USFDA 483 / IPCA / June 2023 /Inadequate OOS investigations, test methods, hold time studies

Observation 1:

Failure to thoroughly review unexplained discrepancies. Out of specification results on multiple batches of tablets were invalidated based on inconclusive hypothesis

• OOS of a batch of tablets for dissolution test at 9 month long term stability condition, no root cause could be established in Phase 1 & Phase II investigations. Hypothesis testing was inconclusive with both higher and lower results, it is not clear why a lower value was obtained though analyst followed the defined test procedure. Based on inconclusive hypothesis testing, the sample was retested and OOS invalidated.
• One of the three batches of tablets analysed in same sequence in HPLC was OOS for assay at the 3^rd month long term stability station. No root cause was established in Phase 1A & Phase IIA investigations. Hypothesis testing was performed for possible sample preparation error, hypothesising the analyst could have used higher pressure causing higher assay in one batch, though all the three sample preparations were by same trained analyst. Based on this inconclusive hypothesis, sample retested and OOS was invalidated.
• One of the three batches of tablets analysed in same sequence in HPLC was OOS for assay. No root cause was established in Phase 1A & Phase IIA investigations. Hypothesis testing suggesting OOS could be due to error in sample preparation stage and/or vial filling, and based on the inconclusive hypothesis, sample retested and OOS invalidated. No manufacturing investigation was performed
• In a Process validation exercise, for initial sample set (1x-3x) OOS was reported in two units. No root cause was established in Phase 1A & Phase IIA investigations. In Phase IIB increased sample size was used (3x-5x) and again two units were OOS. Further an Addendum was created to the protocol and a higher size sample (5x-7x) was taken from Inprocess Bulk Containers (IBC) – which is different from the equipment where initial samples were collected and tested. The samples were collected beyond the hold time. Based on this test, the initial OOS was invalidated and batch proceeded for compression.

Inadequate OOS Investigations and inconclusive hypothesis

OOS with inconclusive root cause continues to be one of the most cited deficiencies in USFDA 483s. While establishing a root cause is by no means an easy task, the investigation and documentation should be thorough to show all possible causes are evaluated. In OOS investigations where a root cause is not conclusively established, a thorough manufacturing investigation is a must, which should unequivocally establish there is no manufacturing / material related causes for the OOS.

When there are multiple samples analysed in same sample set / analysis sequence and one or few of them give out of specification result, assigning analyst error for those failing results is not convincing. Typically the investigations and checklists document analyst is a “ trained analyst”, test procedure is “well defined” and “validated”. But are they? Are analysts aware of all issues that can cause an error; are the test methods detailed with the precautions to be taken while performing the analysis. Root causes like, higher pressure during sample filtration, under filling of sample vial, insufficient extraction due to sonication / shaking, handling of sample solution, handling analysis accessories like cuvettes, cleaning and rinsing of cuvettes in UV analysis are all indicative of poor analytical skills and awareness, inadequate method validation, methods are not adequately detailed with respect to steps and precautions. Procedure and process followed for training and qualification of analysts should be robust. Conduct regular workshops for analysts and laboratory personnel, take up different methods for discussion, debate what are possible areas of errors with contribution from all analysts, take up case studies with past incidents. Pay particular attention to analyst qualification. The rigour required to be followed for qualifying an analyst should not be compromised while rushing the onboarding process for analysts. Rather than only matching the results for blind samples, for critical analysis requiring higher degree of skill, observe the key steps of analysis, correct errors and wrong practices. In method validations, assess all criticalities of the method in robustness studies, example (but not limited to) – sample sonication time, stirring time, temperature, time for shaking. Detail the precautions to be taken in the test procedure.

Blend uniformity samples are taken (normally 1x – 3x, meaning 3 samples each from 10 locations typically from the blender) to verify the uniformity of content in the blend with a Stage 1 (10 samples) and Stage 2 (all 30 samples) criteria for variability (%RSD). For test of Blend Uniformity all the analysis results of the sample shall be within the defined specification range (say 95-105%), while the variability is assessed by %relative standard deviation (RSD). After analysis of initial 10 samples (Stage 1) if the defined %RSD is not achieved, the next 20 samples are analysed to asses %RSD on a larger sample set with a higher limit. But all the samples’ results are to be within the defined specification. Of course an error / mistake during analysis can happen; but then the focus of investigation shall be to identify the cause for the failure. The Stage 2 samples (from the 1x-3x sample) are not a substitute for the initial 10 samples. When there are out of specification results from 1^st set of 10 samples, after establishing the root cause, at best another 10 samples may be analysed to establish that the variability is within limits. If all samples comply and %RSD within Stage 1 limit make a conclusion on the uniformity of the blend. If all samples are within the limit but the %RSD is higher than the Stage 1 limit, test the remaining samples to assess Blend uniformity and based on results make a conclusion, document the rationale. But an approach of expanding the sample size to (3x-5x), (5x-7x) is akin to testing into compliance.

Inadequate hold time studies

Hold time studies are performed for establishing the time limits for holding the materials at different stages of production, to ensure that the quality of the product is not impacted. Holding or storage of the different stages should not have deleterious effects on subsequent processing, stability, safety, efficacy or quality of intermediate products and bulk products prior to final packing. Hold time studies should be conducted for all stages at which a material / product may be stored before further processing; for example in tablet manufacturing, stages like binder solution, wet granules, bulk granules ready for compression, core tablets / bulk tablets ready for packing or coating etc. The General guidance on hold-time studies – Annexure 4, by WHO gives a good overview of considerations for hold time studies for medicinal products.

The parameters tested for establishing hold time should be indicative of possible quality impact due to holding. In bulk granules stage this include -Description, Assay, Water content, Loss on drying, Particle profile, Bulk Density, Uniformity of the blend (impact due to segregation), Impurities, Microbial characteristics. Parameters shall be selected based on a risk assessment and rationale documented. It is best to hold the material in the containers in which the material is stored (intermediate storage drums, IBCs etc) and sample collected at defined intervals from the container for hold time studies. This ensure that the studies cover possible impact due to storage such as moisture ingress, segregation, settling, changes in particle profile, container head space and impact of atmospheric oxidation and so on. Also the drug product processed after material hold time should be loaded for long term stability studies. Skipping some of the parameters on hold time samples and extrapolating these analysis data from the bulk sample collected at the end of the stage is not be justified, as it fails to reveal the possible impact of the storage.

Inadequate system suitability tests in UV method

Specific system suitability requirements and criteria are not defined in Pharmacopeia or regulatory guidance for assay and other methods by UV Spectroscopy. The USP Chapter <851 SPECTROPHOTOMETRY AND LIGHT-SCATTERING > even suggest, in spectrophotometric assays that direct the preparation and use of a standard curve using a reference standard, it is preferable, especially when assay is employed infrequently, not to use a standard curve but to make comparison directly against a quantity of reference standard approximately equal to that in a sample solution.

However, it will be a good practice to define an inhouse system suitability test to verify that the analytical instrument and the method are working properly before an analysis for assay methods (Assay, dissolution etc). A test of repeatability may be performed for verifying system suitability, by measurement of response of a standard solution, for example in five replicates (with %RSD to less than 2.0%). This is also suggested in the EDQM Quality Management Document Annex 3: Qualification of UV-Visible spectrophotometers PA/PH/OMCL (19) 100 R, EDQM.

We all know OOS Investigation with inconclusive root cause and invalidation of the OOS are viewed by agencies as critical. Review all invalidated OOS of batches within expiry. Make an assessment whether the laboratory investigation is thorough. Especially for OOS events where only one or a few samples are OOS in a sample set or chromatographic sequence with several samples, there shall be a thorough evaluation why the error impacted only specific sample(s) and not others. Such random errors point to all or several factors like negligence, poor awareness, poor analytical skill and competence, test method inadequacy, lack of clarity of instructions in test procedure. The CAPA should address programmes to enhance the awareness, skill set and competency of analysts. Also evaluate the method validation data and test method – whether robustness studies are adequate and all possible variabilities in the method are evaluated. Often cited causes in OOS due to laboratory error are insufficient sonication of sample, extraction, crushing of tablets, delay in sample preparation, injection and so on, rather than factors like variations in mobile phase composition or instrument wavelength or pH of mobile phase etc. (which are only routinely covered in method validation). Accordingly take up CAPA for addressing gaps in method validation and robustness studies, clarity of instructions in the test procedures.

Evaluate the hold time studies for all products. Review whether all parameters which can be impacted due to material holding are covered by hold time study data. Where gaps are identified develop a plan for establishing fresh hold time study data. A retrospective assessment may be taken up by reviewing historical data from actual commercial batches for each product – what is the maximum hold time for different stages after which further processing was performed. Also review the history of out of specifications. Document the outcome of assessment – a historical trend showing no impact on product quality after holding and processing the intermediate stages, minimal product failure (OOS) data can lend support to the hold time stability of intermediate stages.

Review and revise the procedure for hold time studies addressing the requirements. The WHO guideline General guidance on hold-time studies – Annexure 4, by WHO is a good reference point.