A review of psychedelics trials completed in depression, informed by European regulatory perspectives

Interest in the therapeutic use of psychedelics for mental health disorders, and depression in particular, has expanded rapidly, prompting regulatory bodies to respond. Classic psychedelics are typically defined by agonism at serotonin 2A receptors, and although no classic psychedelic currently has marketing authorisation, regulators in multiple jurisdictions have issued draft guidance or created special access pathways. Against a backdrop of substantial unmet need in depression, including a high prevalence and a large proportion of patients showing inadequate response to first-line antidepressants, the European Medicines Agency (EMA) has proposed a draft revision of its guideline on clinical investigation of products for depression that includes preliminary considerations specific to psychedelic studies. Silva and colleagues set out to systematically evaluate the methodological approaches used in completed randomised controlled trials of psychedelics for major depressive disorder (MDD) or treatment-resistant depression (TRD), and to discuss how those approaches align with initial regulatory considerations in the EMA draft guideline. The review aimed to identify patterns and gaps across trial design, populations, interventions, outcome measurement and safety assessment that are relevant for future clinical development and regulatory submission of psychedelic medicines in the EU.

The investigators conducted a systematic search of scientific databases (Embase and MEDLINE) and clinical trial registries (ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform). The search combined general and substance-specific terms for classic psychedelics and MDMA with a controlled-trials filter adapted from an existing Cochrane filter; the exact search string is available from the authors on request. The initial search identified both publication records and registry entries and was completed on 19 February 2024. Records were screened at title and abstract level against predefined criteria. Included records had to report a human clinical trial that systemically administered a classic psychedelic or MDMA, be conducted after 1995, and test the psychedelic as part of a therapeutic intervention with at least one efficacy endpoint in a patient population. The reviewers grouped records by trial and retained those that targeted a depression indication and were reported as completed at the time of the search. Exclusions applied to non-trial reports, exclusively open-label or non-randomised trials, studies in healthy volunteers only, and trials reporting only safety or mechanistic endpoints. Ongoing or not-yet-started trials were excluded from most analyses because register entries often lacked detailed methodological information. Data extraction focused on a comprehensive set of pre-defined variables informed by the draft EMA guideline and draft FDA guidance. Extracted items covered trial characteristics and design (phase, allocation, blinding, comparators), population selection (indication, psychiatric comorbidity exclusions, concomitant treatments, prior psychedelic experience), intervention details (substance, dose regimen, number of administrations, and the broader treatment framework including practitioner qualifications and presence of preparatory/integration sessions), and outcome assessments (primary efficacy measures and timing, instruments used to characterise subjective psychedelic effects, and safety monitoring). When available, information from trial protocols and registry attachments was included in the extraction.

After de-duplication, 7,772 records were screened and 421 records relating to 101 controlled psychedelic trials remained; from these, eight completed trials testing psychedelic interventions for depression met the inclusion criteria. The eight trials were conducted between February 2014 and December 2022, comprised a total of 595 recruited patients, and were all early-phase studies: two Phase I/II and six Phase II trials. Five trials evaluated psilocybin, one evaluated intravenous DMT, one evaluated ayahuasca, and one evaluated LSD. Six trials focused on MDD and two on TRD. Design features varied. Sample sizes per trial ranged from 27 to 233 participants. Control conditions included inactive placebo, active placebo, very low (sub-psychoactive) doses of the active substance, and a wait-list control; half of the trials used a placebo control, three used very low dose comparators, and one used a wait-list design. Of the 595 participants, 525 received a psychedelic in a blinded, randomised context (excluding those assigned to inactive placebo or open-label DMT stages). Randomisation was equal allocation in all trials. Reported blinding levels were: four trials described quadruple blinding (participants, care providers, investigators and outcome assessors), two reported triple-blinding, one double-blinding and the waitlist trial reported single-blinding. None of the trials explicitly reported blinding of data analysts, and no trial formally reported systematic assessment of blinding or expectancy among participants. Population eligibility commonly excluded individuals with current or past psychotic or bipolar disorders (seven trials), and several trials also excluded participants with a family history of those disorders. High suicide risk was an exclusion criterion in seven trials. Substance use disorders were excluded in three trials, and most studies prohibited concurrent antidepressant medication; washout periods typically required at least two weeks off antidepressants. Some trials disallowed antipsychotics, psychostimulants, mood stabilisers or anxiolytics, although use of rescue benzodiazepines or certain antipsychotics during dosing was permitted in some protocols. Three trials limited concomitant psychotherapeutic input and three imposed restrictions based on prior psychedelic experience. Interventions were delivered within broader treatment frameworks rather than as isolated drug administrations. Six trials reported a three-stage model of pre-dosing preparation, dosing sessions and post-dosing integration; five trials applied the same framework to both active and comparator arms when reported. Preparatory phases comprised between two and five visits, dosing sessions were typically one or two administrations lasting about 6–8 hours (except intravenous DMT where acute effects lasted roughly 20 minutes), and integration included at least two post-dosing sessions with some protocols offering up to six. Teams of two practitioners were described in the psilocybin trials for which information was available, with minimum qualifications typically at bachelor level in a mental-health-related field and higher qualifications required in some trials; several trials described bespoke practitioner training. Physical settings for dosing were usually homelike, living-room style environments and six trials used standardised music playlists. Three trials planned audio or video recording of dosing sessions for review or qualitative analysis. Primary efficacy outcomes used validated clinician- or self-rated depression scales: MADRS (n = 4), HAM-D or GRID-HAM-D (n = 2), IDS or Quick IDS (n = 2), and BDI (n = 1). The average timing of the last efficacy assessment was 6.25 weeks after the final psychedelic administration; the shortest primary assessment windows were 1–2 weeks in Phase I/II trials, and two trials had blinded efficacy assessments out to 12 weeks. Two trials included open-label extensions with assessments at 13 and 29 weeks, and two long-term follow-ups assessed participants up to 52 weeks post-dosing. All trials incorporated at least one instrument to characterise the acute subjective psychedelic experience: the Mystical Experience Questionnaire (MEQ) was used in six trials, the Altered States of Consciousness scale (ASC) in four, the Emotional Breakthrough Inventory (EBI) in three, the Ego Dissolution Inventory (EDI) in one, and the Hallucinogen Rating Scale (HRS) in one. Additional post-acute measures included the Persisting Effects Questionnaire (PEQ), the Challenging Experiences Questionnaire (CEQ) and measures of mindfulness, humility and personality (e.g. the Ten Item Personality Inventory). Safety approaches commonly included the presence of monitors during dosing (reported in seven trials), availability of a physician nearby (five trials) and periodic vital-sign monitoring (three trials). Standardised suicidality assessments were included in five protocols, chiefly the Columbia Suicide Severity Rating Scale (C-SSRS; n = 4) and the Suicidal Ideation Attributes Scale (SIDAS; n = 1). Two trials used the CEQ to capture challenging acute experiences. The reviewed trials largely prohibited concurrent antidepressant treatment during the study period, and most excluded participants at high risk of suicidal behaviour.

Silva and colleagues interpret their findings as indicating that clinical research on classic psychedelics for depression is progressing but remains at an early stage, with only Phase I/II and II trials completed as of the search cut-off and no Phase III studies finalised. The authors highlight recurring methodological challenges that map onto concerns raised in draft regulatory guidance from the EMA and FDA: functional unblinding and expectancy, the definition and role of the treatment framework or "set and setting", how to characterise and quantify the acute subjective experience and its relationship to outcomes, and systematic collection and interpretation of safety data including psychologically challenging experiences and suicidality. With respect to unblinding and expectancy, the review notes that the pronounced subjective effects of psychedelics make maintaining conventional blinding difficult, particularly when inactive placebos are used or when participants have prior psychedelic experience. The authors describe several mitigation strategies observed or recommended in guidance: use of active placebos or very low-dose comparators, employing blinded independent raters (including remote raters), systematically assessing blinding and expectancy, and trial designs that offer active treatment to all participants at some point (for example dose-ranging designs, wait-list control or open-label extensions). They caution that registry and publication entries sometimes lack clarity about who was blinded, and that none of the reviewed trials reported formal blinding assessments. The review emphasises the centrality of the treatment framework — preparatory sessions, supervised dosing in a controlled environment, integration work, practitioner qualifications and training — and discusses the regulatory implication that developers must disentangle the specific contribution of the pharmacological intervention from that of adjunct psychological support. The question whether the accompanying psychological intervention should be considered a safety measure or an adjunctive psychotherapy with intrinsic efficacy implications is identified as consequential for trial design and for later clinical translation; to address this, trials should standardise and apply the same treatment framework across active and comparator arms and document adherence, for instance via session recordings where ethically acceptable. On outcomes and timing, the authors note that blinded efficacy assessment windows in completed trials were relatively short — commonly up to 12 weeks — while only a subset of studies reported longer-term follow-up. They recommend that later-phase studies address maintenance of effect and dose-response relationships, and balance the ethical considerations of withholding subsequent treatments against the need to characterise durability of effect. Safety considerations discussed include dose-dependent physiological responses (heart rate, blood pressure, anxiety), potential for acute challenging experiences that may have longer-term psychological consequences, and the need for standardised collection of adverse events tied to subjective effects; the CEQ and other instruments were noted as useful tools already employed in some trials. Finally, the authors acknowledge the strengths of their review — a systematic search and regulatory framing — and its limitations, chiefly the cut-off date of 19 February 2024 and the focus on depression which excluded broader psychedelic research. They recommend early and sustained dialogue between developers and regulators through EMA mechanisms such as scientific advice and qualification procedures to ensure that development programmes meet regulatory expectations and to support the medicalisation of psychedelics for unmet patient needs.

The authors conclude that methodological issues are a prominent theme in the early-phase clinical development of psychedelics for depression, but that these challenges are recognised and can be addressed through carefully designed development programmes aligned with regulatory frameworks. They note that EMA platforms exist to support developers, and that multi-stakeholder engagement will be important to advance scientifically robust and regulatorily acceptable trials with the aim of delivering innovative treatments to patients.