A meta-analysis of the placebo rates of remission and response in clinical trials of active Crohn's disease

BACKGROUND & AIMS: Placebo-controlled, randomized clinical trials (PC-RCTs) are commonly used to assess therapies for Crohn's disease (CD). Knowledge of the placebo rates of remission and response and understanding of design factors that influence these rates is important for designing future clinical trials evaluating pharmacotherapy of CD. The aims of this study were to estimate rates of remission and response in patients with active CD receiving placebo and to identify factors influencing these rates. METHODS: We performed a systematic review and meta-analysis of PC-RCTs evaluating therapies for active CD identified from MEDLINE from 1966 to 2001. RESULTS: The pooled estimates of the placebo rates of remission and response were 18% (95% confidence interval, 14%-24%; range, 0%-50%) and 19% (95% confidence interval, 13%-28%; range, 0%-46%), respectively, both with significant heterogeneity among studies (P < 0.01 for remission, P < 0.03 for response). In multivariate models, study duration, number of study visits, and entry Crohn's Disease Activity Index score were important predictors of the placebo remission rate, with study duration the most important. However, no single factor could account for all of the heterogeneity. Factors that influence the placebo response rates were similar to those affecting the placebo remission rates. The absolute benefit of active treatment beyond placebo was generally larger when outcome was measured by response than remission. CONCLUSIONS: Placebo remission and response rates in PC-RCTs for active CD are variable. Study duration, number of study visits, and disease severity at entry have a large influence on placebo remission rates.