Evaluation of partial area under the curve in bioequivalence studies using destructive sampling design

Product blood-level in vivo bioequivalence (BE) studies typically involve complete blood concentration-time profiles generated for each subject. Accordingly, each subject provides the estimates of the rate and extent of drug absorption. However, repeated blood draws are not always feasible for studies using small animals because of handling or blood volume (e.g., fish or in toxicokinetic studies when only single samples can be taken per animal). Although several proposals have been published for comparing the product extent of absorption, the issue of comparative absorption rates remains unresolved. In this paper, we propose to apply the Bailer-Satterthwaite-Fieller confidence interval for estimating the ratio of partial area under the curve in studies that use a destructive sampling design (one blood sample per subject). To characterize the behavior of this alternative approach, we examine the impact of partial area cutoff time, blood sampling schedule, and the number of subjects included at each sampling time. Using simulated situations reflective of the issues encountered with immediate-release veterinary formulations, we compare BE conclusions resulting from the use of this approach with simulated results that assuming repeated blood draws from each study subject.