Psilocybin Therapy Shows Promise for Anorexia Nervosa in Early Clinical Study
A 2026 pilot study has reported sustained improvements in eating disorder symptoms and motivation to change after psilocybin therapy in women with longstanding anorexia nervosa.
The findings are clinically significant.
They do not establish psilocybin as a treatment for anorexia nervosa.
Published in The British Journal of Psychiatry, the study enrolled 21 adult women whose illness had likely been present for more than three years and who had not achieved remission despite previous treatment. Participants received three doses of COMP360, a synthetic psilocybin formulation, over six weeks.
The protocol involved considerably more than administering psilocybin. Each participant received preparation, supervised dosing, integration, talk therapy, continued treatment as usual, and approximately 50 hours of therapeutic contact.
At six months, clinician-rated eating disorder symptoms were significantly lower than at baseline. Motivation to change also improved, with gains still observed at the 12-month follow-up.
The signal deserves further investigation.
The study design does not allow us to conclude that psilocybin caused these changes.
What the Study Actually Tested
This was a single-arm, single-blind pilot study conducted by researchers at Imperial College London.
All 21 participants followed the same dosing sequence:
1 mg of psilocybin
25 mg of psilocybin two weeks later
A second 25 mg dose after another two weeks
Participants knew they would receive psilocybin but did not know the dose assigned to each session.
Every dosing day was preceded by a preparation session and followed by an integration session. Participants also continued receiving treatment through a specialist eating disorder service or private therapist.
Researchers evaluated two primary clinical outcomes:
Eating disorder symptoms measured through the clinician-administered Eating Disorder Examination
Motivation to change measured through the Readiness and Motivation Questionnaire
Participants were between 23 and 52 years old. The estimated duration of illness ranged from four to 22 years, with an average of 10.8 years. Their average baseline body mass index was 16.4 kg/m², and all but one participant were within the underweight range.
The screening process was extensive.
Approximately 450 people expressed interest. Of those, 100 completed telephone screening, 26 attended an in-person assessment, and 21 were enrolled.
Participants had to be connected with a specialist eating disorder team and identify a family member or friend who could serve as a support person. The study excluded people with unstable medical or psychiatric conditions, psychotic disorders, or recent substance dependence.
This matters when interpreting the results.
The study examined a carefully selected group within a highly structured clinical environment. It did not test psilocybin among the broader population of people living with anorexia nervosa, nor did it examine unsupervised or independently facilitated use.
Eating Disorder Symptoms Improved, but Outcomes Varied
Clinician-rated eating disorder symptoms declined significantly over time.
At the six-month follow-up, researchers reported a large within-group effect size of Cohen’s d = 0.98. Improvements were observed in dietary restraint, eating concerns, and shape concerns. Improvements in weight concerns were not maintained at six months.
Using proximity to a community norm as a clinical reference point:
Six participants were within that range at the final six-week visit
Ten participants were within that range after three months
Eight participants were within that range after six months
Two of these participants were already within the community-norm range at baseline despite meeting the diagnostic criteria for anorexia nervosa.
Motivation to change also improved. Reductions in precontemplation scores remained significant at 12 months, with an effect size of d = 0.65.
These aggregate results require context.
The researchers reported considerable variation in how participants responded and whether improvements were maintained. Some experienced sustained changes. Others showed smaller gains or lost improvements during follow-up.
The study therefore does not support the claim that three psilocybin sessions reliably produce recovery from anorexia nervosa.
It shows that a potential therapeutic signal appeared in a small, intensively supported group.
The Dose Findings Are More Complicated Than They Appear
The researchers did not find a significant relationship between dose and incremental changes in eating disorder symptoms or motivation.
That does not mean 1 mg and 25 mg were equally effective.
The study administered the doses in a fixed sequence. Every participant received 1 mg first, followed by two 25 mg doses, with only two weeks between sessions. This made it impossible to separate the effects of dose from accumulated therapy, prior dosing, expectancy, time, or the developing therapeutic relationship.
The study also lacked a parallel placebo group.
Psychedelic trials face a persistent methodological challenge because the noticeable subjective effects of a high dose make blinding difficult. Participants and clinicians may infer whether an active dose was administered, which can influence expectations and reported outcomes.
This is why the findings should be understood as exploratory. The study was designed primarily to evaluate feasibility and safety, not to prove that high-dose psilocybin was superior to placebo or established treatment.
This Was Psilocybin Therapy, Not Psilocybin Alone
The intervention included approximately 50 hours of therapeutic contact over six weeks.
Therapists incorporated elements of emotion-focused family therapy, compassion-focused therapy, and systemic therapy. Participants received preparation before each dose and integration afterward. They remained connected to their existing specialist care, and a family member or friend was included as a support person.
There was no nutritional intervention or specific focus on weight restoration within the study protocol.
This distinction is clinically important.
The changes observed cannot be attributed to psilocybin alone. They may have resulted from the medicine, intensive psychological support, expectations, ongoing care, repeated assessment, the therapeutic relationship, or an interaction among these elements.
The study examined a complete therapeutic system.
Reducing its findings to “psilocybin treats anorexia” removes the clinical conditions under which the outcomes were observed.
For practitioners, the protocol reinforces a central principle of psychedelic-assisted therapy: screening, preparation, therapeutic presence, supervised administration, integration, and continuity of care are not accessories to the intervention.
They are part of it.
Motivation May Be an Important Therapeutic Target
Anorexia nervosa is not defined only by low body weight.
It can involve persistent restriction, fear of weight gain, disturbances in body perception, rigid patterns of thinking, and an overvaluation of weight or shape. Elements of the illness can also become closely tied to identity, safety, achievement, or control.
This can make motivation for recovery clinically complex.
There is currently no medication approved specifically for treating the core symptoms of anorexia nervosa. Psychological and nutritional interventions remain central, but treatment outcomes vary, relapse is common, and some patients develop a chronic course.
The improvement in motivation observed in this study may therefore be one of its most important findings.
One possibility is that psilocybin therapy may help some patients relate differently to rigid beliefs, self-criticism, fear, and illness-related identity. It may create an opportunity for greater psychological flexibility or willingness to engage in recovery.
That remains a hypothesis.
The study did not establish a mechanism, and it did not show that psychological insight produces nutritional rehabilitation or medical recovery.
A change in perspective may open a therapeutic window. It does not correct malnutrition, stabilize cardiovascular risk, or replace specialist eating disorder treatment.
Safety Cannot Be Reduced to Acute Tolerability
The study reported a 95.2 percent retention rate. Twenty of the 21 participants completed the active study period.
The most common adverse events were headache, nausea, and dizziness. One participant withdrew after the second session because of the time and emotional demands of participation.
One participant made two suicide attempts approximately seven and nine months after the final study visit. Investigators classified these as serious adverse events of moderate severity and considered them unlikely to be related to psilocybin.
Both parts of that finding matter.
The investigators did not attribute the events to the study drug. But the events must remain visible in any responsible discussion of safety, particularly because anorexia nervosa is associated with substantial psychiatric and medical risk.
A study involving 21 participants cannot establish the long-term safety profile of an intervention for this population.
Medical vulnerability also requires particular attention. Anorexia nervosa can involve cardiovascular instability, electrolyte disturbances, endocrine changes, impaired thermoregulation, hypoglycemia, and other complications that may affect the safety of an extended psychedelic session.
A previous psilocybin study involving women with anorexia nervosa documented two clinically noteworthy cases of post-dosing hypoglycemia. The 2026 study did not formally assess hypoglycemia. Researchers reported no related symptoms but appropriately noted that an absence of observed symptoms does not establish that it did not occur.
This is why enthusiasm cannot replace multidisciplinary screening and medical oversight.
The Study Does Not Support Independent Psilocybin Use
Nothing in this study supports self-treatment or the use of psilocybin outside specialist care.
Participants underwent psychiatric and physical evaluation, remained connected to an eating disorder treatment team, received extensive psychological support, and completed dosing sessions within a regulated research environment.
They were also selected through strict eligibility criteria that excluded many of the conditions that could increase risk.
Those safeguards limit generalizability, but they also define the conditions under which the research was conducted.
Translating the findings into independent use would not be an extension of the evidence. It would be a departure from it.
What the Next Trial Must Establish
The study provides a reason for further research, not a reason to bypass it.
Future trials will need larger and more diverse samples, randomization, credible comparison conditions, independent outcome assessment, and predefined methods for managing expectancy and multiple statistical comparisons.
Researchers will also need to answer several practical questions:
Which patients may benefit, and which face unacceptable medical or psychiatric risk?
How much of the observed change comes from psilocybin, intensive therapy, or their interaction?
Do improvements lead to durable functional and nutritional recovery?
How should body weight, cardiovascular status, and metabolic risk affect eligibility?
Which preparation and integration practices are clinically necessary?
How do outcomes compare directly with established specialist treatments?
Can this model be delivered safely outside a resource-intensive research center?
What level of training should be required for practitioners working with eating disorder populations?
The current study cannot answer these questions.
Its contribution is showing that they are now reasonable and necessary to ask.
A Clinical Signal Is Not a Small Result
Psychedelic research is often compressed into a binary story: breakthrough or failure.
This study supports neither conclusion.
It provides preliminary evidence that a carefully screened and intensively supported psilocybin therapy protocol was feasible for 21 adult women with longstanding anorexia nervosa.
It found improvements in eating disorder symptoms and motivation to change. It also found considerable variation between participants and included methodological limitations that prevent causal conclusions.
That is not a treatment verdict.
It is a clinical signal.
For a disorder in which rigidity can become intertwined with identity, even the possibility of greater psychological movement matters. The responsibility of the field is to study that possibility without converting vulnerability into certainty.
The promise of this work will not be measured by how quickly psilocybin is called transformative.
It will be measured by whether research can turn an early opening into care that is safe, reproducible, medically integrated, and worthy of the people who place their lives inside it.
