Neonates with seizures prospectively diagnosed in a population-based setting in Newfoundland, Canada, between 1990 and 1995, were followed to a median age of 10 years. The incidence of epilepsy, and physical and cognitive impairments were analyzed for 62 term and 26 premature infants. Outcome was normal in 31 children (35%); 36 (41%) had neurodevelopmental impairments: and 21 (24%) had died. The median age at death was 13 months; range 2 days to 14 years. Gestational age was a better predictor of outcome than birth weight. Normal outcome occurred more often in term than preterm infants (p=0.003). Postneonatal epilepsy developed in 27 (34%), including 29% of term and 48% preterm infants. Median age at onset of epilepsy was 9 months in term and 4 months in preterm infants. Epilepsy complicating mental retardation (MR) and cerebral palsy (CP) was correlated with early death. CP occurred in 35% of those surviving infancy, MR in 32%, and learning disorders in 23%. Neonatal encephalopathy occurred in 42%, usually following hypoxic-ischemic events. Infants with Sarnat stage III encephalopathy either died or had MR, CP, and symptomatic epilepsy. Infections occurred in 19%; outcome was normal in term infants with infection, but poor in premature infants or with CMV infection.
Abnormal neonatal EEGs, usually an abnormal background activity, correlated closely with poor outcome (p<0.0001). Other variables associated with a poor prognosis included cerebral dysgenesis, intraventricular hemorrhage, and need for multiple AEDs. Purely clonic seizures without facial involvement in term infants had a favorable prognosis, whereas generalized myoclonic seizures in preterm infants were associated with high mortality. Subtle seizures, generalized tonic seizures, and number of seizures were not predictive of outcome. The severity and timing of the brain insult are the major determinants of outcome. 
COMMENT. In an editorial, Dlugos D and Sirven JI  refer to the study by Mizrahi E and Kellaway P.  who found, using video-EEG, that focal clonic seizures and focal tonic seizures were consistently associated with ictal activity on EEG. In contrast, generalized myoclonic events showed ictal EEG correlation in 60% of cases. Motor automatisms (subtle seizures), generalized tonic events, and focal myoclonic jerks were not associated with EEG ictal activity, suggesting that these clinical events may not be epileptic. A distinction was made between clinical events that were epileptic and those associated with an encephalopathy but were not neonatal seizures. The report by Ronen and associates stresses the unfavorable prognosis of neonatal seizures, especially in preterm infants, and with the exception of clonic seizures without facial involvement. Outcome is determined primarily by gestational age and the severity and timing of the encephalopathy resulting in seizures. The authors emphasize the need for preventive measures and antiepileptic medications more specific for neonatal seizures.
A study of the etiology and neurodevelopmental outcome of seizures in term newborn infants [4, 5] found that neonatal mortality was 7%, and neurologic outcome of survivors at 1 year was favorable in 72%. Predictors of a favorable outcome were a normal neonatal neurologic exam and normal EEG. Global cerebral hypoxic-ischemia is the most frequent cause (40%) of neonatal seizures and a strong predictor of poor long-term outcome. The evidence favors the etiology as the most important factor in prognosis and not the neonatal seizure per se . in developing laboratory animals, a major tonicclonic seizure with post-ictal depression cannot be induced until the animal is older and cortical organization more mature.