Keto for Migraine: What the Best Trial Actually Found

A very-low-calorie ketogenic diet reduced migraine days more than an equally calorie-restricted balanced diet in overweight adults with frequent episodic migraine. That result matters because both groups restricted calories. The ketogenic group still did better, suggesting that ketosis itself—not weight loss alone—may contribute to the effect.

There is a catch: this conclusion rests mainly on one single-center randomized controlled trial in a specific population. Keto is a reasonable clinical option for some people with migraine, but it is not a universal treatment, a substitute for prescribed prevention, or a project for unsupervised dietary extremism.

The trial that puts keto on the map

The EMIKETO randomized controlled trial enrolled adults with high-frequency episodic migraine who had a BMI of at least 27. It compared an eight-week very-low-calorie ketogenic diet (VLCKD) with a calorie-matched hypocaloric balanced diet. The primary practical outcome was monthly migraine days.

During active ketosis at week 8, the VLCKD group had a significantly greater reduction in monthly migraine days than the balanced-diet group. The study then reintroduced carbohydrates between weeks 8 and 12. After that, both groups followed the same maintenance diet through week 24. The migraine advantage in the original ketogenic group persisted at weeks 12 and 24, after ketosis had ended.

That follow-up design makes the study more interesting than a simple before-and-after diet report. If calorie restriction were the whole explanation, two calorie-matched diets should have produced similar results. They did not. Still, persistence of benefit does not prove a lasting brain reset. The trial measured clinical outcomes, not the causal pathway responsible for them.

EMIKETO also reported lower C-reactive protein, neutrophil-to-lymphocyte ratio, and white blood cell measures alongside the ketogenic intervention. Those are inflammatory markers, not proof that reduced inflammation caused fewer migraines. Weight loss and migraine improvement could have occurred through overlapping or separate pathways.

What might ketosis change?

Several mechanisms fit what migraine biology already suggests, though none has been confirmed as the mediator of the trial result. Beta-hydroxybutyrate, a ketone body, may increase the availability of material used to make GABA, an inhibitory neurotransmitter. Greater inhibitory signaling could raise the threshold for cortical spreading depression, the electrical event linked to migraine aura and trigeminovascular activation.

Ketone bodies may also suppress the NLRP3 inflammasome, reduce mitochondrial reactive oxygen species, and influence pathways involving calcitonin gene-related peptide (CGRP). Excess adiposity may raise CGRP-related migraine susceptibility, so changes in adiposity could matter as well. The 2023 systematic review and meta-analysis by Neri et al. describes these hypotheses and found a statistically significant pooled reduction in migraine attack frequency across ketogenic-diet studies. Biological plausibility is useful. It is not the same as measuring a mechanism in the people who improved.

The rest of the evidence supports the signal, with limits

The meta-analysis helps, but it does not turn the evidence into certainty. Its included studies were often small, non-randomized, or pediatric, and adult randomized evidence was sparse before EMIKETO. Small uncontrolled studies can overestimate benefits because people who tolerate a demanding diet, stay in follow-up, and expect improvement may differ from those who do not.

A 2023 retrospective study by Tereshko et al. adds a useful, but weaker, comparison. In 61 patients—45 with chronic migraine and 16 with episodic migraine—a 2:1 ketogenic diet was compared with a low-glycemic-index diet in a single clinical center. Monthly migraine days fell by about 6.7 days with keto versus 2.1 days with the low-glycemic diet among patients with chronic migraine. In episodic migraine, the corresponding reductions were about 4.2 and 1.8 days.

That gap is consistent with an effect beyond generic glycemic restriction. Because the study was retrospective and involved a selected clinical population, however, it cannot establish that ketosis caused the difference. Differences in adherence, baseline severity, clinical decisions, or other unmeasured factors could contribute.

A broader 2025 systematic review of dietary interventions for migraine places keto ahead of Mediterranean, low-fat, and elimination diets in controlled evidence for the overweight migraine phenotype. That says as much about the thin evidence for other diets as it does about the maturity of keto research.

How does keto compare with migraine drugs?

Across separate controlled trials, established preventive medicines have produced approximate reductions of 2 to 3 monthly migraine days for topiramate, 1.5 to 2.5 days for propranolol, and 1.4 to 3.7 days versus placebo for CGRP monoclonal antibodies such as erenumab and fremanezumab. About 44% to 52% of people taking topiramate achieved at least a 50% response in the cited trial context.

EMIKETO's magnitude is competitive enough to take seriously, but cross-trial comparisons are not head-to-head evidence. The participants, baseline migraine burden, outcome windows, and study methods differ. No randomized trial has directly tested VLCKD against topiramate, propranolol, or a CGRP antibody. A diet and a prescription drug also bring different burdens, risks, costs, and adherence problems.

The metabolic cost is part of the treatment

People sometimes discuss keto for migraine as if migraine days were the only outcome that counts. They are not. A 2025 systematic review and meta-analysis by Turetta et al., conducted in women with PCOS rather than people with migraine, found a consistent LDL-cholesterol elevation across the VLCKD literature. It also documented deficiencies involving selenium, magnesium, and zinc.

This does not refute the migraine finding, and PCOS results cannot be assumed to quantify the same risk in every migraine patient. It does establish a relevant safety signal for the overweight population studied in EMIKETO. Higher LDL can matter greatly for someone with elevated cardiovascular risk. Magnesium matters particularly in migraine care because deficiency can trigger attacks. A restrictive diet that leaves someone magnesium-deficient can work against its own goal.

Medical supervision and planned micronutrient supplementation are part of a responsible VLCKD protocol, especially when migraine prevention is the reason for using it.

Who can reasonably consider it?

The evidence applies to adults with high-frequency episodic migraine and BMI of at least 27, not to all people with migraine. No controlled trial has established efficacy in normal-weight migraineurs. EMIKETO also came from one center, and published reporting does not fully clarify dropout rates or permit a clean comparison of intention-to-treat and per-protocol results. Independent replication would make the estimate more trustworthy.

Its 24-week follow-up offers encouraging evidence that an eight-week ketogenic phase can have benefits that outlast ketosis. It does not answer whether the approach remains effective, safe, or sustainable over years in ordinary practice. Many people eventually return to carbohydrate-containing diets, and real-world adherence to the full supervised protocol has not been established.

The practical verdict

For an overweight or obese person with frequent episodic migraine, an eight-week medically supervised VLCKD is an evidence-grounded option worth discussing with a clinician. It beat a calorie-matched balanced diet in the best available randomized trial, and other studies point in the same direction. The benefit may involve ketones, weight-related biology, or both; the trial cannot separate those explanations completely.

For everyone else, the evidence is thinner than the enthusiasm. Keep established preventive treatment on the table, review cardiovascular risk and medications, monitor LDL, and plan magnesium and other micronutrients before attempting VLCKD. The most defensible claim is specific: keto has a credible role in a defined metabolic migraine phenotype under clinical supervision.