Study links semaglutide to about half the rate of alcohol use disorder diagnoses

People with obesity who were prescribed semaglutide were about half as likely to be diagnosed with alcohol use disorder over the next year as similar patients taking other anti-obesity medications, according to a retrospective study of electronic health records published in Nature Communications in May 2024 ^[1].

The analysis followed 83,825 patients with obesity and no prior diagnosis of alcohol use disorder, or AUD, who received a first prescription for semaglutide (sold as Ozempic, Wegovy and Rybelsus) or for a non-GLP-1 anti-obesity medication — including naltrexone or topiramate — between June 2021 and December 2022 ^[1]. After propensity-score matching, a method that balances groups on traits such as age and sex, each group had 26,566 patients, with a mean age of 51.2 years; 65.9% were women, 15.8% Black, 66.6% white and 6.5% Hispanic ^[1].

Over 12 months, 0.37% of patients prescribed semaglutide received an AUD diagnosis, compared with 0.73% of those prescribed other anti-obesity medications, a hazard ratio of 0.50 (95% confidence interval 0.39–0.63) ^[1]. Compared with naltrexone or topiramate specifically, the rates were 0.35% versus 0.78% (hazard ratio 0.44, 95% CI 0.32–0.61) ^[1]. The lower risk appeared across sex, age group and race, and in patients with and without type 2 diabetes, the researchers wrote ^[1]. The paper's abstract summarizes the result as a 50% to 56% lower risk of both new and recurrent AUD over the 12-month follow-up ^[1].

The researchers also studied 4,254 patients with obesity and a prior AUD diagnosis. Before matching, 1,470 had been prescribed semaglutide and 2,784 other anti-obesity medications; after matching, each group had 1,051 patients, with a mean age of 52.6 years and 41.5% women ^[1]. The team replicated the findings in a separate population of 598,803 patients with type 2 diabetes ^[1].

AUD is common and treatment options are limited. An estimated 29.5 million Americans ages 12 and older, or 10.6%, had AUD in 2021, and it is linked to more than 80,000 deaths a year in the United States ^[1]. Only three medications are approved by the FDA to treat AUD, and the paper describes their benefits as modest ^[1]. Earlier hints that GLP-1 drugs might affect drinking came from rodent studies, case reports, analyses of social media posts and a 127-patient randomized trial of a different GLP-1 drug, exenatide, which reported fewer heavy drinking days in a subgroup of patients with obesity ^[1].

The new work is observational, so it can show an association but cannot prove that semaglutide caused the lower diagnosis rates ^[1]. Before matching, the semaglutide group was older and had more severe obesity, more type 2 diabetes and fewer mental health or tobacco use diagnoses than the comparison group ^[1]. The authors write that the findings call for randomized clinical trials ^[1]. A correction to the paper was published on June 18, 2024 ^[1].

Why it matters for patients

About one in ten Americans age 12 and older had alcohol use disorder in 2021, and the country has only three FDA-approved medicines for it, with modest benefits ^[1]. If semaglutide really does lower the risk, a drug already taken by millions for weight loss or diabetes would have a second effect — reports from patients describe a reduced desire to drink ^[1].

This study does not show that. It counted diagnoses recorded in medical records, so people who drink heavily without ever being diagnosed would not appear in the data, and how many such patients exist has not been reported ^[1]. Whether the size of the effect differs for people without obesity or type 2 diabetes has not been reported either ^[1]. Semaglutide is not among the FDA-approved treatments for alcohol use disorder, and this study does not change its approved uses ^[1].

What happens next

The paper's authors call for randomized clinical trials to test whether semaglutide can prevent or treat alcohol use disorder ^[1]. Results from such trials have not been reported, and no start dates, sponsors or regulatory filings for an AUD indication have been disclosed ^[1]. The correction published on June 18, 2024, is the most recent change to the paper noted in the journal record ^[1].