Efficacy and Safety of Psilocybin in Treatment-Resistant Major Depression: The EPISODE Randomized Clinical Trial

Major depressive disorder affects around 280 million people worldwide and is a leading cause of disability. Treatment-resistant depression, usually defined as failure to respond to at least two antidepressant treatments, carries a particularly high personal and societal burden, creating a clear need for new treatment options. Psilocybin, a classical psychedelic and partial serotonin 2A receptor agonist, has shown promise for depression when combined with psychotherapy or psychological support. Evidence from open-label studies and randomised trials suggests potential benefit, and a recent meta-analysis reported a pooled effect size of 0.66. However, the evidence base remains limited by small samples, weak blinding, expectancy effects and incomplete safety monitoring. This study tested the efficacy and safety of a single 25 mg dose of psilocybin with adjunct psychotherapy in treatment-resistant depression, compared with 5 mg psilocybin and 100 mg nicotinamide as an active placebo. The design also included a second dosing session after the primary endpoint to reduce drop-out, limit disappointment and symptom worsening in comparator groups if unblinding occurred, and ensure that all participants received at least one 25 mg psilocybin dose for both ethical and methodological reasons.

This was a two-centre, triple-blind, Phase IIb randomized clinical trial conducted in Germany testing oral synthetic psilocybin with adjunct psychotherapy in treatment-resistant depression (TRD). Adults aged 25 to 65 years with moderate to severe TRD (HAMD17 ≥17) who had failed at least two adequate antidepressant treatments from different pharmacological classes were eligible. Participants were required to discontinue ongoing psychotherapy and monoaminergic medication before dosing (minimum abstinence 2 weeks, 5 weeks for fluoxetine). Recruitment used self-referral, clinician referral and public advertisement. Randomization used centre-stratified permuted blocks and allocated participants in a 2:2:1:1 ratio to four dosing sequences: (1) nicotinamide 100 mg then psilocybin 25 mg, (2) psilocybin 5 mg then psilocybin 25 mg, (3a) psilocybin 25 mg then 5 mg, and (3b) psilocybin 25 mg twice. Nicotinamide and a 5-mg psilocybin dose were intended as comparators to support blinding. All personnel, therapists, raters and participants were blinded to allocation. Each participant was assigned two therapists (one female, one male) and received a structured psychotherapeutic programme comprising seven 2-hour preparatory/integration sessions (14 hours total) surrounding two dosing sessions (week 0 and week 6). Dosing sessions lasted 6–8 hours with an overnight stay; participants also attended eight in-person visits and received eight weekly safety calls. Assessments included clinician-rated HAMD17, self-report BDI-II, laboratory tests, ECG, neuropsychological testing and functional MRI. The primary outcome was treatment response at week 6 (defined as ≥50% reduction from baseline in HAMD17), assessed one day before the second dose. Key secondary end points were BDI-II response and mean change from baseline on HAMD17 and BDI-II at week 6. Ratings were collected at baseline, day 1, week 1, week 6, week 7 and week 12. Safety assessments (vital signs, UKU, Columbia-Suicide Severity Rating Scale, open questioning, labs, ECG) were performed at each visit and during weekly calls; adverse events (AEs) were coded using MedDRA. The trial was powered to detect a 40% difference in HAMD17 response between psilocybin 25 mg and nicotinamide at week 6 (80% power), based on optimistic prior estimates. The primary efficacy analyses used a while-receiving-treatment population (randomized and treated participants with at least one postbaseline assessment before an intercurrent event). The primary binary outcome was analysed by logistic regression in a predefined fixed-order (sequential) testing procedure comparing psilocybin 25 mg vs nicotinamide and then vs 5 mg, adjusting for centre and baseline HAMD17. Continuous outcomes (HAMD17 and BDI-II scores) were analysed using mixed-effects linear regression models with treatment, time and treatment × time interactions, centre as a fixed effect and participant intercepts as random effects. Sensitivity analyses included a principal stratum approach. Analyses were conducted using R and SAS between April 2024 and November 2025. Of note, 143 participants received at least one investigational medicinal product and 142 were included in the primary efficacy analysis (psilocybin 25 mg n = 47; psilocybin 5 mg n = 48; nicotinamide n = 47).

Sample and analysis populations: A total of 144 participants were randomized and 143 received at least one dose; 142 were included in the primary efficacy analysis with group sizes of 47 (psilocybin 25 mg), 48 (psilocybin 5 mg) and 47 (nicotinamide). By week 12, 10 participants had reinitiated antidepressant medication and 4 withdrew for lack of efficacy or need for additional interventions. Primary outcome (binary HAMD17 response at week 6): The predefined hierarchical comparison of psilocybin 25 mg versus nicotinamide for the primary binary end point did not reach statistical significance. Reported response rates at week 6 were 17% for psilocybin 25 mg, 13% for psilocybin 5 mg and 11% for nicotinamide; the adjusted odds ratio for psilocybin 25 mg versus nicotinamide reported in the paper was non-significant, and because the first hierarchical comparison was nonsignificant, further formal testing in the sequential procedure was not performed. Continuous outcomes and key secondary measures: Mixed-effects models for continuous HAMD17 scores provided evidence of symptom reduction with psilocybin 25 mg. At week 6 the estimated mean difference in HAMD17 change from baseline for psilocybin 25 mg versus nicotinamide was −4.60 (95% CI −7.01 to −2.18; P < .001); an even larger difference was seen at week 1 (−4.89; 95% CI −7.31 to −2.48). Comparisons of psilocybin 25 mg versus 5 mg also favoured 25 mg at week 1 (−3.38; 95% CI −6.23 to −1.42) and at week 6 (−3.09; 95% CI −5.50 to −0.69; P = .02). BDI‑II results were consistent with the HAMD17 findings: a strong effect of psilocybin 25 mg versus nicotinamide was observed 1 day after treatment (estimated mean difference −5.84; 95% CI −10.51 to −1.17) that persisted to week 6 (−7.21; 95% CI −11.86 to −2.56; P < .001). At week 6 BDI-II response (≥50% reduction) occurred in 11 of 47 (23.4%) participants receiving psilocybin 25 mg versus 5 of 47 (10.6%) in the nicotinamide arm and 3 of 48 (6.3%) in the 5-mg arm; nominal odds ratios for BDI-II response were reported (psilocybin 25 mg vs nicotinamide OR 2.64; 95% CI 0.87–9.05; P = .049; psilocybin 25 mg vs 5 mg OR 4.61; 95% CI 1.19–17.83; P = .01), but these were exploratory given the predefined testing sequence. Treatment phase 2 (after the second dose) showed no between-group differences in HAMD17 or BDI-II response rates; all groups demonstrated clinically relevant within-participant improvements by week 12. The pooled estimated mean change from baseline at week 12 across groups was −7.50 (95% CI −8.69 to −6.31) on HAMD17 and −10.53 (95% CI −12.48 to −8.58) on BDI-II, corresponding to substantial symptom reduction across arms when most participants had received at least one 25-mg dose. Post hoc subgroup analyses suggested smaller HAMD17 decreases after psilocybin 25 mg in females, similar effects regardless of antidepressant-withdrawal status, and larger reductions in those with more severe baseline depression. Functional unblinding assessments indicated many participants correctly identified the high-dose psilocybin condition. Safety: Across 280 investigational product administrations (143 individuals), adverse events were common and primarily acute on dosing days. The paper reports AEs in 160 cases (100%) after psilocybin 25 mg, 66 cases (92%) after psilocybin 5 mg and 34 cases (71%) after nicotinamide; most AEs occurred on dosing days. Perceptual and experience-related AEs after psilocybin 25 mg included pseudohallucination (21%), paresthesias/dysesthesias (14%), synesthesia (13%), paranoia (3.8%) and rare occurrences of hallucination, thought disorder or brief psychosis-like states; these events were transient and resolved the same day. Suicidal ideation on dosing day (day 0) was reported in 6 participants (4%) after psilocybin 25 mg, 1 participant (1%) after psilocybin 5 mg and 1 participant (2%) after nicotinamide. Severe AEs on dosing days were more frequent after psilocybin 25 mg (44 of 160 cases, 28%) than after comparators; from day 1 onwards severe AEs were rare. There were four serious adverse events in total, two of which were judged related to psilocybin 25 mg, including one case of hallucinogen persisting perception disorder with broader psychological destabilisation. Suicidal and nonsuicidal self-injurious behaviours were infrequent and worsening of suicidal ideation from baseline was reported as comparable across treatment arms.

The authors interpret the trial as inconclusive: although the primary binary endpoint (≥50% HAMD17 reduction at week 6) did not show a statistically significant benefit of psilocybin 25 mg versus nicotinamide, continuous measures of depressive symptoms (HAMD17 and BDI-II change from baseline) provided evidence of clinically meaningful reductions with psilocybin 25 mg at week 6. They highlight a divergence between the null primary outcome and positive key secondary continuous outcomes, calling for cautious interpretation. The investigators note that antidepressant effects peaked early (around 1 week) with higher response at week 1 for the 25-mg dose, and that no additional between-group benefit was observed after the second dose; by week 12, when most participants had received high-dose psilocybin at least once, symptom reductions were substantial across groups. The authors discuss possible explanations for the low primary response rates and divergence from prior trials: overestimation of treatment effects in power calculations (no prespecified minimum important difference), the chronic and comorbid nature of the TRD sample (including personality disorder traits), strict medication-free status compared with some prior studies, potential effects of expectancy and functional unblinding, and the choice of a 6-week primary time point (whereas prior studies often used earlier endpoints). They also note that the HAMD17 may emphasise somatic symptoms and be less sensitive to psilocybin’s emotional and cognitive effects. On safety, the authors report most AEs were mild to moderate and occurred acutely on dosing days, but they emphasise safety signals including increased suicidal ideation on dosing days after 25 mg and one instance of persisting perceptual disturbance. They attribute higher AE rates than earlier reports at least partly to more systematic and sensitive AE assessment capturing perceptual and emotional changes associated with psychedelic experiences. The authors argue these findings highlight the need for standardised risk monitoring and pharmacovigilance and reinforce that psilocybin should be administered within a psychotherapeutic framework. Key limitations acknowledged by the authors include functional unblinding (most participants correctly identified the high-dose condition), an unexplained small centre effect, overestimation of effect size in the power calculation (no predefined MID), a socioeconomically and ethnically homogeneous sample limiting generalisability, exclusion of participants at high suicide risk, lack of follow-up beyond 12 weeks, possible early underreporting of acute AEs, and missing adherence and therapy-quality ratings. They suggest future confirmatory studies should prespecify a minimum important difference and be adequately powered, include central independent raters, longer follow-up, multiple centres, adherence measures within standardised psychotherapy protocols and further exploration of dosing regimens and psychotherapeutic strategies to prolong effect duration. The authors position their results amid prior research by noting that while the trial did not meet its primary endpoint, the continuous-outcome findings and rapid early antidepressant effects align with previous signals of psilocybin efficacy; nonetheless, the divergence between endpoints and the safety observations mean the findings should be seen as hypothesis generating and warrant replication in larger, well powered trials.

The authors conclude that in this randomized trial in patients with treatment-resistant depression, psilocybin 25 mg with adjunct psychotherapy did not increase HAMD17-defined response rates at week 6 (primary end point) but did produce greater reductions in depressive symptoms on continuous measures at week 6 (key secondary end point). Because primary and key secondary outcomes diverged, the overall findings are inconclusive and require cautious interpretation. No clear additional benefit of a second 25-mg administration was observed within the study timeframe; by week 12 mean symptom reductions were substantial across groups after most participants had received at least one high dose. Psilocybin was generally tolerated but associated with acute adverse events and safety signals, including suicidal ideation on dosing days and one case of persisting perceptual disturbance. The authors state that larger, adequately powered confirmatory trials with longer follow-up are needed to clarify efficacy, durability and mechanisms of action.