Levetiracetam as the first-line treatment for neonatal seizures: a systematic review and meta-analysis

To assess the effectiveness and safety of levetiracetam when used as first-line treatment of neonatal seizures.

Four electronic databases, Medline, Embase, Web of Science, and ClinicalTrials.gov were systematically searched from inception until 20th November 2020. Randomized controlled trials (RCTs) and observational studies that included neonates born preterm and term were eligible for inclusion. The primary outcome measure was levetiracetam effectiveness, defined as seizure cessation within 24 hours of starting treatment. Secondary outcomes included short-term adverse events, mortality before discharge, and long-term neurodevelopmental outcomes.

Fourteen studies assessing 1188 neonates were included: four RCTs, three observational trials with phenobarbital as the control arm, and seven observational studies of levetiracetam with no control arm. Pooled efficacy of levetiracetam from observational studies was 45% (95% confidence interval [CI] 34-57%) (GRADE - very low). Meta-analysis of RCTs evaluating levetiracetam versus phenobarbital showed that both were equally effective (risk ratio [95% CI] 0.6 [0.30-1.20]) (GRADE - very low). Levetiracetam resulted in a lower risk of short-term adverse events compared to phenobarbital (risk ratio [95% CI] 0.24 [0.06-0.92]) (GRADE - moderate).

Very low certainty of evidence suggests levetiracetam might not be more effective than phenobarbital. Moderate certainty of evidence indicates levetiracetam is associated with a lower risk of adverse events. Future trials on neonatal antiseizure medication therapy should include continuous electroencephalogram (EEG) monitoring as standard of care and enrol a homogenous population with similar seizure aetiology. What this paper adds Levetiracetam is effective in 45% of neonatal seizures. Levetiracetam might not be more effective than phenobarbital. Levetiracetam is likely to be safer than phenobarbital. Evidence available is limited and of very low certainty.