FDA's New Process Validation Guidance: Industry Reaction, Questions, and Challenges

By Mike Long,Hal Baseman,Walter D. Henkels

The authors desribe the three-stage approach to validation that is outlined in the new guidance and discuss questions surrounding implementation.

FDA's 2011 Process Validation: General Principles and Practices guidance created a systemic shift in industry's approaches to validation programs. The authors describe the three-stage approach to validation that is outlined in the new guidance and discuss questions surrounding implementation.
In January 2011, FDA published its long-awaited guidance for industry on Process Validation: General Principles and Practices (1) . For many in the pharmaceutical industry, the guidance created a systemic shift in the expectations of their validation programs. Although the new guidance aligns process-validation activities with the product life-cycle concept and with existing harmonized guidelines such as the Internatioanl Confernce on Harmonization's Q8(R2) Pharmaceutical Development, Q9 Quality Risk Management, and Q10 Pharmaceutical Quality System, it may have created as many questions for the industry as it has answered. In this article, the authors briefly describe the three-stage approach to validation that is outlined in the new guidance as well as implications for manufacturers regarding their current approaches to process validation. Specific emphasis is placed on questions surrounding industry implementation.
Design for assurance
The regulatory basis for process validation is contained in a number of places. Process validation is legally enforceable per the Federal Food, Drug, and Cosmetics Act. The requirements are called out in 21 CFR Parts 210 and 211 of the CGMP regulations, more specifically in Part 211.100 (a). This section is what the FDA describes as the regulatory "foundation for process validation" (1). Here, manufacturers are required to have production and process-controls procedures in place that are "designed to assure" drug products have a certain level of quality and that their products are manufactured safely, effectively, and purely.
How is that assurance to be provided, however? How can industry design to assure these qualities? One way is through direct observation; another is through prediction. Direct observation may require the destruction of the products, meaning that industry generally must use predictive methods to determine, with a certain level of confidence, that its processes and controls are designed to assure quality. Validation is the predictive method for providing this assurance.
A brief history of validation
In the 1970s, there was a series of contamination issues in large-volume parenteral bottles that, with a handful of other significant adverse events, led regulatory authorities and manufacturers to focus more on process understanding and quality assurance. The idea that finished product testing was not enough to assure product quality began to grow. Industry needed a better system to determine whether a product was "good" (2).
In the mid-1970s, FDA's Ted Byers and Bud Loftus began to promote the idea that a focus on process design and an evaluation of support processes and functions would assist in assuring that processes were under control (rather than just " testing quality into the product"). At a 1974 conference, Byers called this new predictive approach "Design for Quality" (3).
In 1987, at the request of industry, FDA published its original guideline on General Principles of Validation (4). The agency established a more formal approach to process validation and created the assertion that multiple batches need to be run. This concept eventually translated into the three-batch approach, where the successful running of three consecutive product batches represents a validated process. Herein are fundamental questions that still exist 20 years after the publication of the guidance:

• If companies were following the guidance, why would companies continue to see processes fail during commercial manufacture?
• Were companies really providing high degrees of assurance their processes worked reliably?
• Did companies understand how or why running three batches proved that their processes were adequately controlled and capable of consistently providing required results?

When these questions are asked of people in the industry, the answers, in the authors' experience, are invariably, "Well, stuff happens. Things happen. Things we didn't anticipate." "Stuff happens" tends to refer to "process variation," a key term and a key reason why the 1987 guidance was updated.
The new guidance
The new guidance ushers in a life-cycle approach to validation. This approach is most apparent in the guidance's new definition for process validation: "...the collection and evaluation of data, from the process design through commercial production, which establishes scientific evidence that a process is capable of consistently delivering quality products" (1)
The words used have been selected carefully. The new definition talks about establishing process capability using scientific evidence while previous definitions used the phrase "documented evidence" (4). This earlier definition may have contributed to validation being viewed largely as a late-stage documentation exercise. The new definition, however, describes process validation as a continuous process of collection and evaluation of data, rather than as a three-batch static event. As one FDA representative commented at a recent PDA workshop, the number of batches is not an acceptance criteria; however, the results of the data obtained from the batches are (5). The new definition of validation caused one industry member to lament at the workshop that for the past 30 years, industry has been told that process validation is a documentation exercise. Now, FDA expects industry to consider process validation as a scientific endeavor. That is quite a shift and 30-year habits are hard to break, he noted (5).
This statement underlies the dichotomy that exists within the industry regarding the new guidance. For years, the industry criticized regulators that the industry owned the expertise, knew their processes better than regulatory agencies, and should have flexibility in how they could validate these processes. The agency listened and came back with a guidance that is deliberately non-prescriptive; it doesn't tell industry how many samples to pull or how many batches to run. The industry has to be able to answer these questions and they are not always easy to answer.
Defining the new process-validation stages
As noted, the 2011 guidance enforces the life-cycle model described in ICH Q10 and in FDA's 2006 guidance for industry on Quality Systems Approach to Pharmaceutical CGMP Regulations, which states that, "quality should be built into the product, and testing alone cannot be relied on to ensure product quality" (6, 7). The method of building quality into a product or "designing for assurance" challenges manufacturers to better:

• Understand the sources of variation along the supply chain in the manufacturing process
• Detect the existence and amount of variation that is imparted into the product from sources within the supply chain, the equipment, and personnel
• Understand the impact of the detected variation on the finished product
• Control the variation detected based upon the understanding and knowledge of the sources of the variation that proportional to its risk to product and patient (1).

The new guidance document describes validation activities in three stages using a life-cycle model (see Figure 1). Although explained as discrete stages, some activities can occur in multiple stages while others may overlap between stages. Risk assessments are a good example of such activities. The guidance describes these three stages as follows (see Figure 1):

• Stage 1–Process Design: The commercial-manufacturing process is defined during this stage based on knowledge gained through development and scaleup activities.
• Stage 2–Process Qualification: During this stage, the process design is evaluated to determine whether the process is capable of reproducible commercial manufacturing.
• Stage 3–Continued Process Verification: Ongoing assurance is gained during routine production that the process remains in a state of control (1).

Figure 1: The three stages of the validation life-cycle model based on the new FDA process validation guidance. (Authors)

These stages are defined in more detail in the following sections. Stage 1–Process Design. Process validation is an ongoing process within the product life cycle. This cycle begins at Stage 1. Here, the commercial process starts to be defined using knowledge gained through development and scale-up activities (1). Process knowledge and understanding is captured in Stage 1 to determine initial process capability and to evaluate sources of variation. Design of experiments can be used to identify and establish process-parameter relationships with quality attributes. Risk-management efforts should begin in earnest at this phase to assist in capturing the product and process knowledge and to prioritize development efforts, which can be ranked by the relative importance of the quality attributes. Process-control strategies can begin to be developed during this stage through initial understanding of the risks and sources of variation (8).
Stage 2–Process Qualification. Although the 2011 guidance document's three-stage description of validation may be new to the industry, the content and concepts within Stage 2 should be the most familiar. During this stage, the process design is confirmed as being capable of reproducible commercial manufacturing. This phase involves evaluating the facility and equipment for its fitness for use. Utility systems and equipment are verified to be built and installed properly, and operators ensure that they operate within the intended and anticipated operating ranges. The conclusion of this stage commences with the execution of a process-performance qualification (PPQ) exercise. PPQ is a significant milestone in the product life cycle. The decision to distribute product to the market will be determined by the successful completion of PPQ. The amount and usefulness of the product and process knowledge gained up to this point in the life cycle will determine the approach a company takes with its PPQ (1). The focus at this stage should not be on the number of batches needed to produce a successful PPQ run, but rather, on whether enough data has been produced and evaluated to answer two basic questions:

• What scientific evidence is there to provide the appropriate level of assurance that the manufacturing system has been designed to consistently deliver a quality product to the market?
• Are there systems in place to properly monitor and control that manufacturing system?

Stage 3–Continued Process Verification. Stage 3 of the 2011 guidance is the continuous-improvement phase of the process-validation life cycle. In this phase, data obtained from routine production is used to provide ongoing assurance that the process remains in the state of control. This activity is more than an annual product review; rather, this action involves using a system or multiple systems of assuring control.
Control strategies originally conceived in Stage 1 are implemented during this phase as well. Tracking and trending of data allows for the detection of special-cause variability and the reduction of inherent common-cause variability. Determination on the state of control of the processes in commercial distribution should be calculated using appropriate statistics and derived from appropriate confidence levels. These confidence levels should be based upon risk factors, experience, and attribute criticality (1).
A continuous process of evaluation
With the new life-cycle approach, process validation is no longer a one-time milestone event. Process validation should be considered a continuous process of valuation. The five steps shown in Figure 2 represent stages of validation. These steps require industry to:

• Know the process
• Know the variables
• Have confidence before going into commercial manufacture
• Create vigilance in understanding variation through monitoring and continuous improvement that the process is under control.

Figure 2: The process-validation sequence.(Figures are courtesy of authors)

Initial process understanding for a given product is most often based upon a combination of early development work on the product along with prior knowledge of the existing product and manufacturing platforms that will be used to commercialize the product. It is quite rare that a product entering the validation lif cycle will have a unique product or manufacturing platform. This initial process knowledge allows for a more robust process-design phase, which creates an enhanced understanding of the sources of variation and their impact on the product's quality attributes. Understanding the sources of variation provides confidence to go into commercial manufacture and to create an ongoing monitoring program that will track and trend data for continuous improvement. Industry reaction and challenges
The industry reaction to the 2011 process validation guidance, in general, has been favorable. The document is nonprescriptive. It is relatively flexible. It leaves much to the companies to decide how they wish to meet its expectations. But, as the authors discovered through a series of industry workshops and seminars, there still is a need for further explanation and clarification regarding the guidance as well as a real need for detailed implementation suggestions. Although there are many questions to be answered, this paper is limited to three of the most frequently asked questions about implementing the new guidance.
Common questions. The classic and most frequently asked question about the new guidance is, How many batches do I have to run to validate my processes under the new guidance? The answer does not delight many people: "It depends."
The number of batches needed to validate a process depends on how much data are required to provide a company with an appropriate level of statistical confidence and scientific justification to begin commercial production. There is nothing wrong with a company choosing to run three batches for validation, but that company has to be prepared to explain why the data obtained from three batches is sufficient to determine that the process is ready to move into commercial manufacturing. Likewise, if a company chooses to run fewer than three batches, it has to be able justify the amount of data created by those batches. In short, there is no longer a uniform number of validation batches that can be applied globally to the industry. The approach is now process- and product-specific, depending on a company's ability to test critical parameters, conditions, and attributes.
Overall, the batches are a means to an end, and in this case, that end is data (9).
The discussion of the number of batches required invariably leads to a second frequently asked question, How do we determine the amount of data we need in our validations? The question revolves around the amount of data needed in Stage 2–Process Qualification. FDA has been quite clear that statistical justification for sampling is required. Specifically, the guidance states that:

• The number of samples should be adequate to provide sufficient statistical confidence of quality both within a batch and between batches.
• The confidence level selected can be based on risk analysis as it relates to the particular attribute under examination.
• Sampling during this stage should be more extensive than is typical during routine production (1).

A key to success here will be appropriate use of risk analysis to assist in selection of appropriate confidence levels for a product's critical quality attributes (CQAs). All CQAs are not alike, the higher the risk for a given CQA, the higher the confidence level required during validation. The amount of data required might then increase for those CQAs that require higher confidence levels. This decision will depend on the test methods used and the type of data collected. Test methods that create dichotomous data (pass/fail, go/no go) will require many more samples than those that produce continuous (i.e., variable) data. This difference needs to be taken into consideration early in the validation life-cycle process becaues it also impacts how data will be trended in Stage 3.
Another question brought up frequently is, Do the requirements in the guidance conflict with existing global regulatory requirements? This question has two components to it.
First, there is a specific concern that validation terms within the document do not align with EMA terms, specifically Annex 15 of the European Union Guidelines to Good Manufacturing Practice, or those in ICH Q7 (10, 11). Examples of misalignment would be traditional terms such as installation qualification and operational qualification. These terms are not included in the new FDA process validation guidance, but the concept described in Stage 2 of the guidance does state that, "qualification refers to activities undertaken to demonstrate that utilities and equipment are suitable for their intended use and perform properly." This language is similar in nature to Sections 9–13 of the EU GMP Annex and Section 12.3 of ICH Q7 (1, 10, 11).
Second, there is a concern that certain concepts within the EU GMPs conflict with the new FDA guidance, specifically regarding the three-batch validation approach. Section 25 of the EU GMP Annex 15 states, "It is generally considered acceptable that three consecutive batches/runs within the finally agreed parameters, would constitute a validation of the process" (10). On the surface, this language would seem to conflict with the FDA's new focus on data rather than the number of batches. However, Section 25 of the EU GMP Annex 15 also states that, "...the number of process runs carried out and observations made should be sufficient to allow the normal extent of variation and trends to be established and to provide sufficient data for evaluation." So, while the three-batch approach to validation is embedded in Annex 15, Section 25 also provides for a more flexible approach based upon data (10).
Related to the question above are general questions regarding terminology. The most notable new term in the FDA guidance is process performance qualification (PPQ), described earlier in this paper. The guidance defines PPQ as "the second element of Stage 2, process qualification. The PPQ combines the actual facility, utilities, equipment (each now qualified), and the trained personnel with the commercial manufacturing process, control procedures, and components to produce commercial batches. A successful PPQ will confirm the process design and demonstrate that the commercial manufacturing process performs as expected" (1). PPQ is what most companies call their process validation (PV) and should not be confused with individual equipment process qualifications. Here again, FDA has stated that the concepts are important, not the terms (9). Companies are free to call these activities what they wish as long as the two parts of Stage 2 as outlined in the guidance are met.
Challenges. Changing the existing validation culture to meet the expectations of the new guidance may be the biggest challenge industry faces. The guidance requires companies to expand their current scope of validation by reaching further upstream into development and downstream into day-to-day manufacturing. Companies agree this expansion will foster better communication from the development groups through to manufacturing but are concerned that it also requires additional staffing. To better understand and meet this challenge, companies should create gap analyses of their current state of validation and compare it with the future state based on the new guideline. These gap analyses will better situate companies to create action plans for the new policies and procedures that may need to be put into place. The analyses will also assist in identifying any resource gaps and training needs.
Conclusion
FDA's new process validation guidance has received generally favorable reviews for its nonprescriptive and flexible approach. With this flexibility, however, comes a concern about acceptable approaches for implementation at a practical level and a genuine need for a continued dialogue of explanation and clarification between the agency and industry. Due to limitations, this article could only touch on a handful of the questions being asked at this time. The authors endeavor to create a series of articles in Pharmaceutical Technology to further address issues such as the use of risk management along the three stages of validation, collection and transfer of product and process knowledge along the validation life cycle, and continued process verification.
Dr. Mike Long, MBB,* is director of consulting services, and Walter D. Henkels, Sr., is a consultant, both at ConcordiaValsource. Hal Baseman is chief operating officer of ValSource LLC. Mlong@concordiaValsource.com [mlong@concordiavalsource.com]
*To whom all correspondence should be addressed.
References
1. FDA, Guidance for Industry: Process Validation: General Principles and Practices (Rockville, MD, Jan. 2011).
2. US vs. Barr Laboratories, 1993.
3. J. M. Dietrick and B.T. Loftus, "Regulatory Basis for Process Validation" in Pharmaceutical Process Validation, R.A. Nash and A. H. Wachter, Eds. (Marcel Dekker New York, NY, 3rd ed., 2003) p. 43.
4. FDA, Guideline on General Principles of Process Validation (Rockville, MD, May 1987, reprinted May 1990).
5. PDA Annual Meeting Post-Conference Workshop, "The FDA's Process Validation Guidance: Meeting Compliance Expectations and Practical Implementation Strategies" (San Antonio, TX, April 13–14, 2011).
6. ICH, Q10 Pharmaceutical Quality System (2008).
7. FDA, Guidance for Industry: Quality Systems Approach to Pharmaceutical CGMP Regulations (Rockville, MD, September 2006).
8. M. Long and H. Baseman, presentation at the Elsevier Business Intelligence Webinar, May 18, 2011.
9. G. McNally, presentation at the PDA Annual Meeting Post-Conference Workshop (San Antonio, TX, April 13–14, 2011).
10. EU, Guidelines to Good Manufacturing Practice, Medicinal Products for Human and Veterinary Use, Annex 15: Qualification and Validation (Brussels, July 2001).
11. ICH, Q7 Good Manufacturing Practice Guide For Active Pharmaceutical Ingredients (November 2000).

Figure 1: The three stages of the validation life-cycle model based on the new FDA process validation guidance. (Authors)
Figure 2: The process-validation sequence.(Figures are courtesy of authors)

Analytical Method Validation: Back to Basics, Part I

By Michael Swartz,Ira Krull

Michael Swartz

In any regulated environment, analytical method validation (AMV) is a critical part of the overall process of validation. AMV is a part of the validation process that establishes, through laboratory studies, that the performance characteristics of the method meet the requirements for the intended analytical application and provides an assurance of reliability during normal use; sometimes referred to as "the process of providing documented evidence that the method does what it is intended to do." Regulated laboratories must perform AMV to be in compliance with government or other regulators, in addition to being good science. A well-defined and documented validation process can provide evidence not only that the system and method is suitable for its intended use, but can aid in transferring the method and satisfy regulatory compliance requirements.

Ira Krull

Validation is also the foundation of quality in the high performance liquid chromatography (HPLC) laboratory, and AMV is just one part of a regulatory quality system that incorporates both quality control and quality assurance (1,2). The terms "quality control" and "quality assurance" often are used interchangeably, but in a properly designed and managed quality system, the two terms have separate and distinct meanings and functions. Quality assurance (QA) can be thought of as related to process quality; whereas quality control (QC) is related to the quality of the product. In a given organization, it does not matter what the functions are named, but the responsibilities for these two activities should be defined clearly. Both quality assurance and quality control make up the Quality Unit, and are essential to the production of analytical results that are of high quality and are compliant with the appropriate regulations. QC is the process that determines the acceptability or unacceptability of a product or a product plan, and is determined by the comparison of a product against the original specifications that were created before the product existed. In some organizations, the QC group is responsible for the use of the method to perform analysis of a product. Other tasks related to QC can include documented reviews, calibrations, or additional types of measurable testing (such as sampling) and will reoccur more often than activities associated with quality assurance. QC usually will require the involvement of those directly associated with the research, design, or production of a product. For example, in a laboratory-notebook peer-review process a QC group would check or monitor the quality of the data, look for transcription errors, check calculations, verify notebook sign-offs, and so forth.
QA is determined by top-level policies, procedures, work instructions, and governmental regulations. At the beginning of the validation process, QA can provide guidance for the development of or review of validation protocols and other validation documents. During the analytical stage, QA's job is to ensure that the proper method or procedure is in use and that the quality of the work meets the guidelines and regulations. QA can be thought of as the process that will determine the template and pattern of quality control tasks. As opposed to QC checks, QA reports are more likely to be performed by managers, by corporate level administrators, or third-party auditors through the review of the quality system, reports, archiving, training, and qualification of the staff that performs the work. AMV Guidance
Since the late 1980s, government and other agencies (for example, FDA, International Conference on Harmonization-ICH) have issued guidelines on validating methods. In 1987, the FDA designated the specifications in the current edition of the United States Pharmacopeia (USP) as those legally recognized when determining compliance with the Federal Food, Drug, and Cosmetic Act (3,4). More recently, new information has been published, updating the previous guidelines and providing more detail and harmonization with International Conference on Harmonization (ICH) guidelines (5,6). Guidelines are documents prepared for both regulatory agency personnel and the public that establish policies intended to achieve consistency in the agency's regulatory approach, and to establish inspection and enforcement policies and procedures. For example, the FDA guidance provides recommendations to applicants on submitting analytical procedures, validation data, and samples to support the documentation of the identity, strength, quality, purity, and potency of drug substances and drug products, and it is intended to assist applicants in assembling information, submitting samples, and presenting data to support analytical methodologies. The recommendations apply to drug substances and drug products covered in new drug applications (NDAs), abbreviated new drug applications (ANDAs), biologics license applications (BLAs), product license applications (PLAs), and supplements to these applications.
One final introductory comment: Although for the most part the topic of discussion in most "Validation Viewpoint" columns is HPLC, the guidelines are generic; that is, they apply to any analytical procedure, technique, or technology used in a regulated laboratory (for example, gas chromatography [GC], mass spectrometry [MS], or IR spectroscopy). It also should be noted that the USP publishes official methods, often called compendial methods, which have been accepted by the USP as already validated. However it is common practice to verify these methods, and a separate USP chapter is devoted to this topic (7,8).
The Validation Process
When looking at the guidelines, one observes that AMV is just one part of the overall validation process that encompasses at least four distinct steps: software validation, hardware (instrumentation) validation–qualification, analytical method validation, and system suitability. The overall validation process begins with validated software and a validated–qualified system; then a method is developed and validated using the qualified system. Finally, the whole process is wrapped together using system suitability. Each step is critical to the overall success of the process.
Software Validation

Table I: A time line approach to AIQ

A comprehensive treatment of software validation is outside the scope of this column. However, it is an important topic to at least touch on here as these days, every modern HPLC laboratory makes use of computerized systems to generate and maintain source data and documentation from a variety of instrumentation. These data must meet the same fundamental elements of data quality (for example, attributable, legible, contemporaneous, original, and accurate) that are expected of paper records and must comply with all applicable statutory and regulatory requirements. Two FDA guidelines have appeared recently that address the topic of software validation, and should be consulted for more detailed information (9,10). In addition, in March 1997, the FDA issued 21 CFR part 11, which provided the original criteria for acceptance of electronic records, electronic signatures, and handwritten signatures executed to electronic records as equivalent to paper records and handwritten signatures executed on paper under certain circumstances (11). However, after the effective date of 21 CFR part 11, significant concerns regarding the interpretation and implementation of part 11 were raised by both FDA and the pharmaceutical industry, and as a result, 21 CFR part 11 was re-examined (12). The new Scope and Application Guidance clarified that the FDA intends to interpret the scope of part 11 narrowly and to exercise enforcement discretion with regard to part 11 requirements for validation, audit trails, record retention, and record copying. However, most of the other original Part 11 provisions remain in effect. Analytical Instrument Qualification
Before undertaking the task of method validation, it is necessary to invest some time and energy up-front to ensure that the analytical system itself is validated, or qualified. Qualification is a subset of the validation process that verifies proper module and system performance before the instrument being placed on-line in a regulated environment. In March, 2003, the American Association of Pharmaceutical Chemists (AAPS), the International Pharmaceutical Federation (FIP), and the International Society for Pharmaceutical Engineering (ISPE) cosponsored a workshop entitled "A Scientific Approach to Analytical Instrument Validation" (13). Among other objectives, the various parties (the event drew a cross-section of attendees; users, quality assurance specialists, regulatory scientists, consultants, and vendors) agreed that processes are "validated" and instruments are "qualified," finally reserving the term validation for processes that include analytical methods–procedures and software development.
The proceedings of the AAPS et al. committee have now become the basis for a new general USP chapter, number 1058, on Analytical Instrument Qualification (AIQ) that originally appeared in the USP's Pharmacopeial Forum (14–16). The chapter details the AIQ process, data quality, roles and responsibilities, software validation, documentation, and instrument categories.
Instruments are qualified according to a stepwise process grouped into four phases: design qualification (DQ), installation qualification (IQ), operational qualification (OQ), and performance qualification (PQ). The DQ phase usually is performed at the vendor's site, where the instrument is developed, designed, and produced in a validated environment according to good laboratory practices (GLP), current good manufacturing practices (cGMP), and ISO 9000 standards.
During the IQ phase, all of the activities associated with properly installing the instrument (new, pre-owned, or existing) at the users' site are documented. After the IQ phase is completed, testing is done to verify that the instrument and instrument modules operate as intended in an OQ phase. First, fixed parameters, for example, length, weight, height, voltage inputs, pressures, and so forth are either verified or measured against vendor-supplied specifications. Because these parameters do not change over the lifetime of the instrument, they usually are measured just once. Next, secure data handling is verified. Finally, instrument function tests are undertaken to verify that the instrument (or instrument modules) meets vendor and user specifications.
Instrument function tests should measure important instrument parameters according to the instruments' intended use and environment. For LC, the following types of tests might be included: n pump flow rate

• detector wavelength accuracy
• gradient linearity
• injector precision, linearity, and accuracy
• detector linearity
• column oven temperature.

Relevant OQ tests should be repeated whenever the instrument undergoes major repairs or modifications.

Figure 1

After an IQ and an OQ have been performed, PQ testing is conducted. PQ testing should be performed under the actual running conditions across the anticipated working range. PQ testing should be repeated at regular intervals; the frequency depends upon such things as the ruggedness of the instrument, and the criticality and frequency of use. PQ testing at periodic intervals also can be used to compile an instrument performance history. In practice, a known method with known, predetermined specifications is used to verify that all of the modules are performing together to achieve their intended purpose. In practice, OQ and PQ frequently blend together in a holistic approach, particularly for injector linearity and precision (repeatability) tests, which can be conducted more easily at the system level. For HPLC, the PQ test should use a method with a well-characterized analyte mixture, column, and mobile phase. Figure 1 shows an example of a "vendor" PQ test method that incorporates the essence of a holistic OQ and PQ test. Actual user PQ tests should incorporate the essence of the system suitability section of the general chromatography chapter 621 in the USP (15) to show suitability under conditions of actual use.
System Suitability

Table II: System suitability parameters and recommendations (17)

According to the USP, system suitability tests are an integral part of chromatographic methods (17). These tests are used to verify that the resolution and reproducibility of, for example, a chromatographic system, are adequate for the analysis to be performed. System suitability tests are based upon the concept that the equipment, electronics, analytical operations, and samples constitute an integral system that can be evaluated as a whole. System suitability is the checking of a system to ensure system performance before or during the analysis of unknowns. Parameters such as plate count, tailing factors, resolution, and reproducibility (%RSD retention time and area for repetitive injections) are determined and compared against the specifications set for the method. However, samples consisting of only a single peak (for example, a drug substance assay in which only the API is present) can be used, provided that a column plate number specification is included in the method. Unless otherwise specified by the method, data from five replicate injections of the analyte are used to calculate the relative standard deviation if the method requires RSD > 2%; data from six replicate injections are used if the specification is RSD ≤ 2%. These parameters are measured during the analysis of a system suitability "sample" that is a mixture of main components and expected byproducts. Table II lists the terms to be measured and their recommended limits obtained from the analysis of the system suitability sample (18).

Table III: Maximum specifications for adjustments to HPLC operating conditions

System suitability tests must be carried out before the analysis of any samples in a regulated environment. Following blank injections of mobile phase, water, or sample diluent, replicate system suitability injections are made, and the results compared against method specifications. If specifications are met, subsequent analyses can continue. If the method's system suitability requirements are not met, any problems with the system or method must be identified and remedied (perhaps as part of a formal out-of-specification (OOS) investigation), and passing system-suitability results must be obtained before sample analysis is resumed. To provide confidence that the method runs properly, it is also recommended that additional system suitability samples (quality control samples or check standards) are run at regular intervals (interspersed throughout the sample batch); %-difference specifications should be included for these interspersed samples, to make sure the system still performs adequately over the course of the entire sample run. Alternatively, a second set of system-suitability samples can be included at the end of the run. Quality control check samples are run to make sure the instrument has been calibrated properly or standardized. Instrument calibration ensures that the instrument response correlates with the response of the standard or reference material. Quality control check samples also are used often to provide an in-process assurance of the test's performance during use. In general, AIQ and analytical method validation generally ensure the quality of analysis before conducting a test; system suitability and quality control checks ensure the quality of analytical results immediately before or during sample analysis. As previously mentioned, USP compendial methods are considered to be validated, however, adjustments to USP methods are allowed to meet system-suitability requirements; such instructions can be included in individual monographs. But method changes usually require some degree of revalidation. So an important question is, "At what point does an adjustment become a change or modification" to the method? Historically, if adjustments to the method are made within the boundaries of any robustness studies that were performed, no further actions are warranted as long as system-suitability criteria are satisfied. Unfortunately, the robustness information is not readily available for many USP methods, and any adjustment outside of the bounds of the robustness study constitutes a change to the method and requires a revalidation.
In 1998, Furman and colleagues proposed a way to classify allowable adjustments (19). But it was not until 2005 that guidance appeared on the topic (20–22). Although USP guidance on this topic recently was included into USP Chapter 621 on chromatography (17), the FDA Office of Regulatory Affairs (ORA) has had guidance in place for a number of years (20). Table III summarizes the adjustments allowed for various HPLC parameters taken from both the USP and ORA documents. Adjustments outside of the ranges listed in Table III constitute modifications, or changes, which are subject to additional validation. Sound scientific reasoning should be used when determining method adjustment versus method change for a specific method. For example, if robustness studies have shown that the method conditions allow less variability for a parameter than that listed in Table III, or when robustness testing have shown that more variability is allowed, the robustness results (as summarized in the validation report) should prevail.
A few additional comments: Adjustments to chromatographic systems to comply with system-suitability requirements should not be made to compensate for system malfunctions or column failures. To prevent specification "creep," adjustments are made only from the original method parameters each time the method is run and are not subject to continuous adjustment. Adjustments are permitted only when suitable reference standards are available for all compounds used in the suitability test and only when those standards are used to show that the adjustments have improved the quality of the chromatography so as to meet system suitability requirements. The suitability of the method under the new conditions must be verified by assessment of the relevant analytical performance characteristics. Because multiple adjustments can have a cumulative effect in the performance of the system, any adjustments should be considered carefully before implementation.
Conclusion
In today's global market, validation can be a long and costly process, involving regulatory, governmental, and sanctioning bodies from around the world. A well-defined and documented validation process provides regulatory agencies with evidence that the system (instrument, software, method, and controls) is suitable for its intended use. AMV is a critical part of this process, and will be further addressed in detail in Part II of this two-part series.
Michael Swartz "Validation Viewpoint" Co-Editor Michael E. Swartz is Research Director at Synomics Pharmaceutical Services, Wareham, Massachusetts, and a member of LCGC 's editorial advisory board.
Ira S. Krull "Validation Viewpoint" Co-Editor Ira S. Krull is an Associate Professor of chemistry at Northeastern University, Boston, Massachusetts, and a member of LCGC 's editorial advisory board.
The columnists regret that time constraints prevent them from responding to individual reader queries. However, readers are welcome to submit specific questions and problems, which the columnists may address in future columns. Direct correspondence about this column to "Validation Viewpoint," LCGC, Woodbridge Corporate Plaza, 485 Route 1 South, Building F, First Floor, Iselin, NJ 08830, e-mail lcgcedit@lcgcmag.com [lcgcedit@lcgcmag.com]
.
References
(1) United States Food and Drug Administration, Guideline for submitting samples and analytical data for methods validation, February 1997. US Government Printing Office: 1990-281-794:20818, or at www.fda.gov/cder/analyticalmeth.htm
(2) Current Good Manufacturing Practice in Manufacturing, Processing, Packing, or Holding Of Drugs, 21 CFR Part 210, http://www.fda.gov/cder/dmpq/cgmpregs.htm|~http://www.fda.gov/cder/dmpq/cgmpregs.htm
(3) Current Good Manufacturing Practice for Finished Pharmaceuticals, 21 CFR Part 211, http://www.fda.gov/cder/dmpq/cgmpregs.htm|~http://www.fda.gov/cder/dmpq/cgmpregs.htm
(4) USP 32-NF 27, August 2009, Chapter 1225.
(5) Analytical Procedures and Method Validation. Fed. Reg. 65(169), 52,776-52,777, 30 August 2000. See also: www.fda.gov/cder/guidance
(6) International Conference on Harmonization, Harmonized Tripartite Guideline, Validation of Analytical Procedures, Text and Methodology, Q2(R1), November 2005, See www.ICH.org.
(7) USP 32-NF 27, August 2009, Chapter 1226.
(8) M.E. Swartz and I.S. Krull, LCGC 23(10), 1100–1109 (2005).
(9) FDA, General Principles of Software Validation; Guidance for Industry and FDA Staff. See: http://www.fda.gov/downloads/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/ucm085371.pdf|~http://www.fda.gov/downloads/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/ucm085371.pdf
(10) Guidance for Industry Computerized Systems Used in Clinical Investigations, FDA May 2007. See: http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm070266.pdf|~www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm070266.pdf/
(11) FDA, 21 CFR Part 11, "Electronic Records; Electronic Signatures; Final Rule." Federal Register Vol. 62, No. 54, 13429, March 20, 1997.
(12) FDA, Part 11, Electronic Records; Electronic Signatures — Scope and Application, 2003. See: http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm072322.pdf|~http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm072322.pdf
(13) Qualification of Analytical Instruments for Use in the Pharmaceutical Industry: A Scientific Approach, AAPS PharmSciTech, 2004, 5(1) Article 22 ( http://www.aapspharmscitech.org|~http://www.aapspharmscitech.org/
(14) Pharmacopeial Forum , 31(5), 1453–1463 (Sept-Oct 2005).
(15) USP 32-NF 27, August 2009, Chapter 1058
(16) M.E. Swartz, Analytical Instrument Qualification, Pharmaceutical Regulatory Guidance Book (Advanstar Communications, July 2006), pp. 12–16.
(17) USP 32-NF 27, August 2009, Chapter 621
(18) Center for Drug Evaluation and Research (CDER), Reviewer Guidance: Validation of Chromatographic Methods, US Government Printing Office, 1994 - 615-023 - 1302/02757.
(19) W.B. Furman, J.G. Dorsey, and L.R. Snyder, Pharm. Technol. 22(6), 58–64 (1998).
(20) FDA ORA Laboratory Procedure, #ORA-LAB.5.4.5, USFDA ( 09/09/2005). See also: http://www.fda.gov/ora/science_ref/lm/vol2/section/5_04_05.pdf|~http://www.fda.gov/ora/science_ref/lm/vol2/section/5_04_05.pdf
(21) Pharmacopeial Forum, 31(3), 825 (May-June 2005).
(22) Pharmacopeial Forum, 31(6) 1681 (Nov.-Dec. 2005).

Michael Swartz
Ira Krull
Table I: A time line approach to AIQ
Figure 1
Table II: System suitability parameters and recommendations (17)
Table III: Maximum specifications for adjustments to HPLC operating conditions

A Compliance Perspective on Dissolution Method Validation for Immediate-Release Solid Oral Dosage Forms on Automated Instrumentation

By David Fortunato

An automated dissolution method can be a powerful tool to test drug products at all phases of their development. With minimal automated method validation, this tool can be used early in the drug-evaluation process. And with additional validation efforts, an automated method can be extended to the testing of Phase IV stability batches. Validating an automated dissolution-test method requires an understanding of the potential effects from filtration parameters, system interference, carry-over, cleaning..

As the pace of product development accelerates, the approach to dissolution-method development must advance beyond a manual method and an assay. A natural progression of the method-development process must include the transfer of the manual method onto automated instrumentation.
Validating the automated method is the primary challenge when transferring from the manual method. A dissolution scientist must understand the potential effects from filtration, system interference, carry-over, cleaning parameters, and media replacement. Automated dissolution instrumentation can help generate good manufacturing practice (GMP) data only when validated testing parameters can negate these influences, so the dissolution scientist can be confident that results do not differ between the manual and the automated methods. In addition, all laboratory equipment used to support or generate GMP data about automated instrumentation must follow an instrument "chain of compliance." Proper documentation must exist that proves each piece of equipment has been properly qualified and calibrated for its intended use.
Product development life cycle

Tips for validating automated dissolution parameters

The level of automated dissolution-method validation depends upon a product's phase of development. For early-phase products, minimal validation is required to screen the initial batches. Typically, filtration parameters must be established first to ensure that no amount of active pharmaceutical ingredient (API) is lost with filtration. The initial filtration parameters could be established manually and then transferred to the automated instrument. Next, automated dissolution-profile testing is completed to screen several different dissolution media. Profile sampling could be performed at 10, 20, 30, 45, and 60 min and the results compared. Selecting the various dissolution media that are used in the evaluation depends upon the solubility and stability of the API in each media. This process allows a dissolution chemist to determine quickly the media having the potential to provide the most discriminating dissolution performance for the product. Once these initial parameters have been established, a dissolution chemist can use the automated instrumentation to quickly screen early formulations and help formulators direct their future formulation efforts. As the life cycle of the product progresses, automated dissolution-method parameters must be validated if the generated data have the potential to be included in any type of GMP submission. At this phase of development, a dissolution chemist should have substantial experience performing both manual and automated dissolutions on a product. This experience can be useful to select automated dissolution method parameters, which should be able to generate results equivalent to those of manual dissolution tests.

Ensuring instrument qualification

Because the manual test is considered the "official" dissolution test, side-by-side dissolution-profile testing should be conducted manually and with automation. Results could be compared at 10, 20, 30, 45, 60, and infinity minutes. At the infinity time point, a final sample is taken after the dissolution has progressed with a stirring apparatus speed of 250 rpm for an additional 30 min after the Q time point. The chosen acceptance criteria for the comparison between the two methods should reflect and compensate for both nonvariable and highly variable drug products. The results generated at the earlier time points are the most significant because they have the highest potential for variation between the manual and automated methods. If comparable results are obtained between the two methods at the earlier time points, a dissolution chemist is more likely to be assured that accurate data are obtained at all time points of the automated dissolution-profile testing. At later time points, the percent drug released typically approaches 100%; therefore, not as much variation between the two tests would be expected at these later time points. Acceptance criteria must be established for results generated at the earlier time points ( <85 and="and" dissolved="dissolved" later="later" points="points" the="the" time="time">85% dissolved) with tighter acceptance criteria established at the later time points. Validation of automated method parameters
It is advantageous for a dissolution chemist to validate individual automated parameters even if comparable results are obtained between the automated and manual dissolution tests. This process provides additional information about the product and the automated procedure, which may be useful as the formulation evolves. In addition, by validating individual automated parameters, a dissolution chemist can demonstrate to the US Food and Drug Administration a high level of control and understanding of the automated procedures used to evaluate the performance of the drug product. The individual automated parameters to be validated should include filtration, system interference, carry-over, cleaning parameters, and media replacement for off-line sample collection. A dissolution chemist always must be aware that the dissolution of the product itself is validated, not a particular formula. With this fact in mind, it is beneficial to conduct the validation experiments after final formulas have been determined. If future formulations change drastically, experience and scientific judgment must be used to determine the necessity of revalidation of individual automation parameters.
Filtration. Although automated filtration parameters should be established first, the filtration procedures for manual dissolutions may not always transfer exactly to the automated instrument. Incompatible or inadequate automated filtration procedures are the most likely cause for different results between the manual and automated dissolution tests. A side-by-side comparison between a manual test and an automated test is the quickest way to evaluate compatibility between the two methods. Because results typically approach 100% drug released at later time points, samples taken at the earlier time points provide the best information. At minimum, it is advantageous to test two lots of drug product at the highest and lowest dosage strengths.
Experience in performing manual dissolutions should help determine the type of experiment to be conducted. The comparison can be performed two different ways. For highly variable batches, where the result relative standard deviation is >20% at the 10-min time point or >10% at later time points, USP recommends performing an automated dissolution and collecting manual samples simultaneously during the automated sampling. To ensure that the automated sampling probe is not affecting the hydrodynamics of the dissolution, however, a dissolution scientist should first perform the automated dissolution using only manual sampling. These results could be compared with the manual dissolution. For less variable batches, USP recommends comparing the results between separate manual and automated dissolutions tests. The acceptance criteria proposed under USP " <1092> Intermediate Precision guidelines state the difference should not exceed 10% with less than 85% dissolved and the difference should not exceed 5% for remaining time points above 85% dissolved (1, 2).
System interference. After acceptable filtration parameters have been established, a dissolution chemist should determine whether system interference is affecting the results generated with the automated instrumentation. This parameter is an important variable to investigate and hopefully eliminate as a potential source of difference between the manual and automated tests.
An example of system interference is any binding of the API to tubing or sampling needles when using sampling parameters described in the automated method. Once automated sampling occurs, the API often must travel through very long lines of tubing, thus generating the potential for system interference. The test may be performed by preparing a 20% API and a 100% placebo solution. The amount of API is measured with and without the automated system. A portion of the prepared API solution is placed into each of the six dissolution vessels. Sample aliquots should be withdrawn manually and filtered simultaneously as the automated system is sampling. The difference between the two responses should not exceed 2.0%. Because system interference is more likely to be observed with a 20% API solution, the information generated from this experiment is useful when analyzing results from dissolution-profile testing. Low levels of system interference from higher API concentrations may be less likely to be noticed. If system interference is not observed with a 20% API solution, a dissolution chemist is more likely to be assured that accurate data are obtained at earlier time points of automated dissolution profile testing. Conversely, if the results between the manual and the automated dissolution tests are not equivalent, especially at the earlier time points of profile testing, it may be determined that system interference is the cause of the nonequivalent automated results. Although only the results at Q time will pass or fail a batch, system interference is an important parameter to evaluate because it demonstrates to FDA a high level of control and understanding of the automated procedures used to evaluate the performance of the drug product.
Carry-over. If it has been determined that system interference is not a factor for the product, the elimination of carry-over must be validated for an automated dissolution testing system. Carry-over of an API may occur between sampling time points during dissolution-profile testing and between batches during multiple batch runs. To determine whether carry-over exists between sampling time points, perform a sampling sequence using solutions equivalent to 100% of the highest dosage form and a blank solution. The solutions must be sampled according to the already established automated filtration and sampling procedures. The sequence of the sampling should be as follows: 100% solution, blank, 100% solution.
The API in the blank solution should not exceed 1.0%, and the result for the second 100% sampled solution must be equivalent to 99.0–101.0% of the result of the first 100% sampled solution. If results exceed these acceptance criteria, increased sampling flush volumes, filter changes between time points, different filters, or any combination of these three parameters may need to be altered to obtain acceptable results.
Potential drug product carry-over between batches must be validated and eliminated if possible. This process allows the automated testing system to be used to its fullest potential so that multiple batches can be tested in a single run. The validation can be conducted by performing dissolution tests with the highest dosage strength batch followed immediately with a blank batch (no dosage forms). The API in the blank batch must not exceed an average of 1.0% for six vessels. The samples should be taken and compared at the infinity time point when results are expected to approach 100% API released. Cleaning parameters may need to be increased if results exceed this acceptance criterion.
Cleaning parameters. It is important to clean the automated dissolution-testing system between batches of a single run and between product changes. Clean the dissolution vessels, stirring shafts, sampling needles, and the entire length of all sampling lines. Potential problems may occur, especially between product changes, if a surfactant was used previously. Results from future batches may be inaccurately high if surfactant remains in the system from a previous run.
Adequate cleaning procedures must be validated to ensure no carry-over occurs between batches of a single run or after product changes. Validation of the cleaning parameters may be determined at the same time as the carry-over between batches experiments. If the carry-over results between batches exceed the acceptance criteria, increased cleaning parameters may solve the problem. The amount of dissolution media or hot water flushed through the lines at the end of a dissolution test may be set at the maximum allowable volume for that particular automated dissolution-testing system. In addition, the highest number of vessel washes and the volume of dissolution media or hot water used for the vessel washes may be set at the maximum allowable volume for that particular automated system. If even the most extensive cleaning parameters do not prevent an acceptable level of carry-over from an API, a dissolution chemist may decide that the dissolution procedure for this particular product is not "automatable."
Media replacement. A media-replacement process between time points for off-line sample collection should be validated with an automated dissolution-testing system. The media-replacement option corrects for sampling loss. This option allows a dissolution chemist to replace fresh media into each dissolution vessel after each sampled time point. The replacement media may be the primary dissolution media or a secondary media, typically used to affect the pH of the media already present in the dissolution vessel. The secondary media may be used for enteric-coated products that require a media pH change.
Automated off-line sampling collection differs greatly from manual sampling. Typically, larger sample volumes are removed for automated sampling, which has the potential to affect results for dissolutions with multiple time points. For each time point, a cumulative sample volume is removed. The total volume removed includes the flush volume, the tubing dead volume, the filter-deaeration volume, and the sample-collection volume. The flush volume is the volume of sample used to saturate the filter to prevent loss of the API on the filter. The tubing dead volume is the amount of sample that must fill the lines between the dissolution vessels and the sample-collection vials. The filter-deaeration volume is the amount of sample that is used to prepare the filter for filtration (used in certain automated dissolution testing systems). The sample-collection volume is the amount of sample that is collected for the off-line assay. A dissolution chemist must take into consideration the entire sample volume removed at each time point and decide whether an equivalent amount of fresh media is to be replaced into each vessel after each sampling time point. The large amounts of sample volumes removed and replaced may affect dissolution results. Potentially, large amounts of undissolved drug substance are removed for each sample, which may inaccurately lower the results of subsequent samples. Alternately, large amounts of replacement media may inaccurately dissolve the dosage form, which may affect the results of subsequent samples.
Validation is required to determine the necessity of media replacement. A dissolution chemist should perform dissolutions with and without media replacement and compare the results to manual dissolutions. The technique that produces results that more closely resemble manual results should be used in the automated dissolution test.
Instrument qualification and calibration
Equipment qualification. In addition to all of the validation work that must be completed for each product tested on the automated dissolution system, an instrument "chain of compliance" must be established and well documented for all primary, secondary, and tertiary instruments used to support GMP data generated by the automated system. Instrument qualifications and calibrations must be completed for all components on the entire automated system. These components may include the dissolution apparatus, any on-line ultraviolet or high-performance liquid chromatography instrumentation, and any ancillary equipment. The supporting equipment used to calibrate each component periodically on the automated system includes balances, weights, stopwatches, timers, thermometers, eccentricity meters, and vibration meters. Each piece of supporting equipment must maintain a documented and current calibration status. Although the company ultimately is responsible for GMP compliance when using automated instrumentation, the company may choose to follow qualification acceptance criteria established by the US Pharmacopeia, the instrument vendor, or their own company standard operating procedures (SOPs).
Initially, an installation qualification (IQ) and operation qualification (OQ) must be completed successfully and documented for each component of the automated dissolution-testing system. The customer should request from the vendor the test-script documentation that will be followed to complete initial qualifications. It is advantageous to have the compliance department review the documentation to be certain it fulfills the requirements for the instrument to be used in a GMP environment. If one chooses to have the vendor complete the IQ and OQ activities, one must be certain the company provides training documentation for their service technicians indicating that they are qualified to complete the qualification activities.
Preventive maintenance. The qualification practices do not end after the instrument is initially installed, qualified, and calibrated. Periodic preventive maintenance and calibration schedules must be established according to company SOPs. A qualified vendor is the best choice to perform the preventive-
maintenance activities for the automated instrument. These activities may include a periodic performance qualification of the instrument, which evaluates the overall performance and operation of the system. The preventive maintenance may also include the replacement of general components necessary for the continued smooth operation of the system. These components may include tubing, belts, sampling lines, lamps, and so forth. It is important to understand that the level of maintenance or repair performed on the instrument may necessitate a requalification, recalibration, or a change control. As stricter requirements are placed on the GMP environment, company SOPs should be reviewed to ensure that preventive maintenance activities do not push the instrument out of compliance.
Calibration checks. A periodic calibration schedule must be established for each component on the automated system. The schedule may include semiannual, quarterly, weekly, and daily activities designed to confirm the proper operation of the system. For the automated dissolution-testing system, the quarterly activities may include balance and temperature-probe calibrations. In addition, weekly or daily calibrations may include a quick balance check. For the dissolution apparatus, a semiannual performance calibration must be completed using USP calibrators. Trial dissolutions must be performed on disintegrating ( e.g., prednisone) and nondis-integrating ( e.g., salicylic acid) USP calibrators. Each dissolution test must pass the USP acceptance criteria established for the lot of drug tested. Semiannual physical testing must also pass USP acceptance criteria. The physical specifications include shaft and basket eccentricities, bath level, shaft verticality, and vessel and shaft centering. In addition, even though USP acceptance criteria have not yet been established for vibration, bath vibration is an important variable that should be measured periodically, especially if mechanical components have been changed on any of the components of the automated system. New mechanical components may increase bath vibration, which may increase dissolution results inaccurately. Daily physical specifications that must pass USP acceptance criteria include proper paddle–basket height, initial and final temperatures in all vessels, and shaft rotational speed (rpm).
Conclusion
Automated instrumentation for dissolution testing offers several advantages such as the ability to perform unattended testing and the ability to screen several batches with varying parameters. But, automated instrumentation also poses challenges for a dissolution chemist, including the need to have an overall understanding of the the automated system. Parameters such as filtration, system interference, carry-over, cleaning parameters, and media replacement are factors that must be addressed and validated to ensure equivalent results are obtained with manual and automated methods. The automated system can be used to generate GMP data only if all components on or supporting the system maintain a documented and current qualification and calibration status.
Acknowledgments
The author thanks Ron Mamajek, John Ballard, Ronnie McDowell, Dr. Michael Breslav, Dr. Daniel Kroon, Dr. Weiyong Li, and Dr. Brigitte Segmuller for their valuable suggestions.
References
1. " <1092> The Dissolution Procedure: Development and Validation," Pharmacopeial Forum, 30 (1), 351–363 (Jan.–Feb. 2004).
2. " <1092> The Dissolution Procedure: Development and Validation," Pharmacopeial Forum, 31 (5), 1463–1475 (Sep.–Oct. 2005).
David Fortunato is a scientist in the Chem Pharm division of Analytical Development, US, Johnson and Johnson Pharmaceutical Research and Development, LLC, Welsh and McKean Rds., Spring House, PA 19477, tel. 215.628.5098, fax 215.540.4684, dfortuna@PRDUS.JNJ.com [dfortuna@prdus.jnj.com]

Submitted: Feb. 22, 2006. Accepted: Apr. 7, 2006.
Keywords: Analytical testing, process automation, regulation validation and compliance, solid dosage forms

Correlation of Visible-Residue Limits with Swab Results for Cleaning Validation

By Richard J. Forsyth,Julia Roberts,Tara Lukievics,Vincent Van Nostrand

The use of visual inspection as a criterion for equipment cleanliness has always been a component of cleaning validation programs. Mendenhall proposed the use of only visual examination to determine equipment cleanliness as long ago as 1989 (1). He concluded that visible cleanliness criteria were more rigid than quantitative calculations and clearly adequate. The US Food and Drug Administration limited the use of visually clean criterion between lots of the same product (2). LeBlanc raised the question of whether a visible limit as the sole acceptance criterion could be justified (3). A visible-residue limit (VRL) currently is used in a clinical pilot plant for the introduction of new compounds (4, 5) in cases for which the VRL is lower than the acceptable-residue limit (ARL). The ARL is the amount of a formulation component that can be carried over to the next formulation with no pharmacological or adulteration concerns. The initial use of an active pharmaceutical ingredient (API) in the facility is followed by cleaning and a visual inspection against the previously determined VRL. Visually clean equipment means the current cleaning procedure is effective and the new API is not a new worst case that would require cleaning validation.
The same scientific rationale supports the use of VRLs in a manufacturing facility. The main difference between pilot plant and commercial manufacturing facilities is equipment size. Acceptable viewing parameters for the larger manufacturing equipment, including distance, viewing angle, and light level, consistently detect VRLs for several marketed formulations (6).
The implementation of VRLs in the pilot plant and their potential use in the manufacturing facility were additions to established cleaning programs. The cleaning programs in both the pilot plant and manufacturing facilities established validation based on swab sample data using high-performance liquid chromatography (HPLC) for analysis. Visual inspections were part of the validation, but were qualitative determinations only.
It should be possible to demonstrate the correlation between the quantitatively determined VRL of either the API or formulation on the manufacturing equipment and the analytically determined swab recovery data. The data from the two determinations should be mutually supportive as part of a cleaning process validation. Therefore, a retroactive analysis of the pilot-plant validation study compared the previously obtained swab results with the more recently generated VRLs for the subject compounds.
In addition, a current cleaning validation study conducted in a clinical packaging area included VRL as an integral part of the study. The study used documented worst-case formulations to soil the equipment, followed by cleaning according to a standard operating procedure. Visual inspection used experimentally determined VRLs. Swab samples for the appropriate compound confirmed the equipment's cleanliness. Testing for each formulation was repeated twice to validate the cleaning procedure. The final report includes a comparison between the VRLs and the swab sample results.
Although the cleaning validation passed testing ( i.e., all swab results were lower than the ARL), the swab results for several metformin samples assayed higher than the experimentally determined VRL. An investigation reconciled the discrepancy.
Retrospective analysis
The cleaning-validation study for the pilot plant selected worst-case formulations. One formulation was validated for a dry-granulation equipment train and another for the wet-granulation process. Each piece of cleaned equipment was visually inspected before swab testing and both formulations were tested three times for validation.
The compounds tested were simvastatin and rofecoxib (Whitehouse Station, NJ, Merck & Co., Inc.). Four observers viewed dried solution spots of known concentrations to determine the VRLs. The experimentally derived VRLs for the compounds were 0.485 and 0.871 μg/cm ² , respectively (4). For a swab area of 25 cm ² , the limits were 12.1 and 21.8 μg/swab. The VRLs for the base and neutral detergents used to clean the equipment were <0 .37=".37" and="and" cm="cm" g="g" sup="sup">2

( <9 .3=".3" and="and" g="g" span="span" swab="swab">Of the 78 swabs taken for simvastatin, all were below the VRL of 12 μg/swab. This finding confirms that the equipment was visually clean when swabbed. Of the 78 swabs, 64 had no simvastatin detected (less than 0.1 μg/swab), nine had less than 1 μg/swab, and 5 had more than 1 μg/swab.
Eighty-four swabs were taken for rofecoxib, and all were below the VRL of 21.8 μg/swab. Of these, 55 had no rofecoxib detected (less than 0.02 μg/swab), 23 were less than 1 μg/swab, and only 6 were greater than 1 μg/swab.
Finally, of the 69 detergent swabs taken, 68 had no detergent detected ( <3 14-="14-" and="and" g="g" had="had" less="less" remaining="remaining" span="span" swab="swab" than="than" the="the" vrl.="vrl.">
Although all 231 swabs from the cleaning validation study showed agreement between the swab results and VRLs, only 5% of the swabs were greater than 1 μg/swab. The validated cleaning process reduced residues far below both the ARL and the VRL levels. Therefore, the pilot-plant cleaning-validation study did not seriously challenge the correlation between the swab results and the VRLs.
Experimental
Cleaning validation. The current cleaning validation study established the requirements and acceptance criteria for the following equipment in the clinical packaging area: tablet counter, automatic filler, 8-track filler, and tablet elevator. The validation of the cleaning procedures was on equipment that had been used to fill and package clinical product and then cleaned. A development formulation and 500-mg metformin tablets served as representative worst-case formulations for validation. Both products are non film-coated tablets, which produce dust during the packaging process. The metformin tablets had a very high drug load (95%) and any residual dust was most likely to be API.

Table I: Formulation residue concentration and visibility.

Before the first cleaning trial, a VRL was established for both compounds. Each formulation was dispersed in methanol and spotted on 3 × 6-in. stainless steel coupons at appropriate concentrations. A stream of nitrogen dried the spots to facilitate the drying process and prevent potential degradation of the material. Oxidative degradation had occurred on several in-house compounds in the past. The areas of the dried spots determined the amount-per-unit area (μg/cm ² ) value for each spot of material (Table I). The VRL determined for the development compound was 0.51 μg/cm ² (12.75 μg/swab) and 0.97 μg/cm ² (24.25 μg/swab) for metformin.

Table II: Residual metformin (μg/swab).

The cleaning validation was performed in triplicate on representative pieces of equipment. Validation consisted of overall visual inspection and swab sampling in the areas identified in the protocol. These areas had been identified as hard-to-clean areas of the equipment to confirm that the cleaning procedure removed residue to acceptable levels as predetermined by the established 100 μg/swab (25 cm ² ) acceptance criteria for the API. The swab samples were assayed with a validated HPLC method (Agilent 1100 HPLC, Agilent Technologies, Palo Alto, CA) with ultraviolet diode array detection (Tables II and III).

Table III: Residual development formulation (μg/swab).

The cleaning procedures used the appropriate cleaning standard operating procedure to verify the repeatability of the cleaning processes to control residue. In addition, during one of the three validation trials for each piece of equipment, a seven-day "dirty" hold time established the longest allowable time between the completion of the filling and packaging process and the start of the subsequent cleaning process. The acceptable API limit for the clinical packaging equipment was 100 μg/swab (25 cm ² ). This limit was based on adulteration of the subsequent batch of packaged material. The adulteration limit was used exclusively because any material with a potential safety concern is packaged in an isolated facility and equipment cleaning is addressed on a case-by-case basis. In addition, product contact with the equipment was limited in this case compared with manufacturing equipment because the formulation was in final market image.
Investigation. The investigation into the VRL and swab numbers for metformin targeted the assay method, the cleaning method, and the VRL determination of the API and formulation. The swab samples' testing was by HPLC and the assay run was reviewed as well as the method validation data. The cleaning process used in the clinical packaging area was reviewed for potential chemical interactions. The VRL determination of metformin was investigated for potential issues.
Results and discussion
Cleaning validation. A cleaning-validation study was completed successfully in a clinical packaging facility. The study used the worst-case packaging situations and all swab samples assayed below the ARL of 100 μg/swab (see Tables II and III) for the three executions (including 7-day idle time before cleaning). The cleaning processes used in the clinical packaging area are considered validated for hold times as long as 7 days.
The only problem was the discrepancy between the swab results for metformin and the VRL. The VRL data indicated that anything greater than 25 μg/swab should have been detected visually. A visual inspection of the equipment before sample swabbing concluded that it was visually clean.
Investigation. The investigation into the discrepancy between the metformin VRL and swab data began with the HPLC assay of the samples. Nothing was consistently present in the samples to suggest contamination of the swabs or solvents because the majority of the samples (58 of 69) had assay values below the VRL (see Figures 1a, 2a). Nothing in the sample chromatograms indicated the presence of an intermittent extraneous peak, which might explain the high results. The chromatogram of a control containing a swab, swabbing solvent, and extraction solvent showed no peaks. The system suitability data were within testing parameters. The HPLC assay did not indicate a potential cause.
The method-validation data and documentation review indicated that all validation parameters were within accepted guidelines, including recoveries from stainless steel coupons. Forced degradation studies of metformin showed no degradate peaks to explain the observed swab results.
The cleaning process for equipment in the clinical packaging area consisted of an appropriate amount of equipment disassembly such that all surfaces were accessible. A dry vacuum removed most residue from the equipment. The equipment surfaces were cleaned with an appropriate solvent and then dried before reassembly. A sample of metformin prepared in the cleaning solvent showed no evidence of degradation.

Table IV: Metformin visible residue limit recovery study.

A review of the VRL determination for metformin noted that the VRL samples were dried under nitrogen to prevent potential oxidation of the material, which had been noted on a previous compound. The VRL study was repeated using API and formulation dispersed in alcohol, but the samples were air dried rather than under nitrogen to simulate drying conditions in the packaging area. Sample concentrations were at the VRL of 25 μg/swab level and at the approximate level of the highest sample obtained during the validation studies, 75 μg/swab. Spots were applied in triplicate to 3 × 6-in. stainless steel coupons.

Figure 1

A visual examination of the metformin residues confirmed that the 25-μg/swab spots were visible. The 75-μg/swab spots exhibited a proportionally greater amount of visible residue. The metformin residues were swabbed and assayed by HPLC. The swab results from the formulation (see Table IV) matched the nominal residue amounts. Nonetheless, several swab results for the 25-and 75-μg API spots were significantly higher. As with the HPLC assays for the cleaning validation samples from the packaging area, no contaminants or extraneous peaks were present to explain the intermittent high results.

Figure 2

Diode array scans were taken of two of the 75-μg API samples; one with the expected recovery and one with high recovery. The scans were not significantly different (see Figures 1b, 2b), indicating that the peaks were related, if not the same. A peak purity profile for the sample with the expected recovery indicated that the peak was pure. The peak purity scan for the sample with higher than expected recovery indicated that more than one species was present in the scan (see Figures 1c, 2c). Mass spectrometry (MS) and nuclear magnetic resonance analyses of the samples were inconclusive. The sample concentrations were too low for positive identification.

Figure 3

In portions of some samples, the data indicated that one metformin double bond rearranged to form a set of conjugated double bonds (see Figure 3). The conjugated double-bond system exhibits greater ultraviolet absorbance than metformin where the double bonds are not conjugated. This finding explains the inconsistent, high swab recoveries, the single HPLC peak, and the diode-array impure peak composition. The manufacturer and most literature sources cite the nonconjugated version as the chemical structure for metformin. But, at least one source (7) displayed the alternate structure in Figure 3 as the chemical structure of metformin, supporting the conclusion deduced from the available data.

The swab process.

Several compounds have been tested that degraded during recovery studies. The compounds were known to be susceptible to either by oxidation or hydrolysis. Metformin was the first example of a compound that demonstrated increased absorbance as a result of the cleaning process. Any chemical change during recovery studies will affect recovery results as well as swab assays taken after routine cleaning. The current study demonstrated the value of establishing the VRL along with swab recoveries for cleaning validation. Conclusions
The correlation between swab assay results and visible-residue limits (VRLs) for cleaning validation was examined. A review of previously completed validation studies was inconclusive because swab results were much lower than the more-recently determined VRLs. A current cleaning-validation study evaluated both swab testing and VRLs. Unexpectedly, high swab results led to an investigation that showed the value of establishing the VRL in conjunction with swab recoveries for cleaning-validation programs.
Richard Forsyth* is an associate director in Global Clinical GMP Quality with Merck & Co., Inc., WP53C-307, West Point, PA 19486, tel. 215.652.7462, fax 215.652.7106, richard_forsyth@merck.com [richard_forsyth@merck.com]
Julia Roberts is a research chemist in Vaccine Pharmaceutical Research, Tara Lukievics is a global sourcing project leader in the Global Clinical Supplies Organization, and Vincent Van Nostrand is a research chemist in medicinal chemistry with Merck & Co., Inc.
*To whom all correspondence should be addressed.
Submitted: May 16, 2006. Accepted: May 25, 2006. Keywords: cleaning, validation, visible-residue limits
References
1. D.W. Mendenhall, "Cleaning Validation," Drug Develop. Indust. Pharm. 15 (13), 2105–2114 (1989).
2. US Food and Drug Administration, Guide to Inspection of Validation of Cleaning Processes (Rockville, MD, Office of Regulatory Affairs, 1993).
3. D.A. LeBlanc, "'Visually Clean' as a Sole Acceptance Criteria for Cleaning Validation Protocols," J. Pharm. Sci. and Technol. 56 (1) 31–36 (2002).
4. R.J. Forsyth, V. Van Nostrand, and G. Martin, "Visible Residue Limit for Cleaning Validation and its Potential Application in a Pharmaceutical Research Facility," Pharm. Technol. 28 (10), 58–72 (2004).
5. R.J. Forsyth and V. Van Nostrand, "The Use of Visible Residue Limit for Introduction of New Compounds in a Pharmaceutical Research Facility," Pharm. Technol. 29 (4), 134–140 (2005).
6. R.J. Forsyth and V. Van Nostrand, "Application of Visible Residue Limit for Cleaning Validation in a Pharmaceutical Manufacturing Facility," Pharm. Technol. 29 (10), 152–161 (2005).
7. National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, www.nvbi.nlm.nih.gov.


Table III: Residual development formulation (μg/swab).
Figure 2
Table I: Formulation residue concentration and visibility.
Table II: Residual metformin (μg/swab).
Table IV: Metformin visible residue limit recovery study.
Figure 1
Figure 3
The swab process.