(748) Seizures and Neurologic Injury on ECMO in the Neonatal and Pediatric Populations

Introduction: Seizures have become a topic of interest among the pediatric ECMO community due to their relationship with increased morbidity and mortality. Despite the negative impact on patient outcomes, factors to predict the occurrence of seizure activity in this population remain elusive. Previous literature has identified abrupt changes in PaCO2 to be a risk factor for neurologic injury. We hypothesized that abrupt changes in PaCO2 are related to increased incidence of seizures in pediatric and neonatal ECMO patients.

Methods: A single center, retrospective cohort analysis of pediatric ECMO patients at our center between 01/01/2014 and 06/01/2023 who had continuous electroencephalography monitoring in the first 24 hours of ECMO cannulation, without a previous diagnosis of seizures or epilepsy, was conducted. The primary endpoint was incidence of electrographic seizures. Secondary endpoints were mortality and long-term neurologic outcome.

Results: 53 patients were included in the study, with 13% of patients demonstrating electrographic seizures (n=7), while only one patient with seizures also concurrently experienced an abrupt change in PaCO2 (P=1.00). Patients who experienced seizures while on ECMO support had a mortality rate of 43%, as compared to 37% in the non-seizure group (P=1.00). Of note, 68% of patients in the cohort were exposed to midazolam. Pediatric Cerebral Performance Category scales identified a worse neurologic outcome at discharge for the children who experienced seizures.

Conclusions: The seizure incidence for our cohort at 13% is slightly lower than previously reported (about 15-20%). While there was no significant relationship between abrupt changes in PaCO2 and electrographic seizure activity, there was a high rate of midazolam use in the entire cohort for sedation purposes. It can be hypothesized that this added secondary benefit as a prophylactic anti-epileptic, thus affecting our overall seizure incidence. The effect on overall neurologic outcome remains unclear. Further studies should work to better understand this relationship and if it may confer any benefit in terms of outcome.