Assessment of Algorithms for the Prediction of Metabolic Drug-Drug Interactions

Abstract

The aim of this work was to assess the ability of the static and dynamic (incorporating the time-course of the inhibitor) prediction models to predict drug-drug interactions (DDIs) using a population-based ADME simulator (Simcyp). This analysis focused on fluconazole, ketoconazole, itraconazole, fluoxetine and fluvoxamine, as CYP inhibitors. The rationale for their selection was an abundance of reported DDI studies, involving a wide range of victim drugs. Preliminary analysis focused on the individual victim drug and inhibitor parameters that are utilised in the DDI prediction models. The victim drug properties included in the DDI prediction models are calculated intrinsically in the Simcyp simulator from in vitro data; these values were compared to estimates obtained by different in vivo methods. Estimations of the fraction metabolised by CYP enzymes were generally consistent with 1500 (fluvoxamine from the media loss assay in human hepatocytes). No consistency was observed between methods and human or rat source for any of the inhibitors investigated; therefore, the inclusion of liver uptake into the prediction of DDIs for the current inhibitors was not supported.A database was collated from literature reports of DDIs involving the above named CYP inhibitors (n=97) and used to assess the inclusion of the time-course of inhibition into DDI prediction using the Simcyp simulator. In addition, the impact of active metabolites, dosing time and the ability to predict inter-individual variability in DDI magnitude were investigated using the dynamic prediction model. Simulations comprised of 10 trials with matching population demographics and dosage regimen to the in vivo studies. The predictive utility of the static and dynamic models was assessed relative to the inhibitor or victim drug investigated; both models were employed within Simcyp for consistency in parameters. Use of the dynamic and static models resulted in comparable prediction success, with 67 and 70% of DDIs predicted within two-fold, respectively. Over 60% of strong DDIs (>five-fold AUC increase) were under-predicted by both models, particularly for fluoxetine and fluvoxamine. Incorporation of the itraconazole metabolite into the dynamic model resulted in increased prediction accuracy of strong DDIs (80% within two-fold); no difference was observed for the inclusion of the fluoxetine metabolite. Predicted inter-individual variability in the DDI magnitude was also assessed in healthy, patient and genotyped subjects using a subset of clinical interactions (n=24). Mixed prediction success was observed and the importance of reliable clinical data was highlighted. The differences observed with the dose staggering and the incorporation of active metabolite highlight the importance of these variables in DDI prediction. Finally, the traditional 'two-fold limits' as a measure of the prediction success were reassessed, in particular at AUC ratios approaching one. New limits proposed are applicable for both inhibition and induction DDIs and allow incorporation of th

Date of Award	31 Dec 2011
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Aleksandra Galetin (Supervisor) & James Houston (Supervisor)