Abstract
Background:
Current treatments for obsessive–compulsive disorder (OCD), including serotonin reuptake inhibitors and cognitive-behavioral therapy, are often insufficient. Psilocybin, a 5HT2a agonist psychedelic, has shown promise for treating OCD, but rigorous evidence is still needed.
Aims:
This randomized clinical trial evaluated safety, tolerability, and benefit of multiple psilocybin doses in OCD patients.
Methods:
Fifteen participants were randomized to receive 4 weekly sessions of high-dose (300 µg/kg), low-dose (100 µg/kg) psilocybin, or active placebo (lorazepam) in a double-blind Phase 1 (n = 5 per condition), followed by four additional high-dose sessions (single-blind Phase 2). OCD severity was assessed with the Yale-Brown Obsessive Compulsive Scale (YBOCS) following each session, and prospectively for 6 months. Safety was evaluated via adverse event systematic assessment, suicide severity rating, and psychosis screening.
Results:
Psilocybin was generally well-tolerated, with no serious adverse events, or psychotic symptoms, and no significant changes in suicide severity scores. Psilocybin but not placebo significantly reduced YBOCS scores. At the end of 8-week treatment, after participants had received at least four high doses of psilocybin, 73.3% were responders (⩾35% reduction in YBOCS scores), with 40% in remission. These effects diminished but remained substantial at 6 months. Post hoc analysis of cumulative dosing correlated with YBOCS score reductions at the end of treatment.
Conclusions:
Administration of up to eight doses of psilocybin in a clinical research setting appears to be safe and potentially effective for patients with OCD. Larger trials are needed to further support efficacy and refine treatment protocols.
Clinical trial registration:
ClinicalTrials.gov ID NCT03300947.
Introduction
Obsessive–compulsive disorder (OCD) is a chronic and debilitating psychiatric condition with a lifetime prevalence of 2%–3%, characterized by intrusive thoughts (obsessions) and repetitive behaviors (compulsions; American Psychiatric Association, 2013). OCD symptoms consume significant time, cause considerable distress, impair quality of life and functional ability in affected individuals, resulting also in significant societal costs. Current standard treatments, including psychotherapy and pharmacotherapy, have limitations, such as delayed and incomplete symptom resolution, and challenges with care adherence due to side effects and accessibility issues. This underscores the critical need for alternative and more effective therapeutic approaches (Gosmann et al., 2021).
Psilocybin, a serotonergic psychedelic compound found in certain species of fungi, has emerged as a promising candidate for the treatment of treatment-resistant OCD (Ehrmann et al., 2021). Oral ingestion of psilocybin results in nearly complete metabolism to psilocin. Psilocin binds to several 5-HT receptors, importantly 2-A, 2-C, and 1-A, although partial agonism of 5-HT2A receptors is believed to account for its psychoactive effects, and the mechanism of action for potential neurocognitive changes and therapeutic benefit is more complex than originally thought (De Gregorio et al., 2021). Psilocybin administration has been associated with increased cognitive flexibility and a reduction of rigid thought patterns (Carhart-Harris et al., 2014; Carhart-Harris and Friston, 2019). Contemporary neuroimaging studies have shown that psilocybin alters brain activity and connectivity across networks that interact with the Cortico-Striato-Thalamo-Cortical (CSTC) circuitry, including thalamocortical and frontoparietal control networks (Barrett et al., 2020; Daws et al., 2022; Siegel et al., 2024). The CSTC circuit, including the orbitofrontal cortex, striatum, and thalamus, plays a central role in cognitive flexibility, error monitoring, and habitual responding in OCD (Norman et al., 2018, 2019; Robbins et al., 2012). Taken together, these pharmacological, psychological, and brain network effects support the rationale for evaluating psilocybin as a compelling candidate for the treatment of OCD.
Building on case studies and anecdotal reports (Moreno and Delgado, 1997), an initial pilot study (Moreno et al., 2006) found that various doses of psilocybin produced dose-dependent psychedelic effects were overall well-tolerated, and resulted in temporary reductions in OCD symptoms, with Yale-Brown Obsessive Compulsive Scale (YBOCS; Goodman et al., 1989) scores improving by 23%–100%, when measured within 24 hours. Since then, other preliminary findings have also reported initial signals of safety and efficacy (Kelmendi et al., 2022); A large retrospective survey of 174 participants with OCD who have used psychoactive drugs identified improvements in OCD associated with classic psychedelic use, particularly psilocybin (Buot et al., 2023). A systematic review of psychedelics and OCD elegantly summarized preclinical, case reports, and clinical trials supporting this treatment approach (Graziosi et al., 2024). Since then, two prospective studies have been published: 1 of 18 individuals receiving psilocybin doses of 1 and 10 mg has shown that doses of 10 mg led to greater acute short-term improvement in OCD symptoms when measured 1 week after study drug ingestion and lasting through the 4 weeks of follow-up (Pellegrini et al., 2025). Another study conducted a qualitative analysis of phenomenology experienced after single dose administration of 250 µg/kg of psilocybin for OCD. The study reported post-exposure experiences resulting in variable improvements in obsessions and compulsions among some participants, although formal quantitative analysis was not reported (Ching et al., 2025). Globally, there are multiple ongoing and recently concluded clinical trials exploring the potential of psilocybin and other psychedelics for the treatment of OCD, highlighting the importance of expanding our understanding of tolerability, safety, and efficacy to inform potential clinical approaches to the care of people living with OCD.
Methods
For this manuscript, the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines were used (Schulz et al., 2010, for the CONSORT Group).
Study design
This small Phase 1 clinical trial (ClinicalTrials.gov ID NCT03300947) consisted of two study phases (overviewed in Figure 1). Phase 1 involved a randomized, double-blind trial where participants were assigned to one of three conditions, 4 weekly oral administrations of low (100 µg/kg) or high dose (300 µg/kg) of psilocybin or an active placebo (lorazepam 1 mg). Phase 2 was single-blind (participants) in which participants received four additional administrations of high-dose psilocybin (300 µg/kg). In both phases, high-dose administrations had a minimum dose of 15 mg for individuals weighing less than 50 kg, and a maximum dose of 30 mg for individuals weighing more than 100 kg. We refer to these three groups as P + H, L + H, and H + H (Placebo Phase 1 + High dose Phase 2; Low dose Phase 1 + High dose Phase 2; and High dose Phase 1 + High dose Phase 2). This design provided for a minimum of 4 high doses (for those in the P + H group) to a maximum of eight high doses (for those in the H + H group) by the end of Phase 2. Participants were prospectively assessed for OCD symptoms following the end of treatment for 6 months. This design supported the aims to assess the initial safety, tolerability, and short-term efficacy while exploring various effects such as dose response effects, the impact of frequently repeated doses, and longer-term outcomes. Weekly sessions involved preparatory visits, drug administration, and follow-up days, maintaining uniform procedures across all study conditions. To increase the chances of preserving the blind, participants were informed that they may receive Lorazepam (“a calming medication which does not cause psychedelic experiences”), or one of two doses of psilocybin (“low dose may induce a partial and high dose may induce a full hallucinogenic/psychedelic experience”), and that they would not know what they would receive in any given session. No additional specifics of the trial design were provided.
Study design overview.
Participants
From among an initial pool of 578 individuals that initiated contact with the study team, 18 were deemed eligible for participation and completed the baseline assessment (see Figure 2 for CONSORT diagram). One participant’s symptom level dropped below the inclusion cutoff before Session 1 and was administratively withdrawn. Additionally, one participant withdrew from the study after one session in a placebo condition because of anxiety, intense emotional reaction, and feeling reluctant to continue. Another participant withdrew after one session in the Low-Dose condition due to difficulty with the protocol (being uncomfortable with sitting quietly for hours) and thus opted to pursue another treatment. Fifteen participants thus completed the double-blind placebo-controlled Phase 1 of the study (n = 5 per condition). One participant in the low-dose condition discontinued participation after four sessions, looking for other treatment options, and the remaining 14 participants completed the single blind high-dose Phase 2.
CONSORT diagram.
A power analysis for symptom change estimated that 15 participants would provide 80% power to detect a pre-to-post treatment effect. The effect size to inform this analysis was derived from Moreno et al. (2006), by pooling the symptom change seen in both the 100 and 300 µg/kg dosing sessions. The present study was not powered to adequately detect between-group differences across the Phase 1 RCT.
Participants were recruited through referrals from treatment providers within the community, family members, and self-initiated referrals, and the ClinicalTrials.gov website. All advertisements were Institutional Review Board (IRB)-approved. Enrollment started in Spring 2019 and continued till Spring 2022, with a 17-month pause during the height of the COVID pandemic.
Participants met the following inclusion criteria: (a) age 18–65 years, (b) moderate to severe OCD according to the Diagnostic and Statistical manual of Mental DIsorders, 5th Edition (DSM-5) criteria, confirmed through the Structured Clinical Interview for DSM-5 Research Version (SCID-5-RV), (c) YBOCS score of 16 or higher, (d) failed at least one previous OCD guideline concordant treatment, (e) considered safe for independent living, and (f) after meeting YBOCS severity criteria for entry, discontinued use of certain prescription and over-the-counter products or nutritional supplements at least 2 weeks before initiating the first session in Phase 1 to ensure safety and efficacy. This included discontinuing drugs for depression, lithium, drugs for psychosis, and glutamate-acting medications. Six participants required washout of prohibited medications involving a gradual approach, with weekly individualized dose reductions based on total dose, and tolerability of taper, lasting a minimum of 2 weeks prior to the last dose of discontinued medication. All prohibited medications were discontinued at least 2 weeks prior to the first double blind study drug administration session.
Exclusion criteria included: (a) personal or family history of psychosis or mania; (b) any medical illnesses that may complicate safety or drug metabolism, such as uncontrolled hypertension, severe cardiac disease, severe kidney or liver failure, or QTc greater than 450 ms in EKG; (c) psychiatric comorbidity that may represent an acute risk to their own or other’s safety, and those requiring sedative, narcotic, or neuroleptic medications on a regular basis; (d) pregnancy, breastfeeding, or being unwilling/unable to practice medically acceptable birth control during the study; (e) allergic to lorazepam; (f) history of a suicide attempt within the 12 months prior to screening; (g) any condition for which Magnetic Resonance Imaging (MRI) was contraindicated; (h) lorazepam contraindications; (i) active substance use disorder (within the last year assessed with SCID-5-RV; or (j) use of drugs of abuse listed in the current US DOJ DEA Drugs of Abuse Resource Guide (U.S. Department of Justice Drug Enforcement Administration, 2020) within the week prior to screening, including those categorized by the agency as cannabinoids, stimulants, opioids, sedatives, and hallucinogens.
Demographic and clinical characteristics are displayed in Table 1. Participants had been suffering from OCD symptoms for an average of 18.8 years (SD = 10.0), with current symptom severity in the severe range (x̅ YBOCS score at baseline = 28.6, SD = 5.2; ranged 19–39).
Participant demographic and clinical characteristics.
Comorbid psychiatric diagnoses: MDD: major depressive disorder; ED: eating disorder; PTSD: post-traumatic stress disorder; BPD: borderline personality disorder; GAD: generalized anxiety disorder; ADHD: attention deficit hyperactivity disorder; PMDD: premenstrual dysphoric disorder education: MA: master’s degree; MBA: master’s of business administration; TS: trade school; PhD: doctor of philosophy; HS: high school; JD: Juris doctor; BA: bachelor’s degree.
Relationship status/living situation: M/WC: married with children; NM: never married; RS: in relationship; DV: divorced; LA: living alone; SGL: single; PARENTS: living with parents; COHAB: living together.
Randomization and blinding
Randomization was prospectively stratified by YBOCS symptom severity defined as moderate (score = 16–23) or high (score ⩾ 24). Within each stratum, block randomization (fixed block size 3) with an equal allocation ratio (placebo, low, high) was used, so that over time participants in all three conditions remained relatively balanced. Only one participant qualified for the moderate-randomization schedule. The randomization schedule was computer-generated and under the control of the study pharmacist. The investigators and study personnel with direct participant contact were blinded to condition assignment, having never seen the randomization schedule and unaware of the block randomization approach.
Intervention procedure
Four double-blind testing sessions during Phase 1, and four modified single-blind sessions during Phase 2 were conducted 1 week apart. Each of those sessions consisted of a preparation session, an administration day, and follow-up discussion. At least one special preparatory visit was held prior to the first Test Session (Phase 1 Day 1) to provide the participants with further education about the psychedelic experience, to thoroughly discuss expectations, to establish and improve rapport, and to create comfort with the setting. The testing facilities were shown to the participants. Given the weekly cadence of sessions, subsequent preparatory visits were held at the discretion of the investigator or at the request of the participant.
The administration day consisted of an on-site visit lasting about 10 hours during which study drug ingestion took place, (12 hours for administration sessions 4 and 8). Sessions were conducted in a specially adapted room at the Clinical and Translational Sciences Research Center at the University of Arizona. Participants arrived at 7:00 am, after having had a light breakfast. Testing for drugs of abuse and, when applicable, for pregnancy, was conducted prior to each test session. Participants were escorted to the testing room where they remained for approximately 10 hours. Participants were allowed to use a nearby restroom, but no other outings were permitted. Vital measurements (pulse and respiration rate, blood pressure, and temperature) were taken prior to and at 1, 4, and 8 hours after drug administration. Symptom assessment (YBOCS) was taken prior to and at 4 and 8 hours after drug administration, with the YBOCS at hours 4 and 8 adapted to cover the period since the immediately prior assessment. Also prior to drug administration, suicidality, psychotic symptoms, and adverse events were assessed for the previous week.
Study drug in the form of a capsule was ingested with water at about 7:30 am. A light meal of their choice was made available for participants at about 11:30 am or by request later. Participants were asked to wear eyeshades when lying down, listen to a standardized set of music, and minimize interactions during the sessions. Participants were discouraged from sleeping during test sessions. Two trained facilitators, with at least a bachelor’s degree and training in mental health, were present in the room at all times. Whenever possible, the facilitators were a dyad of one male and one female sitter, including at least one of the study psychiatrists. Their role was observational and supportive, offering reassurance and maintaining participant safety while avoiding directive interventions, dynamic interpretations, or psychotherapy. After approximately 6 hours (around 1:30 pm), participants were allowed to modify the music selection, have more conversation, discontinue use of the eyeshades, and gradually debrief with the investigators regarding aspects of their experience. Debriefing involved a careful discussion about the nature of the experience for the individual. The facilitators heard the description of aspects of the experience the participants wished to disclose and worked to assist the incorporation from a possible altered state of consciousness to a common state of awareness.
After debriefing, at about 3:30 pm, all participants underwent a clinical assessment. In addition to the YBOCS, a psychiatric review of systems and mental status examination were conducted to ensure that participants were not experiencing hallucinogen intoxication or hallucinogen-induced symptoms or suicidal symptoms. Participants then filled out a set of questionnaires about the nature of their psychedelic experience. Once they were assessed to be safe by a study clinician, they were either escorted to their home/hotel or sent home with a trusted individual, and instructed not to drive or operate machinery for the next 24 hours. At the end of Sessions 4 and 8, participants participated in a post Phase 1, and post Phase 2 functional magnetic mesonance imaging (fMRI) and electroencephalogram (EEG) session. Results from these sessions are being reported elsewhere.
On the day after each session day, participants were given the opportunity for an in-person or remote visit for a discussion of the previous day’s experience and whether a preparatory visit prior to the next test session was necessary or desired. Participants were contacted by study psychiatrists for assessment of side effects, suicidality, and psychotic symptoms, and by a blinded rater for YBOCS completion. YBOCS were completed daily until the next session assessing OCD symptom severity since the previous rating session. Thus, from the start of Session 1 until a week following Session 8, participants received a YBOCS interview every day. During these intersession calls, participants were also given the opportunity to convey information to the study team or ask questions during this call. Self-report questionnaires for secondary outcomes were also completed each day between sessions (reported elsewhere).
Primary outcomes
Safety and tolerability were assessed with three clinician-administered tools. The Systematic Assessment for Treatment Emergent Events General Inquiry (SAFTEE-GI; Levine and Schooler, 1986) was administered by a study physician. Under the guidance of the U.S. Food and Drug Administration (FDA), the General Inquiry version of the SAFTEE scale was modified to include specific inquiry of adverse events of interest for psychedelic research such as depression, euphoric mood, elevated mood, aggression, feeling abnormal, feeling drunk, feeling of relaxation, abnormal thinking, psychosis, confusion, hallucination, inappropriate affect, impaired attention, disorientation, and hungover. Clinical raters conducted the Columbia-Suicide Severity Rating Scale (C-SSRS; Posner et al., 2011) to assess suicidal ideation (SI) and behavior, and administered the Psychotic Screening Module from the SCID-5-RV (SCID Screen; First, et al., 2015) to assess emergence of psychotic symptoms. These scales were administered at the beginning of every test session day, inquiring about the last week and recording long-lasting potential adverse events. Additionally, they were administered on the day following a test session inquiring about the time window since the last assessment, capturing acute adverse events surrounding the session. The SCID screen and C-SSRS were also administered during follow-up assessments to determine any effects on suicidality or psychosis associated with psilocybin ingestion.
The primary clinical outcome measure was the clinician-administered YBOCS. The YBOCS is designed to rate symptoms for the week prior to the assessment. In this study, the scale was used to rate the symptoms for the last week at Screening, Baseline, the morning of the session day before ingesting the drug, and during the 6-month of follow-up assessments. Modified timeframes were used for YBOCS ratings during the session day after drug administration, and for the daily ratings in between Sessions 1 and 8 (“since we last assessed these symptoms”). Inter-rater reliability was assessed for a subset of 22 of the YBOCS interviews (one interviewer, 2–5 raters) using the intraclass correlation (Shrout and Fleiss, 1979; ICC(1,1)), obtaining near-perfect agreement rICC = 0.987. These 22 interviews comprised recordings from all phases of the trial (Screening, Baseline, and Weeks 1–8).
Statistical analysis
The main approach to analyzing the time-varying data used mixed linear models with a between-participants factor Condition (Placebo, Low, High) and time as a within-person factor (ranging from baseline to 4 weeks or baseline to end of treatment depending on the specific research question). This approach accommodates missing data (one session for one subject, see Protocol Deviations below); four sessions for one other subject that discontinued after Phase 1. For symptom data a condition-by-time interaction would support the differential of one or both psilocybin conditions relative to placebo, although the limited power due to the small sample size would make finding the interaction unlikely due to low statistical power. Follow-up planned mixed linear model analyses within each condition (only time as a factor) provided more power to examine whether significant reductions occurred within each condition. Finally, response and remission rates were calculated using consensus criteria (Mataix-Cols et al., 2016). These symptom and remission-focused analyses were all preplanned and registered in ClinicalTrials.gov (NCT03300947). Similarly, reporting adverse events, suicidality, and any emergent psychotic symptoms were all preplanned.
Ethical considerations and participant protection
The study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki, Good Clinical Practice, Title 45 Code of Federal Regulations Part 46, and FDA Regulations, including IND approval. Approval for the study protocol and procedures was obtained from the University of Arizona IRB prior to initiation.
Additionally, the study adhered to the University of Arizona’s policies and procedures governing research conduct, including those related to IRB review, conflict of interest, and research integrity. To ensure participant confidentiality and data protection, all data were kept secure and anonymized throughout the study, and a data safety and monitoring board was convened to provide regular oversight. Any data shared publicly do not identify individuals.
Protocol deviations
One participant in the Phase 1 high dose condition (H + H) missed Session 2, but continued all subsequent sessions (3–8). Another participant in the same condition had a 2-week gap between Sessions 3 and 4 due to a holiday scheduling issue, but received all sessions in both phases. One other participant in the P + H condition had a 2-week gap between Sessions 7 and 8 in Phase 2 for personal reasons. All participants were included in all analyses.
Results
Safety and tolerability of the intervention
Adverse events
Events assessed by the SAFTEE were tallied by session type. There were 20 placebo sessions, 20 low-dose sessions, and 75 high-dose sessions (one participant missed one session, and one participant did not complete Phase 2). For each symptom, a chi-square test assessed whether symptom incidence differed by session type. Overall event frequency was low and seldom differed between session types. Table 2 presents individual symptoms as a function of session type. Three symptoms differed significantly between session types: Nausea, which was highest during low sessions, and itching and rash, which were highest during active placebo (lorazepam). Depression was numerically but non-significantly greater during either psilocybin dose assignment compared to placebo.
Adverse events reported in general inquiry by condition.
indicates statistically significant
Psychotic symptoms
The Psychotic Screening Module from the SCID-5-RV was utilized to explore psychotic symptoms. No participant reported any emergence of psychotic symptoms before, right after, or the week after each session, or during 6-month follow-ups.
Suicidality
The C-SSRS was used to assess SI and behaviors. The C-SSRS can provide several metrics. For this analysis, the presence or absence of five levels of ideation was characterized. The five levels are: Passive SI, Nonspecific SI with no plan or intent, Active SI with no plan, Active SI with some intent but no plan, and Active SI with specific plan and intent. Furthermore, it was documented whether there was a suicide attempt and whether there was non-suicidal self-injury (NSSI).
At baseline, the C-SSRS was used to assess the lifetime history of SI and behaviors. Although the sample met the inclusion criteria for non-acute SI, lifetime suicidal and behaviors were assessed at baseline to characterize the sample. The sample had lifetime reports of passive SI (7/15), nonspecific ideation with no plan or intent (2/15), and active ideation with no plan (1/15). The sample reported no lifetime history of Active SI with intent or plan. In terms of behaviors, 1 of the 15 participants had one prior suicidal attempt, and 1/15 had engaged in NSSI.
On the morning of the first session, prior to study drug administration, the C-SSRS inquired about the time since the baseline assessment. There were reports of passive SI (3/15), nonspecific ideation with no plan or intent (1/15), and active ideation with no plan (1/15). The sample reported no Active SI with intent or plan, and no attempts or NSSI.
Session-related changes in SI were quantified by subtracting the C-SSRS ratings taken the morning of the session day from the C-SSRS ratings at the end of the session day. A positive score would reflect higher SI following a session compared to before that session, and a negative score would reflect a decrease in SI from before to after the session. Results are presented by session type (Placebo, Low Dose, High Dose; see Table 3); Chi-square tests for each SI variable revealed no significant differences between the session types (Placebo, Low Dose, High Dose) in the pattern of increase-same-decrease for any of the five SI metrics.
Increases and decreases in suicidal ideation from start to end of session day by session type.
There were no reports at either time of this level of Suicidal Ideation.
Clinical laboratory assessments
The complete blood count (CBC) and Comprehensive Metabolic Panel (CMP) at Baseline vs. at the end of Sessions 4 and 8 yielded no significant change in any variable.
Efficacy of the intervention
Symptom severity
Symptom severity as measured with the YBOCS was examined for both day after session and week after session effects. Analyses are presented first for the first 4-week randomized control trial phase (Phase 1), and subsequently for the entire 8-week treatment duration (Phase 1 + Phase 2), with Phase 2 participants receiving the high dose in the final 4 weeks.
Next day effects
This effect was assessed by examining the baseline YBOCS score along with the scores on the day immediately after each session day in a condition (Placebo, Low Dose, High Dose) by time (5: baseline plus 4 weeks) mixed linear model (see Figure 3). Scores lowered significantly over time (F(4, 21.85) = 6.50, p < 0.001), and conditions differed significantly collapsed across time (F(2, 50.61) = 5.48, p = 0.007), but no significant condition by time interaction was present (F(8, 21.82) = 1.16, ns). The condition main effect reflected that collapsed across all 4 weeks, those receiving high doses (x̅ = 17.8) and those receiving low doses (x̅ = 20.6) had significantly (p < 0.05) lower YBOCS scores than those receiving placebo (x̅ = 25.9). Planned simple slope analyses indicated a significant reduction (F(4, 6.28) = 5.58, p = 0.03) in those receiving the high dose but no significant reduction in those receiving placebo (p = 0.58) or low dose (p = 0.21). Similarly, the analysis across the entire 8 weeks, after participants had received four high doses in Phase 2, with a condition (3) by time (9: Baseline plus 8 weeks) mixed linear model revealed the same pattern of findings: scores decreased over time (F(8, 22.05) = 8.74, p < 0.001) and conditions differed collapsed across time (F(2, 92.479) = 7.60, p < 0.001), but no condition by time interaction (F(16, 22.26) = 0.65, ns). The condition main effect reflected that collapsed across all 8 weeks, those in the H+H condition (x̅ = 14.4) and the L + H condition (x̅ = 17.6) had significantly (p < 0.05) lower YBOCS scores than did the P + H condition (x̅ = 21.9).
YBOCS scores by condition on the day following each session.
Next week effects
This effect was examined by including the baseline YBOCS score along with the YBOCS score 1 week after each session in a condition (Placebo, Low, High) by time (6: baseline, start of week 1, plus 4 weeks) mixed linear model (see Figure 4). Scores lowered significantly over time (F(5, 20.21) = 2.56, p < 0.001) and conditions did not differ significantly collapsed across time (F(2, 59.76) = 2.48, p = 0.092), and no significant condition by time interaction was present (F(10, 20.21) = 0.36, ns). Planned simple slope analyses indicated no significant reduction within any single condition, but the pooled low and high conditions (psilocybin N = 10 vs placebo N = 5) showed a significant reduction across the 4 weeks (F(1, 12.95) = 4.98, p = 0.044). Repeating the analysis with the entire 8 weeks in a condition (3) by time (10: Baseline, start of week 1, plus 8 weeks) revealed that scores decreased over time (F(9, 25.59) = 6.13, p < 0.001) but conditions did not differ significantly collapsed across time (F(2, 95.03) = 2.54, p = 0.084), and no significant condition by time interaction was present (F(18, 25.81) = 0.32, ns).
YBOCS scores by condition at 1 week following each session.
Enduring change: Response and remission
With this small power design, we collapsed the three conditions (P + H, L + H, H + H) to examine response and remission rates 1 week after the 8th (final) session, after all patients had completed at least 4 high-dose sessions, and also at 6-month post-intervention. For the one participant who only completed 4 weeks (in the L + H condition, thus receiving only low doses), we conservatively carried the last observation forward to estimate the end of treatment response and remission rates; this individual was omitted from the rates at 6 months. In line with an international expert consensus report (Mataix-Cols et al., 2016), remission was defined as a YBOCS score of ⩽12, response was defined as a ⩾35% reduction in YBOCS score from Baseline, and partial response was defined as a ⩾25% reduction in YBOCS score from Baseline. Table 4 shows categorical response and remission rates. Figure 5 depicts the individual participant trajectories from baseline to Week 8, and 6-month follow-up.
Remission of response rates for the entire sample at the end of treatment and after 6 months.
⩾equal or greater than.
Trajectory of individual participants from baseline to completion of treatment to 6-month follow-up.
Dose-response relationship
To account for the number of psilocybin ingestions, a post-hoc investigation involved correlating the number of psilocybin doses taken with the percentage of symptom reduction on the YBOCS scale. Rather than merging low and high doses, the low dose was considered as 1/3 of a high dose. This approach enabled a correlation analysis across a range of doses, including increments of 1/3 up to 8 high doses. The YBOCS score 1 week after each session was included along with the cumulative dose up to and including that session (see Figure 6). A significant positive correlation was observed between cumulative psilocybin exposure and OCD symptom reduction (r = 0.59, p < 0.01). To assess whether cumulative dose was related to long-term improvement, we correlated cumulative dose at the end of treatment with percent reduction in YBOCS scores at the 6-month follow-up (relative to baseline) among the 14 participants who completed all 8 weeks of the trial (r = 0.45, p = 0.051, one-tailed). In Figure 5, four of five participants in the H+H condition (highest cumulative dose) had 6-month YBOCS scores that were equal to or lower than their post-treatment scores, compared to 1 of 4 in the L + H condition and 0 of 5 in the P + H condition.
Relationship of symptom reduction and cumulative number of doses.
Exploratory analysis of medication discontinuation for responders and non-responders
To examine whether recent discontinuation of chronic SRI use at the time of study screening impacted treatment outcomes, Chi-square tests were performed to assess the relation of SRI use to responder status (35% reduction) and for partial responder status (25% reduction) in post hoc analysis. As shown in Table 5, individuals who had been on SRIs before the trial were more likely to be non-responders (X2 = 8.18, p = 0.004 and X2 = 5.63, p = 0.018, respectively).
Relationship of SRI treatment discontinuation status to response and remission.
⩾stands for equal or greater than.
Discussion
Safety and tolerability
Overall, multiple low and high doses of psilocybin ingested weekly were well-tolerated, with no serious adverse events and no significant differences in adverse events between the active drug and placebo conditions, except for a higher incidence of nausea during low-dose sessions. Across all psilocybin sessions, adverse event frequency remained low, and adverse event rates were lower than those reported in recent meta-analyses (e.g., Romeo et al., 2024), in spite of our use of a systematic assessment approach with the SAFTEE-GI and specific inquiry of adverse events of interest identified by the U.S. FDA. SI, assessed via the C-SSRS, did not significantly change within or across drug administration session days. Importantly, participants receiving high-dose psilocybin in Phase 2 showed no SI during these final four sessions, arguing against a cumulative risk of suicidality. While these findings are consistent with safety, they derive from non-significant results and a small sample, requiring replication. Psychotic symptoms were absent, and lab blood results remained stable across the study, further supporting tolerability.
These results provide valuable data, currently absent in the clinical research literature, suggesting that 4–8 high doses of psilocybin separated by 1 week are safe and well-tolerated.
Psilocybin’s impact on symptom reduction
RCT Phase 1
Psilocybin produced post-session day reductions in obsessive–compulsive symptoms across the first 4 weeks, with high-dose of psilocybin participants having a significant improvement, unlike their placebo or low-dose counterparts. Furthermore, averaged across the 4 weeks, both the high-dose and low-dose conditions had significantly lower YBOCS scores than the Placebo condition.
Psilocybin also produced meaningful symptom reductions measured 1 week after every session day during the first 4 weeks of the trial. Pooling low- and high-dose participants, significantly decreased OCD symptoms were observed across the 4 weeks.
Enduring effects
Assessing baseline to end of treatment at 8 weeks, there was a significant symptom reduction across all participants with no difference between conditions. With last observation carried forward (LOCF) (N = 15), 73% of participants achieved response, and 40% achieved remission. When considering completer analysis (N = 14), 71.4% achieved response, and 42.9% achieved remission. Prospectively naturalistic follow-up suggests long-term benefit is often sustained at 6 months.
Cumulative dose–response relationship
A cumulative dose–response relationship was observed: higher cumulative psilocybin exposure was significantly associated with greater symptom reduction across treatment (r = 0.59, p < 0.01). This finding suggests that individuals with OCD may experience greater benefit from receiving a higher number of psilocybin doses, an important consideration requiring further study. A non-statistically significant trend for greater symptom reduction from baseline at 6-month follow-up (r = 0.45, p = 0.051, one-tailed) is interesting as well. This trend, although not statistically significant, suggests that higher cumulative doses may predict more durable symptom reduction, with 4 of 5 participants in the H + H condition having 6-month YBOCS scores that were equal to or lower than their post-treatment scores, compared to 1 of 4 in the L + H condition and 0 of 5 in the P + H condition.
Predictors of response
Role of serotonin reuptake inhibitor discontinuation
There were only 3 non-responders (20%), compared to 40%–60% in clinical trials with serotonin reuptake inhibitors (Pallanti and Quercioli, 2006) in a sample of participants who had been characterized as having failed at least one standard treatment. Exploratory analyses suggested that recent treatment discontinuation of stable serotonin reuptake inhibitor use prior to study participation may negatively impact psilocybin response. All three non-responders had recently discontinued use of serotonin reuptake inhibitors as a condition for study participation, and chi-square analyses showed significantly lower response rates among these participants. Although the sample size limits firm conclusions, this aligns with findings from other trials suggesting potential serotonergic receptor desensitization following chronic serotonin reuptake inhibitor use (Erritzoe et al., 2024). This observation is in contrast with findings in healthy volunteers who undergo short-term pretreatment with serotonin reuptake inhibitors and report no impact of serotonin reuptake inhibitors on the subjective mood effects as part of the psychedelic experience (Becker et al., 2022). Interestingly as well, a survey of individuals undergoing psychedelic experiences compared individuals using drugs for depression and those who were not, found that participants who were free of drugs for depression reported greater intensity of subjective, mystical, challenging, and break-through experiences that are often associated with potential mechanism for change. However, there were no differences in positive mood effects among the two groups (Barbut Siva et al., 2024). This highlights the importance of adhering to standardized washout protocols, or stratified designs based on medication history, and underscores that recent serotonin reuptake inhibitor use and discontinuation may impact certain aspects of the experience and the efficacy of psilocybin treatment. Future trials could consider longer washouts or include arms where participants remain on stable medication.
Participant characteristics
No consistent clinical profiles emerged among the three participants who failed to show even a partial response, but in addition to serotonin reuptake inhibitor recent use and subsequent discontinuation, all three lacked prior psychedelic experience. One of them had a Borderline Personality Disorder diagnosis. These observations may warrant further exploration, as the current sample is too small to test these systematically.
Limitations
The small sample size (n = 15) limits statistical power and generalizability. Although large effects of psilocybin in OCD symptoms were observed, replication in larger trials is essential. Participant burden was high, for some requiring extended stays or weekly travel, which may have selected for highly motivated individuals, those with greater resources and support systems, potentially introducing participation bias. These may have impacted the participant socio-demographic characteristics, resulting in limited representation of individuals from groups historically underrepresented in biomedical research. This issue further limits the generalizability of the preliminary findings and must be intentionally addressed in future studies.
The study sample size also precluded a robust assessment of moderators like OCD duration, subtype, or comorbidities. The exploration of recent SRI discontinuation is also limited based on the minimal sample size, and the variability of length since pharmacological treatment among participants.
This study’s assessment burden, including the daily symptom assessment calls and numerous questionnaires, might have affected the participants’ ability to focus on the transformative experiences during sessions. Streamlining assessments and standardizing preparatory and integration protocols could enhance outcomes and reduce variability. Additionally, incorporating measures of expectancy and perceived dose would aid future studies at interpreting placebo effects and blinding fidelity.
Blinding participants in psychedelic trials remains controversial due to the distinctive effects of these substances, which can make it difficult to conceal treatment assignment. However, this study mitigated that concern by including both a low-dose psilocybin arm and an active placebo (lorazepam). We did not conduct formal prospective assessment of blinding, anecdotally, several participants reported difficulty distinguishing between low, high, and placebo doses. One participant who received eight high-dose sessions kept detailed notes of every session, and misidentified their dosage in six of eight instances, underscoring the variability in individual subjective response. This example highlights the perceptual ambiguity even with high doses and suggests that meaningful blinding is possible under carefully designed conditions. It also illustrates the complexity and variability of psychedelic experiences—both across individuals and within the same person across sessions—which presents ongoing challenges for placebo-controlled research in this field.
Future directions
Larger, multi-site trials are needed to replicate and extend these findings, and could focus on optimizing treatment protocols by examining session frequency and dosing schedules (e.g., optimal number of sessions and spacing between sessions). This exploration should incorporate variations in dosing regimens and integration strategies to establish the balance between therapeutic efficacy and sustainability of outcomes over time.
Furthermore, in maximizing the impact for individuals who might otherwise experience minimal psychedelic effects and symptom reduction, several strategies could be pursued. Tailoring interventions by employing personalized dosing protocols or integrating adjunct interventions might enhance the therapeutic outcomes for these individuals. Additionally, understanding the factors influencing the variability in psychedelic experiences and treatment responses, such as psychological predispositions or neurobiological markers, can guide the development of targeted interventions to optimize outcomes.
Conclusion
These findings support the promise of multiple doses of psilocybin as a safe, well-tolerated, and potentially effective treatment for OCD, with greater symptom reduction observed compared to the active control, lorazepam. The study’s design—including the use of a low-dose condition—enhanced blinding integrity and methodological rigor. Importantly, the observed link between greater cumulative psilocybin dose and symptom reduction suggests the causal role that psilocybin has in reducing OCD symptoms in this clinical sample. Together, these results underscore the need for larger trials to further investigate psilocybin’s efficacy and mechanisms and contribute to growing evidence for its potential integration into mainstream OCD treatment.
Footnotes
Acknowledgements
We are indebted to many individuals who assisted with various aspects of planning and implementing this study: Jessica Andrews-Hanna, PhD, Bryan Clines, Maria Ambrose, Maria Gordon, Marisa Fernandez, MD, Sarah Ortega, Ole Thienhaus, MD, Tyler Shifflett, MD, Scott Salomone, MD, Megan Whaley, MA, Madeline Sheffer, BA, Zoe Skinner, BA, Karla Ledezma, BA, Roman Palitsky, PhD, Deanna Kaplan, PhD, Saren Seely, PhD, Kelly Chen, MA, Michael Medrano, PhD, Savannah Boyd, MA, Riley O’Neill, MA, Hatty Lara, MA, Mairead McConnell, PhD, Brian Lord, PhD, Diheng Zhang, PhD, and Jessica Schachtner, MA.
ORCID iDs
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The Usona Institute Inc. provided all psilocybin required for the study free of charge.
Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Francisco A. Moreno, MD receives honoraria from LivaNova for his service as Chairperson of the Data and Safety Monitoring Board for the LivaNova RECOVER study.