Table of Contents
- GLP-1 Agonists vs. AUD Medications: Relapse Rates in Metabolic Disease
- Abstract
- Study at a Glance
- Study Snapshot
- Study Facts Table
- What Researchers Actually Did
- Key Findings: Primary Outcomes
- Key Findings: Secondary Outcomes and Subgroup Analyses
- Adverse Events and Safety Profile
- Statistical Approach and Rigor
- Clinical Takeaway
- Why This Matters Clinically
- Read This Paper Through Nine Different Lenses
- What are GLP-1 receptor agonists?
- How do GLP-1RAs compare to naltrexone for AUD relapse?
- What are the metabolic outcomes for patients treated with GLP-1RAs?
- Does the study address liver outcomes?
- What are the limitations of this study?
- Are there any safety concerns with GLP-1RAs in AUD patients?
- What is the sample size for this study?
- How was propensity score matching used in the study?
- What are the implications for clinical practice?
- Are there any plans for further research?
- Read next
GLP-1 Agonists vs. AUD Medications: Relapse Rates in Metabolic Disease
Exploring the Impact of GLP-1 Receptor Agonists on AUD Relapse
Alcohol Use Disorder
Metabolic Dysfunction
Steatotic Liver Disease
Semaglutide
- How GLP-1 receptor agonists (semaglutide, tirzepatide) compare to naltrexone, acamprosate, and disulfiram for one-year AUD relapse in patients who also have obesity or type 2 diabetes
- What the metabolic and liver-related outcome data show after propensity score matching in a real-world Stanford Health Care cohort
- Where the data are genuinely compelling versus where confounding and study design demand caution
- What this means right now for clinicians treating patients with overlapping metabolic and addiction diagnoses
TL;DR: In a propensity-score-matched retrospective cohort, GLP-1 receptor agonists were associated with a 45% lower rate of one-year AUD relapse compared to naltrexone in patients with concurrent metabolic dysfunction, though the design cannot establish causality and significant baseline differences between groups are not fully eliminable by matching alone.
Want to apply this research to your care?
CED Clinic translates emerging research into individualized clinical care. Dr. Caplan has treated 30,000+ patients.
Book a consultation →Abstract
Background: Metabolic dysfunction (MetD) and alcohol use disorder (AUD) frequently coexist as synergistic risk factors for steatotic liver disease. Glucagon-like peptide-1 receptor agonists (GLP-1RAs) are established therapies for MetD, including type 2 diabetes mellitus (T2DM) and obesity. Recent studies suggested potential beneficial effects of GLP-1RA to decrease addictive behaviours in AUD. The authors evaluated the outcomes of GLP-1RA therapy compared with FDA-approved pharmacotherapies for AUD, including naltrexone, acamprosate, and disulfiram, in patients with dual risk factors of MetD and AUD.
Methods: Retrospective cohort study of patients at Stanford Health Care (2017–2025). Eligible patients had concurrent diagnoses of alcohol-related complications meeting criteria for AUD and MetD, defined as obesity (BMI >25) and/or T2DM with HbA1c >5.7. Patients with advanced liver disease within one year of diagnosis were excluded. Exposure groups required at least six months of GLP-1RA therapy (semaglutide or tirzepatide) or at least six months of an FDA-approved AUD pharmacotherapy. Propensity score matching was employed to reduce confounding.
Results: 1,946 patients with concurrent AUD and MetD were identified. Of these, 274 were exposed to GLP-1RA, 1,272 to naltrexone, 232 to acamprosate, and 168 to disulfiram. Mean follow-up was 1,341 days. GLP-1RA patients had higher baseline BMI (35.5 vs. 30.1) and higher T2DM prevalence (66% vs. 14%). GLP-1RA was associated with lower one-year AUD relapse (IRR 0.55, 95% CI 0.42–0.73; p<0.01), greater BMI reduction (−1.3 vs. −0.3; p=0.004), and HbA1c improvement (−1.0 vs. +0.1; p=0.02). Decompensated cirrhosis trended lower but did not reach statistical significance (HR 0.52, p=0.09). Mortality was similar across groups.
Conclusions: GLP-1RAs are a promising option for patients with concurrent MetD and AUD, improving relapse rates and metabolic outcomes compared with currently FDA-approved pharmacotherapies for AUD. Trends toward better liver outcomes support further prospective evaluation.
DOI: https://doi.org/10.1111/apt.70596
Study at a Glance
| Study Design | Retrospective propensity-score-matched cohort study |
| Population | Adults aged 18–75 with concurrent AUD and metabolic dysfunction (obesity BMI >25 and/or T2DM with HbA1c >5.7), without advanced liver disease at baseline; Stanford Health Care, 2017–2025 |
| Total N | 1,946 (GLP-1RA: 274; naltrexone: 1,272; acamprosate: 232; disulfiram: 168) |
| Primary Endpoint | One-year incidence of AUD relapse (ICD-10-coded healthcare encounter for recurrent alcohol use, intoxication, withdrawal, or AUD-related complication) |
| Key Finding | GLP-1RA associated with 45% lower relapse rate vs. naltrexone after propensity score matching (PS-IRR 0.55, 95% CI 0.42–0.73; p<0.01) |
| Mean Follow-Up | 1,341 days (~3.7 years) |
Study Snapshot
| Outcome | GLP-1RA | Naltrexone | Acamprosate | Disulfiram |
|---|---|---|---|---|
| 1-Year AUD Relapse | 34% | 65% | 67% | 73% |
| Change in BMI | −1.27 | −0.26 | +0.51 | +0.42 |
| Change in HbA1c | −0.72 | +0.18 | −0.01 | +0.12 |
| Decompensated Cirrhosis | 3.65% | 5.9% | 9.05% | 10.71% |
| All-Cause Mortality | 2.92% | 1.57% | 0.86% | 2.98% |
Study Facts Table
| Authors | Gougol A, Kwo P, Pike W, Farokhnia M, Hui G, Gombar S, Mirminachi B |
| Journal | Alimentary Pharmacology & Therapeutics |
| Year | 2026 |
| Design | Retrospective propensity-score-matched cohort |
| Total N | 1,946 |
| Intervention | Semaglutide or tirzepatide for at least 6 months |
| Comparators | Naltrexone, acamprosate, or disulfiram for at least 6 months (mutually exclusive groups) |
| Primary Endpoint | One-year incidence of AUD relapse by ICD-10 code |
| Secondary Endpoints | Change in ALT, incident decompensated cirrhosis, all-cause mortality, change in BMI and HbA1c |
| Key Results | PS-IRR for relapse vs. naltrexone: 0.55 (95% CI 0.42–0.73; p<0.01) PS-IRR for relapse vs. acamprosate: 0.57 (95% CI 0.46–0.72; p<0.01) PS-IRR for relapse vs. disulfiram: 0.65 (95% CI 0.44–0.97; p=0.03) Decompensated cirrhosis HR vs. naltrexone: 0.52 (95% CI 0.24–1.11; p=0.09, NS) Mortality HR vs. naltrexone: 0.65 (95% CI 0.24–1.71; p=0.38, NS) |
| Adverse Events | Not formally assessed in this study |
| Funding | None reported |
| Conflicts of Interest | Not declared in the paper; two authors affiliated with Atropos Health Inc. |
What Researchers Actually Did
Gougol and colleagues mined Stanford Health Care’s longitudinal electronic data warehouse, spanning more than three million patients from 2017 through 2025, to identify adults aged 18–75 with concurrent ICD-coded AUD and metabolic dysfunction. MetD was operationalized as BMI greater than 25 kg/m² and/or T2DM with HbA1c above 5.7%, a narrower definition than the full five-criterion CMRF framework endorsed by AASLD in 2023, justified by the authors on clinical and methodological grounds. Patients were required to have the two diagnoses within 90 days of each other, at least one year of follow-up, and no prior compensated or decompensated cirrhosis or alcohol-associated hepatitis within 12 months of the index date. Those receiving GLP-1RAs must have done so for at least six months; the same duration requirement applied to each comparator AUD drug. Groups were mutually exclusive: no prior exposure to other AUD pharmacotherapies was permitted.
To address confounding, the team applied two strategies sequentially: basic Mahalanobis-distance matching on age and sex, then propensity score (PS) matching incorporating a broad covariate set including Charlson Comorbidity Index, BMI, laboratory markers, number of healthcare visits, and multiple surrogates for AUD severity (alcohol withdrawal history, alcohol-related hospitalizations, AUD-related end-organ damage, and prior AUD pharmacotherapy exposure). Pairwise comparisons were conducted independently for GLP-1RA versus each of the three AUD drugs. Outcomes were analyzed via incidence rate ratios for relapse, Cox proportional hazards models for cirrhosis and mortality, and mean difference or standardized mean difference for continuous metabolic parameters.
Key Findings: Primary Outcomes
- Overall one-year AUD relapse occurred in 58% of all patients across groups, underscoring the difficulty of sustained remission in this population regardless of therapy.
- Relapse rate in the GLP-1RA group was 34%, versus 65% for naltrexone, 67% for acamprosate, and 73% for disulfiram in the unmatched cohort.
- After PS matching, GLP-1RA remained superior to naltrexone for relapse reduction: PS-IRR 0.55 (95% CI 0.42–0.73; p<0.01).
- GLP-1RA was also superior to acamprosate after PS matching: PS-IRR 0.57 (95% CI 0.46–0.72; p<0.01).
- Against disulfiram, basic matching showed no significant difference; PS matching revealed superiority for GLP-1RA: PS-IRR 0.65 (95% CI 0.44–0.97; p=0.03).
Key Findings: Secondary Outcomes and Subgroup Analyses
- BMI: GLP-1RA produced a mean BMI reduction of −1.27 kg/m². Naltrexone produced −0.26, acamprosate +0.51, and disulfiram +0.42. After PS matching vs. naltrexone, mean difference was −1.1 (95% CI −1.8 to −0.3; p=0.004). Results were significant in all PS-matched pairwise comparisons.
- HbA1c: GLP-1RA produced a mean reduction of −0.72%; naltrexone showed a mean increase of +0.18% (p=0.003). After PS matching vs. naltrexone, HbA1c remained significantly lower in the GLP-1RA group (MD −1.1; 95% CI −1.9 to −0.07; p=0.02). The comparison vs. acamprosate did not reach significance after PS matching due to limited sample size. PS matching was not performed for the GLP-1RA vs. disulfiram HbA1c comparison due to small patient numbers.
- Decompensated cirrhosis: Occurred in 3.65% (GLP-1RA), 5.9% (naltrexone), 9.05% (acamprosate), and 10.71% (disulfiram). Versus naltrexone: HR 0.52 (95% CI 0.24–1.11; p=0.09, not significant). Versus acamprosate: significant in unadjusted comparison (HR 0.48; p=0.029) but attenuated to non-significance after PS matching (PS-HR 0.60; 95% CI 0.20–1.18; p=0.37). Versus disulfiram: not significant (HR 0.68; p=0.25).
- ALT: Mean change was −4.82 U/L for GLP-1RA, −2.83 U/L for naltrexone, −9.54 U/L for acamprosate, and +5.82 U/L for disulfiram. No formal statistical comparison on ALT was reported as a primary analysis.
- All-cause mortality: 35 total deaths. PS-matched HR for GLP-1RA vs. naltrexone: 0.65 (95% CI 0.24–1.71; p=0.38). No statistically significant difference versus acamprosate or disulfiram either.
- Subgroup note: The authors cite a prior RCT of exenatide (first-generation GLP-1RA) in which benefit for alcohol reduction was confined to the subgroup with BMI >30 kg/m², which is relevant context for interpreting this study’s MetD-restricted population.
Adverse Events and Safety Profile
No formal safety or adverse event analysis was conducted within this study. Adverse events are not reported. This is a meaningful gap for a population in which GLP-1 receptor agonists carry known risks of nausea, vomiting, pancreatitis, and potential interactions with alcohol use patterns. Any assessment of comparative safety between GLP-1RAs and AUD pharmacotherapies in this specific population requires prospective data.
Statistical Approach and Rigor
The analytic framework is appropriately layered: unmatched, basic covariate-matched, and PS-matched results are all reported, which allows for internal consistency checking. Continuous outcomes were assessed with t-tests and expressed as mean differences with 95% CIs; time-to-event outcomes used Cox proportional hazards with Kaplan-Meier curves; count outcomes used incidence rate ratios. The PS model incorporated a clinically sensible and broad covariate set, including several AUD-severity surrogates not always present in administrative data studies. However, residual confounding cannot be excluded, particularly given that GLP-1RAs were prescribed primarily for metabolic indications while AUD drugs were prescribed for addiction, creating a fundamental asymmetry in the treatment context that propensity matching cannot fully resolve. After PS matching, GLP-1RA patients still had higher BMI and HbA1c, indicating incomplete covariate balance on key metabolic variables. Matched sample sizes for the acamprosate (n=96 per group) and disulfiram (n=89 per group) comparisons are small, reducing precision. E-values are reported for BMI comparisons, providing some quantification of the unmeasured confounding that would be required to explain away the findings.
Clinical Takeaway
For clinicians managing patients who carry both a metabolic diagnosis (obesity or T2DM) and an active AUD, this study provides real-world evidence that GLP-1 receptor agonists, specifically semaglutide and tirzepatide, may reduce one-year AUD relapse rates more effectively than naltrexone, acamprosate, or disulfiram, while simultaneously improving weight and glycemic control. The liver-outcome data trend favorably but do not reach statistical significance. The findings are hypothesis-generating, not practice-defining: the retrospective design, differential treatment indications, and residual confounding preclude causal conclusions. That said, for a patient already receiving semaglutide for T2DM or obesity who also carries an AUD diagnosis, the evidence base for potential dual benefit is strengthening and warrants explicit clinical attention.
Clinical Bottom Line: GLP-1 receptor agonists were associated with substantially lower one-year AUD relapse rates than all three FDA-approved AUD medications in propensity-matched patients with concurrent metabolic dysfunction, but this is observational data requiring prospective validation before it changes prescribing practice.
Why This Matters Clinically
Up to one in five patients with AUD also carries concurrent metabolic dysfunction, yet this population is systematically excluded from most MASLD trials and remains therapeutically underserved. Current FDA-approved AUD pharmacotherapies have modest effect sizes: naltrexone’s number-needed-to-treat is approximately 12 in some meta-analyses, disulfiram fails to outperform placebo in blinded trials, and fewer than 5% of patients with AUD receive any pharmacotherapy at all. Against that baseline, a therapy that reduces relapse by roughly 35–45% in a matched real-world cohort
Read This Paper Through Nine Different Lenses
The same evidence can produce very different conclusions depending on the question being asked. Explore this study through multiple physician-guided interpretive frameworks.
Overview
This study demonstrates that GLP-1 receptor agonists, such as semaglutide and tirzepatide, are associated with a significantly lower rate of one-year alcohol use disorder (AUD) relapse compared to traditional AUD medications like naltrexone in patients with concurrent metabolic dysfunction.
The findings highlight the potential for GLP-1RAs not only to improve metabolic health but also to enhance addiction treatment outcomes, suggesting a promising dual-benefit approach for this patient population.
- GLP-1RA reduces AUD relapse by 45% compared to naltrexone.
- Metabolic improvements include BMI reduction and HbA1c improvement.
- Trends toward better liver outcomes suggest further research is warranted.
Patient Takeaway
If you have both alcohol use disorder and metabolic issues like obesity or diabetes, GLP-1 receptor agonists may offer a dual benefit by helping manage your addiction while also improving your metabolic health.
These medications can lead to weight loss and better blood sugar control, which might make it easier to maintain sobriety in the long term.
- GLP-1RAs help with both AUD and metabolic issues.
- Weight loss and improved blood sugar are potential benefits.
- Further research is needed for full understanding of liver outcomes.
Clinician’s POV
Clinicians treating patients with concurrent alcohol use disorder (AUD) and metabolic dysfunction may consider GLP-1 receptor agonists as a promising treatment option.
These medications are associated with lower AUD relapse rates and improved metabolic outcomes, offering a dual-benefit approach to care.
- GLP-1RAs reduce AUD relapse by 45% compared to naltrexone.
- Metabolic improvements include BMI reduction and HbA1c improvement.
- Trends toward better liver outcomes suggest further research is warranted.
A Skeptical Read
While the study suggests that GLP-1 receptor agonists may reduce AUD relapse in patients with metabolic dysfunction, it is important to consider the limitations of a retrospective design and potential residual confounding.
Further prospective studies are needed to confirm these findings and fully understand the mechanisms behind the observed benefits.
- Retrospective design limits causal inference.
- Residual confounding from baseline differences remains.
- Prospective studies are needed for confirmation.
Study Critic
Critics may point out that the study’s reliance on propensity score matching to address confounding introduces potential biases and does not fully eliminate baseline differences between groups.
Additionally, the lack of formal safety analysis is a significant limitation, as GLP-1RAs carry known risks in this patient population.
- Propensity score matching has limitations.
- Safety concerns with GLP-1RAs need further investigation.
- Baseline differences may not be fully addressed.
Compared to Past Research
Prior research has suggested potential beneficial effects of GLP-1 receptor agonists on addictive behaviors, but this study provides real-world evidence in a patient population with concurrent metabolic dysfunction and AUD.
Previous studies have focused more on the metabolic benefits of GLP-1RAs rather than their impact on addiction outcomes.
- Prior research suggested potential benefits for addictive behaviors.
- This study focuses on real-world outcomes in patients with both AUD and metabolic issues.
- Historical context highlights the novelty of this dual-benefit approach.
Practical Considerations
Practically, clinicians can consider GLP-1 receptor agonists as a potential treatment option for patients with both AUD and metabolic dysfunction.
The medications are associated with lower AUD relapse rates and improved metabolic outcomes, offering a dual-benefit approach to care.
- GLP-1RAs reduce AUD relapse by 45% compared to naltrexone.
- Metabolic improvements include BMI reduction and HbA1c improvement.
- Trends toward better liver outcomes suggest further research is warranted.
Future Directions
Future research should focus on prospective studies to confirm the findings of this retrospective cohort study.
Investigating the mechanisms behind the observed benefits and conducting safety analyses in this patient population are also important areas for further investigation.
- Prospective studies are needed for confirmation.
- Mechanisms behind GLP-1RA benefits need exploration.
- Safety analysis is crucial given known risks of GLP-1RAs.
Misreadings & Bad-Faith Takes
A common misunderstanding is that the study establishes causality between GLP-1 receptor agonist use and AUD relapse reduction. However, due to its retrospective design, the study cannot definitively prove causation.
Another potential misreading is that GLP-1RAs are a cure-all for both AUD and metabolic issues. While they offer benefits, further research is needed to fully understand their impact.
- Study does not establish causality.
- GLP-1RAs are not a cure-all for AUD or metabolic issues.
- Further research is needed for full understanding.
Have thoughts on this? Share it:
What are GLP-1 receptor agonists?
GLP-1 receptor agonists (GLP-1RAs) are medications that mimic the effects of glucagon-like peptide-1, a hormone that helps regulate blood sugar and reduce appetite.
How do GLP-1RAs compare to naltrexone for AUD relapse?
In this study, GLP-1RA was associated with a 45% lower rate of one-year AUD relapse compared to naltrexone after propensity score matching.
What are the metabolic outcomes for patients treated with GLP-1RAs?
GLP-1RA produced a mean BMI reduction of -1.27 kg/m² and a mean HbA1c reduction of -0.72% compared to naltrexone.
Does the study address liver outcomes?
The study shows a trend toward lower decompensated cirrhosis rates in GLP-1RA patients, though it did not reach statistical significance.
What are the limitations of this study?
The study cannot establish causality due to its retrospective design and residual confounding from baseline differences between groups.
Are there any safety concerns with GLP-1RAs in AUD patients?
No formal safety analysis was conducted, but known risks of GLP-1RAs include nausea, vomiting, and potential interactions with alcohol use patterns.
What is the sample size for this study?
The study included 1,946 patients with concurrent AUD and metabolic dysfunction, divided into groups based on their treatment regimen.
How was propensity score matching used in the study?
Propensity score matching was employed to reduce confounding by balancing covariates across treatment groups.
What are the implications for clinical practice?
The findings suggest that GLP-1RAs may be a promising option for patients with concurrent metabolic dysfunction and AUD, improving relapse rates and metabolic outcomes.
Are there any plans for further research?
Trends toward better liver outcomes support the need for further prospective evaluation of GLP-1RAs in this patient population.
