Contopoulos-Ioannidis Despina, Tseretopoulou Xanthippi, Ancker Megan, Walterspiel Juan N, Panagiotou Orestis A, Maldonado Yvonne, Ioannidis John P A
Department of Pediatrics, Division of Infectious Diseases, Stanford University School of Medicine, 300 Pasteur Drive, Stanford, CA, 94305-5107, USA; Palo Alto Medical Foundation Research Institute, 795 El Camino Real, Ames Building, Room 2A027B, Palo Alto, CA, 94301, USA; Meta-Research Innovation Center at Stanford (METRICS), 1070 Arastradero Road, Palo Alto, CA, 94304, USA.
Leeds Teaching Hospital, NHS Trust, Great George Street, Leeds LS1 3EX, UK.
J Clin Epidemiol. 2016 Oct;78:10-21. doi: 10.1016/j.jclinepi.2016.02.032. Epub 2016 Apr 6.
We set up to evaluate the relative risk of harms in trials performed in less developed vs. more developed countries.
Meta-epidemiologic evaluation using the Cochrane Database of Systematic Reviews. We considered meta-analyses with at least one randomized clinical trial (RCT) in a less developed country and one RCT in a more developed country. We targeted severe adverse events (AEs), discontinuations due to AEs, any AE, organ system-specific AEs, individual AEs, and all discontinuations due to any reason. We estimated the relative odds ratio (ROR) of harms between more and less developed countries for each topic and the summary ROR (sROR) across topics under each category of harms.
We identified 42 systematic reviews (128 meta-analyses, 521 independent RCTs). Summary sRORs did not differ significantly from 1.00 for any harm category. Nominally significant RORs were found in only 6/128 meta-analyses. However, in 27% (35/128) of meta-analyses the ROR point estimates indicated relative differences between country settings >2-fold. Considering also ROR 95% confidence intervals, in 92% (118/128) of meta-analyses one could not exclude a 2-fold difference in both directions.
We identified limited comparative evidence on harms in trials from these two country settings. Substantial differences in the risk point estimates were common; the potential for modest differences could rarely be excluded with confidence.
我们着手评估在欠发达国家与发达国家进行的试验中伤害的相对风险。
使用Cochrane系统评价数据库进行的Meta流行病学评估。我们纳入了在欠发达国家至少有一项随机临床试验（RCT）且在发达国家至少有一项RCT的Meta分析。我们将严重不良事件（AE）、因AE导致的停药、任何AE、器官系统特异性AE、个体AE以及因任何原因导致的所有停药作为研究对象。我们估计了每个主题下发达国家与欠发达国家之间伤害的相对比值比（ROR）以及各伤害类别下所有主题的汇总ROR（sROR）。
我们识别出42项系统评价（128项Meta分析，521项独立RCT）。任何伤害类别的汇总sROR与1.00相比均无显著差异。仅在6/128项Meta分析中发现名义上显著的ROR。然而，在27%（35/128）的Meta分析中，ROR点估计表明国家背景之间的相对差异>2倍。同时考虑ROR的95%置信区间，在92%（118/128）的Meta分析中，无法排除两个方向上2倍的差异。
我们发现这两种国家背景下试验中伤害的比较证据有限。风险点估计存在实质性差异很常见；很少能有把握地排除适度差异的可能性。
