Successful Prediction of Human Pharmacokinetics by Improving Calculation Processes of Physiologically Based Pharmacokinetic Approach.

ABSTRACT

The physiologically based pharmacokinetics (PBPK) model is a major mechanistic approach for predicting human pharmacokinetics (PK) using drug-specific and physiological parameters but has been difficult to use for human PK prediction with acceptable accuracy. Here, we report a newly developed PBPK approach that incorporates the mechanism of albumin-mediated membrane penetration in the liver and interspecies correlation for unbound tissue fractions. To verify the utility of our PBPK approach, we used 12 drugs that are mainly eliminated by hepatic metabolism to compare the prediction accuracy with a conventional PBPK approach and to observe human PK parameters. We found the predictive accuracy for total clearance (CLtot), distribution volume at the steady state (Vss), elimination half-life (t1/2), and plasma concentration at the last measurable time point (Clast) of our PBPK approach to show better absolute average fold error and percentage within 2-fold error (1.6-1.8 and 67%-92%, respectively) compared with values obtained from the conventional PBPK approach (2.1-2.4 and 42%-67%, respectively). As our approach can use parameters obtained in early drug screening, it could help accelerate successful nomination of drug candidates by optimizing the pharmacokinetics of new chemical entities by directly using predicted human PK profiles.