Modelling and Simulation of Ketoconazole inhibition when co-administration time is not sufficient: role of CYP3A function recovery
Gianluca Nucci, Italo Poggesi and Roberto Gomeni
Clinical Pharmacokinetics /Modelling & Simulation
Introduction: In this study ketoconazole (keto), a potent CYP3A inhibitor, was co-administered with GSK drug X (X) to assess the extent of interaction as well as the relevance of CYP3A (3A) on X metabolism. Given the prolonged elimination of X (t1/2 ~ 40-60 hours), the study was amended part-way to increase the duration of keto dosing so that two groups with different keto durations were examined (200 mg BID from day -4 to day 4 in group 1 and from day -4 to day 10 in group 2). Given the maximum recommended duration of keto dosing, it was not possible to fully assess the extent of inhibition with standard methods, especially considering that the decline of X plasma concentration was markedly different after stopping keto administration, returning back to the non-inhibited values. This lead to the paradoxical observation of higher AUClast ratios than AUCinf ratios.
Objectives: The objective of this study was to model keto-X interaction data to estimate the extent of the full metabolic inhibition and to simulate X profile with different keto regimens.
Methods: Methods: A two-compartment model with first-order absorption and lag-time was used for X . It was fitted to the non-inhibited state data of 36 healthy volunteers (20 in group 1 and 16 in group 2) using NONMEM VI. Individual covariate used to refine the model estimates were age, weight and sex. The inhibition phase was then introduced using the time of keto co-administration as covariate. Different keto inhibition structures were employed and the best one (based on GOF considerations) was retained.
Results: The best base model retained a correlation between inter-individual variability in CL and central distribution volume. In the best inhibition model keto was found to both decrease CL and increase first pass of X. As expected, subjects with higher baseline clearance had higher inhibition (modelled as a correlation between non-inhibited CL and extent of inhibition). The time course of 3A recovery was introduced as an exponential function of time after stopping keto.
Conclusions: The model provided good fit of the observed data and enabled to assess individual and population inhibition results. Therefore, using simulations, it was possible to evaluate the full extent of 3A involvement in drug X metabolism as well as the influence of different durations of keto co-administration.
