An open-label, dose-escalation trial of psilocybin-assisted therapy for bipolar 2 depression

Downey and colleagues situate the study within a clinical context in which bipolar II disorder (BD-II) presents persistent depressive episodes, high psychiatric comorbidity, and limited effective treatments. Earlier clinical trials of psilocybin plus psychotherapy have shown antidepressant effects in other populations, but people with BD-II and those with family histories of bipolar disorder have largely been excluded because of theoretical concerns about precipitating mania or psychosis. The authors note that adverse event reporting in psychedelic trials has been inconsistent and that restrictive eligibility criteria limit generalisability to real-world, comorbid populations. This paper reports an open-label, single-arm, dose-escalation pilot designed to evaluate safety, tolerability, and preliminary efficacy of psilocybin-assisted therapy for individuals with BD-II currently experiencing at least moderate depression. The trial emphasises systematic adverse-event capture, a conservative dosing strategy (initial 10 mg with optional 25 mg if depressive symptoms persisted), and inclusion of participants with greater psychiatric complexity than in many prior trials, with the aim of informing whether larger, randomized studies are warranted.

The study used an open-label, single-arm dose-escalation design. Eligible participants were adults aged 18–70 with DSM-defined BD-II in a current major depressive episode lasting more than four weeks and a baseline MADRS score >18; they were required to have had at least one prior unsuccessful medication trial of >6 weeks. Key psychiatric exclusions included current hypomania, history of psychotic disorder, recent (past 12 months) moderate-or-worse substance use disorder, psychedelic use in the past six months, and imminent suicide risk on the C-SSRS. Medical exclusions included serious cardiovascular or central nervous system disease. Four participants were taking medications with potential psilocybin interactions; these were tapered under psychiatric supervision to achieve clearance (≥5 half-lives) before dosing. To support safety, participants identified a support person who saw them at least five days per week. Diagnostic and baseline assessments included serum labs, ECG, physical exam, vital signs, and the SCID-5; the investigators used an adapted hypomania criterion of ≥2 consecutive days of symptoms to diagnose BD-II. Each participant received preparatory psychotherapy with a licensed therapist, then an initial oral dose of synthetic psilocybin (10 mg). If the 10 mg session was well tolerated and MADRS remained ≥7 at day 21, participants could receive a second administration of 25 mg. Participants were not told the specific MADRS threshold used to determine progression to the second session. Administration took place in a standardised, comfortable setting with a lead and assisting therapist present; vital signs and urine drug screens were checked before dosing and rescue medications were available to the on-site medical provider. Regulatory approvals included the FDA, DEA, relevant state regulators, and the UCSF IRB. Outcome assessments were scheduled at 7, 14, and 21 days after each administration (A7/A14/A21 for 10 mg; B7/B14/B21 for 25 mg), with a long-term follow-up 90 days after the participant's final session (AB90). Safety and tolerability were the primary focus: adverse events were captured prospectively at every visit, the Clinical Global Impression of Improvement (CGII) was used, suicidality was measured with the C-SSRS (including Ideation severity subscale), and treatment-emergent (hypo)mania, psychosis, and insomnia were assessed with the YMRS, PANSS, and ISI. The primary preliminary efficacy endpoint was change in MADRS from baseline to 21 days after the final administration; quality of life used the QoL-BD. Exploratory measures included treatment expectations (SETS), PTSD symptoms (PCL-5), attachment style (ECR-M16), adverse childhood experiences (ACE), and the MEQ, EBI and CEQ to characterise the acute psychedelic experience. For statistical analysis, mixed-effects linear models were fitted with timepoint as a fixed effect and participant ID as a random effect; model assumptions were checked but small sample size limited confirmatory inferences. Paired Wilcoxon signed-rank tests were used as a non-parametric check, and Spearman correlations (with Pearson and Kendall checks) examined associations between baseline/acute-experience measures and outcomes. Paired t-tests compared phenomenological scores between 10 mg and 25 mg sessions. Results are reported as estimated mean differences (standard errors) with Hedges' g effect sizes where applicable; analysis scripts were noted as available but the detailed statistical analysis plan was not finalised prior to database lock, so analyses are presented as exploratory.

Fourteen of 58 screened participants received at least one psilocybin administration; mean baseline MADRS was 27.7 (SD 6.6). All 14 completed the 21-day assessment after 10 mg; nine received a second (25 mg) session and completed the corresponding 21-day follow-up. Mean positive expectancy on the SETS was low (2.0 [1.4]) while negative expectancy items averaged 4.4 (1.8). Physiologic effects were transient and self-limiting: mean peak heart rate was 84.8 (17.0) bpm after 10 mg and 84.4 (17.3) bpm after 25 mg; peak systolic/diastolic pressures averaged 141.6 (15.5)/86.3 (8.9) mmHg for 10 mg and 144.7 (18.3)/88.9 (8.3) mmHg for 25 mg. Eight of eleven respondents (73%) to the Treatment Satisfaction Questionnaire said they would recommend the treatment. Adverse events (AEs) were frequent but mostly mild: of 143 total AEs, 83% were mild, 15% moderate, and one was severe; no events were classified as serious. Most events occurred within 24 hours of dosing and commonly included mild-to-moderate anxiety, nausea and headache. On the C-SSRS Ideation severity subscale there was no increase at A21/B21 versus baseline; rather a small average reduction (~0.7 points) was observed at post-administration timepoints. Notably, three psychiatric adverse events were judged concerning: one participant experienced worsening suicidal ideation with intent 37 days after a challenging 10 mg session (this participant declined the 25 mg session and later accessed external mental health care); a second participant had suicidal ideation 11 days after 10 mg that resolved in two days with therapist support; and a third participant showed elevated YMRS scores consistent with hypomania 14 days after 25 mg that returned to baseline by B21 without emergency services. Mean YMRS and PANSS scores did not show significant group-level worsening. On the primary preliminary efficacy outcome, MADRS scores improved significantly. After the 10 mg session participants showed a mean change at A21 of -12.7 (SE 2.7), n=14, p<0.001, with 4/14 (28.5%) meeting remission criteria (MADRS ≤6) and therefore not proceeding to the 25 mg session. Among those receiving 25 mg, B21 change was -18.6 (3.1), n=9, p<0.001. At the 90-day follow-up (AB90) sustained improvement versus baseline was observed (AB90: -14.3 [2.8], n=12, p<0.001), with Hedges' g=1.9. Quality of life (QoL-BD) also improved substantially at A21 and B21 (A21 reported as 35.5, p<0.001; B21 55.9 [10.9], p<0.001) and remained improved at AB90 (31.2 [10.2], p=0.004; Hedges' g=1.6). Exploratory analyses noted baseline PTSD symptom burden in many participants (mean PCL-5 31.9 [20.9]), with a significant reduction at AB90 (change -13.9 [6.1], p=0.048; Hedges' g=-0.87). Phenomenological measures of the acute experience (MEQ, EBI, CEQ) were numerically similar between 10 mg and 25 mg sessions; normalized MEQ scores were 0.53 (0.28) vs 0.63 (0.29), CEQ 0.20 (0.13) vs 0.30 (0.19), and EBI 0.78 (0.14) vs 0.76 (0.18). Attachment anxiety and avoidance on the ECR-M16 decreased at A21 and B21; avoidance remained significantly lower at AB90 while anxious attachment showed a non-significant trend at AB90. No robust associations emerged between baseline measures (ACE, SETS, PCL-5) or acute-experience measures (EBI/MEQ/CEQ) and clinical outcomes.

Downey and colleagues interpret the trial as providing initial evidence that psilocybin-assisted therapy can be administered to people with BD-II depression under controlled conditions with a safety profile broadly comparable to that reported in other psilocybin trials. Group-level worsening to frank mania or psychosis was not observed, but systematic adverse-event monitoring revealed a substantial burden of mostly mild, transient events concentrated around the day of administration. The authors highlight three psychiatric adverse events of concern—one case of worsening suicidal ideation with intent, one brief episode of suicidal ideation, and one transient hypomanic episode—and acknowledge that attributing these events to psilocybin is complicated by participants' clinical complexity, recent medication tapers, and the natural course of BD-II. The investigators emphasise the sizeable and sustained reductions in depressive symptoms (MADRS) and improvements in quality of life through 90 days, and note that 4 of 14 participants remitted after a single 10 mg session. They caution that some remissions after the low dose may reflect placebo responses, but point out that baseline expectancies were not associated with outcomes in this sample. Additional exploratory findings included reductions in PTSD symptoms and attachment insecurity, which the authors suggest may merit further study given the resistance of these domains to standard treatments. Limitations acknowledged by the study team include the small sample size, open-label design, and exclusionary criteria that reduce generalisability to BD-I or more complex presentations; the required tapering of most serotonergic medications further limits applicability to routine clinical populations. The absence of a finalised statistical analysis plan prior to database lock led the authors to present analyses as exploratory. Practical and ethical lessons for future trials are discussed: the authors recommend routine, extended monitoring for suicidality in the weeks to months after dosing, greater transparency and flexibility in decisions about additional dosing to reduce participant distress, standardised and complete adverse-event reporting in psychedelic trials, and randomised, placebo-controlled and dose-response studies to confirm efficacy and refine safety protocols.