Psilocybin with psychological support for treatment-resistant depression: six-month follow-up

Psilocybin, a naturally occurring serotonergic psychedelic, has been investigated for a range of psychiatric indications and prior trials have reported improvements in depressive symptoms following psilocybin‑assisted psychotherapy. Previous work has also suggested therapeutic potential for other serotonergic psychedelics in conditions such as obsessive–compulsive disorder, addiction and anxiety related to terminal illness, and mechanistic interest centres on 5‑HT2A receptor agonism. Earlier clinical and historical studies, as well as laboratory findings, indicate both promise and complexity in how 5‑HT2A signalling relates to antidepressant effects. Carhart‑Harris and colleagues present an extension of their earlier pilot study in treatment‑resistant depression, increasing the sample from 12 to 20 patients and lengthening follow‑up from 3 to 6 months. The paper reports safety and efficacy outcomes from an open‑label feasibility trial in which two oral psilocybin doses were administered with structured psychological support; the main aim was to assess symptom change over time, tolerability and relationships between acute subjective experiences and longer‑term clinical outcomes.

This was an open‑label feasibility study in 20 patients meeting criteria for treatment‑resistant, unipolar major depression. Regulatory and ethical approvals were obtained and psilocybin (formulated as 5 mg capsules) was sourced from THC Pharm and prepared by a hospital pharmacy unit. Treatment comprised two supervised oral psilocybin sessions given 7 days apart: an initial 10 mg dose followed by a 25 mg dose. Psychological support followed a three‑component model: preparation, acute/peri‑acute support and post‑session integration (acronym PSI). The primary outcome was change in self‑rated depressive symptoms measured with the 16‑item Quick Inventory of Depressive Symptoms (QIDS‑SR16). QIDS‑SR16 assessments were collected at 1–3 weeks, 5 weeks (primary endpoint), and at 3 and 6 months post‑treatment. Secondary measures included the Beck Depression Inventory (BDI), Spielberger Trait Anxiety Inventory (STAI‑T), Snaith–Hamilton Pleasure Scale (SHAPS), clinician‑rated HAM‑D and Global Assessment of Functioning (GAF); the latter two were collected at 1 week only. For the six post‑treatment QIDS contrasts a Bonferroni‑corrected α of 0.0083 was applied; the team elected not to correct other correlated clinical measures to avoid excessive type 2 error. Prospective participants underwent telephone screening and a full screening visit including MINI‑5 diagnostic interview, physical assessments, ECG and drug/pregnancy testing. Main inclusion criteria were unipolar major depression of at least moderate severity (HAM‑D ≥ 16) and failure to improve after at least two pharmacologically distinct antidepressant courses in the current episode. Main exclusions included current or first‑degree family history of psychotic disorder. Most patients on antidepressants underwent supervised taper and washout prior to dosing. Patients also attended baseline MRI and preparation visits; post‑session debriefing and MRI scans occurred the day after dosing. Subjective acute effects were measured with the 11‑dimension Altered States of Consciousness (11D‑ASC) questionnaire at the end of each dosing day with ratings referenced to the peak experience. Side effects were documented from direct patient reports at each post‑treatment visit. Statistical analysis used two‑tailed paired t tests for pre‑vs‑post contrasts with Cohen's d for dependent data reported; Bonferroni correction was applied only to the primary QIDS timepoint contrasts.

Recruitment began from 120 initial enquiries; 74 proceeded to telephone screening, 29 to in‑person screening and 20 were enrolled, with 19 completing all assessment time points. Eighteen of 20 patients had severe or very severe depression at baseline (QIDS‑SR16 ≥ 16). Patients reported a long illness history (mean duration 17.7 ± 8.4 years) and a high burden of prior treatment failure (median 4 failed medications, mean 4.6 ± 2.6, maximum 11). QIDS‑SR16 scores were significantly reduced relative to baseline at all six post‑treatment time points. The maximal mean change occurred at 5 weeks (mean reduction = -9.2, 95% CI = -11.8 to -6.6, t = -7.2, p < 0.001), with a large effect size (Cohen's d = 2.3). All 19 completers showed some reduction in QIDS‑SR16 at 1 week; these reductions were sustained in the majority across weeks 3–5. Other clinical measures mirrored this pattern: BDI scores fell markedly at 1 week (mean reduction = -22.7, 95% CI = -17.6 to -27.8, p < 0.001) and remained significantly lower at 3 and 6 months; STAI‑T anxiety scores and SHAPS anhedonia scores showed significant reductions at early and some later time points; HAM‑D decreased and GAF scores improved at 1 week. Safety findings indicated good tolerability and no serious adverse events. Common transient effects during or shortly after dosing included anxiety (n = 15), headaches lasting 1–2 days (n = 8), transient nausea (n = 5) and brief paranoia during the acute experience (n = 3). Fourteen patients reported autobiographical visions, which were generally experienced as insightful. One patient became uncommunicative at the peak of the 25 mg session but normalised afterwards; this individual declined most further follow‑up assessments but provided QIDS and BDI scores at 6 months (QIDS 25, BDI 40). Suicidality items on the QIDS‑SR16 were significantly reduced at weeks 1 and 2 (mean reductions -0.9 and -0.85, respectively) with trend reductions at weeks 3 and 5; HAM‑D suicide item scores were significantly decreased at 1 week with 16 of 19 patients scoring 0. Acute subjective effects measured by the 11D‑ASC were greater after 25 mg than 10 mg for experience of unity, spiritual experience, blissful state, insightfulness and complex imagery (all surviving Bonferroni correction). Because experience of unity, spiritual experience and blissful state were highly intercorrelated (r > 0.92), the investigators combined them into a USB factor and found that USB and insight scores during the 25 mg session correlated with greater reductions in QIDS‑SR16 at 5 weeks (USB r = -0.49, p = 0.03; insight r = -0.57, p = 0.01). At the 6‑month endpoint some patients had received additional treatments: six began new antidepressant courses after 3 months and five received various psychotherapies; five participants had independently obtained psilocybin outside the study between 3 and 6 months. Excluding those five did not materially alter the primary 3‑ and 6‑month results (effect sizes remained large and p values remained significant). Of the nine responders at 5 weeks, three met conservative criteria for relapse by 6 months while six maintained response.

Carhart‑Harris and colleagues present extended data from an open‑label trial that reinforce earlier findings of rapid and substantial reductions in depressive and anxiety symptoms following two psilocybin sessions with psychological support. Many participants showed large, early improvements, with nominal peak effects at 5 weeks and durable reductions evident at 3 and 6 months for a substantial proportion of the sample. The correlation observed between acute subjective experiences labelled as insightfulness (and a composite of unity/spiritual/blissful experiences) and later symptom improvement is noted as supportive of a model in which qualities of the acute psychedelic experience relate to therapeutic benefit; the authors point to complementary fMRI data collected in the trial as relevant to refining mechanistic understanding. Importantly, the investigators emphasise limitations that constrain causal inference. The open‑label design and absence of a control condition prevent definitive efficacy claims and the sample size is small. Regulatory and logistical constraints prevented a planned placebo‑controlled RCT, leading to the present feasibility design. Generalisability is limited in part because the final eight participants were all male and because duration of the current depressive episode was based on patient estimates, introducing uncertainty. The authors also acknowledge that subsequent, non‑trial interventions among participants confounded outcomes at the 3‑ and 6‑month timepoints, although sensitivity checks excluding those who sought additional psilocybin did not substantially change results. Looking forward, the paper argues for better characterisation, measurement and control of the psychological components of the treatment model—such as expectations, interpersonal context and music—plus clearer specification or manualisation of psychotherapeutic support. The investigators caution that experiments manipulating psychological context must respect the vulnerability of individuals under psychedelics and maintain minimum standards of care. Concluding remarks frame the findings as proof‑of‑principle: psilocybin administered with psychological support appears feasible and tolerable in severe treatment‑resistant depression and warrants further investigation in larger, double‑blind, controlled trials, including studies addressing comparative and cost‑effectiveness questions.