In a retrospective cohort of 90 infants with genetically confirmed developmental and epileptic encephalopathy (DEE) whose seizures began before two years of age, ketogenic diet therapy (KDT) achieved seizure freedom in one‑third of the children. Earlier start of the diet and variants in non‑ion‑channel genes were linked to better outcomes, while the lipid‑to‑non‑lipid ratio did not differ between seizure‑free and uncontrolled groups. Adverse events were common but mostly gastrointestinal and rarely led to discontinuation.
Study and findings
The authors retrospectively reviewed records of 117 children with DEE who began KDT between 2016 and 2023 at four Chinese pediatric hospitals; 90 met inclusion criteria. Median age at seizure onset was 2 months and the median interval to KDT start was 8 months. Genetic testing identified 30 causative genes, most frequently ion‑channel genes (54%). After a median follow‑up, 31 children (34%) became seizure‑free for at least six months (Grade 1), 19 (21%) achieved 75–<100% reduction (Grade 2), 13 (14%) 50–<75% reduction (Grade 3), and 27 (30%) <50% reduction (Grade 4). Overall, 50 children (56%) were classified as significantly effective (Grades 1 + 2). Children who attained seizure freedom had a shorter latency to diet initiation (median 6 months vs 10 months in the uncontrolled group) and were more likely to carry non‑ion‑channel gene variants (61% vs 37%). The maintenance lipid‑to‑non‑lipid ratio did not differ between seizure‑free and uncontrolled groups (2.2 ± 0.7 vs 2.5 ± 0.7). Adverse events occurred in 66% of participants, chiefly gastrointestinal symptoms; 14% discontinued KDT because of adverse reactions.
Clinical interpretation
These data suggest that KDT can produce clinically meaningful seizure reduction in a substantial proportion of infants with genetic DEE, especially when introduced early after seizure onset. The lack of association between higher lipid‑to‑non‑lipid ratios and seizure freedom indicates that maximal ketosis is not a prerequisite for efficacy in this population. The higher response among children with non‑ion‑channel gene variants (e.g., STXBP1) aligns with the notion that metabolic modulation may be more effective when the underlying pathophysiology is not directly related to channel dysfunction. Gastrointestinal discomfort is the most frequent side effect, but most children remain on therapy, implying that a standardized, gradual titration of the ketogenic regimen can mitigate intolerance.
Limitations and open questions
The retrospective design limits causal inference; unmeasured confounders such as concurrent antiepileptic drugs or variations in diet implementation could influence outcomes. The sample size for individual gene sub‑analyses is small, preventing firm conclusions about genotype‑specific responsiveness. Long‑term neurodevelopmental outcomes were not reported, leaving uncertainty about whether seizure control translates into functional benefit. Future prospective, controlled trials with standardized KDT protocols and systematic follow‑up of developmental milestones are needed to clarify optimal timing, lipid ratios, and genotype‑guided indications.
Source
Song TY, Deng J, Chen CH, Wang XH, Tian XJ, Wu ML, et al. [Effectiveness and tolerability of ketogenic diet therapy in children with genetic developmental and epileptic encephalopathy onset within 2 years of age]. Zhonghua Er Ke Za Zhi. 2026;64(10):1130-1136. doi: 10.3760/cma.j.cn112140-20260601-00420. PMID: 42742013.