Psilocybin for alcohol use disorder: what the trials show

Does psilocybin help people with alcohol use disorder? The honest answer is that researchers are still finding out. A handful of small trials have been published, and they do not all agree. This article lays out what each one tested, how many people took part, what it was compared with, and what it could not show. It describes research. It does not tell anyone what to do.

Every study below paired psilocybin with psychotherapy or structured therapy sessions, in supervised clinical research settings. None of this describes taking mushrooms on your own. For help with drinking, see where to find help in Canada at the end.

What do the trials of psilocybin for alcohol use disorder show?

Five published human studies matter most so far. Two are led by Michael Bogenschutz, one is from Zurich, one from a French inpatient programme and one is an open-label study from Denmark. The table puts them side by side. Figures come from each paper's abstract or open full text.

The 2015 pilot: ten people, no control group

The first modern test was a proof-of-concept study of ten volunteers with alcohol dependence, as defined in the older DSM-IV manual. Each had one or two supervised psilocybin sessions plus motivational enhancement therapy and sessions to prepare for and debrief from the psilocybin. Abstinence did not rise significantly during the first four weeks, before psilocybin. It rose after, and the gains were largely kept up to 36 weeks. The strength of the first session's effects predicted later drinking change. The authors reported no significant treatment-related adverse events and called for larger controlled trials (Bogenschutz et al., 2015). Evidence: preliminary. With ten people and no comparison group, it cannot separate psilocybin from therapy, expectation or natural change.

The 2022 randomized trial: 95 people versus an active placebo

The best-known trial is the double-blind randomized study published in JAMA Psychiatry in 2022, run at two US academic centres. Adults aged 25 to 65 with alcohol dependence were offered 12 weeks of manualized psychotherapy (motivational enhancement and cognitive behavioural therapy) and randomly assigned to psilocybin or diphenhydramine across two day-long medication sessions at weeks 4 and 8. Diphenhydramine was chosen as an active placebo: a drug with noticeable effects, so participants cannot easily tell they got nothing. People with major psychiatric disorders, other drug use disorders or current treatment for alcohol use disorder were excluded (Bogenschutz et al., 2022).

A total of 95 people were randomized (49 to psilocybin, 46 to diphenhydramine), and 93 received at least one dose and were analysed. The primary outcome was the percentage of heavy drinking days, recorded by a timeline followback interview over the 32 weeks after the first dose. It was 9.7% in the psilocybin group and 23.6% in the diphenhydramine group, a difference of 13.9 percentage points (95% CI 3.0 to 24.7; P = .01). Mean daily drinks were also lower with psilocybin (Bogenschutz et al., 2022). Evidence: this is the largest randomized result to date and it favoured psilocybin, but it has not been replicated at scale.

Both groups improved. The authors note that people on the active placebo, who got the same psychotherapy, cut their heavy drinking days by more than half relative to screening. The psilocybin effect was measured over and above that (Bogenschutz et al., 2022). The trial was funded by the Heffter Research Institute, among others, and the psilocybin was supplied by the Usona Institute and two university laboratories. The authors state the funders had no role in conducting the study.

Newer trials: a null result and a small positive pilot

A Swiss team randomized 37 people (18 psilocybin, 19 placebo analysed) who had recently finished inpatient withdrawal treatment. They received a single psilocybin or placebo (mannitol) session and brief psychotherapy. At the 4-week primary endpoint and at 6 months, the groups did not differ significantly in abstinence or alcohol use per day. Both groups reported less craving after dosing, with a bigger drop in the psilocybin group. There were 13 adverse events with psilocybin and 7 with placebo; one serious event with psilocybin and four with placebo, all linked to inpatient withdrawal treatment. The authors wrote that a single dose with five therapy sessions may not be enough to reduce relapse in severely affected patients, and that larger trials are needed (Rieser et al., 2025). Evidence: preliminary, and it points the other way from the 2022 trial. The two differ in many ways, including one session versus two, brief therapy versus 12 sessions, and people fresh from withdrawal treatment.

A French pilot randomized 30 inpatients with severe alcohol use disorder and depressive symptoms 2:1 to two sessions of psilocybin or a very low-dose control, as an add-on to standard care. Its main aim was feasibility, not proof of effect. At 12 weeks, 11 of 20 people in the psilocybin group were abstinent versus 1 of 9 in the control group (one control participant was lost to follow-up). Relapse was 35% versus 50%, a difference that was not statistically significant (Luquiens et al., 2025). Evidence: preliminary. Efficacy was secondary, the sample is tiny, and the control group's blinding problems (see below) matter.

A Danish open-label study gave 10 people a single psilocybin session with preparation and integration sessions. Heavy drinking days fell by 37.5 percentage points over the following 12 weeks, but with no comparison group the result cannot be attributed to the drug (Jensen et al., 2025). Secondary analyses of psilocybin-versus-diphenhydramine data also exist. One, in 84 adults with alcohol use disorder, reported shifts in personality-trait scores after psilocybin-assisted therapy compared with the active placebo (Pagni et al., 2025). Such analyses describe what else changed, not whether the treatment works.

Why are blinding and expectation such a problem in these trials?

A trial is blinded so that neither the participant nor the staff knows who got the drug. Psilocybin makes that hard, because its effects are obvious. After each medication session in the 2022 trial, participants correctly guessed their assignment in 93.6% of first sessions and 94.7% of second sessions. Study therapists guessed correctly 92.4% and 97.4% of the time. The authors wrote that diphenhydramine was ineffective in maintaining the blind and that biased expectancies could have influenced results (Bogenschutz et al., 2022). Researchers call this functional unblinding.

It recurs. In the Swiss trial, participants guessed correctly in 91.89% of cases and therapists in 97.30% (Rieser et al., 2025). In the French pilot, patients guessed correctly 93.3% of the time and investigators 86.7%; four control participants refused the second session after guessing they were in the control group (Luquiens et al., 2025). Evidence: established as a problem. What is unresolved is how much it inflates results. If people who know they got psilocybin expect to do well, some of the difference could come from that expectation.

How much of the effect is the therapy?

In every study above, psilocybin was given inside a programme of psychotherapy: 12 sessions of motivational enhancement and cognitive behavioural therapy in the 2022 trial, five brief sessions in the Swiss trial, an add-on to inpatient relapse-prevention care in the French pilot. In the 2022 trial both arms got the same therapy, so psilocybin's contribution is the difference between arms. But the authors point out that a two-group design cannot evaluate the effects of the psychotherapy itself, or how it interacts with the drug (Bogenschutz et al., 2022). Evidence: unresolved. Any result describes psilocybin plus a specific therapy package, in a trial setting.

What else limits the evidence so far?

• Small samples. The randomized trials had 95, 37 and 30 participants. Results from small trials can swing widely, and the Swiss and French trials point in different directions.
• Who was studied. The 2022 trial excluded people with major psychiatric disorders and drug use disorders and people in current treatment for alcohol use disorder, and its authors note drinking at entry was lower than in most medication trials, so results cannot be assumed to apply to more severe cases (Bogenschutz et al., 2022).
• Verification of drinking. Most outcomes are self-reported. In the 2022 trial, ethyl glucuronide (EtG) samples to check self-reports were available for only 53.8% of treated participants.
• Early stopping. The 2022 trial was designed for up to 180 people, but recruitment was halted at 95 by the 2020 COVID-19 suspension.
• Short follow-up. The 2022 trial followed people for 32 weeks of double-blind observation, and says nothing about effects beyond that, which matters for a condition with a relapsing course.
• Side effects. In the 2022 trial, headache was reported by 21 of 48 people given psilocybin (43.8%) and 2 of 45 given diphenhydramine; anxiety and nausea were also more frequent with psilocybin during sessions, and two participants given psilocybin received diazepam for anxiety in the second session. There were no serious adverse events in the psilocybin group; the three serious events were in the diphenhydramine group (Bogenschutz et al., 2022).

Are there newer or Canadian trials of psilocybin for alcohol use disorder?

Yes: several are registered. The table lists those we could read in the ClinicalTrials.gov registry on 4 October 2026. A registry entry is a plan, not a result, and sponsors do not always update status. One Canadian entry still shows a last-verified date of 2023. The list is not exhaustive and is not an invitation to enrol: eligibility and recruitment are decided by each trial, and we cannot confirm whether any site is currently accepting participants.

So Canadian sites are in the registry: eight in one multi-site entry, one in Calgary and one planned at CAMH in Toronto. The registry does not say what any of them will find. Until their results are published and peer reviewed, Canadian trials add no evidence to this article.

What does this research not show?

• It does not show that psilocybin treats alcohol use disorder. One randomized trial favoured it over an active placebo, and a second did not find a difference at 4 weeks or 6 months.
• It does not show that the benefit comes from the drug alone. Therapy was part of every study, and participants largely knew which arm they were in.
• It does not show how long any effect lasts, who it works for, or who it might harm. The 2022 trial, for example, excluded people with major psychiatric disorders, other drug use disorders and current treatment for alcohol use disorder.
• It does not show that results from supervised trials carry over to use outside them. These were research settings with screening, preparation and monitoring.
• It is not an approval. Health Canada's page, last modified 20 October 2026, says there are no approved therapeutic products containing psilocybin in Canada or elsewhere (Health Canada). Regulatory news changes; see our article on what COMP360 and the FDA mean for Canada.

Where can someone find help with alcohol use in Canada?

This list points to official pages we opened on 4 October 2026. It is not medical advice. If you are in immediate danger, call 9-1-1.

• Crisis: the Government of Canada says you can call or text 9-8-8, the Suicide Crisis Helpline, 24 hours a day, 7 days a week, and that the service is bilingual (Government of Canada).
• Ontario: ConnexOntario lists a free, 24/7 line at 1-866-531-2600 for mental health, addiction and problem gambling information. The Government of Canada's substance-use help page also lists it (Government of Canada).
• Other provinces: the same federal page lists lines such as BC's Alcohol & Drug Information Referral Service and the Alberta Addiction Helpline. Check it for your province (Government of Canada).
• General information on alcohol use disorder, withdrawal and usual treatments: CAMH's alcohol page. More numbers are on our safety page.

What we still don’t know

• Whether psilocybin reduces drinking in larger, independent trials (unresolved). Published results from the completed Danish trial, and from the registered Phase 3 trial, would add data.
• Whether one session works as well as two (unresolved). The Swiss trial used one, the 2022 trial two.
• How much of any effect is expectation, given broken blinding (unresolved).
• How long effects last past the follow-ups so far, which run to roughly 6 to 9 months (unresolved).
• Whether results hold in people with more severe drinking, other psychiatric conditions, or who were excluded from the trials (unresolved).

The research is real and moving fast, but it is still research. Check the date on this page and read our overview of what psilocybin is, or compare how the same questions play out in depression trials.

Shroomverse sells mushroom products. This article is general information about clinical research. It is not medical, legal or purchasing advice, and it does not recommend any product.