Psilocybin therapy for adult females with anorexia nervosa: pilot study

Anorexia nervosa is described as a severe eating disorder with high mortality, chronicity and limited effective treatment options. The authors note that talking therapies are first-line, but there are no approved pharmacological treatments, and many patients remain unwell or become chronically ill. They also emphasise that anorexia nervosa can be ego-syntonic, meaning it may be experienced as part of the person’s identity, which contributes to resistance to treatment and lower readiness to change. Previous psychedelic research, including psilocybin therapy supported by psychological care, has suggested possible benefits for psychiatric conditions and early signals in eating disorders, but evidence in anorexia nervosa remains limited. Against this background, the study aimed to further assess the safety, feasibility and potential efficacy of psilocybin therapy in adult females with anorexia nervosa. Douglass and colleagues state that they wanted to see whether this protocol could be recruited to, tolerated and associated with improvement in eating disorder symptoms and motivation to change. The paper presents this as a pilot study intended to generate preliminary data for larger, more rigorous trials.

Douglass and colleagues conducted a single-arm, single-blind, within-individual pilot study in adult females with anorexia nervosa. The intervention was adjunctive to treatment as usual, meaning participants continued receiving their existing NHS eating disorder care and/or private therapy during the study. The trial used COMP360, a pharmaceutical-grade synthetic psilocybin formulation, under appropriate UK regulatory approvals and ethical review. Eligible participants were females assigned at birth, aged 21-65 years, with a primary DSM-5 diagnosis of anorexia nervosa likely present for more than 3 years. They had to be registered with a UK specialist eating disorder team and have failed to achieve remission despite previous treatment. Exclusion criteria included psychotic disorders, unstable medical or psychiatric illness, contraindications to MRI or EEG, and substance dependence within the previous 6 months. Participants also identified a support person. Recruitment came from two NHS specialist eating disorder services and online self-referral via social media and charity websites. The target was 20 study completers. The protocol involved three fixed-order oral dosing sessions over 6 weeks: 1 mg, then 25 mg, then 25 mg, each separated by 2 weeks. Participants were told they would receive psilocybin and that the maximum dose would be 25 mg, but they were not told the order of the remaining doses, so the study was single-blind. Each dosing day was preceded by a preparation session the day before and followed by an integration session the next day. There were remote monthly follow-ups for 6 months and a final remote 12-month follow-up. Feasibility was assessed through recruitment, screening and retention. Safety and tolerability were tracked through adverse events and body mass index, although BMI was not a preregistered outcome. The two primary efficacy outcomes were global scores from the Eating Disorder Examination Interview (EDE) and the precontemplation subscale of the Readiness and Motivation Questionnaire (RMQ), which measures lack of readiness to change. The Eating Disorder Examination Questionnaire (EDE-Q) was a key secondary outcome. Primary time points were the final 6-week visit plus 3-month and 6-month follow-ups for EDE, and the final visit for RMQ; RMQ and EDE-Q scores were also compared between dosing days. Analyses used intention-to-treat methods with last observation carried forward for missing data, linear mixed-effects models controlling for baseline scores, repeated-measures ANOVA for dose comparisons, paired Cohen’s d effect sizes, and a significance threshold of p<0.05. No adjustment was made for multiple comparisons because the study was exploratory.

Recruitment was substantial: from roughly 450 expressions of interest, 161 were considered potentially eligible, 100 attended telephone screening, 26 attended in-person screening and 21 were enrolled. One participant withdrew after dose 2, leaving a retention rate of 95.2% for the active study period and meeting the recruitment target of 20 completers. The enrolled sample was small and relatively homogeneous, with mean age 32 years, mean BMI 16.4 kg/m2 and mean illness duration 10.8 years. Most participants were White British, employed and had postgraduate education. Psychiatric comorbidity was common, particularly generalised anxiety disorder and major depressive disorder. Nearly half were withdrawn from contraindicated medication such as SSRIs. Psilocybin was reported to be well tolerated. All participants had at least one adverse event, with 107 events in total. Most were mild or moderate and were judged possibly related to the intervention. The most common events were transient headache and nausea. No participant required lorazepam. Two serious adverse events, both suicide attempts by one participant about 7 and 9 months after the final visit, were reported; these were judged unlikely to be related to psilocybin. Eating disorder symptoms improved significantly on the clinician-rated EDE. Global EDE scores fell over time, with significant reductions at the final visit, 3-month follow-up and 6-month follow-up compared with baseline. The strongest effects were seen at 3 months and were still present at 6 months. Higher baseline symptom severity predicted greater reduction. Antidepressant withdrawal was not a significant factor. Most participants improved, and the number falling within one standard deviation of the community norm increased at follow-up, although two were already within that range at baseline. Subscales showed significant time effects for dietary restraint, eating concerns, shape concerns and weight concerns, although weight concerns did not remain significant at 6 months. Self-reported global EDE-Q scores also improved significantly over time, with reductions visible from the final visit through 12 months compared with baseline. The largest numerical change occurred at 3 and 12 months. Again, antidepressant withdrawal was not a significant fixed effect. However, there was no significant main effect of dose on incremental change in EDE-Q scores. Motivation to change, measured by RMQ precontemplation, improved significantly over time. A notable reduction appeared after dose 2, and improvements were sustained across the 12-month follow-up. The largest reductions were seen at 2 and 3 months. Antidepressant withdrawal was again not a significant factor. As with EDE-Q, there was no significant main effect of dose on incremental RMQ change. Overall, the observed improvements appeared to accumulate across the protocol rather than being clearly linked to any single dose.

The authors interpret the findings as preliminary evidence that psilocybin therapy, delivered with psychological support as an adjunct to usual care, is feasible in adult females with anorexia nervosa. They emphasise the successful recruitment of the planned sample, high retention, and generally favourable tolerability as important findings, especially given the high dropout rates often seen in eating disorder treatment and research. They also state that the adverse event profile was consistent with earlier psilocybin studies and that the lack of observed hypoglycaemia-related symptoms is reassuring, although they note that this was not formally assessed. Douglass and colleagues compare their findings with previous anorexia nervosa psilocybin work and note both similarities and differences. They report that their lower-BMI cohort suggests safety may extend to more physically vulnerable patients, but they caution that the strict eligibility criteria produced a relatively homogeneous sample and limit generalisability. They also note that their symptom improvements were broadly consistent with earlier studies, although the pattern across eating disorder subscales differed. In their view, this may reflect differences in illness stage, baseline symptom profile and motivation for recovery. They also point out that the EDE and EDE-Q gave broadly consistent results. The authors are careful not to overstate causal inference. Because the study was single-arm, fixed-order and brief between doses, they say it was not possible to determine whether high-dose psilocybin was superior to the 1 mg dose or to separate drug effects from the broader treatment package. They note that improvements may reflect the entire protocol, time, or post-study life changes as much as the drug itself. They therefore argue that future trials need between-participant designs comparing psilocybin therapy with placebo or with established psychological treatments, ideally stratified by illness severity and subtype, with longer follow-up and better blinding approaches such as an active placebo or ethically acceptable mild deception about dose. The discussion also highlights several limitations: the small feasibility sample, no correction for multiple comparisons, the non-blinded EDE interviewer, and the predominance of highly educated White women. The authors mention that this demographic may have had more resources to engage in integration and concurrent care, which could affect reproducibility. They also note that the absence of a nutritional intervention and weight-restoration focus distinguishes this protocol from standard anorexia nervosa care. Despite these limitations, they conclude that the findings support further research into psychedelic therapy for anorexia nervosa and related disorders.

The authors conclude that this psilocybin therapy protocol appears feasible, safe and potentially efficacious as an adjunct treatment for adult females with anorexia nervosa. They state that recruitment and retention were successful, adverse events were in line with existing psychedelic research, and improvements in eating disorder psychopathology and motivation to change were observed and generally sustained over follow-up. They caution against broad extrapolation because of the sample and design, but argue that the results strengthen the case for larger studies of psychedelic therapy in anorexia nervosa and related disorders.