Placebo rates in randomized clinical trials of ulcerative colitis: an individual patient data meta-analysis

BACKGROUND AND AIMS: We assessed placebo rates and associated factors using individual patient data (IDP) from randomized clinical trials (RCTs) in ulcerative colitis (UC). METHODS: We conducted an IPD meta-analysis using Vivli and Yale University Open Data Access data-sharing platforms. Phase 2 and 3 RCTs of advanced biologics in adults with moderate-to-severe UC published since 2010 were included. Pooled placebo rates and 95% CIs were estimated using one- and two-stage meta-analytical approaches. Significant patient-level factors (P < .05) were identified using regression analyses. Primary outcomes were clinical response and remission. RESULTS: Data were available for 1703 patients from nine studies. For induction trials, overall placebo response and remission rates were 33% (95% CI 29%-38%) and 9% (95% CI 7%-11%). Overall placebo response and remission rates in maintenance trials were 28% (95% CI 18%-41%) and 14% (95% CI 9%-20%). A lower body mass index reduced the odds of placebo response and remission, while higher baseline albumin levels and left-sided (compared to extensive) UC increased the odds of these outcomes. A 1-point increase in the Mayo Clinic Score (MCS) and adapted MCS was associated with a 26% and 27% reduction in odds of clinical remission. For induction trials, prior biologic exposure was associated with lower odds of response and remission. Multicenter trials have lower placebo effects than single-center trials. CONCLUSIONS: These results enable future trials to incorporate design elements that reduce placebo rates as well as a precise benchmark for expected rates in clinical trials that do not include placebo.