Most people meet keto as a weight-loss diet, but that is not where it came from. The ketogenic diet was developed in the 1920s as a treatment for epilepsy, and it has been used for that ever since, long before anyone thought of it for the waistline. That history matters, because epilepsy is the one area where the evidence for keto is not new or tentative but decades deep.
Why it started with seizures
A seizure is, very loosely, a storm of excess electrical activity in the brain. Researchers noticed early on that fasting could calm that activity, and the ketogenic diet was designed to recreate the fasting state, with its high ketone levels, while still letting a person eat. Ketones appear to steady the brain’s energy supply and dampen the runaway excitability that drives seizures. The exact mechanisms are still being worked out, but the clinical effect has been recognised for a century, particularly in children with hard-to-treat epilepsy.
The modern question: does it work for adults?
For a long time the strong evidence was mostly in children, and ketogenic therapy was seen as a children’s treatment. The open question was whether it helps adults and teenagers whose seizures resist medication too. A randomised controlled trial set out to test exactly that, enrolling 160 people, half of them adolescents and half adults, all with epilepsy that had not responded to anti-seizure drugs, and adding a modified Atkins diet, a more liveable, less rigid form of ketogenic eating, alongside their usual medication.
What the trial found
The difference was clear. Over six months, about 26 per cent of the diet group achieved at least a halving of their seizures, compared with roughly 2.5 per cent of the control group, and a small number became seizure-free. Notably, the adults responded as well as or better than the adolescents, putting to rest the idea that this is only for children. Quality of life also improved in the diet group. For people whose seizures have shrugged off the available drugs, a measurable extra reduction from diet alone is a meaningful thing.
The honest limits
A few points keep this in proportion. The diet was an add-on to medication, not a replacement for it, and that is exactly how it should be understood: nobody in these studies stopped their anti-seizure drugs in favour of food. Sticking to the diet is hard, which is why the more flexible modified Atkins version was used, and adherence is the main obstacle in practice. Researchers also flag longer-term issues to watch, such as blood lipids and bone health, which is part of why this is done under medical supervision rather than alone.
This is genuinely a medical therapy
It is worth being blunt here in a way the other articles on this site need not be. Ketogenic therapy for epilepsy is a clinical treatment delivered by neurology and dietetics teams, not a self-help diet. It involves careful set-up, monitoring of bloods and ketones, and coordination with medication, because getting it wrong, or changing seizure medication without guidance, can be dangerous. If you or someone you care for has drug-resistant epilepsy, this is a real and evidence-backed option to raise with the treating neurologist, and that is the route to it.
The bottom line
Keto’s oldest and best-evidenced use is epilepsy, going back a hundred years. A recent randomised trial confirmed that a modified ketogenic diet meaningfully reduces seizures in adolescents and adults with drug-resistant epilepsy, not just children, when added to their medication under medical care. Of all keto’s claimed benefits, this is the one resting on the firmest ground.
This is general information, not medical advice. Ketogenic therapy for epilepsy is a medical treatment that must be set up and monitored by a neurology and dietetics team. Never start it for seizures on your own, and never change anti-seizure medication without your doctor.
Source: Cervenka MC. Keto Is Not Just for Kids: A Randomized Trial of a Modified Atkins Diet for Adolescents and Adults With Anti-Seizure Medication-Resistant Epilepsy. Epilepsy Currents. 2023;23(3):147-149. Read it here.