Neonatal Seizures: Diagnosis, Etiologies, and Management

Julie Ziobro, MD, PhD; Renée A. Shellhaas, MD, MS


Semin Neurol. 2020;40(2):246-256. 

In This Article

Management—Acute Intervention Versus Long-term Treatment

New-onset seizures in a neonate are often a symptom of a serious underlying neurologic condition and should be treated as a medical emergency. In most cases, treatment with ASMs should be initiated as soon as EEG seizures are confirmed. However, in patients with high-risk conditions such as HIE, ICH, or repetitive focal clonic seizures (a semiology often associated with perinatal stroke), ASM treatment should be initiated as quickly as possible, even before the seizures are confirmed on EEG.[52] The goal of treatment should be seizure cessation (including EEG-only seizures) and medications should be titrated rapidly to attain this goal.


Few clinical trials are available to guide treatment algorithms, but phenobarbital remains the most commonly used agent for first-line treatment of neonatal seizures.[53] Seizures are controlled in approximately half of patients with acute symptomatic seizures after a single loading dose of 20 mg/kg.[54] In patients whose seizures do not respond to an initial bolus of phenobarbital, additional bolus doses are usually indicated prior to moving on to second-line agents. Various factors have been shown to be associated with resistance to phenobarbital (e.g., extreme prematurity and seizures due to hemorrhage).[55] Nonresponders with HIE tend to have more severely abnormal EEG backgrounds, higher seizure burdens, and a more severe pattern of injury on MRI (e.g., including injury to the white matter, parenchymal cortex, and more severe watershed injury).[56]


Intravenous fosphenytoin has similar efficacy to phenobarbital as an initial treatment for neonatal seizures,[54] but the pharmacokinetics of phenytoin can be unpredictable in neonates. In addition, fosphenytoin can lead to hypotension and cardiac arrhythmias, making it less desirable than phenobarbital as a first-line agent. However, fosphenytoin is frequently used as a second-line agent if phenobarbital fails to control seizures, as long as the infant has appropriate cardiovascular monitoring and laboratory monitoring. Of note, phenytoin is a protein-bound drug. As critically ill neonates often have low albumin levels, measurement of both free and total phenytoin levels can be informative.


There is considerable disagreement among observational clinical studies of levetiracetam as a first-line agent for neonatal seizures.[57] Phenobarbital has important side effects including sedation, abnormal neuronal apoptosis in rodent models, and memory problems when used long term.[58] Such data have led to increased interest in the use of levetiracetam as a first-line agent for neonatal seizures, despite a lack of clinical trial data or rigorous studies to demonstrate efficacy or safety in this population. The popularity of levetiracetam is likely related to its availability in oral and intravenous formulations, its relative safety in older infants and children, and published pharmacokinetic data in newborns.[59,60] Until recently, studies assessing the efficacy of levetiracetam versus phenobarbital to treat neonatal seizures have been limited. Retrospective studies suggested that levetiracetam may provide superior seizure control when compared with phenobarbital, based on time to seizure cessation. However, retrospective studies are limited by their nonrandom treatment allocation and the natural history of acute symptomatic seizures (e.g., seizures related to HIE are often frequent in the first day and decrease in the second day[61]).

Recently, the NEOLEV2 trial (NCT01720667) has provided a rigorous, multicenter, randomized, blinded, controlled phase IIb trial comparing levetiracetam with phenobarbital as the first-line treatment of acute symptomatic neonatal seizures. Seizures were defined by gold standard EEG monitoring. The study found that 80% of neonates randomized to phenobarbital (20 mg/kg loading dose with an additional 20 mg/kg if seizures persisted) as the first ASM and remained seizure free for 24 hours, compared with only 28% of subjects randomized to levetiracetam (40 mg/kg loading dose with an additional 20 mg/kg if seizures persisted). Phenobarbital had an increased rate of adverse effects including hypotension, respiratory suppression, and sedation, but these were easily managed in the NICU setting. Dose escalation from 40 to 60 mg/kg of levetiracetam increased efficacy of levetiracetam only by 7.5%.[62] Levetiracetam appeared to be most effective for neonates with a lower seizure burden, though even in this scenario the efficacy remained inferior to phenobarbital. This latter finding may still be consistent with retrospective studies of levetiracetam versus phenobarbital, as patients with a lower seizure burden tended to receive levetiracetam.[61]

Results of NEOLEV2 are practice-changing as they establish that levetiracetam is not as effective as phenobarbital for initial neonatal seizure management. As such, levetiracetam is not recommended as a first-line agent for acute symptomatic neonatal seizures. Importantly, well-designed studies such as NEOLEV2 which refute the secular trend of treatment with newer ASMs underscore the urgent need for high-quality treatment trials for neonatal seizures.

Second-line Treatments

Approximately 66% of patients with acute symptomatic seizures have an incomplete response to the initial loading dose of ASM, regardless of the medication selected. Incomplete response to an initial loading dose in acute symptomatic seizures is more likely among neonates with extreme prematurity or whose seizures are due to intracerebral hemorrhage.[55] There are no data from clinical trials to inform decisions about second- or third-line treatments for neonatal seizures. Local resources vary; so, a carefully thought out local protocol that is individualized for particular hospitals is suggested as the best approach to ensure timely treatment and appropriate escalation of management. A sample protocol is provided in Figure 1. If seizures are refractory to repeated loading doses of standard ASMs, infusions may be indicated. Midazolam infusion may be used in refractory seizures in the setting of acute brain injury or in status epilepticus. Lidocaine infusions may be indicated for refractory acute symptomatic seizures, though it is contraindicated in neonates with congenital heart disease or those who have been treated with fosphenytoin due to the risk of arrhythmia.[63]

Figure 1.

Sample algorithm for the treatment of neonatal seizures. The algorithm should generally be followed until seizure freedom is attained. ASM, antiseizure medication.

Duration of Treatment of Acute Symptomatic Seizures

Acute symptomatic neonatal seizures typically arise within the first 24 to 48 hours of life and persist for 48 to 96 hours before naturally dissipating.[64,65] Video EEG monitoring can help determine when seizures arise, as well as their tempo and resolution, which is important to guide acute medication management. Once acute symptomatic seizures resolve, there is as much as a 25% chance of postneonatal epilepsy (emergence of repeated unprovoked seizures later in infancy or childhood).[4] There is no clear consensus on the ideal duration of ASM following resolution of acute symptomatic seizures; in one large study, duration of treatment was best predicted by the study center rather than specific patient characteristics.[66] Ongoing work is designed to evaluate the impact of short versus prolonged ASM treatment on later epilepsy and neurodevelopmental outcomes (NCT 02789176).

If ASMs have been administered for less than 1 week, they can be stopped without a wean. This is particularly true of phenobarbital, given the long half-life of this medication. If a patient is to be continued on maintenance medications upon discharge from the NICU, the treatment regimen should be simplified (usually not more than two standing ASMs prescribed at the time of discharge from hospital) and there should be a clear plan to reassess the infant with a goal of discontinuing ASMs during the first few months of life. This will avoid prolonged, unnecessary exposure to medications, and allow for transition to a different ASM should the child develop postneonatal epilepsy (e.g., for new-onset nonsyndromic epilepsy in infants, levetiracetam is more likely than phenobarbital to result in seizure control—a finding that is in contrast with the evidence for neonates).[67]

Neonatal Epilepsy Treatment

Initial treatment of neonatal-onset epilepsy frequently follows the same algorithm as acute symptomatic seizures. However, once the diagnosis of neonatal epilepsy has been made, more precise, etiology-specific treatment may be utilized. Except for self-limited neonatal seizures, long-term treatment for neonatal epilepsy is usually indicated. Among neonates with structural epilepsy (e.g., lissencephaly and focal cortical dysplasia), chronic ASMs such as topiramate, oxcarbazepine may be considered, though data are limited.[2,68–70]

The advancement of clinical genetics has led to increased opportunities for precision medicine. Neonatal epilepsy due to metabolic etiologies (e.g., pyridoxine-dependent epilepsy) may be vitamin responsive or require specific metabolism-based treatment strategies, as discussed earlier. Infants with KCNQ2 and KCNQ3-related epilepsies have a response to sodium-channel blocking agents (e.g., carbamazepine or oxcarbazepine), whereas children with SCN1A-related epilepsies may have an adverse response to those medications. We anticipate an increasing number of effective, targeted therapies as access to genetic testing improves and translational science advances. Resources such as the Rare Epilepsy Network can direct providers and families to disease-specific resources and relevant research opportunities.