Risk factors for electrographic seizures in infants after cardiac surgery and their association with outcomes: a retrospective cohort study

Neonates and infants undergoing cardiac surgery are prone to cerebral injuries. Neuromonitoring with continuous conventional electroencephalography (cEEG) and/or amplitude integrated electroencephalography (aEEG) is able to detect subclinical abnormal cerebral patterns, especially seizures, which have been found to correlate with medium-long term neurocognitive deficits. The endpoints of this study are (i) to define a risk model for electrographic subclinical epileptic seizures (ES) and (ii) to investigate the relationship between ES and neurologic and non-neurologic outcomes. Retrospective analysis of 373 neonates and infants < year undergone cardiac surgery and receiving a pre/postoperative combined cEEG/aEEG exam. The risk factors for postoperative ES were investigated and pooled in a predictive model, and postoperative ES were assessed for association with neurologic and non-neurologic (major morbidity) postoperative patterns. The independent risk factors for postoperative ES were the complexity of surgery (RACHS-2 score), an open sternum, and the hematocrit at the arrival in the intensive care unit. The predictive model based on these factors yielded a good discrimination (c-statistics 0.81). Postoperative ES were significantly associated with major non-neurologic morbidity, but once corrected for other confounding factors they lost significance. Conversely, postoperative ES were the only predictor of cerebral injuries, with an odds ratio of 8.0 (95% confidence interval 2.25–28.3, P = 0.003). Postoperative ES are associated with an hemodynamic compromise and low arterial oxygen content, with a consequent inadequate oxygen delivery to the brain and other organs. In presence of these factors, EEG monitoring is useful for the early diagnosis of cerebral injuries in patients sedated and paralyzed.