Stymulacja neuronalna wrażliwa na napady w pediatrycznej epilepsji opornej na leki: przegląd systematyczny i metaanaliza wyników napadów, bezpieczeństwa i czynników predykcyjnych odpowiedzi na leczenie

OBJECTIVE: To provide a comprehensive quantitative synthesis of seizure outcomes, adverse events, non-seizure outcomes, and predictors of response in pediatric patients with drug-resistant epilepsy (DRE) treated with the responsive neurostimulation (RNS) system. METHODS: A systematic search of PubMed, EMBASE, and OVID/MEDLINE (from inception through March 2026), conducted in accordance with PRISMA guidelines, identified studies reporting RNS outcomes in patients ≤18 years. Primary outcome was seizure responder rate (≥50% reduction). Secondary outcomes included seizure freedom (100% reduction), super-responder rate (≥90% reduction), adverse events, and non-seizure outcomes. Pre-specified meta-regression examined mean age, follow-up duration, proportion with generalized epilepsy, and proportion with thalamic lead placement as predictors of seizure response. Risk of bias, certainty of evidence, and publication bias were assessed. RESULTS: Thirty-nine studies (Total n = 413 patients; median n= 7; median follow-up 18.7 months) were included, predominantly retrospective, with one prospective study. The pooled seizure responder rate (≥50% reduction) was 64.8% (95% CI: 59.7%-69.5%; I² = 0%). The pooled seizure freedom rate was 2.9% (95% CI: 1.0%-8.0%) and the super-responder rate (≥90% reduction) was 28.4% (95% CI: 20.5%-37.9%). Longer follow-up duration was the only significant predictor of seizure freedom on meta-regression (β = 0.046, p = 0.009); mean age, epilepsy type, and thalamic lead placement did not significantly predict seizure freedom, and no predictors were identified for seizure responder or super-responder outcomes. The pooled adverse event rate was 15.8% (95% CI: 10.9%-22.4%); device-related infection (∼4%) was the most common. Non-seizure outcomes - including quality of life, cognition, and behavior - were generally stable or improved across 12 reporting studies. Publication bias was noted (Egger's p = 0.029). SIGNIFICANCE: In this most comprehensive review to date, RNS provides meaningful seizure reduction in ∼two-thirds of pediatric DRE patients and a manageable safety profile, supporting selective off-label use while highlighting the need for prospective, multicenter, long-term data.