Julien Péron, Denis Maillet, and Benoit You, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, Pierre-Bénite; Julien Péron, Hospices Civils de Lyon; Julien Péron and Benoit You, Université de Lyon, Lyon; Julien Péron, Centre National de la Recherche Scientifique Unité Mixte de Recherche 5558, Laboratoire de Biométrie et Biologie Evolutive, Equipe Biostatistique-Santé, Villeurbanne; Benoit You, EMR UCBL/HCL 3738, Faculté de Médecine Lyon-Sud, Oullins, France; Hui K. Gan, Joint Austin-Ludwig Oncology Unit, Austin Hospital, Melbourne, Victoria, Australia; Eric X. Chen, Princess Margaret Hospital, University Health Network; and Eric X. Chen, University of Toronto, Toronto, Ontario, Canada.
J Clin Oncol. 2013 Nov 1;31(31):3957-63. doi: 10.1200/JCO.2013.49.3981. Epub 2013 Sep 23.
The Consolidated Standards of Reporting Trials (CONSORT) guidance was extended in 2004 to provide a set of 10 specific and comprehensive guidelines regarding adverse event (AE) reporting in randomized clinical trials (RCTs). Limited data exist regarding adherence to these guidelines in publications of oncology RCTs.
All phase III RCTs published between 2007 and 2011 were reviewed using a 16-point AE reporting quality score (AERQS) based on the 2004 CONSORT extension. Multivariable linear regression was used to identify features associated with improved reporting quality.
A total of 325 RCTs were reviewed. The mean AERQS was 10.1 on a 16-point scale. The most common items that were poorly reported were the methodology of AE collection (adequately reported in only 10% of studies), the description of AE characteristics leading to withdrawals (15%), and whether AEs are attributed to trial interventions (38%). Even when reported, the methods of AE collection and analysis were highly heterogeneous. The multivariable regression model revealed that industry funding, intercontinental trials, and trials in the metastatic setting were predictors of higher AERQS. The quality of AE reporting did not improve significantly over time and was not better among articles published in journals with a high impact factor.
Our findings show that some methodologic aspects of AE collection and analysis were poorly reported. Given the importance of AEs in evaluating new treatments, authors should be encouraged to adhere to the 2004 CONSORT guidelines regarding AE reporting.
2004 年，《临床试验报告统一标准》（CONSORT）指南进行了扩展，提供了一套关于随机临床试验（RCT）中不良事件（AE）报告的 10 项具体和全面的指南。关于这些指南在肿瘤 RCT 出版物中的遵循情况，相关数据有限。
使用基于 2004 年 CONSORT 扩展的 16 点不良事件报告质量评分（AERQS），对 2007 年至 2011 年期间发表的所有 III 期 RCT 进行了回顾。使用多变量线性回归来确定与改善报告质量相关的特征。
共审查了 325 项 RCT。AERQS 的平均得分为 16 分制的 10.1 分。报告最差的常见项目是 AE 收集方法（仅有 10%的研究充分报告）、导致退出的 AE 特征描述（15%）以及 AE 是否归因于试验干预（38%）。即使有报告，AE 收集和分析的方法也高度异质。多变量回归模型显示，工业资助、洲际试验和转移性疾病试验是 AERQS 较高的预测因素。AE 报告的质量并没有随着时间的推移而显著提高，在高影响因子期刊上发表的文章中也没有更好。
我们的研究结果表明，AE 收集和分析的某些方法学方面报告较差。鉴于 AE 在评估新治疗方法中的重要性，应鼓励作者遵守 2004 年 CONSORT 指南关于 AE 报告的规定。
