Anatomy of an FDA Surveillance Facility Inspection - Biocon, Bangalore, India - May 25 to June 3, 2017

Introduction

This review article analyzes the U.S. Food and Drug Administration (FDA) facility inspection, conducted at Biocon Ltd, Bangalore, India from May 25 to June 3, 2017, in the context of the FDA marketing approval of the biosimilar OGIVRI™ (Trastuzumab-dkt). The FDA Form 483 issued at the close of the inspection, initially attracted the attention of the author, because of its emphasis on microbiology and related aseptic processing. The review provides background on the organization of facility inspections, a timeline of the OGIVRI™ marketing application approval, and comments on information obtained from the heavily redacted Form 483 Inspectional Observations and Establishment Inspection Report obtained through the U. S. Federal Freedom of Information Act.

Types of FDA Inspections

It is useful to review current FDA inspectional practices, as they have been re-engineered recently. In general, there are four main types of regulatory inspection of pharmaceutical manufacturing facilities, i.e., surveillance, pre-approval, post-approval and for-cause facility inspections (FDA Concept of Operations Whitepaper, June 6, 2017). It is assumed that the FDA April 7, 2017 inspection is best classified as a pre-approval facility inspection while the June 6, 2017 was selfdescribed as a surveillance facility inspection.

Surveillance Facility Inspections

This type of inspection focuses on facilities that manufacture U.S. marketed prescription and over-the-counter drug products or drug substances monitoring Current Good Manufacturing Practices (CGMP) compliance and not necessarily addressing any specific product. It is a system-based inspection that is used to identify quality problems and adverse trends at facilities and encourage their mitigation. Prior to the globalization of the pharmaceutical industry with the vast majority of drug substances and products imported from Europe, Japan, India and China, the statutory requirement was that these inspections be conducted at least every two years. To rationalize the workload, the FDA has moved to a risk-based model managed by the Office of Surveillance (OS). The highest risk facilities are prioritized annually on past inspection histories, recalls, drug shortages, customer complaints, field alerts, foreign regulator outcomes, submitted quality metric data, and listings of products manufactured at the facility (sterile drug products considered more `critical than non-sterile drug products because of their invasiveness). The recent mutual recognition agreement between the FDA and the European Medicine Authority (EMA) may reduce the frequency of surveillance facility inspections and place more emphasis on foreign inspections in countries like India and China.

The Office of Regulatory Affairs (ORA) leads the surveillance inspections with CDER or CBER participation when considered necessary. If the inspection uncovers critical conditions, which may result in an imminent health hazard, these conditions may be discussed with subject matter experts in the Office of Manufacturing Quality (OMQ) before the close of the inspection.

Objectionable findings are documented on a Form 483 Inspectional Observations and are discussed with the manufacturer before the inspection is closed out. The ORA informs the appropriate division usually within two days of an adverse finding termed Official Action Indicated (OAI). Within 45 days of the close of the inspection, ORA completes the Establishment Inspection Report (EIR) with the classification of the inspection as identified in Field Management Directive 86. It should be remembered that the inspectional observations represent the opinions of the FDA investigators at the time of the inspection and may not be supported on further review.

The three classifications are Official Action Indicated (OAI), Volunteer Action Indicated (VAI) and No Action Indicated (NAI). OAI classifications may result in FDA warning letters, voluntary recalls, import alerts, product seizures or criminal prosecutions.

Pre-Approval Facility Evaluations and Inspections

This type of inspection supports the assessment of marketing applications (regulatory submissions) and focuses on determining that manufacturing facilities named in the sponsor’s application are capable of manufacturing the drug in compliance with cGMP requirements and that data submitted in the application are accurate and complete.

At the time of submission, an Integrated Quality Assessment (IQA) Team led by an Application Technical Lead and facilitated by a Regulatory Business Project Manager, consisting of a drug substance and drug product reviewer, a process/facility assessor, and for pre-approval inspections, an ORA representative. The IQA team is responsible for providing the quality recommendation for the marketing application and identifying the need for a pre-approval facility inspection. The ORA investigator leads the inspection with divisional representation focusing on the areas of concern outlined by the IQA team, documenting any objectionable findings, issues a FDA Form 483, and reports back to the IQA team who recommends on the approvability of the application. During the facility assessment, communication may be sent direct to the facility management to resolve inspection issues, especially if the facility owner is different from the sponsor of the marketing application.

The inspection team provides the Establishment Inspection Report (EIR) to the IQA Team after the ORA Director of Investigations has reviewed it. The IQA Team through a regulatory meeting may address outstanding issues with the sponsor or with an Information Request (IR), Discipline Review (DR) or Complete Response (CR) letter, and provides an overall recommendation on the approvability of the application.

For-Cause Facility Inspections

This type of inspection is initiated in response to a new registrant, specific event or information that brings in question the cGMP compliance and/or quality of a manufacturing practice, facility, process or drug and is meant to gather information to determine the quality of a marketed drug product, and justify enforcement actions including warning letters, voluntary market withdraws, import alerts or product seizure.

ORA, Office of Process and Facilities (OPF), Office of Surveillance (OS), or Office of Compliance (OC) can initiate a request for, forcause inspections. Once these offices determines the inspection is warranted, ORA prepares and gets approval per Field Management Directive 17 for the assignment, it is then scheduled, dependent on risk and other priorities. The inspection team, based on their initial observation, develops an on-site inspection strategy. If major deficiencies are encountered, the inspection team discusses their findings with the impacted FDA offices whether to continue the inspection to gather additional information or close the inspection, and initiate prompt regulatory action.

If objectionable conditions are observed, an FDA Form 483 is issued and the investigators discuss it with the company’s management at the close of the inspection. This is the skin and the bones of the inspection. Within 45 days, the EIR is completed, and electronic documents are available to the initiating office for review. The EIR may be considered the flesh of the inspection. ORA will not issue an FMD-145 letter (Cover letter for a copy of the EIR and a statement that the inspection is considered closed) without concurrence from the initiating office.

In terms of foreign inspections, the breakdown by inspection type is summarized in Table 1. Points of interest are the steady growth in the number of foreign inspections reflecting the globalization of the pharmaceutical industry and the common practice of combining surveillance and pre-approval inspections. The frequency of postapproval inspections was not documented.

Table 1. FDA Foreign Facility Inspections conducted Fiscal Year (FY) 2010 through June 30, 2016 (GAO-17-143)

The recent FDA inspections of the company are summarized in Table 2. Over a nine-year period, Biocon was inspected eight times, at two or three year intervals, with 75% of the inspections classified as Volunteer Action Indicated (VAI) and 25% No Action Indicated (NAI). Despite the size and complexity of their pharmaceutical manufacturing operations, this FDA inspection record can be viewed as barely satisfactory.

Table 2. Biocon Ltd, Bangalore, India FDA Inspection History (October 1, 2008 to December 13, 2017)

During the period from March, 2011 through July, 2017, EMA regulators inspected Biocon Bangalore facilities, at least eleven times with only the July 5, 2017 resulting in a GMP non-compliance notification, which resulted in the temporary withdrawal of the Mylan/Biocon EMA application. At the time of writing this review article, the biosimilar Trastuzumab had not been approved for sale in the European Union (E.U.). This brings up the question of different conclusions reached as a result of FDA and EMA inspections.

During the period from May, 2015 through August, 2016 there have been at least nine examples where the FDA issued warning letters to pharmaceutical manufacturers, while the EMA issued GMP compliance notifications for the same pharmaceutical facilities. This difference in the conclusion reached after an inspection have confounded industry observers. The mutual recognition agreement between the two regulatory agencies became official in November 1, 2017. These differences will need to be resolved. Mutual recognition will avoid the cost of duplicated inspections and enable resources to be re-allocated from facilities in the U.S. and eight European countries initially included in the agreement to the parts of the world with greater cGMP risks and uninspected facilities. The author expects that the two agencies may retain pre-approval and for-cause inspections and share the responsibility for surveillance inspections.

What was at Stake with the FDA Approval of the Mylan/Biocon Biosimilar Trastuzumab?

Although the European patents on the Roche drug Herceptin (Trastuzumab) expired on July 28, 2014 and the U.S. patent is due to expire June 18, 2019, it still is the third biggest drug product with 2016 global sales of U.S. $6.7 billion with no biosimilar competition. With four other E.U. and three other FDA marketing applications under review, there is intense competition to be first to market. Past experience has shown that price erosion with biosimilars is much less than with generic pharmaceuticals with 60 to 80% of the innovator’s price maintained after generic competition. However, a recent Quintiles IMS report commissioned by the E.U. shows that in therapy classes where more than one biosimilar has been launched, the first to market usually takes over 70% of the market share with the second entry taking 30 to 40%, and third entry only 5 to 22% of the market. Analysts from Morgan Stanley believe that Mylan/Biocon could achieve $200 million in sales within two years after the product launch, factoring in 60% erosion in drug price and a market share of 15%. This was translated to a potential $40 million profit share for Biocon by the financial analyst (Sahu, 2017). Despite the multiple adverse inspection reports, Mylan/ Biocon are still expected to be the first to bring a Herceptin biosimilar to market in both the U.S. and European markets. This will be a stunning achievement for Biocon Ltd that less than 40 years ago began as a producer of industrial enzymes for the U.S. and European market.

OGIVRI™ (trastuzumab-dkst) is a HER2/neu receptor antagonist biosimilar to the Roche Herceptin indicated for the treatment of HER2- overexpressing breast cancer and the treatment of HER2-overexpressing stomach cancer. According to the CDC, in 2014 237,000 women were diagnosed and 41,000 died of breast cancer in the U.S. with 20 to 25% potential candidates for Herceptin treatment. The recommended 12-months treatment costs may be up to $70,000.

How important are biosimilars to healthcare in the U.S.? At the time of the application approval on December 1, 2017, the Office of the FDA Commissioner issued the following statement: “The FDA continues to grow the number of biosimilar approvals, helping to promote competition that can lower health care costs. This is especially important when it comes to diseases like cancer, that have a high cost burden for patients,” said FDA Commissioner Scott Gottlieb, M.D. “We’re committed to taking new policy steps to advance our biosimilar pathway and promote more competition for biological drugs.”

Currently because of its later start, the FDA lags behind the EMA in terms of overall number of biosimilar approvals. Since 2006 the EMA has approved 38 biosimilar applications, while since 2015 the FDA has approved nine biosimilars. As indicated in the statement of the FDA Commissioner, it is the FDA policy to accelerate biosimilar approval and they may eventually catch up to the EMA.

Mylan/Biocon OGIVRI™ (Trastuzumab-dkt) Approval Timeline

This timeline is useful to the reader as it provides context around the June 6, 2017 Biocon Surveillance Facility Inspection. Key milestones in the timeline are listed as follows:

January 18, 2014 – CANMAB™ a generic Trastuzumab manufactured by Biocon for treating metastatic breast cancer was launched in India. November 3, 2016 – The FDA received the Biologics License Application (BLA) for the biosimilar OGIVRI™ (Trastuzumab-dkt), 420 mg/vial for injection from the sponsor Mylan.

March 17, 2017 - The French regulatory authorities (ANSM) issued a statement of non-compliance with GMP for the OGIVRI™ (Trastuzumabdkt) Biocon, Bangalore facility. The EU marketing authorization application for three products including Ogivri (Trastuzumab-dkt) would not be approved while this Non-Compliance Statement (NCS) was in force.

April 7, 2017 – And eight item FDA Form 483 (Inspectional Observations) was issued to the CEO and Managing Director of Biocon Limited at the conclusion of the March 27 to April 7, 2017 pre-approval facility inspection. The emphasis of the inspection was on biological drug substance manufacture.

July 11, 2017 – FDA Oncology Drug Advisory Committee unanimously recommended the approval of the Herceptin and Avastin biosimilars applications from the sponsor Mylan.

June 6, 2017 - A 10-item FDA Form 483 (Inspectional Observations) was issued to the CEO and Managing Director of Biocon Limited at the conclusion of the May 25 to June 3, 2017 surveillance facility inspection. It is assumed that Biocon replied within 15 business days outlining their corrective actions to the 483 observations.

July 31, 2017 – The inspection team issued their Establishment Inspection Report (EIR) internally. The opening statement the EIR was: “This CDER surveillance inspection of a foreign sterile drug manufacturer was conducted per FY17 work plans and FACTS assignment #55866. The firm is a manufacturer of finished sterile parenteral drug products. Coverage was given to compliance program 7356.002A, Sterile Drug Process Inspections. The current inspection utilized the systems approach and covered Quality, Facilities & Equipment, Materials, Production Packaging & Labeling and Laboratory Control systems”. The recommended classification in the EIR obtained via the U.S. Freedom of Information Act was Official Action Indicated (OVI) with CDER review and concurrence.

August 16, 2017 – Biocon announced, as a procedural requirement linked to the re-inspection of their Bangalore facility, their withdrawal of their E.U. marketing applications for the biosimilars trastuzumab and pegfilgrastim.

September 26, 2017 – The Health Canada Inspection Tracker: Drug Manufacturing Establishments listed Biocon, Bangalore facility as reported by a regulatory partner(s) was continuing to review evidence submitted related to corrective actions on GMP observations including data integrity. This item was subsequently closed.

November 20, 2017 – Biocon reported in a brief letter to the National Stock Exchange of India that the FDA issued a Complete Response (CR) letter and the EIR in relation to the cGMP inspection of their aseptic drug product facility May 25-June 3, 2017. The FDA classified the outcome of the inspection as VAI (Voluntary Action Indicated), not as recommended by the FDA investigators in the EIR, and that the inspection was closed.

November 30, 2017 – Mylan and Biocon jointly announced the acceptance of the previously withdrawn EMA Marketing Authorization Application for the biosimilars trastuzumab and pegfilgrastim, as Biocon completed the Corrective Actions, Preventative Actions (CAPAs) resulting from the EMA audit observations. The CAPAs will be confirmed during a re-inspection as part of the regulatory review.

December 1, 2017 – Mylan received FDA marketing approval for the multi-dose OGIVRI™ (Trastuzumab-dkt) 420 mg/vial for injection that will be manufactured, filled, labeled and packaged at Biocon in Bangalore, Karnatake, India. On the same date, the EMA reaccepted Mylan’s marketing authority applications for the biosimilars trastuzumab and pegfilgrastim suggesting that the cGMP compliance issues had been successfully resolved.

January 18, 2018 – Sandoz, a Division of the Swiss company Novartis, and a leader in biosimilars with five marketed worldwide, announced an exclusive global collaboration with Biocon on the development, manufacture and distribution of next-generation biosimilars. The announcement highlighted that Biocon was the first Indian company to have a biosimilar approved by the FDA.

Details of the Biocon Adverse Regulatory Inspections

The eight observations in the FDA Form 483 issued on April 7, 2017 and obtained by the author can be summarizes as follows:

  • Discrepancies were found between the information submitted in the BLA and manufacturing process as performed at Biocon.
  • The Quality Assurance Unit has approved and implemented documents that do not assure appropriate production, testing, and deviation review and release of commercial material.
  • The handling of in-process samples is inadequate.
  • A lack of Quality oversight in the review of quality control testing procedures of the drug substance and product.
  • Out-of-Specification (OOS) procedures do not ensure conformance to batch release specifications, appropriate investigations and tracking.
  • Material used for drug substance formulation has inadequate microbial control.
  • Bioburden sampling during the drug product manufacturing is inadequate.
  • Manufacturing deviation are not open and closed in a timely manner as required by Standard Operating Procedures (SOP).

Davio (2017A) reported that a non-compliance statement (NCS) issued on July 5, 2017 by the French Regulatory Authority (ANSM) identified a number of concerns with the drug product manufacturing activities for Biocon’s pegfilgrastim (Fulphila), trastuzumab (Ogivri), and insulin glargine (Semglee). The French inspection, a preapproval audit related to pending European Medicines Agency (EMA) marketing authorization applications for the three biosimilars, was conducted from March 13 to March 17, 2017. During the course of that inspection, the ANSM found 35 problems related to drug product manufacturing and quality control operations for the three biosimilars and recommended that the product not be approved while the NCS is in force (Report 17MB002NCR, July 5, 2017). Of these problems, the French agency considered the following eleven to be classified as major deficiencies:

  • Environmental monitoring
  • Training
  • Out-of-specification results management
  • Cleaning validation
  • Process validation
  • Vendor qualification
  • Media fill test
  • Cross-contamination risks
  • Batch manufacturing record
  • Differential pressure alarms’ management in classified areas
  • Access management in systems, applications, and products (SAP) for batch certification

Unlike the FDA Form 483, which when made public, provides details of the deficiencies observed by the FDA investigators, the NCS merely lists the areas of concern and it impact on any outstanding marketing authorization applications. The French authorities requested that a follow-up GMP inspection to verify that a Corrective Action, Preventative Action (CAPA) plan has been implemented prior to approving the three biosimilars.

Davio (2017B) reported that FDA inspectors listed ten observations made during the agency’s inspection conducted during May 25 to June 6, 2017:

  • Inadequate investigation of unexplained discrepancies among batches.

The report notes that Biocon did not properly investigate the presence of visible particulate material that had been observed in an injectable drug since August 2015. The drug in question continues to be sold in Uruguay, Russia, and the Dominican Republic.

  • Inadequate validation of sterilization processes.

Procedures designed to prevent microbiological contamination of drugs do not include validation of sterilization processes. Aseptic personnel are not tracked, there are no requirements for qualifying personnel during media fills, the process used to sterilize gloves has not been validated, filling times are not tracked during microbiological growth media fills, and the procedure used to validate the Heating Ventilation and Air Conditioning (HVAC) system does not take into account microbial levels.

  • Deficient monitoring of environmental conditions.

Aseptic operators are held to lower qualifications than necessary, there are no specified locations for environmental monitoring, and a protocol for microbiological monitoring of controlled environments is not followed.

  • No establishment of procedures designed to prevent microbiological contamination of sterile drug products.

Validation of the effectiveness of disinfectants is not performed, there are no disposal records for expired cleanroom garments, and staff were observed picking up cleaning cloths that had been dropped on the floor.

  • Incomplete laboratory records.

An FDA microbiologist reported different readings on equipment from those reported by the facility’s microbiologist, and not all internal records can be reconciled. Data from the leak testing of sterile gloves is not documented.

  • Deficient procedures for cleaning and maintaining equipment.

The firm lacks a scientific rationale for its disinfecting practices, numerous surfaces are not routinely disinfected, and inquiries into microorganisms found during previous investigations were delayed by months.

  • Lack of authority in the quality control unit.

The facility’s quality control department lacks the authority or responsibility to approve or reject drug products or components. The inspectors noted that vendor qualification procedures are not enforced for sterile gloves and that expired materials are in unrestricted use.

  • Lack of training for employees who manufacture and process drugs.

The FDA inspectors found that colony growth on the facility’s environmental and personnel monitoring plates are often enumerated incorrectly, and that personnel with administrator rights to critical information technology systems do not have documented training.

  • Lack of scientifically sound laboratory controls.

The firm lacks procedures adequate to assure the identity, strength, quality, and purity of drug products. The firm’s bacterial endotoxin testing for finished products is deficient and could lead to false negative results. The pH testing protocol of samples is also deficient.

  • Labeling procedures are not followed.

Rejected labels can be found in an unsecured waste bin, rather than under lock and key.

Detailed Analysis of the FDA Form 483 Observations Received on June 3, 2017 by Biocon Limited, Bangalore, India During a Surveillance/PreApproval Inspection

Two FDA investigators, a Consumer Safety Officer (CSO) who works out of the FDA Wauwatosa, WI office and a Microbiologist (Investigational Analyst) who works out of the FDA Alameda, CA office conducted the inspection. The CSO has industry experience as a quality manager. The emphasis of the inspection was on aseptic filling and the supporting microbiology programs, which makes it interesting to pharmaceutical microbiologists. As we have come to expect, the FDA Form 483 observations were often relatively poorly organized and so heavily redacted to be difficult to read and comprehend. Redaction is considered necessary to eliminate confidential information, but it significantly reduces the usefulness of the FDA Form 483 as a learning tool to outside interests. Division as to areas of non-compliance with the GMP regulations, which should be cited in each section as in a warning letter, would be helpful. In terms of organization, the EMA policy of classifying the observations into critical, major and minor and listing them accordingly is preferred. It has been suggested that the 483 observations are listed in order of importance, but we cannot be certain that is so in this case or any other case. Biocon’s reply to the 483 observations was not available to the author.

The 64-page EIR provided additional details about the inspection including the division of labor between the two FDA investigators, an inventory of supporting documents and photographs collected, and was structured with each of the Form 483 observations followed by a Supporting Evidence and Relevance/Discussion with Management section. The inspection was conducted over seven days.

Observation 1 - Investigations

A. The extension of a failure investigation or a risk assessment to other batches within a specified time frame is a consistently cited issue during FDA inspections. What should that time frame be? How conservative should a company be in terms of extending their investigations, so as not to be secondguessed by future FDA investigators? The EIR documents the investigation of visible particulates in long term and accelerated stability batches of a single product that went back to August 2015. Based on external particle analysis and comparisons to the innovator’s product, the company concluded that the particles were intrinsic to the product and not a safety concern. The investigators determined that the investigation was protracted, poorly documented and inadequate. The visual inspection reports prior to product release and customer complaints did not discover these particles outside of the stability samples. Intrinsic particulates are difficult to eliminate as they are related to both the properties of the drug substance and the drug formulation and might only appear late in the product shelf life.

The FDA has taken an aggressive stance on foreign matter in injectable products, going as far as encouraging companies to recall batches with isolated customer complaints. In the view of the author, this stance has led to drug shortages. Visual inspection is a probabilistic activity; so it is not possible to exclude every defect from a large-sized batch of product. Furthermore, the author believes that visible foreign matter particles occasionally found in individual vials, when injectable products have limits for subvisible particulates that may penetrate the vascular system, i.e. USP <787>, is not a safety issue. However, they as extrinsic matter can be mitigated by attention to the procurement, cleaning, and sterilization of primary packaging components.

B. Reviewing how an SOP update would affect marketed batches implicated in an investigation appears to be a new requirement. This issue would be addressed during an investigation in terms of the change control documentation, which drives the SOP revisions, not the other way around. However, SOPs must be current with compendial and regulatory changes and subject to periodic review and revision.

Observation 2 - Sterilization Process Validation

A. Participation in a media fill that is conducted periodically, i.e., as infrequently as twice a year depending on the number of products, aseptic filling lines and production shifts, should not be an absolute requirement for working in an aseptic filling area as is gowning qualification. However, the lack of participation due to scheduling contraints may define the range of activities of an operator are qualified to conduct in each area and their inclusion in a future media fill must be ensured by the company. The challenges in staffing an aseptic processing area may require some flexibility. The EIR showed that the media fill participation of the operators was assumed and their involvement in media fills including critical interventions poorly documented. The company committed to define the requirement of the operators.

B. The FDA investigators are correct. The interventions conducted by an operator during a media fill should be defined in the media fill protocol and documented in the batch record. They may be routine interventions like refilling a stopper bowl or non-routine like responding to a power outage.

C. The media fill batch record should capture any stop-start times not just the total duration of the media fill. Media fills must be fully documented.

D. As the integrity of sterile gloves is critical in preventing microbial contamination, the glove integrity testing should be validated to its effectiveness.

E. Investigations into environmental monitoring (EM) excursions should trigger a review of environmental controls, batch record review, result in the implementation of corrective actions, and in limited cases revalidation of the aseptic processing.

Observation 3 - Environmental Monitoring

A,B. Operators that work in ISO 5 areas must meet the personnel monitoring (PM) requirements for both gloves and cleanroom uniforms for an ISO 5 area. This would include operators who mainly work in the surrounding ISO 7 area but conduct any interventions in the ISO 5 aseptic processing areas such as EM and removing rejected vials. The observations in the FDA Form 483 and the EIR used the E.U. GMP cleanliness classifications, i.e., Grade A, and not ISO 5. FDA staff should be encouraged to use global International Organization for Standardization (ISO) standards.

C. EM locations should be risk based and defined in the environmental monitoring procedures. A map of the floor plan showing the sampling location is helpful.

D. Swab sampling conducted in the aseptic filling areas must comply with the procedure found in the SOP. Compliance with current SOPs is a fundamental cGMP requirement.

E. The location of the particle counter when monitoring the filling operation should be based on the risk of product exposure. Sampling probes during dynamic monitoring should not draw from “first air” upstream from critical aseptic operations. Smoke studies may be helpful in determining the location of the particle counter, however, their capacity to distort laminar airflow during routine monitoring should be considered. The EIR documented a discrepancy with the 2004 FDA Aseptic Processing Guidance recommendation of using a location within one foot of the filling station. Companies should be able to establish sampling locations based on practical considerations.

Observation 4 - Aseptic Process Procedures

A. The rationale for an in-process bioburden program should be available for regulatory review. This would include sampling after each critical processing step and the bioburden limits adopted.

B. Conducting disinfectant validations inside a cleanroom, as recommended in the EIR makes no sense, if they involve taking live organisms into critical areas in the facility. Sampling surfaces prior to and after sanitization is unlikely to demonstrate adequate log reductions expected for validation purposes because of the low counts on these surfaces but will demonstrate the overall sanitization effectiveness. The ability of a disinfectant to eliminate representative microorganisms, including spore-former, from different surfaces found in the facility can be determined in laboratory studies. Observation 4, B is a very questionable observation.

C. Cleanroom garments must be managed to control microbial contamination originating from operators. Multiple uses of the same garments in an aseptic processing area should be discouraged. If they are clean, sterilized and reused when they should be inspected for damage prior to cleaning and, if practical, repaired. The use of expiration dating is probably unmanageable with cleanroom garments, but first in, first out inventory control is recommended. The use of singleuse, disposable garments should be encouraged.

D. Incubation schema validation and growth promotion testing have never required the use of mixed cultures, e.g., USP<61>, so the FDA investigators are mistaken on this issue. Many microbiologists believe that high- low temperature incubation makes sense because bacteria derived from people are the most common EM isolates. However, there is always the possibility that fungal colonies at lower incubation temperatures and bacteria at higher temperatures will crowd out other colonies on a plate reducing the count and diversity of isolation. The literature is inconclusive as to the incubation sequence (Cundell, 2016). Recommendation to use mixed cultures is not a requirement and should not have been included in the Form 483 observations.

E. Major interventions should always be captured in the batch record.

F. It is difficult to comment without the knowledge of the SOP.

G. Cleanroom behavior is important in preventing microbial contamination, as people are the major source of microorganisms.

Observation 5 – Laboratory Records

A. Citing a company for a deficiency that occurred seven years ago and was subsequently corrected is completely misguided. cGMPs requirements evolve and companies improve their procedures. A company should be judged by current observations made during an inspection and not a long passed cGMP violation. This observation should not have been included in the FDA Form 483.

B. The observation that the company was reporting no ColonyForming Units (CFU) and the FDA investigator was finding 1 CFU is disconcerting and may be or not a data integrity issue. The delay in examining the plates may have been a factor in the appearance of colonies - the plates were read 5/24 and the inspection conducted between 5/25 and 6/1/2017 or there was falsification of the EM results. The author believes it was the former. The FDA microbiologists should be aware that only the CFUs reported within the recommended incubation time are considered valid, as colonies will continue to appear and grow in size on the plates with increasing incubation times. Furthermore plates with numerous colonies had slightly higher counts when reviewed by the FDA investigators beyond the specified incubation time reinforcing this pattern. This observation should not have been included in the FDA Form 483.

C. The EM and PM plates distributed and collected for incubation must be reconciled. Damaged and mishandled plates may be declared invalid, but the number of these plates should be small and their occurrence infrequent.

D. Audit trails should be comprehensible to an external auditor.

E. escribed in the FDA Form 483 seems reasonable.

F,G. The glove inspection and testing procedures need to be tightened up. Raw data generated must be documented and results scored as a pass or fail.

H. Batch records used for routine aseptic filling operations may not be adequate for media fills. Informational sheets may contain statements that a requirement is met, i.e., passes, prior to the final review and approval of the media fill validation. This is a questionable observation.

Observation 6 – Cleaning and Disinfection Practices

A. The pharmaceutical industry understood that the FDA was going to back off on disinfectant effectiveness validations! Apparently we were wrong. The inclusion of an environmental isolate like Cladosporium oxysporum found in both a media fill and routine monitoring in disinfectant program revalidation is appropriate, especially if the isolation was part of an adverse trend, but the author believes that this mold would not be expected to be more resistant than other representative molds included in the original disinfectant validation. The log reduction would depend on the choice of disinfectant, application method, contact time, and the nature of the surface. Revalidation of cleaning and disinfection programs are typically conducted annually based on the annual review of the environmental monitoring from the previous year and then scheduled, so this time lag is not unexpected.

B. Tracking the use of multiple spray bottles is a difficult task. Instead of documenting the number of times a spray bottle is used, assigning an expiration date may work better. Single-use sterile spray bottles would be a suitable, but a more expensive solution. The documentation of the cleaning of all items used in a cleanroom setting may be excessive.

C. Clearly SOPs must be followed. The frequency of changing a disinfectant wipe and the manner of wiping would be important in terms of the disinfectant effectiveness. Dropping wipes on the floor and retrieving them as reported in the EIR is unacceptable and may suggest insufficient wipes are available to the operators.

D. Representative surfaces in a cleanroom, especially those adjacent to critical aseptic processing activities, should be subject to cleaning and disinfectant validation in laboratory studies. The extension of the validation to ceiling panels, light features and wall panels outside the critical aseptic filling area is excessive and is not industry practice. This is a questionable observation.

Observation 7 – Responsibility of the Quality Control Unit

A. Sterile gloves are a critical processing item and the vendor should be subject to qualification that may include a site audit.

B. The outer packaging is important with respect to items entering a cleanroom and can be inspected for integrity. Double or even triple wrapping is common.

C. AQL-based sampling plans do accommodate the incoming batch size. However, statistical sampling of incoming sterile gloves seems excessive. The EIR documented incoming lots of sterile gloves are subjected to sterility testing, inspection for pinholes and general description and appearance inspection for appearance. The author agrees that more attention should be given to the integrity of outer packaging.

D. Reagents requiring expiration dating must not be used passed their dating.

Observation 8 – Employee Training

A. This brings into question the integrity of EM, PM and bioburden data and the appropriate training to read a plate. The additional CFUs could be due to the time lag in the investigator reading the plate, i.e., 1 or 2 days later. The count is defined by the incubation period not what you observe day(s) later. Personnel with an educational background in microbiology and training on a related SOP should be able to read a plate. However, small, translucent, or colorless colonies may be occasionally overlooked and the literature suggests that multiple plate readings by different microbiologists will vary in counts up to 10%. The EIR established that the two QC microbiologists had missed their 2016 annual cGMP training and the laboratory lacked a SOP that specifically covered plate reading. That deficiency must be corrected.

B. IT personnel with system administrator responsibilities must have GMP training.

Observation 9 – Specifications

A. The Bacterial Endotoxin Test sample dilution must be adjusted in accordance with the calculated MVD.

B. USP<85> states, “If necessary, adjust the pH of the solution (or dilution thereof) to be examined so that the mixture of the LAL reagent and sample falls within the pH range specified by the LAL reagent manufacturer, i.e. 6.0 to 8.0.” This means if the test solution is in range, you do not need to test the reagent-product mixture. Although technically correct, this is a questionable observation.

Conclusions

The question that must be asked is why was the recommendation of the FDA investigators for an Official Action Indicated (OAI) classification in the EIR downgraded by CDER to Volunteer Action Indicated (VAI). Without the knowledge of the internal discussions, the author must believe that the CDER and the ORA management did not accept the observation of a clear case of lack of EM data integrity in plate counting, and they either discounted other questionable observations or accepted the proposed CAPAs to correct valid observations in Biocon’s response to the Form 483 observations. Also, CDER must have been satisfied by the response to observations during the April 7, 2017 facility inspection that the discrepancies found between the information submitted in the BLA and manufacturing process as performed at Biocon were adequately explained. The re-opening of the EMA application authorization submission in December 1, 2017 re-enforces the view that Biocon have improved their cGMP compliance and implemented appropriate CAPAs. The EIR documented the openness of the management and staff to the inspection and their willingness to take corrective actions, further justifying a VAI classification. However, CDER must have felt pressure to approve the application in light of unanimous recommendation for approval by the FDA advisory panel and the smaller number of biosimilars approved by the FDA compared to EMA. Overall, though I disagreed with some of the FDA Form 483 observations, my analysis of the record shows a high level of professionalism and judgment exercised by the FDA, which is staffed by dedicated and hard-working public servants. Based on the limited information available, the author agrees with the FDA decision to approve the biosimilar.

References

  1. Biocon Ltd Investor Presentation - Ahead of the Curve November 2017 https://www.biocon.com/docs/Biocon_IR%20PPT_Nov%202017.pdf
  2. Cundell, T. 2016 Mold Monitoring and Control in Pharmaceutical Manufacturing Areas. Amer. Pharm. Rev. 19(5):10-19
  3. Davio, K. 2017A French inspectors find major deficiencies at Biocon’s manufacturing facilities. July 10, 2017 www.centerforbiosimilars.com/news/french-inspectors-findmajor-deficiencies-at-biocons-manufacturing-facility
  4. Davio, K. 2017B FDA Inspection Notes 10 Problems With Biocon’s Manufacturing Facility August 8, 2017 www.centerforbiosimilars.com/news/fda-inspection-notes-10-problemswith-biocons-manufacturing-facility
  5. FDA Whitepaper Integration of FDA Facility Evaluation and Inspection Program for Human Drugs: A Concept of Operations June 6, 2017
  6. Mylan Announcement – U.S. FDA approves Mylan and Biocon’s Ogivr, the first biosimilar for Trastuzumab for treatment of HER2-positive breast and gastric cancers
  7. www.mylan.com/newsroom2017-12-01
  8. Quintitles/MS Report IMS Biosimilar report - The Impact of Biosimilar Competition in Europe May, 2017
  9. Sahu, Ram Prasad 2017. For Biocon, windfall seen from drug launch. Business Standard, Bangaluru Edition, March 17, 2017
  10. U. S. Government Accountability Office Report to the Committee on Energy and Commerce GA)-17-143 FDA’s Foreign Offices and Drug Inspections December 2016
  11. U.S. Cancer Statistics Working Group. United States Cancer Statistics: 1999–2014 Incidence and Mortality Web-based Report. Atlanta (GA): Department of Health and Human Services, Centers for Disease Control and Prevention, and National Cancer Institute; 2017.
  12. USP 40/NF35 <61> Microbiological Examination of Non-sterile Products: Microbial Enumeration Tests
  13. USP40/NF35 <787>Sub-visible Particulate Matter in Therapeutic Protein Injections
  • <<
  • >>

Join the Discussion