Comparison of Metrics for Investigating Bioequivalence in Onset of Action in Incurred Bioequivalence Studies

Background: In those products where onset of action is clinically relevant, different metrics have been proposed to demonstrate bioequivalence, and different metrics, statistical methods and acceptance criteria have been employed in the United States, the European Union and Canada in the past. Recently, a harmonised approach has been recommended by the ICH M13A guideline. Methods: This work investigates the outcomes of several metrics in studies conducted in the Clinical Trials Unit of Hospital Universitario de La Princesa, Madrid, Spain, from 2000 to 2016, to identify the most adequate metric to reflect the onset of action. Results: An optimal single metric cannot be identified in these studies. The metrics show very high variability in most cases, making the conclusion of bioequivalence very unlikely with the usual sample sizes required for demonstration of bioequivalence based on Cmax and AUC. This is particularly evident if 90% confidence intervals are required to aim at inference of the study results to the whole population. Optimisation of the sampling times around tmax may be needed to reduce variability and increase reliability of the estimations. Conclusions: pAUC and tmax have been considered acceptable metrics under the ICH M13A guideline. For tmax, a non-parametric 90% confidence interval of the difference between the two formulations within a widened acceptance range may be sufficient for concluding BE in the early onset of action. Regarding pAUC, its high intra-subject variability can be addressed with the current scaling.