Equivalence Test for Two Emax Response Curves

Abstract

ABSTRACTThe evaluation of equivalence of response profiles including both dose–response and time–response curves plays an important role in the development of biosimilars. The 2014 FDA draft guidance of clinical pharmacology data to support a demonstration of biosimilarity to a reference product indicates that “if clinical pharmacology similarity between products is demonstrated, in some instances this may complete the clinical evaluation, and in others it may support a more targeted clinical development program.” The methodology how to demonstrate the equivalence on two response profiles, however, seems to be unclear in the regulatory guidance. In this article, we study the problem of equivalence test for two Emax response curves. First, we find the location where the largest absolution difference between the two Emax curves occurs, and show that the location only depends on the ratios of the parameters between the two Emax models. Then we propose a new method to test the equivalence between the two Emax curves. We compare this new method with several other methods using simulation, and show that the new method has good statistical properties. At the end, we discuss the implications of our findings in the application to the development of biosimilars.