Efficacy and safety of psilocybin-assisted treatment for major depressive disorder: Prospective 12-month follow-up

Major depressive disorder (MDD) is highly prevalent, recurrent, and a major cause of disability; first-line treatments often take weeks to work and many patients fail to achieve durable remission. Earlier clinical studies have reported rapid antidepressant effects after one or two doses of psilocybin administered alongside psychotherapy, but follow-up intervals in those trials were short (up to 6 months) and evidence about longer-term efficacy and safety in MDD remains limited. A recent double-blind trial comparing high-dose psilocybin with standard antidepressant treatment did not find a significant difference on its primary outcome at 6 weeks, underscoring uncertainty about comparative and sustained benefits. Davis and colleagues set out to evaluate the durability of antidepressant effects and safety of psilocybin-assisted treatment across a 12-month follow-up in adults with moderate to severe unipolar MDD. Using a randomised waitlist-controlled design in which participants received two supervised psilocybin sessions with supportive psychotherapy, the study aimed to measure clinician- and self-rated depression severity over one year, to examine whether features of the acute psilocybin experience predicted long-term outcomes, and to monitor adverse events and other safety signals.

This was a randomised, waiting-list controlled trial carried out under institutional review board approval with written informed consent. Eligible participants were adults aged 21–75, medically stable, and experiencing a moderate to severe major depressive episode defined by a GRID-Hamilton Depression Rating Scale (GRID-HAMD) score of ⩾17 assessed by blinded clinician raters. People with personal or first- or second-degree family histories of psychotic or bipolar I/II disorders were excluded. Participants were required to discontinue psychoactive medications at least five half-lives before screening and to refrain from such medications for at least one month after the second psilocybin session. After baseline assessments and screening, 27 participants were randomised to immediate or delayed (8-week) treatment; 24 completed both psilocybin sessions and all follow-up visits. Preparatory psychotherapy comprised 6–8 hours with two facilitators (at least one with master’s or doctoral clinical training). Psilocybin was administered in a supervised, comfortable setting using a nondirective therapeutic stance. Dosing consisted of two oral administrations approximately two weeks apart: 20 mg/70 kg for the first session and 30 mg/70 kg for the second. Follow-up assessments occurred at 1 day and 1 week after each session and then at 1 (4 weeks), 3, 6, and 12 months after the second session; functional MRI was obtained at baseline and one week after the second session. The pre-specified primary outcome was clinician-rated depression severity on the GRID-HAMD, assessed by blinded raters. Self-report measures included the Quick Inventory of Depressive Symptoms (QIDS) and the Beck Depression Inventory-II (BDI-II). Definitions of treatment response (⩾50% reduction from baseline) and remission (GRID-HAMD ⩽7; QIDS ⩽5; BDI-II ⩽9) were used. Acute-session measures included the Mystical Experience Questionnaire (MEQ30) and single-item ratings of personal meaning, spiritual significance, psychological insight and psychological challenge; the Persisting Effects Questionnaire provided an overall well-being score at follow-up (grand mean of five positive-change subscales expressed as percentage of maximum). Safety monitoring comprised adverse-event recording at each visit, the Columbia Suicide Severity Rating Scale for suicidal ideation, and direct solicitation for hallucinogen persisting perceptual disorder (HPPD) symptoms. Analyses began with a repeated-measures ANOVA on GRID-HAMD scores with time and condition (immediate vs delayed) as factors; because there was a significant effect of time but no condition or time-by-condition interaction, data were pooled across groups. Paired t tests (Bonferroni-adjusted) compared baseline with each follow-up timepoint and Cohen’s d quantified effect sizes. Correlations between acute-session measures and follow-up outcomes used Spearman’s rho, taking each participant’s highest session score. Categorical comparisons used χ2 and continuous comparisons used t tests; statistical analyses were performed in SPSS (versions 26 and 27).

Twenty-seven participants were randomised and 24 completed both psilocybin sessions and all follow-up visits through 12 months. The analysed sample was 67% female and 92% White, with a mean age of 39.8 (SD 12.2) years. Participants had chronic illness (mean years since diagnosis 21.5) and long current episodes (mean 24.4 months). Prior antidepressant treatment had been attempted by 88% and 58% reported antidepressant use during the current episode; 25% reported past psychedelic use. The repeated-measures ANOVA on GRID-HAMD showed a significant effect of time (F(4.4, 96.3) = 34.9, p < 0.001, ηp2 = .61) but no effect of treatment condition or interaction, so results were collapsed across arms. Mean GRID-HAMD decreased from 22.8 (SD 3.9) at baseline to 8.7 (7.6) at 1 week, 8.9 (7.4) at 4 weeks, 9.3 (8.8) at 3 months, 7.0 (7.7) at 6 months, and 7.7 (7.9) at 12 months; all follow-up timepoints differed from baseline (paired t tests, p < 0.001). Effect sizes were very large: Cohen d (95% CI) = 2.3 (1.5–3.1) at 1 week, 2.3 (1.5–3.1) at 4 weeks, 2.0 (1.3–2.7) at 3 months, 2.6 (1.7–3.4) at 6 months, and 2.4 (1.6–3.2) at 12 months. Self-reported depression (QIDS and BDI-II) showed similar reductions and sustained effects (reported in supplement tables and figures). Clinically meaningful change was sustained: 17 of 24 participants (71%) met criteria for response (⩾50% reduction) and 14 (58%) were in remission (GRID-HAMD ⩽7) at 1 week; at 12 months overall response was 75% and remission 58%. Three participants (13%) never met response criteria at any post-treatment timepoint. Use of daily antidepressant medication increased over follow-up: reported daily use was 0% at 4 weeks, 12.5% (3/24) at 3 months, 20.8% (5/24) at 6 months, and 33.3% (8/24) at 12 months. Participants who began antidepressants during follow-up had higher baseline GRID-HAMD (mean 25.3 vs 21.5, p = 0.02) but, at 12 months, did not differ significantly in GRID-HAMD or overall well-being scores from those who did not start medication. Acute-session measures showed differential predictive patterns. MEQ30 scores correlated significantly with overall well-being at all long-term follow-ups, and single-item ratings of personal meaning and spiritual significance correlated with well-being at most timepoints. Ratings of personal meaning and spiritual significance correlated with improvement in GRID-HAMD at the first long-term follow-up (week 4) only; beyond that, session measures did not significantly predict depression change. Psychological challenge during sessions did not correlate with subsequent well-being or depression outcomes. Safety monitoring through 12 months found no serious adverse events judged related to psilocybin, low levels of suicidal ideation, no self-injurious behaviour, no reported illicit psilocybin or other psychedelic use, and no participants meeting criteria for HPPD. Further adverse-event detail was cited as available in the supplement.

Davis and colleagues interpret their findings as evidence that two supervised doses of psilocybin given with supportive psychotherapy produced large, rapid reductions in depression that were maintained across 12 months in this sample of moderate to severe MDD. Blinded clinician ratings (GRID-HAMD) and participant-reported measures showed large effect sizes and sustained response and remission rates (75% response, 58% remission at 12 months). The authors note these results extend prior trials with shorter follow-up and that the observed magnitude and persistence of effect at 6 and 12 months exceeded results reported in some earlier studies, though the reasons for such differences (participant characteristics or procedural factors) are unknown. The study team also reports that subjective features of the acute experience—particularly mystical-type experience (MEQ30) and ratings of personal meaning and spiritual significance—were associated with enduring increases in participant-attributed well-being, but except at the first long-term assessment these session measures did not predict depression change. The authors acknowledge limited power as one possible reason for the absence of a stronger predictive relationship with depressive symptoms and note that prior work in cancer-related samples did find such associations. Safety findings were favourable over the 12-month period, with no serious adverse events related to psilocybin, no HPPD, and no observed increases in illicit psychedelic use in this cohort. Nonetheless, the investigators stress that larger Phase III and IV studies are required to characterise safety more fully. Key limitations the authors acknowledge include the small, predominantly White non-Hispanic sample, exclusion of individuals at elevated suicide risk, and the lack of an active comparator during the long-term follow-up because of the waitlist design. They also note that expectancy, psychotherapy and placebo effects—highlighted by other work using double-blind designs—cannot be ruled out as contributors to the observed outcomes. The fact that 33% of participants initiated antidepressant medication during follow-up complicates attribution of long-term effects solely to psilocybin, although measures of persisting well-being attributed to psilocybin did not differ between those who did or did not later use antidepressants. In terms of clinical implications discussed by the authors, psilocybin is compared with ketamine: both have rapid antidepressant effects, but psilocybin here showed sustained effects after only two administrations, which the authors suggest may be an advantage over treatments requiring repeated dosing. They caution that questions remain about abuse potential, optimal dosing strategies including repeat administration for relapsing patients, and the need for larger, controlled trials in more diverse populations.

The study team concludes that psilocybin-assisted treatment for moderate to severe MDD produced large and stable antidepressant effects that persisted for at least 12 months in this cohort of participants who were unblinded to treatment. They emphasise that these results demonstrate larger and longer-duration effects than previous psilocybin studies in depressed patients but recommend further research with active or placebo-controlled designs in larger, more diverse samples to confirm efficacy and fully characterise safety.