Effects of Psilocybin-Assisted Therapy on Major Depressive Disorder

Major depressive disorder (MDD) is described as a leading global cause of disability with substantial personal and economic burden and limitations in current pharmacotherapies, including incomplete response, adverse effects, and adherence problems. Earlier research highlighted rapid but short-lived antidepressant effects of ketamine and preliminary evidence that one or two administrations of psilocybin, given with psychological support, produced antidepressant effects in patients with cancer-related distress and in small samples of treatment-resistant depression. Psilocybin was noted to act via serotonergic and possibly glutamatergic pathways and to have lower addiction liability and less long-term toxicity than ketamine, motivating further investigation in broader MDD samples. Davis and colleagues designed the present study to test whether psilocybin-assisted therapy reduces depressive symptoms in adults with moderate to severe MDD. The trial used a randomised, waiting list–controlled design to compare immediate administration of two psilocybin sessions (with psychological support) against an 8-week delayed treatment control, with clinician-blinded assessment of depression severity as the primary outcome.

This single-centre randomised clinical trial was conducted at the Center for Psychedelic and Consciousness Research in Baltimore. Eligible adults were aged 21–75 years, met DSM-5 criteria for current moderate or severe MDD (GRID-Hamilton Depression Rating Scale [GRID-HAMD] score ≥17), and were not using antidepressant medications at screening. Key exclusions included personal or first-/second-degree family history of psychotic or bipolar disorders, recent moderate-to-severe substance use disorders, substantial prior use of ketamine or classic hallucinogens, unstable medical conditions, and pregnancy or nursing in women. Recruitment proceeded from August 2017 to April 2019; 870 people were screened by phone or online, 70 underwent in-person screening, and 27 were randomised. Participants were randomised after baseline assessment using urn randomisation to balance sex, age, baseline depression severity, and treatment-resistance level; 2 participants were allocated to the immediate treatment group and the delayed group using this method (the extracted text reports final allocation of 13 and 11 completers respectively). The immediate treatment period lasted 8 weeks and included preparatory meetings (about 8 hours total) with two facilitators, two daylong psilocybin administration sessions a mean 1.6 weeks apart (session 1: 20 mg/70 kg; session 2: 30 mg/70 kg, given in opaque capsules with water), and follow-up integration meetings (2–3 hours total). Facilitators came from a range of professional backgrounds and remained present during sessions while participants were encouraged to lie down, wear eyeshades and headphones, and focus inward with a music playlist. The delayed treatment (waiting list) group underwent weekly safety monitoring by brief telephone or in-person check-ins during the 8-week delay; GRID-HAMD assessments were obtained at weeks 5 and 8 before they received the active intervention. The primary outcome was GRID-HAMD score assessed by clinician raters blinded to condition at baseline and at the predefined postrandomisation time points that corresponded to week 1 and week 4 after intervention in the immediate-treatment group (these were weeks 5 and 8 for the delayed group). Secondary symptom measures included the Quick Inventory of Depressive Symptomatology–Self Report (QIDS-SR), Beck Depression Inventory II, Patient Health Questionnaire-9, clinician-rated Hamilton Anxiety Rating Scale, State-Trait Anxiety Index, and the Columbia–Suicide Severity Rating Scale. Blood pressure and heart rate were monitored during sessions. Interrater reliability for depression assessments through postsession week 4 was reported as 85%. Analyses were conducted on the evaluable population (participants who completed the intervention), with data analysis dated July 1, 2019 to July 31, 2020. The primary between-group test used a repeated-measures analysis of variance with time and condition factors; follow-up independent samples t tests compared corresponding time points between groups. Within-participant changes and rapid effects were examined with paired t tests on GRID-HAMD and QIDS-SR, with effect sizes reported as Cohen's d for t tests and partial eta squared for ANOVA. The trial sample size was justified by prior psilocybin data suggesting a large effect and consequent high power with approximately 24 participants. All tests used a two-sided P < .05 threshold.

Twenty-seven participants were randomised and 24 (89%) completed the intervention and postsession assessments; the evaluable sample comprised 13 participants in the immediate treatment condition and 11 in the delayed condition. Among completers the mean (SD) age was 39.8 (12.2) years, 16 were women (67%), and the mean (SD) baseline GRID-HAMD score was 22.8 (3.9), in the moderate-to-severe range. There were no statistically significant baseline differences between conditions on stratification variables reported in the extracted text. A significant time-by-condition interaction on GRID-HAMD scores was observed (partial ηp2 = 0.57; 90% CI, 0.38–0.66; P < .001). Mean (SD) GRID-HAMD scores in the immediate treatment group changed from 22.9 (3.6) at baseline to 8.0 (7.1) at week 5 and 8.5 (5.7) at week 8 (these correspond to weeks 1 and 4 after intervention). The delayed treatment group, before receiving psilocybin, had mean (SD) GRID-HAMD scores of 22.5 (4.4) at baseline, 23.8 (5.4) at week 5, and 23.5 (6.0) at week 8. Between-group effect sizes were large: Cohen d = 2.2 (95% CI, 1.4–3.0; P < .001) at week 5 and Cohen d = 2.6 (95% CI, 1.7–3.6; P < .001) at week 8. Within-participant comparisons among completers showed large reductions from baseline to postsession follow-ups (baseline to week 1: Cohen d = 3.6; 95% CI, 2.2–5.0; P < .001; baseline to week 4: Cohen d = 3.6; 95% CI, 2.2–4.9; P < .001). Rapid change on the QIDS-SR occurred by day 1 after the first psilocybin session, falling from mean (SD) 16.7 (3.5) at baseline to 6.3 (4.4) at day 1 (Cohen d = 3.0; 95% CI, 1.9–4.0; P < .001) and remaining reduced at week 4 after the second session (6.0 [5.7]; Cohen d = 3.1; 95% CI, 1.9–4.2; P < .001). Clinically significant response (≥50% reduction in GRID-HAMD) was observed in 16 participants (67%) at week 1 and 17 (71%) at week 4, while remission (GRID-HAMD ≤7) occurred in 14 participants (58%) at week 1 and 13 (54%) at week 4. Secondary depression and anxiety measures showed similar patterns of improvement after entry into the active treatment period. Suicidal ideation was reported as low overall and trended lower after enrollment. Safety data indicated no serious adverse events. One session involved a transient blood pressure elevation beyond the protocol threshold (diastolic >100 mm Hg) that resolved without medical intervention. Nonserious adverse events included challenging emotional and physical experiences during sessions, transient headache reported during 33% of sessions and after 29% of sessions, and other events detailed in supplementary tables referenced by the authors. Initiation of antidepressants or psychotherapy after the intervention was documented in supplementary material but specific counts are not clearly reported in the extracted text.

Davis and colleagues interpret their findings as evidence that two administrations of psilocybin given with supportive psychotherapy produced rapid, large, and sustained antidepressant effects in adults with moderate to severe MDD. The authors contrast psilocybin with ketamine by noting that ketamine's antidepressant effects are typically brief (days to about 2 weeks) whereas the current study documented clinically significant response that persisted at least through the 4-week follow-up after the second session, with 71% of participants still meeting response criteria at that time. Psilocybin was further described as having lower addiction potential and a comparatively mild adverse-event profile in this trial, which the authors suggest could confer therapeutic advantages relative to some existing pharmacotherapies. The trial's results were presented as consistent with prior reports of antidepressant effects of psilocybin in patients with cancer-related distress and in open-label studies of treatment-resistant depression. The authors also highlight associations observed between intensity of mystical-type, personally meaningful, or insightful experiences occasioned by psilocybin and reductions in depressive symptoms, and they reference emerging evidence that psilocybin may reduce negative affect and related neural correlates as a possible mechanism. Consequently, the authors recommend further research into psychological and neural mechanisms that might underlie transdiagnostic efficacy. The paper acknowledges several limitations. The delayed treatment control did not account for non-drug elements of the intervention such as preparatory and integration meetings or expectancy effects; the investigators noted difficulties in implementing placebo or active control conditions given the highly discriminable effects of psilocybin. The study had a small sample, short follow-up, limited demographic diversity (predominantly White, non-Hispanic), and excluded individuals at higher suicide risk, restricting generalisability. Facilitators varied in training and professional background, and the authors signalled that the optimal psychotherapy approach and therapist characteristics require further study. They also called for larger, longer, and placebo- or active-controlled trials to better determine safety concerns (including abuse potential, suicide risk, and psychosis emergence) and efficacy across more diverse clinical populations.

The authors conclude that psilocybin-assisted therapy produced large, rapid, and sustained antidepressant effects in this randomised trial of adults with MDD, extending prior findings from cancer-related distress and treatment-resistant depression samples. They recommend further research using active or placebo controls and larger, more diverse samples with longer follow-up to more fully characterise efficacy and safety.