Chen Kaiyi, Du Yu, Zhu Yuxin
Global Statistical Science, Eli Lilly and Company, Indianapolis, Indiana, USA.
Armstrong Institute for Patient Safety and Quality, Johns Hopkins University, Baltimore, MD, USA.
Ther Innov Regul Sci. 2025 Jul 22. doi: 10.1007/s43441-025-00840-9.
The use of composite endpoint is a common strategy often employed to enhance statistical power and address the low incidence of individual outcomes, particularly in cardiovascular and kidney outcome studies. By merging multiple clinically relevant events into a single variable, these endpoints negate the need for multiple testing adjustments and augment the event rate, thus enabling studies of reasonable size and duration. However, as underscored by the FDA's guidance, a thorough evaluation of each component's impact is equally important to ensure the clinical relevance of these endpoints. This article delves into controversies surrounding the interpretation of hazard ratios derived from analyzing the composite endpoint and its individual components, exemplified by an observation from the CLEAR outcome trials. It highlights a paradoxical scenario where the combined treatment effect for the composite endpoint appeared less favorable than when assessing individual components separately. Moreover, we did a re-evaluation of the suitability of using Cox proportional hazards model in this context through theoretical investigation and simulation studies.
使用复合终点是一种常用策略，常用于增强统计效能并应对个体结局发生率较低的情况，尤其是在心血管和肾脏结局研究中。通过将多个临床相关事件合并为一个单一变量，这些终点无需进行多次检验调整，并提高了事件发生率，从而能够开展规模和时长合理的研究。然而，正如美国食品药品监督管理局（FDA）指南所强调的，对每个组成部分的影响进行全面评估同样重要，以确保这些终点的临床相关性。本文深入探讨了围绕分析复合终点及其各个组成部分得出的风险比解释的争议，以CLEAR结局试验的一项观察结果为例。它突出了一种自相矛盾的情况，即复合终点的联合治疗效果似乎比分项评估各个组成部分时更不理想。此外，我们通过理论研究和模拟研究，重新评估了在此背景下使用Cox比例风险模型的适用性。
