Ketogenic diet and ketamine infusion treatment to target chronic persistent eating disorder psychopathology in anorexia nervosa: a pilot study

Anorexia nervosa (AN) is a severe, often chronic psychiatric disorder characterised by self-starvation, intense fear of weight gain, distorted body image and entrenched food- and exercise-related compulsions. Existing treatments, including psychotherapies and available medications, have limited effectiveness for severe and enduring forms of AN, and there is no FDA‑approved pharmacotherapy. Recent preclinical and clinical observations implicate biological mechanisms in AN's pathophysiology: in activity‑based anorexia (ABA) animal models, high‑fat low‑carbohydrate diets prevented self‑starvation and promoted recovery, and ketamine (an NMDA receptor antagonist) improved obsessional symptoms in prior human reports. Both ketogenic diet and ketamine modulate dopaminergic systems, which have been proposed as centrally involved in AN, suggesting a mechanistic rationale for combining metabolic (therapeutic ketogenic diet, TKD) and pharmacological (ketamine) interventions. Calabrese and colleagues designed an open‑label pilot trial to test whether the therapeutic sequence observed in a prior case report—TKD to induce nutritional ketosis followed by a course of ketamine infusions—could be replicated in a small group of adults with long‑standing AN who were weight recovered but continued to experience severe eating‑disorder psychopathology. The primary aim was to assess safety, feasibility and change in AN‑related symptoms (shape/weight concerns, eating‑related preoccupations and functional impairment) over a 6‑month follow‑up period; the TKD was deliberately aimed at maintaining, not reducing, body weight.

This was an open‑label, single‑arm pilot study enrolling adult females with a history of AN of at least 10 years who were weight restored but had persistent AN‑typical behaviours. Recruitment occurred via community advertisements; 25 people were screened and five met eligibility criteria. Eligibility included medical clearance, normal baseline laboratory values and confirmation of lifetime AN diagnosis using DSM‑5 criteria and the EDA‑5 interview. Participants continued any outpatient psychiatric medications under their usual providers. The study received institutional review board approval and was registered on ClinicalTrials.gov. The intervention sequence began with a 2‑day immersive TKD programme led by an experienced ketogenic dietitian and multidisciplinary staff, followed by home maintenance of nutritional ketosis for 4–8 weeks and then six racemic ketamine infusions delivered in an outpatient psychiatric clinic. The TKD targeted macronutrient proportions of 70% fat, 20% protein and 10% carbohydrate and was explicitly intended to induce nutritional ketosis without promoting weight loss. Each participant attended the immersion with a designated adult support person; meals were prepared and consumed together with staff and peer support, including the index case from the prior case report. Ketosis was monitored non‑invasively with a commercial breath acetone meter (Keyto) twice daily for 2 weeks then daily, with a goal breath score of ≥4 on the device's 0–6+ scale. Weight measurements were collected but blinded to participants via a scale that transmitted data to the team. After at least 4 weeks of nutritional ketosis, participants could extend TKD for a further 4 weeks before ketamine administration. Ketamine was given as six intravenous infusions within a 17‑day period; initial dose was 0.75 mg/kg with a titration target around 1.2 mg/kg administered over 45 minutes, doses modelled on the index case and adjusted to elicit dissociation. Standard physiological monitoring occurred during infusions, ondansetron 8 mg was given pre‑infusion, and nausea/dizziness were treated as needed. Follow‑up assessments occurred weekly for 4 weeks after the final infusion, then monthly for 3 months and every 3 months up to 6 months for the outcomes reported here. Outcome measures included the Eating Disorder Examination Questionnaire (EDE‑Q), the Clinical Impairment Assessment (CIA), the Eating Disorders Recovery Questionnaire (EDRQ), and the Patient Health Questionnaire‑9 (PHQ‑9), together with weight and qualitative interviews. Data analysis used repeated measures ANOVA with false discovery rate (FDR) correction for multiple comparisons; missing data were handled with last observation carried forward (LOCF).

All five enrolled participants were white females aged 29–45 years (mean 38.4 ± 7.4). All had a history of restricting‑type AN and were weight restored at enrolment; comorbidities included major depressive disorder in all five participants and variable anxiety, OCD, PTSD or body dysmorphic disorder. Final ketamine doses varied by participant from 0.75 mg/kg to 1.3 mg/kg. Of the 25 initially screened, five entered the trial and all completed the protocol. Adoption of nutritional ketosis was feasible in this sample: three participants reached breath acetone levels above 5 by the end of the 2‑day immersion, a fourth achieved ketosis within the first week and one required an additional 4 weeks to reach sustained ketosis. All participants maintained nutritional ketosis for at least 4 weeks prior to ketamine. Two participants elected to continue TKD for 8 weeks before ketamine and remained on a liberalised TKD throughout follow‑up; three participants completed 4 weeks of TKD before ketamine and tapered off TKD after the final infusion. No serious or unexpected adverse events were reported during TKD immersion, home adherence or ketamine infusions; specifically, no participant reported suicidal ideation. Weight (BMI) remained stable across time points for the group and did not change significantly in repeated measures analysis, although one participant showed weight decline and clinical relapse 4 months after treatment and was admitted to residential care; that participant had discontinued TKD prior to relapse. On psychometric outcomes, the group showed statistically significant improvements over time on multiple measures: Clinical Impairment Assessment (CIA) (p = 0.008), EDE‑Q Global score (p = 0.006), EDE‑Q Eating Concern (p = 0.005), EDE‑Q Shape Concern (p = 0.016), EDE‑Q Weight Concern (p = 0.032), EDRQ Acceptance of Self and Body (p = 0.027) and EDRQ Social and Emotional Connection (p = 0.001). Missing questionnaire data occurred intermittently across participants and time points; LOCF was used for imputation. The authors report large effect sizes for symptom change despite the small sample. Two participants who remained on TKD through 6 months reported continued symptom remission.

Calabrese and colleagues interpret the observed pattern—improvement after TKD and further benefit following ketamine infusions—as preliminary evidence that the TKD-plus‑ketamine sequence can reduce persistent AN‑related psychopathology in weight‑restored adults. Symptom improvement occurred without significant weight loss or severe adverse effects in this small sample, supporting feasibility and an acceptable short‑term safety profile. The investigators emphasise that while TKD alone yielded meaningful relief for some participants, the addition of ketamine produced additional symptomatic gains, suggesting additive effects rather than a ceiling effect from diet alone. The authors situate these findings within emerging biological models of AN that implicate altered neurotransmission and metabolic processes, especially dopamine circuitry. They note that ketogenic metabolism provides ketone bodies as alternative brain fuel (supplying an estimated 60% of cerebral energy in ketosis), which can alter neuronal physiology via mitochondrial function, lipid signalling, adenosine modulation, greater glucose stability and changes in the gut microbiome; ketamine may further modulate dopaminergic and emotional‑processing circuits. However, the study cannot determine causal mechanisms, and the neural bases of the clinical changes remain speculative and require targeted mechanistic research. Key limitations acknowledged by the investigators include the very small sample size, lack of a control group, heterogeneity and titration of ketamine dosing, variable duration of TKD before ketamine, reliance on breath acetone rather than blood beta‑hydroxybutyrate to quantify ketosis, and the presence of comorbid major depressive disorder in all participants which could confound whether improvements were primarily antidepressant effects versus direct effects on AN psychopathology. Missing data were managed with LOCF. The authors conclude these pilot results are promising and justify larger, controlled studies to evaluate efficacy, safety and mechanisms, and to determine whether longer‑term TKD maintenance is therapeutic for some individuals with chronic AN.

The study concludes that a therapeutic ketogenic diet aimed at inducing and maintaining nutritional ketosis, followed by a short course of ketamine infusions, was feasible, well tolerated and associated with symptomatic improvement in a small sample of adults with long‑standing AN who were weight restored but continued to experience severe preoccupations with weight, shape and self‑acceptance. The investigators recommend larger, controlled trials to confirm effectiveness and to explore mechanisms underlying these effects.