Bioequivalence: An overview of statistical concepts

Bioequivalence (BE) means the absence of a greater-than-allowabledifference between the systemic bioavailability of a test product andthat of a reference product. Studies to test the BE of drug products,and the statistical basis for their design, analysis andinterpretation, have evolved over the last two decades. A crossoverdesign is preferred over a parallel-group design as it segregates theinter-subject variation (which is not product-dependent) from theintra-subject variation (which is product-dependent). The value oftesting two one-sided null hypotheses of non-equivalence at asignificance level of 0.05, and the importance of estimating a 90%confidence interval of the ratio (test/reference) of mean AUC and Cmaxvalues, and of the difference between mean Tmax values, are nowrecognized and form the current standards for BE. The number ofsubjects required for a BE study with the desired power (at least 0.80)and significance level (0.05), depends on the expected deviation of thetest product from the reference product and the error varianceassociated with the bioavailability parameters (AUC, Cmax, Tmax etc.)of the drug substance. At present, according to the Indian regulatoryauthority, the number of subjects required to conduct a BE study is 12which is inadequate for most drug substances by the currentinternational standards and criteria. This review expounds theforegoing principles of BE testing.