Semaglutide for alcohol use disorder: the primary endpoint failed, the drinking still fell

Design

Phase 2, double-blind, randomised, parallel-arm, placebo-controlled trial over 8 weeks

Population

50 treatment-seeking adults with moderate to severe alcohol use disorder

Primary outcome

Laboratory alcohol cue-elicited craving at week 6

Effect

Primary outcome not significant; heavy drinking days b = -0.580 (95 per cent CI -1.012 to -0.148), drinks per drinking day -1.177 (-2.307 to -0.047), cannabis use days -1.434 (-2.568 to -0.301)

Fifty treatment-seeking adults with moderate to severe alcohol use disorder were randomised double-blind to oral semaglutide - 3 mg daily for four weeks, then 7 mg for four weeks - or placebo. The primary outcome was laboratory-based alcohol cue-elicited craving at week 6.

It did not move, and neither did drinks per day. What did move were the preregistered secondary outcomes and the exploratory ones: heavy drinking days (b = -0.580, 95 per cent CI -1.012 to -0.148), drinks per drinking day (-1.177, -2.307 to -0.047), naturalistic craving reported in daily life rather than in the laboratory (-2.195, -4.174 to -0.216), alcohol-related consequences (-4.618, -8.651 to -0.585) and cannabis use days (-1.434, -2.568 to -0.301). More participants on semaglutide dropped at least one World Health Organization risk drinking level.

Read the discrepancy rather than picking a side of it. A cue-reactivity paradigm in a laboratory and a diary of what somebody actually drank measure different things, and it is not obviously the laboratory that is right - but a trial whose primary outcome is null and whose secondaries are positive is, formally, a negative trial with encouraging signals. Fifty patients, eight weeks, several outcomes tested. This warrants the phase 3 the authors call for; it does not warrant prescribing semaglutide for alcohol use disorder, and a patient asking about it should be told the trial exists, that it missed its main target, and that the drug's licensed indications are unchanged.

• Do not prescribe a GLP-1 receptor agonist for alcohol use disorder outside a trial; the licensed indications are unchanged.
• If a patient raises it, describe the trial honestly: primary endpoint null, drinking outcomes positive, 50 participants.
• Note the cannabis signal as genuinely interesting and entirely exploratory.
• Continue established treatment - naltrexone, acamprosate, disulfiram where appropriate, and psychosocial treatment.
• Watch for the phase 3; this is the second trial pointing the same way, now in a more severe treatment-seeking group.

The statistics, in plain English

When the primary outcome is null and several secondaries are positive, the trial has not shown what it set out to show; each additional outcome tested raises the chance that one crosses significance by accident, and the intervals here run close to zero at their upper bounds - -0.047 for drinks per drinking day, -0.216 for craving. With 25 patients per arm, even the significant effects are estimated loosely. The consistency of direction across several different drinking measures is what keeps this interesting despite the formal result.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.