A ketogenic diet may sound like an especially counterintuitive therapy for anorexia nervosa. The eating disorder is associated with food restriction, low body weight and an intense fear of gaining weight. A diet that limits carbohydrates could easily be mistaken for another form of restriction.
But researchers are studying a carefully supervised version of ketogenic therapy as a possible way to address persistent psychological symptoms in some adults with anorexia nervosa. In a small pilot trial, participants received individualized support intended to help them maintain their weight while following the diet. The findings were encouraging enough to justify more research, but they do not show that a keto diet is an effective treatment for anorexia nervosa or that anyone should attempt the approach without specialized medical care.
The peer-reviewed feasibility trial, published in Communications Medicine, enrolled 22 adults with anorexia nervosa who were either weight-restored or mildly underweight. All of the participants were female. Eighteen completed the 14-week outpatient intervention.
Anorexia nervosa is a serious psychiatric disorder. Weight restoration is an important part of treatment, but it does not always resolve symptoms such as an intense fear of eating, body dissatisfaction or a persistent focus on weight and shape. Those symptoms can increase the risk of relapse.
“We urgently need new approaches to anorexia nervosa,” said Guido Frank, the study’s lead author and a professor of psychiatry at the University of California San Diego School of Medicine.
Researchers are exploring whether metabolism may play a role in the disorder alongside its psychological and behavioral dimensions. Genetic studies and other research have raised questions about whether differences in glucose and lipid metabolism may be associated with anorexia nervosa. Scientists do not yet know what those findings mean for treatment.
The new study tested whether ketogenic therapy could be delivered safely and tolerated by adults with anorexia nervosa who were not severely underweight. It was not a trial of a typical consumer keto diet.
Participants were advised to eat three meals and two snacks each day. The diet was designed to provide about 70% of calories from fat, 20% from protein and 10% from carbohydrates. Participants received guidance from a dietitian, and meals were provided through a ketogenic meal-delivery service when needed. Some participants prepared their own food with dietitian support.
The research team closely monitored participants’ weight, symptoms and ketone levels. Participants used connected scales that transmitted weight information to the study team and completed weekly assessments covering eating-disorder symptoms, mood, anxiety and suicidal thoughts. They also met regularly with a psychiatrist and a dietitian.
The goal was weight maintenance, not weight loss. Participants were instructed to eat more if they began losing weight and could not remain in the study if their body mass index fell below a set threshold.
At the end of the intervention, 72% of the participants who completed the study had eating disorder and depression scores within the normal range on standardized questionnaires. Depression scores improved among all completers. The researchers reported that the intervention did not lead to clinically significant weight loss overall or a worsening of symptoms.
Those results are notable, but they require careful interpretation. Having symptom scores within the normal range is not the same as proving that someone has fully recovered from anorexia nervosa. The study also cannot establish that ketogenic therapy caused the improvements.
The trial did not include a control group receiving another diet or a placebo intervention. That means the researchers could not rule out other explanations, including expectation effects, changes in medication, contact with outpatient providers or other changes in eating behavior. Some participants continued receiving psychotherapy, and its influence on the results was not systematically assessed.
The study was also small. All participants were female, and most were white and non-Hispanic. Only three participants were mildly underweight. The results cannot be generalized to everyone with anorexia nervosa, particularly people who are more severely underweight or medically unstable.
The findings also do not establish that ketogenic therapy is nutritionally appropriate or sustainable over the long term. The study did not include extensive laboratory testing before and after the intervention, including cholesterol measurements. Some participants had difficulty eating enough ketogenic food to reach ketosis, and one found the protocol too restrictive in social settings.
The underlying explanation remains uncertain. Researchers have proposed that ketogenic therapy could affect how the brain uses energy, but this study did not directly test that mechanism.
Frank said researchers are hopeful that a metabolic intervention could eventually help address psychological symptoms. For now, the appropriate next step is a larger controlled trial, not a new diet trend.
That distinction matters because ketogenic diets are widely promoted online for purposes ranging from weight loss to mental health. A medically supervised experimental therapy for a serious eating disorder should not be treated as a do-it-yourself eating plan.
Anyone experiencing signs of an eating disorder should seek care from a qualified health professional. People with anorexia nervosa or a history of restrictive eating should not begin a ketogenic diet based on the findings of this early trial.
The study was funded by Baszucki Group, which supports research into metabolic interventions for mental health conditions, including ketogenic therapy.