An exploratory randomised, double‑blind, placebo‑controlled trial of a patient‑informed low‑dose psilocybin pulse regimen for cluster headache found no statistically significant reduction in attack frequency overall, although a moderate overall effect size (d = 0.69) and a large, sustained effect in chronic patients (d = 1.25) were observed. Psilocybin was well tolerated, therapeutic effects did not correlate with acute psychotropic intensity, and the authors conclude larger definitive trials are warranted.
Papers cited by this study that are also in Blossom
Andersson, M., Persson, M., Kjellgren, A. · Harm Reduction Journal (2017)
Objective
Using a patient‐informed regimen, we conducted an exploratory randomized, double‐blind, placebo‐controlled study to systematically investigate the effects of psilocybin in cluster headache.
Background
Sustained reductions in cluster headache burden after limited quantities of psilocybin‐containing mushrooms are anecdotally reported, although to date there are no controlled studies investigating these effects.
Methods
Participants were randomized to receive psilocybin (0.143 mg/kg) or placebo (microcrystalline cellulose) in a pulse of three doses, each ~5 days apart. Participants maintained headache diaries starting 2 weeks before and continuing through 8 weeks after the first drug session. A total of 16 participants were randomized to receive experimental drug and 14 were included in the final analysis.
Results
In the 3 weeks after the start of the pulse regimen, the change in cluster attack frequency was 0.03 (95% confidence interval [CI] −2.6 to 2.6) attacks/week with placebo (baseline 8.9 [95% CI 3.8 to 14.0]) and −3.2 (95% CI −8.3 to 1.9) attacks/week with psilocybin (baseline 9.6 [95% CI 5.6 to 13.6]; p = 0.251). Group difference in change from baseline had a moderate effect size ( d = 0.69). The effect size was small in episodic participants ( d = 0.35) but large in chronic participants ( d = 1.25), which remained over the entire 8‐week period measured ( d = 0.81). Changes in cluster attack frequency were not correlated with the intensity of acute psychotropic effects during psilocybin administration. Psilocybin was well‐tolerated without any unexpected or serious adverse events.
Conclusions
Findings from this initial, exploratory study provide valuable information for the development of larger, more definitive studies. Efficacy outcomes were negative, owing in part to the small number of participants. The separation of acute psychotropic effects and lasting therapeutic effects underscores the need for further investigation into the mechanism(s) of action of psilocybin in headache disorders.
Cluster headache is an uncommon but disabling primary headache disorder notable for severe unilateral pain and autonomic features, and it carries a high burden for affected patients. Previous case reports, survey data and some peer-reviewed work have suggested that indoleamine 5-HT2A receptor agonists (for example psilocybin and LSD) may produce sustained reductions in cluster headache frequency or even brief remissions after only one or a few doses. Many patients self-administer low or subpsychedelic doses and report benefit, raising the possibility of a mechanism that is at least partially independent of acute psychedelic experiences. Schindler and colleagues designed an exploratory, controlled study to test a patient-informed low-dose psilocybin ‘‘pulse’’ regimen—three doses given about 5 days apart—modelled on practices used by people with cluster headache. The investigators hypothesised that, compared with placebo, this regimen would reduce several measures of cluster headache burden and be safe when administered under laboratory conditions; as the first randomised, double-blind, placebo-controlled investigation in this indication, the study also aimed to generate effect-size estimates to guide future trials.
This was an exploratory randomised, double-blind, placebo-controlled trial. Adults aged 21–65 meeting International Classification of Headache Disorders criteria for cluster headache and with a minimum attack frequency of approximately one attack per day were recruited from clinical and community sources. Both episodic participants (typical cluster periods of about 2 months or more) and chronic participants were eligible. Past psychedelic exposure was permitted provided use had not occurred within the prior 3 months. Participants had to be off serotonergic antidepressants for at least 6 weeks; triptans were allowed with restrictions. Full procedural details were stated to be available in the Supporting Information, and the extracted text does not clearly report every inclusion/exclusion detail. Participants were randomised 1:1, stratified by headache subtype, to receive either oral psilocybin 0.143 mg/kg (approximately 10 mg for a 70 kg person) or placebo (microcrystalline cellulose). Randomisation used computer-generated block lists and a Research Pharmacy maintained blinding under an enhanced blinding procedure. Each participant completed three experimental sessions separated by 5 ± 2 days and received the same study capsule at each session. Sessions occurred in a clinical research unit with vital-sign monitoring; participants remained for at least 6 hours post-dose and were followed by telephone out to 6 months. Participants kept a standardised headache diary for 14 days before the first session (baseline) and for 56 days after the first session. Because the trial was exploratory, no single primary outcome was pre-specified; emphasis was placed on change in weekly cluster-attack frequency in the 3-week period after starting the pulse regimen. Other outcomes included attack duration and pain intensity over the same 3-week window, subgroup analyses by episodic versus chronic status (extending to 8 weeks for chronic participants), acute physiological effects, general subjective drug effects, psychedelic experiences measured by the 5-Dimensional Altered States of Consciousness (5D-ASC) scale, and adverse events (AEs). Statistical analyses used SAS and GraphPad Prism. Tests were two-sided with p < 0.05, and no covariate adjustment was performed in primary analyses. Diary-derived weekly attack rates were calculated per participant and group means with 95% confidence intervals were reported; between-group comparisons of change-from-baseline used independent t-tests. Acute session measures (vital signs, general drug effects) and 5D-ASC scores were analysed with linear mixed models including group, session day and time factors and random participant effects. Correlations between acute psychotropic effects and change in attacks used Spearman analysis. Fisher’s exact test compared AE counts. The authors note some missing session data for a small number of measures and indicate that mixed models used all available data. No a priori power calculation was conducted due to lack of prior controlled data.
Schindler, E. A. D., Gottschalk, C. H., Weil, M. J. et al. · Journal of Psychoactive Drugs (2015)
Sewell, R. A. · Neurology (2006)
Schindler, E. A. D., Sewell, R. A., Gottschalk, C. H. et al. · Neurotherapeutics (2021)
Studerus, E., Gamma, A., Vollenweider, F. X. · PLOS ONE (2010)
Grob, C. S., Danforth, A. L., Chopra, G. S. et al. · JAMA Psychiatry (2011)
Hasler, F., Grimberg, U., Benz, M. A. et al. · Psychopharmacology (2004)
Aday, J. S., Heifets, B. D., Pratscher, S. D. et al. · Psychopharmacology (2021)
Karst, M., Halpern, J. H., Bernateck, M. et al. · Cephalalgia (2010)
Garcia-Romeu, A., Griffiths, R. R., Johnson, M. W. · Current Drug Abuse Reviews (2015)
Carhart-Harris, R. L., Erritzoe, D., Williams, T. et al. · PNAS (2012)
Papers in Blossom that reference this study
Schindler, E. A. D., Gottschalk, C. H., Pittman, B. P. et al. · Headache (2025)
Aday, J. S., Mcafee, J., Conroy, D. A. et al. · Frontiers in Pain Research (2025)
Rucker, J., Butler, M., Hambleton, S. et al. · Headache (2024)
Schindler, E. A. D. · Current Pain and Headache Reports (2023)
Recruitment between November 2016 and August 2021 yielded 238 pre-screened candidates; 20 entered secondary screening, 16 underwent study procedures, and 14 participants were included in the final analysis (six placebo, eight psilocybin). Two subjects were excluded post-procedure for protocol violations (use of psilocybin-containing mushrooms during follow-up and failure to provide a diary after baseline). The analysed sample comprised nine males and five females with mean age 49.1 years (SD 10.7, range 27–61); six participants had episodic cluster headache and eight had chronic disease. None reported satisfaction with their current treatment regimen. The prespecified primary window was the 3-week period after starting the pulse. In that interval the change in weekly attack frequency was 0.03 attacks/week (95% CI −2.6 to 2.6) for placebo (baseline 8.9 attacks/week, 95% CI 3.8 to 14.0) and −3.2 attacks/week (95% CI −8.3 to 1.9) for psilocybin (baseline 9.6 attacks/week, 95% CI 5.6 to 13.6). The between-group difference did not reach statistical significance (p = 0.251). The authors report a moderate overall effect size (Cohen’s d = 0.69). By subgroup, effect sizes were smaller in episodic participants (d = 0.35) and larger in chronic participants (d = 1.25); the larger chronic-participant effect was reported to persist across the full 8-week observation period (d = 0.81). Responder analyses showed heterogeneous individual responses. Across all participants over 3 weeks, placebo yielded ≥25% and ≥50% reductions in weekly attacks in 33.3% (2/6) of participants, whereas psilocybin produced ≥25% reductions in 50% (4/8) and ≥50% reductions in 37.5% (3/8); 37.5% (3/8) of psilocybin recipients achieved ≥75% reduction and none achieved 100% remission. In the small episodic subgroup both arms had similar percentages for 25% and 50% responders (66.7% each), and psilocybin had one 75% responder (1/3); reported differences were not statistically significant. The authors emphasise notable heterogeneity across individual trajectories and that small subgroup sizes limit inferences. Acute physiological and subjective effects were consistent with pharmacologic activity. Psilocybin produced a mean maximal increase in mean arterial pressure of 8.5 mmHg (95% CI 2.9–14.1) between 60 and 90 minutes post-dose; heart rate and oxygen saturation did not show significant group×time interactions. On a 0–3 scale, psilocybin induced peak overall drug effects between 90 and 120 minutes with a mean increase of 2.6 (95% CI 2.2–3.0). Significant drug×time interactions were observed for sleepiness, anxiety and joy/intense happiness, but maximal overall drug-effect ratings (averaged across the three sessions) were not correlated with percent change in weekly attacks (Spearman r = 0.063, p = 0.883). Psychedelic effects measured by the 5D-ASC were markedly higher after psilocybin: average percent total score across sessions was 20.6% (95% CI 14.8%–26.5%) for psilocybin versus 0.9% (95% CI 0.1%–1.7%) for placebo (p < 0.0001). No association between 5D-ASC total score and change in weekly attacks was found (r = 0.527, p = 0.224). Test-day 1 produced higher 5D-ASC scores than days 2 and 3. Safety data showed no unexpected or serious adverse events. The most common acute AEs with psilocybin were nausea, anxiety and fatigue; only nausea incidence during sessions was significantly higher with psilocybin than placebo (p = 0.031). One participant experienced transient paranoia during a session that resolved with staff support and subsequently declined the third session; this individual nevertheless had rapid termination of their cluster period. At six-month follow-up no lasting physical or neuropsychological changes were reported by participants.
Schindler and colleagues report that their patient-informed low-dose psilocybin pulse regimen did not produce statistically significant reductions in cluster-attack frequency compared with placebo in this small exploratory trial, although the observed overall effect size was moderate. The investigators highlight a larger effect size among chronic participants compared with episodic participants, but note that small subgroup sizes and natural variability in episodic cluster periods (including spontaneous tapering) complicate interpretation. Safety and tolerability were acceptable in the controlled setting. The authors discuss important methodological challenges for psychedelic trials in cluster headache. Expectation and blinding are emphasised: heightened community interest in psychedelics and dissatisfaction with existing treatments might influence expectations, and the observable acute effects of psilocybin may unblind participants. The placebo agent used (microcrystalline cellulose) produced minimal acute effects and may not have been an adequate active comparator; the authors suggest using an active control with similar acute subjective effects in future work. They also note that acute psychotropic intensity did not correlate with therapeutic outcome in this and related controlled studies, supporting the possibility that therapeutic mechanisms can be dissociated from the namesake psychedelic experience. Mechanistic considerations raised by the authors include circadian and hypothalamic systems relevant to cluster headache, potential effects on pineal/melatonin signalling and REM sleep, and acute reductions in hypothalamic blood flow reported in other psilocybin studies; the trial was not designed to probe mechanisms, but ongoing actigraphy and imaging studies are mentioned. Strengths of the work include the randomised, double-blind, placebo-controlled design and use of a regimen informed by patient practice. Limitations cited by the authors are the small and non-representative sample (all Caucasian, more chronic than episodic cases), allowance of past psychedelic exposure, a single fixed low dose (~0.143 mg/kg) without capacity to titrate, imperfect blinding, and lack of power to detect modest effects. The investigators conclude that larger, better-blinded and more representative trials are needed to define efficacy, optimal dosing/regimen and moderators of response.
In this first controlled investigation of a patient-informed three-dose low-dose psilocybin pulse regimen for cluster headache, the authors observed a moderate effect size but no statistically significant efficacy compared with placebo. The regimen was well tolerated and no serious adverse events were reported. The authors recommend larger, more representative trials with improved blinding and alternative comparator strategies to clarify clinical efficacy and mechanism.