Early prediction of outcome with aEEG in preterm infants with large intraventricular Hemorrhages

Background: The electrocortical background contains prognostic information in full-term asphyxiated newborn infants already during the first postnatal hours. In preterm infants with intraventricular hemorrhages (IVH) the background activity in EEG and amplitude-integrated EEG (aEEG) is depressed during the first days of life, and the extent of the depression correlates with the degree of IVH. However, it has not been previously evaluated whether very early aEEG can predict later outcome also in preterm infants. Objective: To investigate if early prediction of outcome is possible from aEEG in preterm infants with large IVH. Methods: aEEG recordings from the first postnatal week were investigated in 64 preterm infants with IVH grade III - IV. For every 24-hour period the aEEG background pattern was classified, and the maximum and minimum numbers of bursts/h, respectively, were counted. Outcome was divided into three categories: died (n = 36), survived (n = 28) with "poor" outcome, i.e., severe cerebral palsy and not able to walk and/or mental retardation (n = 8), and survived with "fair" outcome, i.e., healthy or mild cerebral palsy (n = 19). One surviving child was lost in the follow-up. Results: There were significant differences in maximum bursts/h (MaxB) at 0 - 24 hours (p = 0.033), 24-48 hours (p = 0.011), 48 72 hours (p = 0.049) and 72 - 96 hours (p = 0.032), respectively, between the infants who died and the surviving infants. At 2448 hours the median (range) MaxB in the surviving infants with "fair" outcome was 156 (103 - 179) versus 102 (73 - 156) in the surviving infants with "poor" outcome (p = 0.002). With the assumption that MaxB < 130 was predictive of death or survival with "poor" outcome, 68% and 78% of infants were correctly predicted at 0 - 24 hours and 24 - 48 hours, respectively. Conclusions: This study shows that outcome may be predicted with aEEG already during the first days of life in preterm infants with large IVH. The findings should be confirmed in prospective studies since they may have clinical implications if specific medical interventions become available.