Electrographic seizures in neonates correlate with poor neurodevelopmental outcome

OBJECTIVE: To quantify the number, duration, and intensity of electrographic seizures (ESz) in neonates and to compare the outcome of neonates with ESz with those who were at risk but did not have ESz recorded. METHODS: The EEG and outcome data were reviewed from 68 infants who met at-risk criteria for neonatal seizures and underwent prolonged continuous EEG monitoring. Forty infants had ESz. The control group contained 28 infants monitored for at least 18 hours and found not to have ESz. Outcomes for both groups were evaluated using hospital and follow-up clinic records and a standardized telephone interview. RESULTS: The etiology of ESz included asphyxia (n = 23), stroke (n = 7), and other (n = 10, intraparenchymal, subdural, and subarachnoid bleeding; meningitis; sepsis; hyponatremia; and unknown). The cumulative recorded ESz duration was 8 minutes to 30 hours. Forty-three percent of infants with ESz spent 38 minutes to 32 hours in electrographic status. Despite doses of 40 mg/kg of phenobarbital and 20 mg/kg of phenytoin, 30% of infants continued to have ESz. Ten infants with ESz and one without died from causes related to neurologic instability. The occurrence of ESz was correlated with microcephaly (p = 0.04), severe cerebral palsy (CP) (p = 0.03), and failure to thrive (p = 0. 03). In the subgroup of infants with asphyxia, those with ESz were more likely to die of neurologic causes (p = 0.02) and have microcephaly (p = 0.05) or severe CP (p = 0.04). Additionally, those with the greatest number of ESz were more likely to have these severe outcomes. CONCLUSION: The authors' data indicate an association between the amount of electrographic seizure activity and subsequent mortality and morbidity in at-risk infants in general and in infants with perinatal asphyxia. Only with more effective treatment of neonatal electrographic seizures can their potential contribution to poor neurodevelopmental outcome, independent of degree of insult, be ascertained.