Treatment of Neonatal Seizures
Phenobarbital is the recommended first-line antiseizure medication for neonatal seizures, regardless of etiology (HIE, stroke, or hemorrhage), with the exception of suspected channelopathy with positive family history where a sodium channel blocker should be used first. 1
First-Line Treatment Algorithm
Standard First-Line: Phenobarbital
- Dose: 20 mg/kg IV loading dose
- Efficacy: Controls seizures in approximately 80% of neonates when used as first-line therapy 1
- Use for shortest duration possible with early discontinuation for acute provoked seizures responding to treatment, as phenobarbital has been associated with potential adverse effects including hypotension (17% incidence) 1
Exception - Suspected Channelopathy with Family History
- If positive family history suggests self-limited familial neonatal epilepsy (KCNQ2/KCNQ3 variants), use phenytoin or carbamazepine (sodium channel blockers) as first-line instead 1
- This avoids treatment delays while awaiting genetic confirmation
- Without positive family history, still use phenobarbital first to avoid delays 1
Second-Line Treatment Algorithm
When seizures persist after first-line therapy, choice depends on clinical context:
For Most Etiologies (HIE, Stroke, Hemorrhage)
Phenytoin or levetiracetam are the preferred second-line options 1
- Phenytoin: 39% response rate when added after phenobarbital 1
- Levetiracetam: 54% response rate when added after phenobarbital 1
- Less favorable alternatives: midazolam (0% in small trial) or lidocaine (60% response rate) 1
Special Clinical Scenarios
If channelopathy suspected by clinical/EEG features (without prior family history):
- Use sodium channel blocker (phenytoin or carbamazepine) as second-line 1
- Choice depends on clinical state (critically ill vs. stable) and regional drug availability 1
If cardiac disorders present:
- Prefer levetiracetam as second-line to avoid cardiac effects of phenytoin 1
If vitamin B6-dependent epilepsy suspected (based on clinical/EEG features) OR seizures unresponsive to second-line without identified etiology:
- Trial of pyridoxine (add-on to ASM) 1
- If unsuccessful, trial of pyridoxal 5'-phosphate (PLP) may be considered 1
- Do not delay treatment to collect diagnostic samples 1
Duration of Treatment
Discontinue ASM before hospital discharge following cessation of acute provoked seizures (electroclinical or electrographic) without evidence for neonatal-onset epilepsy, regardless of MRI or EEG findings 1
This represents a change from prior guidelines and is supported by evidence showing no difference in neurodevelopment or epilepsy at 24 months between neonates whose ASMs were discontinued versus continued at discharge 1
Critical Diagnostic Considerations
- EEG or amplitude-integrated EEG (aEEG) confirmation is essential for reliable diagnosis 1
- Only focal clonic and focal tonic seizures can be diagnosed by clinical observation alone; all other seizure types require EEG/aEEG confirmation 1
- Many neonatal seizures are electrographic-only without clinical manifestations 2
- Treatment of electrographic-only seizures to reduce seizure burden may be associated with improved outcomes, though direct evidence is limited 1
Common Pitfalls to Avoid
- Do not continue ASM beyond discharge for acute provoked seizures that have resolved, even with abnormal MRI or EEG background 1
- Do not use phenytoin in neonates with cardiac disorders 2
- Do not delay phenobarbital to pursue genetic testing unless there is a positive family history suggesting channelopathy 1
- Avoid emerging agents (oxcarbazepine, brivaracetam, lacosamide) as no controlled efficacy/safety data exist in neonates 1
- Recognize that partial response to ASM does not rule out vitamin B6-dependent epilepsy 1
Evidence Quality Note
The 2023 ILAE Task Force guidelines 1 represent the most comprehensive and recent evidence-based recommendations, superseding the 2011 WHO/ILAE/IBE guidelines. Evidence for first-line phenobarbital is moderate quality from RCTs, while second-line evidence remains very low quality with substantial variability across studies 1.