“…For the sensitivity analyses, studies were excluded from the analysis as follows: one study 32 due to a high risk of bias from attrition, seven studies 35,38,39,45,48,65,68 that calculated healing rates based on PPS, and two studies 50,62 with only life table (or Kaplan-Meier) estimated healing rates. Only 20 studies (22 trials) 14,16,30,31,35,[37][38][39][40]42,43,[49][50][51]62,64,66,[68][69][70] used the Los Angeles scale for endoscopic assessment and were eligible for this analysis. The results (both estimates and orders) of these analyses were consistent with the main analysis, showing its robustness (Table S3).…”