Clinical, Endoscopic, and Safety Placebo Rates in Induction and Maintenance Trials of Crohn's Disease: Meta-Analysis of Randomised Controlled Trials

Background Precision in estimating placebo rates is important for clinical trial design. Aim To quantify placebo rates across relevant endpoints in Crohn's disease [CD] trials and identify the factors influencing these rates in a contemporary meta-analysis. Methods We searched MEDLINE, EMBASE, and CENTRAL from inception to March 2021. Eligible studies were placebo-controlled trials of pharmacological interventions for CD. Placebo response and remission rates for induction and maintenance trials were extracted and pooled by random-effects to quantify placebo rates across studies. Mixed-effects meta-regression was used to evaluate the effects of study-level characteristics on placebo rates. Results In 125 studies [91 induction, 46 maintenance], placebo clinical remission and response rates for induction studies were 18% (95% confidence interval [CI] 16, 21%], and 32% [95% CI 29, 35%], respectively, and for maintenance studies were 28% [95% CI 23, 34%] and 30% [95% CI 24, 37%], respectively. Endoscopic remission and response rates in induction studies were 8% [95% CI 4, 18%] and 16% [95% CI 11, 23%], respectively. Trials enrolling patients with prior biologic exposure, longer disease duration, and higher CD activity index scores were associated with lower placebo clinical remission rates. Increased duration of follow-up, more follow-up visits, and a greater proportion of patients with colonic disease distribution were associated with higher clinical placebo rates. Conclusions Placebo remission and response rates in CD trials vary according to the phase of the trial, endpoint assessed, and induction or maintenance design. These contemporary estimates will help to inform future CD trial design.

Automated summary

This study quantifies the placebo rates in Crohn's disease trials and identifies factors influencing these rates. The findings suggest that placebo remission and response rates vary based on the trial phase, endpoint assessed, and induction or maintenance design.