Abstracts

Rationale: Hypoxic-ischemic encephalopathy (HIE) is a major cause of neurological morbidity in neonates, including cerebral palsy, mental retardation, and epilepsy. Clinical trials have shown that therapeutic hypothermia (TH) improves motor outcomes in children with HIE, but separate epilepsy outcomes are not well characterized after TH. We aim to identify clinical, radiographic, and 72-hour EEG features that predict development of postneonatal seizures by age 2 in children with HIE who received selective head cooling,. We developed a model based on these features to predict which patients are at risk for development of epilepsy, to better inform prospective follow-up and surveillance of patients at risk. Methods: Chart review was performed on neonates who underwent selective head cooling at a tertiary care center neonatal ICU. Subjects were monitored on amplitude EEG at screening (0-6 hours of life) and during cooling. LTM-video EEG with neonatal montage was initiated at 72 hours, post-cooling, for a minimum of 24 hours and up to 10 days of life. MRI was performed at 7-10 days of life. EEGs were re-reviewed for background features and epileptiform activity by two pediatric epileptologists. MRIs were re-reviewed by a blinded pediatric neuroradiologist to identify regions of injury. A multivariable logistic regression model was built to identify significant independent risk factors, using variable selection with the "best subsets" method. The probability of developing epilepsy was calculated based on this model. Results: 49 neonates had 2 year follow up. 47/49 (95%) had seizures in the neonatal period, initially identified clinically or on aEEG, or on cEEG after 72 hours. 9 (18%) developed epilepsy outside of the neonatal period, 3 of which developed epileptic spasms. In multivariable analysis using "best subsets" approach, significant predictive factors were pH ≤6.8 on Day of Life 1 (p = 0.02, OR 21.4 [1.5,312]) cEEG at 72-96h with burst suppression pattern (p=0.03, 13.8 [1.3,151]) MRI with evidence of basal ganglia injury (p = 0.25, OR 4 [0.4,42.4]) Predictive modeling found a low-risk (0-1 risk factors, probability of epilepsy 3%), medium (2 risk factors, probability of epilepsy 30%), and high-risk group (3 risk factors, probability of epilepsy 80%) for the development of epilepsy. One baby who developed post neonatal epilepsy was low risk, five babies were medium risk, and three babies were high risk based on this model. Conclusions: Based on ongoing work in development of a predictive model, neonates with HIE who undergo TH are at greatest risk of future epilepsy if they have 2-3 identified significant risk factors: (1) significant acidemia in the first 6 hours of life, (2) burst suppression pattern present on cEEG at 72-96 hours of life, and evidence of DWI or T1 basal ganglia injury on MRI. The same risk factors are associated with other important sequelae of HIE, such as cerebral palsy. The presence of these identified risk factors, stratified into low, medium and high-risk groups, can inform clinician follow up for the risk of development of epilepsy.