Do GLP-1s curb addictions?

The question everyone is asking#

People on semaglutide and tirzepatide keep saying the same thing in slightly different words: the drinks stopped calling as loudly. The evening smoke felt optional. The junk-food loop went quiet, and then some of the other loops did too.

That is a big claim for a gut hormone drug. It is also the kind of claim that, for once, science has started catching up with rather than laughing at.

So: do GLP-1s curb addictions?

Short answer: for alcohol, the human evidence is now strong enough to take seriously. For other substances, the signal is promising and incomplete. For gambling, shopping, and nail-biting, you are still mostly in anecdote territory. None of these drugs is currently licensed as an addiction treatment. Educational overview, not medical advice.

Why a weight-loss drug would touch addiction at all#

GLP-1 is a gut hormone involved in insulin release, satiety, and gastric emptying. The modern drugs are long-acting versions of that signal.

The part that matters for addiction is where the receptors live. GLP-1 receptors are not only in the gut and pancreas. They show up in brain regions that handle reward, motivation, and cue-driven behaviour: ventral tegmental area, nucleus accumbens, and related circuitry. Those are the same pathways alcohol, nicotine, cocaine, and opioids hijack when they flood the system with dopamine and teach the brain to chase the next hit.

Researchers have argued for years that obesity and substance use disorders share overlapping reward biology: not "food is a drug" as a slogan, but a real overlap in how the brain assigns value to cues and consummatory behaviour. If a drug turns down the hedonic pull of highly palatable food, it is not wild to ask whether it also turns down the pull of a pint, a cigarette, or a line.

Preclinical work already pointed that way. In animals, GLP-1 receptor agonists reduce alcohol intake, blunt nicotine reinforcement, and dampen responses to several drugs of abuse. The social-media stories came later. The biology came first.

Alcohol is where the evidence got real#

This is the substance with the cleanest human data.

Small phase 2 trial, 2025#

A double-blind randomised trial in JAMA Psychiatry gave once-weekly semaglutide or placebo to 48 adults with alcohol use disorder who were not seeking treatment. Low-dose semaglutide reduced alcohol consumed in a lab self-administration task, cut drinks per drinking day, and reduced weekly craving. It did not cleanly reduce the number of drinking days. In smokers in the sample, cigarettes per day also fell relative to placebo.

That is not a miracle cure. It is a medium-to-large signal on craving and intensity of drinking, in a small study, over nine weeks. Exactly the sort of result that justifies bigger trials.

Larger trial in people with obesity and AUD, 2026#

Then came the one that moved the conversation. In a 26-week randomised, double-blind, placebo-controlled trial published in The Lancet, 108 treatment-seeking patients with moderate to severe alcohol use disorder and obesity received weekly semaglutide (up to 2.4 mg) or placebo on top of cognitive behavioural therapy.

Semaglutide cut heavy drinking days by 41.1 percentage points from baseline, versus 26.4 with placebo: an estimated treatment difference of 13.7 percentage points. Blood-alcohol biomarkers backed the self-reports. GI side effects were more common on drug, mostly mild to moderate and transient. NIH coverage of the trial noted a number needed to treat around 4.3 in this study, better than the usual figures quoted for existing AUD medicines.

Important caveats, because they matter:

• Everyone in that Lancet trial also had obesity. An earlier GLP-1 alcohol study had suggested the benefit may be clearer in people with higher BMI. Do not assume the same effect size in a lean person with AUD.
• CBT was given to both arms. Semaglutide looked like an add-on, not a replacement for psychosocial care.
• Duration was six months. Addiction is a longer game.

Still: this is no longer vibes. For alcohol use disorder with obesity, weekly semaglutide has randomised evidence of a clinically meaningful effect on heavy drinking.

What about everything else?#

Cannabis#

A large retrospective EHR study in Molecular Psychiatry compared semaglutide with other anti-obesity or anti-diabetes drugs. In people with obesity, semaglutide was associated with lower risk of a new cannabis use disorder diagnosis (hazard ratio 0.56) and lower risk of recurrent CUD diagnosis (HR 0.62). Similar patterns showed up in a type 2 diabetes cohort.

That is association, not proof of causation. EHR diagnoses are messy. Propensity matching helps and does not erase every confounder. It is still a real-world signal consistent with the reward-dampening story, and there are still no approved medicines for cannabis use disorder, so researchers are paying attention.

Nicotine#

Signals exist: preclinical reductions in nicotine intake, exploratory cigarette reductions nested inside the alcohol trial, and small questionnaire work in people with type 2 diabetes suggesting semaglutide may blunt tobacco craving more than some older GLP-1s. Dedicated, adequately powered cessation trials are still the missing piece.

Opioids and stimulants#

Here the story is mostly "trials underway." There are active protocols testing semaglutide in opioid use disorder and cocaine use disorder, plus NIH HEAL-linked work pairing semaglutide with CBT for cocaine. Animal data and mechanism make the hypothesis credible. Human efficacy is not settled.

Behavioural addictions#

Reddit threads and mixed-methods social media scrapes describe quieter compulsive shopping, less nail-biting, and the occasional gambling comment. Treat that as hypothesis generation. Controlled evidence for behavioural addictions is thin. If your brain uses the same reward dial for a slot machine that it uses for a pastry, the prediction is interesting. Prediction is not data.

What this is not#

A few lines worth saying plainly.

Not approved for addiction. In 2026, GLP-1 receptor agonists remain diabetes and obesity medicines. Using them for AUD or other substance use disorders is off-label research-adjacent territory, not standard care.

Not a personality transplant. People report less pull, not a rewritten life. Therapy, mutual aid, sleep, environment, and honesty about triggers still do the heavy lifting. The Lancet trial layered drug on CBT for a reason.

Not risk-free. Nausea, vomiting, constipation, gallbladder issues, rare but serious risks, cost, supply, and muscle loss on rapid weight drop are all part of the real conversation. Turning down reward can also feel flat for some people. "I want food less" and "I want life less" are not the same sentence, but they sit closer together than marketing admits.

Not a grey-market free-for-all. Compounded or research vials are not the same product as a licensed pen. Purity, dose accuracy, and sterility are not optional details when you are talking about weekly injectables and addiction vulnerability.

Not "food addiction proved." Shared circuitry does not mean every overeating pattern is a substance use disorder, and it does not mean every person on a GLP-1 has an addiction that got cured as a side effect.

How to hold the finding#

The useful frame is narrower than the headlines and wider than the sceptics.

GLP-1 drugs appear to turn down a reward-and-cue system that addiction exploits. Alcohol has the best clinical evidence so far, especially when obesity and AUD sit together. Other substances have enough signal to fund serious trials. Behavioural addictions remain speculation with anecdotes attached.

If you are already prescribed a GLP-1 for diabetes or obesity and you notice alcohol or nicotine getting quieter, that is a clinically interesting observation worth telling your clinician. If you are hoping a pen will replace detox, relapse prevention, and the boring work of changing a life, it will not.

The optimistic read is still the honest one: for the first time in a long while, a widely used medicine class may give addiction medicine a new lever. The next step is larger trials across substances, clearer guidance on who benefits, and enough humility to keep psychosocial care in the room.

If you need help now#

In the UK: Alcohol Change UK, Frank on 0300 123 6600, NHS alcohol support, and Samaritans on 116 123. If you are in immediate danger, 999.

References#

• Hendershot et al., Once-weekly semaglutide in adults with alcohol use disorder (JAMA Psychiatry, 2025)
• Klausen et al., Once-weekly semaglutide versus placebo in AUD with obesity (The Lancet, 2026)
• NIH, Adding weekly GLP-1 to CBT further reduces heavy drinking
• Wang et al., Semaglutide and cannabis use disorder in real-world populations (Molecular Psychiatry, 2024)
• ClinicalTrials.gov, GLP-1R agonist treatment for opioid use disorder (NCT06548490)
• ClinicalTrials.gov, Semaglutide in cocaine use disorder (NCT06691243)
• Arillotta et al., GLP-1 receptor agonists, substance use, and compulsive behaviour: social media mixed-methods study
• Boston University, Do obesity and diabetes drugs also curb drinking and other addictions?