Loading ...

Before the development of antiseizure medications (ASMs), the ketogenic diet was the gold standard for treating epilepsy.

Developed in the 1920s to mimic the metabolism of fasting, the high-fat, low-carbohydrate regimen is highly effective. Among those who try ketogenic diet therapies, nearly half achieve at least a 50% reduction in seizures. Some show freedom from seizures altogether.

After 105 years of clinical use, the diet remains an important neurologic tool, although guidelines primarily recommend its use in the estimated 30% of patients with drug-resistant epilepsy (DRE). Despite its proven efficacy, only young children have good adherence rates. The diet requires near-total carbohydrate elimination, constant monitoring, and a level of dietary restriction that most adolescents and adults find unsustainable.

Efforts to make adherence more palatable by developing pharmacologic alternatives that mimic the diet’s effects have failed, in large part because why the diet works remains largely a mystery.

And there’s the paradox: The medical community knows the diet works but doesn’t know how and can’t make it practical enough for most patients until they find out how it works.

A Century of Use — and Unanswered Questions

The ketogenic diet was the work of Russell Wilder, an endocrinologist at the Mayo Clinic who proposed in 1921 that the antiseizure benefits of fasting depended on ketonemia — the presence of ketone bodies in the blood. If a diet could keep the body in a fasting-like metabolic state indefinitely, he reasoned, it might control seizures without requiring that patients stop eating.

The approach worked well enough to become the primary epilepsy treatment for two decades. Then ASMs arrived, and the diet was gradually abandoned. By the late twentieth century, it was used by only a small number of children’s hospitals.

Its revival came through increased interest in the diet for children with severe DRE whose seizures did not respond to two or more ASMs. The pediatric evidence is now substantial. A Cochrane meta-analysis of randomized controlled trials suggested that children on ketogenic diets had roughly sixfold greater odds of achieving meaningful seizure reduction than those on standard care, though the authors cautioned that these estimates carried substantial uncertainty.

The adult evidence remains thin. No randomized controlled trial in the past 5 years has compared ketogenic diet therapies in adults. And even where the evidence is strong, there is disagreement over when the diet should be introduced. Currently, patients must typically fail several ASM trials before guidelines recommend ketogenic diet therapy.

Only in glucose transporter type 1 deficiency— a rare metabolic condition in which the brain’s glucose transporter is dysfunctional — is the ketogenic diet considered a first-line treatment. For every other form of epilepsy, it is a last resort.

By the time patients have failed multiple medications, their brains may have become more treatment-resistant, potentially limiting the diet’s effectiveness. “It makes sense that the earlier the person goes on the diet, the better the outcomes,” Manisha Patel, PhD, professor of pharmaceutical sciences at the University of Colorado Anschutz Medical Campus, Aurora, Colorado, told Medscape Medical News.

Patel recently published a review that synthesizes 5 years of clinical trials and basic science. The review outlined how the diet appears to work through multiple overlapping and potentially synergistic pathways that are unaffected by current ASMs — and whether understanding those pathways could eventually yield a medication that mimics the diet’s effects.

More Than Ketones

One thing most experts agree on is that the efficacy of the diet is about more than just ketones.

“It’s something about the change that your body goes into that helps seizures,” said Eric Kossoff, MD, a pediatric neurologist and director of the ketogenic diet program at Johns Hopkins University, Baltimore. Kossoff was not part of the new review.

“It’s more complicated than just adding ketones and experiencing less seizures,” he told Medscape Medical News.

Indeed, among the review’s key findings is that the diet’s antiseizure effects do not appear to depend on any single mechanism. Higher ketone levels do not necessarily mean fewer seizures. Instead, the evidence points to at least two broad shifts that may help clinicians explain how the diet works to patients.

The first involves the brain’s fuel source. During seizures, neurons become overly dependent on the rapid breakdown of glucose for energy — a metabolic pattern the review compared to the Warburg effect in cancer.

The ketogenic diet counters this by supplying ketone bodies that fuel the brain’s energy-producing machinery directly, bypassing glucose metabolism. In practical terms, the diet switches the brain off the dysfunctional pathway that seizures exploit and onto a more stable energy supply.

The second involves the balance between excitation and inhibition. Preclinical studies showed that ketogenic diets increased the ratio of gamma-aminobutyric acid (GABA) — an inhibitory neurotransmitter that dampens electrical activity — to glutamate, an excitatory one. The ketone body beta-hydroxybutyrate appeared to drive this shift by boosting the enzyme that converts glutamate into GABA. In effect, the diet helps put the brakes on.

“Antiseizure drugs target ion channels and excitation-inhibition balance, but they touch upon the surface,” Patel said. “What lies underneath is this entire metabolic spectrum.”

The Gut Connection

Among the most intriguing findings is emerging preclinical evidence that the gut microbiome may actively mediate the diet’s effects. The review cited studies showing that mice raised in sterile, bacteria-free environments did not benefit from ketogenic diets unless they were colonized with gut bacteria from mice whose seizures had responded to the diet.

A 2025 study included in the review found that seizure resistance in mice was driven primarily by dietary fiber rather than fat content alone — suggesting the microbiome’s role extends beyond ketosis itself. That link may involve short-chain fatty acids — metabolic byproducts of gut bacteria that appear to support neurotransmitter balance and blood-brain barrier integrity.

Patel cautioned against drawing clinical conclusions from animal data. “It’s really too premature to make any kind of conclusion,” she said. “We’re just fascinated with the fact that there is an effect at the preclinical level.”

If gut bacteria are actively mediating the diet’s antiseizure effects, anything that disrupts them — such as antibiotics or other common interventions — could undermine treatment.

This is also something Kossoff discusses with families in his practice. “We don’t know what effect antibiotics are going to have on someone on the diet. You may be doing harm.”

He added that some families respond by asking about probiotics, but the bacterial strains in commercial supplements are not the same ones that appear to matter. “You eat yogurt, it’s lactobacillus. That’s great. But it’s actually not the bacteria that flourish on the ketogenic diet.”

These are the kinds of queries clinicians hear regularly from patients, Kossoff said — questions for which the field still has no answer. While the latest review offers an important analysis of decades of evidence, Kossoff said the paper succeeded less as a mechanistic explanation of how the diet works and more as an honest analysis of how metabolically complex the answer is likely to be.

Keto Diet in a Pill?

Clinicians’ interest in understanding how the diet works goes beyond a need to explain its therapeutic benefits to patients. The review noted several strategies being explored to replicate aspects of the diet’s effects, including the development of compounds that target energy metabolism, antioxidant pathways, and neuroinflammation. The authors described a “primary goal of basic science research” as identifying therapeutic targets that could enhance the diet’s efficacy or lead to pharmacologic alternatives.

Kossoff was skeptical. “For 105 years, people have tried to figure out, ‘Can you make a pill?’” he said. “And the answer is no.” The diet’s efficacy arises from the interplay of its many mechanisms, he argued, not any single pathway. “You’re not going to find one compound that’s going to do everything that being on a dietary therapy does.”

However, he was open to the idea that reviews like this could open the door to “more therapeutic drugs that could mimic aspects of the diet.” If a patient is unable or unwilling to follow the diet, he said, a compound that replicates even part of its effects would be worth trying.

Patel pushed back. The adherence problem, she argued, is precisely the reason this research matters.

“One of the major barriers for neurologists in prescribing ketogenic diets is patient adherence, given the diet’s restrictive nature,” she said. “If we can identify the key biological drivers of its antiseizure effects, we may eventually design therapies that are more practical and acceptable to patients.”

Research suggests this may not be purely aspirational. A recently published study found that beta-hydroxybutyrate — the primary ketone body produced by the diet — suppresses seizures in mice by binding to a specific receptor called HCAR2 on hippocampal neurons. When the receptor was genetically deleted, beta-hydroxybutyrate lost its antiseizure effects. Niacin, a vitamin long used to treat cholesterol, activated the same receptor and produced similar seizure suppression.

Though the finding doesn’t prove a pill can replace the diet, it offers a specific molecular target — and an existing compound that can hit it.

Patel noted that researchers at University of California, Los Angeles; the University of Virginia; the University of California San Diego; and elsewhere are pursuing similar work. “If we could home in on what types of medications we could convert the diet into,” she said, “we might be able to really make a difference in people with epilepsy.”

When to Diet?

The review also called for clinicians to rethink when the diet is introduced. Currently, ketogenic diet therapy is typically offered only after patients have failed multiple ASMs. But the authors argued that this delay may allow the brain to become more treatment-resistant and called for trials examining the diet earlier in the treatment sequence.

Patel framed the stakes in terms of research priorities. If the field can execute large-scale comparative effectiveness trials and validate biomarkers for treatment response, “we may be able to actually bypass some of the waiting times that individuals have to be advised to go on the diet,” she said. “We also may come across some more tolerable dietary therapies.”

Compliance with the diet among adults may continue to prove difficult. “It’s much easier to give a bottle to a baby that has a different formula,” Patel said. “It is much harder for adults to go on this type of very strict diet.”

Funding constraints compound the problem. Dietary interventions lack the commercial incentive that drives pharmaceutical development — you cannot patent a diet. And epilepsy research overall has long been underfunded relative to disease burden, with dietary trials representing a small fraction of registered studies.

“There’s not as much funding or money that goes into these areas,” Patel said. “It really becomes an uphill battle.”

Kossoff offered a longer view.

“Ten years from now, a patient with epilepsy will come to our office. Based on their genetics, we can tell them with one hundred percent certainty that a ketogenic diet therapy will be helpful — tailored to their specific variant, with side-effect prevention that’s built in.”

But first, the field will have to crack the problem it has circled for 105 years — explaining why a therapy that clearly works remains so difficult to reduce to its parts.

The study was independently supported. Patel reported receiving funding from the National Institute of Neurological Disorders and Stroke and the Eunice Kennedy Shriver National Institute of Child Health and Human Development. Kossoff reported having no relevant financial relationships.