BackgroundShorter half-life glucagon-like peptide-1 receptor agonists (GLP-1 RAs) delay gastric emptying (DGE) more than GLP-1 RAs with longer half-lives. DGE is a known risk factor for gastro-oesophageal reflux disease (GERD) and its complications.AimTo determine whether short-acting or long-acting GLP-1 RAs are associated with an increased risk of new GERD or GERD-related complicationsDesignWe used the TriNetX global database to identify adult patients with type 2 diabetes mellitus and generated two cohorts totalling 1 543 351 patients on (1) GLP-1 RA or (2) other second-line diabetes medication. Using propensity-score matching, Kaplan-Meier Analysis and Cox-proportional hazards ratio (HR), we analysed outcomes and separately examined outcomes in patients starting short-acting (≤1 day) and long-acting (≥5 days) GLP-1 RAs.Results177 666 patients were in each propensity-matched cohort. GLP-1 RA exposure was associated with an increased risk (HR 1.15; 95% CI 1.09 to 1.22) of erosive reflux disease (ERD). However, this was solely due to short-acting (HR 1.215; 95% CI 1.111 to 1.328), but not long-acting (HR 0.994; 95% CI 0.924 to 1.069) GLP-1 RA exposure. Short-acting GLP-1 RAs were also associated with increased risk of oesophageal stricture (HR 1.284; 95% CI 1.135 to 1.453), Barrett’s without dysplasia (HR 1.372; 95% CI 1.217 to 1.546) and Barrett’s with dysplasia (HR 1.505; 95% CI 1.164 to 1.946) whereas long-acting GLP-1 RAs were not. This association persisted in sensitivity analyses, and when individually examining the short-acting GLP-1 RAs liraglutide, lixisenatide and exenatide.ConclusionStarting shorter-acting GLP-1 RAs is associated with increased risks of GERD and its complications.