The Statistical Analysis of Data from Bioequivalence Studies Using Stable Isotopes

This paper examines the statistical analysis of bioequivalence studies using stable isotopes based on the relative bioavailabilities of test and reference formulations with respect to a third labeled formulation. It is shown that the analysis of relative bioavailabilities requires the assumption that the labeled formulation is not differentially affected by the presence of the test and reference formulations, and that the properties of the statistical test of bioequivalence are very sensitive to this assumption. The reduction in sample size using stable isotope data is dependent on the relative values of the correlations of the test and reference formulations with the simultaneously administered labeled formulations compared to correlations between days. Consideration of a simple compartment model shows that a reduction in sample size will usually occur but that the degree of reduction cannot be predicted.