A systematic review of non-invasive biomarkers for seizure forecasting in pediatric epilepsy patients

Introduction: Epilepsy is one of the most common neurological disorders globally. While medications, surgical interventions, and dietary changes can be successful in controlling seizures, a subset of individuals experience refractory epilepsy and are at increased risk for sudden unexpected death in epilepsy (SUDEP). Efforts to provide a detection system using devices have been successful at identifying seizures once they start, but there are no devices or systems on the market that predict seizures. The purpose of this systematic review (PROSPERO ID: CRD42024444250) is to determine non-invasive physiologic and environmental biomarkers that can be used to forecast seizures in pediatric epilepsy patients. Methods: A systematic search of relevant literature was conducted in PubMed, Web of Science, CINAHL Ultimate, and EMBASE in August 2023. Articles were reviewed by two investigators in a two-step process. Data extraction occurred using two independent extractors to identify study characteristics, patient characteristics, and forecasting results. Evidence quality was evaluated by two investigators using the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) tool. Results: Eleven observational cohort studies were included and cardiovascular biomarkers using electrocardiogram (ECG) measurement were most commonly used. Pre-ictal anticipation algorithm times ranged from 21.8 s to 32 min, while correlational studies observed cardiovascular biomarker changes 3.59 s to 40 min before seizures. This systematic review provides a comprehensive overview of the current evidence for seizure forecasting. However, the evidence in 9/11 of reviewed studies were rated as either low or very low certainty using the GRADE tool due to methodological flaws, risk of bias, inconsistent results, and indirect or sparse evidence. Discussion: There are ongoing opportunities to build on our findings, including further testing of cardiovascular biomarkers with other physiologic and environmental factors, larger sample size studies, and a precision medicine approach to tailoring algorithms and biomarker measurements to individual patients. Systematic review registration: PROSPERO (ID: CRD42024444250).