Psilocybin efficacy and safety
8 evidence topics · 18 sources
Summary: COMP005 phase 3 design and Week 6 primary result
COMP005 is a phase 3, randomized, double-blind, placebo-controlled trial in which adults with TRD (2 to 4 inadequate prior treatment responses) were randomized 2:1 to a single dose of COMP360 25 mg or matched placebo, administered with psychological support. A total of 258 participants were dosed (171 COMP360, 87 placebo) at 40 sites in the United States according to the Form 10-K; the 2025 Form 8-K topline filing states 32 sites. Part A was blinded through Week 6, Part B was blinded from Week 6 to Week 26, and Part C was open-label from Week 26 to Week 52. The primary endpoint was the difference between arms in change from baseline in MADRS total score at Week 6, from a mean baseline score of 31.5 in both arms. The least squares mean difference at Week 6 was -3.6 points (95% CI -5.7 to -1.5; p<0.001). Differences at earlier visits were -4.7 at Day 2, -7.2 at Week 1, and -5.2 at Week 3 (each p<0.001). In a post hoc analysis, 25% of participants in the 25 mg arm had a MADRS reduction of at least 25% at Week 6. These results come from Form 8-K filings, the Form 10-K, a November 2026 investor presentation that describes them as preliminary, and Psych Congress Elevate congress abstracts.
Summary: COMP005 Part B and Part C through Week 52
In Part B, participants who did not remit by Week 6 or who later met relapse criteria could receive a second dose of their randomized treatment; 70% of the COMP360 arm and 53% of the placebo arm received a second administration. The congress abstract states that “numerical separation from placebo was maintained through Week 26”. No statistical analyses were performed after the Week 6 primary analysis, and Part B results are based on observed data only. The company reports that over 40% of participants who had a clinically meaningful MADRS reduction but had not remitted by Week 6 remitted after the second dose. Approximately 70% of participants entered open-label Part C, where 90 participants from the 25 mg arm (approximately 80% of those entering) and 52 from the original placebo arm (approximately 90%) received a 25 mg dose. The company reports an average 13-point MADRS reduction from baseline at Week 52 in the 25 mg arm and an average 10-point reduction at the end of Part C in the original placebo arm, with response rates of 40% to 45% and remission rates of approximately 30% between Day 1 and Week 6 after the Part C dose, based on the 142 participants dosed in Part C. These results come from company documents and a congress abstract.
Summary: COMP006 phase 3 design and Part A results at Weeks 6 and 9
COMP006 is a phase 3, randomized, double-blind trial in which adults with TRD were randomized 2:1:1 to two doses, 3 weeks apart, of COMP360 25 mg, 10 mg, or 1 mg. The company reports 581 dosed participants (296, 142, and 143) at 116 sites in the United States, Canada, and Europe; the registry lists actual enrollment of 572. Part A was blinded through Week 9, Part B through Week 26, and Part C is an open-label part from Week 26 to Week 52. The primary endpoint was the difference in change from baseline in MADRS total score between the 25 mg and 1 mg groups at Week 6, and Week 9 was a key secondary endpoint. At Week 6, the least squares mean difference was -3.8 points for 25 mg versus 1 mg (95% CI -5.8 to -1.8; p<0.001) and -2.5 for 10 mg versus 1 mg (95% CI -4.9 to -0.2; p=0.016), which the November 2026 presentation reports as -2.5 (-4.8, -0.2) with a p-value of 0.017. At Week 9, the differences were -2.9 (p=0.002) and -2.6 (p=0.016). In a post hoc analysis, 39% of the 25 mg arm had a MADRS reduction of at least 25% at Week 6. These results come from Form 8-K filings, company presentations, and a congress abstract.
Summary: COMP006 Part B through Week 26
In Part B, participants meeting prespecified criteria could receive retreatment with their randomized dose or an antidepressant, and protocol-allowed antidepressants were permitted in Parts B and C; up to 4 doses of COMP360 are possible across 52 weeks. In the 25 mg arm, 58% of participants received retreatment after Week 9. The company reports “consistent separation between the 25 mg and the 1 mg arm” through Week 26, and Part B results after Week 9 are observed data only. The 39% of the 25 mg arm with a clinically meaningful MADRS reduction at Week 6 maintained benefit, on average, through at least Week 26, and 28% of participants with a clinically meaningful reduction who had not remitted by Week 6 remitted after an additional dose in Part B. Participants had current depressive episodes lasting on average over three years and an average of more than six lifetime depressive episodes. The registry lists study completion as estimated for 2026-12 and shows no results posted. These results come from a Form 8-K filing and a company investor presentation.
Summary: phase 2b COMP001 and the COMP004 follow-up
COMP001 was a phase 2b, randomized, double-blind trial in which 233 adults with TRD received a single dose of COMP360 25 mg (n=79), 10 mg (n=75), or 1 mg (n=79) with psychological support at 22 sites in 10 countries. The mean baseline MADRS total score was 32 or 33 in each group. In the peer-reviewed publication, least squares mean changes from baseline to Week 3 were -12.0, -7.9, and -5.4; the difference was -6.6 for 25 mg versus 1 mg (95% CI -10.2 to -2.9; P<0.001) and -2.5 for 10 mg versus 1 mg (95% CI -6.2 to 1.2; P = 0.18). The authors state that in the 25 mg group, response and remission at Week 3, “but not sustained response at 12 weeks,” were generally supportive of the primary results. The company reports remission in 29.1% of the 25 mg group at Week 3 and sustained response at Week 12 in 20.3% versus 10.1% with 1 mg. COMP004, a peer-reviewed 52-week observational follow-up of 66 participants (58 from COMP001), reported median time to a depressive event for all 233 COMP001 participants of 92 days for 25 mg, 83 days for 10 mg (86 days in the Form 10-K), and 62 days for 1 mg; a post hoc analysis of enrolled participants gave 189, 43, and 21 days, based on 10 events after enrollment.
Summary: supporting studies COMP003, PsiDeR, and the phase 1 study
COMP003 was a phase 2, exploratory, open-label study without a comparator in which 19 participants with TRD received a single 25 mg dose while continuing a single SSRI. In its peer-reviewed report, the mean MADRS change from baseline at Week 3 was -14.9 (95% CI -20.7 to -9.2), and 42.1% of participants met response and remission criteria at Week 3. PsiDeR was a single-site, double-blind, placebo-controlled phase 2 feasibility trial sponsored by the South London and Maudsley NHS Foundation Trust and King’s College London, in which 60 participants were randomized 1:1 to COMP360 25 mg or placebo after withdrawal of antidepressants; COMPASS Pathfinder supplied the capsules and had no role in design, conduct, or analysis. The adjusted MADRS difference was -10.41 at Week 3 and -12.92 at Week 6, and at Week 3, 13 (43%) psilocybin versus 1 (3%) placebo participants met response criteria and 12 (40%) versus 1 (3%) met remission criteria. The COMPASS-sponsored phase 1 study (EudraCT 2018-000978-30) randomized 89 healthy participants to 25 mg, 10 mg, or placebo, given simultaneously to up to six participants, and found no clinically relevant between-group differences in CANTAB or emotional processing scores. All three studies are reported in peer-reviewed publications.
Summary: safety in the phase 3 trials
In COMP005, any treatment-emergent adverse event (TEAE) occurred in 120 (70.2%) COMP360 and 34 (39.1%) placebo participants through Week 6, and in 145 (84.8%) and 52 (59.8%) through Week 26. Headache, nausea, and visual hallucination were the most common TEAEs; 66% of TEAEs occurred on dosing days and 88% resolved within a day. Serious TEAEs through Week 26 occurred in 8 (4.7%) COMP360 and 2 (2.3%) placebo participants, including suicidal ideation in 4 (2.3%) COMP360 participants. In COMP006 Part A, any TEAE through Week 9 occurred in 274 (92.6%), 125 (88.0%), and 116 (81.1%) participants in the 25 mg, 10 mg, and 1 mg arms; 73% of TEAEs occurred on dosing days and 83% resolved within a day. Nausea (39.9% versus 10.5%) and blood pressure increased (10.5% versus 2.8%) occurred more often with 25 mg than with 1 mg. Through Week 26, serious TEAEs occurred in 17 (5.7%), 5 (3.5%), and 9 (6.3%) participants, with suicidal ideation in 4 (1.4%), 2 (1.4%), and 4 (2.8%), one suspected suicide in the 25 mg arm determined by the investigator as “not related to the treatment”, and one suicide attempt in the 1 mg arm. The data safety monitoring board found no evidence of a clinically meaningful imbalance in suicidality between arms in either trial. These data come from company filings and presentations.
Summary: safety in the phase 2 and supporting studies
In COMP001, adverse events occurred in 179 of 233 participants (77%), including headache, nausea, and dizziness, and suicidal ideation or behavior or self-injury occurred in all dose groups. Serious adverse events occurred in 5/79 (6.33%), 6/75 (8.00%), and 1/79 (1.27%) participants in the 25 mg, 10 mg, and 1 mg groups, including suicidal behaviour in 3/79 (3.80%) with 25 mg; the company states that these behaviours were reported at least one month after dosing in three non-responders. In COMP003, 12 of 19 participants (63.2%) had a TEAE and none was serious; three participants had blood pressure increased, of whom two had severe events treated with clonidine. In PsiDeR, four serious adverse events were recorded (one placebo, three psilocybin), all but one deemed not related to study medication, with the remaining event, in the placebo arm, deemed possibly related. In the phase 1 study in healthy participants, there were no serious TEAEs, 67% of adverse events started and resolved on the administration day, one participant had substance-induced psychotic disorder and received midazolam, and there were no clinically significant findings in vital signs. The COMP001 figures come from the peer-reviewed publication and the registry, except the timing of suicidal behaviour, which comes from a company press release; the other three studies are reported in peer-reviewed publications.