Quantitative Property–Property Relationship for Screening-Level Prediction of Intrinsic Clearance: A Tool for Exposure Modeling for High-Throughput Toxicity Screening Data
Christopher R. Kirman, Lesa L. Aylward, Barbara A. Wetmore, Russell S. Thomas, Mark A. Sochaski, Stephen Ferguson, Susan A. Csiszar, Olivier Jolliet · Applied In Vitro Toxicology · 2015
Abstract A quantitative property–property relationship model was developed to predict intrinsic clearance (Cl int ) for 403 chemicals across a broad range of chemical properties and validated using a published set of Cl int measurements. Cl int measurements were predicted using octanol–water and water–air partition coefficients that are readily available or can be quickly estimated for most chemicals. These two chemical properties were determined to be statistically significant predictors of Cl int , with >90% of predictions being within a factor of 10 of their measured value, and >60% being within a factor of 3. The resulting Cl int estimates, with appropriate confidence bounds, can be readily incorporated into a generic physiologically based pharmacokinetic model for a wide range of chemicals, which can be used to estimate steady-state blood concentrations resulting from environmental exposures. The level of precision associated with this approach (i.e., within an order of magnitude) allows a screening-level estimation of this important metabolic parameter for a wide range of chemicals.