Clinical Pharmacology Asia, Bayer Yakuhin Ltd., Osaka, Japan.
Drug Metab Pharmacokinet. 2013;28(4):321-31. doi: 10.2133/dmpk.dmpk-12-rg-109. Epub 2013 Jan 22.
This study was designed to confirm the appropriateness of the dose setting for a Japanese phase III study of rivaroxaban in patients with non-valvular atrial fibrillation (NVAF), which had been based on model simulation employing phase II study data. The previously developed mixed-effects pharmacokinetic/pharmacodynamic (PK-PD) model, which consisted of an oral one-compartment model parameterized in terms of clearance, volume and a first-order absorption rate, was rebuilt and optimized using the data for 597 subjects from the Japanese phase III study, J-ROCKET AF. A mixed-effects modeling technique in NONMEM was used to quantify both unexplained inter-individual variability and inter-occasion variability, which are random effect parameters. The final PK and PK-PD models were evaluated to identify influential covariates. The empirical Bayes estimates of AUC and C(max) from the final PK model were consistent with the simulated results from the Japanese phase II study. There was no clear relationship between individual estimated exposures and safety-related events, and the estimated exposure levels were consistent with the global phase III data. Therefore, it was concluded that the dose selected for the phase III study with Japanese NVAF patients by means of model simulation employing phase II study data had been appropriate from the PK-PD perspective.
本研究旨在确认基于 II 期研究数据的模型模拟来设定非瓣膜性心房颤动（NVAF）日本 III 期研究 rivaroxaban 剂量的适宜性。先前开发的混合效应药代动力学/药效学（PK-PD）模型，由一个口服一室模型组成，根据清除率、体积和一级吸收速率进行参数化，使用来自日本 III 期研究 J-ROCKET AF 的 597 名受试者的数据进行了重建和优化。NONMEM 中的混合效应建模技术用于量化无法解释的个体间变异性和个体间变异性，这是随机效应参数。最后对 PK 和 PK-PD 模型进行评估以确定有影响的协变量。来自最终 PK 模型的 AUC 和 C（max）的经验贝叶斯估计值与来自日本 II 期研究的模拟结果一致。个体估计暴露量与安全性相关事件之间没有明确的关系，并且估计的暴露水平与全球 III 期数据一致。因此，从 PK-PD 的角度来看，通过使用 II 期研究数据的模型模拟选择日本 NVAF 患者的 III 期研究剂量是合适的。
