Therapeutic Use of LSD in Psychiatry: A Systematic Review of Randomized-Controlled Clinical Trials

Fuentes and colleagues situate lysergic acid diethylamide (LSD) within a long, interrupted history of psychiatric investigation: discovered in 1938, studied clinically from the late 1940s through the 1970s, then largely prohibited and neglected until a recent resurgence of interest. Earlier work used LSD for a range of psychiatric problems including anxiety, depression, psychosomatic disorders and addiction, but most of those studies predated contemporary methodological standards. The authors note LSD's pharmacology as a classical hallucinogen acting principally at serotonin 2A receptors, its typical subjective effects (altered time and self‑perception, visual phenomena, ego dissolution), and relevant safety considerations (challenging acute reactions, possible exacerbation of psychosis in vulnerable individuals, modest cardiovascular effects). They also introduce the key therapeutic variables historically emphasised in LSD research: dose, set (patients' mindset and expectations), and setting (physical and interpersonal environment), and three historical therapeutic approaches—psycholytic, psychedelic‑chemotherapy, and psychedelic‑peak therapy. This paper aims to identify and synthesise randomized controlled clinical trials (RCTs) that tested LSD as a therapeutic agent in psychiatric populations, extracting trial characteristics and variables that investigators controlled which might relate to outcomes. The stated objective is to inform discussion about integrating classical hallucinogen research into modern clinical trial designs and to provide guidance for future LSD research under contemporary standards.

The review followed PRISMA guidance. Searches were performed in PubMed and the MAPS Psychedelic Bibliography, covering publications from 1 January 1950 to 5 May 2019, and were supplemented by manual review of reference lists. Two authors independently screened titles and abstracts and retrieved full texts for potentially eligible reports; disagreements were resolved by discussion. The extracted text does not report whether additional databases beyond PubMed and MAPS were searched. Eligibility was restricted to randomized controlled clinical trials of LSD administered to patients diagnosed with mental illness; experimental studies in healthy volunteers and non‑randomized or uncontrolled designs were excluded. Data items extracted included dosage, regimen and route; participant characteristics (age, gender, diagnosis) and inclusion/exclusion criteria; country, trial design and size; length of follow‑up; comparator details (active and non‑active placebos included); outcome measures (change scores from baseline or endpoint across diverse diagnoses); and adverse events. Study quality was assessed with the Cochrane Collaboration risk of bias tool across six domains: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, and selective reporting/other bias. Two authors applied the tool independently with a third author resolving discrepancies. The extracted text indicates that risk assessments were summarised in a table, and that many trials had unclear or high risk in specific domains, particularly blinding and some other sources of bias.

The search yielded 3,668 PubMed records plus 12 additional records from other sources. After screening and exclusion, 11 randomized controlled trials met the inclusion criteria and were included in the review. Across these trials, LSD was administered to a total of 567 patients, with doses ranging from 20 to 800 mcg. The oral route was most commonly used; one study used intravenous administration and one did not specify the route. Most trials delivered a single LSD dose, though some used dose escalation or permitted repeat dosing at 2–3 week intervals. A few studies combined LSD with other pharmacological agents (for example, dextroamphetamine before LSD or antipsychotics after LSD), which the authors note could obscure LSD's therapeutic effects. Trials were geographically concentrated in North America and Europe: seven in the USA, three in Canada and one in Switzerland. Trial programmes and preparations varied widely: inpatient treatment durations ranged from 24 hours to 90 days; pre‑session preparation ranged from a few hours to five weeks; therapeutic approaches during sessions included active psychotherapy, supportive observation, and ‘psychedelic therapy’ sessions of 12–14 hours with continuous therapist/nurse support. Five trials used no‑drug control groups, several used usual treatment comparators, and many used active placebos; four studies used low‑dose LSD as an active comparator. The authors report consistent difficulty in maintaining blinding, noting that therapists could often correctly identify treatment due to observable psychoactive effects. Quality assessment findings showed that all trials used random assignment, but allocation concealment was often unclearly reported. Five trials had high risk of bias due to inadequate blinding of patients or staff; other common concerns included incomplete outcome data (two trials rated high risk) and selective reporting (three trials flagged). Four trials presented other sources of bias, such as baseline imbalances or unequal treatment exposure between groups. Efficacy outcomes varied by diagnostic category. Alcohol use disorder comprised the largest body of evidence: several trials reported within‑group improvements in abstinence and drinking behaviour, but between‑group differences were inconsistent. The review notes significant LSD effects in four studies, though some gains related more to quality of life and general health than to clear improvements in alcohol abstinence. One trial (Smart et al.) found substantial improvement across all three groups (high‑dose LSD 800 mcg, ephedrine active placebo, and no drug) with no significant between‑group differences. Other alcohol trials reported short‑term advantages for LSD at 2 months or 6 months in specific measures, and one later trial reported higher abstinence rates in the LSD group at three months and sustained superiority at one year versus one control group but not versus the usual treatment control. The authors summarise that LSD showed a significant effect in four alcoholism studies but the effects were heterogeneous across measures and follow‑up times. Other diagnostic findings included: a single eligible trial in heroin use disorder that reported higher total abstinence rates at 12 months for the LSD group; two trials addressing neurotic symptoms (anxiety, depression, psychosomatic presentations) that showed short‑term improvements for the LSD groups versus some comparators but little consistent superiority at six months; and one modern trial in patients with anxiety related to life‑threatening illness that reported significant reductions at two months in state anxiety (ANOVA p = 0.021) and trait anxiety (ANOVA p = 0.033), with non‑significant trends at 12 months. Safety data across the 567 treated patients identified two serious adverse events: a tonic–clonic seizure in a patient with prior seizure history and an episode of prolonged psychosis in a patient with a prior history of recurrent psychotic episodes; both recovered with treatment. No other serious adverse events were reported among the remaining 565 subjects. Common exclusion criteria across trials included severe organic disease, active psychosis, pregnancy and epilepsy. The authors also highlight procedural variables associated with therapeutic response: higher LSD dose and the occurrence of mystical or ‘peak’ experiences correlated with better global adjustment in some trials; musical accompaniment and more extensive preparatory work (set and setting) were associated with improved outcomes in modern trials. Aftercare practices varied substantially, ranging from brief discharge with friends' support to prolonged inpatient follow‑up and optional second LSD sessions in some studies.

Fuentes and colleagues interpret the body of randomized evidence as indicating LSD has therapeutic potential across several psychiatric domains, with the strongest and most consistent signal for alcohol use disorder. They stress that many trials showed important short‑term patient improvements, although in several studies treatment and control groups converged at longer follow‑up, while other trials reported benefits sustained to 6–12 months. The authors note that a previously published meta‑analysis (cited in their discussion) pooled six trials and reported an odds ratio of 1.96 (95% CI 1.36–2.84, p = 0.0003) favouring LSD over placebo for alcoholism; they further comment that a single LSD dose compared favourably with standard daily pharmacotherapies for alcohol dependence in that synthesis. A major theme is methodological heterogeneity and trial quality. The review emphasises inherent difficulties in blinding and placebo control for a profoundly psychoactive drug, and documents variable allocation concealment, incomplete outcome data and other biases in many trials. Differences in dosing, preparation, therapeutic approach, setting, integration/aftercare and concurrent pharmacotherapies are presented as likely contributors to divergent outcomes. The authors argue that set and setting, the quality of the acute psychedelic experience (notably mystical or ego‑dissolution phenomena), and post‑session integration are plausible mediators of therapeutic benefit; they reference available psychometric tools (MEQ‑30, Ego‑Dissolution Inventory) and recent neuroimaging literature that links LSD‑induced changes in brain connectivity with ego dissolution and subjective effects. The discussion acknowledges limitations of the review itself: inclusion of English‑language articles only, the predominance of older studies with heterogeneous designs, variable patient populations and diagnostic methods, and the consequent restricted strength of conclusions. At the same time, the authors position the review as a first systematic synthesis focused exclusively on randomized LSD trials, noting methodological rigor in study selection and the value of identifying trial variables that may predict outcome. They conclude by calling for modern, standardised randomized controlled trials that control for set, setting and integration, and that employ contemporary outcome measures and safety monitoring to better define LSD's therapeutic utility and address lingering stigma and regulatory barriers.

Fuentes and colleagues conclude that LSD demonstrates potential as a therapeutic agent in psychiatry, with the most robust evidence to date for alcohol use disorder. They emphasise the difficulty of conducting double‑blind trials with LSD given its conspicuous psychoactive effects, and call for new, methodologically sound studies under modern standards to strengthen the evidence base, reduce stigma and clarify LSD's clinical role.