Statistical Considerations in Assessing Bioequivalence with Baselines: A Case Study of Endogenous Drugs

Abstract

Objectives To assess bioequivalence between two endogenous drugs in 2 × 2 crossover trial with baseline measurements. Methods Two statistical methods are applied to assess bioequivalence between two endogenous drugs in 2 × 2 crossover trials. The first method is based on the current regulatory guideline published by Ministry of Food and Drug Safety (MFDS), which is based on the difference between baseline measurements and responses. The second method is more general approach, so-called general linear model method, which is defined the baseline measurements as covariates. Results The first method based on current guideline shows that two drugs are not bioequivalent; however, the second method by general linear model shows that two drugs are bioequivalent. When the baselines of the subjects are expected to be highly variable, general linear model approach is more suitable to assess the bioequivalence by adjusting high subjects’ variations. Conclusions General linear model with covariates should be considered in assessing bioequivalence of endogenous substances when highly subject variations of baseline measurements are expected. Key words: 2 × 2 crossover design, Analysis of covariance, Baseline measurements, Bioequivalence, Endogenous drug