Predictors of Clinical Remission to Placebo in Clinical Trials of Crohn's Disease.

Background: In placebo-controlled clinical trials for Crohn's disease (CD), some placebo-treated patients demonstrate improvement. However, it is unclear what factors contribute to placebo response and remission.

Methods: This was a post hoc analysis of 3 placebo-controlled clinical trial programs (GEMINI-2, UNITI-1/2, and CLASSIC-1) of moderate-severe CD evaluating the efficacy of vedolizumab, ustekinumab, and adalimumab. Baseline predictors of clinical remission at the end of induction (week 4/6), defined as Crohn's Disease Activity Index <150 were evaluated among placebo-treated patients. Clinical response (decrease in Crohn's Disease Activity Index ≥100 points from baseline) at the end of induction was also evaluated. Univariate analyses were performed and predictors with P < .10 were included in multivariable analyses.

Results: A total of 683 patients (148 from GEMINI-2, 470 from UNITI-1/2, and 65 from CLASSIC-1) treated with placebo were included. Of the predictors evaluated, C-reactive protein <5 mg/L (odds ratio [OR], 1.66; 95% confidence interval [CI], 1.04-2.67; P = .035), albumin >40 g/L (OR, 1.57; 95% CI, 1.05-2.93; P = .023), and disease duration <5 years (OR, 1.70; 95% CI, 1.05-2.75; P = .032) were included in the multivariable model for clinical remission. Disease duration <5 years was the only variable that retained significance on multivariable analysis (adjusted OR, 1.67; 95% CI, 1.02-2.73; P = .040). For clinical response, isolated ileal disease and disease duration <1 year were included in the multivariable model, of which only the latter retained significance (adjusted OR, 1.84; 95% CI, 1.04-3.24; P = .035).

Conclusions: Strategies that reduce placebo response rates in clinical trials of CD should be considered, including stratification or exclusion of subjects by disease duration and mild disease severity as measured by objective biomarkers.