March 20, 2021


Sharpe C, Reiner GE, Davis SL, Nespeca M, Gold JJ, Rasmussen M, et al. Levetiracetam Versus Phenobarbital for Neonatal Seizures: A Randomized Controlled Trial. Paediatrics 2020; 145(6):e20193182. PMID: 32385134.


Hannah Penner
James Cook University
Mackay Base Hospital, Queensland Health, Australia

Gopakumar Hariharan MD FRACP
Consultant Neonatologist
Senior Lecturer, James Cook University
Mackay Base Hospital, Queensland Health, Australia




(P) Among term neonates less than 2 weeks of age, (I) is levetiracetam more effective and safer compared to (C) phenobarbital when used for the (O) treatment of electrographically-confirmed seizures (T) up to 48 hours after treatment?


  • Design: Randomised controlled trial with primary outcome of 24-hour seizure freedom in neonates with seizures of any aetiology.
  • Allocation: 60:40 randomised allocation ratio
  • Blinding: Double blinded
  • Follow-up period: Infants were followed up until death, transfer, discharge, study withdrawal due to alternative therapy commencement, or exclusion because of not meeting inclusion criteria.
  • Setting: The multicentre study recruited neonates from 6 hospital sites in the USA and New Zealand between March 2013 and October 2017.
  • Patients:
    • Eligible patients included all neonates considered at risk or suspected of having seizures, if aged between 36 and 44 weeks corrected gestational age (< 2 weeks of age) and ≥2.2kg. Neonates with previous anticonvulsant therapy or likely to die imminently were excluded
  • Intervention: Patients were randomised to one of two arms.
    • Phenobarbital group: Treatment was intravenous (IV) phenobarbital load of 20mg/kg.
    • Levetiracetam group: Treatment was IV levetiracetam load of 40mg/kg.
    • Additional doses of the treatment were given if electrographic seizure cessation was not obtained within 15 minutes.
    • A switch to the alternate treatment method was administered after failure of seizure control at 30 minutes following second dose of treatment. Second doses of this were prescribed if seizure persisted at 45 minutes.
  • Outcomes:
    • Primary outcome: Primary outcome was complete cessation of seizures for 24 hours.
    • Secondary outcomes:
      • Rate of achieving and maintaining seizure control for 48 hours
      • Rate of achieving and maintaining seizure freedom for 1 hour
      • Sub-analyses of neonates with hypoxic-ischemic encephalopathy specifically undergoing therapeutic hypothermia
      •  Levetiracetam dose-escalation component to determine the number of neonates with seizure control at 60mg/kg after initial treatment with a loading dose of 40mg/kg.
  • Analysis and Sample Size: The sample size included 280 neonates initially of which 106 met inclusion criteria for the trial and were randomly assigned. 83 patients had outcome data collected and were included in final efficacy analysis.
  • Patient follow-up: (% included in analysis): 83 of the initial 106 neonates (78.3%) were included in the final results.


Patient characteristics and demographics of interest n (%): Table 1

Levetiracetam Phenobarbital Overall
HIE as seizure aetiology 35 (55%) 22 (52%) 57 (54%)
Received hypothermia treatment 24 (38%) 18 (43%) 42 (40%)
Male sex 31 (48%) 24 (57%) 55 (52%)
Cord pH mean 7.07 (0.2 SD) 7.15 (0.17 SD) 7.1 (0.19 SD)
5 min APGAR score mean 6.52 (3.01 SD) 6.47 (2.4 SD) 6.5 (2.78 SD)
Gestational age mean weeks 39.3 (1.3 SD) 39.1 (1.3 SD) 39.3 (1.3 SD)
Birth weight mean 3342 (557 SD) 3317 (501 SD) 3332 (546 SD)
Pre-treatment seizure severity mean 12.3 (12.0 SD) 9.1 (9.3 SD) 11.1 (11.2 SD)

Primary and secondary outcome measures: Table 2

Phenobarbital (% cessation) Levetiracetam Fischer’s Exact P value Relative Risk (95% CI)
Primary outcome: 24 hours cessation of seizures 24 of 30 (80%) 15 of 53 (28%) <0.001 0.35 (0.22-0.56)
Secondary outcome: 48-hour cessation 18 of 28 (64%) 8 of 47 (17%) <0.001 0.26 (0.13-0.53)
Secondary outcome: 1 hour cessation 28 of 30 (93%) 26 of 53 (49%) <0.001 0.53 (0.39-0.7)
Secondary outcome:  Sub analysis HIE hypothermia cohort: 24-hour cessation rate 9 of 10 (90%) 6 of 17 (35%) 0.014 0.39 (0.2-0.77)


In the randomised controlled trial of neonates treated with either levetiracetam or phenobarbital, phenobarbital demonstrated greater efficacy in termination of electrographic seizures compared to levetiracetam. Adverse effects were more common in phenobarbital-treated neonates. Current evidence is unsupportive for levetiracetam as a first line therapy in neonatal seizures.


Neonatal seizures pose a diagnostic and therapeutic challenge in neonatal care.(1) Prolonged seizures are associated with developmental and cognitive impairment and higher neonatal mortality.(2) Rapid recognition and effective treatment are essential in reducing these harmful impacts on the developing brain.(3)

Neonatal seizures often lack definitive clinical features and are under or over diagnosed.(4) Although continuous electroencephalogram (cEEG) remains the diagnostic gold standard, poor availability and resource intensive nature limits its routine use in many neonatal units.(2,3) Furthermore, the controversy regarding the endpoints of neonatal seizure management (clinical or electroencephalographic) is not resolved.

Phenobarbital is the most common anticonvulsant used in neonatal care.(4) However, phenobarbital is associated with neuronal toxicity in animal models and poor neurocognitive and motor development in neonates.(2) Emerging evidence suggests that levetiracetam may be a promising alternative.(2) Levetiracetam has been identified to have neuroprotective effects on the developing brain and seems to have a better safety profile compared to other anticonvulsant drugs.(2,4) However, most prior levetiracetam studies in neonates examine the medication as a second-line treatment.(5)

A systematic review by Sharma et al (includes pre-print NEOLEV2 trial) suggested levetiracetam achieved better safety and side effect profile compared to phenobarbital, but evidence for its clinical use is unclear.(2) A recent study examined levetiracetam use in clinical seizures. The open-labelled RCT by Gowda et al found increased efficacy of levetiracetam compared to phenobarbital in clinical seizures, but did not account for a possible electrographic-uncoupling effect (electrographic seizure activity without clinical manifestations) of levetiracetam.(6)

NEOLEV2 is the first RCT directly comparing the efficacy and safety of levetiracetam to phenobarbital in EEG detected seizures. The authors conclude that phenobarbital is a better anticonvulsant of choice in electrographic seizures after analysing both clinical and electrographic cessation of seizures. The adverse effects reported in the levetiracetam arm were fewer compared to phenobarbital although not statistically significant. Neonates experienced increased rates of respiratory depression, hypotension and sedation in the phenobarbital arm. A potential source of selection bias may exist due to only neonates with cEEG confirmed seizures being recruited as participants. This probably resulted in exclusion of neonates with short lasting clinical seizures that may have terminated prior to cEEG application. A previous study estimated 12 minutes for application of the electrodes to obtain an EEG record.(7)

The NEOLEV2 RCT was a rigorously conducted study and provides important evidence regarding use of levetiracetam and phenobarbital in neonatal seizures. However, its generalisability into all clinical contexts is inconclusive considering the lack of availability of cEEG, and trained personnel for interpreting them in many neonatal units.(8) The study is also limited by the lack of long term neurodevelopmental follow up.

The current evidence regarding levetiracetam’s use as a first line therapy and optimal dosing in neonatal seizures is inconclusive.(2) Rigorous studies investigating EEG-uncoupling effects of anticonvulsants, and randomised controlled trials investigating long term neurodevelopmental effects of treating clinical versus EEG seizures are important to determine optimal treatment strategies for neonatal seizures.


  1. Mohammadi M, Rezaei Z. Neonatal seizure: how reliable is its diagnosis and treatment? A mini review of previous knowledge. Iran J Child Neurol 2020; 14(4):9-16.
  2. Sharma D, Hussain AM, Sharma SS. Efficacy of Levetiracetam in neonatal seizures: a systematic review. J Matern Fetal Neonatal Med 2020; 1-8.
  3. Boylan GB, Stevenson NJ, Vanhatalo S. Monitoring neonatal seizures. Semin Fetal Neonatal Med 2013; 18(4):202-8.
  4. Hellstrom-Wastas L, Boylan G, Agren J. Systematic review of neonatal seizure management strategies provides guidance on anti-epileptic treatment. Acta Pediatr 2015; 104(2):123-9.
  5. Giva S, Boyle MA, Gorman KM. Should levetiracetam rather than phenobarbitone be the first-line treatment for neonatal seizures? Arch Dis Child 2020; 56(8):643-646.
  6. Gowda VK, Romana A, Shivanna NH, Benakappa N, Benakappa A. Levetiracetam versus phenobarbitone in neonatal seizures – a randomised controlled trial. Indian Pediatr 2019
  7. Lloyd R, Goulding R, Filan P, Boylan G. Overcoming the practical challenges of electroencephalography for very preterm infants in the neonatal intensive care unit. Acta Paediatr 2015; 104(2):152-7.
  8. Sharpe C, Davis SL, Reiner GE, Lee LI, Gold JJ, Nespeca M, et al. Assessing the feasibility of providing a real-time response to seizures detected with continuous long-term neonatal electroencephalography monitoring. J Clin Neurophysiol 2019; 36(1):9-13.

Leave a comment