Psilocybin with psychotherapeutic support for treatment-resistant depression: a pilot clinical trial

Depressive disorders are a leading cause of global non-fatal disease burden and a substantial subset of people with major depressive disorder (MDD) do not remit after successive courses of conventional treatments. Meikle and colleagues note that after two unsuccessful pharmacological interventions the label 'treatment-resistant depression' (TRD) is commonly applied, and that 10%-30% of people with MDD may meet this threshold. Previous clinical trials of psilocybin-assisted therapy in MDD and TRD have generally shown improvements from baseline, though trial results vary and some well-blinded studies have failed to demonstrate superiority over placebo or active comparators, raising concerns about expectancy effects and heterogeneity across protocols (for example single-dose versus two-dose regimens, and differing amounts of preparatory and integration therapy). Safety outside healthy volunteer samples is still being characterised, and unique interpersonal risks in psychedelic therapy (such as therapeutic touch) require further empirical attention. This pilot study aimed to evaluate a two-dose psilocybin protocol (two 25 mg doses) delivered with psychotherapeutic support for people with TRD in an Australian setting. Meikle and colleagues set out to examine effectiveness and safety, to explore individual trajectories of clinical response with mixed quantitative and qualitative methods, and to report a fully disclosed treatment protocol to inform practice and future research in this jurisdiction.

Meikle and colleagues conducted an open-label, single-arm pilot trial at Swinburne University using two 25 mg oral doses of psilocybin administered 4–6 weeks apart, combined with preparatory and integration psychotherapy delivered by a co-therapist dyad. The investigational product was supplied by Usona Institute. The trial operated under the Australian Clinical Trial Notification scheme and followed TREND reporting guidance. Eligibility targeted adults aged 18–65 with moderate-to-severe MDD (Montgomery-Åsberg Depression Rating Scale, MADRS ⩾ 30) who met the study's definition of TRD (failed response to two or more antidepressant medications in the current episode). Exclusions included medical contraindications, pregnancy/breastfeeding, recent psychedelic or MDMA use (within 12 months, microdosing exceptions noted), very severe anxiety or suicidality, current or past psychotic or bipolar I/II disorders (or first-degree relatives with these conditions), recent substance dependence, dissociative or eating disorders, or other complex psychiatric/medical profiles judged unsafe. Screening proceeded in four phases (online prescreen, video interview, medical exam, and consultation with the participant's treating clinician), and participants tapered contraindicated medications under supervision before dosing. The therapeutic package comprised three 60–90 minute preparatory psychotherapy sessions across three weeks, two full-day dosing sessions with eyeshades and music and participant-led, non-directive support (minimal therapeutic touch permitted under prespecified conditions), and three 60–90 minute integration sessions following each dose. Sessions were video recorded; therapists were trained mental healthcare professionals following treatment guides rather than a manualised protocol. Assessments included a primary outcome of change in self-rated depressive symptoms measured by the Quick Inventory of Depressive Symptomatology—Self-Report (QIDS-SR) at 3 weeks post-dose 2. Secondary measures covered quality of life (WHOQoL-Bref), anxiety (GAD-7), experiential avoidance (BEAQ), self-efficacy subscales (RSES), connectedness (Watts Connectedness Scale), persisting effects (PEQ), expectancy (PATHEV), Mindset (preparedness) on dosing days, and acute experience measures (11D-ASC, Psychological Insight Questionnaire, Emotional Breakthrough Inventory) administered the day after dosing. Adverse events (AEs), vital signs during dosing, blood tests, and the Hallucinogen Persisting Perceptual Disorder (HPPD) questionnaire at long-term follow-up were also collected. One-hour qualitative interviews occurred at baseline, 3 weeks post-dose 2 and long-term follow-up. Analyses were conducted in R. Because of the small sample, exact paired permutation tests assessed change in QIDS-SR and secondary outcomes at the primary endpoint; effect sizes used Hedges' g with 95% confidence intervals. Clinically meaningful response was defined as ⩾50% reduction in QIDS-SR and remission as QIDS-SR ⩽ 5. Participants were grouped descriptively by treatment trajectory and multinomial logistic regression models tested predictors of trajectory (PATHEV subscales, 11D-ASC subscales, PIQ, EBI, and Mindset), rotating reference categories to explore relationships. Secondary/exploratory outcomes were treated descriptively, prioritising effect sizes and confidence intervals; qualitative quotes were used selectively to contextualise findings.

Recruitment ran from February 2022 with enrolment between August 2022 and August 2023. From 209 complete prescreen responses, eight participants were enrolled but one was withdrawn prior to dosing after identification of a 'complex case' that posed safety concerns; the remaining seven participants completed the full protocol and were included in analyses. Recruitment stopped early due to COVID-19–related delays and funding/time constraints, so the planned sample size (n=15) was not reached. On the primary outcome (QIDS-SR), all seven participants showed reductions from baseline to 3 weeks post-dose 2. The mean reduction was 7.14 points (SD = 4.41), representing an average decrease of 47.96%; this change was statistically significant by permutation test (p = 0.02) with Hedges' g = -1.27 (95% CI [-2.40, -0.37]). Clinically meaningful response (⩾50% reduction) and remission were observed in three of seven participants at the primary endpoint (identified as participants 3, 4 and 6). At long-term follow-up (20 weeks post-dose 2) the aggregated reduction in QIDS-SR remained statistically significant (mean reduction 7.14 points; SD = 5.84; p = 0.02; Hedges' g = -1.16; 95% CI (-2.77, -0.46)); however, three participants (IDs 1, 2 and 5) had resumed or started a new antidepressant before that follow-up. Individual trajectories were categorised descriptively into three groups: sustained response (two participants, IDs 3 and 6) who maintained ⩾50% reduction through long-term follow-up; relapsing (three participants, IDs 1, 4 and 7) who showed initial ⩾50% reduction that was not sustained; and non-response (two participants, IDs 2 and 5) who never achieved ⩾50% reduction. Secondary health and well-being measures showed mean improvements at the primary endpoint across outcomes reported, with confidence intervals excluding zero for most measures except anxiety scores. Persisting Effects Questionnaire data indicated predominantly positive lasting changes in attitudes, mood and behaviour. The pattern of change on both primary and secondary outcomes varied by treatment trajectory. Multinomial logistic regression models indicated that pre-dosing Mindset and certain elements of the acute psychedelic experience (11D-ASC subscales) were associated with treatment trajectory. Specifically, higher Spiritual Experience scores were associated with decreased odds of relapsing and of non-response compared with sustained response. Higher Changed Meaning of Percepts scores increased the odds of being in the sustained or relapsing groups versus non-response. Higher Mindset scores (greater preparedness/rapport/intention) reduced the odds of non-response compared to sustained and relapsing responses, and increased the odds of sustained or relapsing responses relative to non-response. The models reportedly displayed good fit, but a combined multivariable model was not run because of the small sample size and multicollinearity among predictors; other candidate predictors showed less certain effects. Qualitative interview data were used to validate the trajectory categories and to illustrate variation in acute subjective experiences. Safety and feasibility findings indicated no serious adverse events. Four participants reported adverse events overall; one participant (ID 2) experienced anxiety, insomnia and suicidal ideation after a very challenging first dose, required daily phone contact and external mental health support, and commenced psychiatric medication; all these adverse events had resolved by long-term follow-up except for ongoing anxiety in that participant which required ongoing treatment. A second participant (ID 5) received an additional integration session between doses for well-being concerns. All participants consented to therapeutic touch and five received it, reporting it as helpful; no adverse events related to touch were reported. The study also noted practical challenges such as withdrawal-related adverse events and symptom exacerbation between screening and baseline that could inflate baseline scores; AE monitoring used open-ended questioning, which the authors acknowledge may underreport certain events.

Meikle and colleagues interpret their findings as consistent with prior psilocybin therapy research in TRD: the trial observed statistically significant and clinically meaningful reductions in self-reported depressive symptoms with response/remission rates of 43% at the primary endpoint, comparable to other TRD trials. The authors highlight the clinical context that participants had typically undergone multiple prior treatment attempts in the current episode, and contrast the observed rates with lower response/remission rates typically seen at this stage with conventional treatments. The investigators note sustained depression reductions at 20 weeks in two participants, and compare this to previous reports of median time to relapse of 11–20 weeks in similar populations, suggesting some enduring effects for a subset of participants. They caution, however, that the trial differed from others in using a two-dose schedule and more extensive preparation and integration, and that the present design and small sample prevent causal claims about whether these protocol differences improved outcomes. The use of a self-report primary outcome (QIDS-SR) is discussed as a limitation: self-report measures can be more susceptible to response bias and may overestimate change relative to clinician-rated instruments, and there are questions about the QIDS-SR's suitability in psychedelic research. Regarding predictors of outcome, the authors emphasise the apparent role of psychological readiness (Mindset) and certain acute experiential features—particularly spiritual experience and changes in the meaning of percepts—in distinguishing sustained responders from relapsing or non-responding participants. They present Changed Meaning of Percepts as a potentially novel predictor that aligns with theories that therapeutic benefit may arise from emotional insight or perspective shifts during the acute experience. Treatment expectations did not reliably predict trajectory in this sample, which the authors suggest may reflect restricted variability in expectancy scores among participants. On safety, the trial's adverse events were described as generally congruent with previous psilocybin studies, but the occurrence of transient suicidal ideation and one case of persisting anxiety underscore the need for vigilance and consistent, standardised AE monitoring. The authors discuss challenges related to antidepressant discontinuation, noting withdrawal-related AEs and potential inflation of baseline symptom severity; they observe that emerging evidence questions the necessity of stopping antidepressants before psilocybin therapy and recommend further study of this issue. Key limitations acknowledged include the small sample size, open-label single-arm design, reliance on a single self-report measure, self-referral recruitment strategy, and extensive exclusion criteria that limit generalisability. The authors conclude that exploratory findings—particularly regarding predictors and individual trajectories—are tentative and require confirmation in larger, controlled studies that use more standardised AE assessment tools and consider masking or alternative trial designs to reduce expectancy bias.

The authors conclude that this pilot trial adds to evidence suggesting psilocybin with psychotherapeutic support is feasible, generally safe in the context studied, and can produce clinically meaningful improvements in depressive symptoms for some people with treatment-resistant depression, while emphasising the need for larger, controlled trials and more systematic safety monitoring.