RESUMO
OBJECTIVE: To determine whether the seizure clinical type is predictive of etiology. METHODS: This is a prospective study from 9 neonatal and pediatric intensive care units. Neonatal seizures were defined clinically as described by Volpe. The seizures etiology was defined by positive clinical date, cranial sonogram and laboratory investigation routinely obtained in all neonates with seizures. Correlation between seizures type and etiologies was determined by Odds ration (OR). RESULTS: During one year 89 neonates develop neonatal seizures and the etiologies were determined. The mains seizures types were: subtle (n = 59; 66.3%), tonic (n = 27; 30.4%), clonic (n = 20; 22.4%), myoclonic (n = 4; 4.5%). The most frequent etiologies determined were: asphyxia (n = 60; 67.4%), hypocalcemia (n = 19; 21.3%), hypoglicemia (n = 14; 15.7%), peri-intraventricular hemorrhage (n = 13; 14%). The OR estimate that clonic seizures (OR 5.65 - 1.34 < OR < 23.61%) is predictive to peri-intraventricular hemorrhage. CONCLUSION: The OR estimate high risk to periventricular hemorrhage when the seizure is the clonic type. When the seizure is subtle type the risk is low to this etiology.