Neonatal Seizures

by Michael Wong, MD, PhD
Dr. Wong is an Assistant Professor of Neurology at Washington University School of Medicine and an epileptologist in the Pediatric Epilepsy Center at St. Louis Children’s Hospital.


Seizures during the neonatal period (generally defined as the first month of life) are relatively common, occurring in approximately 1% of all neonates. Neonatal seizures represent an age-specific seizure disorder, which is usually considered to be in a separate category from epilepsy. While in children seizures often occur in the absence of another neurological disorder, neonatal seizures frequently are a non-specific sign of an underlying disease. Neonatal seizures also have many other characteristics that are quite different from seizures in children and adults. This article will review the unique clinical and EEG features, causes, treatment, and prognosis of neonatal seizures.

Clinical and EEG features

Seizures in neonates have a number of clinical features that differ significantly from seizures in older patients. Generalized tonic-clonic or “grand-mal” seizures, which are common in adults, are rare in neonates. This is probably due to the anatomical and physiological immaturity of the neonatal nervous system, such as the incomplete myelination pattern of neonates, which tends to prevent highly-organized, synchronized, generalized seizure activity. Instead, neonatal seizures usually have much more focal or subtle features. Based on clinical symptomatology alone, neonatal seizures are usually divided into four main categories: clonic, myoclonic, tonic, and subtle. Clonic seizures are characterized by rhythmic, repetitive shaking of isolated parts of the body. Focal clonic seizures involve one part of the body and multifocal clonic seizures involve several parts simultaneously or in sequence with migration in a random fashion. Myoclonic seizures represent very brief isolated jerks of parts of the body. Tonic seizures involve stiffening of parts of the body and can be focal or generalized, although generalized tonic stiffening is usually not due to true electrographic seizure activity (see below). Finally subtle seizures consist of other miscellaneous paroxysmal clinical phenomena, such as eye deviation, oral-buccal movements, or patterned movements of the extremities like bicycling.

For years, all four of the above clinical categories were believed to be epileptic seizures based purely on clinical grounds. Landmark work by Mizrahi and Kellaway in the 1980’s showed that not all such clinical spells in neonates represented seizures. Using video-EEG, it was found that most clonic seizures, myoclonic seizures, and focal tonic seizures had definite epileptiform EEG correlates and thus were definitely epileptic seizures. However, generalized tonic seizures and most forms of subtle seizures, except for isolated eye deviation, did not have any EEG correlate and thus were not felt to be true epileptic seizures. Mizrahi and Kellaway proposed that these non-epileptic spells represented abnormal brainstem release reflexes, possibly due to loss of cortical inhibition from cortical damage. The practical importance of these findings was that the neonatal spells without an epileptiform EEG correlate probably did not warrant aggressive treatment with antiepileptic drugs.

Similarly, electroclinical dissociation is a common phenomenon in neonates, where the clinical component of a seizure can occur at some times with an EEG correlate and other times without an EEG correlate in the same patients. Conversely, it has also been shown that neonates frequently have electrographic seizures on EEG with no clinical correlate. The frequent dissociation between clinical and EEG findings in neonate raises complicated issues about how to monitor and treat seizures in neonates. (See below).


A specific etiology for neonatal seizures can usually be identified in the majority of cases. In contrast to children who often develop idiopathic epilepsy without another identified cause, it is relatively rare for an otherwise healthy baby to have epileptic seizures. In this respect, neonates more closely resemble older adults in that most new-onset seizures among the elderly are also symptomatic, such as from a stroke, brain tumor, or neurodegenerative disease. In fact, because the neonatal nervous system has a relatively limited repertoire of behaviors and symptoms, seizures are often the only neurological manifestation of a serious underlying neurological or systemic disease. Not surprisingly, neonatal seizures are most commonly encountered in the neonatal ICU in critically-ill babies. Therefore, if a neonate develops seizures and the etiology is not readily apparent, an extensive diagnostic work-up is often warranted.

Overall, the most common etiology for neonatal seizures is hypoxic-ischemic encephalopathy. Other major etiological categories include intracranial hemorrhage or stroke, meningitis or other infections, acute metabolic derangements, such as hypoglycemia or hypocalcemia, inborn errors of metabolism or other metabolic diseases, such as amino or organic acidurias, urea cycle defects, mitochondrial diseases and pyridoxine dependency, and with improving neuroimaging techniques, various structural malformations of cortical development. Although relatively rare, there are some idiopathic or benign familial epilepsies in neonates, such as benign familial neonatal convulsions, which was one of the first epilepsies to be identified as a channelopathy, involving a potassium channel.


The general principles of treating seizures in neonates are similar to children and adults, but there are some important differences. Like seizures in older patients, if there is a treatable cause for the seizures, such as hypoglycemia, infection, or intracranial hemorrhage, then the underlying etiology should be treated with the appropriate specific therapy. Otherwise, neonatal seizures are generally treated with antiepileptic drugs. The most common antiepileptic drugs used in neonates are phenobarbital, phenytoin, and various benzodiazepines. The repertoire of drugs used for neonatal seizures is relatively limited for several reasons. First, few drugs have been formally tested in the neonatal population. Second, oral preparations often cannot be used in neonates. Finally, the complicated metabolism and pharmacokinetics of neonates makes use of some drugs difficult. Phenobarbital is especially popular in neonates because of its long half-life and favorable pharmacokinetic properties. A recent study in the New England Journal of Medicine in 1999 was one of the first controlled studies to compare efficacy of antiepileptic drugs in neonatal seizures. This study compared phenytoin and phenobarbital and found them to be equally effective, each stopping neonatal seizures in about 45% of patients. When both treatments were combined, ~60% of patients’ seizures were controlled. This still left about 40% of patients whose seizures were not adequately controlled with these drugs, indicating that there is still room for improvement in developing new therapies for neonatal seizures.

Given that neonates have frequent subclinical electrographic seizures, there is some controversy as to how the effect of the antiepileptic drug should be titrated. Should the goal of treatment be stopping only clinical seizure activity or should the goal be cessation of all electrographic seizure activity, which requires continuous EEG monitoring? Ideally, cessation of all electrographic seizures is probably desirable. Theoretically, ongoing subclinical seizure activity may cause neurological damage through excitotoxicity or other cellular mechanisms. On the other hand, subclinical seizure activity in itself has not been proven to alter long-term neurological prognosis. Furthermore, to completely suppress all electrographic seizure activity in some babies may require extremely high doses of multiple antiepileptic drugs, resulting in making the baby deeply comatose and risking serious systemic side effects. The decision on how aggressive to be in treating electrographic seizures in neonates will depend on multiple considerations, such as the underlying etiology, predicted prognosis, and other systemic factors.

After recovery from neonatal seizures, the duration of treatment with antiepileptic drugs beyond the neonatal period varies, depending most on the predicted prognosis for seizure recurrence off medication (See below). If a patient has been seizure-free on medication and is not at high risk for seizure recurrence, it is reasonable to try and wean the medication off after just a few months. This is based on the rationale that if the acute stressor causing the seizures has resolved, the patient is unlikely to have epilepsy. Even if a patient is in a high risk group, it might still be reasonable to give them a careful trial off medication around a year of age if they have been seizure-free on medication. Obviously, patients who continue to have seizures beyond the neonatal period on medicine have epilepsy and should remain on medication.


There have been numerous studies looking at the outcome and prognosis of patients with neonatal seizures. Overall, ~10-50% of patients with neonatal seizures die, mostly during the neonatal period, and ~50% of survivors develop long-term neurological complications, like epilepsy, mental retardation, and cerebral palsy. With regard to both mortality and morbidity, most studies agree that the most important prognostic factor in predicting outcome in patients with neonatal seizures is the underlying etiology. Patients with hypoxic-ischemic encephalopathy, intraventricular hemorrhage, and structural cerebral anomalies have the worst prognosis. Patients with meningitis have an intermediate prognosis. Patients with transient metabolic disturbances or presumed idiopathic genetic or familial etiologies have the best prognosis.

Neonatal neurological exam is also a good predictor of outcome in patients with neonatal seizures, with a normal neurological exam being associated with a good prognosis and an abnormal exam predicting poor outcome. Neonatal EEG has also been found to be a useful predictor of outcome in patients with neonatal seizures. Severe EEG abnormalities, such as burst suppression, severe low voltage, or multifocal abnormal discharges, are associated with abnormal neurological outcome in >80% of cases. On the other hand, a normal background activity on EEG despite electrographic seizures has a normal developmental outcome in ~80%.


Seizures are relatively common in the neonatal period. The clinical and EEG features of neonatal seizures can be unique and differ from seizures in older patients. Neonatal seizures often represent a sign of a serious underlying neurological or systemic disorder and therefore if there is not a readily-identifiable cause, an extensive diagnostic work-up is often warranted to look for various etiologies. Treatment for neonatal seizures includes specific therapy for the underlying cause and use of antiepileptic drugs. While ideally treatment with antiepileptic drugs will stop all electrographic seizure activity, the goals of treatment need to be tailored to the specific situation. Prognosis following neonatal seizures is extremely variable and depends most on the underlying etiology of the seizures.

Selected References

  1. Mizrahi EM, Kellaway P. Characterization and classification of neonatal seizures. Neuerology 1987;37:1837-1844.
  2. Painter MJ, Scher MS, Stein AD, et al. Phenobarbital compared with phenytoin for the treatment of neonatal seizures. N Engl J Med 1999;341:485-489.
  3. Rose AL, Lombroso CT. Neonatal seizure states: a study of clinical, pathological, and electroencephalographic features in 137 full-term babies with a long-term follow-up. Pediatrics 45:404-425.
  4. Rust RS, Volpe JJ. Neonatal seizures. In: Pediatric Epilepsy: Diagnosis and Therapy, Dodson WE, Pellock JM, eds. New York: Demos, pp. 107-128.
  5. Volpe JJ. Neurology of the Newborn, 3rd edition. Philadelphia: WB Saunders Co, 1995.