Summary of the FDA, UMD Meeting on June 15, 2023, regarding “Mitigation Strategies for NDSRIs: Quality and BE Considerations”
Aloka Srinivasan, Ph.D.
Principal and Managing Partner
Purpose Of the Workshop
The current workshop was FDA’s attempt to communicate their current thinking and efforts to address the NDSRIs. The presentations which were from FDA researchers as well as industry stakeholders, discussed the risks of forming NDSRIs in certain drug products, strategies to mitigate these risks, and considerations in assessing the safe levels of NDSRIs. The workshop provided some approaches to prevent or mitigate the formation of NDSRIs in drugs substances and drug product and finally, the workshop discussed the potential impacts of such reformulation on the bioequivalence of generic products, and strategies to efficiently address these issues. The sessions are provided below:
Session 1: Risk of Forming NDSRIs and Strategies to Mitigate the Risk
Session 2: Safety & Risk Assessment of NDSRIs for Human Health
Session 3: Impact of Reformulation on the Bioequivalence of Generic Products
The presentations are available in the website, https://www.eventbrite.com/e/mitigation-strategies-for-ndsris-quality-and-be-considerations-tickets-602756409647.
Summary of Session 1: Risk of Forming NDSRIs and Strategies to Mitigate the Risk
Speakers summarized: Andre Raw (FDA), Justin Moser (Merck), Jan Yang (FDA), Sander van Gessel (DFE Pharma), Martin Ehlart (Apotex), Diaa Shakleya (FDA), Marko Trampuz (Sandoz), Rok Grahek (Sandoz)
The presentations were insightful for this session. There were presentations from FDA and industry on the impact of nitrite scavengers on blocking NDSRIs in pharmaceuticals. Well known nitrite scavengers like ascorbic acid, caffeic acid, ferulic acid, para-aminobenzoic acid were studies. It was also shown that using excipients with low levels of nitrite reduced the levels of NDSRIs drastically. One of the presentation showed that the spatial location of nitrosation-susceptible amine impurities relative to the API can impact nitrosation kinetics. The solubility-limited impurity purge (SLIP) test was employed to identify when an impurity forms its own pure solid phase during crystallization. The conclusion was that controlling secondary amine impurities in the final product is crucial in pharmaceutical manufacturing.
Overall, this session showed that both industry and agency are far from finding a solution to the NDSRI problem. At the current time, excipient manufacturers have not made progress towards reducing nitrite in their excipients. Thus, the possibility of using excipients with low levels of nitrite is not practical. Also, the presentations showed, there is no single nitrite scavenger that can help in controlling NDSRIs. Every API and formulation may need to be evaluated to better understand which excipient works. Comments were made regarding the fact that the scavengers are sometimes leading to an increase in other impurities in the drug substances and drug products. It was acknowledged that the studies were very rudimentary and focused on simple dosage forms like IR tablets. Also, many of these nitrite scavengers are not in the FDA’s IID (Inactive Ingredient Database) and will need a full battery of non-clinical testing to qualify as excipients and additives. FDA was questioned if they will accept information available in public domain to support the use of some of the non-IID nitrite scavengers. To this FDA remained silent. The take home message from the discussion was that both agency and pharmaceutical industry have a lot to understand before considering nitrite scavengers as the solution to formation of NDSRIs.
Summary of Session 2: Safety & Risk Assessment of NDSRIs for Human Health
Speakers Summarized: Dr. Robert Heflich (FDA), Dr. Naomi Kruhlak (FDA). David Ponting (Lhasa), Andrew Teasdale (AstraZeneca), Raphael Nudelman (Teva)
FDA on the basis of their own research have confirmed that the modified Ames assay with hamster S9, preincubation, and careful selection of solvents is the most promising protocol for initial testing of nitrosamine drug substance-related impurities (NDSRIs) and that tester strain TA1535 was the most sensitive strain. However, the details of the protocol will possibly be available during later meetings.
It was also communicated that FDA is working on using other mammalian cell mutagenicity assays (e.g., TK6 and HepaRG cells) using metabolically competent human cells to further corroborate Ames Assay findings.
FDA confirmed the collaboration with MultiCASE on QSARFlex/nearest neighbor analysis to guide their initial surrogate candidate selection process. FDA appears to be moving towards using this surrogate selection process in combination with assessing the relative potency of structural features:
Industry presented that some of the same NDSRIs have been evaluated in BigBlue transgenic assay, Comet assay, and Duplex Sequencing assay with interesting results:
- Duplex Sequencing appears more sensitive at low in vivo doses than BigBlue Transgenic (TGR) because of lower inter-animal variability.
- Comet assay might be a suitable biomarker of in vivo genotoxicity of NDSRIs
Industry made an argument that a negative finding in the hepatic Comet assay should allow control of a NDSRI to the ICH M7 TTC of 1.5 µg/day and that a negative TGR/Duplex assay should allow control at ICH Q3 limits.
SAR Working Group (Lead by LHSASA) discussed the following:
- Combination of LHASA, CPDB, and Instem Nitrosamine Databases
- Will include proprietary and donated nitrosamine data from Industry including 147 Ames results for 104 complex nitrosamines
- Identifying nitrosamines that are mechanistically excluded from cohort of concern
- Identifying which historical carcinogenicity data may be dependable even though limited doses etc. – Need to be able to use these older datasets
- Working further on SAR – how should best read-across analogue be selected
- Creating AI/PDE monographs for reference nitrosamines/Harmonized guidance
- More research on whether less than lifetime is applicable to NDSRIs
There was general consensus among Industry participants that M7 LTL approach and MW correction approach is conservative and applicable to nitrosamines. Future focus could be on the following:
- Work on the Permissible Daily Exposure (PDE) approach for NDSRIs – finding an appropriate point of departure value (POD; e.g. In vivo mutagenicity data) and using adjustment factors in accordance with ICH Q3C.
- Work on computational models incorporating quantum mechanics to improve AI assessment. During the panel discussion, FDA was lukewarm on this topic, as they think that the computational models should be supported by adequate data.
It was agreed by industry and FDA that the transgenic mouse studies are not encouraged without discussion with FDA, due to challenges with this study and minor changes in protocol affecting the outcome.
From industry’s presentation, it appears that the modified AMES test followed by Duplex sequencing as a key in vivo assay moving forward which can provide both sensitive mutagenicity data, a usable in vivo dose response and POD, can be used with non-transgenic strains of rodents, all in a more cost effective/timely format than currently available transgenic assays.
FDA remained non-committal regarding the use of LTL and MW adjustment in determining acceptable intake of NDSRIs.
From the presentations, it was apparent that agencies may be coming around to allowing the use modified Ames assay as a strong line of evidence for mutagenicity of NDSRIs – but will probably need a confirming mammalian mutagenicity assay of some type. There are some publications on the way for the AMES Assay and the in silico studies.
Summary of Session 3: Impact of Reformulation on the Bioequivalence of Generic Products
Speakers: Chris Brode (Pharmaron), Sook Wah (UCSF), Fang Wu (FDA), DongMei Lu (FDA)
This was a short session as the work was still nascent in nature. The speakers focused on the potential impact of reformulation on the bioequivalence of the drug products. The speakers used modeling and simulation approaches to assess bio inequivalence risks.
At the end, FDA did a candid presentation on the gaps in their guidance on nitrosamines when it comes to NDSRIs. However, there was no communication regarding updating of the guidance or effective mode of data sharing. The FDA mentioned that they are considering an extension of the deadline to completion of risk evaluation and submission of amendments to dossiers related to nitrosamine risk, including risk of NDSRIs. The FDA stated that they are aware that the current deadline of the end of September 2023 may be difficult to meet for the industry and requested the industry to come to the agency with proposals on this issue.
The workshop was well conducted and informative. However, the information shared demonstrated that neither industry nor FDA has a clearly mapped path forward when it comes to blocking NDSRIs in pharmaceuticals. It was amply clear that a significant amount of work will be needed to reduce NDSRIs in drugs or control them appropriately based on practical acceptable intakes and suitable analytical methods. Industry will need to wait for FDA’s publication on the modified AMES assay and the in silico studies, that were mentioned in the workshop. There needs to be negotiation with the industry regarding how the data that the FDA accumulates on potency or lack of it in NDSRIs is shared with the entire industry. Nitrosamines are a safety issue, and the consensus is that it is the responsibility of FDA to keep the industry and American public aware of the safety levels of these impurities.
The take home message from the workshop was that it could take industry several years to complete this herculean task of controlling NDSRIs in drug substance and drug products and FDA’s transparency related to the NDSRIs is critical for their success. Also, the deadline for completion of confirmatory testing and updating of dossiers with NDSRI information should be deferred significantly so that that the industry can work better on root cause analysis, control, and mitigation strategies.
Acknowledgements: The author thanks Chuck Lambert, Principal of Intrinsik, for his valuable inputs.