Furukawa Toshi A, Cipriani Andrea, Atkinson Lauren Z, Leucht Stefan, Ogawa Yusuke, Takeshima Nozomi, Hayasaka Yu, Chaimani Anna, Salanti Georgia
Department of Health Promotion and Human Behavior, Kyoto University Graduate School of Medicine/School of Public Health, Kyoto, Japan; Department of Clinical Epidemiology, Kyoto University Graduate School of Medicine/School of Public Health, Kyoto, Japan.
Department of Psychiatry, University of Oxford, Oxford, UK.
Lancet Psychiatry. 2016 Nov;3(11):1059-1066. doi: 10.1016/S2215-0366(16)30307-8. Epub 2016 Oct 7.
Previous studies have shown that placebo response rates in antidepressant trials have been increasing since the 1970s. However, these studies have been based on outdated or limited datasets and have used inappropriate statistical methods. We did a systematic review of placebo-controlled randomised controlled trials of antidepressants to examine associations between placebo-response rates and study and patient characteristics.
In this systematic review, we searched for published and unpublished double-blind randomised placebo-controlled trials of first-generation and second-generation antidepressants for acute treatment of major depression in adults (update: Jan 8, 2016). The log-transformed proportions of placebo response, defined as 50% or greater reduction in depression severity score from baseline, were meta-analytically synthesised for each year. We then looked for a structural break point in the secular changes in these characteristics through the years and examined the influence of the study year and other trial and patient characteristics on the response rates through meta-regression.
We identified 252 placebo-controlled trials (26 324 patients on placebo) done between 1978 and 2015. There was a structural break in 1991, and since then, the average placebo response rates in antidepressant trials have remained constant in the range between 35% and 40% (relative risk [RR] 1·00, 95% CI 0·97-1·03, p=0·99, for every 5-year increase). The length of the study and the number of study centres were significant factors (RR 1·03, 95% CI 1·01-1·05 for 1 more week in trial length; 1·32, 1·11-1·57 for multicentre vs single-centre trials).
Contrary to the widely held belief, the average placebo response rates in antidepressant trials have been stable for more than 25 years. This new evidence should have an effect on the interpretation of the scientific literature and the future of psychopharmacology, both from a clinical and methodological point of view.
Japan Society for Promotion of Science, Great Britain Sasakawa Foundation.
既往研究表明，自20世纪70年代以来，抗抑郁药物试验中的安慰剂反应率一直在上升。然而，这些研究基于过时或有限的数据集，并使用了不恰当的统计方法。我们对安慰剂对照的抗抑郁药物随机对照试验进行了系统评价，以研究安慰剂反应率与研究及患者特征之间的关联。
在这项系统评价中，我们检索了已发表和未发表的第一代和第二代抗抑郁药物用于成人重度抑郁症急性治疗的双盲随机安慰剂对照试验（更新日期：2016年1月8日）。对每年安慰剂反应的对数转换比例进行荟萃分析综合，安慰剂反应定义为抑郁严重程度评分较基线降低50%或更多。然后，我们通过多年来这些特征的长期变化寻找结构断点，并通过荟萃回归研究研究年份以及其他试验和患者特征对反应率的影响。
我们纳入了1978年至2015年期间进行的252项安慰剂对照试验（26324例服用安慰剂的患者）。1991年出现了一个结构断点，自那时起，抗抑郁药物试验中的平均安慰剂反应率一直稳定在35%至40%之间（每增加5年，相对风险[RR]为1.00，95%置信区间为0.97 - 1.03，p = 0.99）。研究时长和研究中心数量是显著因素（试验时长每增加1周，RR为1.03，95%置信区间为1.01 - 1.05；多中心试验与单中心试验相比，RR为1.32，95%置信区间为1.11 - 1.57）。
与广泛持有的观点相反，抗抑郁药物试验中的平均安慰剂反应率在25年多的时间里一直保持稳定。从临床和方法学角度来看，这一新证据应会对科学文献的解读以及精神药理学的未来产生影响。
日本学术振兴会、英国笹川基金会。
