Prediction of hepatic microsomal intrinsic clearance and human clearance values for drugs.

Twenty-nine drugs of different structures were used in theoretical QSAR analysis of human hepatic microsomal intrinsic clearance (in vitro T(1/2) and in vitro CL'(int)) and whole body clearance (CL(blood)). The examined compounds demonstrated a wide range of scaled intrinsic clearance values. Constitutional, geometrical, physico-chemical and electronic descriptors were computed for the examined structures by use of the Marvin Sketch 5.1.3_2, the Chem3D Ultra 7.0.0 and the Dragon 5.4 program. Partial least squares regression (PLSR), has been applied for selection of the most relevant molecular descriptors and development of quantitative structure-activity relationship (QSAR) model for human hepatic microsomal intrinsic clearance (in vitro T(1/2)). Optimal QSAR models with nine and ten variables, R(2)>0.808 and cross-validation parameter q(pre)(2)>0.623, were selected and compared. Since the microsomal in vitro T(1/2) data can be used for calculation of in vitro CL'(int) and in vivo CL(blood), the developed QSAR model will enable one to analyze the kinetics of cytochrome P450-mediated reactions in term of intrinsic clearance and whole body clearance. A comparison is made between predictions produced from the QSAR analysis and experimental data, and there appears to be generally satisfactory correlations with the literature values for intrinsic clearance data.