New 52-week data from a Phase 3 psilocybin trial suggest that some people with treatment-resistant depression may experience additional improvement after receiving another dose.
The findings are encouraging.
They do not establish that a few psilocybin treatments can reliably control depression for an entire year.
On September 9, 2026, Compass Pathways reported topline results from the final portion of COMP005, a Phase 3 trial evaluating COMP360, the company’s proprietary synthetic psilocybin formulation.
Among participants initially assigned to receive 25 milligrams of COMP360 who later received another active dose, depression scores were an average of 13 points lower than baseline at Week 52. Participants who originally received placebo and received their first 25-milligram dose during the open-label phase showed an average 10-point reduction at the end of the study.
Across everyone who received COMP360 during this final phase, response rates ranged from 40 to 45 percent during the six weeks after dosing. Approximately 30 percent met the company’s definition of remission.
These results offer an important early look at how repeat dosing might work over time.
But the 52-week findings came from an open-label phase without a concurrent placebo comparison. They were released by the trial sponsor as topline data and have not yet undergone the scrutiny of a complete peer-reviewed publication.
The appropriate conclusion is not that COMP360 has demonstrated one year of continuous benefit from a single treatment.
It is that a structured model involving infrequent, clinically supervised dosing may help some patients maintain or deepen improvement over a year, and that this possibility now deserves careful examination of the complete data.
What Treatment-Resistant Depression Means
Treatment-resistant depression generally refers to major depressive disorder that has not responded adequately to at least two appropriate courses of antidepressant treatment.
It is not simply depression that has lasted a long time or felt particularly severe. The term describes a population whose symptoms have persisted despite multiple treatment attempts.
These patients may cycle through medications, psychotherapy, transcranial magnetic stimulation, electroconvulsive therapy, ketamine, or other interventions. Some experience partial relief, while others remain significantly impaired.
Compass estimates that approximately four million people in the United States live with treatment-resistant depression, although prevalence varies depending on how resistance is defined and measured.
This is an important context for interpreting the trial. COMP005 was not testing psilocybin among healthy volunteers or people with mild, temporary distress. It enrolled patients whose depression had not responded adequately to established treatments.
A meaningful improvement in this population would be clinically important.
The question is whether the study design allows us to determine how much improvement resulted from COMP360, how durable it was, and which patients were most likely to benefit.
How the COMP005 Trial Was Structured
COMP005 enrolled 258 adults with treatment-resistant depression across the United States.
The trial was divided into three parts.
During Part A, participants were randomly assigned in a two-to-one ratio to receive either a single 25-milligram dose of COMP360 or an inactive placebo. This portion remained blinded through Week 6.
During Part B, which continued through Week 26, eligible participants could receive another dose of the treatment to which they had originally been assigned. Someone in the COMP360 group could therefore receive a second active dose, while someone in the placebo group could receive another placebo.
Part C covered Weeks 26 through 52. This phase was open-label, meaning participants and researchers knew that the treatment being offered was COMP360. Eligible participants from either original group could receive one 25-milligram dose.
Approximately 70 percent of the original study population entered Part C.
Among those who originally received COMP360, about 90 people received another active dose during Part C. Depending on whether they had also been redosed during Part B, this was their second or third COMP360 treatment.
Among participants originally assigned to placebo, approximately 52 received COMP360 during Part C. This was their first active 25-milligram dose.
The design makes it possible to observe what happened after repeat treatment and after placebo recipients crossed over to active treatment. It does not preserve the randomized placebo comparison through the full year.
That difference is essential.
Depression Scores Improved After Additional Dosing
The trial used the Montgomery-Åsberg Depression Rating Scale, or MADRS, to measure depression severity.
MADRS is a clinician-administered scale containing ten items. Total scores can range from zero to 60, with higher scores indicating more severe depression.
Compass reported that participants initially assigned to the 25-milligram group who received another dose during Part C had an average MADRS reduction of 13 points from their original baseline at Week 52.
Participants originally assigned to placebo who received their first COMP360 dose during Part C showed an average 10-point reduction from baseline by the end of the study.
These changes are potentially meaningful, particularly in a treatment-resistant population.
However, an average change does not tell us how every participant responded. Some may have improved substantially, some modestly, and some not at all. A mean score can also be affected by who remained in the study and qualified for further treatment.
Compass separately reported response and remission rates among all participants who received 25 milligrams during Part C.
A response was defined as a reduction of at least 50 percent from baseline in the participant’s MADRS score. Between 40 and 45 percent met this threshold at time points during the six weeks after dosing.
Remission was defined as a MADRS score of 12 or lower, with no individual item rated above a specified level. Approximately 30 percent met this definition during the same period.
Those percentages also mean that most participants did not meet the remission threshold.
That does not imply they received no benefit. A person may experience a meaningful reduction without reaching formal remission. It does show why response distributions, relapse patterns, functional outcomes, and individual treatment histories will matter when the complete data become available.
The Results Do Not Mean One Dose Lasted for a Year
The phrase “durability out to one year” can easily be misunderstood.
Many participants received more than one treatment during the study. Someone initially assigned to COMP360 could receive a second dose during the blinded period and another during the open-label phase, for a potential total of three doses over 52 weeks.
The reported Week 52 improvement therefore should not be presented as evidence that a single psilocybin dose continued working unchanged for a full year.
A more accurate interpretation is that the treatment model allowed additional dosing when participants met the study’s eligibility requirements, and some participants showed sustained or deepened improvement across that longer course.
This distinction matters clinically and economically.
A treatment that requires one session per year would create a different care model from one requiring two or three administration sessions, each accompanied by screening, preparation, several hours of supervision, and follow-up.
The number and timing of repeat doses may also differ among patients. Some may maintain improvement after one treatment, while others may relapse quickly, require redosing, or never respond sufficiently.
The trial may help inform a future retreatment strategy, but the topline announcement does not yet provide enough detail to define one.
The Open-Label Phase Limits What Can Be Concluded
The first portion of COMP005 was randomized, double-blind, and placebo-controlled. Those features help reduce bias and allow a more credible comparison between groups.
Part C did not maintain those conditions.
Every participant receiving treatment during this phase knew they were receiving 25 milligrams of COMP360. Researchers also knew the treatment being administered.
That creates several interpretive challenges.
Expectations may influence symptom reporting, behavior, and therapeutic interactions. Clinicians rating depression may also be affected, consciously or unconsciously, by knowledge of treatment assignment.
Participants who continued into Part C may differ from those who left the study. Approximately 30 percent of the original sample did not enter this phase. Without complete information about discontinuation, eligibility, relapse, adverse events, and missing data, it is difficult to know how attrition affected the results.
Retreatment was also not randomly assigned during Part C. Eligibility depended on clinical circumstances and the trial protocol. People receiving another dose may therefore represent a selected subgroup rather than the entire original population.
The placebo group’s improvement after crossing over to COMP360 is supportive but not equivalent to a randomized comparison. By that point, participants had spent months in a clinical trial, received repeated assessments, and may have developed strong expectations about finally receiving active treatment.
These limitations do not make the results meaningless.
They mean the open-label data are better suited to exploring longer-term treatment patterns, feasibility, and safety than to producing a definitive estimate of efficacy at Week 52.
The Data Have Not Yet Been Peer Reviewed
The 52-week findings were announced through a Compass Pathways press release and discussed by Psychedelic Alpha.
A company announcement is a primary source for what the sponsor reported, but it is not the same as a complete scientific paper.
Topline releases usually provide selected headline outcomes before all analyses, tables, adverse-event details, protocol deviations, missing-data methods, and subgroup findings are publicly available.
Compass itself cautioned that the full results may not be consistent with the initial topline description.
Independent evaluation will require more information about:
The number of participants assessed at every time point
Reasons for withdrawal and exclusion from retreatment
Outcomes among participants who were not redosed
Differences between second and third doses
Time to relapse after each treatment
Concomitant medications and additional care
Functional outcomes beyond symptom scores
Suicidal thinking and behavior
Serious and treatment-emergent adverse events
Whether benefits differed by demographic or clinical characteristics
Until those details are published, the results should be described as preliminary Phase 3 topline data.
Safety Requires More Than the Absence of a New Signal
Compass reported that the safety profile during Part C was consistent with earlier phases and that no new safety signals were identified.
That is reassuring, but it is not a complete safety analysis.
“No new safety signal” generally means that investigators did not identify an unexpected pattern requiring a new category of concern. It does not mean there were no adverse events, no serious events, or no participants who worsened.
Psilocybin can produce acute anxiety, fear, confusion, increased blood pressure, nausea, headache, and psychologically intense experiences. People with depression also require careful monitoring for deterioration, suicidality, mania, psychosis, and prolonged distress.
Repeat dosing introduces additional questions. An adverse event may occur after a later exposure even when earlier sessions were tolerated. A participant’s medical condition, medications, support network, or psychiatric state may also change between treatments.
The complete safety report will need to show not only what happened on dosing days, but what emerged during the weeks and months that followed.
For clinicians and training programs, safety should be understood as a longitudinal responsibility rather than an event limited to the administration room.
Repeat Dosing Could Change the Care Model
One of the most consequential aspects of the COMP005 data is the possibility that psilocybin treatment may become an episodic intervention rather than a single transformative event.
The public conversation often focuses on whether one dose can produce lasting change. Clinical practice may be more complex.
If approved, some patients may receive an initial treatment, improve, and remain stable for an extended period. Others may need another dose after symptoms return. Some may require a different treatment entirely.
This resembles other areas of medicine in which response determines what happens next. The difference is that psychedelic administration requires substantial infrastructure.
Each treatment may involve medical and psychiatric screening, preparation, medication review, a supervised session lasting several hours, transportation planning, follow-up assessment, and psychological support.
Repeat dosing therefore affects staffing, scheduling, cost, insurance coverage, and workforce capacity.
A viable model will need clear criteria for retreatment. Symptoms alone may not be sufficient. Clinicians may also need to consider functioning, prior response, adverse effects, medical changes, patient preference, and the availability of appropriate support.
The field should avoid assuming that more doses necessarily produce more benefit. Retreatment needs to be based on evidence rather than enthusiasm.
Psychological Support Remains Part of the Intervention Context
COMP360 is being developed as a drug product, but it is not administered like a conventional daily antidepressant.
Participants receive preparation and support within a structured clinical environment. They remain under supervision during the acute psychedelic experience and receive follow-up afterward.
This means the study did not test a capsule in isolation.
The therapeutic environment, professional relationships, expectations, and post-session processing may all influence outcomes. The trials have not fully separated the contribution of each element.
That uncertainty has practical consequences.
If regulators eventually approve COMP360, the product label and any risk-management requirements may define where it can be administered, who must be present, and what monitoring is required. Payers will then need to decide which surrounding services they will reimburse.
A repeat-dosing model increases the importance of these decisions. Patients may need to navigate the entire care pathway more than once.
The treatment’s real-world value will depend not only on its effect on MADRS scores, but on whether health systems can deliver it safely, consistently, and affordably.
What This Means for Psychedelic Practitioners
The 52-week data reinforce several principles relevant to psychedelic care.
First, long-term outcomes require long-term measurement. A strong response several weeks after treatment does not show whether someone remained well, relapsed, received another intervention, or experienced delayed harm.
Second, retreatment should not be treated as a routine continuation session. Each additional dose requires renewed screening and informed consent. Medical conditions, medications, psychiatric symptoms, and life circumstances may have changed.
Third, nonresponse needs its own care pathway. If most participants did not reach remission, programs must be prepared to identify when another dose is unlikely to be sufficient and when referral or a different treatment is needed.
Fourth, outcome monitoring should extend beyond symptom scales. Work, relationships, sleep, substance use, quality of life, and healthcare use may reveal benefits or problems that a depression score does not capture.
Finally, scope of practice remains essential. Facilitators and support providers should understand the boundaries between psychological support and the diagnosis or treatment of a depressive disorder. People with treatment-resistant depression may require medical oversight and coordinated psychiatric care before, during, and after psychedelic treatment.
An Important Signal, Not the Final Answer
The COMP005 follow-up moves the field closer to understanding what longer-term psilocybin treatment might look like.
The results suggest that another 25-milligram dose may deepen or extend improvement for some people with treatment-resistant depression. They also suggest that patients crossing from placebo to active treatment may experience a rapid and potentially durable reduction in symptoms.
But the one-year headline needs context.
The final 26 weeks were open-label. Only about 70 percent of the original participants entered that phase. Retreatment was selective rather than randomly assigned. Some participants received multiple doses. The findings are topline results released by the sponsor and have not yet been presented in a complete peer-reviewed paper.
These facts do not erase the potential value of the data.
They define the questions that should come next.
How many patients maintain improvement after one dose? Who benefits from a second or third treatment? When should redosing occur? What happens to people who do not respond? Which adverse events emerge between sessions? How much supportive care is necessary? Can the model be delivered equitably?
If COMP360 receives regulatory approval, these will become practical questions for clinicians, payers, training programs, and patients.
The promise of infrequent treatment is compelling. Its success will depend on whether the field can match that promise with transparent evidence, careful patient selection, reliable follow-up, and systems designed for both response and nonresponse.
Sources
Compass Pathways: 52-Week Topline Results from the Phase 3 COMP005 Trial
Psychedelic Alpha: Analysis of Compass Pathways’ Second Phase 3 Readout
This article is for educational purposes and does not constitute medical advice. COMP360 is an investigational treatment and has not been approved by the FDA. Psilocybin remains legally restricted in many jurisdictions.
