Proposed Validation Standard VS-1: Nonaseptic Pharmaceutical Processes -- Introduction and Preamble

Kenneth G. Chapman, Drumbeat Dimensions Inc.
Gamal Amer, Validation and Process Associates Inc.
Cheryl Boyce, Boyce Regulatory & Quality Consulting
George Brower, Analex Corp.
Cindy Green, Northwest Regulatory Support
William E. Hall, Hall & Associates
Dan Harpaz, PCI, Pharmachem International
Barbara Mullendore, AstraZeneca

This proposed validation standard is presented in two parts: Introduction and Preamble (this article) and the actual VS-1 Standard.

Contents
Introduction
Regulations, Guidance Documents, and Standards
Style of the IVT Validation Standard
Contents of the Validation Standard
Future Validation Standards by IVT/SC
Preamble
Frequently Encountered Questions
The Validation Master Plan (VMP)
A Brief History and Status of Process Validation Guidelines
Process Validation Life Cycle (Time Line)
The Three Lot Controversy
Critical Operating Parameters and the Pyramid
Establishing the End Values
Use of Statistics in Process Validation
IQ, OQ, and PQ
Nonresearch-Based Firms
Legacy Processes

Introduction (Back to Top)
This Preamble introduces the first in a series of new proposed validation standards by the Institute of Validation Technology Standards Committee (IVT/SC). The purpose of the Preamble is to explain the rationale behind the proposed standard.

The Process Validation Standard VS-1 itself is intended to help practitioners worldwide who develop, implement, control, and validate processes that produce APIs (Active Pharmaceutical Ingredients—a.k.a. drug substances) and drug products. All of the new validation standards will also be used by reviewers of manuscripts intended for publication in the Journal of Validation Technology (JVT.)

Readers are encouraged to offer comments, questions, and recommendations. Such feedback should be useful to the IVT/SC and JVT editors updating this document and in developing future standards. Technologies are continually changing, sometimes in ways that can influence the way validation is best conducted. Therefore, the IVT/SC plans to periodically update each validation standard, including its corresponding Preamble and reference list. IVT/SC also plans to make all standards electronically available in order to be dynamically responsive to changing industrial practices and regulatory requirements, and make it easier for readers to cut and paste the contents for their use.

A fundamental need the IVT/SC intends to meet with its new standards stems from the fact that most GMP regulations today call for numerous written procedures; for example, more than 100 different kinds of written procedures are required to comply with current Good Manufacturing Practice (cGMP) Regulations in the United States. Many firms find it helpful to issue written policies in order to coordinate and reduce the number and length of required Standard Operating Procedures (SOPs). Thus, the IVT Validation Standards format includes statements that can be excised and used directly or with minor editing in a firm's policies and SOPs.

Regulations, Guidance Documents, and Standards (Back to Top)
There are marked differences between regulations, guidance documents, and standards. In the United States, cGMP regulations evolve only through due process and are considered binding and legally substantive. Food and Drug Administration (FDA) guidance documents do not usually involve due process and are no more legally binding than are industrial guidance documents. (At one time, FDA Guidelines and Compliance Policy Guides were considered legally binding on FDA itself, but that ruling was reversed by FDA General Counsel in August 1990.) Rules Governing Medicinal Products in the European Community, counterpart of U.S. cGMPs, are also portrayed as nonbinding; however, such rules in the European Union (E.U.) and, within a few years, many FDA Guidance Documents in the U.S., although technically nonbinding, often become regarded as de facto law.

In a somewhat related area of software validation, professional organizations like the Institute of Electronic and Electrical Engineers (IEEE) have demonstrated the value of setting standards. The rapid evolution of computer technology in today's pharmaceutical operations calls for focus on validation of computer-related systems. As with process validation, numerous guidance documents, published articles, and even regulations have appeared in recent years on the subject of computer-related system validation, most of which rely on IEEE standards for the software portion of this important subject.

Style of the IVT Validation Standard (Back to Top)
Most pharmaceutical firms like to have lists of succinct, unambiguous, and explicit rules about quality assurance against which to audit. It is preferable for such rules to contain unambiguous, imperative verbs like "shall," "will," and "must," rather than passive verbs like "should," "may," and "can" to avoid interpretive debates with auditees, including suppliers and contract vendors. Standards usually satisfy this need, whereas guidance documents often do not. However, it is also important for those who are to follow the rules to have access to some kind of interpretive documents that accompany and explain the rules. FDA provides a preamble to each of its regulations for that purpose. Similarly, the IVT/SC plans to preface each validation standard with a Preamble like this one to explain the subjective rationale behind some of the more complex rules offered in the standard.

Validation has proven to be a complex subject, clear understanding of which depends largely on use of a common language by everyone concerned. Thus, the IVT/SC has attempted to provide a glossary that reflects the clearest and most accurate contemporary definitions possible. Moreover, each definition is developed in ways that will enable that term to have the same meaning when used in any IVT validation standard. With time and sustained effort, this approach is expected to help improve worldwide harmonization of validation terminology.

Contents of the Validation Standard (Back to Top)
Each IVT validation standard will include the following five sections after the Preamble and list of contemporary references:

I. Policy Statements—Standards that indicate what is required
II. Procedural Statements—Standards that describe how to meet requirements
III. Acronyms—Meaning of each acronym used in the document
IV. Glossary—Definition of key terms, which are highlighted and asterisked (*) when first used in the validation standard
V. Regulatory Excerpts—Regulatory language (United States, Australia, Canada, World Health Organization [WHO], Japan, and European Union) related to each Standard

Future Validation Standards by IVT/SC (Back to Top)
The IVT/SC authors hope to deliver several new sets of IVT validation standards for publication in the JVT over the next few years. Future standards under consideration include the following (not necessarily in the order of proposed publication):

• Validation of Analytical Test Methods
  
• Aseptic Pharmaceutical Process Validation
  
• Biopharmaceutical Process Validation
  
• Medical Device Process Validation
  
• Equipment Cleaning Validation
  
• Computer-Related System Validation
  
• Water Treatment System Validation
  
• Terminal Sterilization Validation

Preamble (Back to Top)
In recent years, several worldwide guidance documents and some regulations have been published that address pharmaceutical process validation. Not all such documents are harmonized and, in fact, some diametrically contradict each other. The Global Harmonization Task Force (GHTF) (ref. 1) appears to be making good progress, especially with regard to standardizing process validation for medical devices. Corresponding worldwide efforts to harmonize process validation for nonaseptic pharmaceuticals are at present less advanced, less consistent, and less comprehensive. In fact, significant confusion seems to exist about the subject. A primary objective of VS-1 is to help alleviate such confusion.

Frequently Encountered Questions (Back to Top)

• When does process validation begin, and when does it end?
  
• What is a Validation Master Plan (VMP)?
  
• What kind of validation data can be created in the lab or pilot plant, and when must data be created at commercial scale?
  
• How important are statistical methods, and when must they be employed?
  
• What are the purposes of Installation Qualification (IQ), Operational Qualification (OQ), and Performance Qualification (PQ)?
  
• When and why must three consecutive, commercial-size lots be run?
  
• How do requirements for existing (legacy) processes differ from those for new processes?
  
• What is revalidation, and when must it be conducted?

The Validation Master Plan (VMP) (Back to Top)
The Validation Master Plan (VMP) is a master document that begins with the initiation of any validation project and is regularly updated as needed, at least until the product becomes commercial. Although the VMP is specifically called for by most contemporary draft and approved validation guidelines, it has become a confusing term because two basic definitions exist; both are used by different (and sometimes even the same) regulatory officials. One definition calls for the VMP to be project-oriented, as in this standard, VS-1 (Sec. IV—Glossary). The other definition describes a more global document embracing a firm's overall validation philosophy.

Most pharmaceutical firms use policies and/or procedures (SOPs) to address such global matters individually. IVT/SC finds the second (global) VMP definition workable, but cumbersome and inefficient. To minimize confusion, a firm should clearly define its use of the term VMP (e.g., by written policy or SOP), while ensuring that global and project-related matters are both adequately covered in some way. For firms preferring the global VMP, a term such as Validation Project Plan can be used in place of the VMP defined in VS-1.

A Brief History and Status of Process Validation Guidelines (Back to Top)
FDA made available its first draft Guideline on Process Validation in March 1983, shortly after which the Pharmaceutical Manufacturers' Association (PMA, now PhRMA—Pharmaceutical Research and Manufacturers of America) appointed its Validation Advisory Committee (VAC) to respond. FDA's guideline (ref. 2) was not finalized until May 1987. During the intervening four years, three PMA Committees (VAC, the Deionized Water Committee, and the Computer System Validation Committee) each published an industrial validation guideline. (refs. 3,4,5) Prior to each publication, PMA/FDA meetings were held and draft documents discussed at length until regulatory and industrial experts reached basic agreement on all major issues. Nonetheless, in subsequent years, some misunderstandings and contradictory interpretations have evolved, the majority of which are based on semantics, rather than technical differences.

Two recent guidance documents, one by FDA (ref. 6) (Guidance for Industry—Manufacturing, Processing, or Holding Active Pharmaceutical Ingredients—Mar 1998), and the other by the European Agency for the Evaluation of Medicinal Products (EMA) (ref. 7) (Note for Guidance on Process Validation—30 Sep 1999), describe the importance of Research & Development (R&D) roles in process validation and offer a full life-cycle approach that begins with R&D. Two other draft guidance documents by Pharmaceutical Inspection Convention (PIC) (ref. 8) (Mar. 1999) and by the EC9 (Validation Master Plan Design Qualification, Installation, and Operational Qualification, Non-Sterile Process Validation, Cleaning Validation—30 Oct. 1999), one a minor modification of the other, seem out of harmony with the life-cycle approach, suggesting that "…three consecutive batches/runs within the finally agreed upon parameters… would constitute a proper validation of the process." The same two documents also contain certain terminology that is outdated, contradictory, or otherwise in need of global harmonization.

Process Validation Life Cycle (Time Line) (Back to Top)
VS-1 Policy Statement 1.3 lists 12 steps in the process validation life cycle, starting with definition of the product and the process. Critical pharmaceutical product specifications are determined chiefly by safety (animal and human) studies. Critical process operating parameters are a function of process capability and are determined by process development, which includes process validation.

A significant semantics issue today involves two different perceptions of the scope and definition of the term process validation. The leading perception, to which IVT/SC subscribes, is that process validation embraces an entire life cycle, beginning in R&D, including IQ, OQ, and PQ, and ending only when the related product is no longer commercial. The other, now outdated, perception is that process validation starts after IQ and OQ; in fact, some authors (refs. 8,9) equate process validation exclusively with PQ.
Unfortunately, the three PMA guidelines (refs. 3,4,5) of the early 1980s failed to recognize the future semantic problem by portraying validation as something following IQ and OQ. FDA had included performance qualification in its 1983 draft guideline, but associated the term with medical devices, and not with pharmaceutical processes.

In the early 1990s, PDA's Committee on the Validation of Computer-Related Systems recognized the importance of the life cycle, since good software design and development are essential to ultimate system validation and must start at the beginning of the life cycle. PDA defined computer-related system validation as all-embracing, with PQ representing a late stage of the validation life cycle. The GAMP Forum (refs. 10,11) soon thereafter adopted a similar approach.

Practitioners and regulators have learned that, just as "quality must be built into a product" (i.e., "it cannot be tested in") robustness also has to be built into a process. The analogy between process validation and validation of computer-related systems is apt; in both cases, considerable interaction by multiple disciplines throughout the life cycle, including effective technology transfer from R&D to Production and Quality Control, have proven essential.

The Three Lot Controversy (Back to Top)
During 1983–1984, representatives of FDA and Industry debated at length over the value of positioning three consecutive, commercial-sized lots as pivotal evidence of process validation. Industry agreed that FDA's argument for three lots might be suitable for medical devices, but argued successfully that it was not appropriate for pharmaceutical processes for several reasons:

1. Unnecessarily costly and risky to perform prior to regulatory submission;
   
2. Limited statistical benefit from three lots; and
   
3. Establishing critical process parameter ranges and probable adverse consequences of exceeding range limits (ref. 12) represents a better investment of resources and contributes more to process robustness and reliability, while the three-lot requirement can detract from such efforts.

In 1990, when FDA launched its Pre-Approval Inspections (PAI) program, the three-lot issue again arose. PAI's chief architects (Richard Davis and Joseph Phillips, FDA, Newark District Directors) announced they would require evidence of three consecutive, successful lots of commercial size prior to shipment of a new product across state lines, as "final" evidence of process validation, even when the firm had already received its New Drug Application (NDA) Approvable Letter.

This time, Industry did not protest the requirement. Several reasons made the requirement logical:

• Three commercial lots add some degree of assurance that the process works and at least a limited indication of reproducibility.
  
• Three lots can be made in a practical period of time, compared with the number of lots required to gather statistical evidence of reproducibility.
  
• The overall approach forces focus of validation emphasis on process development measures that occur earlier in the life cycle and, thus, do not jeopardize market launch timing.

Since 1990, most firms have found the predistribution three-lot requirement practical and useful. Some have made the mistake of believing that critical parameters should be varied during the three runs in order to develop new validation evidence, usually of the kind that can be developed in the laboratory or pilot plant more economically and with less risk of failure.

Critical Operating Parameters and the Pyramid (Back to Top)
Another concept that appears ambiguous in most current guidance documents is whether a process should be validated against regulatory or "internal" process parameters and product specification limits.

It is important that the proven-acceptable, regulatory, and operating ranges all be recognized and considered when writing validation protocols. Many firms also use control ranges that lie between operating and regulatory ranges for added insurance against, and control over, minor plant deviations. Regulatory range limits represent those that a firm includes in its registration (e.g., NDA). The firm's basic commitment is that product safety and efficacy will be ensured when all regulatory limits are met. Regulatory range limits must fall within the upper and lower edges-of-failure. In order to define edges-of-failure, it is essential to identify what the probable adverse consequences are of exceeding the edges-of-failure in each direction. For example, exceeding the upper edge-of-failure for tablet hardness might cause an unacceptable dissolution rate, while exceeding the lower edge-of-failure could lead to friability problems. Overheating an API solution may cause predictable degradation reactions, while underheating could cause premature crystallization or failure to complete a desired reaction.

Many firms employ more than one range of internal limits, such as control ranges for quality monitoring and approvals, as well as the usual, and somewhat tighter, operating ranges for shop-floor directions. Each internal range must lie within the corresponding regulatory range for compliance. Control ranges are often found to be convenient, especially for in-process control test limits, but need not be regarded as essential.

In its initial 1983 draft guideline, FDA proposed that process validation should be based on FDA's definition of "worst-case," which, at that time, extended from one edge-of-failure value to the other. The industry objected to the proposal and pointed out in a 1984 article (ref. 12) that it is unnecessary to have either edge-of-failure value available, so long as one can establish a Proven-Acceptable Range (PAR) that embraces the regulatory range. In its final 1987 guideline, FDA redefined worst-case to equate with the operating range, a move that facilitated future process validation planning.

Establishing the End Values (Back to Top)
The major objective of process validation is to ensure development of a robust process that will produce a product having the intended quality every time the process is run. The first five steps of the Process Validation Life Cycle (VS-1 Policy 1.3) involve establishing end values of regulatory and internal ranges.

Although not absolutely essential, it is also useful to identify edges-of-failure, since the difference between edge-of-failure ranges and regulatory ranges helps determine sensitivity of the process as to causing product rejections. Edge-of-failure data, as well as all other limit values, can frequently be determined in the laboratory or pilot plant (using aliquot samples) long before the process is fully scaled up.

For APIs, reaction kinetics are often used to predict thermal and pH end values. Other studies that help ensure robustness can be created in the early stages of API process development; for example:

1. Determining conditions under which API polymorphs, isomers, hydrates, solvates, and degradation products might form (also important for process patent reasons);
   
2. Isotherms of pH and temperature versus API solubilities, degradation rates, and other variables; and
   
3. Similar studies involving major impurities specific to the API process.

In the case of drug products, developmental pharmaceutics, which include physicochemical profiles and excipient interaction studies, similarly provide information that is needed to determine edges-of-failure and reliable end values. Stability studies and behavior of various lots of clinical supplies contribute further insight to drug product end value design.

Final determination or confirmation of operating ranges for some unit operations, such as blending, will require exploratory studies in larger equipment. In the case of blending, such studies should be preceded by particle size measurements and crystal morphology studies in the laboratory, since the tendency to blend or deblend is often predictable. Blending also represents a case where commercial-scale experiments can usually be run at low risk, for example, to optimize rotational speed and time periods by testing aliquot samples taken at various time intervals.

Use of Statistics in Process Validation (Back to Top)
Some current publications address process validation from an almost exclusively statistical approach. The effect of such articles on non-statisticians occasionally ranges from dismay to panic and, unfortunately, drives them away instead of toward use of statistics. Statistical Process Control (SPC) can be especially valuable when applied to process validation, both before and after the validated process enters commercial use. By statistically analyzing critical process parameter data throughout a batch or continuous process, SPC provides the opportunity to predict problems (trend analysis) and even take corrective action (trend control) before the problems occur. Yet today relatively few firms actually appear to be implementing SPC universally across all processing. Diagnosing reasons for this apparent anomaly is beyond the scope of this Preamble. Nonetheless, all firms need to be aware of where statistical tools are and are not needed in process validation work and to have statistical expertise available, regardless of whether SPC itself is broadly used.

Statistical analysis is routine and taken for granted in most laboratory work, including the validation and implementation of analytical test methodology, and in the design of most sampling plans. A question that frequently arises is when statistical tools need to be applied to determine adequacy of operating and regulatory ranges (i.e., process capability.) If a proven-acceptable range exists for a given parameter that is 20% wider than the regulatory range, and if the regulatory range is 20% wider than the operating range, the process is likely to be robust enough to obviate need for statistical analysis for the given parameter. Conversely, if the same ranges appear to be within 2% of each other, the process may or may not require more development, but statistical analyses should certainly be considered. Between those two extremes, judgment is needed of the kind that can usually be provided only by statistical experts.

Another common situation in which statistical analyses may be essential occurs when multiple critical process operating parameters display interactive effects and none of the parameters can be analyzed in isolation. Factorial design experiments may be needed, design and interpretation of end results of which are likely to demand statistical analyses.
The bottom line is that most process validation teams should include or have access to a statistics expert. Because use of SPC offers many opportunities to improve costs and quality through trend analyses and trend control, SPC is highly recommended as a measure to be included in a process validation program.

IQ, OQ, and PQ (Back to Top)
Rather widespread confusion seems to have accompanied the great variety of definitions that have evolved for the three "qualification" terms, Installation Qualification, Operational Qualification, and Performance Qualification. Here are a few basics that may help clarify the terms (all definitions can be found in the VS-1 Section IV Glossary):

• Systems and processes are validated; equipment and materials are qualified; persons are trained and qualified.
  
• IQ is intended to ensure that all critical equipment has been purchased and installed correctly.
  
• OQ is intended to ensure that all critical equipment works as intended for the process in which it is to be used.
  
• It is not unusual for some IQ and OQ activities to overlap, an occurrence that presents no problem as long as it is recognized and addressed systematically.
  
• IQ and OQ data records must be adequate to support ongoing and future change control and revalidation requirements.

PQ is intended to demonstrate that the process will function correctly in its normal operating environment. The demonstration may involve pilot lots, commercial-scale lots, or carefully designed simulations of either. PQ protocols often involve individual modules (i.e., steps, unit operations) of a new process prior to pilot or commercial scale-up of the full process. When a given critical process parameter cannot be simulated at less than commercial scale, all other process parameters are often established first to avoid potential interference with the first commercial batch that must involve the sensitive parameter. The three commercial lots required to authorize future distribution can theoretically represent the final PQ experiments; however, there is no limit to the number of subsequent commercial lots that can continue to be considered part of the PQ step in the validation life cycle.

Instrument calibration is an example of an activity that overlaps IQ, OQ, and other steps in the life cycle. In validation work, instruments frequently need more extensive calibration (e.g., concerning linearity) than in subsequent process control applications. The step in which the records are included is unimportant as long as the records are available and consistently documented.

Nonresearch-Based Firms (Back to Top)
The question may be raised, "How do nonresearch-based firms, such as those that produce off-patent generic products rather than proprietary drugs, develop basic developmental pharmaceutics, such as physicochemical profiles?" IVT/SC's response is that such firms have the same responsibility for understanding and providing robust, reliable processes as do the larger, research-based firms. By the time proprietary products are off patent, they have often become compendial (e.g., United States Pharmacopoeia [USP], BP), with much of the required basic information readily available. However, minor differences in processing equipment or materials can affect process robustness. Contract development work may be necessary if the firm lacks process development resources.

Legacy Processes (Back to Top)
A validation life cycle for an established (legacy), or altered process (revalidation) will be the same as for a new process, except for those steps already adequately completed and directly applicable. Legacy processes should have created numerous batch records which, with appropriate retrospective statistical review, can be translated into a series of PQ experiments, provided that a well-defined process and adequate change control measures are in effect.

References

1. Global Harmonization Task Force "Final Document - Process Validation Guidance for Medical Device Manufacturers,"—GHTF/FD-99-10 Feb. 1999 [Finalized 29 Jun 1999].
   
2. FDA, "Guideline on General Principles of Process Validation" Final. May 1987.
   
3. PMA Validation Advisory Committee, "Process Validation Concepts for Drug Products," Pharmaceutical Technology, 9 (9), 78-82. Sept. 1985.
   
4. PMA Deionized Water Committee, "Validation and Control Concepts for Water Treatment Systems," Pharmaceutical Technology, 9 (11), 50-56. Nov. 1985.
   
5. PMA Computer Systems Validation Committee, "Validation Concepts for Computer Systems Used in the Manufacture of Drug Products," Pharmaceutical Technology, 10 (5), 24-34. May 1986.
   
6. FDA CDER, CBER, CVM "Guidance for Industry—Manufacturing, Processing, or Holding Active Pharmaceutical Ingredients," March 1998.
   
7. The European Agency for the Evaluation of Medicinal Products, CPMP, CVMP "Note for Guidance on Process Validation," DRAFT 30 September 1999.
   
8. Pharmaceutical Inspection Convention, "Recommendations on Validation Master Plan, Installation and Operational Qualification, Non-Sterile Process Validation, Cleaning Validation"– PR 1/99-1 01 March 1999.
   
9. European Commission, "Validation Master Plan Design Qualification, Installation, and Operational Qualification, Non-Sterile Process Validation, Cleaning Validation," [E-3 D(99)]. 30 Oct. 1999.
   
10. GAMP Forum, "Supplier Guide, Version 2.0," May 1996.
    
11. GAMP Forum, "GAMP Guide for Validation of Automated Systems in Pharmaceutical Manufacture, Version 3.0, Volumes One and Two," March 1998.
    
12. Chapman, K. G., "The PAR Approach to Process Validation," Pharmaceutical Technology, 8 (12), 22-36, Dec. 1984.

Suggested Reading

1. PhARMA QC Section, Bulk Pharmaceuticals Committee, "Concepts for the Process Validation of Bulk Pharmaceutical Chemicals," Pharmaceutical Technology, Europe, 6 (1), 37-42, Jan. 1994.
   
2. PDA Committee on Validation of Computer-Related Systems, "Validation of Computer-Related Systems" Technical Report No. 18, Supplement to PDA Journal of Pharmaceutical Science and Technology, Vol. 49 (S1) 14 Oct. 1994.
   
3. FDA, "Current Good Manufacturing Practice: Amendment of Certain Requirements for Finished Pharmaceuticals"—21 CFR Parts 210 and 211 pages 20103 to 20115. Federal Regulation, May 1997.
   
4. ABPI, Draft PIC GMP Guide for Active Pharmaceutical Ingredients," 7 Oct. 1997.
   
5. FDA, "Guidance for Industry; BACPAC I: Intermediates in Drug Substance Synthesis; Bulk Actives Post-approval Changes: Chemistry, Manufacturing, and Controls Documentation," DRAFT GUIDANCE, Whole Document 1-24. 17 Nov. 1998.

This article originally appeared in the Journal of Validation Technology.

For more information: Terri Kulesa, Institute of Validation Technology, 200 Business Park Way, Suite F, Royal Palm Beach, Florida 33411. Tel: 561-790-2025. Fax: 561-790-2065.