Editorial: Two for the Price of One— GLP ‐1 Receptor Agonists Could Be More Effective Than MAUD in Patients With AUD and Metabolic Dysfunction

Glucagon-like peptide-1 (GLP-1) receptor agonists have emerged as a promising new class of medications for alcohol use disorder (MAUD), as anecdotes of reduced drinking have grown and GLP-1 receptors have been identified in brain regions associated with reward and addiction [1-3]. A recent prospective, randomized clinical trial found semaglutide significantly reduced alcohol cravings and consumption in a laboratory self-administration procedure compared to placebo [4]. Gougol and colleagues add to this evidence in their large retrospective cohort study of patients in the Stanford University Healthcare System (California, USA) from 2017 to 2025 with AUD and concurrent metabolic dysfunction (type 2 diabetes and/or obesity) [5]. Patients treated with GLP-1s (semaglutide or tirzepatide) for at least 6 months demonstrated a 45% relative reduction in one-year AUD relapse risk compared to FDA-approved MAUD (naltrexone, acamprosate, or disulfiram) (p < 0.01). GLP-1 use, compared to naltrexone, was also associated with reductions in BMI (mean difference (MD) −1.1, p = 0.004) and HbA1c (MD −1.1 p = 0.02), though other secondary endpoints including liver decompensation and mortality did not reach statistical significance. Several limitations temper this study's findings. Relapse was defined by ICD codes rather than objective measures such as timeline followback or phosphatidylethanol (PEth) levels. AUD severity was inferred from clinical documentation of AUD-related complications rather than structured assessments, raising the possibility that GLP-1 recipients had milder AUD severity and were less likely to have AUD relapse coded. While propensity score matching was used to control for confounding across variables, it did not account for behavioural treatments—including twelve-step programs or motivational interviewing—which play a central role in AUD treatment and may substantially influence outcomes. Adherence and dosing of MAUDs were also not assessed, potentially underestimating the effectiveness of conventional treatments. Finally, the exclusion of patients with advanced liver disease limits generalizability to the population arguably most in need of effective MAUD. Nonetheless, these findings are consistent with another study reporting a 50%–56% lower risk of AUD incidence and recurrence among patients with obesity initiated on GLP-1 agonists [6]. Whether certain subgroups may benefit more from GLP-1-s remains unclear. While GLP-1-s have generally been associated with improvements in alcohol outcomes, one study found increased alcohol use among patients with a BMI below 25, possibly reflecting greater drops in blood glucose driving alcohol cravings [7]. However, it also raises the possibility that AUD pathways are distinct in the setting of comorbid obesity [8, 9]. Sex-based differences in the relationship between AUD and obesity have also been identified but remain incompletely explored in studies examining GLP-1-s [10]. Future studies should conduct subgroup analyses with BMI, sex, race, and AUD severity to understand how the effect of these medications may vary across populations, and incorporate objective alcohol biomarkers to more rigorously assess relapse. GLP-1 receptor agonists represent a potential breakthrough in AUD treatment. Their widespread use for metabolic dysfunction may reduce treatment stigma and broaden patient acceptance. While this retrospective study does not provide definitive, practice-changing results, it adds to the groundswell of evidence suggesting that GLP-1s can improve metabolic dysfunction plus AUD: two for the price of one. Laila Fozouni: conceptualization, writing – original draft, writing – review and editing. Gene Im: conceptualization, writing – original draft, writing – review and editing. The authors have nothing to report. This article is linked with Gougol et al. paper. To view this article, visit https://doi.org/10.1111/apt.70596. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.