Development and validation of a predictive model for seizure recurrence following discontinuation of antiseizure medication in children with epilepsy: a systematic review and meta-analysis, and prospective cohort study.

Abstract

Seizure relapse in pediatric patients with epilepsy after antiseizure medication (ASM) withdrawal is a critical concern, yet the risk factors are not fully understood. Identifying these factors is essential for personalized treatment planning. In this systematic review and meta-analysis, and prospective cohort study, we conducted a meta-analysis of cohort studies to derive a predictive model for seizure recurrence post-ASM discontinuation, then validated it in a prospective cohort study. The derivation cohort was derived from a systematic, search of PubMed, Web of Science, Embase, and Cochrane Library (from inception to May 1,2024) for English-language cohort studies on risk factors for seizure recurrence after ASM withdrawal in pediatric epilepsy, focusing on children initiating ASM tapering with documented relapse, while excluding case reports, and non-pharmacological interventions. Risk factors were selected and weighted according to the statistical significance of pooled relative risks (RRs), with β coefficients derived from log-transformed RRs to establish weighted scores in the predictive model. The validation cohort included children with epilepsy enrolled between February 16, 2015 and November 15, 2024, from two Chinese hospitals. Inclusion criteria comprised first-time ASM withdrawal candidates aged <18 years with ≥24-month follow-up, while exclusion criteria focused on incomplete data, protocol deviations, and non-pharmacological interventions. This study is registered at https://www.medicalresearch.org.cn/ (MR-50-24-042059). A total of 26 cohort studies were identified from the systematic review and included in the meta-analysis. The derivation cohort included 4080 children with epilepsy, of whom 959 (23.50%) experienced seizure recurrence. The predictive model identified nine significant risk factors: intellectual disability, abnormal neurological examination or motor deficit, history of febrile seizures, only focal onset seizures, overall number of ASM used, duration of epilepsy ≥3 years, abnormal electroencephalogram (EEG) at the start of ASM tapering, abnormal EEG after ASM tapering, and age at first seizure ≥10 years. β coefficients were derived from the logarithm of pooled relative risks for each factor and converted into weighted scores, yielding a maximum total risk score of 17. The validation cohort comprised 341 patients with a median follow-up duration of 2.84 (0.27-9.75) years, and 122 (35.8%) out of them had seizure relapses. The model demonstrated robust performance in the validation cohort, with an AUC of 0.85 (95% CI: 0.81-0.91), sensitivity of 0.74 (95% CI: 0.68-0.80), and specificity of 0.82 (95% CI: 0.75-0.89). Our evidence-based predictive model offers a robust tool for estimating the risk of seizure recurrence in pediatric patients with epilepsy after ASM withdrawal, aiding clinicians in personalized treatment decisions. While this tool enhances personalized treatment decisions in epilepsy management, its predictive thresholds require external validation across diverse clinical settings and populations to ensure broad clinical applicability. Chongqing Medical University (CQMU) Program for Youth Innovation in Future Medicine.