire process.
• Saving on timeline of analytical method validation at this stage
• The documents are updated each time significant process changes occur. The technical review reference document and associated specifications and procedures are filed
in a central archiving system, and are then used by manufacturing for generation of batch production records. Upon site transfer, lab analysts will be present for method validation according to internal
SOPs. These will also meet ICH Q2, USP or other regulatory or compendial standards.
Process Validation Master Developed a detailed Validation Master Plan (VMP) that identified specific studies to be performed. Individual Process Validation protocols were written for each batch. The PPQ The process validation plan was initiated prior to Stage 2 to identify supportive information
Plan batches were completed just before the expected NDA approval. needed from Stage 1. However, the formal Validation Master Plan was finalized during
• In addition to new process validation studies, the plan identified studies and appropriate references that had been executed for other projects, but would be used to support Stage 2, when all attributes, parameters, and systems were known.
this product.

78 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Table 7.2-2 Stage 2: Process Qualification

Category Activities Outputs/ Deliverables Rationale/Examples

Process Qualification Equipment, Utilities, and The extent of the qualification and verification of the equipment was based on risk assessment. The critical aspects (e.g., critical Qualification was performed for:
Facility Qualification functions, controls, and attributes) and other system components or functions were verified to be fit for their intended use. Equipment • Agitator mixing speeds
and utilities had to be in qualified states of their own for any product used. This activity was carried out according to plant procedures • Sensors, such as level, volume, and temperature measurements
to maintain a state of control.
• In-tank pH measurements
Qualifications and calibrations were confirmed. Any system for which proper operation was fully ensured through routine calibration • Oxygen measurements
and/or preventive maintenance programs may not have required formal qualification. • Solution filling system
• Storage chambers (frozen).

Process Performance Technology Transfer and • Manufacturing, analytical, and biological procedure specs were transferred to the manufacturing site based on process evalua- The demonstration batch included verifying:
Qualification Engineering Runs tion batch results. • Solution mixing process
• Three stability batches of DP strength were produced at 10-15% of commercial batch volume. • Filling process
• Three different batches from various API suppliers were factored in (matrix approach) among all stability batches. One batch with • Sterilization process (as applicable)
the highest strength per API supplier was performed using the worst case scenario for CPP (e.g., solution hold-time). Stability
• Packaging and confirmation of finished product meeting final specifications.
studies were initiated using tank release, in-process testing, and finished product release testing assays.
• Analytical and microbiological methods were validated. Assays were performed by a dedicated stability operations group. Long For the registration, the different suppliers provided matrix; everything else in the process remained the same.
term stability studies for the aforementioned batches at 2-8° C/60% RH; 30° C/65% RH and 40° C/75%RH were initiated. For one PPQ batches verified the same process parameters and quality attributes used in the demonstration (pre-validation) batch.
batch of each strength, stability data were generated for the products, which were stored in an inverted orientation. A formal
stability plan was prepared prior to entering stability production, and was issued prior to submission. Formal stability production
protocols were issued for each code prior to stability production.
• A full-scale commercial ‘demo” batch was followed by multiple PPQ batches at a manufacturing facility for launch quantities
Process Performance • Sites for the commercial production process were identified in Stage 1. Readiness for PPQ was confirmed at ‘Stage gate’ meet- Ensure finalization of:
Qualification Readiness ings. PPQs runs took place under nominal, routine conditions. • Specifications • Previous batches done at worst case scenario
Assessment
• Previous reports (e.g., Formulation, PE) • The readiness of other items, (e.g., labeling)
• Equipment and facility qualifications
PPQ campaign Material requirements to support commercial runs determined the number of runs for the campaign. The DP PPQ campaign consisted • Solution Mixing step: Time, Temp., dissolve oxygen, agitator speed, solution homogeneity by drug assay, and pH– mixing valida-
of 5 runs, covering 3 batches with highest strength, 1 batch with mid-range strength, and 1 batch with the lowest strength of finished tion
product made at the commercial production scale. Additional sampling was performed for all the runs. All runs that meet commercial • Sterile fill or terminally sterilized finished product testing: finished product assay, degradation and impurities, pH, particulate,
release criteria could be used to support commercial supply. All PPQ batches were performed at nominal conditions. microbial and sterility testing.
• Hold study was performed on one of the three highest strength product batches to establish and validate hold intervals for solu- The number of batches in the entire process validation Stages 1 and 2 were:
tion mix, fill, and hold prior to sterilization (as applicable).
• 3-6 feasibility batches
• Data report of initial analysis of the variation of outputs such as quality and performance attributes in stages 1 and 2
• 3-6 formulation batches
• Cleaning validation that was specific for the new process was performed concurrently with the PPQ runs.
• 1-2 PE batches
• 3 stability batches
• 5 PPQ batches
Data were analyzed for the total of both stages.
A slight variation in the number of development and PPQ batches can depend on dosage strengths, complexity of formulation and
process, and results of PE and stability batches.
Five batches provided 3 at worst-case of most concentrated conditions. Lower concentrations were confirmed with 2 batches.
Timing of PPQ batches were scheduled in advance of the targeted NDA submission date to allow for initial data in the application. In
this case, 1 month of real-time and accelerated stability data was available. This led to the respective start of the DS and DP PPQ runs
12 and 9 months prior to the anticipated approval date. PPQ batches were thus able to be commercialized.
Stability All batches of DP from the PPQ campaign were put into the stability program. In addition to real-time testing and the designated
storage temperature, stability at accelerated conditions was performed per ICH guidelines. In addition to the primary stability data
obtained during the Stage 1 runs, supportive stability data acquired during PPQ runs was also used in the submission package on an
as-needed basis.

79 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 Table 7.2-3 Stage 3: Continued Process Verification

Category Activities Outputs/ Deliverables Rationale/Examples

Continued Process Process Monitoring A process monitoring plan and trending were developed during the commercial phase. The monitoring plan was used during routine • Performance metrics (e.g., yields, complaints, and deviations) continued during commercial production.
Verification manufacturing to help ensure that the process remained in a state of control. The process capability metric and trend analysis were • Process robustness contour plots are used when the number of data points is small (e.g., less than 20-25). Process performance capabil-
performed with positive outcomes. ity indices, such as PpK and /or CpK, are used for 25 or greater data points.
Product Technical Teams Each product has a Product Technical Team (PTT) that helps to oversee the process for the remainder of the product’s lifetime. Additional studies, including PAT, DOEs, continuous processing experiments, and clinical studies were carried out in a long Stage
The PTT is cross-functional, with representatives from manufacturing, process development, analytical, quality, and statistics. The 3 to improve the product line.
team is responsible for reviewing the processing data that accumulates during commercial production. It can recommend process
changes and help ensure continuous improvement. The PTT is also responsible for reviewing data from multiple production sites to
ensure consistent process performance and product quality.
Specification File/NDA Numerous supplements to the registrations were made to add new manufacturing and testing facilities. Transfers and process valida-
Supplements tions were carried out during Stage 3.
Manufacturing knowledge documentation files generated at the time of development are updated regularly with all pertinent studies.
Critical quality attributes and parameters have been agreed to by the development and quality organizations. The process understand-
ing file is maintained throughout the product lifetime and is updated to include any process and/or specification changes.

80 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

 8.0 Appendices

8.1 Appendix 1: Statistical Methods for

Determining the Number of Lots
Listed below are statistical approaches used to determine the number of lots that may be required at
the PPQ stage. Other approaches may also be suitable. As there is no standard industry approach to
statistically determine the number of lots, multiple options are offered. This section will provide ap-
plied statistical methods for determining the number of lots. It will also stimulate further discussion
on this issue. Regardless of the number of lots selected and the acceptance criteria used, the data col-
lected during PPQ as well as CPV should be statistically analyzed to help understand process stability,
capability, and within (intra-) and between-lot (inter-lot) variation.

8.1.1 Average Run Length (ARL) to detect a p×100% lot failure rate
The average number of lots until the first lot failure is ARL = 1/p, where p is the lot failure rate that
is important to detect.

Example: A lot failure rate of 20% is deemed unacceptable for a given process. A lot failure rate of 20%
would be detected on average in 1/0.2 = 5 lots.

Common choices for p would be 25%, 20%, 10%, and 5%, depending on the other factors given ear-
lier (e.g., prior knowledge, risk, production rate) Five (5%) would generally be the tightest value to
consider since a process running right at the Acceptance Quality Limit is still expected to have a 5%
lot rejection rate. If applicable, this approach can also be used to determine the number of lots to use
with tightened sampling during CPV (continued process verification).It may be particularly useful
when there are many quality attributes to assess. Rather than determine the number of lots required
separately for each attribute, the PPQ stage is complete when all attributes pass for the required num-
ber of lots.

8.1.2 Range of between-lot (inter-lot) variation expected to be covered in nL lots

Table 8.1.2-1 outlines the expected between-lot variation coverage in nL lots.
Table 8.1.2-1 Expected Between-Lot Variation Coverage in nL Lots
Expected Coverage Number of lots nL
33% 2
50% 3
60% 4
67% 5
75% 7
80% 9
85% 12
90% 19
95% 39

Example: It is desired to represent two-thirds = 67% of the between-lot variation during PPQ. The number
of lots required is nL=5 lots.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 81

 Expected coverage is calculated as (nL – 1) / (nL + 1). This follows from the expected percentile of an
order statistic being its rank divided by n+1 (53). This approach does not require between-lot nor-
mality. The method may be modified to provide confidence levels of coverage instead of expected
coverage. The approach may be used to determine “step-down sampling” during CPV. For example,
highly tightened sampling may be used in PPQ for the first three lots until 50% coverage is reached.
At that time, the PPQ is considered complete. Moderately tightened sampling for critical characteris-
tics could continue into Stage 3 CPV for four more lots until 75% coverage is reached, at which point
routine sampling begins.

8.1.3 Within and Between Lot Normal Tolerance Intervals

Statistical tolerance intervals are commonly used in validation. For example, a capping process may
have a validation criterion of “demonstrate with 90% confidence that at least 99% of the removal
torques for the lot are within specification limits.” Tables of normal tolerance interval factors for
variables data are widely available and also implemented in statistical software. Specialized software
is available to optimally calculate the desired confidence statement. Normal tolerance intervals for
the total process variation over time are more complicated; they include both within- and between-
lot variation. Standard normal tolerance interval factors assume that there is only one population in
the data. However, most PPQs contain multiple populations, since each lot is a separate population.

If there are no significant differences between the lots, the simplest way to deal with multiple lots is to
combine the data. ANOVA may be used to compare lot means; within-lot variation may be compared
with the Levene/Brown-Forsyth, Bartlett, Cochran, or Fmax tests (54-57). An omnibus test may also
be used. If there are no significant differences between lots or if the between-lot variance component
is not statistically significant, the standard normal tolerance interval for the combined data may be
used. The sample size per lot and number of lots should be statistically determined to have adequate
power to detect any between-lot variation.

Example: The specification for cap removal torque for a small volume parenteral (SVP) product is 8.0-12.0
inch-pounds. Limited data from Stage 1 showed a standard deviation of about 0.5. The production AQL
(Acceptance Quality Limit) for removal torque is 1.0%. The acceptance criterion for the PPQ is to show with
90% confidence that at least 99% (1 minus the AQL) of the cap torques are within specifications.

Three lots are included in the PPQ to evaluate the within- and between-lot variations. A sample size of 30
units per PPQ lot was tested to detect between-lot variation as large as the within-lot variation with 90%
confidence (58). Samples were tested from throughout each of the three lots, and the acceptance criteria for
each lot was met. An I/MR SPC chart indicated that the process was in control during each lot. Normal-
ity tests for each lot did not indicate significant non-normality. Since ANOVA and Levene’s test showed no
significant difference between the three lots, the data were combined. The 90 test results had a mean of 9.59
and standard deviation of 0.51.

A 90% confidence normal tolerance interval for 99% of the population is 9.59 ± 2.872 x 0.51 = (8.13,
11.05). This interval is within the specification limits of 8.0-12.0. Thus, the PPQ has shown with 90%
confidence that at least 99% of torque results are expected to meet specifications.

If there are statistically significant differences between lots, the tolerance interval should be construct-
ed with more advanced methods that take the between-lot variance component into account (56,57).

8.1.4 Statistical Process Control Charts (45)

Most SPC references suggest obtaining data from 20-30 time periods before calculating control limits
to assess whether the process is in control. Samples could be taken at 30 time periods across three or

82 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

more lots. For three lots, 10 sets of “rational subgroup” samples could be selected from each lot. The
SPC chart limits are then calculated and the process assessed for statistical control. The number of
lots to use can be based on the power of the SPC chart to detect undesirable between-lot variation.

A potential problem with the use of SPC charts, such as Xbar/S charts plotted across lots, is that they
define a process as being in statistical control if there is no underlying lot-to-lot variation (Figure
6.2.2.1-1). This is often not the case, and some lot-to-lot variation is typical and expected, especially
for lot means. In these cases, an I/MR chart for the mean and/or standard deviation or three-way
between/within chart can be used to detect out-of-statistical-control between-lot variation.

If there is only one test result per lot, such as lot assay or pH of a tank of solution, the 20-30 time peri-
ods become 20-30 lots. This is seldom feasible for PPQ. An alternative is to select a smaller number of
lots, perhaps 5-10, and construct a preliminary I/MR control chart. If it shows an in-control process,
the PPQ would be complete and the control chart extended into Stage 3 to verify longer term statisti-
cal control during CPV.

8.1.5 Ppk , Cpk Process Capability Metrics (59)

Ppk (see Figure 6.2.2.1.3-1) is the most common statistic used to assess long-term process capability.
Acceptable values of Ppk depend on the criticality of the characteristic, but 1.0 and 1.33 are commonly
used. Smaller or larger values may be used depending on the risk factors involved. The Ppk acceptance
criterion may be based on a point estimate or a one-sided lower confidence interval. If there is signifi-
cant between-lot variation, caution should be exercised in using confidence intervals for Ppk calculated
by statistical software. Most statistical software programs do not take between-lot variation into ac-
count, and may provide optimistic confidence intervals that are too narrow.

Example: Fill volume specification limits for a small-volume parenteral product are 98-102. PPQ acceptance
criteria are that each lot’s Ppk≥1.0; also, that the overall process Ppk is ≥1.0 with 95% confidence. To detect a
between-lot standard deviation that is half of the within-lot standard deviation with 90% confidence, 33
units will need to be tested from across each of five PPQ lots.

The data from the five lots were analyzed by control charts, histograms, normality tests, Levene’s test, and
ANOVA. These analyses indicated that the data from the five lots could be combined. Each of the five lots’
Ppks were > 1.0. The calculated Ppk from the combined data was 1.14, with a lower 95% confidence interval
of 1.03. Since each lot met its Ppk requirement and the lower confidence interval for the overall process, Ppk
was above the acceptance limit of 1.0. Thus, the PPQ acceptance criteria were met.

An alternative to calculating a parametric confidence interval for Ppk is to require four or five lots in a
row to each meet the Ppk acceptance criteria. For example, four PPQ lots, each with Ppk ≥ 1.0, provides
over 90% confidence that the process median Ppk is ≥1.0. Five lots provide over 95% confidence.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 83

 8.1.6 Assure The Lot Conformance Rate is Above an Acceptable Rate
With Specified Confidence
This approach demonstrates that the percent of lot conformance is acceptable. It identifies unaccept-
able variation due to either common or special causes. Table 8.1.6-1 shows the required number of
conforming lots. This method is sometimes called “confidence for reliability.”
Table 8.1.6-1 Number of lots to demonstrate confidence for lot conformance rate
Conformance
Confidence Accept # n
Rate
90% 0 7
50% 95% 0 14
99% 0 69
90% 0 22
90% 95% 0 45
99% 0 230
90% 0 29
95% 95% 0 59
99% 0 299

Example: To demonstrate the process is acceptable, the PPQ acceptance criterion will be to show with 90%
confidence that the process lot conformance rate (the lot pass rate) is at least 90%. A total of 22 passing lots
in a row will demonstrate this.

Requiring such a large number of lots during PPQ to reach 90% or 95% confidence is difficult. An
alternative is to use 50% confidence in PPQ and monitor the process further during CPV to reach the
final desired confidence. Crossing the 50% confidence threshold is the point at which it is more likely
that the selected lot conformance rate is met. For the example above, once 7 passing lots are reached,
it is more likely that the conformance rate is greater than 90% rather than less than 90%, and the
PPQ could be considered complete. An additional 15 lots during early CPV would reach the required
22 lots. This approach may be particularly useful when there are many quality attributes to assess.
Rather than determine the number of lots required separately for each attribute, the PPQ stage is
complete when all attributes pass for the required number of lots.

8.1.7 Wald Sequential Probability Ratio

Wald’s sequential probability ratio (SPR) test can be used to determine when a suitable number of
PPQ lots have been made for decision-making. No fixed number is specified in advance; the PPQ
continues until either the failure or pass decision line is crossed. The SPR test is most often used for
percent of passing units or lots, but it can also be used for other statistics as well as for CPV to deter-
mine when to revert to normal sampling. An example is shown in Figure 8.1.7-1.

Example: A 5% lot failure rate is considered minimally acceptable, while a 25% lot failure rate is not ac-
ceptable. The SPR decision chart below was made using a=0.05, b=0.2, p1=0.05, p2=0.25. The failure
decision line was crossed at lot 7; the PPQ failed due to too many lot(3 in 7) failures.

84 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

Figure 8.1.7-1 Wald’s Sequential Probability Ratio Example

5
FAIL
4
Lots Failed

X
3
X X
2
X X X
1
X PASS
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Lot Number
8.1.8 Narrow Limit Gauging
Narrow limit gauging can be used to reduce the sample size or number of lots required in PPQ. The
basic idea is to use narrowed pseudo-specification limits during PPQ to obtain more statistical power.
An example is a case in which the assay specification for an active ingredient in a solution is 95–105, and
only one assay result is determined per lot. If five lots all are within narrowed limits of 97.5–102.5, this
gives the same confidence as a larger number of lots being within the unadjusted 95-105 specification
limits in detecting the lot nonconformance rate (see Farnuma and Stantona for calculation details) (57).

One form of narrow limit gauging is called PRE-Control. It is often used as a QC procedure for fill
volume. If 5 units in a row fall in the middle 50% of the specification limit, then the lot is qualified for
startup. This concept can be extended to quality characteristics (e.g., lot assay, pH) ) where there is one
result per lot. For an assay specification of 95–105 for an active ingredient, the PRE-Control narrow
limits would be 97.5–102.5. If five PPQ lots in row meet the 50% narrow limits, then the PPQ is com-
plete. Note that the narrow limits are not used to determine lot acceptance, but only to determine
whether the PPQ acceptance criteria are met.

8.1.9 Demonstrate Between-Lot Variation is

Less Than Within-Lot Variation (Anova) (60)
Under the classical one-factor random effects variance components model, the total process standard
deviation is calculated as . Due to the squaring under the radical, if σb < σw , the im-
pact of the between-lot variation σb on the total process variation decreases rapidly the smaller it is
compared to the within-lot variation σw. Table 8.1.9-1 shows this impact.
Table 8.1.9-1 Effect of between-lot variation on the total process variance

Total Between as
Within Between Total
Variance % of Total
σw σb σt
σt2 σb2 / σt2
1.00 2.00 2.24 5.00 80%
1.00 1.50 1.80 3.25 69%
1.00 1.00 1.41 2.00 50%
1.00 0.75 1.25 1.56 36%
1.00 0.50 1.12 1.25 20%
1.00 0.25 1.03 1.06 6%

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 85

 If the between-lot variation (σb) is half (50%) of the within-lot variation (σw), the former only ac-
counts for 20% of the total process variance. Reasonable PPQ acceptance criteria for between-lot
variation would typically be 75%, 50%, or 25% of the within-lot variation. The sample size within
lots and number of lots required may be determined by the statistical power to detect the differences
of interest. Acceptance criteria could be based on point estimates of the variance components or
confidence intervals.

8.1.10 Sample Size

Table 8.1.10-1 shows the sample n size required to estimate a standard deviation to within a specified
% of its true value with 90% and 95% confidence (45,54). This method does not require a previous
estimate or reference sigma since the error is expressed in relative rather than absolute terms.
Table 8.1.10-1 Sample Size to estimate a standard deviation to within ±X% of true value
Confidence ± % relative error n
90% 20% 35
90% 25% 23
90% 33% 14
95% 20% 49
95% 25% 32
95% 33% 18

Table 8.1.10-1 indicates that a minimum of 32 lots are required for the estimated between-lot stan-
dard deviation σb to be within ±25% of its true value σb with 95% confidence. Since the table assumes
the lot means are estimated exactly, more than 32 lots may be required if the sample size per lot is
small or there is substantial within-lot variation. Table 8.1.10-1 shows the difficulty in estimating a
standard deviation: large sample sizes are required to obtain precise estimates. Again, a phased ap-
proach could be used where the PPQ is based on five lots, and additional data is collected during CPV
to obtain a more precise estimate.

8.1.11 Demonstrate the Between-Lot Standard

Deviation σb ≤ Acceptable Value X
It is generally easy to test enough samples to estimate the within-lot standard deviation σw with
reasonable precision. Also, estimates of the within-lot variation are often available before PPQ
from Stage 1 data or other similar production processes. The total process standard deviation is
. The PPQ acceptance criterion for σb may be selected to show that the total process
variation is acceptable (e.g., 3-sigma capable or other desirable value).

8.1.12 Demonstrating equivalence between lots

Differences between PPQ lots can be statistically detectable, but they might be small enough to con-
sider the lots equivalent. To use this method, an equivalence test using multiple TOST (two one-sided
tests) or other extension of the equivalence concept may be used. The number of lots required would
be based on the statistical power for the equivalence procedure chosen (52).

86 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60

8.2 Appendix 2: Types of Control Charts
8.2.1 Control Charts for Variables Data
Commonly used control charts for variables data are:
• Xbar/R or Xbar/S “Shewhart” chart using ±3-sigma limits
• Individuals / Moving Range (I/MR) chart
• Exponentially weighted moving average (EWMA) chart
• Cumulative sum (CUSUM) chart
• Moving average chart
• Each chart has strengths and weaknesses. The Xbar/R and Xbar/S charts are good at picking up
large process shifts quickly. They are also easy to apply when sample sizes vary, which may be
important when the variance is based on the batch size. The Individuals /Moving Range chart
can be used when there is only one test result per time period, such as % theoretical yield, and
also when between-lot variation is typical and expected. The EWMA, cumulative sum, and mov-
ing average charts are good at detecting smaller shifts in a process. Xbar/R and Xbar/S charts can
detect smaller shifts almost as well as these more complex charts by adding runs rules in addition
to the one point beyond 3-sigma limits rule (61). Other less commonly used charts for variables
data are multivariate charts, three-way I/MR between-within charts, zone charts, short-run charts,
ARIMA (auto regressive integrated moving average) time-series analysis charts, tool wear charts,
acceptance charts, median charts, and others.

• One problem in using control charts during validation is that most only determine if the process is
in classical statistical control (Figure 6.2.2-1). If the process has the more complex form of statisti-
cal control with intra- and inter-lot variation (Figure 6.2.2-3), most charts will incorrectly indicate
that the process is “out of statistical control.” This problem is often exacerbated during validation
because of the large sample sizes used. This increases the statistical power of the control chart in
detecting small differences between subgroups or lots. The simplest solution for this problem is to
use an I/MR chart for the subgroup means and standard deviations, which will take the between-
lot variation into account. Other solutions are to use a three-way I/MR/S between/within control
charts (62) or use separate control charts for each lot and do not control chart across lots.

• Another problem in determining if a process is in control relates to the overall probability of find-
ing one or more subgroups beyond the 3-sigma limits. The false alarm probability of a subgroup
exceeding the 3-sigma limits or failing a runs test (if several of the commonly used runs tests are
used) is often as high as 1%. If a PPQ takes samples from 32 time periods from each of four PPQ
lots, a total of 4x32 = 128 subgroups will be plotted on each control chart. If Xbar/R charts are
used, there will be a total of 256 plotted values, each with as much as a 1% chance of indicating
“out of statistical control” even when the process is actually in statistical control. The probability
that all 256 plotted values will show a process to be in control is only 0.99256 = 0.08 = 8%. There is
only about an 8% chance that zero “out of statistical control” events will occur even though the
process is in statistical control. If acceptance criteria are used for control charts during PPQ, they
should take the number of sampled points into account and allow a statistically determined small
number of values to exceed the control limits or fail one of the runs tests to control the overall
false alarm rate.

Technical Report No. 60 © 2013 Parenteral Drug Association, Inc. 87

 8.2.2 Control Charts For Attributes Data
The most commonly used control charts for attributes data are:
• p-chart for fraction or percent nonconforming (binomial data)
• np-chart for number of units nonconforming (binomial data)
• c-chart for number of defects (Poisson data)
• u-chart for average number of defects (Poisson data)

In addition, the Individuals/Moving Range chart can be used for attributes data if most of the plotted
values are not 0. For very large sample sizes per period, say greater than 500 or 1,000, theI/MR chart is
usually preferred to the corresponding attributes chart. This is because the assumption of a binomial
or Poisson distribution that attributes charts use is often violated for large sample sizes and long time
periods. For example, if an automatic metal detector is used to inspect 100% of all tablets, an I/MR
control chart would be preferred to a p-chart for plotting the percent of tablets rejected across lots if
1,000 or more tablets are inspected per lot.

8.2.3 Performance Of Control Charts: Average Run Length (ARL)

The ability of a control chart to detect a shift in the process is measured by the average number of
sample periods before the shift is detected. This is called the Average Run Length (ARL). When the
process has not changed (a shift of 0), it is desirable to have a large ARL to keep the false alarm rate
low. However, if the process has shifted, it is desirable to have a small ARL to detect the change quickly.
ARL is most important in CPV, when it is desirable to detect a significant process change within a short
period of time. The ARL can be used to help select the sample size and frequency of sampling. Tables
of ARLs comparing different control chart types are available and can also be calculated by software.

88 © 2013 Parenteral Drug Association, Inc. Technical Report No. 60