Published by Unseen Progress, an independent publisher of caregiver research. Last reviewed 2026-05-10. Part of the epilepsy caregiver research overview.
Short answer. Yes, for a meaningful share of patients with drug-resistant epilepsy. The randomised controlled trial evidence (Neal et al., 2008) and large cohort studies show that roughly 30–50% of children on the classic ketogenic diet achieve at least 50% seizure reduction at 3 months, and 10–15% become seizure-free. The diet is most effective in certain childhood syndromes (Dravet, Lennox-Gastaut, glucose transporter deficiency, infantile spasms), and the modified Atkins protocol (Kossoff et al., 2006) makes a less restrictive version accessible for older children, adolescents, and adults.
The Neal trial (Neal et al., 2008, Lancet Neurology) was the first properly randomised controlled trial of the ketogenic diet for childhood drug-resistant epilepsy. At 3 months, 38% of children on the diet had at least 50% reduction in seizures, compared to 6% in the control group; 7% had at least 90% reduction, compared to 0% in controls. The effect size was substantial enough that the diet is now standard-of-care to consider in drug-resistant childhood epilepsy after two failed medications, alongside surgical evaluation.
Subsequent cohort and meta-analytic data (Kossoff et al., 2009 international consensus update) have replicated these findings in larger samples and over longer follow-up. The headline numbers across the literature converge on:
The diet is not a panacea. About half of children do not respond, and the diet is demanding enough that drop-out rates are non-trivial. But the response rate is comparable to or better than most third-line ASMs in the same population.
The literature consistently identifies several syndromes and seizure patterns where the ketogenic diet is particularly effective:
For focal epilepsies that are surgical candidates, surgery typically takes precedence over diet. For generalised epilepsies in the syndromes above, the diet is a credible parallel option to additional ASMs.
The classic ketogenic diet is medically supervised, weighed, and ratio-controlled (typically 4:1 fat-to-(carbohydrate+protein) by weight), and is genuinely demanding for families. The modified Atkins diet (MAD), developed by Kossoff and colleagues at Johns Hopkins (Kossoff et al., 2006), is a less restrictive version: carbohydrates are limited to ~10–20g per day, fats are encouraged, protein is unrestricted, and meals are not weighed. MAD is roughly comparable in efficacy to the classic ketogenic diet for many indications, and substantially easier to sustain — particularly for adolescents and adults.
The choice between classic ketogenic diet, MAD, and the low glycemic index treatment (LGIT) is now part of the dietary-therapy decision. The international consensus paper (Kossoff et al., 2009; updated 2018) treats them as a family of related interventions matched to patient age, family capacity, and syndrome.
The diet is demanding. Every meal is calculated. Eating outside the home requires planning. Birthday parties, school lunches, and travel become measurable logistical events. Families typically describe the first 1–3 months as the hardest, with the load decreasing as the routine stabilises. The effort cost is real and is one of the main reasons for discontinuation alongside non-response.
The benefit, when it is present, can be transformative. A child whose seizures had not responded to three ASMs and now has 80% fewer events on the diet, with the freedom and cognitive presence that comes with that, is the case the literature is built around. The case where the diet is tried, doesn't work in 3 months, and is discontinued is also well-documented and not a failure of the family — it is the response distribution at work.
Most of the trial data is paediatric, but observational evidence in adults (particularly in drug-resistant focal epilepsy and in genetic generalised epilepsies) suggests roughly comparable response rates, with adherence being the main differentiator. MAD and LGIT are typically the protocols offered to adults rather than the classic diet. For adults considering the diet, the conversation is the same as for children: indication, expected response rate, 3-month evaluation, and adherence capacity.
1. Raise the diet as an option after two failed ASM trials, alongside comprehensive epilepsy center referral. Both are standard-of-care; neither rules out the other. 2. Identify whether the syndrome is one of the high-response indications (GLUT1-DS, Dravet, Lennox-Gastaut, infantile spasms, tuberous sclerosis) — if so, the diet should be discussed early. 3. Insist on a registered dietitian with epilepsy experience as part of any diet trial. Self-implementation is not supported by the literature. 4. Pre-define the 3-month evaluation criteria — what would count as enough response to continue, what would justify stopping. Run the same structured seizure diary throughout. 5. If the diet is discontinued, do so via a planned wean, not abrupt stopping.
---
Unseen Progress publishes long-form caregiver research and builds research-backed daily trackers for the families covered. See the full epilepsy caregiver research overview for the complete framework.