EMA addresses comparative quality assessment methods

Regulatory NewsRegulatory News | 27 July 2021 |  By 

An updated reflection paper from the European Medicines Agency (EMA) tackles how statistics should be used in drug development when a comparative assessment of quality attributes needs to be made.
The paper touches on issues that arise both in the development of small molecules and in work being done in the area of biosimilars, where tricky manufacturing challenges and mutable organic molecules make defining the concept of “similarity” a challenge.
EMA’s human medicines committee (CHMP) had considered recommendations made during the public consultation period and in a 2018 workshop to issue a guideline rather than finalize the reflection paper (RP). However, the committee rejected the notion when they realized thinking is still unsettled regarding many aspects of the questions addressed in the draft.
In all, about 300 general and 1,000 specific comments were received by the time consultation ended in March 2018. The paper was cleared by the Biostatistics Working Party in June 2021 and adopted by CHMP on 22 July.
“[T]he decision was made to stick to the format of a RP, as many fundamental methodological aspects in relation to the comparative assessment of quality attributes (QA) continue to require in-depth considerations in an open manner,” wrote CHMP in discussing how the comments received were incorporated into the draft. “Further exchange with stakeholders and within regulatory expert groups revealed that it would be premature to provide explicit guidance for the diverse settings of QA comparison tasks.”
At the workshop, case studies in pre- and post-manufacturing changes and biosimilars were presented. Also on the agenda was a discussion of new statistical strategies and methodologies, as well as a review of the operating characteristics of similarity criteria in current use.
One recurring critique in the comments received during the consultation period and at the workshop, wrote EMA, was that the draft paper did not answer the question, “What is similarity?” Accordingly, a new concept of the “similarity condition” is introduced in the revised RP. This concept, which is distinct from a “similarity criterion,” makes clear that the similarity condition needs to be agreed on and defined before an appropriate similarity criterion can be selected.
During the consultation period, EMA also received concerns from “many” stakeholders that a too-prominent role of statistical methodology could result in “potentially overriding or ignoring the expertise of pharmaceutical development and manufacturing,” wrote EMA, adding, “These concerns were acknowledged.” The revised RP now puts forth “a rather open framework” for how to approach comparative assessment of quality attributes.
Although the revised RP does not dictate which similarity criteria should be used, it does provide an overview of frequently used criteria.
Mixed feedback was received during the consultation period about the optimal scope of the guideline, with some commenters advocating for a focus on biosimilars alone, while others “supported the more general context of the problem description.”
In the end, the committee felt there were enough commonalities between the statistics involved in how biologics and small molecules are compared, and in how manufacturing changes are compared, to warrant issuing a unified reflection paper.
One concrete recommendation that was added to the updated reflection paper is to plan prospectively for any data comparison of quality attributes that might influence regulatory decision-making. CHMP advises that a “quality attributes data comparison control” be constructed as part of the planning. In the same section of the RP, the content of a checklist of methodologic issues that had been in an appendix is now incorporated into the main body of the paper.


© 2022 Regulatory Affairs Professionals Society.

Discover more of what matters to you