Abstracts

Rationale: Describe clinical and electroencephalographic (EEG) characteristics of neonates who underwent continuous EEG monitoring (cEEG) in the neonatal intensive care unit (ICU), and to determine clinical and EEG biomarkers of outcome. Methods: Retrospective descriptive study of patients < 1 month of age who underwent a cEEG (>3 hours) in the surgical, cardiac, and neonatal ICU at a tertiary care pediatric site from 2011-2013. Patients were excluded if the cEEG was performed as part of an elective admission. Results: 211 patients (116 males) met the inclusion criteria. 158 (75%) were term (defined as at least 37 weeks of gestational age) and had a median gestational age of 38.7 weeks (range 24.9 - 42.1). Monitoring began at a median age of 4 days (Table 1). The main indications for cEEG were event characterization in 191patients (91%), and encephalopathy/concern for nonconvulsive seizures in 22 (8%). The etiology of the episodes leading to monitoring was symptomatic of a structural lesion in 137 patients (65%). Of those, 125 were acute symptomatic and 12 were chronic symptomatic (Table 2). Most patients had an abnormal EEG background with increased discontinuity in 91 patients and slow/disorganized features in 94 patients being the most common. Seventy five (35%) patients had electrographic seizures. Seizures were focal in 44 patients, multifocal in 30 patients, and generalized in 1 patient. Most seizures lasted less than 5 minutes. Electrographic status epilepticus, defined as seizures lasting greater than 30 minutes or seizures occurring in over 50% of a 1 hour epoch occurred in 8 (12%) patients. Of the 8 patients with status epilepticus, 6 (75%) had no clinical correlate while 2 (25%) had some degree of a clinical correlate. 9 patients (12%) met criteria for burst suppression. 30 of 75 (40%) patients had electrographic only seizures without any clinical correlate, while the remainder had both electroclinical and electrographic only seizures. One patient only had all electroclinical events. 30 patients (14%) died before leaving the hospital. In a logistic regression model with in-hospital death as the outcome and controlling for gestational age and structural/non-structural etiology as potential confounders, the factors that influenced outcome were status epilepticus (OR 9.6; 95%CI 2.1-43.2) and lack of change in sleep-wakefulness states (OR3.2; 95%CI 1.3-8). In our cohort, the following factors did not influence outcome: electrographic seizures (OR 1.5; 95%CI: 0.6-3.5), EEG continuity (OR 0.6; 95%CI 0.3-1.4), and interictal epileptiform discharges OR 0.9 (95%CI 0.4-2). Conclusions: In our cohort, cEEG revealed subclinical seizures in 14% of neonates that did not occur in conjunction with clinical seizures, and subclinical status epilepticus in 2.8%, thereby influencing clinical management. Furthermore, status epilepticus and lack of change in sleep-wakefulness states on cEEG are important prognostic features regarding in-hospital mortality. Future interventional studies will need to determine longitudinal modifiability of these biomarkers of outcome.