Long-term outcome prediction after pediatric cardiac arrest by continuous electroencephalography

Continuous electroencephalography (cEEG) features after cardiac arrest (CA) are associated with short-term outcome predictions. However, the association between cEEG features and long-term neurobehavioral outcomes in children remains unclear. This study aimed to identify cEEG features that can accurately predict neurobehavioral outcomes in children 6 months after CA.

A single-center, retrospective, observational study was performed at the tertiary care pediatric intensive care unit (PICU) of a national children's hospital. This study included a consecutive cohort of children admitted to the PICU after CA who underwent cEEG monitoring. The initial EEG background, electrographic seizures, and changes in EEG background over time were documented for each time point. The primary outcome was neurobehavioral outcomes at 6 months after resuscitation. To assess the reliability of the predictive model, the area under the receiver operating characteristic curve (AUROC) of models was compared for different combinations of EEG features.

The study included 62 infants and children. Unfavorable neurobehavioral outcomes were reported in 39 patients (63 %), including mortality in 20 patients. The combination of initial EEG background category and the EEG change from initial EEG initiation to 24 h, or variability/reactivity alone, were minimal optimal predictive models for unfavorable neurobehavioral outcomes at 6 months (AUROC, 0.99 [95 % confidence interval, 0.98-1.00], and 0.99 [95 % confidence interval 0.97-1.00], respectively).

Early cEEG features in the PICU can accurately predict neurobehavioral outcomes in children at 6 months after CA. These results can enhance prognostication and guide clinical decision-making in pediatric CA cases.