RESUMO
BACKGROUND: The incidence of seizures in the neonatal period is thought to be high due to a lower seizure threshold of the immature brain. Data on seizures in extremely premature infants are scarce. OBJECTIVES: The aim of this study was to determine whether seizures are an independent risk factor for in-hospital death and to determine the incidence of seizures in extremely premature infants. METHODS: This was a retrospective cohort study. Included were infants born under 28 weeks' gestation and monitored with amplitude-integrated electroencephalography (aEEG) over the first 3 days of life. The number and duration of seizures was retrieved from aEEG recordings together with clinical data. The association of seizures and other parameters with mortality was assessed using univariable analyses methods. Relevant parameters were used for a multivariable Cox regression analysis. RESULTS: Overall, 229 infants were included in the study. Forty-six infants had at least one seizure episode yielding an incidence of 20%. In univariable analyses, gestational age (p < 0.001), birthweight Z-score (p < 0.001), seizures (p = 0.025), suppressed background aEEG (p < 0.001), and severe intraventricular hemorrhage (IVH; p < 0.001) were associated with death before discharge. In multivariable analysis, gestational age (HR = 0.61, p < 0.001), background aEEG activity (HR = 0.30, p < 0.001), birth weight Z-score (HR = 0.51, p = 0.04), and severe IVH (HR = 2.60, p < 0.001) were found to be significant predictors of mortality while the presence of seizures in the first 3 days of life trended to significantly predict an increased risk of mortality (HR = 1.53, p = 0.09). CONCLUSIONS: Although seizure incidence was relatively high in this cohort of extremely preterm infants and infants with seizures were more likely to die, seizures alone are not a predictor for early death. However, they may be an important indicator of pathologies that are not immediately diagnosed yet could eventually lead to death among this vulnerable population.
