T87. EEG and simultaneously recorded SEPs in evaluation of newborns with hypoxic ischemic encephalopathy or stroke in the NICU

Introduction Differential diagnostics and outcome prediction in newborns presenting with seizures or encephalopathy remain difficult in the bedside. The objectives of the present study were (i) to determine the accuracy of simultaneously recorded EEG and somatosensory evoked potentials (SEPs) in differentiating between hypoxic-ischemic encephalopathy (HIE) and stroke as the underlying etiology of neonatal seizures or encephalopathy and (ii) to determine the accuracy of EEG and SEPs and their combination for predicting outcome in newborns with HIE or stroke. Methods The study comprised of 133 newborns (gestational age > 34 weeks) that underwent a clinically indicated EEG and simultaneous SEPs in the acute phase of their illness. 102 of the newborns had HIE and 31 had stroke. EEG background was scored as continuous (sleep stages identifiable and EEG continuous at least in active sleep or awake state) or discontinuous (sleep stages not identifiable and EEG not continuous at any time during the recording). The SEPs were classified as bilaterally absent, unilaterally absent, or bilaterally present. Outcome was evaluated from medical records at approximately 1–1.5 years age. Death, cerebral palsy, severe mental retardation, and epilepsy were considered as unfavorable outcomes. Normal or mildly abnormal development (e.g. mild motor or language delay) were considered as favorable outcomes. Results The combinations of EEG and SEP abnormalities were indicative of etiology (HIE vs stroke) and predictive of outcome. If the EEG was discontinuous, the etiology was always HIE (33 newborns) regardless of SEPs. Furthermore, if the EEG was continuous, but SEPs were uni- (9 newborns) or bilaterally (1 newborn) absent, the etiology was always stroke. Normal SEPs and continuous EEG occurred in 69 newborns with HIE and 21 newborns with stroke, and hence in these newborns the distinction between the etiologies could not be made with the present EEG-SEP criteria. This combination was, however, highly predictive of a favorable outcome regardless of the etiology: The outcome was favorable in 67/69 of those HIE and 20/21 of those with stroke. On the contrary, bi- or unilaterally absent SEPs were associated with poor outcome in all newborns with HIE (22/22) and all but two newborns with stroke (8/10). Furthermore, in newborns with HIE, even if the EEG was discontinuous but SEPs were bilaterally present, the outcome was likely to be favorable (8/11 newborns). Conclusion EEG-SEP is useful in bedside differential diagnostics between HIE and stroke in newborns presenting with seizures or general signs of encephalopathy. Furthermore, EEG-SEP is accurate in outcome prediction in both etiologies.