Safety, feasibility, and preliminary clinical outcomes of psilocybin-assisted therapy for veterans with severe, treatment-resistant PTSD: an open-label pilot clinical trial

Post-traumatic stress disorder (PTSD) is described as a severe and often chronic condition, with particularly high prevalence, symptom burden, treatment resistance, and suicide risk among U.S. military Veterans. The authors note that standard psychotherapies and SSRIs often do not work well enough in Veterans, and that psychedelic-assisted therapy has shown promise in other populations, but there had not yet been a U.S. Veteran-specific trial of psilocybin-assisted therapy for PTSD. They also highlight unresolved questions about the contribution of the psychotherapy component and the role of expectancy in psychedelic treatment outcomes. Against this background, Armstrong and colleagues set out to test the safety, feasibility, and preliminary clinical effects of psilocybin-assisted therapy in Veterans with severe, treatment-resistant PTSD. The study was designed as an open-label pilot trial using two psilocybin dosing sessions combined with preparation and integration therapy, with primary attention to adverse events and suicidal ideation/behaviour, and secondary attention to clinician- and participant-rated PTSD symptom change. The paper frames this as a proof-of-concept study intended to inform larger controlled trials.

Armstrong and colleagues conducted an open-label pilot trial at the Center for Psychedelic Drug Research and Education in Columbus, Ohio, with ethical approval from The Ohio State University Biomedical Sciences Review Board and preregistration at ClinicalTrials.gov. The study recruited U.S. military Veterans aged 21-64 with a DSM-5 diagnosis of PTSD for at least six months, a CAPS-5 severity score of at least 35, and treatment-resistant symptoms despite prior SSRI/SNRI treatment and psychotherapy. Key exclusions included psychotic spectrum or bipolar disorders, epilepsy, recent moderate or severe substance use disorder, significant suicide attempts, serious cardiovascular disease, and recent or extensive psychedelic use. Participants also needed to discontinue serotonergic medications before enrolment when applicable. Recruitment ran from January 2023 to April 2025. Of 3,628 people who clicked the online screening link, 668 completed eligibility assessment, 52 were invited for in-person screening, 13 enrolled, and 12 completed the intervention and primary outcome assessment. The intervention period lasted about 11 weeks and included at least 13 visits. Participants received four preparatory therapy visits totalling 8 h before dosing, then two psilocybin sessions spaced 2-3 weeks apart: 15 mg in the first session and 25 mg in the second. These were followed by integration therapy after each dosing day, with 6-8 h of post-session therapy in total. Sessions took place in a comfortable, living room-like setting with two facilitators present, while participants were encouraged to use eyeshades, headphones, and a music playlist. Facilitators were trained study staff, supervised by the principal investigator, and at least one licensed mental health professional was present at each therapy visit. The primary endpoint was safety, including adverse events and suicidal ideation/behaviour measured with the Columbia Suicide Severity Rating Scale (C-SSRS), from baseline to one month after the second psilocybin session. Secondary outcomes were PTSD severity measured by the CAPS-5 and PCL-5 at baseline, during preparation, and at follow-up. Expectancy was assessed with a single item from the Credibility/Expectancy Scale before the first dosing session. Physiological safety was monitored through blood pressure and heart rate during dosing sessions. The statistical plan included descriptive summaries, paired t-tests for peak physiological measures, repeated-measures ANOVAs for CAPS-5 and PCL-5, mixed-effects models to account for within-person change, expectancy, and pre-dose symptom reduction, and calculation of response, clinically meaningful response, and remission rates. Because this was a small pilot study, the authors used a conservative 50% reduction threshold for treatment response.

Twelve participants completed the study and were analysed. They were on average 37.7 years old, mostly White (92%), male (75%), and educated to bachelor’s degree level or above (75%). Most had combat-related trauma, and the average baseline CAPS-5 score was 39.7, indicating severe PTSD. Only two participants reported previous minimal psychedelic use. Therapy adherence was high, with an aggregated adherence rating of 98% across visits. No serious adverse events occurred. The study-related adverse events were mild to moderate and transient, most commonly headache, anxiety, and dizziness. These events appeared within a week of dosing and resolved spontaneously or with non-medical support. Heart rate and blood pressure remained within acceptable limits during dosing sessions, and no concerning physiological trend was observed. The highest recorded heart rate was 121 bpm, and the highest recorded blood pressures were 166 mmHg systolic and 103 mmHg diastolic. Suicidality did not worsen over the trial. C-SSRS scores did not significantly change from baseline to the final preparatory visit or from the final preparatory visit to one month after the second psilocybin session. Baseline lifetime suicidal ideation was reported by 83% of participants and past-month ideation by 17%. After dosing, 75% reported no suicidal ideation, and at one month 25% reported some ideation, all of it passive. PTSD symptoms improved substantially. CAPS-5 scores showed a significant time effect, with a mean reduction of 27.5 points from baseline to one month after session 2 (95% CI 19.9 to 35.1, p < 0.001, Cohen’s d = 2.30). There was also a smaller but significant reduction from baseline to the final preparatory visit, suggesting some improvement during preparation therapy before psilocybin administration. The mixed-effects model confirmed a significant time effect, while expectancy was not a significant predictor. Pre-dose symptom reduction during preparation did predict later CAPS-5 change at one month. Self-rated PTSD symptoms on the PCL-5 also declined significantly over time. Scores fell significantly from baseline to one month after session 2 and also at one week after session 2, indicating a rapid effect. Unlike CAPS-5, baseline-to-preparation change on the PCL-5 was not statistically significant. In the mixed-effects model, neither pre-treatment symptom change nor expectancy predicted PCL-5 outcomes. By the primary endpoint, 75% of participants met the authors’ conservative response criterion and were in remission, and 83% met the criterion for a clinically meaningful response.

The authors interpret this open-label pilot as the first clinical trial of psilocybin-assisted therapy specifically in U.S. military Veterans with severe, treatment-resistant PTSD. They argue that the combination of two synthetic psilocybin doses and about 16 h of psychotherapy produced large reductions in PTSD symptoms at one month, without increases in suicidal ideation and without severe adverse events. They present the findings as consistent with, and extending, earlier psychedelic-assisted therapy research in depression, anxiety, MDMA-assisted PTSD treatment, and prior psilocybin studies in trauma-affected groups and adults with PTSD. Armstrong and colleagues emphasise that the Veteran population is distinct because of high symptom chronicity, treatment resistance, and suicide risk. They suggest that the high response and remission rates indicate meaningful clinical change and that psilocybin-assisted therapy may be well suited to the symptom cluster of severe PTSD in this group. At the same time, they note that the effect sizes were very large and compare favourably with psychotherapy-alone and pharmacological PTSD studies, while also acknowledging that small uncontrolled trials can overestimate effects. A key interpretative point in the Discussion is that symptoms improved during the preparatory psychotherapy phase, before drug administration, and that greater pre-dose improvement predicted better one-month outcomes on CAPS-5. The authors take this as evidence that the therapeutic context, alliance, and preparation may contribute materially to outcomes, not just the drug itself. They also report that expectancy did not significantly predict outcomes, but caution that the small sample and open-label, non-blinded design mean this cannot be taken as definitive evidence that expectancy is unimportant. The main limitations acknowledged are the very small sample size, lack of a control group, open-label design, short follow-up, self-referred recruitment, possible selection bias, and restricted generalisability due to stringent exclusion criteria. They also note that expectancy was measured with only a single item, the sample was predominantly White, and the two-dose escalating regimen may have influenced outcomes and cost. Strengths included the independent rater for CAPS-5 and the ability to assess symptom change during preparation before the first psilocybin dose. The authors conclude that larger randomised controlled trials are needed to confirm efficacy, clarify mechanisms, examine dosing strategies, and determine whether different psychotherapies or specific subgroups might benefit most.

The authors conclude that this proof-of-concept trial suggests psilocybin-assisted therapy can be delivered safely to Veterans with severe, treatment-resistant PTSD and may produce remission in a substantial proportion of patients. They state that the findings justify larger, better-powered, randomised and blinded trials to test efficacy more rigorously, improve understanding of mechanisms, and refine dosing and psychotherapy approaches.