A pharmacovigilance study published online in Gut and Liver on 2 September pooled the national adverse-event databases of the United States, Japan and Canada and found gastroesophageal reflux disease reported roughly two to five times more often for GLP-1 drugs than for a comparator diabetes drug class. The signal held in all three countries. It is a reporting signal, not a rate: these databases cannot say how many people taking a GLP-1 develop reflux, and obesity itself raises reflux risk — a confounder the authors put first among their own limitations.
What the three databases show
The analysis covers 260,417 GLP-1 receptor agonist reports: 252,664 in the FDA Adverse Event Reporting System (2020 Q1 to 2025 Q4), 2,881 in Japan’s JADER (2004 Q1 to 2024 Q4) and 4,872 in Canada Vigilance (from the database’s 1965 inception to 2024). The comparator was DPP-4 inhibitors, another diabetes drug class — an active comparator rather than an everything-else denominator, which is the stronger choice.
The measure is the reporting odds ratio: how much more often reflux turns up among one drug’s reports than among the comparator’s. In FAERS it was 2.83 (95% CI 2.39 to 3.35), from 9,245 reflux reports against 139. Canada Vigilance gave 2.91 (1.97 to 4.31), from 206 against 29. JADER gave the largest number, 4.76, but on 19 reports against 16, and its interval runs from 2.45 to 9.27 — nearly four times as wide as the American one, which is what 19 reports buys you.
Within FAERS the drug-level results were semaglutide 3.83 (3.23 to 4.55), liraglutide 3.16 (2.58 to 3.87), tirzepatide 2.76 (2.33 to 3.27) and dulaglutide 2.35 (1.97 to 2.80). Exenatide ran the other way at 0.60 (0.46 to 0.77), reported less often than the comparator. Signals strengthened with age: 2.10 (0.67 to 6.59) under 40, an interval that crosses 1 and settles nothing; 2.89 (1.99 to 4.20) at 40 to 64; and 3.01 (2.33 to 3.89) at 65 and over, with a trend p-value of 0.028.
What it adds, and what it cannot
Reflux on these drugs is not new information. It is already on Wegovy’s label, which lists gastroesophageal reflux disease among the reactions occurring at higher incidence in treated adults than on placebo, with the two arms two percentage points apart in the adverse-reaction table. The same label notes in its warnings that the drug “delays gastric emptying” — the pharmacology that makes a reflux signal unsurprising. We cover the wider side-effect picture on our page on semaglutide side effects.
What this study adds is consistency. The same signal shows up in three separately collected national datasets, on two continents, against an active comparator, and it survives the authors’ sensitivity analyses. That is a stronger footing than any single database analysis, and it is the reason a fourth or fifth FAERS-only paper on the same question would not have been worth reporting.
What it cannot give you is a rate. A reporting odds ratio is not an incidence, a relative risk or a probability; these systems hold reports, not patients, and nobody knows the denominator of people taking each drug. The authors name confounding by indication as the central limit: obesity is itself a risk factor for reflux, and spontaneous-reporting databases carry no BMI, so it cannot be adjusted away. Diagnoses are not clinically adjudicated, concomitant medications are not controlled for, and semaglutide’s public profile plausibly inflates reporting about semaglutide. Exenatide’s inverse result is a useful check on how much of this is biology and how much is reporting behaviour. The practical reading is narrow: reflux that starts or worsens after beginning a GLP-1 is plausibly related and worth raising with the prescriber, and is not a reason to stop on your own.
Sources: Jeong-Gyu Choi, Eun Jeong Gong, Chang Seok Bang and Jae Jun Lee, “Multi-Database Pharmacovigilance Analysis of Gastroesophageal Reflux Disease Associated with GLP-1 Receptor Agonists: A Cross-National Signal Validation Study”, Gut and Liver, published online 2 September 2026 (abstract also on PubMed). Funded by Korea’s National Research Foundation and a Hallym University research fund; the authors reported no conflicts of interest. Label detail from Wegovy prescribing information on DailyMed.
