Psilocybin-assisted therapy for major depressive disorder: An exploratory placebo-controlled, fixed-order trial

Earlier clinical work has reported rapid and sometimes durable antidepressant effects after one or two doses of psilocybin given with psychological support, but many of those studies lacked placebo control, had limited measures to preserve blinding, used heterogeneous psychotherapeutic approaches, and in some cases assessed long-term outcomes after crossover or delayed-treatment periods that complicate interpretation. In particular, uncertainty remains about how much of observed clinical benefit is attributable to the pharmacological effects of psilocybin versus expectancy, nonspecific “set and setting” influences, or the psychosocial elements of therapy. Sloshower and colleagues designed an exploratory trial to address some of these gaps by combining a placebo-controlled design with enhanced blinding procedures and a manualised psychotherapeutic framework that incorporated elements of Acceptance and Commitment Therapy (ACT). Although the study was powered primarily to investigate electrophysiological markers of neural plasticity, the current report presents the clinical outcomes from that trial, focusing on depressive symptoms, anxiety, quality of life, acute subjective effects, blinding success, and safety following placebo and a single moderate dose of psilocybin administered within a structured therapy protocol.

This was a placebo-controlled, within-subject, fixed-order trial conducted under institutional review board approvals and an FDA investigational new drug application. Participants who met DSM-5 criteria for Major Depressive Disorder via SCID-5-CT, were aged 18–65, had a current moderate to severe depressive episode of at least 6 weeks (GRID-HAM-D-17 score ≥17), and had failed at least one adequate antidepressant trial in the current episode were eligible. Candidates were required to discontinue conventional antidepressant or antipsychotic medications at least 2 weeks prior to enrolment (4 weeks for fluoxetine). Exclusions included a primary psychiatric diagnosis other than MDD, active substance use disorder, personal or family history of psychosis or bipolar disorder, unstable medical/neurological disease, or psilocybin exposure within the prior year. The extracted text refers to supplemental material for full criteria. Recruitment occurred from autumn 2018 to spring 2021 via online postings, referrals, and word of mouth. Interested individuals underwent telephone prescreening, formal consent, structured psychiatric assessment, medical screening (labs, ECG), and communication with outpatient clinicians as needed. Of 949 assessed, 22 were enrolled; 19 completed at least one dosing session and 15 completed both sessions. The study used a within-subject fixed-order design with placebo administered first and psilocybin second, with dosing sessions approximately 4 weeks apart and embedded within an eight-session psychotherapy protocol. A primary rationale for the fixed order was to reduce functional unblinding and carryover effects that can follow psychedelic doses. To enhance blinding and reduce expectancy effects, participants and most staff were told they would receive two of three possible dosing conditions in random order—placebo, low-dose psilocybin (0.1 mg/kg), or moderate-dose psilocybin (0.3 mg/kg)—but only placebo then moderate-dose (0.3 mg/kg, maximum 35 mg) were actually administered. The low dose was never given; only pharmacists and a limited set of investigators knew the true fixed order. Psychotherapeutic support was manualised and ACT-informed. Each dosing phase was flanked by a 2-hour preparatory psychotherapy session and two 1-hour debriefing/integration sessions (1 day and 1 week post-dose), with two additional integration sessions after the primary outcome period; therapists and a study psychiatrist attended all sessions. Dosing procedures included on-site monitoring, music playlists, support from therapist and psychiatrist in a non-directive manner, vital-sign checks, and minimum 6-hour observation. Rescue medications were available. Primary clinical outcomes included the clinician-rated GRID-HAM-D-17 collected before dosing and for 2 weeks after each dose (through week 6), and the self-rated QIDS-SR-16 collected throughout the study to week 16. Anxiety was assessed with the HAM-A through week 5 and health-related quality of life with the RAND-36 through week 16. Acute subjective effects were measured with the MEQ-30 and CEQ on dosing days, plus questionnaires about abuse liability and guessed dosing condition. Safety monitoring included adverse event recording and frequent vital signs. Analyses were conducted on participants who completed at least one dosing session (intention-to-treat), using linear mixed models with drug and time as within-subject factors, selection of the best-fitting variance-covariance structure by information criteria, and effect sizes (d′) calculated from raw means and SDs. Correlations between MEQ and symptom change were assessed with Pearson or Spearman tests as appropriate. Given the exploratory nature and small sample, tests used a two-tailed alpha of 0.05 with no correction for multiple comparisons.

Enrollment and sample. Of 949 people screened by phone, 42 were screened in person and 22 enrolled; 19 completed at least one dosing session and 15 completed both. The analysed sample (n = 19) had a mean age of 42.8 (SD 13.8) years, included 13 women and 6 men, and averaged 20 ± 12 years since depression diagnosis. Mean GRID-HAM-D-17 at screening was 23.9 ± 5.0. Prior lifetime psychedelic exposure was reported by 8/19 (42.1%), but none within the preceding 3 years. Mood and depression outcomes. On the clinician-rated GRID-HAM-D-17 there were significant main effects of time (F(3,99)=17.8, p<0.0001) and drug (placebo versus psilocybin; F(1,99)=49.9, p<0.0001), but no significant drug-by-time interaction (F(3,99)=1.2, p=0.31). Mean reductions in GRID-HAM-D-17 over the 2-week period after dosing were larger following psilocybin (Δ = 6.3–8.7) than after placebo (Δ = 4.4–5.8), yet this difference did not reach statistical significance. Reported effect sizes (d′) were greater post-psilocybin (1.02–2.27) than post-placebo (0.65–0.99). Among the 15 participants who completed both sessions, all achieved a clinically significant response (≥50% reduction in GRID-HAM-D) by the end of week 6: 5 participants (33.3%) responded after placebo prior to psilocybin, and 10 (66.7%) responded only after psilocybin. Remission (GRID-HAM-D-17 ≤7) occurred in 3 (20%) after placebo and in 7 (46.7%) only after psilocybin; by week 6, 10 (66.7%) retained response and 5 (33.3%) remained in remission. Self-reported depression (QIDS-SR-16) decreased over time (F(3,101)=16.5, p<0.0001) and was significantly lower following psilocybin than post-placebo through week 6 (F(1,101)=43.3, p<0.0001), though there was no drug-by-time interaction (F(3,101)=1.3, p=0.29). Extending QIDS-SR-16 to week 16 showed a rapid decrease one day after psilocybin that remained statistically significant through week 12 (time effect F(6,75)=7.47, p<0.0001). Anxiety and quality of life. HAM-A scores declined over time (F(2,75)=14.3, p<0.0001) and were lower following psilocybin versus placebo (F(1,75)=52.9, p<0.0001), but no significant drug-by-time interaction was found (F(2,75)=0.04, p=0.97), and mean anxiety reductions in the week after dosing were similar for placebo and psilocybin. RAND-36 analyses showed significant improvements over time on seven of eight subscales maintained through week 16 (except general health); four subscales (role limitations due to emotional problems, social functioning, general health, emotional well-being) were significantly improved after psilocybin compared with after placebo. Acute subjective effects and blinding. MEQ-30 scores were substantially higher on the psilocybin day (mean 70.0 ± SE 4.9) than placebo (8.4 ± 4.5) (F(1,13)=97.3, p<0.0001). CEQ scores were also higher after psilocybin (28.9 ± 4.8) versus placebo (9.6 ± 2.0) (F(1,18)=17.7, p=0.0005). Maximum MEQ correlated negatively with change in HAM-D after placebo at 1 day (r = -0.55, p = 0.017) and 2 weeks (r = -0.75, p = 0.005), but there were no significant correlations between MEQ or CEQ and symptom changes after psilocybin. Blinding was limited: after placebo sessions 15/19 (78.9%) participants correctly guessed they had received placebo, and after psilocybin sessions 12/15 (80%) correctly guessed they had received the higher dose. Masking among therapists and other staff was not formally assessed. Safety and adverse events. No severe adverse events occurred during dosing sessions and no rescue medications were administered acutely. One participant required psychiatric hospitalisation about 2 weeks after psilocybin for lack of improvement. Adverse events were more frequent on psilocybin days (25 events) than placebo days (3 events) (z = 3.01, p = 0.003); the most common were transient mild-to-moderate headaches, transient anxiety, and dysphoria. Transient elevations in systolic blood pressure (>160 mmHg at one or more time points) occurred in 2/19 (10.5%) placebo sessions and 5/15 (33.3%) psilocybin sessions; elevated diastolic pressure (>100 mmHg) occurred in 2/19 (10.5%) placebo sessions and none during psilocybin sessions. None of the cardiovascular events required medical intervention. No prolonged perceptual changes, persistent psychotic symptoms, or dependency-related behaviours were observed, though several participants expressed a desire for further supervised psilocybin sessions.

Sloshower and colleagues interpret the findings as evidence that a single moderate psilocybin dose given within a manualised ACT-informed psychotherapy protocol produced sizable clinical effects on depressive symptoms, anxiety, and multiple quality-of-life domains, but that these improvements were also observed following placebo and thus could not be attributed solely to the drug in this design. Effect sizes for depression reductions were larger after psilocybin than after placebo, and antidepressant effects following psilocybin persisted on average for about 8 weeks, with several participants achieving response or remission only after the active dose. The investigators emphasise several methodological factors that complicate causal attribution. A notable carryover effect from the placebo phase reduced symptomatic severity prior to the psilocybin session, potentially limiting the detectable additional benefit of psilocybin. The fixed-order design, while intended to limit carryover from an active first dose, therefore may have constrained ability to detect drug-placebo differences. Blinding and expectancy also influenced outcomes: despite enhanced blinding procedures and withholding the low dose, most participants correctly identified dosing condition, and the sample was largely self-referred with generally high expectations for psilocybin's benefit. The authors suggest that expectancy, engagement in psychotherapy during the placebo phase, and anticipation of receiving an active dose likely contributed substantially to the clinical gains observed after placebo. Sample size and generalisability are further limitations cited by the researchers. The trial was primarily powered for electrophysiological endpoints and was small for clinical efficacy comparisons, which increases the risk of type II error. In addition, the sample lacked diversity in race and education, limiting the ability to generalise findings. The investigators note that because the ACT-based psychotherapy was not compared with a minimal or different psychotherapeutic approach, it is not possible in this trial to disentangle the relative contributions of therapy versus dosing sessions to the observed improvements. Looking forward, the authors recommend future trials with parallel-group or factorial designs (for example, with and without psychedelic dosing crossed with with and without psychotherapy) to separate drug, therapy, and expectancy effects. They also highlight the need to test repeated or escalating dosing regimens, given variable durability of response and participants’ expressed desire for additional sessions, and to improve masking strategies—potentially by increasing the number of possible dosing conditions or using active controls—while acknowledging the feasibility challenges these approaches present. Finally, the researchers call for more diverse samples and further investigation into neurobiological and psychosocial mechanisms of action, noting that mystical-type experiences did not correlate with clinical improvement after psilocybin in this sample but were associated with placebo-phase improvement, underscoring the complex role of expectancy.

In this exploratory, placebo-controlled, within-subject fixed-order trial, administration of a single moderate dose of psilocybin within a manualised ACT-informed psychotherapy protocol was associated with clinically meaningful and sometimes durable improvements in depression, anxiety, and several quality-of-life domains. However, comparable improvements were observed after placebo, and differences in primary outcomes between psilocybin and placebo were not statistically significant. The authors conclude that the interplay of expectancy, psychotherapeutic support, and drug effects complicates interpretation, and they recommend that future studies prioritise improved masking, measurement and mitigation of expectancy, evaluation of repeated dosing strategies, and designs that can disentangle the contributions of psychotherapy and pharmacology.