Abstract
BACKGROUND: High placebo responses have limited drug development in eosinophilic oesophagitis. The optimal configuration of trial outcomes is uncertain. AIMS: To inform more efficient future trial designs, to characterise clinical, endoscopic and histologic placebo responses in eosinophilic oesophagitis randomised controlled trials (RCTs). METHODS: We updated a Cochrane systematic review and meta-analysis, searching multiple databases to January 1, 2024, to identify placebo-controlled RCTs evaluating medical therapies for patients with eosinophilic oesophagitis. The primary outcome was the pooled proportion of study-defined clinical, endoscopic and histologic responders and remitters randomised to placebo, using an intention-to-treat approach and random-effects model. Sources of heterogeneity were explored using meta-regression. RESULTS: We included 25 RCTs. The pooled proportion of clinical response was 41.0% [95% CI: 29.7%-52.8%] with substantial heterogeneity (I(2) = 74.9%). On meta-regression, older age and a higher probability of being randomised to placebo reduced the likelihood of clinical response to placebo. The pooled proportion of histologic remission defined as a peak eosinophil count [PEC] ≤ 6 eosinophils per high power field [HPF] or ≤ 1 eosinophil/HPF was 4.3% [95% CI: 2.6%-6.2%] (I(2) = 23.6%) and 1.3% [95% CI: 0.5%-2.5%] (I(2) = 0%), respectively. The standardised mean difference in the Eosinophilic Oesophagitis Endoscopic Reference Score to placebo was -0.25 [95% CI: -0.41, -0.10]. CONCLUSIONS: Over 40% of patients in eosinophilic oesophagitis trials respond clinically to placebo, and this is associated with trial design factors such as randomisation ratio and trial population. Objective endoscopic and histologic measures are associated with very low placebo responses.