Comparison of the use of liver models for predicting drug clearance using in vitro kinetic data from hepatic microsomes and isolated hepatocytes

Purpose. To compare three liver models (well-stirred, parallel tube, and dispersion) for the prediction of in vivo intrinsic clearance (CLint), hepatic clearance (CLh), and hepatic availability (F-h) of a wide range of drugs in the rat using in vitro data from two in vitro sources. Methods. In vitro CLint was obtained from studies using isolated rat hepatocytes (35 drugs) or rat liver microsomes (52 drugs) and used to predict in vivo CLint using reported scaling factors, and subsequently CLh and F-h were predicted based on the three liver models. In addition, in vivo CLint values were calculated from the reported values of CLh based on each of the three models. Results. For all of the parameters, predictions from hepatocyte data were consistently more accurate than those from microsomal data. Comparison of in vitro and in vivo CLint values demonstrated that the dispersion model and the parallel tube model were comparable and more accurate (less bias, more precise) than the well-stirred model. For CLh and F-h prediction, the three models performed similarly. Conclusions. Considering the statistics of the predictions for three liver models, the use of parallel tube model is recommended for the evaluation of in vitro CLint values both from microsomes and hepatocytes. However, for the prediction of the in vivo drug (hepatic) clearance from in vitro data, as there are minimal differences between the models, the use of the well-stirred liver model is recommended.