On the Question of Sample Size Planning in Bioequivalence Studies from the Perspective of Desired Statistical Power

Vollständiger Abstract

Worum geht es in dieser Arbeit?

INTRODUCTION. The sample size (number of volunteers or target animals) is a significant element of in vivo bioequivalence (BE) studies; however, the algorithm for its calculation is not specified in regulatory documents. Increasing the sample size beyond the minimum allowable number may be unethical and lead to unreasonable financial costs. AIM. This study aimed to review scientific approaches to justifying the sample size in bioequivalence studies from the perspective of the desired statistical power and to retrospectively evaluate our own experiments to improve the planning of such studies. MATERIALS AND METHODS. The study analyzed regulatory documents of the Eurasian Economic Commission, the Ministry of Agriculture of Russia, scientific publications and other open-access sources. Using the intra-individual coefficient of variation (CV intra ) and RStudio software, we estimated the statistical power of the performed studies for the pharmacokinetic parameters maximum concentration (C max ) and area under the concentration-time curve (AUC) and calculated the sample size required to achieve a power of at least 80% (if this power level was not achieved). RESULTS. The use of an evidence-based approach to sample size estimation using CV intra (for crossover design studies) or the coefficient of total variation (CV total ) (for parallel design studies) is limited by the lack of a unified calculation algorithm and insufficient available data on previous BE studies. A retrospective evaluation of BE studies conducted with our participation revealed that in many cases, the recommended power level was not achieved. For low-molecular-weight-heparin (LMWH) products for human medical use, the calculated sample size was up to 100 or more subjects, while for veterinary drugs, it ranged from 14–18 to 50–80 animals. CONCLUSIONS. To correctly estimate the sample size when planning clinical trials of LMWH products, which may be highly variable, it is advisable to accumulate data on the CV intra parameter and to consider two-stage (adaptive) designs. Increasing the sample size and introducing adaptive designs for veterinary medicinal products may lead to an unjustified increase in study duration and cost and to the market introduction of such products. Therefore, analysis and comprehensive discussion of the planning of preclinical and clinical studies of veterinary medicinal products are advisable.