For analytical similarity assessment in biosimilar studies, a two one-sided tests (TOST) procedure and a confidence interval (CI) approach are commonly employed to demonstrate biosimilarity between a proposed biosimilar (test) product and an innovative (reference) biological product. TOST is considered an official method for biosimilarity assessment by most regulatory agencies including the United States (US) Food and Drug Administrations (FDA) and the European Medicines Agency (EMA) of the European Union (EU). Since TOST is operationally equivalent to the CI approach, the CI approach is often used for biosimilarity assessment. However, FDA and EMA use the CI approach based on relative change and absolute change in mean response, respectively. This difference may lead to different conclusions regarding biosimilarity between the test product and the reference product, which causes some confusion among sponsors, regulatory agencies, and researchers in academia. This article attempts to clarify the confusion points and compare the FDA and EMA’s approach in terms of false positive rate and false negative rate when performing analytical similarity assessment.
Author(s): Yuqing Liu, and Shein-Chung Chow