Jiansong Yang, Fan Zhang
Department of Clinical Pharmacology Modeling and Simulation, GSK, China
Objectives: A single-dose, randomized, parallel-group, open-label study in 126 healthy subjects was conducted to investigate BE of 250 mg lamotrigine XR relative to the combination of 200 and 50 mg lamotrigine XR. The 90% CIs for geometric mean ratio of AUC0-∞ and Cmax were (0.735, 0.956) and (0.894, 1.045), respectively. Although AUC0-∞ failed to meet the BE criterion (0.80, 1.25), the cause of AUC0-∞ difference between two groups seemed to stem from clearance (CL) rather than absorption. The aim of this study was to perform post hoc analyses using population pharmacokinetic (PK) modeling approach.
Methods: Pop PK model for lamotrigine XR was developed using nonlinear mixed-effect models in Monolix (version 4.3.3) by the SAEM algorithm combined with a MCMC procedure. Prior data indicated a polymorphic metabolism of lamotrigine, and a bimodal distribution of CL was also observed in this study. Thus, the Pop PK model incorporated a mixture variable on CL with two populations. Formulation (250 mg lamotrigine XR & 200+50mg lamotrigine XR) was tested as a covariate for PK parameters.
Results: The PK of lamotrigine XR was best described using a one-compartment model with 0-order oral absorption and an absorption lag time. The final mixture model separating subpopulation with different CL was able to significantly reduce the between subject variability and objective function value. CL values for subjects with slow and quick clearance were 0.573L/h and 1.96 L/h, respectively. Formulation wasn’t a significant covariate for PK parameters, suggesting the two formulations were bioequivalent. Bayesian post-hoc subpopulation classification indicated there were 9 and 3 subjects with slow clearance in (200 mg + 50 mg) and 250 mg treatment groups, respectively. This uneven numbers of subjects with slow clearance between the two BE groups might explain why AUC0-∞ failed to meet BE criterion. Excluding the subjects with slow CL (11% of overall subjects) resulted in meeting BE criterion for both AUC0-∞ and Cmax, with 90% CIs for geometric mean ratios of (0.832, 1.062) and (0.915, 1.076), respectively.
Conclusions: The AUC difference between two BE groups was a result of elimination difference, instead of absorption difference. Uneven ratios of high:low clearance subjects between the two BE groups were the main cause of failure to demonstrate BE for AUC0-∞. Population approach, although not commonly used in BE analysis, proved a valuable tool for post hoc analyses.
Reference: PAGE 24 () Abstr 3387 [www.page-meeting.org/?abstract=3387]
Poster: Drug/Disease modeling - Other topics