Chamod Danosha Perera¹, Tehan Siriwardana¹, Kasun Lakmal², Dineshani Hettiarachchi³
1- Faculty of Medicine, University of Colombo, Sri Lanka
2- Post Graduate Institute of Medicine, University of Colombo, Sri Lanka
3- Department of Anatomy, Genetics and Bio-informatics, University of Colombo, Sri Lanka
Abstract Body
Background: Epileptic encephalopathies in childhood are severe disorders characterized by drug-resistant seizures and progressive neurodevelopmental impairment. Early nonpharmacological interventions are increasingly explored as adjunctive strategies to improve seizure outcomes. However, evidence on their effectiveness across different encephalopathy subtypes remains fragmented, warranting a systematic evaluation.
Objectives: To evaluate the impact of early non-pharmacological interventions on improving seizure control in children with epileptic encephalopathies.
Methods: We conducted a systematic review of non-pharmacological interventions for epileptic encephalopathies, up to November 26, 2025, using relevant MeSH terms. Data was manually extracted, and the risk of bias was assessed.
Results: 19 studies met the inclusion criteria. 3 out 7 cohort studies assessing vagal nerve stimulation (VNS) in Lennox– Gastaut syndrome (LGS); showed meaningful long-term benefit. Eight studies evaluated the effect of ketogenic diet (KD); Cohort data showed significant seizure reduction in LGS, Doose syndrome, and GLUT-1 deficiency and in infants with IESS. One crossover study found significant short-term improvement with KD. Among RCTs, short-term KD provided better early control. The modified Atkins diet (MAD) improved seizures in children aged 1–2 years in one RCT. A case series of cannabidiol showed no significant improvement.
Conclusion: Early non-pharmacological interventions showed mixed effects in epileptic encephalopathies. Overall, dietary therapies appear most effective. The KD demonstrated the most consistent effectiveness, particularly in GLUT-1 deficiency , Doose syndrome, and IESS. VNS produced variable outcomes, with some benefit in LGS. Evidence for MAD was inconsistent, and cannabidiol showed no meaningful benefit.