文献速递流行病学与生物统计临床试验与 GCP

药物干预的活性安慰剂对照组很少使用但值得认真考虑:一项方法学概览

Active placebo control groups of pharmacological interventions were rarely used but merited serious consideration: a methodological overview.
PubMed 文献 · J Clin Epidemiol · 2017-03-22

该概览由三个主题关联的子研究组成,评估随机药物试验中活性安慰剂对照的使用情况。基于2013年10月发表的200项安慰剂对照随机药物试验随机样本,活性安慰剂组患病率为1/200(95%CI:0-2),即0.5%(0-1%)。共识别并特征化了89项使用活性安慰剂的随机试验(1961-2014年发表),并分析了25篇方法学文献;作者建议在副作用明显、预期疗效 modest 且揭盲偏倚风险高的试验中更多使用活性安慰剂。

要点

  • 2013年发表的200项安慰剂对照随机药物试验中仅1项使用活性安慰剂组,患病率0.5%(95%CI:0-1)。
  • 共识别89项使用活性安慰剂的随机试验(1961-2014年),常用抗组胺药、抗胆碱能药和镇静剂。
  • 作者建议在副作用明显、预期疗效 modest 且揭盲偏倚风险高的试验中更多使用活性安慰剂。

核心结论

  • 药物活性安慰剂对照干预在随机临床试验中很少使用,但构成一种值得认真考虑的方法学工具。
  • 作者建议在副作用明显的药物试验中更常使用活性安慰剂,尤其是在预期治疗效应 modest 且因揭盲导致偏倚风险高的情况下。

研究设计与样本

  • 本概览由三个主题关联的子研究组成:一项观察性子研究基于2013年10月发表的200项PubMed索引的安慰剂对照随机药物试验随机样本评估活性安慰剂组的患病率。
  • 一项系统综述识别并特征化了使用活性安慰剂对照组的试验,不论发表时间。
  • 第三项子研究回顾了关于活性安慰剂组有实质性方法学评论的出版物,检索PubMed、Cochrane图书馆、Google Scholar和HighWirePress。

主要发现

  • 2013年发表的200项试验中,使用活性安慰剂组的患病率为1/200(95%置信区间:0-2),即0.5%(0-1%)。
  • 识别并特征化了89项使用活性安慰剂的随机试验(1961-2014年发表),例如抗组胺药、抗胆碱能药物和镇静剂。
  • 此类试验通常涉及交叉设计,实验干预具有明显副作用,且结局为患者报告。
  • 活性安慰剂的使用集中在特定研究环境,且似乎并不一致反映实验干预的副作用特征,例如选择性5-羟色胺再摄取抑制剂在疼痛试验中与活性安慰剂比较,但在抑郁症试验中未如此。
  • 识别并分析了25篇有实质性评论的方法学出版物。

关键数值与推荐

  • 支持活性安慰剂的主要论点是降低揭盲风险;反对的主要论点是意外治疗效应的风险。

临床怎么用

  • 作者建议在副作用明显的药物试验中更常使用活性安慰剂,尤其是在预期治疗效应 modest 且因揭盲导致偏倚风险高的情况下。

局限性

  • 原文未报告局限性。

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

OBJECTIVES: Active placebos are control interventions that mimic the side effects of the experimental interventions in randomized trials and are sometimes used to reduce the risk of unblinding. We wanted to assess how often randomized clinical drug trials use active placebo control groups; to provide a catalog, and a characterization, of such trials; and to analyze methodological arguments for and against the use of active placebo. STUDY DESIGN AND SETTING: An overview consisting of three thematically linked substudies. In an observational substudy, we assessed the prevalence of active placebo groups based on a random sample of 200 PubMed indexed placebo-controlled randomized drug trials published in October 2013. In a systematic review, we identified and characterized trials with active placebo control groups irrespective of publication time. In a third substudy, we reviewed publications with substantial methodological comments on active placebo groups (searches in PubMed, The Cochrane Library, Google Scholar, and HighWirePress). RESULTS: The prevalence of trials with active placebo groups published in 2013 was 1 out of 200 (95% confidence interval: 0-2), 0.5% (0-1%). We identified and characterized 89 randomized trials (published 1961-2014) using active placebos, for example, antihistamines, anticholinergic drugs, and sedatives. Such trials typically involved a crossover design, the experimental intervention had noticeable side effects, and the outcomes were patient-reported. The use of active placebos was clustered in specific research settings and did not appear to reflect consistently the side effect profile of the experimental intervention, for example, selective serotonin reuptake inhibitors were compared with active placebos in pain trials but not in depression trials. We identified and analyzed 25 methods publications with substantial comments. The main argument for active placebo was to reduce risk of unblinding; the main argument against was the risk of unintended therapeutic effect. CONCLUSION: Pharmacological active placebo control interventions are rarely used in randomized clinical trials, but they constitute a methodological tool which merits serious consideration. We suggest that active placebos are used more often in trials of drugs with noticeable side effects, especially in situations where the expected therapeutic effects are modest and the risk of bias due to unblinding is high.

来源:J Clin Epidemiol,查看原文。版权归原作者所有。

引用

  1. Jensen JS, Bielefeldt AØ, Hróbjartsson A, et al. Active placebo control groups of pharmacological interventions were rarely used but merited serious consideration: a methodological overview. J Clin Epidemiol. 2017. PMID: 28342907. doi:10.1016/j.jclinepi.2017.03.001

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。