Purpose: Neonatal seizures are common, especially in prematurity. Phenobarbital (PB) currently represents the antiepileptic drug (AED) of choice, despite being related to increased neuronal apoptosis in animal models and cognitive impairment in human subjects. Levetiracetam (LEV) may have a more favorable profile since it does not cause neuronal apoptosis in infant rodents. Methods: In a prospective feasibility study, LEV was applied as first-line treatment in 38 newborns with EEG-confirmed seizures, after ruling out hypoglycemia, hypocalcaemia, hypomagnesaemia and pyridoxin dependency. Initial intravenous doses of 10 mg/kg LEV were gradually increased to 30 mg/kg over 3 days with a further titration to 45-60 mg/kg at the end of the week. Acute intervention with up to 2 intravenous doses of PB 20 mg/kg was tolerated during LEV titration. LEV was switched to oral as soon as the infants' condition allowed. Based on clinical observation, EEG tracings (aEEG/routine EEGs), and lab data, drug safety and anticonvulsant efficacy were assessed over 12 months. Results: In 19 newborns a single PB dose of 20 mg/kg was administered, while 3 newborns received 2 PB doses. 30 infants were seizure free under LEV at the end of the first week and 27 remained seizure free at four weeks, while EEGs markedly improved in 24 patients at 4 weeks. In 19 cases, LEV was discontinued after 2-4 weeks, while 7 infants received LEV up to 3 months. No severe adverse effects were observed. Conclusions: These results illustrate the safety of LEV treatment in neonatal seizures, including prematurity and suggest LEV anticonvulsant efficacy. Additional PB treatment admittedly constitutes a methodological shortcoming due to the prolonged anticonvulsive efficacy of PB. Double blind prospective controlled studies and long-term evaluation of cognitive outcome are called for.