Incorporating Monte Carlo simulation into physiologically based pharmacokinetic models using advanced continuous simulation language (ACSL): a computational method.

Abstract

Biologically based models with physiological parameters are becoming more popular as a tool to estimate target tissue doses from chemical exposures. However, the majority of current physiologically based pharmacokinetic (PBPK) models do not take into account the uncertainty and/or variability within the various model parameters. Consideration of uncertainty is important to evaluate the predictive ability and complexity of a model as well as identification of parameters which contribute disproportionately to variability in model output. In order to estimate the uncertainty in PBPK model output, a versatile and simple computational method is presented which can be readily incorporated into the majority of PBPK models without extensive additions to model computer code. In this paper, a separate computer program for Monte Carlo simulation is furnished that randomly samples values for model parameters and writes them into a run-time language (command file) format which can then be utilized to execute individual PBPK models. Modifications to the PBPK model allow the desired output to be written to a data file for statistical analysis. The method presented in this paper is applied to a simple PBPK model for benzene disposition.