Robustness of the Conclusion of Bioequivalence; A Non-Parametric Comparison

Abstract

Generics are substituted on an individual level when bioequivalence has been established. However, it is known that similarity in exposure may be different in individual subjects. The goal of the presented study is to investigate if pharmacokinetic subpopulations exist when subjects are exposed to bioequivalent formulations of gabapentin. Data was gathered in a 4-way crossover study in 24 subjects with 4 different oral formulations of gabapentin previously concluded to be bioequivalent by non-compartmental analysis.1 The non-parametric pharmacokinetic modelling software Pmetrics was used, which is capable of identifying unprespecified subpopulations in pharmacokinetic data.2 Models were validated using visual predictive checks (VPCs) and the Akaike information criterion (AIC). In Pmetrics, 4 different structural models describing gabapentin pharmacokinetics were compared, i.e. a 1- and 2-compartment model with absorption constant, with or without absorption lag time (Tlag). Based on VPCs as well as the reduction in AIC (129, 20 and 107 respectively) it was concluded that a 2-compartment model with Tlag best explained observed data. The regression equation of the observed versus the predicted concentration is y=0.96x + 0.74 with an R2 of 0.78 (bias -0.083: and imprecision: 0.615). Using the final Pmetrics model, differences in absorption parameter estimates were observed for a subpopulation that was administered one of the gabapentin formulations, suggesting an altered pharmacokinetic profile of the gabapentin formulation in these subjects. The developed model appears to indicate differences in gabapentin pharmacokinetics for a subpopulation of subjects, despite demonstrated bioequivalence between the used formulations. Following further optimisation and validation of the Pmetrics model, we plan to perform simulations to investigate the extent of the influence of these subpopulations on the pharmacokinetic similarity between the bioequivalent formulations.