M. Marchand (1), M. Chenel (2), H. Merdjan (1), S. Mouksassi (1)
(1) Pharsight, a CertaraTM company, St. Louis, MO, USA; (2) Department of Clinical Pharmacokinetics, Institut de Recherches Internationales Servier, France
Objectives: The objective was to perform population pharmacokinetic (PK) modeling and clinical trial simulations (CTS) to inform the study design of a future bioequivalence study comparing a reference, marketed, drug product with an innovative supra-bioavailable formulation of the same active moiety, exhibiting non-linear absorption characteristics.
Methods: A pilot clinical bioavailability study comparing three doses of the novel formulation (equivalent to 25%, 51% and 102% of the reference dose) was performed in 16 healthy subjects. A population PK model was developed using NONMEM based on combined drug concentrations resulting from the administration of both formulations (novel and reference). Following successful qualification of the population PK model, CTS were performed using various sources of variability including uncertainty on fixed effect parameters. The design selected to simulate a definitive bioequivalence study was a 2 x 2 Latin square comparing the 2 formulations after single dose, complete washout between periods. To determine the sample size, a target statistical power of 90% was selected. AUC and Cmax were derived on each simulated profile for each simulated trial. An ANOVA model was used to derive 90% CI for AUC and Cmax for each simulated trial. 1000 trials were simulated for each sample size.
Results: In the PK model, drug concentrations were analyzed simultaneously with different absorption models and a common disposition model for both formulations. The bioavailability of the novel formulation relative to the reference was adequately described by a power model with a population estimate of -0.339 for the exponent. Three dose levels of the novel formulation (equivalent to 70%, 76% and 82% of the reference, expressed as the active moiety) were selected. Thanks to the supra-bioavailability, a dose containing 76% of the reference dose is likely to demonstrate bioequivalence with the reference. Moreover, 20 subjects would be sufficient to demonstrate bioequivalence on both Cmax and AUC, with a statistical power in excess of 90%.
Conclusions: The modeling and simulation approach supported the design (i.e. optimal dose and sample size) of a definitive bioequivalence study between a novel formulation and the reference one. The preliminary results of a confirmatory PK study performed with the selected dose showed a close agreement with the CTS-based predictions.
Reference: PAGE 21 () Abstr 2317 [www.page-meeting.org/?abstract=2317]
Poster: Study Design