EMA对CYP抑制介导药物相互作用预测的资格认定意见:ICH M15时代PBPK可信度评估的启示
本文解读EMA对Simcyp Simulator(V19R1)预测CYP介导药物相互作用的资格认定意见,并将其映射至ICH M15可信度框架。资格认定数据显示,在三种使用背景下预测性能一致,绝对平均折叠误差为1.23至1.25,246项预测中98.8%落在观测值2倍范围内。文章提出整合资格认定证据、药物特异性验证、预测性能、不确定性表征和前瞻性应用的方法,以在ICH M15框架内建立PBPK支持监管决策的信心。
要点
- EMA资格认定覆盖CYP1A2、2C8、2C9、2C19、2D6、3A4/5的竞争性和机制性抑制,涉及victim和perpetrator药物及三种使用背景。
- 资格认定数据显示预测性能一致,绝对平均折叠误差1.23–1.25,98.8%的246项预测在观测值2倍内。
- 该意见早于ICH M15,未提供可信度概念映射或平台证据外推至具体研究药物的操作基础。
核心结论
- EMA对Simcyp Simulator(V19R1)的资格认定意见可映射至ICH M15可信度概念,并需结合药物特异性验证等整合方法将平台证据外推至具体药物。
研究设计与样本
- 本文为对EMA资格认定意见的解读与方法学论述,未报告原始研究设计或样本量。
主要发现
- 资格认定数据覆盖六种CYP酶、竞争性和机制性抑制、victim和perpetrator药物及三种使用背景。
- 平均折叠误差接近1,绝对平均折叠误差为1.23至1.25。
- 246项预测中98.8%落在观测值2倍范围内。
- 该意见早于ICH M15,未提供可信度概念映射或平台证据外推至具体研究药物的操作基础。
关键数值与推荐
- EMA资格认定覆盖CYP1A2、CYP2C8、CYP2C9、CYP2C19、CYP2D6、CYP3A4/5的竞争性和机制性抑制,涉及victim和perpetrator药物及三种使用背景。
- 资格认定数据显示绝对平均折叠误差为1.23至1.25,246项预测中98.8%在观测值2倍内。
临床怎么用
- 临床 pharmacology 和 pharmacometrics 团队可参考该映射方法,在ICH M15框架下整合平台资格认定证据与药物特异性验证,支持PBPK监管决策。
局限性
- 该资格认定意见早于ICH M15,未提供可信度概念映射或平台证据外推至具体研究药物的操作基础。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
Clinical pharmacology and pharmacometrics teams are now being encouraged to implement the ICH M15 framework in drug development to support model-informed decisions in regulatory submissions. Although model performance is an important component of model evaluation, regulatory confidence also depends on the intended context of use (CoU) and the supporting level of evidence. This is particularly relevant for physiologically based pharmacokinetic (PBPK) modeling, where increasing regulatory application has strengthened the need for structured approaches to evaluate predictive performance, extrapolation, uncertainty, and model applicability. The Simcyp Simulator (V19R1) was recently qualified by the EMA for prediction of CYP-mediated drug–drug interactions (DDIs) involving six enzymes (CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP3A4/5), competitive inhibition, and mechanism-based inhibition for both victim and perpetrator drugs across three CoUs. Qualification data demonstrated consistent predictive performance across the three CoUs, with average fold errors close to unity, absolute average fold errors ranging from 1.23 to 1.25, and 98.8% of 246 predictions falling within 2-fold of observed values. The opinion predates ICH M15; however, it does not provide either a mapping to its credibility concepts or an operational basis for extending platform-level evidence to a specific investigational drug. We address this gap with an explicit mapping of the qualification opinion onto ICH M15 credibility concepts and an integrated approach linking qualification evidence, drug-specific validation, predictive performance, uncertainty characterization, and prospective application, illustrating how confidence in PBPK-supported regulatory decisions can be established within an ICH M15-aligned framework.
引用
- Clinical Pharmacology & Therapeutics:The EMA Qualification Opinion for CYP Inhibition‐Mediated Drug–Drug Interaction Predictions: Implications for PBPK Credibility Assessment in the Era of ICH M15
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。