Efficacy, tolerability, and safety of serotonergic psychedelics for the management of mood, anxiety, and substance-use disorders: a systematic review of systematic reviews

Dos Santos and colleagues followed PRISMA guidance and an established guide for identifying and appraising systematic reviews. They searched PubMed to 11 April 2018 using terms for LSD, mescaline, DMT, psilocybin and ayahuasca combined with "systematic review" or "meta-analysis", with no language restriction, and also hand-searched reference lists of retrieved papers. Inclusion criteria required systematic reviews (with or without meta-analysis) that reported at least one RCT evaluating serotonergic psychedelics for mood, anxiety or substance-use disorders; systematic reviews of observational studies were also eligible when focused on adverse reactions. Reviews published only as abstracts, those not focused on the target topics, or those citing fewer than one RCT were excluded. Screening of titles, abstracts and full texts was performed by two independent reviewers, with a third reviewer used to resolve disagreements. Extracted variables comprised bibliographic details, type of review (RCTs/observational/mixed), counts of included studies and patients, target disorder, psychedelic type, risk-of-bias/quality features and main findings (therapeutic or adverse). For reviews that conducted meta-analyses, reported effect measures (odds ratios), heterogeneity (I2) and p values were recorded. Methodological quality of the included systematic reviews was assessed using a six-item modified AMSTAR tool; two independent reviewers performed this assessment with consensus resolution procedures. The search initially identified 531 records, 14 of which were potentially relevant after abstract screening, and 10 systematic reviews met inclusion criteria (one meta-analysis and nine descriptive systematic reviews).

The authors included 10 systematic reviews: one meta-analysis (focused on LSD) and nine descriptive reviews. Most reviews were descriptive and incorporated only a small number of RCTs and, in several cases, observational or open-label studies. The extracted evidence consistently suggested that single or a few controlled administrations of serotonergic psychedelics are associated with rapid and, in some reports, sustained anxiolytic and antidepressant effects; evidence for antiaddictive effects was more variable and often limited by study design. Key efficacy findings reported across the included reviews were: three RCTs of psilocybin (total N = 92) showed reductions in anxiety and depressive symptoms in patients with advanced-stage cancer or other life‑threatening illness; LSD data included a meta-analysis of six RCTs (combined N = 536) that reported decreases in alcohol misuse and increased abstinence in alcoholic patients; single small RCTs and open-label studies reported anxiolytic and mood improvements with LSD in advanced cancer populations (examples cited include one RCT with n = 12 and open-label data). Ayahuasca/ DMT evidence included one RCT in healthy experienced users (n = 9) and an open-label clinical trial in treatment‑resistant major depressive disorder (n = 17) showing rapid antidepressant and anxiolytic effects after a single dose; the authors also cite a later RCT with 29 patients reporting a rapid (24 h) and sustained (7 days) antidepressant effect versus placebo. Small open-label studies further suggested improvements in tobacco and alcohol dependence after psilocybin (examples: n = 15 for tobacco, n = 10 for alcohol in the cited reports). Additional findings included a dose‑escalation, double‑blind study in nine patients with obsessive‑compulsive disorder that reported symptom reductions, and several reports (observational and experimental) of changes in personality measures such as increased openness and reduced novelty-seeking or impulsivity in ritual ayahuasca users, although the evidence was mixed and inconsistent. Safety and tolerability data from the reviews indicated that, in controlled experimental and clinical settings, single or few doses of these psychedelics were generally well tolerated. The most common adverse events in trials were transient dysphoric/anxiety or psychotic‑like reactions, transient moderate increases in blood pressure and heart rate, nausea, vomiting and other gastrointestinal effects (particularly with ayahuasca). Prolonged psychotic reactions were reported in some case reports and observational (often recreational or ritual) contexts, but none were described in the controlled RCTs reviewed for the preceding 25 years. The authors emphasise that psychiatric screening to exclude individuals with a personal or family history of psychotic illness appears to reduce the risk of severe adverse psychiatric events in trials. Concerning methodological quality, the mean modified AMSTAR score across the 10 included systematic reviews was 4.4 (range 3 to 5), indicating moderate quality overall. Common methodological shortcomings were the frequent absence of quality assessment of included studies (reported in nine reviews), occasional lack of clear conflict‑of‑interest declarations, and in some reviews unclear reporting about whether study selection and data extraction were performed by dual reviewers. The authors also note heterogeneity in older LSD studies (conducted mainly between 1966 and 1970) with variable diagnostic criteria, psychometric measures and therapeutic techniques, which may limit comparability with more recent trials.

Dos Santos and colleagues interpret the aggregated evidence as cautiously optimistic: serotonergic psychedelics show anxiolytic, antidepressant and antiaddictive potential and, when administered in controlled settings with appropriate screening, have a generally favourable short‑term safety and tolerability profile. They highlight stronger levels of evidence in two areas: LSD for alcohol dependence (the LSD meta‑analysis of six RCTs) and psilocybin for existential anxiety and depression in patients with advanced cancer (three RCTs). However, the authors repeatedly stress that the literature is limited — most included reviews contained few RCTs, sample sizes were small, follow‑up durations short, and many primary studies were open‑label or observational rather than controlled. The discussion draws attention to several important limitations identified by the reviewers: older LSD trials differ in era‑specific diagnostic and measurement standards and varied in dose and therapeutic techniques, reducing comparability; many positive signals arise from observational or open‑label designs vulnerable to bias; and routine exclusion of individuals with personal or family histories of psychosis in trials limits generalisability to typical psychiatric clinic populations. The authors also note that while ritual ayahuasca users have not shown consistent increases in psychopathology in observational studies, those studies may be subject to selection bias and prospective data on new users are needed. For future research, the paper argues for more rigorous Phase II and Phase III RCTs with larger, multicentre samples, longer follow-up and varied dosing schemes, and for trials that investigate the contribution of psychotherapeutic approaches and contextual factors (for example music) to outcomes. The authors indicate multiple ongoing or planned RCTs (registered clinical trials) at the time of writing and discuss regulatory and ethical considerations: potential re-examination of drug scheduling in some countries, and respect for the botanical and cultural patrimony surrounding ethnobotanical substances such as ayahuasca. Finally, they recommend extending investigation to other disorders (for example PTSD, obsessive–compulsive and personality disorders) and exploring non‑psychiatric applications such as creativity or meditative practices, as proposed by the authors.

The authors conclude that serotonergic hallucinogens/psychedelics (ayahuasca/DMT, psilocybin and LSD) have shown anxiolytic, antidepressant and antiaddictive effects in experimental, open‑label and, to a lesser extent, randomised controlled trials. They caution that controlled‑trial evidence remains limited by small sample sizes, short durations and heterogeneous methods, so clinical evidence is still preliminary. Notwithstanding these limitations, single or few doses have produced rapid and sometimes sustained therapeutic effects, and controlled studies report a generally good short‑term safety and tolerability profile when participants are carefully screened. Dos Santos and colleagues call for larger, longer Phase II/III RCTs to establish efficacy and safety more conclusively and note that clinical implementation will need to address screening exclusions (notably psychotic histories) and ethical/regulatory issues, particularly in relation to plant‑based substances with cultural significance.