Pilot study of psilocybin treatment for anxiety in patients with advanced-stage cancer

Interest in addressing the psychological, spiritual and existential distress experienced by patients with advanced-stage cancer has re-emerged after a long hiatus in clinical research. Earlier trials from the 1950s to early 1970s reported that hallucinogens could produce psychospiritual experiences associated with sustained improvements in mood and reduced anxiety in terminally ill patients, but research largely ceased for political and cultural reasons. Psilocybin, a naturally occurring psychedelic metabolised to psilocin and acting as an agonist at serotonin 5-HT1A/2A/2C receptors (with 5-HT2A activation linked to hallucinogenic effects), has been re-examined because recent studies in healthy volunteers and in selected clinical populations suggest it can safely induce profound subjective experiences and, in some settings, durable improvements in well‑being. Grob and colleagues conducted the present investigation to explore the safety and preliminary efficacy of a moderate dose of psilocybin for anxiety reactive to advanced-stage cancer. The study aimed to assess physiological safety during sessions, acute subjective effects, and changes in mood and anxiety over a 6-month follow-up period using within-subject, double-blind, placebo-controlled methods in a small clinical sample. By reviving a controlled approach to this question, the study sought to establish feasibility and to inform the design of future trials.

This was a within-subject, double-blind, placebo-controlled pilot study in which each participant received two experimental sessions several weeks apart: one with psilocybin (0.2 mg/kg, oral) and one with active placebo (niacin 250 mg). Twelve adults with advanced-stage cancer and a DSM‑IV diagnosis of an anxiety-related disorder (acute stress disorder, generalized anxiety disorder, anxiety disorder due to cancer, or adjustment disorder with anxiety) were enrolled. Recruitment used internet postings, flyers, presentations, clinician referrals and patient-support contacts; medical and psychiatric screening included brain MRI and communication with treating oncologists. Eleven participants were women; ages ranged from 36 to 58. Primary cancers included breast, colon, ovarian, peritoneal, salivary gland and multiple myeloma; duration of illness ranged from 2 months to 18 years. Eight subjects completed the 6‑month follow-up, 11 completed at least 4 months, and all 12 completed at least 3 months. Two subjects died during follow-up and two became too ill to continue; by 2010 ten of the twelve had died. Key exclusion criteria were central nervous system involvement of cancer, severe cardiovascular disease, untreated hypertension, abnormal hepatic or renal function, diabetes, lifetime psychotic or bipolar disorders, recent (within 1 year before cancer onset) anxiety or affective disorders, and certain medication contraindications (active chemotherapy, antiseizure drugs, insulin/oral hypoglycaemics, and psychotropic drugs within the prior 2 weeks). Subjects were asked to avoid most medications on the day of and the day after sessions, with defined exceptions for pain medications. Four participants had no prior hallucinogen experience; the remainder had varying historic exposure. Preparatory meetings with study staff established rapport and clarified session structure and intentions. Sessions occurred in a hospital clinical research unit in a comfortable room; participants were admitted the day before, wore eye shades and listened to preselected music during the acute session, and were monitored for 6 hours with hourly check-ins including heart rate and blood pressure measurements. Holter cardiac monitoring covered 24 hours starting at admission. Self-report instruments were administered at prespecified times: Beck Depression Inventory (BDI), Profile of Mood States (POMS) and State-Trait Anxiety Inventory (STAI) before sessions and at repeated follow-ups; the 5-Dimension Altered States of Consciousness profile (5D-ASC) and Brief Psychiatric Rating Scale (BPRS) were given at session end. Monthly telephone contact and scheduled follow-up assessments continued for 6 months after the second session. Statistical analyses used two-way analysis of variance (ANOVA) with drug as a within-subject factor and day or time as a repeated measure for BDI, POMS and STAI; 1-way ANOVA was used for 5D-ASC and BPRS comparisons. When two-way ANOVA indicated significant main effects or interactions, post hoc one-way ANOVAs or paired tests were performed for individual time points. Paired t tests compared follow-up scores with baseline. The report defines significance thresholds for each set of comparisons (typically α=.05, with a stricter threshold where multiple comparisons of cardiovascular measures were made). ANOVA is a method to test for differences between group means while accounting for repeated measures and interactions.

Safety and cardiovascular monitoring indicated only modest physiological effects of psilocybin at the chosen dose. Heart rate and blood pressure rose modestly after psilocybin compared with niacin, with mean (SEM) peak systolic blood pressure during psilocybin sessions of 138.9 (6.4) mm Hg versus 117.0 (4.3) mm Hg during niacin sessions, and mean (SEM) peak diastolic blood pressure of 75.9 (3.4) mm Hg versus 69.6 (2.7) mm Hg. Holter monitoring showed no sustained tachyarrhythmias or heart block and was comparable between psilocybin and placebo sessions. There were no clinically significant adverse psychological events reported; participants tolerated sessions well and investigators observed no instances of severe acute anxiety or prolonged adverse reactions. Acute subjective effects measured by the 5D-ASC showed marked differences between psilocybin and placebo. Psilocybin produced large effects on the oceanic boundlessness dimension (F1,11=33.12, P<.001) and visionary restructuralization (F1,11=18.95, P=.001), with smaller but statistically significant effects on anxious ego dissolution (F1,11=4.91, P=.049) and auditory alterations (F1,11=5.93, P=.03). Item clusters that increased significantly with psilocybin included positive derealization, positive depersonalization, altered sense of time, positive mood, manialike experiences, elementary hallucinations, visual pseudohallucinations, synesthesia, changed meaning of percepts, facilitated recollection and imagination. Subscales showing no appreciable differences included anxious derealization, thought disorder, delusion and fears of loss of control. Mood and anxiety measures yielded mixed but some promising signals. For the BDI there was an interaction of drug and day that approached but did not reach conventional significance (F1,11=3.75, P=.08). Mean BDI scores trended downward after psilocybin from 16.1 (3.6) one day before treatment to 10.0 (2.7) two weeks after; BDI scores fell by almost 30% from the first session to 1 month after the second session (t11=-2.17, P=.05) and this improvement was sustained and statistically significant at the 6-month follow-up (t7=2.71, P=.03). The POMS showed a trend toward reduced adverse mood following psilocybin (drug×time interaction F3,33=2.71, P=.06). Notably, mean POMS scores were elevated one day before psilocybin treatment (F1,11=7.48, P=.02) relative to placebo sessions, a difference that disappeared by 6 hours post‑administration; POMS scores were reduced after psilocybin in 11 of 12 subjects, but overall POMS values were not changed across the full 6‑month follow-up compared with baseline. Anxiety outcomes differed by STAI subscale. State anxiety showed no significant changes from baseline to 2 weeks, though a non-significant reduction at 6 hours after psilocybin was observed. In contrast, STAI trait anxiety demonstrated a sustained decrease across the 6‑month follow-up, reaching significance at 1 month (t11=4.36, P=.001) and at 3 months (t10=2.55, P=.03) after the second treatment session. The BPRS recorded at session end revealed no appreciable differences between psilocybin and placebo. Pain and somatic symptom effects were not robustly altered by the moderate psilocybin dose. Some participants reported lessened pain in the two weeks after sessions while others did not, and no consistent difference between psilocybin and placebo was evident in the data presented.

Grob and colleagues framed the study primarily as establishing feasibility and safety for administering a moderate psilocybin dose to patients with advanced-stage cancer and reactive anxiety. The selected 0.2 mg/kg dose reflected regulatory and institutional discussions; it was intended to balance the potential for an altered state of consciousness with minimised risk. Physiological data supported safety at this dose: only mild sympathomimetic effects were observed and continuous cardiac monitoring did not reveal increased arrhythmias relative to niacin placebo. Psychological safety was also supported, with no reports of severe anxiety or so‑called "bad trips" and only modest effects on measures of anxious ego dissolution. The investigators interpreted the subjective and longer‑term psychological data as suggestive but preliminary. Acute psilocybin sessions produced clear phenomenological effects (oceanic boundlessness, visionary phenomena) and many participants reported meaningful insights about life, relationships and mortality during sessions and in monthly follow-up discussions; however, the frequency of these qualitative reports was not quantified. Quantitative measures showed a sustained reduction in STAI trait anxiety with significance at 1 and 3 months, and BDI scores demonstrated a significant improvement by 6 months. POMS results suggested mood improvement in most participants but did not reach robust statistical significance over the follow-up period. The absence of consistent pain reduction contrasted with some earlier reports and may reflect the lower dose used here. Several limitations were acknowledged. The small sample size and heterogeneity of the cohort constrain generalisability and statistical power. Although the within-subject, double-blind design was intended to be rigorous, unblinding was common because the active drug's effects were generally apparent to participants and staff; many participants perceived placebo sessions as less valuable. Variability in post‑session contact and support (beyond a minimum monthly hour) may have influenced outcomes, and the ethical choice to allow all participants to receive the experimental medicine removed the option of an independent control group receiving only placebo or standard pharmacotherapy. The moderate dose may have limited the intensity of psychospiritual experiences and thus therapeutic effects compared with historical higher-dose protocols. Despite these constraints, the investigators concluded that carefully controlled administration of psilocybin in this population appears feasible and reasonably safe, and that the observed reductions in trait anxiety and trends toward improved mood warrant further research. They recommended future studies with larger samples, consideration of higher doses or repeated dosing, improved blinding strategies or independent control groups, and standardised post-session support to better assess efficacy and mechanisms while maintaining stringent safety protocols.