A method for estimating pharmacokinetic risks of concentration-dependent drug interactions from preclinical data

This article evaluates a novel approach for estimating the pharmacokinetic risks associated with drug interactions in populations. Preclinical pharmacokinetic and metabolism data are analyzed with a stochastic differential equation-based pharmacokinetic model that recognizes that the risks associated with known drug interactions involve deterministic and stochastic components. Specifically, a Bernoulli jump-diffusion pharmacokinetic model that accounts for the pharmacokinetics, the variability inherent in the pharmacokinetics, and the idiosyncratic nature of the possibility of drug interactions is proposed. In addition, the variability inherent in the extent of drug interaction is explicitly accounted for. The approach provides useful mechanistic insights into the stochastic processes that "drive" drug interactions in populations because it yields analytical results. The validity of the model predictions was tested with experimental data from two previously investigated systems: N-1 and N-3 caffeine demethylation in populations with smokers and in the terfenadine-ketoconazole system.