Psilocybin for treatment resistant depression in patients taking a concomitant SSRI medication

Goodwin and colleagues situate this study in a growing literature showing antidepressant effects of psilocybin in preclinical models and small clinical trials of cancer-related depression, major depressive disorder (MDD), and treatment-resistant depression (TRD). Prior work has linked the intensity of the acute psychedelic or mystical-type experience to therapeutic outcomes, and mechanistic models emphasise 5-HT2A receptor activation. Because chronic treatment with selective serotonin reuptake inhibitors (SSRIs) can downregulate 5-HT2A receptors and surveys and animal models have suggested that antidepressants might attenuate subjective psychedelic effects, most clinical trials of psilocybin have required withdrawal from serotonergic antidepressants prior to dosing. However, a recent controlled study in healthy volunteers found that two weeks of SSRI treatment did not significantly reduce the acute subjective effects of a 25 mg dose of psilocybin, raising the possibility that requiring antidepressant withdrawal may be unnecessary and may create practical and safety barriers to clinical implementation and research participation. This exploratory Phase II, open-label, fixed-dose trial therefore set out to examine the safety, tolerability, and preliminary efficacy of a single 25 mg dose of COMP360, a synthetic pharmaceutical-grade psilocybin, when administered adjunctive to a stable SSRI in adults with TRD. The trial aimed to test whether ongoing SSRI therapy meaningfully compromises the acute subjective effects or the antidepressant response to psilocybin, and to collect safety data relevant to conducting psilocybin dosing without mandated antidepressant withdrawal.

The study was a Phase II, exploratory, fixed-dose, open-label clinical trial conducted at two sites in Ireland and the United States between August 2020 and September 2021. COMP360 (25 mg single oral dose) was administered once with psychological support to outpatients meeting DSM-5 criteria for a current episode of MDD that met the study definition of TRD (failure to respond to two to four adequate pharmacological treatments for the current episode). The protocol was approved by local ethics committees and conducted according to Good Clinical Practice. Eligible participants were adults (≥18 years) recruited from psychiatric services or primary care, required to have at least moderate depression severity (HAM-D-17 >18) at screening and baseline, and to be taking a single, stable therapeutic dose of a permitted SSRI (citalopram, escitalopram, fluoxetine, paroxetine, sertraline, vilazodone, or vortioxetine) with ≥75% adherence for at least six weeks prior to enrolment; participants on multiple antidepressants were excluded. High suicide risk and major comorbid psychiatric disorders were exclusion criteria as assessed by structured instruments (Mini International Neuropsychiatric Interview, McLean screening instrument, C-SSRS) and clinical review. A full eligibility list was reported in the Supplementary Appendix (not reproduced here). Procedures included a three-week screening/preparation period with weekly clinic visits and three preparatory therapy sessions delivered by trained mental health professionals. Baseline assessments were completed the day before dosing (Day -1). The administration session (Day 1) comprised a single 25 mg dose of COMP360 given in a controlled clinic setting; each participant had a lead therapist (and access to an assisting therapist and on-site psychiatrist), used eyeshades and a music playlist, and was continuously monitored (blood pressure via finger cuff). Integration sessions occurred on Day 2 and Week 1. Participants continued their SSRI throughout the study and were followed for three weeks post-dose. Safety endpoints were standard clinical and laboratory measures including ECGs, vital signs, clinical laboratory tests, adverse events coded by MedDRA v23.0, and suicidal ideation/behaviour (C-SSRS). The primary efficacy endpoint was change in Montgomery-Åsberg Depression Rating Scale (MADRS) total score from Baseline (Day -1) to Week 3; MADRS assessments were performed by a remote, blinded independent rater. Key secondary endpoints included proportion meeting response (≥50% MADRS improvement) and remission (MADRS ≤10) at Week 3, change in Clinical Global Impression-Severity (CGI-S) at Week 3, and CGI-S response. Exploratory measures included EQ-5D-3L and EQ-VAS (quality of life), GAD-7 (anxiety), QIDS-SR-16 (self-reported depression), CGI-Improvement, PANAS (positive and negative affect), and the 5D-ASC to capture acute subjective psychedelic phenomena. No formal sample size calculation was performed. Analyses were descriptive: safety analyses included all participants who received psilocybin (safety analysis set) and efficacy analyses used the full analysis set (participants who received psilocybin and had at least one post-Baseline efficacy assessment). No imputation was performed for missing efficacy data; the primary endpoint and secondary/exploratory endpoints were summarised with descriptive statistics.

Nineteen of 24 enrolled participants were dosed and completed the three-week follow-up; all 19 were included in the safety and full analysis sets. One participant was excluded from the per-protocol set for a prior electroconvulsive therapy/ketamine exposure. The sample was majority female (68.4%), predominantly White (78.9%), with a mean age of 42.2 years (SD = 10.8). Mean baseline MADRS total score was 31.7 (SD = 5.77), with 57.9% classified as severe (MADRS ≥31). Participants had been on SSRIs for a mean duration of 14.68 months (SD = 13.213); two participants increased escitalopram dose during the study. Safety: Twelve participants (63.2%) experienced a total of 17 treatment-emergent adverse events (TEAEs). Most TEAEs were mild and transient; 11 TEAEs in nine participants (47.4%) began on the day of dosing and eight resolved the same day. The most common TEAE was headache (six events in six participants; 31.6%), typically mild or moderate and resolving with or without standard analgesia. Three participants had a TEAE of increased blood pressure; two of these were classified as severe and treated with clonidine. The investigators note that blood pressure measurement relied on a finger cuff device that may not have been accurately calibrated, and no conventional cuff auscultation values are provided. No TEAEs were serious, none led to study withdrawal, and all resolved by Week 3 with durations not exceeding seven days. There were no reports of active suicidal ideation with plan or intent, attempts, or self-injurious behaviour during the study. No clinically meaningful laboratory or ECG abnormalities, including QTcF changes, were reported. Primary and key secondary efficacy: Mean MADRS at Week 3 was 16.8 (95% CI 11.2–22.4), corresponding to a mean change from Baseline of −14.9 (95% CI −20.7 to −9.2). Improvement in MADRS was apparent by Day 2 and maintained through Week 3. Response (≥50% reduction in MADRS) at Week 3 was observed in eight participants (42.1%), and the same eight participants met remission criteria (MADRS ≤10) at Week 3. Response and remission rates were higher at earlier time points for some participants (e.g. 63.2% response on CGI-S at Week 1). Mean CGI-S decreased from 4.3 (SD = 0.99) at Baseline to 2.9 (SD = 1.84) at Week 3, a mean change of −1.3 (SD = 1.29). Per-protocol analyses were consistent with the full analysis set. Exploratory outcomes: Quality of life and related measures improved from Baseline to Week 3. EQ-5D-3L index increased by a mean of 0.18 (SD = 0.310), and EQ-VAS increased by 13.7 points (SD = 18.09). Anxiety (GAD-7) decreased by a mean of 4.1 (SD = 4.90), which the authors consider clinically meaningful. Self-reported depression (QIDS-SR-16) fell by a mean of 6.9 (SD = 6.05), moving the group mean from the severe to the mild range. PANAS scores showed increases in positive affect (mean +8.1 by Day 2) and reductions in negative affect (mean −7.5 by Day 2). On the clinician-rated CGI-Improvement, 14 participants (73.7%) were at least minimally improved by Week 3 and seven (36.8%) met the CGI-defined responder threshold at Week 3. The 5D-ASC indicated acute alterations across dimensions, with greatest effects in visual restructuralization, oceanic boundlessness, and reduction of vigilance, and lesser effects in anxious ego dissolution and auditory alterations. The authors report that improvement was observed across all individual MADRS items in a post-hoc item-level analysis. No quantitative long-term follow-up beyond three weeks is provided in the extracted text.

Goodwin and colleagues interpret the findings as indicating that a single 25 mg dose of psilocybin administered with psychological support alongside a stable SSRI showed an acceptable safety profile and produced clinically meaningful improvements in depressive symptoms in this small, open-label sample of participants with TRD. The pattern of adverse events—mostly mild, transient, and resolving without intervention—along with no serious TEAEs or increase in suicidal ideation or behaviour, led the authors to conclude that safety was not compromised by continuing SSRI treatment in this context. They note the blood pressure elevations in three participants but caution that the finger cuff measurement system may not have been accurately calibrated and that conventional cuff measurements were not available for confirmation. On efficacy, the authors emphasise that ongoing SSRI therapy did not appear to attenuate the therapeutic effects of psilocybin: mean MADRS improvement (−14.9 at Week 3) and Week 3 response and remission rates (both 42.1%) are presented as broadly similar to, or in some respects higher than, those observed in a larger Phase IIb trial (COMP 001) that required antidepressant withdrawal. The 5D-ASC profile in this study resembled that reported in healthy volunteer studies and in trials where participants had withdrawn from antidepressants, leading the authors to argue that chronic SSRI exposure may not meaningfully downregulate the acute subjective psychedelic response. They outline possible explanations, including lower binding affinity of psilocybin at 5-HT2A compared with some other psychedelics and potential compensatory downstream mechanisms. The authors acknowledge important limitations: small sample size, open-label design, absence of a comparator or randomisation, limited demographic diversity, and exclusion of participants with high suicide risk or current hospitalisation. They further note that participant expectations could have influenced outcomes and that definitive comparisons with studies requiring antidepressant withdrawal are limited by design differences. Accordingly, the authors call for larger, double-blind, comparator-controlled trials with longer follow-up to confirm whether psilocybin can be used safely and effectively without antidepressant withdrawal. Finally, they highlight pragmatic implications: the feasibility of simultaneous administration sessions with appropriate staffing could reduce resource requirements and, if confirmed, allowing continued antidepressant use might broaden accessibility by avoiding the risks and burdens associated with medication tapering.