Assessment of Intrasubject Variability of Pharmacokinetic Parameters in Clinical Studies

According to the current bioavailability/bioequivalence regulations by the Food and Drug Administration (FDA), comparison of mean bioequivalence values between two formulations of the same drug in pharmacokinetic studies is sufficient for approval of the new formulation in a new drug application. Intrasubject variability of pharmacokinetic parameters may still statistically differ between two bioequivalent formulations which cannot be proven based on the typical 2 × 2 crossover design if no repeated measures for the same formulation can be evaluated. When repeated meaures are required in a study protocol, a parallel group study design may be preferred over a crossover study design due to a pharmacological and/or clinical reason. This paper evaluates two issues regarding intrasubject variability based on a parallel group design with repeated measures by two different approaches: 1. Conventional analysis of variance and 2. Individual subjects with respect to: an estimation procedure for variance and coefficient of variation (CV) with an emphasis on the different estimates for the CV, and different methods (parametric or nonparametric) for the comparison of variability with regard to variance and CV.