Use of a physiology-based whole-body population model to simulate the influence of anthropometric-, clearance-, and physiological-variance on the pharmacokinetics of drugs
Edginton, A.N. (1), S. Willmann (1), M. Kleine-Besten (2)
(1) Competence Center Systems Biology, Bayer Technology Services GmbH, 51368 Leverkusen, Germany; (2) Institute of Pharmaceutical Technology, J.W. Goethe University, Frankfurt am Main, Germany
Objective: The objective of this study was to determine the accuracy of a physiology-based population pharmacokinetic model to predict interindividual variability by comparing simulated pharmacokinetic profiles with experimental data. The model was previously developed using a population algorithm to simulate interindividual pharmacokinetic variability using known distributions of anatomical and physiological properties of individuals.
Methods: The predictive population model, PK-Pop, as incorporated into the PK-Sim® software (Bayer Technology Services, Leverkusen, Germany), was used for all simulations. Simulations were compared to human experimental data using the compounds paclitaxel, cimitedine and ciprofloxacin. Populations were generated in PK-Pop based on the age, weight and height range of each experimental group. The addition of clearance variability and the variability of fraction unbound in the case of the specific paclitaxel formulation and gastrointestinal parameters in the case of oral cimitedine were included and the relative contributions of these parameters to interindividual variability were assessed.
Results: Both the mean and the variability of the observed plasma concentrations were well represented by the generated population concentration time curves. The addition of clearance variability only slightly increased the geometric standard deviation in comparison to a simulated population all having identical intrinsic clearances on a per kilogram organ weight basis. The addition of variability in the unbound fraction of paclitaxel was important for describing the observed variability at Cmax. For cimitedine, gastric emptying time and intestinal transit time significantly influenced oral absorption of a slow release formulation and was not important for the accurate prediction of an immediate release formulation.
Discussion: The PK-Pop algorithm creates virtual individuals that are reasonable representatives of real humans. The comparison of pharmacokinetic population simulation results for ciprofloxacin, paclitaxel and cimitedine to experimentally observed plasma concentration time profiles demonstrates that the population model is able to simulate the pharmacokinetic variability of these three drugs.