Semaglutide and tirzepatide linked to fewer alcohol-related hospital visits

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Jul 24 2026

A large US health-record study suggests newer GLP-1 drugs may reduce alcohol-related acute-care visits, but the findings remain observational and need confirmation in clinical trials.

In a recent study published in the journal BMJ Open, a group of researchers evaluated whether newer Glucagon-like Peptide-1 Receptor Agonists (GLP-1 RAs), including semaglutide and tirzepatide, are associated with a lower risk of alcohol-related emergency department (ED) visits or hospitalizations among adults with Alcohol Use Disorder (AUD) and Type 2 Diabetes (T2D) or obesity.

Background

Excessive alcohol consumption remains one of the most significant preventable causes of diseases and deaths in the United States. Thus, millions of people suffer from the effects of alcohol, which imposes considerable financial costs for the provision of health care services. Although medications for AUD are available, their use is limited by adherence and tolerability challenges. At the same time, GLP-1 RAs, approved for the treatment of T2D and obesity, may affect patients' alcohol consumption, but the data provided by human research remain mixed. Further research is needed to determine whether these medications are linked to fewer alcohol-related health events.

About the study

The researchers conducted a retrospective cohort study using a target trial emulation framework based on de-identified Electronic Health Record (EHR) data obtained from Truveta, a network of 30 United States healthcare systems representing more than 120 million patients. The study population consisted of adults diagnosed with AUD and either T2D or obesity, starting treatment for either condition between January 1, 2018, and December 31, 2024. Four separate target trials were designed to reflect different clinical settings: an Anti-Diabetic Medication (ADM) trial, an Anti-Obesity Medication (AOM) trial, a Medications for AUD with T2D (MAUD-T2D) trial, and a Medications for AUD with obesity (MAUD-obesity) trial. Together, these trials compared prescribing situations in which patients were starting treatment for diabetes or obesity with situations in which patients were actively receiving medication for AUD. Participants on new GLP-1 RAs, namely semaglutide and tirzepatide, were compared with appropriate active comparator drugs.

All enrolled participants were monitored for up to one year to identify alcohol-related ED visits or hospitalizations. Non-alcohol-related hospitalizations served as a negative control outcome. In order to reduce the effect of confounding variables, the researchers used propensity score weighting or matching, inverse probability of treatment weighting, inverse probability of censoring weights, and Cox proportional hazards models. Sensitivity analyses were also performed to evaluate the robustness of the findings.

Study results

A total of 40,703 adults met the eligibility criteria across the four target trials. These included 18,676 participants in the ADM trial, 9,391 in the AOM trial, 8,942 in the MAUD-T2D trial, and 11,198 in the MAUD-obesity trial. Participants with T2D were generally older than those with obesity, while individuals enrolled in the medications for AUD trials showed indicators of more severe and more recent AUD. Newer GLP-1 RAs were initiated more recently than comparator therapies in all studies. Most alcohol-related ED visits or hospitalizations were identified using diagnostic codes, while the remainder were detected through laboratory testing.

The ADM trial found that 11.9% of the people in the trial group who had received treatment with newer GLP-1 RAs had an alcohol-related ED visit or hospitalization within a year after the start of the trial. This number was 14.6% in participants treated with sulfonylureas and 14.0% in individuals treated with other forms of diabetes medications. Treatment with newer GLP-1 RAs was associated with significantly lower hazards of alcohol-related ED visits or hospitalizations than sulfonylureas, equivalent to about a 26% lower hazard (HR 0.74; 95% CI 0.62-0.89), and other ADM, equivalent to about a 22% lower hazard (HR 0.78; 95% CI 0.65-0.92). However, no such differences were found compared with older GLP-1 RAs (HR 1.09; 95% CI 0.87-1.37). Sensitivity analyses using stricter outcome definitions produced similar findings.

Similar results were obtained in the AOM study, which included adults with obesity but not T2D. When controlling for pre-study differences, alcohol-related ED visits or hospitalizations in the first year were recorded for 10.1% of participants receiving the new GLP-1 RAs compared to 12.8% of patients receiving other AOM. Newer GLP-1 RAs were associated with a significantly lower risk of alcohol-related ED visits or hospitalizations than other AOM, equivalent to about a 32% lower hazard (HR 0.68; 95% CI 0.54-0.85), although no significant difference was observed compared with older GLP-1 RAs. Sensitivity analysis conducted only with the alcohol-related diagnosis code showed the same results. Exploratory analysis indicated that the use of new therapies was associated with larger reductions in alanine aminotransferase and aspartate aminotransferase levels than other medications, although these liver enzymes are nonspecific markers and were assessed only in the ADM trial.

Among participants actively receiving treatment for AUD, the associations appeared numerically stronger but required more cautious interpretation. In the MAUD-T2D trial, alcohol-related ED visits or hospitalizations occurred in 13.5% of participants treated with newer GLP-1 RAs compared with 31.4% of those receiving approved medications for AUD, meaning the newer GLP-1 RA group had about a 63% lower hazard than the approved AUD medication group, with the confidence interval suggesting the reduction could plausibly range from 54% to 71% (HR 0.37; 95% CI 0.29-0.46). In the MAUD-obesity study, event rates were 8.0% in participants receiving new GLP-1 RAs and 20.4% in participants using approved AUD therapies, equivalent to about a 65% lower hazard, with the confidence interval suggesting a plausible reduction of 53% to 74% (HR 0.35; 95% CI 0.26-0.47). Further analyses based specifically on diagnosis codes showed similar reductions, supporting consistency across clinically distinct populations, although negative control findings and high treatment discontinuation in comparator groups suggested possible residual confounding.

Conclusion

The study found that initiating newer GLP-1 RAs, including semaglutide and tirzepatide, was consistently associated with a lower observed risk of alcohol-related ED visits or hospitalizations among adults with AUD and either T2D or obesity. These associations were evident across multiple clinically distinct populations and remained consistent in several sensitivity analyses, suggesting that the findings were generally consistent. While the observational design cannot establish causality, the findings show that newer GLP-1 RAs may be promising candidates for further evaluation through randomized controlled trials to determine their effects on AUD.