Warning letter / Centaur / July 2023 / Cleaning deficiencies, failure of Quality unit

1. Failure to clean equipment to prevent contamination / cross contamination

During inspection, residues of different product were observed in direct and indirect contact surfaces of manufacturing equipment. Firm acknowledged some sections of the equipment were not cleaned and verified since they were installed more than 14 years ago. Analytical testing of the residues showed multiple active ingredients. The residue samples collected after running placebo batches and subsequent cleaning also demonstrated active ingredient cross contamination on surfaces. The equipment also caused backflow leading to equipment contamination. Following the observations, Centaur informed its client Breckenridge Pharmaceutical Inc., and initiated recall of several within expiry batches of Alprazolam and Clobazam tablets.

Further in communications with FDA, Firm stated it has improved cleaning and verification processes. It also stated based on testing of placebo batch samples and control samples, it can be concluded there is no product quality impact on commercially distributed batches. FDA contended that the response is inadequate as cross contamination cannot be uniform, it did not sufficiently address contamination recovered from product contact surfaces; Other locations and sampling may reveal higher levels of contamination. Firm is asked to provide comprehensive, independent retrospective assessment of cleaning effectiveness covering each piece of manufacturing equipment, CAPA plan addressing improvements to cleaning program, cleaning verification and cleaning validation.

2. Failure to establish an adequate quality unit

Quality unit (QU) failed to adequately implement the facility’s quality function and ensure quality oversight.

A. Equipment cleaning procedures lacked requirements to clean equipment sufficiently -sections of equipment were not cleaned since installation more than 14 years ago; visual verification procedures for ensuring cleanliness failed to identify visible contamination. In the response to FDA Firm stated cleaning and visual verification procedures are improved and updated; but failed to provide evidence that CAPA measures are implemented to ensure written procedures are sufficiently written and reviewed prior to implementation.

Firm is asked to provide a comprehensive assessment and remediation plan to ensure:

• Quality Unit is given authority and resources to effectively function,
• All procedures used by the Firm are robust
• QU oversight throughout all operations, investigations, batch release &
• authority and resources to Quality unit (QU) in discharging all other QU duties

B. Inadequate Investigations

Investigation of extraneous peaks in a dissolution testing by HPLC during long term stability testing was inadequate. Additional peaks eluted with some tablets but not in others from same batch, the peaks were attributed to excipients without adequate rationale. Further in the response to FDA Firm concluded the inconsistent peaks are due to insufficient saturation of filter used in sample preparation. The response did not provide assessment of (adequacy) procedures for investigation of discrepancies and actions for ensuring investigations contain adequate root cause determination, CAPA and effectiveness checks.

Firm is asked to provide a comprehensive assessment of procedures for investigation of discrepancies & action plan for significant improvement of investigation procedures, investigation competencies, root cause determination and CAPAs. Firm is also asked to provide a detailed independent assessment of all test methods for instructions to ensure repeatability and system suitability and is supported by adequate validation / verification, and test methods used in stability programme are stability indicating.

Manufacturing equipment and their operations can be complex. When cleaning procedures are developed, it should be in discussion with a cross functional team (CFT) involving Production, Engineering, QA and QC, even involving equipment vendor. The procedure should address the cleaning and cleaning verification of all sections and components of the equipment. This require a comprehensive understanding  of the design, engineering drawing and operating principles of the equipment, areas  / sections where the product residue will carry during operation and cleaning. For example equipment like FBE (fluidised bed equipment), coating machine etc can have airsupply / exhaust ducts, dust catch pots, other inaccessible areas like pipings. A CFT will help in identifying all potential areas where product residue can accumulate and cause contamination in next product.

For each type of manufacturing equipment, develop cleaning procedure logically, addressing the equipment section by section, with detailed checklist for cleaning each section (For e.g. in an FBE the fluidisation chamber, bowl, bottom dish, air inlet duct, exhaust duct, FBE bags, dust collector, external body). Similarly the checklist for visual verification should also be addressing each section. A rudimentary checklist with check points like, for example – Is FBE clean, Is external surface clean – do not give direction for the personnel who clean, who verify the cleaning for doing a thorough job. The sampling points for visual verification should cover difficult to clean and difficult to reach areas in each section of the equipment. And build awareness among the team about criticality and thoroughness of cleaning, patient impact and also the consequences of audit findings on cleanliness of equipment. (The case in discussion is a tell-tale example where the Firm recalled more than 100 distributed batches).

Extraneous peaks in a product batch analysis in batch release and / or stability analysis requires careful examination. Causes could be several – degradation of the product, contamination during analysis, sample preparation, chromatographic issues, excipient related peaks generated during shelflife, peaks due to leaching from packing materials and so on. Comprehensive investigation with appropriate investigation tools and involving a cross functional team only can identify potential causes and identify CAPAs. For example, investigation tools like Ishikawa (fishbone) analysis with brain storming by a cross functional team (CFT) following a 6M method –  Material, Method, Machine, Mother nature, Measurement, Man can help in listing various possible factors which can be a cause for the additional peaks. Each of the potential causes thus identified shall be evaluated before ruling out as a potential cause and the rationale documented. Inadequate failure investigations, especially OOS investigations is one of the most cited observations in FDA audits. Investigation procedures and checklist should be thorough and elaborate to ensure all possible causes are evaluated before concluding on the probable root causes. In out of specifications (OOS), extraneous peaks etc, if the investigations are not conclusive with clearly attributable reasons for batch failure (OOS, deviations etc), full scale investigation including manufacturing investigation (Phase II) should be mandatorily performed; it should be part of the OOS review checklist for QA, for signing off the OOS investigation. The Phase 1B investigation checklists for laboratory error shall be elaborate covering checkpoints for each type of analysis – Impurities, extraneous peaks, assay failure, dissolution failure, colour and description and so on. When each possible factor is ruled out, it shall be scientifically justified. This will help avoid discovering new causes for a failure, each time an investigation is reviewed / audited. Equally important is to improve the investigation competencies of QC, QA and Tech-Ops teams. Train the team in different investigation tools – Brain storming, Ishikawa analysis, Why-Why method, Fault tree analysis. Take up workshop / case studies using the observations in different USFDA 483s, Warning letters which are publicly available. Take up the workshops in small forums of 10-15 people, reviewing the observations and team debating whether similar situations could be potentially present in their areas. Where the debates indicate vulnerabilities, follow it up with deeper evaluation, address the issues.

Apart from this, there are practices that are useful for investigations. For stability analysis, along with a (finished) product batch load a placebo batch also for stability for all timepoints; and with each stability station analysis, the placebo batch is also injected (chromatographed) to identify if any additional peaks generated are due to sample matrix in placebo. Ensure method validations are comprehensive and meaningful – robustness studies covering different variables in sample preparation, stress studies (forced degradation studies) can identify what changes in the method and settings can cause disturbance in analysis, chromatography, peak shape and so on. Excipient compatibility studies during development can help identifying potential degradant peaks due to sample matrix.

Ensure adequate, competent resources in the Quality Unit (QU). For e.g. IPQA personnel in all shifts for verifying cleaning process and cleanliness, reporting deficiencies. Competent QA review process for coordinating investigations, review investigations. Adequate authority for QU oversight over all operations, for ensuring batch releases, investigation closures only when the investigations are complete, scientific with appropriate root cause determination; initiate remediation actions like product recall when there is potential risk / impact on product quality, integrity, and patient safety. 

As cited in the Warning letter, deficiencies in equipment cleaning practices, and the deficiencies getting identified only during regulatory inspections has serious consequences. Hundreds of batches might have to be recalled, their logistics to be handled, impact on reputation of the Firm / site. Remediation and close out of Warning letter could take very long time, impacting business.

Firms will do well to do a comprehensive re-evaluation of their cleaning processes covering each piece of equipment, review the cleaning processes by a CFT involving Production, Engineering, QA, QC, R&D. Discuss threadbare whether the cleaning and cleaning verification cover all sections of the equipment (direct and indirect contact surfaces). Inspect each piece of equipment critically by a CFT to verify and identify areas and sections which are not getting adequately cleaned. Accordingly elaborate the cleaning procedures and checklists. After such an assessment if there are significant changes in cleaning procedures, worst case products, sampling locations etc, revamp the cleaning validation and reinitiate the cleaning validation -address worst case products with respect to cleanability, solubility, toxicity, potency; review the sampling plan and sampling locations to cover the most difficult to clean and sample areas; perform recovery studies and incorporate a recovery factor into the residue acceptance limits. During such assessment if risk of contamination in distributed products are identified, perform a comprehensive impact assessment. It is not sufficient to test control (reserve) samples and conclude contaminants were not observed and hence no risk. As FDA notes in the Warning letter contamination will not be uniform and hence chances of detecting it in control samples is very low. What is important is to perform a scientific assessment of potential contamination and extent of contamination – What could be the contaminant, composition of the contaminant (for e.g. in a tablet manufacturing process, the residues in an equipment like FBE can be the granulated mixture), what could be the maximum quantity that can contaminate each batch, whether the contaminants will be uniformly distributed in the product or not (due to the manufacturing process and operation in a FBE the contaminant could be more uniformly distributed), what is maximum contamination that can happen, amount of contaminants and whether such contamination has risk of patient safety (considering extent and potency of contaminants). If the assessment shows there is no risk and reserve sample analysis also support the risk assessment, there is a justification to conclude nil or minimal impact of cross contamination. Or else, the risk and impact is not quantifiable and remediation actions like recall need to be initiated.

Similarly review all invalidated OOS investigations (of batches distributed and within expiry) to assess whether the investigations and root cause are conclusive: whether the investigations have evaluated all possible factors and documented with scientific rationale; where investigation is not conclusive whether a manufacturing investigation is performed. Where gaps are identified, reinvestigate the incidents with appropriate investigation tools, and based on investigation and root cause establish CAPAs. Perform impact assessment if any OOS are confirmed during such evaluation, and take up remediation actions.

Review and enhance the investigation procedures, checklists as elaborated above; initiate programmes for enhancement of investigation skills and competencies of Quality and Tech-Ops personnel with trainings and workshops.

Perform an assessment of all test methods, their status of validations, robustness studies, sufficiency of system suitability parameters to ensure repeatability; Where there are gaps, perform additional method validation studies and accordingly upgrade the methods with additional instructions and precautions.