Marzena Ingram is senior manager, quality and compliance at Eurofins CDMO. She led and hosted several cGMP audits, and has diverse quality assurance and technical operations experience. 

Marzena Ingram

The pharmaceutical regulatory landscape and pharmaceutical development have been reshaped by International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) Harmonised Guidelines. ICH guidelines Q8–Q12 (1–5), and those in development such as ICH Q14 (6), have applied science, risk management, and quality systems to enhance process and product quality. In 2011, FDA’s revised process validation guidance (7) extended ICH’s concepts to pharmaceutical product lifecycle, an approach that is now being applied to cleaning validation and other areas of manufacturing, as well as to clinical trials (8).

Legacy drug products developed before any these guidelines were established, however, have not undergone the same rigor during development as newer products. As a result, companies must categorize legacy products as being in Stage 1, 2, or 3 of the lifecycle and make a data-driven assessment to determine whether remediation or one of the following plans must be developed:

Stage 3b monitoring (for formulations deemed to be robust)

Stage 3a monitoring that calls for additional data to be gathered (in order to determine robustness)

This article proposes an independent structured review process for legacy products—those pharmaceuticals that predate FDA’s updated process validation and quality by design [QbD] guidance. This approach uses data to develop insights into product sustainability and risks. It focuses on the evaluation of intra- and inter-batch variability for uniformity and other critical quality attributes (9–10) and can be applied to small-molecule manufacturing processes.

This review process involves comprehensive process assessment (i.e., historical data review and the gathering of end-to-end manufacturing process data); extensive process monitoring, and characterization exercises that enable lifecycle quality gaps to be addressed.

By allowing sound pharmaceutical manufacturing science principles to be applied, as well as data analysis, this assessment permits legacy product quality to be on par with that of any products developed by applying QbD approaches. Going through this assessment process results in a heightened level of product and process knowledge, while substantially reducing the risk of potential failure.

Based on the 2011 FDA process validation guidance and relevant ICH guidelines, this assessment method requires the use of analytical and interpretive methods based on standard statistical and data visualization tools. The approach allows results to be verified at any time by both product owner and regulator, utilizing standard statistical tools. All data sets prepared, transposed, and analyzed should be subject to data verification procedures to assure data accuracy and integrity.

Documents required for the assessment include the following:

executed batch records for process and packaging

current specifications for in-process, finished product, and stability

analytical test methods and test method validation

analytical results (in-process,
finished product, microbiological, stability)

process performance qualification (PPQ) protocols and reports

The assessment uses current regulatory standards for validation (e.g., from FDA, the European Medicines Agency, Health Canada, and the World Health Organization). It also uses tools and methods such as cause-and-effect Ishikawa diagram, scientific rationale checklists, decision trees, as well as supporting analytical and statistical tools such as Minitab or JMP software.

The assessment focuses on analyzing the information and knowledge gained from the three stages of the process validation lifecycle: process design, process performance qualification, and continued process verification during commercial manufacturing. It is performed in the following two parts:

Part A involves a holistic regulatory compliance review solely focusing on the compliance with the lifecycle-associated regulatory guidance—such as ICH Q8, Q9, and Q10 requirements—and future industry trends—such as ICH Q12, Q13 (11), and Q14—and how they work in tandem

Part B is a diagnostic study of the current process control (a seven-step product assessment process) that results in determining the adequacy of the current control strategy.

Part A. Compliance review: Adherence to FDA process validation and ICH guidelines

A comprehensive compliance review is designed to summarize and confirm compliance with the FDA process validation guidance and ICH and other emerging guidance, based on observation and the available data. The recent ICH Q12 document allows use of lifecycle guidelines for seamless post-approval change management. Adherence to lifecycle guidance, therefore, becomes more important for ongoing operational flexibility and compliance.

Stage I, process design assessment, involves the following steps:

determining the quality target product profile and justification of critical quality attributes and critical process parameters (as per ICH Q8)

applying quality risk management tools (as per ICH Q9)

documenting a critical manufacturing equipment suitability assessment

determining degree, and understanding variability

determining the availability of scale-up trials to determine effects of scale.

Stage 2, process performance qualification review, covers aspects such as:

design considerations of a facility and qualification of utilities and equipment

validation of analytical methods per analytical QbD (ICH Q 14)

utilization of cumulative data from all relevant studies to establish the PPQ manufacturing conditions

number of batches and samples to provide sufficient statistical confidence

application of statistical process performance indicators

clear, documented justification for manufacturing process to be in a state of control.

Stage 3, continued process verification of the product, includes the following:

Continued monitoring and sampling of critical process parameters and quality attributes is at the level established during the process qualification stage until sufficient data is available.

An ongoing program is in place to collect and analyze product and process data, detecting unplanned departures from the process.

The data collection plan and statistical methods to determine process stability and capability is adequate in identifying variability in the process to signal potential process improvements.

Review of failures, non-conformances, and trends during commercial manufacturing are conducted.

Part B. Control strategy review: The seven-step legacy product assessment 

The seven steps involved in Part B are as follows:

Identify sources of variability (Step 1).


Variability is caused by contributing factors associated with materials, methods, machine, measurement, manpower, and mother-nature (12). In pharmaceutical manufacturing, this relates to attributes of material, formulation and associated processing parameters including scale-up factors, processing equipment, test method, personnel, and facility/utilities (

Figure 1. Sources of variability. CQA is critical quality attribute. All figures courtesy of the authors.

The overall critical quality attribute (CQA) variability is a function of individual component variability.

A cause-and-effect technique is utilized to identify underlying factors that could have an impact on CQA that would affect patients (i.e., label claim CQAs). The assessment includes factors and attributes of materials, formulation, processing parameters and scale-up, processing equipment, method and analytical, personnel, and facility/utilities as shown in 

Figure 2. Ishikawa example (material attributes, formulation, process parameter/scale, equipment, method/analytical, personnel, and facility/utilities/environmental). CMA is critical material attribute. CQA is critical quality attribute.


The rationale for the criticality assignment is discussed qualitatively (

). Upon identifying the risk factors through cause-and-effect techniques, a risk-rating decision tree (

) is used to analyze the legacy process upon risk assessment. The criticality assignment will be commensurate with the product/formulation characteristics. In addition to any product-specific CQAs that would affect the patient, the review should, at a minimum, assess the impact of variability factors on CQAs (i.e., assay, uniformity, weight variation, dissolution, and related substances). Process variability may affect the assay of the drug product, thus the impact of sources of variability on assay should be evaluated. Variability in uniformity and weight variation can affect safety and efficacy.

Figure 3. Example section of factors affecting uniformity variability. CQA is critical quality attributes. CPP is critical process parameters. PSD is particle size distribution. Ref. Std. is reference standards.

Formulation, process, and other variables may impact content uniformity and weight variation. Thus, the impact of sources of variability on content uniformity should also be evaluated. Related substances can impact safety and must be controlled based on FDA, compendia, or ICH requirements. Formulation, process, and other variables can impact related substances. Risk assessment must cover all the formulation and process attributes as shown in the below example for uniformity.

The decision tree with a data-driven semi-quantitatively risk-rating system (

) is used to prioritize the unexamined influencing factors labeled in yellow (

) identified for each element. Note that the example provided in Figure 3 pertains to factors affecting one quality attribute only (uniformity). Additional review is warranted when the rating value is higher.


The decision tree-based assessment determines whether the established current control strategy for each attribute is supported by the cumulative documented data from relevant Stage 1 studies (e.g., design of experiments; ranging studies, laboratory or pilot batches). The control strategies established to control and mitigate the identified risk factors are then reviewed to determine the missing parts.

This gap assessment also evaluates whether the established current control strategies are derived based on objective information/rationale and data from original and subsequent process design studies. The need for establishing additional controls, including manufacturing process controls, is determined at the end of the gap assessment.

The gap assessment also determines whether the established current control strategy is supported by the cumulative data from all relevant studies (e.g., designed experiments; laboratory, pilot, and PPQ batches). The control strategies established to control and mitigate the identified risk factors are reviewed to arrive at the conclusions. The residual risks and the need for establishing additional controls, including manufacturing process controls, are then identified to initiate a heightened sampling and testing Stage 3a plan for the product.

Lifecycle review and Stage 3a monitoring batch execution (Step 4).


The lifecycle review should determine, based on available lifecycle data, whether the current process is operating consistently within the originally established control strategy. The control strategy gap assessment step also may identify the need for additional controls or identify gaps in supporting data. The lifecycle review will determine if any commercial product data supports the identified gaps. The data from all three stages of the process validation lifecycle (e.g., design of experiments, manufacturing trials, statistical analysis, annual product review, and commercial batch data) may enable the development of a risk-based control strategy that would suitably minimize sources of variability. Data from previous credible experience with sufficiently similar products and processes may also be utilized to support the control strategy.

Data gaps for legacy products at Stage 1 QbD, Stage 2 equipment/facility/utility qualification and PPQ, and Stage 3 continued process verification require that a sampling and testing plan be developed under a Stage 3a protocol, to further enable prediction with adequate statistically significant data sets.

The type of Stage 3a sampling/testing plan developed is based on the product/process gaps that have been identified. The Stage 3a study focuses on processes that have not been characterized, or for which variability has not been determined. Extensive data must be collected and tested, and a monitoring protocol executed by the technical operations team responsible for Stage 1 lifecycle of the product. The Stage 3a lifecycle monitoring data that are collected supports the identified gaps or enables closure of existing Stage 1 data gaps for the product.

Data-driven, scientifically sound Stage 3a monitoring enables the following:

continual enhancement of the product control strategy

collection and evaluation of information and data, such that sources of variability can be detected, as recommended by FDA process validation guidance

closure of the characterization gaps that were not addressed during legacy product development and to meet QbD standards, per ICH Q8 (R2)

determination of intra and inter-batch uniformity as recommended by FDA process validation guidance

development of a baseline for Stage 3b monitoring as recommended by FDA’s process validation guidance.

ICH Q10 recommends management of product and process knowledge from development through commercialization and until discontinuation. The monitoring batches add to the body of knowledge on the product. FDA’s process validation guidelines allow for use of credible experience with sufficiently similar products, hence a bracketing approach for monitoring may be applied and conclusions can be used for control strategy enhancement of similar products.

The Stage 3a data collection protocol (13) developed after gap assessment enables in-depth statistical assessment of the legacy product. The Stage 3a assessment will also provide a baseline for continual Stage 3b program for the product. The approach supports the organization’s ability to make science- and risk-based continual process improvement decisions for the product, as well as to identify areas where QbD development should focus for similar manufacturing processes or products. The Stage 3a monitoring measures are developed in line with upcoming revisions to ICH Q2/Q14 and the proposed I General Chapter <1220> framework for analytical QbD, which requires identification of adverse trends, allowing proactive measures and facilitation of continued improvements and change control through continued monitoring.


The extensive data collection step includes batch data from development, process performance qualification, and commercial manufacturing. The new database will include potential variables that can impact the CQAs. This will include, but will not be limited to, API and excipient material attributes, process parameters, equipment parameters, hold times, operational variables, operation/analytical personnel, instrumentation, documented weights/time, facility/utility parameters, and incidents. This step will involve extensive data set collection, data set review, and data preparation for the subsequent statistical analysis.

As a result, the database must be developed so that it maximizes the utilization of existing product/process data for multivariate statistical analysis such as correlations. The database is created for variables that can impact CQAs. The database is utilized as a starting point for Stage 3b trending and ongoing signal detection of the legacy formulations that are determined to be robust post assessment.


The data collected from Stage 3a heightened sampling and the historical batches are statistically analyzed to identify patterns and insights into the impact of potential sources of variability on the CQAs. Statistical software tools such as Minitab and JMP may be utilized. The analysis determines the adequacy and suitability of the collected data sets to support the current control strategy. The analysis first attempts to determine whether the original control strategy developed for the legacy product adequately meets the current standards.

Enhancement requirements are then determined, based on the correlations and trends observed from batch data across commercial manufacturing. This step includes an assessment of within-batch and between-batch variability, and the associated controls available to minimize the variability.

The statistical assessment can identify patterns and insights showing the impact of potential sources of variability on the CQAs. The correlation between the uniformity results, such as blend uniformity (BU) or content uniformity (CU), can suggest whether adequate mixing is happening and therefore justify continuing stratified CU sampling in place of error-prone and disruptive BU sampling. A further statistical criterion can also be developed using the assessment: removing the CU sampling once adequate uniformity data have been generated. To augment the historical data, for example, the variability of uniformity is determined with Stage 3a monitoring batch data generated (14). Correlation of process parameters are conducted to determine the degree of their influence on the CQAs at manufacturing stages.

The statistical assessment tries to address each of the attribute gaps identified though Stage 3a heightened monitoring and analysis. The rating table is then updated based on the newly identified and generated data, as shown in 


Interpretation generates actionable recommendations based on the statistical analysis. The data-driven estimations allow for in-depth product/process understanding, proving product robustness and are in line with regulators’ expectations that science and data be used for product and process decision making. The determination is based on data signals identified through assessment of the historical data and the generated heightened results from Stage 3a monitoring batches.

When adequate data are available, the signals seen (i.e., CQAs, key performance parameters, and CPPs), can be categorized based on risk. All signals are not created equal, and different action plans are required for the yellow flags. The assessment suggests actions based on practical relevance and the statistical strength of the signal (15). As a result, this review process enables organizations to make data-driven justification on the product sustainability while implementing continuous improvement remediation for legacy drug products.

The protocol-driven legacy product assessment closely dissects the manufacturing process to determine the current state of control, to comprehend and address Stage 3a gaps for completeness of validation status. The exercise also determines the process robustness, allowing for business decision making. The review provides clear data-driven input for one of three actions:

developing a Stage 3b monitoring program, in cases where product has been determined to be robust,

developing a Stage 1 continuous improvement plan, or

Articles

A data-driven strategy can assess the quality of legacy drugs developed before 2011 process-validation requirements were established.

Assessing Legacy Drug Quality 

To provide insight into regulatory trends and factors contributing to noncompliance, the authors analyzed data from Form 483 observations issued by FDA during routine investigations of finished drug and API manufacturing sites between 2014 and 2018. The goal was to provide pharmaceutical quality compliance professionals, managers, and operators with insights into how they can best strengthen compliance and develop appropriate solutions for key regulatory compliance issues. This article summarizes results of the study.

	To review the basics, form 483s are issued by FDA’s Office of Regulatory Affairs (ORA), which is responsible for field activities such as site inspections and enforcement (1). ORA inspectors issue Form 483 observations when they find current good manufacturing practice (cGMP) violations during a facility inspection. These observations are then presented to the company’s senior managers to notify them of the need to correct compliance deficiencies.

	Inspection observations are crucial, because they define the site’s state of compliance. Companies that receive 483s are expected to ensure that steps are taken to correct problems and address the observations and any auxiliary systems that are also deficient but not cited. The final establishment inspection report (EIR) then determines the resulting regulatory actions, considering the observations, proposed corrective measures, and evidence collected.

	Three different types of observation classifications (2) are possible:

		No Action Indicated (NAI) indicates that no objectionable conditions were found during inspection and that no further regulatory action required.

		Voluntary Action Indicated (VAI) shows that objectionable conditions were found, but the agency is not yet willing to take any regulatory action.

		Official Action Indicated (OAI) shows that ORA will recommend specific actions.

	This study used data from FDA’s Inspection Classification Database, which was established in 2009 as part of a transparency initiative (3). The analysis used data classified as final, which is updated by FDA and posted on the agency’s website each month. Any undisclosed inspections are not captured in the list. The inspectional observation data set files for years 2014, 2015, 2016, 2017, and 2018 provided the full list of audit observations. For instance, the 2014 list of inspectional observations included inspections that ended between Oct. 1, 2013 and Sept. 30, 2014.

	FDA recorded a total of 2997 audit observations in 2014, including 645 Form 483s issued for finished formulation and API sites, which, in turn, were classified into nine categories (4):

		Subpart E, control of components and drug product containers and closures

		Subpart F, production and process controls

		Subpart G, packaging and labeling control

	The inspection classification database detailed the number of official and voluntary actions (OAIs and VAIs), whether at US manufacturing sites or at sites outside of the United States.

	The authors used the Pearson correlation coefficient to identify any positive correlation between the variables, where positive correlation indicates that as the values of one variable increase the values of the other variable increase (5).

	Between 2014 and 2018, the total number of FDA Form 483 audit observations was found to range from 2997 to 3626, with an average of 3362 observations per year. The agency issued a total of 3424 Form 483s, an average of 685 per year. However, 716 483s were issued in 2018, a higher number than seen in previous years (

 summarize the nine categories of inspection observations and associated subsections.

The data provide a clear indication that most 483 observations were related to Subpart I–laboratory controls, which accounted for 2603 observations, and Subpart J–records and reports, with 2530 observations. The other top contributors were Subpart F–production and process controls (with 2088 observations), Subpart B–organization and personnel (with 1902 observations), and Subpart D–equipment (with 1621 observations).

 illustrates the percentage changes in the identified observation categories. The most relevant fluctuations would be the increase in FDA observations for packaging and labeling control (Subpart G) in 2017 and 2018. Also noteworthy is the spike in equipment (Subpart D) observations, which went from 311 to 370 between 2017 and 2018. However, laboratory control observations decreased during the same period, which may reflect the implementation of improved system and data integrity laboratory controls.

VAI and OAI are critical for companies, because they indicate non-compliance and the need for critical actions, which may have implications for product quality and supply. Data showed that there has been a consistent reduction in the issuance of VAI and OAI since 2017, suggesting a higher rate of compliance with regulatory guidance requirements, as found during inspections.

	The finished formulation and API site inspection trend (

) indicates a declining domestic inspections trend; however, there was no corresponding increase in foreign site inspections, which may reflect the fact that FDA has adopted a risk-based inspection management process. The data were also analyzed to determine whether there were any correlations between the variables: number of VAI/OAI, number of domestic inspections, and number of foreign inspections.

	Based on the Pearson correlation coefficient calculation of X values (domestic inspections) and Y values (number of VAI/OAI), the value of R is 0.9443, suggesting a strong positive correlation between X and Y variables. The p-value is 0.015648, hence statistically significant (<0.05). The correlation was not seen when VAI/OAI numbers were compared with foreign inspection trend. It can be, therefore, concluded that there is a higher tendency to generate VAI/OAI observations if the domestic inspection rate increases.

	The Pearson Correlation Coefficient for individual GMP observations associated with the nine subparts (B through J) were analyzed for 2014 through 2018 (

). A positive correlation could be ascertained across laboratory control observations (Subpart I) and records and reports (Subpart J). This may suggest that increased regulatory scrutiny of laboratory controls may have resulted in improved good documentation practices, because there appeared to be an improvement in records and reports observations as well. A summary of statistics is found in 

Although additional analysis might be required to confirm the correlation, it is safe to assume that laboratory data integrity-based inspection assessment extends from GMP sections Subpart J to Subpart I. As an inspection readiness strategy, organizations should provide equal weight to laboratory electronic controls, procedures, specifications, and methods as well as to data access and integrity (6), and to the storage of raw data and reports, whether electronic or paper.

	Analysis of FDA Form 483 observations issued from 2014 to 2018 revealed that the number of 483 forms issued has increased during this time period, while the number of observations has declined. The number of VAI and OAI observations issued has also fallen since 2017.

	Over this period, there has been a decrease in domestic inspections, and a positive correlation was found between the number of domestic inspections conducted and the number of VAI/OAI observations issued. The highest number of observations were related to Subpart I–laboratory controls and Subpart J–records and reports, Subpart F–production and process controls, Subpart B–organization and personnel, and Subpart D–equipment.

	Increased inspection of laboratory controls (Subpart I) may also result in increased good document practices, records and reports observations (Subpart J) based on the positive correlation observed. It is also noteworthy that there was an increase in packaging and labeling control (Subpart G) observations in 2018.

Note: This article was prepared by the authors in their personal capacities. The opinions expressed are the authors’ own and do not reflect the view of their employer, government, or any agency with which they are affiliated.

	1. M. Hamburg, Justification of Estimates for Appropriations Committees Fiscal Year 2016, US Department of Health and Human Sciences, FDA (2015).
	2. FDA,

 Integration of FDA Facility Evaluation and Inspection Program for Human Drugs: A Concept of Operations

 (June 6,  2017).
	3. FDA, Transparency Initiative: Draft Proposals for Public Comment to Increase Transparency by Promoting Greater Access to the Agency’s Compliance and Enforcement, FDA Data Transparency Task Force, US Department of Health and Human Services (2011).
	4. Code of Federal Regulations, Title 21: Part 210, 211: Food and Drugs, Current Good Manufacturing Practices.
	5. J. Freeman, T. Young, Correlation Coefficient: Association Between Two Continuous Variables, Scope Tutorial, (2009).
	6. D. Hodgson, F. Maini, et al., Under the Spotlight: Data Integrity in Life Sciences, (Deloitte, London, UK 2017).

	Ajay Pazhayattil is a management consultant (ajpazha@gmail.com). Marzena Ingram is senior manager, quality and compliance at Eurofins CDMO. Naheed Sayeed is deputy director, validation services at Sanofi.

Vol. 43, No. 10
	October 2019
	Pages: 56–61

	When referring to this article, please cite it as A. B. Pazhayattil, M. Ingram, and N. Sayeed, “Lessons from FDA 483s and cGMP Inspection Data,” 

Production and process controls, organization and personnel were the top problems found, while packaging and labelling citations increased in 2017 and 2018.

Lessons from FDA 483s and cGMP Inspection Data

A year’s worth of FDA warning letters suggest that API and finished drug manufacturers should strengthen their approach to continued process verification.

	In 2011, when FDA revised its process validation guidance (1), it introduced the idea that validation must continue throughout the lifetime of a pharmaceutical product. FDA advocated three stages:

		Process design, using the principles of pharmaceutical quality by design (QbD)

		Process performance qualification (PPQ)

	Pharmaceutical manufacturers have been developing novel methods and solutions to support this new definition of process validation, and FDA has been encouraging science- and risk-based approaches for all three stages.

	Other regulatory agencies have also embraced the lifecycle approach to process validation, including the International Council for Harmonization (ICH) in its Q8, Q9, and Q10 Questions and Answers (R4) (2) guidance for industry. ICH views the entire product lifecycle as: development, technology transfer, manufacturing, and discontinuation (3). 

	Best practices call for approaches such as:

		Taking a design-of-experiments (DoE) approach to product development, sampling, and testing plans

		Using data-driven statistical tools for CPV. 

	The European Medicines Agency’s (EMA) 2014 process validation guidance, the World Health Organization’s 2015 Appendix 7, and the Pharmaceutical Inspection Cooperation Scheme (PIC/S) 2015 Annex 15 all reflect the expectation that process validation should continue throughout the lifecycle of the process. Pharmaceutical manufacturers would do well to take an integrated lifecycle approach to process validation that works across multiple markets. 

	As the industry’s approach to lifecycle-based process validation has evolved, some companies now divide Stages 1, 2, and 3 into the following: 

		Stage 1A – Product Development, Stage 1B- Scale Up

		Stage 2A – Equipment/System/Utility/Facility Qualification, Stage 2B – Process Performance Qualification

		Stage 3A–a heightened verification program for newly introduced molecules and Stage 3B-routine continued process verification.  Implementing the process validation lifecycle approach globally demands that pharmaceutical manufacturers be especially careful in selecting the most effective statistical tools and Stage 3 strategies (4).

	A good knowledge-management strategy is essential, starting from Stage 1 to Stage 3 of the process validation lifecycle. Scientifically identifying sources of variability requires a robust QbD-based product development program along with a Stage 3 CPV program that allows for continual gathering of product and process knowledge. 

	Process drifts, or times when the process fails to perform at its best, trigger identification of sources of variability, ensuring improved control and allowing for continuous improvement. Ongoing collection of product and process knowledge is required for continuous improvement, and also to improve the development process for similar products (5).

	This article asks how well pharmaceutical manufacturers have been addressing new process validation requirements by analyzing FDA warning letters (WLs) from 2017 to date. The goal of this study is to understand which parts of the process may need to be strengthened in order to improve regulatory compliance and minimize risk.

	Since the FDA process validation guidance is applicable to drug substances, drug products, and biologics, insights from WL assessments (6) can provide ample supporting evidence to justify an integrated approach to process validation.

	For this research, 85 FDA WLs issued in 2017 and 2018 were analyzed, 61 for 2017 and 24 for 2018. Inadequacies were found in both API and finished pharmaceutical facilities, but occurred mainly in finished pharmaceutical facilities, as outlined in 

	In 2017 and 2018 FDA warning letters, process validation issues were more frequently noted in finished pharmaceutical than in API manufacturing sites, with noncompliance increasing by 5% between 2017 and 2018.  Of the warning letters analyzed, 40% contained observations that were related to process validation. Since 2017, 29 finished formulations sites have been found to be noncompliant with process validation approaches, and the number of facilities that were out of compliance increased in 2018 (

	Approximately 20% of the API sites that received warning letters during the period under examination had problems that were related to PV noncompliance. Two of the five warning letters related to API sites cited inadequate process control monitoring, resulting in the inability to ensure stable manufacturing operations and consistent drug quality. 

	Regulators who had inspected the facilities identified the strong need for engaging a qualified consultant to assist the organization in meeting current good manufacturing practice (cGMP) requirements. Another warning letter cited inadequacies in the sampling plan during the PPQ stage. It highlighted the need for a statistically sound sampling and testing plan as stipulated in the PV guidance. 

	In addition, inspectors found that the manufacturer was not monitoring critical process parameters adequately. In a third API process-related observation, the regulator cited that the organization did not demonstrate adequate process understanding, which includes failing to consider operating parameters for critical operations. Regulators focus on all three PV stages during an inspection. In its inspection, FDA staff called for hiring external help, especially when the manufacturer could not implement the Stage 3 CPV program independently.

	In 14 warning letters, regulators noted the manufacturers should follow FDA process validation guidance for the process validation elements. Additional assessment revealed a consistent pattern. As the warning letters make clear, the following are very important in process validation:

		Having a data-driven and scientifically sound process-validation program that appropriately identifies sources of variability

		Performing process performance qualification studies

		Establishing an ongoing program for monitoring process controls to ensure stable manufacturing operations and consistent quality. 

	The problem that inspectors most often found at finished pharmaceutical facilities was failure to establish a continued process verification program to identify sources of variability.  This gap was cited in 17 warning letters. Eight of the citations pointed out that the manufacturer did not perform PPQ studies.

	Observations about Stage 3 CPV-related operations were found across prescription, over-the-counter (OTC), and sterile manufacturing sites alike. The regulator used consistent language and quoted from the 21 US Code of Federal Regulations (CFR) Part 211, Subpart F- Production and Process Controls, Sec. 100–Written procedures, deviation (a). The 21 CFR 211.100 (a) section states: “There shall be written procedures for production and process controls designed to assure that the drug products have the identity, strength quality, and purity they purport or are represented to possess. Such procedures shall include all requirements in this subpart. These written procedures, including any changes, shall be drafted, reviewed, and approved by the appropriate organizational units and reviewed and approved by the quality control unit” (7).

	There were five instances in which regulatory inspectors noted a failure to prevent microbial contamination in sterile manufacturing, including lack of established procedures as well as inadequate validation of aseptic manufacturing processes and sterilization. The drug manufacturer is required to implement elements of cGMPs for sterile drug products using aseptic processing. Additional observations that were cited included deficiency in performing smoke studies for classified areas, which include Restricted Access Barrier Systems (RABS).

	For this assessment, warning letter trends were analyzed to provide manufacturers with some insight into regulators’ objectives. Both API and finished pharmaceutical manufacturing sites are consistently being cited for violating best practices in recent process validation guidance.  Analysis of the observations confirms the need for a well-defined continued process verification program that continually monitors the sources of variability and enables an organization to embark on continuous improvement initiatives prior to encountering quality issues. 

	Regulators around the world are encouraging the use of data-driven, science-based approaches to process validation. In some cases, inspectors have suggested that manufacturers enlist the help of qualified consultants, which will require that they invest time and resources in establishing a Stage 3 CPV program and supporting team. 

	More biopharmaceutical manufacturers are now striving for data-driven approaches to Stage 3 process validation, and solution providers are offering integrated software to help manage such programs. The convergence of regulatory guidance, current inspection trends, and the emergence of solutions allows existing monitoring programs to be enhanced, but there will likely be a need for additional elaboration and training across quality, operations, R&D, and regulatory functions. 

	Ensuring that process validation programs and training bridge different functional groups within the organization helps all groups understand their role in enabling continuous improvement. This kind of approach will help manufacturers prevent process failures, and can also help guard against having people overreact to individual events. It can also ensure that they catch sources of variability that might otherwise go undetected. 

: This article was prepared by the authors in their personal capacity. The opinions expressed are their own and do not reflect the view of their employer, government, or any agency with which they are affiliated.

Guidance for Industry: Process Validation-General Principles and Practices

, (Silver Spring, MD, Jan 2011).
	2. ICH, 

Q10 Harmonized Tripartite Guideline, Guidance for Industry: Q8, Q9 and Q10 Questions & Answers (R4)

(ICH, November 2010), ich.org.
	3. ICH, 

Q10 Harmonized Tripartite Guideline, Pharmaceutical Quality Systems

, September 2017.
	5. Parenteral Drug Association (PDA), “How Global Pharma Organizations Can Achieve Process Validation Success,” 

Warning Letters and Notice of Violation Letters to Pharmaceutical Compliance, Silver Spring, MD

Title 21, Food and Drugs, 21 Code of Federal Regulations 211.100(a)


	Supplement: Outsourcing Resources
	Vol. 42
	August 2018
	Pages: s22–s25

	When referring to this article, please cite it as A. Pazhayattil, N. Sayeed-Desta, and M. Ingram, “Lifecycle-Based Process Validation Emphasizes the Need for Continued Process Verification," 

Pharmaceutical Technology Outsourcing Resources Supplement

	Ajay Babu Pazhayattil is director, quality and compliance, at Eurofins Alphora Inc. (ajay.pazhayattil@alphoraresearch).He has held key management roles with brand name, generic, and contract manufacturing pharmaceutical organizations. Naheed Sayeed-Desta has experience in product development, technical services, and validation. Marzena Ingram has extensive quality assurance and technical operations experience. She developed a specialized continued process verification team and spearheaded the implementation of a product lifecycle program.