Proposal of a prediction model describing the metabolic interaction between warfarin and sorafenib using a population pharmacokinetic approach.

OBJECTIVE: We attempted to examine the applicability of a population pharmacokinetic-pharmacodynamic (PK-PD) model describing the metabolic interaction between warfarin and sorafenib due to CYP2C9 inhibition and to predict the plasma concentrations of sorafenib and S-warfarin, the international normalized ratio (INR), and the optimal maintenance dose of warfarin in the presence of sorafenib in vivo. MATERIALS AND METHODS: The sorafenib inhibition constant for S-warfarin metabolism was determined in vitro, and the unbound fraction in the liver was estimated using the published equations. A population PK-PD model describing the interaction between warfarin and sorafenib assuming competitive metabolic inhibition of S-warfarin by sorafenib was developed using NONMEM. The model was evaluated using clinical data and INR collected from the literature. RESULTS: The observed time courses of INR retrieved from Japanese and Caucasian patients given warfarin and sorafenib were mostly within the 90% range of the predicted values. Then, we predicted the plasma sorafenib and S-warfarin concentrations and INR after administration of warfarin (3 mg/day) alone and warfarin + sorafenib (800 mg/day). The predicted mean plasma S-warfarin concentration and INR at steady state were almost 6 and 2 times greater, respectively, in the presence of sorafenib than those for warfarin alone. The predicted S-warfarin concentrations and INR after reduction of the warfarin dose (0.5 mg/day) in the presence of sorafenib were comparable to those after 3 mg/day warfarin alone. CONCLUSION: The proposed population PK-PD model has the potential to predict an increase in INR quantitatively after concurrent administration of warfarin and sorafenib.