Factors Affecting Clinical and Endoscopic Outcomes of Placebo Arm in Trials of Biologics and Small Molecule Drugs in Ulcerative Colitis: A Meta-Analysis

Background: Previous systematic cumulative analyses of the placebo arm in ulcerative colitis (UC) were limited by the wide heterogeneity in the scores and definitions of response. We aimed at estimating the placebo rates of remission, response, and mucosal healing (MH) in phase 2 and 3 randomized placebo-controlled trials of biologics and small molecule drugs that used homogeneous criteria for the assessment of outcomes. Methods: PubMed Central, Embase, and reference lists of articles were systematically searched through July 2017. Only studies that employed the Mayo score were included. Results: Thirty-one randomized placebo-controlled trials consisting of 2702 patients met the inclusion criteria. At induction, the pooled estimates of the placebo rates of remission, response, and MH were 9% (95% confidence interval [CI], 7%-12%; range: 0%-29%; I-2 = 75.0%), 34% (95% CI, 31%-38%; range: 12%-75%; I-2 = 61.3%), and 26% (95% CI, 22%-30%; range: 2%-65%; I-2 = 77.7%), respectively. At maintenance, the pooled estimates of the placebo rates of remission, response, and MH were 14% (95% CI, 10%-18%; range: 6%-30%; I-2 = 73.0%), 23% (95% CI, 20%-27%; range: 18%-36%; I-2 = 53.0%), and 19% (95% CI, 15%-23%; range: 12%-30%; I-2 = 65.0%), respectively. Among the variables assessed by logistic regression analysis, multiple factors influenced the outcomes of placebo arms, including concomitant systemic steroids at baseline, endoscopic central reading, being naive or non-naive to anti-TNFs, and disease duration. Conclusions: Despite the wide use of homogeneous criteria for the assessment of clinical and endoscopic outcomes, a high heterogeneity among placebo arms of modern trials in UC still exists.