Xuejun Chen, Al Corey, Suresh Mallikaarjun, Steve Bramer
Otsuka Maryland Research Institute
Purpose: QTc prolongation has been extensively investigated as a predictor of cardiac toxicity, and has been shown to correlate with plasma concentrations for some drugs. Exploratory analyses may indicate trends in QTc prolongation as a function of plasma concentration but cannot account for covariate effects, high inter-individual variability and prediction of this relationship for new doses. The purpose of this study is to investigate the relationship between the pharmacokinetics and QTc prolongation, using population-based techniques.
Methods: Data from 2 phase III studies were used for a population pharmacodynamic analysis using NONMEM. Based on the lack of hysteresis, and since the relationship between QTc and concentration appeared to be stationary, it was decided that the QTc – concentration relationship would be analyzed directly. The types of models to be evaluated were limited to those that reflect direct action (Emax type models, simple linear relationships). Different models and covariate effects were tested and compared in terms of the objective function and diagnostic plots.
Results: The best model, Sigmoid Emax with a Circadian Baseline, provided adequate fit to the pooled data. Baseline effect parameters (placebo dose) are consistent with literature. Baseline QTc and gender have effect on baseline effect parameters. The typical values (% Relative standard error) for maximum possible effect (Emax, msec), concentration required to attain 50% of Emax (EC50, µg/mL), exponent describing steepness of the effect-concentration curve (γ), Midpoint of the range of circadian fluctuation in QTc (Eavg, msec), amplitude of the circadian fluctuation in QTc, relative to Eavg (Amp, msec) and time of the maximum circadian deviation from Eavg (tcircad,h) are 61.5 (23%), 4.72 (28%), 1.3 (8%), 381 (0.3%), 4.28 (9%) and 8.09 (3%), respectively. The value for Eavg was predicted by a combination of the subject’s first QTc measurement on study (BSF) and the subject’s gender. Inter-subject variability was characterized for the key parameters except EC50 and was estimated to be approximately 27, 25, 2, 54 and 3 %CV for Emax, γ, Eavg, Amp and tcircad , respectively.
Conclusions: The model developed is suitable for the PK/PD analysis of QTc prolongation. It accounts for covariate effects and inter-individual variability, which cannot be explained by a noncompartmental analysis.
Reference: PAGE 13 (2004) Abstr 505 [www.page-meeting.org/?abstract=505]
Poster: poster