Keto diet for schizophrenia

A dietary solution for symptoms of psychosis?

This week, I’m preparing a short lecture with the title “Complementary, alternative, and nutritional treatments for schizophrenia.” I’ll be giving this lecture to an audience of clinicians who care for people affected by schizophrenia. (more about this lecture at the end of the letter)

One of the reasons I chose this topic is because “what about complementary, alternative, or nutritional treatments?” is one of the most common questions I get from affected people. Many such individuals are already trying out such treatments on their own. It’s a great question that has not received the scientific attention or and publicity it deserves.

So for this post, I’m going to steal from my upcoming lecture and focus on one of topics I’ll discuss: the ketogenic diet.

What is the ketogenic diet?

The ketogenic diet (also called keto diet) was developed about 100 years ago as a treatment for epilepsy. Most of the energy in a ketogenic diet comes from fat and (to a lesser extent) protein. Carbohydrate intake in the keto diet is very low, in the neighborhood of 20 grams per day. For reference, an average American diet contains around 300 grams of carbohydrates per day (far more than 300 grams if the diet is heavy in sweets or processed foods).

The diet is called ‘ketogenic’ because it causes the body to generate ketone bodies. “Ketone bodies” is an antique biochemical term for specific kinds of byproducts from fat metabolism. Having evolved in a time when food supplies were unpredictable, our bodies are perfectly able to run on fat instead of carbohydrates. When the body enters ‘fat-as-primary-fuel’ mode, it generates these ketone bodies.

The keto diet is an effective epilepsy treatment

Prolonged fasting will put a person’s body into ‘fat-as-primary-fuel’ mode, and prolonged fasting was recommended as a treatment for epilepsy in Greek medical texts from the Hippocratic era.

Prolonged fasting would be recommended, off and on again, for the next 2,000 years. The fasting treatment (the original pathway to ketogenesis) was ‘on-again’ in the early 20th century. But around that time, physicians reasoned that a high fat, minimal carbohydrate diet might produce similar results but without the hassle of long-term fasting. Such a diet, it was thought, would put the body into ‘fat-as-primary-fuel’ mode – but without the hassle or pain of prolonged fasting.

The ketogenic diet thus entered medical service in the 1920s as a treatment for epilepsy – and it was remarkably effective. A meta-analysis of 19 observational studies (1084 patients) found that after six months, approximately 60 percent of children started on the ketogenic diet had a greater than 50 percent seizure reduction, with 30 percent having greater than 90 percent seizure reduction. A 2009 expert consensus panel recommended the ketogenic diet as an option for people whose seizures are not well-controlled by medications.

If the keto diet is an anticonvulsant treatment, and if anticonvulsant drugs are often effective in psychiatric disorders, then could the keto diet address psychiatric symptoms?

Anticonvulsant drugs like valproic acid, carbamazepine, or lamotrigine are widely used in the treatment of bipolar disorder. They are also sometimes used in the treatment of schizophrenia. So, one wonders: would an anticonvulsant diet have any useful effects on symptoms of schizophrenia? This question, surprisingly, has not been systematically investigated. However, the data that do exist are encouraging.

Keto diet data in schizophrenia-spectrum disorders

Pacheco et al (1965) implemented a ketogenic diet in 10 adult women with longstanding symptoms of schizophrenia. The researchers observed significantly lower symptom severity (based on a schizophrenia symptom rating scale) after two weeks on the diet. Symptoms worsened in 7 of the 10 volunteers when they returned to normal-carb diet.

Kraft and Westman (2009) – Open Access Article report striking success of a ketogenic diet in a 70 year old woman who had experienced nearly-continuous symptoms since age 17. Her hallucinations went away after eight days on the diet, and her symptoms stayed in remission for a full year while she practiced the diet.

Palmer (2017) describes the case of a 33 year old man suffering from symptoms of schizoaffective disorder for approximately 14 years. The patient had been treated with multiple antipsychotic medications, with incomplete results. His PANSS score (a widely-used symptom rating scale) was 96 at the time that he decided to start a ketogenic diet. He maintained significant improvement for the following year. His PANSS score would eventually fall to 49 and achieved significantly higher levels of social and occupational functioning. The PANSS scale consists of 30 questions and the lowest possible score on each question is 1 – so the lowest possible PANSS score is 30. Going from 96 to 49 is a significant achievement). His symptoms returned quickly and dramatically within days of stopping the ketogenic diet, and symptoms remitted each time after re-establishing a state of ketosis.

Palmer also describes the case of a 31 year old woman with symptoms of schizoaffective disorder lasting approximately 8 years. Multiple therapeutic trials of mood-stabilizing and antipsychotic medications failed. And the patient remained symptomatic despite ECT treatment. The patient decided to go on a ketogenic diet for weight loss. Within 4 weeks, her psychiatric symptoms were noticeably less severe. Her PANSS score fell from 107 to 70. As in the prior case, symptoms worsened when the diet was stopped and symptoms improved with the re-establishing of ketosis.

Meanwhile, Phelps et al. (2013) present two case reports of multi-year stabilization of type 2 bipolar symptoms.

Mechanism of action for the keto diet’s neuropsychiatric effects

The mechanism(s) whereby the keto diet might be causing these possible improvements is not exactly clear.

Modern scientific studies (Open Access article) suggest that the ketogenic diet may affect the activity of the excitatory neurotransmitter glutamate and/or may reduce brain excitability by affecting the rate of potassium flux across nerve cell membranes.

I am most intrigued by the hypothesis that the keto diet helps to normalize the brain’s insulin sensitivity. Insulin resistance is a common feature of both schizophrenia and depression. And the brain possesses insulin receptors. Drugs that work to normalize insulin response may possess antidepressant actions (Kemp et al., 2012 – Open Access article).

Others have suggested that the reported benefits of the keto diet for schizophrenia might have nothing to do with ketosis. Rather, they say, it might be as simple as removing certain foods from the diet. The keto diet is usually gluten-free.

 The gluten connection to schizophrenia

I’ve written before about cases where gluten sensitivity (celiac disease, to be more precise) produced symptoms that were identical to schizophrenia. In both of the published case reports, schizophrenia-identical symptoms showed up years before the other signs of celiac disease. In both cases, antipsychotic treatment did little to reduce the symptoms of psychosis. But in each case, the psychosis went away when gluten was removed from the diet. These case reports are especially interesting in light of large clinical studies showing that gluten-related antibodies are about 7 times more likely among people with schizophrenia compared to the general population.

In addition to the case reports, Dohan and colleagues conducted case-control studies of a grain-free, dairy-free diet among hospitalized patients with schizophrenia. They found that the dietary intervention caused more rapid resolution of symptoms. They also observed rapid re-appearance of psychosis if gluten was added back to the diet.

Should we be recommending the keto diet to people with schizophrenia?

When it comes to putting medical literature findings into clinical action, there are three types of response: scientific detachment, independent adoption of the practice by highly motivated individuals, or silence on the part of professional or government oganizations. Almost invariably, scientists or the study authors who report groundbreaking findings will end their papers with a call for more research (‘These findings are interesting, potentially amazing actually… but we need more studies before any practice recommendations can be made’). On the other hand, affected people – the ones with the most at stake – will want to go ahead and adopt the finding. Especially if the intervention is seen as low risk. This is partly why complementary, alternative, or nutritional treatments are so popular. Meanwhile, professional organizations like the American Psychiatric Association are usually silent. And regulatory agencies like the FDA (the ones most people and all insurance companies look to for approval) are not even called into play.

To my way of thinking, the keto diet’s risk level is low; and where there are risks, they are well-known (recall that it’s been in use for about 100 years now as a neurological therapy). And on the other hand, its potential benefit (remission of psychosis for at least some individuals) is enormous. Most likely, in my view, it will work well for some people and not for others. In this way, it’s really no different than almost every other therapy in Medicine. The only way to know if it will work for someone would be to try it.

About that upcoming lecture

Every Tuesday, at noon (eastern time), I host a videoconference called SZconsult. SZconsult has its own informational website.

SZconsult is a clinician telementoring service, offered at no cost by the Best Practices in Schizophrenia Treatment (BeST) Center at Northeast Ohio Medical University. Additionally, the University awards free Category 1 Continuing Medical Education credit to SZconsult participants.

SZconsult sessions consist of a short lecture followed by case consultation. Any clinician is welcome to ask any questions related to the care of their patients. We work toward helping clinicians solve the issues that might limit the full recovery of their patients.

If you’re a clinician who works with people with schizophrenia-spectrum illness (or know someone who is), feel free to contact me for conference login details.


Receive email notices of new posts to The Unsponsored Psychiatry Report. It's free, and you can unsubscribe at any time. You will be required to verify your email address.