IVTJVT1107.qxd 10/8/07 3:27 PM Page 36 The Future of Process Validation - Going Back to Basics The New FDA Perspective as a Refocus on the Original Definition of Process Validation BY MIGUEL MONTALVO INTRODUCTION The concept of Process Validation (PV) has evolved during the last 30 years and will continue to evolve in the future as our industry matures and the regulatory agencies move toward implementation of quality systems, Quality by Design (QbD), and global harmonization. The area of Process Validation was near and dear to Mr. Ken Chapman's interests and the perspectives presented in this article agree with his own point of view in relation to the topic. Just to give you an historical perspective on the evolution of PV thus far: During the 1980's - companies (especially sterile and parenteral manufacturers) started to perform Process Validation without guidance from the Food and Drug Administration (FDA). During this time, companies developed aseptic processing and sterilization validations. 1987 - PV Guideline document from FDA is released 3. Solid dosage and oral products manufacturers start implementation of guideline. There is a great deal of confusion created by the terms worst case versus set-up at nominal value with a range of variability (parameter to remain within standard operating procedure (SOP) limits) during process validation. During his talks to industry, Ken Chapman also stressed the negative effects resulting from this confusion. Development areas were not documenting their processes adequately - Process validation was forced into challenging processes at worst cases, setting parameters at process limits, and even risking failure during validation. The focus was to comply with documentation requirements. FDA Compliance Guide 7132c.08 Sec. 490.100 The FDA has been working, during the last few years, on the next step in the evolution of the concept of process validation. One of the documents that they created which indicates the new direction was the Compliance Policy Guide (CPG) 7132c.08 Sec. 490.100 Process Validation Requirements for Drug Products and Active Pharmaceutical Ingredients subject to Pre-Market Approval released in March 2004. In this document, the agency includes very clear statements that describe where and how they would like the concept of PV to evolve. Some of the most important indications included: 36 Journal of Validation Technology
IVTJVT1107.qxd 10/8/07 3:27 PM Page 37 Proof of Validation: Proof of validation is obtained through rational experimental design and the evaluation of data, preferably beginning from the process development phase continuing through the commercial production phase. Prior to Commercial Distribution: Manufacturer has accumulated enough data and knowledge about the commercial process to support post-approval distribution. The data normally includes: Satisfactory product and process development Scale-up studies Equipment and system qualification Successful completion of initial conformance (validation) batches Prior to the manufacture of the conformance batches, the manufacturer should have identified and controlled all critical sources of variability. Conformance batches (sometimes referred to as validation batches and demonstration batches) are prepared to demonstrate that, under normal conditions and defined ranges of operating parameters, the commercial scale process appears to make acceptable product. It is clear that the Agency's intent is to change the concept of PV from a testing and documentation focus to a design, confirm, control, and monitoring focus. In the Compliance Policy Guide (CPG), FDA also mentions their objective of revising the PV Guideline from 1987. CGMPs for the 21st Century A second document from the FDA, the Pharmaceutical current Good Manufacturing Practices (cgmps) for the 21st Century - A Risk Based Approach Final Report from September 2004, enforces their intent when referring to the revised CPG as stressing the importance of rational experimental design and ongoing evaluation of data. Also, Achieving and maintaining a state of control for a process begins at the process development phase and continues throughout the commercial phase of a product's lifecycle articulating more clearly the role of conformance batches in the product lifecycle. The document clearly signals that a focus on three full-scale production batches would fail to recognize the complete story on validation. In agreement with these concepts, Chapman frequently stressed the need to use the data and process information collected during development to minimize or eliminate testing at the commercial scale. Quality Systems Approach to Drug GMPs In addition to these documents, the FDA has implemented the Quality Systems Approach to Drug GMPs Guideline from September 2006. The concepts in this guideline are very similar to the ones described earlier - Quality by Design, control of the critical parameters, and effective process monitoring. Then, the questions become: Why is this different from what we are doing now? Are these totally new concepts? In reality, they are not new. Ken Chapman proposed similar concepts in 1984 when he introduced the Proven Acceptable Range (PAR) approach to process validation. In his approach, Chapman emphasized use of knowledge gathered during development and the use of proven acceptable ranges, which are similar to today's concept of Design Space. Our fault was focusing on the testing and documentation instead of the process design, understanding, and effective monitoring and control. Traditional Approach Let us try to describe the actual, traditional process validation concepts: Emphasis on replication at full-scale - Validation lots viewed as evidence of process reproducibility Significant milestone - just prior to commercial launch November 2007 Volume 14, Number 1 37
IVTJVT1107.qxd 10/8/07 3:27 PM Page 38 Over time, validation became centered on documentation instead of on ensuring quality Validation protocols, reports, and related documentation proliferated - still, some processes do not work! Lost sight of the goal which should have been to demonstrate the process's reliability Failure to understand our processes - Why do we agree on validating them? False assumptions: do not change anything and everything will remain the same. The fact is neither ingredients nor processing conditions remain the same with time. Quality approach relies on discarding bad lots to keep quality high. This is inefficient and costly. This traditional approach encourages fixed component and processing conditions (change is bad), discourages fuller process understanding, and a root-cause analysis on a failure or an out of specification (OOS) situation is more difficult because the process is not well understood. The other problem with the traditional approach was the total focus on getting the PV (three batches) completed (documentation package). The traditional approach rarely asked the question: Are we monitoring, controlling, and assessing the reliability of our processes after the initial PV? The fact is that the commercial operation phase generates large volumes of process knowledge/data. The question is - Do we use that data in the most effective manner? Do we analyze the data using scientifically-based tools to provide indications on how our process is behaving? Do we determine based on the data, how effective are our controls and our overall process design? BACK TO BASICS Process Validation Definition In order to understand the deficiencies with the traditional PV approach, and be able to move forward with the new FDA Process Validation Life Cycle approach, we must go back to the original definition of Process Validation as stated in the 1987 guideline. The definition reads: Process validation is establishing documented evidence which provides a high degree of assurance that a specific process will consistently produce a product meeting its predetermined specifications and quality attributes. The traditional approach focuses on the documented evidence part of the definition but, in reality, is it providing adequate evidence? Does it provide a high degree of assurance that our process will meet its pre-determined specifications? The rest of the definition is even more in doubt if we are following the traditional PV approach - high degree of assurance that a specific process will consistently produce a product meeting its pre-determined specifications and quality attributes. Let us address each component of the definition one at a time: High degree of assurance - Do we quantify the level of assurance with the traditional approach? How well our process meets our specifications can be measured and even quantified, but that was not part of our traditional PV protocol. Specific process - Do we have a well-defined specific process? In 23 years of experience, I have seen many occasions where management was asking for a PV protocol and execution when they knew the process was not well-defined or specified. Did they know the critical steps, the critical parameters? Were the control systems on the equipment adequate for the operational ranges on the specific process? That is the process and product understanding that the FDA is discussing as part of the QbD and Process Analytical Technology (PAT) principles and the proposed PV Lifecycle Concept. As we discussed earlier, the FDA expects that 38 Journal of Validation Technology
IVTJVT1107.qxd 10/8/07 3:27 PM Page 39 prior to the manufacture of the conformance batches, the manufacturer should have identified and controlled all critical sources of variability. That is part of defining a specific process to be validated. Will consistently - We still argue that, with three batches and all the samples meeting the specifications, this requirement of the PV definition is being met when clearly it is not. To demonstrate that we can predict or estimate that our process will consistently produce a product meeting the specifications, we need to use such tools as calculating a process capability. The problem with calculating the process capability is that people often forget that there are pre-requisites for this tool to be applied - the process must be in control, must follow a normal distribution, and the methods used for the testing must be reliable and adequate for the analysis (minimal variability). Predetermined specifications and quality attributes- The specifications must be developed during the design process and must be justified with data created from scientifically-based studies. Based on the entire PV definition from 1987, the Process Validation Lifecycle concept was a necessity and not merely a new approach. FUTURE OF PROCESS VALIDATION The evolution of the new PV Lifecycle Concept will address all of the key principles in the Process Validation definition from 1987. The lifecycle, including four phases on a continuous feed-back loop, includes: design, confirmation (three batches), monitoring, and assessment. Based on the continuous monitoring and assessment of our commercial process, we will provide the recommendations for the design changes and adjustments. This process also occurs in two levels - internally, within each batch, and as an accumulative response from several batches. The desired PV state will include: Product quality and performance achieved and assured by the design of effective and efficient manufacturing processes Figure 1 Effective characterization and monitoring Monitoring the right attributes in real-time Input materials CQA Process In-process or final product CQA Adjusting process parameters Process parameters Variables Control models and systems November 2007 Volume 14, Number 1 39
IVTJVT1107.qxd 10/8/07 3:27 PM Page 40 Product specifications based on the mechanistic understanding of how formulation and process factors impact product performance An ability to effect continuous improvements and the continuous real time assurance of quality How do we get there? Perhaps by implementing the following: Eliminate and control special causes of variability - develop effective Corrective and Preventive Action (CAPA) systems Reduce and control common causes of variability - use process capability analysis Focus on the critical few - acquire ability to predict Critical-to-Quality Attributes (CQA) Establish CQA target values and acceptable variability Monitor to demonstrate state of control Based on critical material and equipment attributes Not end-product testing From Fixed to Variable Process Another concept which may be harder to comprehend is that Change will be GOOD. To understand this concept, we need to understand our traditional way of thinking as follows: Variable Input + Fixed Process = Variable Output (OOS lots are scrapped) Results: The control on the product critical quality attributes (CQA) was to be achieved via a fixed, static process. Defects are corrected after they occur. Investigate deviations and conduct root cause analysis with limited data. The proactive approach will look like this: Variable Input + Variable Process = Consistent Output (greater efficiency) Results: CQA control attempted by constantly varying the process to adapt to minor changes. Continuous Quality Verification There are other industry representatives even more audacious who discuss a Continuous Quality Verification Model as follows: Step 1 Process Understanding Design phase develops an adequate level of product and process understanding which is a key requirement to establish a Continuous Quality Verification system. During design, the manufacturer defines the critical parameters, the levels of control for each, the in-process testing for controlling the process, and the instrumentation required (accuracy, precision, reliability). Step 2 Continuous Quality, Monitoring, and Feedback Adequate testing and instrumentation to monitor the process on-going and provide feedback to the controlling system to adjust the parameters based on the results (internal - within the process itself). Step 3 Periodic Data Analysis Accumulative data compilation and analysis to determine needs for changes or adjustments on our process or systems. Step 4 Modifications Based on Data Analysis Changes and adjustments as a result of the previous step for data analysis. CONCLUSION The optimal approach to validation considers process parameters, product attributes and their relationships. The link between parameters and attributes is established 40 Journal of Validation Technology
IVTJVT1107.qxd 10/8/07 3:27 PM Page 41 during the developmental process. The future of PV will bring more focus on the development phase and on the monitoring and assessment phases in addition to the confirmation step. This will include application of real-time quality control and assurance eliminating the need for concepts such as time-based revalidations or testing after the fact. We are on the right road, but have not reached our destination as yet. A clear pathway forward is desirable. One of those paths will be determined by the FDA's revised process validation guideline expected sometime this year. REFERENCES 1. FDA Compliance Policy Guide 7132c.08, Sec. 490.100, Process Validation Requirements for Drug Products and Active Pharmaceutical Ingredients Subject to Pre-Market Approval, - Revised March 12, 2004. 2. FDA, Pharmaceutical cgmps for the 21st Century - A Risk Based Approach. Final Report, September 2004. 3. FDA, Guideline on General Principles of Process Validation, Final, May 1987. 4. Presentation by Ms. Grace E. McNally, CDER Office of Compliance, Division of Manufacturing and Product Quality, FDA at the PDA/FDA meeting in Washington, DC, September 2006. 5. Chapman, K.G., The PAR Approach to Process Validation, Pharmaceutical Technology, 8 (12), 22-36 (1984). Mr. has over 24 years of experience in the areas of cgmp compliance, quality operations/systems and validation functions/responsibilities. He is the owner and President of Expert Validation Consulting, Inc, a firm specialized in focused and practical consulting for the pharmaceutical and OTC drug industry on cgmp compliance areas. Before forming EVC, Mr. Montalvo held positions of increasing responsibility in the areas of Validation, Technical Services and Quality Operations in companies such as AAC Consulting Group, Inc, Millipore Corporation, Raytheon Engineers and Constructors, Mova Pharmaceutical Corp., Bristol-Myers Squibb and Baxter Healthcare Corporation. He has developed comprehensive and compliant quality and validation programs and systems for numerous companies and provided support, audits, and assessments for many existing operations. His extensive areas of expertise include development and implementation of quality functional procedures (QA/QC), internal and external cgmp audits including remediation efforts for companies under a consent decree, non-conformance evaluation and documentation, Quality Systems, handling of all kinds of investigations and deviations, validation maintenance, change control, CAPA, international GMP compliance, all types of validations, calibrations, Risk Management, and start-up manufacturing facilities. He holds a BS in Chemical Engineering from Rensselaer Polytechnic Institute and an MBA. He has been a frequent speaker and has chaired hundreds of validation and quality related conferences around the world for such groups as PDA, PTi, CfPIE, IVT, Barnett International and the CTFA. His articles and papers have been published in the American Pharmaceutical Review and the Journal of Validation Technology. He is a member of the Journal of Validation Technology Editorial Advisory Board. Miguel can be reached by telephone at (407) 587-6540 and by email at: mmontalvo@expertvalcon.com. Article Acronym Listing ABOUT THE AUTHOR CAPA cgmp CPG CQA FDA OOS PAR PAT PV QbD SOP Corrective And Preventive Action Current Good Manufacturing Practice Compliance Policy Guide Critical Quality Attribute Food and Drug Administration Out Of Specification Proven Acceptable Ranges Process Analytical Technology Process Validation Quality by Design Standard Operating Procedure November 2007 Volume 14, Number 1 41