Safety, Tolerability, and Efficacy of Psilocybin in 9 Patients With Obsessive-Compulsive Disorder

OCD affects roughly 2% to 3% of the population and carries a high burden of disability; many patients remain symptomatic despite standard treatments. Potent serotonin reuptake inhibitors (SRIs) and cognitive behavioural therapy are effective for some patients but typically produce only partial improvement, with an estimated 40% to 60% of adequately treated patients failing to respond fully. Anecdotal case reports and historical observations have suggested that indole-based hallucinogens such as psilocybin and LSD can produce acute reductions in obsessive and compulsive symptoms, and pharmacologically these compounds are notable agonists at 5-HT1A, 5-HT2A, and 5-HT2C receptors — sites implicated in serotonergic treatments for OCD. Moreno and colleagues designed a small proof-of-concept study to evaluate the safety, tolerability, and potential therapeutic effects of psilocybin in symptomatic, treatment-resistant OCD patients. The study also aimed to explore whether the intensity of the psychedelic experience related to changes in obsessive-compulsive symptom severity during testing. This investigation was positioned as a Phase I, exploratory evaluation to inform whether further controlled trials would be warranted.

This single-centre Phase I study was approved by the institutional Human Subjects Committee and the FDA, and conducted from November 2001 to November 2004. Nine adults (7 male, 2 female) meeting DSM-IV criteria for current OCD were recruited. Inclusion required at least one prior adequate SRI trial of at least 12 weeks without significant improvement; subjects had a mean of 3.4 ± 1.9 prior treatment failures and baseline Yale-Brown Obsessive Compulsive Scale (YBOCS) scores between 18 and 36 (mean 24.1 ± 5.9). Exclusion criteria included current major psychiatric disorders other than OCD, current substance abuse/dependence, personal or family history of psychosis, and relevant medical or neurological disorders. For safety, participants were required to have tolerated at least one prior exposure to indole-based psychedelics. Antidepressants were discontinued for at least 2 weeks (6 weeks for fluoxetine) and other medications/supplements were stopped before testing. Testing took place in an outpatient psychopharmacology setting with overnight observation in a hospital unit. Rapport-building sessions occurred before dosing. Psilocybin was administered orally in a modified dose-escalation, partially blinded protocol with up to four sessions per subject spaced at least 1 week apart. Doses were 25 µg/kg (very low dose, VLD), 100 µg/kg (low dose, LD), 200 µg/kg (medium dose, MD), and 300 µg/kg (high dose, HD). LD, MD, and HD were given in ascending order; VLD was inserted randomly after the first dose and was double blind. Subjects wore eyeshades, listened to a standardised music programme, and were accompanied by trained sitters throughout each 8-hour session. Primary clinical measures were YBOCS and a visual analogue scale (VAS) for overall obsessive-compulsive symptom severity, recorded at baseline and 4, 8, and 24 hours post-ingestion. The Hallucinogen Rating Scale (HRS) was administered at 8 hours to characterise subjective psychedelic effects across subscales (somaesthesia, affect, volition, cognition, perception, intensity). Vital signs were monitored at baseline and 1, 4, 8, and 24 hours. Statistical analysis used repeated-measures multivariate analysis of variance (ANOVA) to assess main effects of time, dose, and time-by-dose interactions; within-subjects linear contrasts tested dose–response on HRS intensity. Pearson correlations evaluated relationships between HRS intensity and changes in YBOCS. Significance was set at p ≤ .05, with trends at p ≤ .1.

Safety and tolerability: Overall the procedures were tolerated. One subject experienced transient hypertension after a medium dose (readings up to 142/105 mm Hg) without associated anxiety or somatic symptoms. Two participants declined further participation after the first session because of discomfort with hospitalisation. No other adverse reactions were reported. All nine subjects received the LD; seven received VLD and MD; six completed all four doses. Primary clinical outcomes: Repeated-measures ANOVA across all YBOCS values showed a significant main effect of time (Wilks lambda F = 9.86, df = 3,3; p = .046), indicating change in symptom severity over the assessment period, but there was no significant main effect of dose (F = 2.25, df = 3,3; p = .261) nor a significant time-by-dose interaction (F = 0.923, df = 9,45; p = .515). Individual subject responses varied, with marked decreases in YBOCS scores of 23% to 100% during one or more sessions. A contrast comparing baseline versus post-ingestion YBOCS for all doses combined was significant (F = 9.37, df = 1,5; p = .028). Examining individual doses, time effects were significant for LD (p = .004) and MD (p = .006), but not for VLD (p = .128) or HD (p = .406). Changes in obsessive and compulsive symptom subscales mirrored total YBOCS findings. Analysis of baseline to 24-hour YBOCS scores also showed a significant main effect of time (F = 7.89, df = 1,5; p = .038) without dose or interaction effects. Secondary and exploratory findings: VAS scores showed mixed results; the omnibus repeated-measures ANOVA for all doses combined was nonsignificant, but a contrast of baseline versus post-ingestion VAS was significant (F = 11.41, df = 1,5; p = .020), and LD produced a significant time effect (p = .010). The HRS intensity subscale demonstrated a significant linear dose effect (F = 12.36, df = 1,5; p = .017), and total HRS and all subscales except volition showed significant linear dose effects, indicating that subjective psychedelic intensity increased with dose. Importantly, there were no significant correlations between HRS total or subscale scores and reductions in YBOCS or VAS scores. Duration of effect and follow-up: OCD ratings were obtained only up to 24 hours after dosing prior to the next session, so duration beyond that was not prospectively measured. Nevertheless, 88.9% of subjects had at least one testing dose that produced a ≥ 25% decrease in YBOCS at 24 hours, and 66.7% had at least one dose giving ≥ 50% decrease at 24 hours. Two participants reported symptomatic improvement lasting most of the week after testing. One subject achieved long-term remission as assessed at 6-month follow-up. There was no statistically significant order effect across test days (p = .133). The investigators noted that the VLD elicited stronger-than-expected clinical effects, undermining its validity as a placebo comparator.

Moreno and colleagues conclude that, when administered in a supportive clinical environment with monitoring, psilocybin appeared safe and was generally well tolerated in this small sample of treatment-resistant OCD patients. The study found transient reductions in obsessive-compulsive symptoms following psilocybin administration, with many subjects experiencing clinically meaningful decreases at 24 hours after at least one dose. These observations are consistent with prior anecdotal reports and case histories suggesting rapid symptomatic improvement after exposure to indole hallucinogens. The investigators highlight several methodological limitations that temper interpretation. The sample was small and the design was an open, modified dose-escalation Phase I protocol rather than a fully randomised, placebo-controlled trial. The ascending order of LD, MD, and HD (with VLD inserted randomly) and the partial blinding could have influenced expectations of both participants and raters. Requirement that participants tolerate prior psychedelic exposure and overnight hospitalisation may have introduced selection bias. The unexpectedly active response to the VLD precluded its use as an effective placebo control. Measurements of symptoms were limited to 24 hours post-ingestion, so duration of benefit beyond that window was not systematically assessed. Regarding mechanisms, the authors note that although psilocybin is a 5-HT1A/2A/2C agonist and preclinical data support adaptive receptor and gene-expression changes after hallucinogen exposure, it remains unclear which receptor(s) or pathways mediate any anti-obsessional effect. They also acknowledge alternative explanations such as expectancy/placebo effects, distraction or pleasurable experience during intoxication, or psychological insights and altered cognition arising from the psychedelic state. Several participants described psychologically or spiritually meaningful experiences, and some reported symptomatic relief extending beyond the acute psychedelic period, a finding that raises mechanistic questions but cannot be resolved from this dataset. The investigators recommend cautious future research: screening to minimise risk in vulnerable individuals, careful preparation and debriefing, and continuity of support post-session. They propose that, given the burden of treatment-resistant OCD and the limitations of current therapies, further studies using standard randomised, blinded, placebo-controlled designs — including assessments of repeated dosing and longer-term outcomes — are warranted to better characterise efficacy, safety, and duration of benefit.

In this small proof-of-concept Phase I study, psilocybin administered in a monitored clinical setting was generally safe and associated with transient reductions in obsessive-compulsive symptoms among treatment-resistant OCD patients. Methodological constraints — notably small sample size, the modified dose-escalation and blinding scheme, selection criteria, and short prospective follow-up — limit definitive conclusions. The authors suggest that the observed acute benefits, together with the high unmet need in refractory OCD, justify further investigation with rigorous, randomised, placebo-controlled trials to determine efficacy and the duration of therapeutic effects from repeated psilocybin exposure.