Objective To assess the predictive value of amplitude-integrated electroencephalography EEG (aEEG) and near-infrared spectroscopy (NIRS) during therapeutic hypothermia.
Patients and Methods We studied 39 cooled, asphyxiated infants. We assessed aEEG and calculated mean regional cerebral oxygen saturation (rcSO (2) ) during and after treatment. At 30 months, we performed a neurological examination and administered the Bayley Scales of Infant and Toddler Development, 3rd edition. We calculated the odds ratios (ORs) of abnormal aEEG and rcSO (2) for severely abnormal outcome.
Results At 6 and 12 hours, severely abnormal aEEGs predicted severely abnormal outcomes (OR, 7.7 [95% confidence interval, CI, 1.39-42.6] and 24.4 [95% CI 4.2-143] respectively), as did epileptic activity (OR 28.9, 4.6-183). During the first 48 hours, rcSO (2) was not associated with outcome, but at 72 hours after birth and after rewarming it was, with ORs for severely abnormal outcomes of 12.8 (1.31-124) and 21.6 (1.05-189), respectively. In multivariate analyses, aEEG and rcSO (2) remained independently predictive in the model at 48 hours and significantly from 72 hours after birth onward.
Conclusion aEEG was a strong predictor of adverse outcome. After 48 hours of cooling, a higher rcSO (2) was associated with a severely abnormal outcome, adding to the predictive value of aEEG in cooled, asphyxiated infants.