OR WAIT 15 SECS
James P. Stumpff, RPh, is principal consultant with Parexel International, firstname.lastname@example.org.
Investigating the root causes of OOS conditions, product defects, and batch failures requires a systematic, thorough approach.
Investigations, deviations, complaints, and laboratory investigations are critical to assessing the compliance of a company’s quality systems. Deficiencies in the timeliness and thoroughness of these investigations can jeopardize a company’s compliance status with potentially serious consequences. This article describes a challenging biopharmaceutical investigation into a quality control (QC) test deviation. In this case, thorough evaluation involving a multidisciplinary team determined root cause, suggesting strategies that pharmaceutical companies can use to probe beneath the surface to improve laboratory investigations and overall quality systems.
Whenever product defects, batch failures, process deviations, or out-of-specification (OOS) situations come up, regulatory agencies require that pharmaceutical manufacturers determine the root cause of those problems, using quality risk-management and other methods (1,2). Corrective and preventive actions must be taken in response to investigations, and their effectiveness monitored and assessed. As shown in Table I, FDA issued 264 Form 483 citations to pharmaceutical manufacturers in 2019 for inadequate responses to quality failures (3). Regulatory citations offer insights into failures that many pharmaceutical manufacturers make when it comes to investigations (Sidebar). Consider an FDA warning letter (4) issued in 2013. FDA inspectors cited the company for inadequate investigation of critical deviations and batch failure. The company had manufactured a number of lots of an API for more than one year, but six of those lots were severely contaminated with the bacterium, Bacillus thuringiensis, and a number of other microbes including Acinetobacter radioresistens.
Although contamination levels had exceeded the company’s action limit, FDA inspectors found, the manufacturer did not thoroughly investigate the root causes of the contamination, and did not assess the potential impact on final product, including the ability of its manufacturing process to clear any non-host cell contamination.
Instead, regulators found, the company had responded to the problem by increasing its use of a sporicidal agent in its clean rooms but did not even attempt to verify the effectiveness of the sporicidal treatment or cleaning procedures before resuming API production.
Subsequently, the company performed additional tests, which it used to justify releasing the API for additional processing. However, it failed to assure the absence of other contaminants. In short, the company had failed to identify underlying root causes of these problems, and, thus could not correct them or demonstrate that it could prevent them from recurring in the future.
Other regulatory citations show the importance of clear and complete standard operating procedures (SOPs) that spell out the steps that must be followed during an OOS event or other product or process failure. In 2019, for example, European Medicines Agency (EMA) regulators cited a manufacturer for inadequate corrective action and preventive action (CAPA) and SOPs (5). Although the company’s CAPA plan met regulators’ requirements in some areas, it did not address the handling of OOS events, and failed to respond adequately to two important problems that regulatory inspectors had found at the facility.
Regulators then determined that the company’s SOPs for investigating OOS conditions were not sufficiently defined. SOP descriptions were not clear and complete, they found, so that they could not ensure that OOS situations would be adequately handled in the future.
Regulations and guidance documents do not specify a set timeframe for completing OOS, root cause, and other investigations, although many companies use the 30 business days specified in the Barr Laboratories legal decision of 1994 (6) as a guideline. Currently, it normally takes companies anywhere from 30 to 45 days to complete an OOS or other investigation in the pharmaceutical industry.
When an OOS occurs, it must be investigated in the laboratory. Requirements for conducting laboratory investigations have been set by the United Kingdon’s Medicines and Health Care Products Regulatory Agency (MHRA) and FDA. Most companies aim to complete the initial Phase I lab investigation within 10 days.
Phase I of the investigation is commonly conducted by QC laboratory personnel and management, and commonly includes the completion of a checklist and an interview with the analyst who performed the work. Generally, this interview should occur as soon as possible to ensure that he or she recalls clearly what occurred during the analysis. Phase II of the investigation typically involves additional personnel (e.g., from QC, quality assurance [QA], and manufacturing departments) and expanded investigations. During this stage, sampling procedures, sample management and the manufacturing batch record will be reviewed.
The following case study, which describes an actual OOS investigation, demonstrates the right way to approach the challenge of determining root cause and taking appropriate corrective action.
In this case, an established biologics manufacturer that produces both liquid and lyophilized products, had a deviation occur after using an established bioburden test, membrane filtration, for final fill load sampling. The manufacturer had used this method without incident for years. According to procedure, the specification for the final fill load sample had been set at ≤1 colony forming units (CFU)/10 mL. However, 3 CFU/10 mL was observed in the trypticase soy agar media incubated in an anaerobic condition. The microorganism was identified as Propionibacterium sp.
A laboratory investigation report was initiated, and the investigational tests and retest were performed using remaining sample and retained sample. There was no microbial growth in either sample. To confirm that results of the retest were credible (the retest had been performed more than 195 hours after sampling), a hold-time study for 195 hours was performed. However, this study failed to recover results, and was declared invalid. Because there were no confirmed laboratory errors and the investigational test result was not valid due to hold-time study failure, a deviation was initiated.
In this type of situation, the entry of any microorganisms into the sampling process is most likely to occur, either during the procedure (when the sampling bag and ultrafiltration/diafiltration [UF/DF] pool bag are assembled) or as a result of inadequate sampling procedures.
Review of relevant SOPs confirmed that sampling procedures had been properly established. Also, considering that the microorganism involved was a human- derived anaerobic bacteria and that the closed manufacturing process strictly prevented the process from coming in contact with the outside environment, it was difficult to say that the microorganism had been introduced during production.
Fishbone root-cause analysis ruled out personnel, equipment, the environment, material or measurement as potential root causes of the deviation. The four operators who had performed the sampling were interviewed, and it was determined that the sampling had been done properly and that contamination had not been introduced during bag assembly. An interview with the QC analyst then confirmed that the bioburden test had been performed correctly, according to the procedure. To complete investigation into root cause, the SOP for the bioburden tests was reviewed, including gowning and aseptic procedures in the biosafety cabinet. This is where the problem was found. As it turned out, the procedures did not require face covering, which could allow contamination to take place.
The exposure of the operator’s face likely occurred during the test when the operator adjusted safety googles. Thus, the most probable root cause for the bioburden test OOS result was found to be a deficient gowning procedure. Phase II investigation tests were run as follows:
Re-performing the hold-time study to confirm the hold time for the retest result. The hold time for 195 h could not be confirmed, so the retest could not be validated.
Simulating the sample-hold condition study to reproduce the stored conditions before the sample was tested. In this case, recovery of the microorganism was maintained. The microbial hold-time study (168 h in 2–8 °C conditions) did not allow the recovery of contaminated microorganisms, but in a short period of storage, Propionibacterium sp. could be found through the bioburden tests.
Performing a QC operator monitoring study to test three QC analysts, including the first tester to determine whether microorganisms could have been introduced during the test.
The identical microorganism (Propionibacterium sp.) was found from the same analyst. Only Staphyloccus sp. were found from the other analysts, so contamination could have been due to the microorganism from the QC operator. The procedure was revised to include the wearing of a facial mask during the bioburden test. This was a first-time event despite many years of bioburden testing with the established procedure and personnel. QC analysts and supervisors had not recognized the deficient gowning procedure.
The usual checklist and analyst interview failed to identify the root cause. QA does not typically conduct routine oversight of laboratory operations as they do for aseptic manufacturing operations. QA internal audits had not identified this deficiency either. Phase II of the investigation involved more, and a more diverse group of, specialists, who brought their knowledge to the investigation. They reviewed the manufacturing process and event, which led to the hypothesis that the most likely root cause was sample contamination.
Lab procedures were reviewed again with the additional investigational staff involved, and the root cause identified. Finally, a PowerPoint presentation was prepared to make this investigation easy to explain to the regulatory authorities.
A number of lessons were learned from this event. For one thing, supervisors and QA should spend time watching critical operations in the lab just as they do in production.
It was also determined that multi-disciplinary teams should be used for investigations and internal audits and QA staff included in Phase I investigation for OOS investigations. The use of isolators should also be considered, to further reduce the likelihood of sample contamination occurring in the future.
1. EU, EudraLex Volume 4, Chapter 1, Pharmaceutical Quality Systems (Jan. 31, 2013).
2. US CFR Title 21, Part 211.192 (Government Printing Office, Washington, DC) 162.
3. FDA, “Inspection Observations,” www.fda.gov.
4. FDA, Warning Letter, CMS #352798, March 22, 2013.
5. EMA, Inspections Database, eudragmdp.ema.europa.eu, accessed June 10, 2020.
6. United States vs. Barr Laboratories, 812 F. Supp. 458 (D.N.J. 1993).
James P. Stumpff, RPh, is principal consultant with Parexel International, email@example.com.
Vol. 44, No. 7
When referring to this article, please cite it as J. Stumpff, "Succeeding With OOS and Root-Cause Investigations," PharmTech 44 (7) 2020.