Analytical Similarity Statistical Guidance Out

Share

FDA has published a draft guidance, Statistical Approaches to Evaluate Analytical Similarity, with agency advice on the evaluation of analytical similarity for sponsors interested in developing biosimilar products under the Biologics Price Competition and Innovation Act of 2009. Specifically, the document says, it describes the type of information a sponsor of a proposed biosimilar product should obtain about the structural/physiochemical and functional attributes of the reference product, how that information is used in developing an analytical similarity assessment plan for the proposed biosimilar, and the statistical approaches recommended for evaluating analytical similarity.

In general, FDA says in the guidance, an analytical similarity assessment involves a comparison of structural/physiochemical and functional attributes using multiple lots of the proposed biosimilar product and the reference product.

“Conducting appropriate statistical analyses in the evaluation of analytical similarity can provide a high degree of confidence in the results and reduce the potential for bias,” the document says. “However, there are many challenges in designing the statistical analyses to be performed. First, there may be a limited number of reference product lots, and those obtained may be the result of biased sampling, leading to imprecise and possibly inaccurate estimates of the distribution of important quality attributes for the reference product. Second, there may also be a limited number of proposed biosimilar lots, and the available lots may not reflect the true variability of biosimilar product manufacturing. Third, there are a large number of potential quality attributes that can be compared in an evaluation of analytical similarity, and subjecting all of these attributes to formal statistical tests in the context of limited lots could lead to concluding incorrectly that a large number of truly highly similar products are not highly similar.”

FDA says it recommends addressing the challenges through a risk-based approach involving several steps including determining the quality attributes characterizing the reference product in terms of structural/physiochemical and functional properties, ranking the attributes according to their risk of potential clinical impact, and evaluating the attributes/assays according to one of three tiers of statistical approaches based on a consideration of risk ranking and other factors.

Read more