Prediction of drug‐drug interactions between various antidepressants and ritonavir using a physiologically based pharmacokinetic model

Abstract

Depression can impact on the treatment of HIV infection, and effective treatment of depressive conditions can have a beneficial effect improving adherence. However antidepressant treatment requires long‐term maintenance, and is prone to pharmacokinetic drug‐drug interactions (DDI) with antiretrovirals. The aim of this study was to predict the magnitude of DDI between ritonavir (RTV) and the most commonly prescribed antidepressants using a physiologically based pharmacokinetic (PBPK) model simulating virtual clinical trials. In vitro data describing the physiochemical properties, absorption, metabolism, induction and inhibitory potential of RTV and five antidepressants were obtained from published literature. Interactions between RTV and antidepressants were evaluated using the full PBPK model implemented in the Simcyp Population‐based Simulator (Version 11.1, Simcyp Limited, UK) and virtual clinical studies were simulated on 50 Caucasian subjects receiving 100mg bid of RTV for 21 days plus sertraline (100mg qd), citalopram (40mg qd), fluoxetine (20mg qd), venlafaxine (25mg qd) and then from day 14–21. Simulated pharmacokinetic parameters were compared with observed values available in the literature. The simulated PK parameters of RTV, sertraline, citalopram, fluoxetine, mirtazepine and venlafaxine given alone at standard dosage were similar to reference values obtain from published clinical studies. The effect of simulated RTV co‐administration on sertaline, fluoxetine and venlaflaxine was an AUC decrease of 40%, 26% and 6%, respectively and on mirtazepine and citalopram, an AUC increase of 60% and 20% respectively. The magnitude of the simulated DDI between RTV and the antidepressants was overall weak to moderate according to the classification of the FDA. The modest magnitude of these drug‐drug interactions could be explained by the fact that antidepressants are substrates of multiple isoforms thus metabolism can still occur through CYPs that are not or weakly impacted by RTV. Although from a pharmacokinetic point of view, venlafaxine or citalopram represent better candidates for patients on RTV, there are many considerations in seeking to optimize antidepressant therapy. The next stage in this work is to simulate DDI between boosted protease inhibitors and antidepressants. IVIVE is a useful tool for both prediction of drug‐drug interactions and design of prospective clinical trials, simulating optimal sample size, and selection of doses.