Figure 4
Comparison of the effects of twelve inhibitor concentration combinations on the predictions of increase in the AUCs in clinical DDI studies. Reversible inhibition and time-dependent inhibition of CYP3A were predicted using uncorrected intestinal inhibitor concentrations, and uncorrected (left panel) or Fcyto-corrected (right panel) hepatic inhibitor concentrations. Twelve intestinal and hepatic inhibitor concentration combinations were evaluated for prediction bias (GMFE) (a,b), precision (RMSE) (c,d), and for the percentage of the predictions that were within two-fold of the observed AUC values (e,f). A perfect prediction of all actual values would give a GMFE value of 1, and a lower RMSE value denotes a greater precision of the prediction. Input parameters are given in Methods and in Supplementary Table S5. See Supplementary Table S2 for more details and results of predictions based on the mixed term approach. GMFE, geometric mean-fold error; Fcyto, cytosolic bioavailability; [I]ave,u, unbound average inhibitor concentration; [I]g, intestinal inhibitor concentration; [I]inlet,ave,u, unbound hepatic inlet inhibitor concentration based on [I]ave; [I]inlet,max,u, unbound hepatic inlet inhibitor concentration based on [I]max; [I]max,u, unbound peak inhibitor concentration; RMSE, root-mean square error.
