All Times EDT
Keywords: Equivalence tests, Multivariate analysis, Hierarchical Sampling, API sameness
The assessment of active pharmaceutical ingredient (API) sameness between a test product and the reference listed drug (RLD) product is an important component of pharmaceutical equivalence (PE) assessment for generic products. API sameness assessment can be challenging, especially for drug products with complex APIs (i.e., API with heterogenous chemical structures and/or heterogenous mixtures), where the API sameness assessment often involves analytical methods that generate complex multi-dimensional data for assessment, e.g., by liquid chromatography–mass spectrometry (LC-MS). Thus, one challenge for demonstrating API sameness, among other things, is the comparison of generated multi-dimensional data (e.g., representing multiple components of interest) from the test and reference products. In this study, we developed a two-stage statistical approach that can be used to conduct quantitative assessment of API sameness for drug products with complex heterogenous API mixtures. Briefly, in Stage 1, an equivalence testing for summation of all API components is considered and a well-accepted margin in Stage 1 is applied. In Stage 2, we constructed a statistical model to describe the measured data from the RLD product, by which datasets of acceptable and unacceptable test products can be simulated. Subsequently, we developed a multivariate statistical equivalence test suitable for comparing multi-dimensional data from Test and RLD products. Given the simulated data and developed test, statistical criteria were evaluated and established to reach reasonable type I/II errors. As a case example, the LC-MS data from a drug product with complex API obtained from a natural source were used to demonstrate the potential application of the developed statistical approach for comparing complex heterogenous mixtures.