The 2012 Meta-Analysis of LSD Treatment for Alcohol Misuse helped turn a disputed mid-century treatment into a renewed clinical question. By combining six of the best randomized controlled studies conducted in the 1960s and 1970s, it found a significant benefit from a single dose of LSD, lasting as long as six months. The result was important not because it settled every question about psychedelic therapy, but because it showed that a promising signal had survived a record long clouded by poor controls, inconsistent practice, and institutional neglect.

A treatment with a longer history

Psychedelic approaches to alcohol misuse did not begin with modern clinical research. Native Americans had long used peyote as both sacrament and treatment for alcoholism, and in the 1950s and 1960s thousands of people were treated with LSD and other psychedelics. At the time, the proposed treatment drew on the possibility that a powerful experience might interrupt habitual drinking or prompt a lasting change in a patient’s outlook. The evidence, however, was difficult to interpret: clinical reports could be enthusiastic, while many formal studies were poorly designed or inadequately controlled.

The treatment became established in Saskatchewan, where the provincial government helped make LSD therapy a standard option for alcoholism. That policy reflected the confidence generated by the Saskatchewan work of Humphry Osmond and Abram Hoffer, but it did not end disagreement about whether the results were credible. In the early 1960s, the Addiction Research Foundation in Toronto set out to repeat the trials under more rigorous controls, hoping to separate the drug’s effects from other influences on treatment.

How the treatment conditions shaped the evidence

The Toronto replication deliberately altered the circumstances of the sessions. Clinicians placed patients in neutral rooms, told them not to engage with the participants during their trips except to administer an extensive questionnaire, and in some cases restrained or blindfolded them. The outcomes failed to match those reported in Saskatchewan; several volunteers had terrifying experiences. Critics took the replication as evidence that LSD treatment did not work reliably under controlled conditions. Its supporters saw a different lesson: the contrast showed how much the therapist’s presence and the surrounding environment might matter.

Osmond and Hoffer had learned from volunteers that the session environment could shape the experience, and that an engaged, empathetic therapist could help prevent a difficult session. They also suspected that some psychotic reactions might arise partly from the sterile clinical atmosphere itself. The later language of set and setting was not yet in use in their work, but the principle was already emerging: a drug effect could not be understood apart from the person taking it, the expectations brought to the session, and the setting in which the experience unfolded.

The meta-analysis helped clarify the record without erasing those complications. Rather than treating every past report as equally persuasive, it combined data from the six strongest randomized controlled studies, covering more than 500 patients. Across that evidence base, one LSD dose had a statistically robust and clinically significant beneficial effect on alcohol misuse, with benefits lasting up to six months. The result gave the earlier Saskatchewan findings support beyond the most enthusiastic individual reports, even as the divergent trials remained a reminder that study conditions and therapeutic practice mattered.

A finding returned to clinical view

The review’s publication made the prior disappearance of LSD treatment harder to explain. Its authors noted the evidence for benefit and questioned why the treatment had been largely overlooked. That question carried particular force because LSD therapy had once been regarded as effective enough to become standard treatment in Saskatchewan, yet the uneven quality of the historical literature had left later readers unsure what the research as a whole showed. The meta-analysis did not make that history straightforward; it made the neglected evidence legible enough to warrant renewed investigation.

Afterward, psychedelic treatment for alcohol and other addictions underwent a modest revival in both university studies and underground practice. In a 2015 pilot at the University of New Mexico, ten people with alcoholism received psilocybin alongside motivational enhancement therapy, a form of cognitive behavioral treatment designed for addiction. Psychotherapy alone had little effect on drinking behavior, but drinking decreased significantly after the psilocybin session, and the changes continued through thirty-six weeks of follow-up. The lead investigator, Michael Bogenschutz, reported a strong correlation between the strength of the experience and its effect on drinking. Encouraged by those results, Bogenschutz began a larger phase 2 trial involving 180 volunteers at NYU with Stephen Ross and Jeffrey Guss.

Promise and limits

The broader return of psychedelic research brought renewed attention to the therapeutic session as a whole, not simply to the drug dose. Practical guidance for participants and guides addresses preparation, possible drug interactions, intentions, difficult experiences, and post-session integration; clinical protocols likewise make the session’s structure and the presence of trained sitters part of the treatment. Such practices do not by themselves establish efficacy, but they reflect the historical lesson of the alcohol trials: outcomes can vary when researchers change the relational and environmental conditions around the drug.

The evidence also calls for restraint. The LSD meta-analysis drew on the best available randomized studies, but the historical literature had substantial methodological weaknesses, and the later revival initially rested on small trials. More generally, early psychedelic studies can be affected by small samples, selective recruitment, unusually attentive therapists, and strong expectations; a promising signal is not the same as definitive proof of cure. The meta-analysis therefore occupies a careful middle ground: it gives a previously muddled treatment history a measurable positive result, while leaving open how reliably that benefit can be reproduced, for whom it works, and which features of the drug experience and its setting are essential.