Semaglutide cut heavy drinking in AUD trial

A randomised, double-blind, placebo-controlled trial published in The Lancet on 2 May 2026 found that once-weekly subcutaneous semaglutide reduced heavy drinking days by a significantly greater margin than placebo in adults with alcohol use disorder (AUD) and comorbid obesity^[1]. An ACP Journal Club commentary in the Annals of Internal Medicine summarised the findings on 4 August 2026^[2]. The trial, known as SEMALCO and led by Klausen and colleagues, is one of the most rigorous controlled studies of a GLP-1 receptor agonist in AUD to date.

What the trial tested

SEMALCO enrolled 108 adults with AUD and comorbid obesity (54 per arm) at a specialist clinic in Denmark. Participants were randomly assigned to either once-weekly semaglutide, escalated to a 2.4 mg maintenance dose, or saline placebo. Both groups received cognitive behavioural therapy alongside their injection. The primary outcome was the percentage of heavy drinking days over 26 weeks, where a heavy drinking day is defined as more than four standard drinks for men or more than three for women^[1]. Eighty-one percent of participants completed the full intervention.

What the results showed

Participants in the semaglutide arm reduced their percentage of heavy drinking days by 41.1 percentage points from baseline (95% CI: -48.7 to -33.5). The placebo arm reduced by 26.4 percentage points (95% CI: -34.1 to -18.6). The between-group difference was 13.7 percentage points, and this was statistically significant (p = 0.0015)^[1]. Both groups improved, which the researchers attribute in part to the cognitive behavioural therapy all participants received; the drug-specific effect sits on top of that shared benefit.

Secondary outcomes on alcohol-related and somatic measures also favoured semaglutide, though the full secondary dataset has not yet been published in peer-reviewed form. Adverse events in the semaglutide group were predominantly gastrointestinal, transient, and mild to moderate in severity, consistent with the class profile seen in obesity and diabetes trials^[1].

The mechanism hypothesis: GLP-1 and the reward pathway

GLP-1 receptors are expressed not only in the pancreas and gut but also in brain regions that regulate reward and compulsive behaviour, including the nucleus accumbens, ventral tegmental area, and central amygdala. Preclinical work has shown that semaglutide modulates these circuits. A 2023 study in JCI Insight found that semaglutide dose-dependently reduced binge-like alcohol intake in mice and rats and altered inhibitory (GABA) neurotransmission in the central amygdala and infralimbic cortex^[3]. The working hypothesis is that GLP-1 receptor activation dampens the reward salience of alcohol, making the reinforcing pull of drinking weaker without eliminating the capacity for reward entirely.

This is the same mechanism invoked to explain why patients on semaglutide for obesity sometimes report spontaneous reductions in alcohol craving, a pattern noted in observational data and now given a controlled-trial signal by SEMALCO. An earlier Phase 2 RCT in JAMA Psychiatry (Hendershot et al., 2025) tested lower-dose semaglutide (up to 1.0 mg weekly) in 48 non-treatment-seeking adults with AUD and found reductions in laboratory alcohol self-administration, drinks per drinking day, and weekly craving scores, with medium to large effect sizes^[4]. SEMALCO extends that signal to a treatment-seeking sample at the clinical obesity dose.

What the trial does not resolve

SEMALCO is a Phase 2 trial with 108 participants. It is large enough to establish a signal but not large enough to characterise the full benefit-risk picture, confirm which subgroups respond best, or determine how the effect holds up beyond six months. The trial enrolled adults with both AUD and obesity, so the findings do not straightforwardly generalise to people with AUD who are not obese; the obesity requirement was partly a practical constraint (semaglutide 2.4 mg is licensed for obesity) and partly a design choice that could limit how far the findings travel.

Both arms received cognitive behavioural therapy, which is an active treatment in its own right. The 26-percentage-point reduction in the placebo arm illustrates that CBT alone does meaningful work. The drug effect is an additive benefit over CBT, not a replacement for it. Trials without a therapy component would be needed to isolate the pharmacological contribution fully.

Duration of effect is also open. STEP-4, the semaglutide discontinuation trial in obesity, showed substantial weight regain after stopping the drug. Whether the reduction in heavy drinking days is similarly contingent on continued treatment is not yet known. Long-term safety and the question of whether semaglutide could become a licensed AUD treatment are both contingent on Phase 3 data that does not yet exist.

What this means now

Semaglutide is currently approved for type-2 diabetes and chronic weight management. It is not approved for alcohol use disorder in any jurisdiction. Prescribing it specifically for AUD would be off-label, and the trial data, while promising, is not sufficient to support that as a routine clinical step. Any decision about treatment for alcohol use disorder belongs with a clinician who knows the individual patient. For people already on semaglutide for obesity or diabetes, the potential incidental benefit on alcohol craving is worth discussing with a prescriber, not acting on unilaterally.^[1]