Abstracts

Rationale: Electrographic seizures are common in encephalopathic critically ill children, even without a pre-existing diagnosis of epilepsy, but identification requires resource intense continuous EEG monitoring (CEEG).  Development of a seizure prediction model would enable more efficient use of limited CEEG resources.  We aimed to develop and validate a seizure prediction model. Methods: We developed a seizure prediction model using a retrospective multicenter database of children with acute encephalopathy without an epilepsy diagnosis, who underwent clinically indicated CEEG.  We performed model validation using a separate single center prospective database.  Predictor variables were chosen to be readily available to clinicians prior to the onset of CEEG and included: age, etiology category, clinical seizures prior to CEEG, initial EEG background category, and inter-ictal discharge category. Results: The model has fair to good discrimination ability and overall performance.  At the optimal cut-off point in the validation dataset, the model has a sensitivity of 59% and specificity of 81%.  Varied cut-off points are described which could be chosen to optimize sensitivity or specificity depending on available CEEG resources. Conclusions: A model developed from multi-center CEEG data can guide the use of limited EEG resources when applied at a single center.  Depending on CEEG resources, centers could choose lower cut-off points to maximize identification of all patients who will experience electrographic seizures (but with lowered CEEG resource efficiency by monitoring more patients) or higher cut-off points to reduce resource utilization by reducing monitoring of lower risk patients (but sacrifice identification of some patients experiencing seizures).