Population modeling of the relationship between the pharmacokinetics of the oral thrombin inhibitor dabigatran etexilate and coagulation biomarkers in patients with non-valvular atrial fibrillation from the RE-LY trial
A. Schäftlein (1), H. Maas (2), P. A. Reilly (3), A. Staab (2) and T. Lehr (1)
(1) Saarland University, Germany, (2) Boehringer Ingelheim Pharma GmbH&Co. KG, Biberach, Germany, (3) Boehringer Ingelheim Pharmaceuticals, Inc. Ridgefield, CT, USA
Objectives: To develop population models describing the relationships between total dabigatran plasma concentrations and the Ecarin clotting time (ECT) as well as the Thrombin time (TT).
Methods: 27399 (ECT) and 8066 (TT) biomarker observation each with an associated dabigatran concentration from 9473 (ECT) and 2173 (TT) patients in the RE-LY trial [1] were included for data analysis in NONMEM 7.2. Time-dependent and independent linear as well as maximum effect (Emax) models were compared based on the results of the Likelihood-ratio test and graphical diagnostics. A covariate analysis was performed using a stepwise forward inclusion (p=0.05) and backward elimination (p=0.001) procedure. Data from two thirds of the patients were used for model development. The remaining observations were used evaluating the predictive performance of the final covariate models.
Results: Time-dependent models were not superior to time-independent models for both biomarkers. Hence, no delayed effect of dabigatran concentration on the coagulation pathway could be detected. A baseline dependent linear model for ECT as well as a combination of a baseline dependent linear and an Emax model for TT described the data best. Interindividual variability as well as intraindividual variability on model parameters was moderate for TT (< 37 %CV) and low for ECT (< 15 %CV). The concentration-ECT and concentration-TT relationships were statistically significant affected by sex and age (ECT), and by weight (TT) on the slope, respectively. Nevertheless, simulations revealed that all statistically significant covariates showed only minor effects on the interaction between dabigatran concentration and the measured biomarkers. Adequate predictive performance was confirmed for all developed models based on predictions for the evaluation set.
Conclusions: Two population models relating the dabigatran concentrations to the coagulation biomarkers ECT and TT were successfully developed. Despite the excellent descriptive and predictive performance of the models, the concentration biomarker relationship of anticoagulant drugs in general is dependent on the anticoagulant assay used [2]. This needs to be taken into account in applying the results from RE-LY to clinical practice.
References:
[1] Connolly, SJ et al. N Eng J Med (2009)
[2] Banfi, G et al . Semin Thromb Hemost (2009)