Psilocybin-Assisted Psychotherapy for Treatment-Resistant Depression in Bipolar II Disorder

Over the past two decades, psilocybin and other classic serotonergic psychedelics have re-emerged as potential treatments for psychiatric disorders, particularly depression. Previous psilocybin-assisted psychotherapy (PAP) trials have suggested rapid and robust antidepressant effects in major depressive disorder (MDD), treatment-resistant depression (TRD) and in patients with life-threatening illness, but most trials used small, highly selected samples and intensive psychotherapy protocols. These exclusions—commonly omitting individuals with substantial comorbidity, bipolarity, suicidality or very high degrees of treatment resistance—limit the generalisability of existing evidence to real-world clinical populations. Important open questions also remain about durability of effect, the role of repeat dosing for relapse, and the minimal necessary ‘‘dose’’ of adjunctive psychotherapy for safe and scalable delivery. Rosenblat and colleagues designed a randomised, waitlist-controlled clinical trial to address these gaps by testing a brief four-session PAP model in a more complex TRD population. The trial deliberately broadened eligibility to include both MDD and bipolar II disorder (BDII), participants with comorbid psychiatric conditions and baseline suicidality, and imposed no upper limit on prior failed treatment trials. A secondary aim was to evaluate the feasibility, safety and preliminary efficacy of offering repeat psilocybin doses to participants who showed clinical relapse after an initial response. The investigators framed this study as a feasibility trial intended to reflect how PAP might be implemented in real-world clinical practice.

This single-centre, randomised, waitlist-controlled trial was conducted at a community clinic in Mississauga, Ontario, and registered on ClinicalTrials.gov (NCT05029466). Blocked randomisation with permuted blocks (block size ten) allocated participants to immediate treatment or to a two-week delayed (waitlist) start; the trial was open-label with no blinding. The study was designed as a feasibility trial and recruited adults aged 18–75 years with a primary diagnosis of MDD or BDII and a current major depressive episode of at least three months. Treatment resistance required failure of at least two adequate pharmacological trials for the current episode, with no upper limit on the number of failed trials. Candidates were required to be medically stable and to refrain from antidepressants, antipsychotics, ketamine/esketamine and augmenting medications for at least five half-lives before screening and for the 6-month follow-up, although BDII participants could continue mood stabilisers where clinically indicated. Psilocybin (synthetic, 25 mg per session) was administered orally dissolved in water. Each dose was accompanied by a brief preparatory therapy session (1–2 h), a supervised dosing day (6–8 h) with a therapist dyad, and two integration sessions (1–2 h each), amounting to approximately 4.5 h of psychotherapy per dose outside the dosing day. Therapists were multidisciplinary and underwent trial-specific training; the therapeutic model was transtheoretical and adapted from prior PAP protocols. Eligibility for repeat dosing (second or third dose) required demonstration of prior clinical benefit, adequate tolerability, and signs of relapse of depressive symptoms lasting two weeks or more; repeat doses followed the same therapy protocol as the first. Primary feasibility outcomes were prespecified: (1) less than 30% all-cause dropout before the week-2 primary endpoint, (2) zero or minimal worsening of baseline suicidality post-dose, (3) fewer than 15% experiencing serious adverse events (SAEs), and (4) greater than 80% of adverse events resolving within 48 h of dosing. For preliminary efficacy, the primary clinical comparison contrasted change in clinician-rated Montgomery-Åsberg Depression Rating Scale (MADRS) scores from baseline to week 2 between the immediate treatment arm and the delayed/waitlist period of the control arm. Statistical analyses used descriptive statistics for feasibility and safety, two-tailed t-tests for between-group MADRS change with significance at p<0.05, 95% confidence intervals, and Hedges' g to report effect sizes appropriate for small samples. No formal power calculation was performed; the planned sample aimed for roughly 12 completers per arm as a feasibility rule-of-thumb.

Between 1 November 2021 and 1 February 2023, 135 patients were assessed and 31 were enrolled and randomised; one withdrew before receiving the intervention because of difficulties tapering antidepressants, leaving 16 allocated to immediate treatment and 14 to delayed treatment. One further participant dropped out prior to the week-2 primary endpoint, so 29 participants were included in the primary outcome analysis. Over the 6-month follow-up, eight additional participants discontinued to begin other treatments, leaving 21 participants who completed the full follow-up. The sample was clinically severe and highly treatment resistant. Mean age was 44.4 years (SD 13.7) and mean duration of depression was 18.3 years (SD 11.5). The mean number of failed medication trials was 11.27 (SD 5.59), and most participants had failed multiple psychotherapies; 12/30 (40%) had previously failed intensive interventions such as electroconvulsive therapy or ketamine. Baseline mean MADRS was 30.5 (severe range), with the immediate treatment arm more severely affected at baseline by an average of 5.4 MADRS points. Primary diagnoses were mainly MDD (n = 26), with four participants diagnosed with BDII. All participants met criteria for at least one comorbid psychiatric condition; nine participants had comorbid personality disorders. All four predefined feasibility criteria were met. Dropout before the week-2 primary endpoint was 1/30 (3%), two participants experienced transient worsening of suicidality for 24–48 hours post-dosing without requiring further intervention, zero participants experienced SAEs, and all but one adverse event resolved within 48 hours. The single persistent treatment-emergent adverse event was ongoing genital arousal that persisted through the 6-month follow-up and, according to the authors, has not been previously reported with psilocybin. For the primary clinical comparison, change in MADRS from baseline to week 2 differed between groups: the waitlist group had a mean change of −3 (SE 0.9) versus −9.6 (SE 1.9) in the immediate treatment group, yielding a between-group difference of 6.6 MADRS points. This corresponded to a Hedges' g of 1.1 (95% CI 0.3–1.8) and a p value of 0.005. Open-label longitudinal data over the 6-month follow-up showed sustained reductions in MADRS scores; 17 participants received a second psilocybin dose and 5 participants received a third dose. The cumulative reduction in mean MADRS at two weeks after the most recent dose increased with each subsequent dose. Self-reported depressive symptoms (QIDS-SR) declined by a magnitude similar to the MADRS, whereas self-reported anxiety (GAD-7) showed only small, short-lived improvements and largely returned to baseline by two months. There was no statistically significant change on the MADRS suicidal ideation item between baseline and the primary endpoint in either group. Adverse events reported were mostly mild to moderate and transient, consistent with expectations for PAP, aside from the single persistent adverse event described above.

Rosenblat and colleagues conclude that a brief PAP protocol is feasible and can be safely evaluated in a more complex, treatment-refractory clinical population that included both MDD and BDII, substantial comorbidity and baseline suicidality. Feasibility was shown both for broadened eligibility criteria and for an abbreviated psychotherapy model consisting of one brief preparatory session and two integration sessions per psilocybin dose, amounting to roughly 4.5 hours of non-dosing psychotherapy per dose. The investigators observed clinically meaningful reductions in depressive symptoms that were consistent across clinician-rated (MADRS) and self-reported (QIDS-SR) measures, while anxiety improvements were limited and transient. The trial also demonstrated feasibility for offering flexible repeat doses upon signs of relapse, and exploratory data suggested incremental antidepressant benefit with subsequent doses, although the authors caution that indication bias limits causal interpretation of that finding. When positioned against prior trials, the authors note that the between-group MADRS difference of 6.6 points at the primary endpoint mirrors that reported in a large phase 2b RCT (Goodwin et al.), but absolute reductions were smaller than in some prior studies. Rosenblat et al. attribute this partly to the greater complexity and higher degree of treatment resistance in their sample—29/30 participants would have been excluded from that previous TRD RCT—while acknowledging that abbreviated psychotherapy could also have contributed to smaller effect sizes. Inclusion of BDII (n = 4) and participants with personality disorders (n = 9) was feasible; no treatment-emergent mania, hypomania or psychosis was observed, although the authors stress that the small subsamples preclude firm safety conclusions for these diagnoses. Key limitations are openly acknowledged: the open-label design, small sample size, use of a waitlist rather than placebo control, sample heterogeneity, lack of expectancy measurement, and deliberate modifications to trial procedures compared to prior PAP studies. These limitations impair causal inference and preclude contribution toward regulatory approval. The authors therefore recommend adequately powered, controlled RCTs to evaluate PAP in complex clinical populations, to determine optimal numbers of psilocybin doses for acute and maintenance phases, and to clarify the necessary intensity of adjunctive psychotherapy for safe, effective and scalable implementation.