Abstract
Objective: The aim of the study was to identify the predominant predictors of seizure relapse following discontinuation of ASM in epileptic children.Methods: The study cohort consisted of 403 epileptic children who had a withdrawal process of ASM (monotherapy: 344; dual therapy or polytherapy: 59) after at least a 2‐year seizure‐free period. Patients were categorized if they had a well‐defined epileptic syndrome. Epileptic children with ongoing ketogenic diet, vagal nerve stimulation, or surgery were excluded from the cohort due to the additional withdrawal process related to other therapy modalities.Results: The cohort's seizure relapse rate was 12.7% (51/403). The highest rates of seizure relapse were defined for genetic etiology at 25% and structural etiology at 14.9%. An epilepsy syndrome was defined in 183 of 403 children (45.4%). There was no difference in the seizure relapse rate between the subgroups of well‐defined epileptic syndromes; 13.8% for self‐limited focal epileptic syndromes, 11.7% for developmental and epileptic encephalopathies, and 7.1% for generalized epileptic syndromes. Five predictors were defined as the most powerful predictors of seizure relapse in univariate analysis: age at epilepsy diagnosis >2 years (hazard ratio [HR]: 1.480; 95% confidence interval [CI]: 1.134–1.933), defined etiology (HR: 1.304; 95% CI: 1.003–1.696), focal seizure (HR: 1.499; 95% CI: 1.209–1.859), ≤3 months duration of the withdrawal process (HR: 1.654; 95% CI: 1.322–2.070), and a history of neonatal encephalopathy with or without seizures (HR: 3.140; 95% CI: 2.393–4.122). In multivariate analysis, the main predictor of seizure relapse was a history of neonatal encephalopathy with or without seizures (HR: 2.823; 95% CI: 2.067–3.854).Significance: The duration of seizure freedom before discontinuation of ASM was not a predominant risk factor for seizure relapse: 2–3 years versus >3 years. The predictive values of five predictors of seizure relapse rate should be evaluated for patients with different epilepsy subgroups.

This publication has 25 references indexed in Scilit: