Application of unbound liver-to-plasma concentration ratio to quantitative projection of cytochrome P450-mediated drug-drug interactions using physiologically-based pharmacokinetic modelling approach.

1. This study evaluated the prediction accuracy of cytochrome P450 (CYP)-mediated drug-drug interaction (DDI) using minimal physiologically-based pharmacokinetic (PBPK) modelling incorporating the hepatic accumulation factor of an inhibitor (i.e. unbound liver/unbound plasma concentration ratio [K-p,K-uu,K-liver]) based on 22 clinical DDI studies. 2. K-p,K-uu,K-liver values were estimated using three methods: (1) ratio of cell-to-medium ratio in human cryopreserved hepatocytes (C/M-u) at 37 degrees C to that on ice (K-p,K-uu,K-C/M), (2) multiplication of total liver/unbound plasma concentration ratio (K-p,K-u,K-liver) estimated from C/M-u at 37 degrees C with unbound fraction in human liver homogenate (K-p,K-uu,K-cell) and (3) observed K-p,K-uu,K-liver in rats after intravenous infusion (K-p,K-uu,K-rat). 3. PBPK model using each K-p,K-uu,K-liver projected the area under the curve (AUC) increase of substrates more accurately than the model assuming a K-p,K-uu,K-liver of 1 for the average fold error and root mean square error did. Particularly, the model with a K-p,K-uu,K-liver of 1 underestimated the AUC increase of triazolam following co-administration with CYP3A4 inhibitor itraconazole by five-fold, whereas the AUC increase projected using the model incorporating the K-p,K-uu,K-C/M, K-p,K-uu,K-cell, or K-p,K-uu,K-rat of itraconazole and hydroxyitraconazole was within approximately two-fold of the actual value. 4. The results indicated that incorporating K-p,K-uu,K-liver into the PBPK model improved the accuracy of DDI projection.