Psilocybin for End-of-Life Anxiety Symptoms: A Systematic Review and Meta-Analysis

Yu and colleagues frame the problem by noting that patients with life‑threatening illnesses such as cancer and HIV commonly experience substantial psychological distress, including high rates of anxiety and depression. The introduction summarises biological and psychosocial contributors to this distress — for example, the threat of disease progression, stigma, and neuroinflammatory processes — and observes that conventional psychopharmacological and psychosocial treatments show mixed and limited efficacy in these populations. Against this background, the authors introduce psilocybin, a serotonergic (5‑HT2A agonist) psychedelic derived from psilocybin mushrooms, which has been reported to induce profound subjective experiences and long‑lasting positive changes in cognition, affect, and behaviour. Early clinical trials have suggested benefits for cancer‑ and HIV‑related anxiety; the present study therefore set out to synthesise evidence from clinical trials to evaluate the effectiveness and tolerability of psilocybin for end‑of‑life anxiety symptoms, and to examine the persistence of effects at multiple follow‑up points after a single treatment session.

The investigators conducted a systematic review and meta‑analysis following PRISMA guidelines. Two researchers independently searched Medline, Embase, the Cochrane Central Register of Controlled Trials (CENTRAL), and PsycINFO up to 25 November 2020; a third reviewer resolved disagreements. The review included peer‑reviewed clinical trials of psilocybin administered to patients with life‑threatening conditions (for example cancer or HIV), accepting both controlled (placebo‑comparative) and uncontrolled (open‑label single‑arm) trial designs. Non‑clinical reports, case series, conference abstracts and non‑English publications were excluded. Primary outcomes were changes in state anxiety (a transient emotional condition) and trait anxiety (a relatively stable predisposition to anxiety). Secondary outcomes comprised all‑cause discontinuation, peak changes in heart rate (HR), systolic blood pressure (SBP) and diastolic blood pressure (DBP), and serious adverse events. Two authors independently assessed study quality using the Cochrane risk‑of‑bias tool across the standard seven domains. For quantitative synthesis the authors used random‑effects meta‑analysis. Effect sizes for primary outcomes were expressed as Hedges' g (a standardised mean difference; negative values indicate anxiety improvement). Weighted mean differences (WMDs) were used for physiological outcomes (HR, SBP, DBP) and odds ratios (ORs) for discontinuation. Anxiety was assessed at day 1 and at 1, 3 and 6 months after a single psilocybin session; pre–post change at the last available follow‑up was also examined. Heterogeneity was evaluated with Cochran's Q and the I2 statistic (I2 >75% considered high). Meta‑regression was not performed because fewer than 10 studies were available. Publication bias was inspected with funnel plots and assessed using trim‑and‑fill; statistical significance was set at two‑tailed p<0.05. Selection results are reported: 45 full texts were reviewed and five studies met inclusion criteria. The extracted text includes study‑level characteristics and methodological quality summaries that were incorporated into the analyses.

Five trials were included in the review. Across these studies the mean sample size was 26.4 (SD 16.2; range 12–56), mean participant age was 54.4 years (SD 4.2), and mean female proportion was 52.5% (SD 29.8). All five studies reported state and trait anxiety outcomes at day 1 and at 1, 3 and 6 months after a single psilocybin session. Risk‑of‑bias assessments indicated that three randomised controlled trials had low or unclear risk in most domains but generally lacked clarity on blinding of outcome assessment. The single‑arm study by Anderson et al. had high risk across several domains (randomisation and blinding), and a post‑RCT follow‑up study by Agin‑Liebes et al. showed high risk for blinding of outcome assessment. For state anxiety, psilocybin was superior to placebo at day 1 (Hedges' g = -0.70; 95% CI -1.01 to -0.39; I2 = 0.0%) and at 1 month (Hedges' g = -0.73; 95% CI -1.25 to -0.20; I2 = 62.8%). Effects were not statistically significant at 3 months (Hedges' g = -0.47; 95% CI -1.31 to 0.37; I2 = 93.3%) or at 6 months (Hedges' g = -0.88; 95% CI -1.78 to 0.03; I2 = 86.5%). Two longer‑term assessments reported by Agin‑Liebes et al. at ~3.2 and 4.5 years indicated large reductions in state anxiety (Hedges' g ≈ -1.05 and -0.90), but these were from a single study and are presented separately. For trait anxiety, psilocybin outperformed placebo at day 1 (Hedges' g = -0.71; 95% CI -1.15 to -0.25; I2 = 39.2%) and at 1 month (Hedges' g = -1.04; 95% CI -1.66 to -0.42; I2 = 65.7%). The advantage persisted at 3 months (Hedges' g = -0.60; 95% CI -1.16 to -0.05; I2 = 84.9%) and at 6 months (Hedges' g = -1.03; 95% CI -1.54 to -0.52; I2 = 68.8%), though heterogeneity at the longer time points was substantial. Pre–post analyses at the last follow‑up across studies showed significant reductions for state anxiety (Hedges' g = -0.63; 95% CI -1.24 to -0.03) and trait anxiety (Hedges' g = -0.83; 95% CI -1.25 to -0.42), with significant heterogeneity for both. Regarding physiological and tolerability outcomes, psilocybin did not differ significantly from placebo in peak heart rate (WMD = 2.85 bpm; 95% CI -4.51 to 10.20; I2 = 51.9%). Systolic blood pressure increased during sessions (WMD = 19.00 mm Hg; 95% CI 13.58 to 24.41; I2 = 8.9%) as did diastolic blood pressure (WMD = 8.66 mm Hg; 95% CI 5.18 to 12.15; I2 = 19.1%); the trials reported these elevations as transient and self‑limited, requiring no medical intervention. All‑cause discontinuation did not differ between groups (OR = 1.53; 95% CI 0.37 to 6.22; I2 = 0.0%). No significant differences were reported for serious adverse events in the extracted text. Publication bias analyses (funnel plots and trim‑and‑fill) did not indicate small‑study effects and suggested that imputing potential missing studies would not change the direction of pooled effects.

Yu and colleagues interpret their findings as evidence that a single session of psilocybin‑assisted therapy can reduce end‑of‑life anxiety symptoms. They highlight a clear short‑term benefit on state anxiety (day 1 and 1 month) and more persistent improvements in trait anxiety that were observed up to 6 months post‑treatment in pooled analyses. The authors suggest that sustained reductions in trait anxiety may reflect improvements in stress‑coping strategies and personality dimensions such as neuroticism. In contextualising these results, the paper notes consistency with earlier clinical findings that psychedelics can produce long‑lasting changes in affect and wellbeing. Proposed biological mechanisms discussed include 5‑HT2A receptor agonism and altered activity or blood flow in medial prefrontal and anterior cingulate cortices, regions implicated in emotion regulation and anxiety. The authors acknowledge that the precise neural mechanisms in humans remain incompletely understood. Safety and tolerability observations are tempered: although no differences were seen in discontinuation or heart rate and no serious adverse events were reported, psilocybin sessions produced transient elevations in systolic and diastolic blood pressure that may pose risks in patients with cardiovascular comorbidity. The possibility of abuse is raised; while psilocybin has not been shown to cause dependence or withdrawal, prior work has flagged potential for misuse on some subjective scales, indicating a need for controlled scheduling and monitoring if used clinically. The authors list several limitations: only five eligible trials were available, reducing power for publication‑bias analyses; heterogeneity was substantial for several long‑term outcomes; generalisability is limited because the included populations all had life‑threatening illnesses; and some trials combined psychotherapeutic support with psilocybin, making it difficult to disentangle the contribution of psychotherapy to sustained effects. On this basis the authors call for larger randomised controlled trials with longer follow‑up, and recommend cautious, closely monitored clinical use tailored to individual patient status.