New RCT Adds Evidence for Ketogenic Diets in Psychosis

If you live with schizophrenia, schizoaffective disorder or bipolar disorder, you have probably experienced firsthand how antipsychotic medication, while often essential, can bring weight gain, high blood sugar and other metabolic problems along with it. Building on earlier pilot and case study work, a new study, just published in Schizophrenia Bulletin, provides the first randomised controlled evidence that a ketogenic diet can improve not just metabolic health but psychiatric symptoms and thinking ability too. The results are worth understanding. 

Who ran the study, and why

The trial was led by researchers at the San Francisco VA Medical Center and the University of California, San Francisco, in collaboration with Stanford University. Their starting point was a well known problem, antipsychotic medications are linked to high rates of metabolic syndrome (a combination of excess weight, high blood pressure and disrupted blood sugar and cholesterol) and this metabolic dysfunction is itself associated with worse quality of life, more severe symptoms and impaired cognition. Some smaller, uncontrolled studies and case reports had already hinted that a ketogenic diet, the very high fat, low carbohydrate diet best known for treating epilepsy, might ease both metabolic problems and psychiatric symptoms. This trial set out to test that idea properly, with a randomised control group, for the first time in people with psychosis.

How the study was designed

58 adults with schizophrenia, schizoaffective disorder, or bipolar disorder with psychotic features, were randomly assigned to either follow a ketogenic diet for one month (28 people) or continue eating as they normally would (30 people, the diet as usual group). Those on the ketogenic diet received ready made low carbohydrate meals each week and were coached on how to structure the diet, aiming for roughly 70 to 80 percent of calories from fat, 10 to 20 percent from protein and no more than 10 percent from carbohydrates.

Partway through, additional funding allowed the trial to extend, and everyone who wished to continue was offered a further period on the ketogenic diet, so that a subgroup of 25 participants ultimately followed the diet for four months in total. Throughout the study, the researchers measured metabolic markers such as weight, blood sugar control and insulin resistance, psychiatric symptoms including positive symptoms (such as hallucinations and delusions), negative symptoms (such as reduced motivation and social withdrawal) and depression, and performance on a standard battery of cognitive tests.

What actually improved

After just one month, people on the ketogenic diet, compared with those eating their usual diet, showed a meaningful drop in weight, along with improvements in haemoglobin A1c (a blood marker reflecting average blood sugar over the past two to three months) and insulin resistance. These are genuine metabolic gains, achieved in a single month.

The psychiatric and cognitive benefits took longer to emerge. At one month, symptom and cognition scores had moved in a favourable direction but had not yet reached statistical significance. By four months, however, the participants who stayed on the ketogenic diet showed significant improvements across the board, in depressive symptoms, in positive symptoms of psychosis, in negative symptoms and in overall cognitive performance. The authors note that of the participants who improved on depressive symptoms after the first month, most improved further by four months, and a similar pattern of gradual, continued improvement was seen for cognition and negative symptoms. In other words, the diet's psychiatric benefits build over time rather than appearing immediately.

The power of ketosis, not just weight loss

An important finding concerns what seemed to be actually driving the improvements. The researchers measured beta hydroxybutyrate, a ketone body that rises when the body is in ketosis, and found that participants who achieved higher ketone levels, and who sustained ketosis more consistently, showed the biggest improvements in haemoglobin A1c and depressive symptoms. Weight loss on its own, by contrast, did not explain these improvements. This suggests that being in a state of ketosis, rather than simply losing weight, may be the key therapeutic ingredient. Because weight loss alone did not track with the same benefits, the improvements are unlikely to be explained by general metabolic or anti-inflammatory change alone; the authors point to prior evidence that beta hydroxybutyrate acts directly on brain cells, supporting neural plasticity and dampening excitotoxic glutamate signalling, mechanisms already implicated in psychosis, which raises the possibility that ketone bodies affect the brain in ways that go beyond metabolic improvement. This study was not designed to test that neuronal pathway directly, so the authors are appropriately cautious and call for larger studies to confirm it.

Feasibility, and important caveats

The diet appeared feasible for outpatients with serious mental illness over the first month, with the large majority of those assigned to KETO completing it, though retention fell over the longer four month extension, where only about half of the participants originally assigned to the ketogenic diet went on to complete the full course. No serious adverse events occurred; minor issues like headaches resolved with more fluids and electrolytes. Twenty five of the twenty eight people originally assigned to the ketogenic diet completed the first month, and adherence was confirmed by home ketone monitoring, not just self report.

That said, the study has real limitations the authors are upfront about. Because of limited funding, the diet as usual comparison group did not receive any prepared meals, so the ketogenic diet group had the advantage of free, easy, fresh food delivered to them, and the researchers cannot fully rule out that expectations about receiving a novel treatment influenced some results. The four month findings come from a single arm extension without its own matched control group at that time point, so those results are less rigorous than the one month, randomised comparison. The sample, under thirty people per arm, precluded the researchers from testing more complex statistical models and from performing diagnostic subgroup analyses, and blood markers such as cholesterol and electrolytes, which matter for cardiovascular and kidney safety on a high fat diet, were not measured. Medication changes during the study were not tracked. Inflammation, measured by C-reactive protein, did not improve, which the authors note runs against some expectations for ketogenic diets.

What this means if you are considering it

This trial does not prove that a ketogenic diet treats psychosis. What it does show is that a carefully supported ketogenic diet is feasible for people with schizophrenia spectrum and bipolar disorders, produces real metabolic improvements within a month, and is associated with broader psychiatric and cognitive gains that appear to strengthen the longer ketosis is sustained. The authors themselves are careful to frame this as motivation for larger, longer, better controlled trials, some of which are already underway or recently completed.

If you live with a psychotic disorder and are curious about this approach, the sensible next step is a conversation with your psychiatrist or psychologist before making any dietary change, particularly given the medication interactions, monitoring needs and, for some people, the potential unsuitability of a restrictive diet if there is a history of disordered eating.

This research marks an exciting step forward for the growing field of metabolic psychiatry and the evidence base linking metabolic and mental health looks set to expand considerably in the years ahead. Further RCT evidence is on its way, with additional trials of ketogenic metabolic therapy in serious mental illness about to be published soon, so watch this space! Exciting times!

Citation: Abram SV, Kyner JM, Vu A, Naeem Z, Sethi S, Jacob MS, Fryer SL, Mathalon DH, Ford JM. Metabolic Improvements with a Ketogenic Diet Correlate with Symptom Improvement in Psychosis: A Randomized Controlled Trial. Schizophrenia Bulletin, 2026;52(4):1-11. https://doi.org/10.1093/schbul/sbag082