Development and verification of a physiologically based pharmacokinetic model of dronedarone and its active metabolite N-desbutyldronedarone: Application to prospective simulation of complex drug-drug interaction with rivaroxaban

Abstract

Aims: Despite potential enzyme- and transporter-mediated drug-drug interactions (DDIs) between dronedarone and rivaroxaban in atrial fibrillation (AF) patients, pharmacokinetic/pharmacodynamic data remain limited to guide clinical practice. We aimed to develop, verify and validate a physiologically based pharmacokinetic (PBPK) model of dronedarone and its major metabolite, N-desbutyldronedarone (NDBD), to prospectively interrogate this clinically relevant DDI in healthy and mild renal impairment populations.

Methods: The middle-out development of our PBPK model combined literature-derived or in-house in vitro data, predicted in silico data and in vivo clinical data. Model verification was performed for intravenous and oral (single and multiple) dosing regimens. Model validation for the accurate prediction of cytochrome P450 (CYP)3A4- and P-glycoprotein-mediated DDI utilized simvastatin and digoxin as respective victim drugs. Rivaroxaban-specific inhibitory parameters of dronedarone and/or NDBD against CYP3A4, CYP2J2, OAT3 and P-glycoprotein were incorporated into the PBPK-DDI model for prospective dronedarone-rivaroxaban DDI simulation.

Results: Dronedarone and NDBD PK following clinically relevant doses of 400 mg dronedarone across single and multiple oral dosing were accurately simulated by incorporating effect of auto-inactivation on dose nonlinearities. Following successful model validation, nondose-adjusted rivaroxaban-dronedarone DDI in healthy and mild renal impairment populations revealed simulated rivaroxaban area under the plasma concentration-time curve up to 24 h fold change greater than dose exposure equivalence (0.70-1.43) at 1.65 and 1.84, respectively. Correspondingly, respective major bleeding risk was 4.24 and 4.70% compared with threshold of 4.5% representing contraindicated rivaroxaban-ketoconazole DDI.

Conclusion: Our PBPK-DDI model predicted clinically significant dronedarone-rivaroxaban DDI in both healthy and mild renal impairment subjects. Greater benefit vs. risk could be achieved with rivaroxaban dose reductions to at least 15 mg in mild renal impairment subjects on concomitant dronedarone and rivaroxaban.

Keywords: N-desbutyldronedarone; dronedarone; drug-drug interactions; physiologically based pharmacokinetic modelling; rivaroxaban.