Seizures are most common in the neonatal period (defined as the first 28 days of life or up to 44 weeks conceptional age) compared to any other period in life with an incidence of 1.5–3.5 per 1000 live births.^1,2,3 The incidence may vary according to birth weight, gestational age, and both antenatal and intranatal factors;^3,4,5 most notably, there is an increase risk for seizures in premature compared to full-term infants.

The time of seizure onset is usually within the first few days of life, with 80% of all neonatal seizures presenting within the first week of life.^1,4 Seizures in the neonatal period are also the most common neurological emergency and are associated with high potential mortality and morbidity.^6,7,8,9,10

In contrast to seizures in infancy and childhood, most neonatal seizures are acute and symptomatic with suspected specific causes; relatively few seizures are idiopathic or symptomatic of a clearly defined epilepsy syndrome. Although neonatal seizures have many causes, a limited number account for most seizures (Table 9–1^{11,12,13,14,15}). In term newborns, hypoxic–ischemic encephalopathy (HIE) is the most common underlying factor, typically beginning in the first 2 days of life. In preterm infants, intracranial hemorrhage (ICH) is the most common associated risk factor, although a direct relationship to seizure generation is unclear. Meningitis, focal cerebral infarction, transient metabolic disorders, and congenital abnormalities of the brain may cause seizures at any conceptional age.

The developing brain is particularly susceptible to developing seizures in response to injury; several mechanisms are likely to be involved. Overall the hyperexcitable state of the immature brain is based upon enhanced excitatory neurotransmission, paucity of inhibitory mechanisms, developmental expression of neuronal ion channels, age-dependent modulation of neuropeptides, and age-dependent early microglial activation.^16,17 An important contributing factor is that gamma-aminobutyric acid (GABA), considered the major inhibitory neurotransmitter in older children and adults, is initially excitatory in neonates and only later in development becoming inhibitory.^16,18 These mechanisms may also explain the different clinical semiology of neonatal seizures and the poor response to conventional antiepileptic drugs (AEDs) compared to older infants and children.

The immature brain was traditionally regarded to be less susceptible to seizure-induced injury. This has recently been challenged with the demonstration of long-term changes in behavior and cognition following seizures in animals with early-onset seizures.^19,20 This may in part may reflect cell death in specific neuronal populations, alteration of neuronal networks, and epigenetic changes that may contribute to neurodevelopmental disability and enhanced epileptogenesis following neonatal seizures in animals.^21,22,23,24 More recent clinical data present more conflicting data (see Prognosis below).^25,26

The clinical diagnosis of neonatal seizures may be difficult because clinical manifestations are highly variable with some features unique to that period of life.^27,28 Even among trained observers, clinical neonatal seizures may be difficult to recognize and differentiate from either normal behaviors or abnormal movements of nonepileptic origin.²⁹ Additional problems arise when the relationship between clinical and electroencephalographic (EEG) seizure are considered. There may be temporal overlap of the two (so-called “electroclinical seizures”). However, in some clinical settings up to 85% of electrographic seizures are clinically silent (i.e., EEG discharges with no clinical accompaniment, referred to as “electrical only seizures”), leading to significant underestimation of seizure burden.³⁰ In addition, some clinical seizures may occur in the absence of EEG seizure activity (“clinical only seizures”), but their clinical relevance is controversial.³¹ For these reasons, in the broadest terms a seizure in this age group can be defined either in clinical terms as an abnormal paroxysmal event (with or without EEG seizure activity) or electrographically as a sustained epileptiform change in the EEG, which may or may not be accompanied by paroxysmal alteration in neurological function. However, unless all neonates at risk for seizures are subject to continuous neurophysiologic monitoring—an impractical strategy—clinicians must rely on bedside observation for initial recognition in susceptible infants, followed by diagnostic and then surveillance neurophysiologic studies.

Several features of seizure semiology in the neonatal period differ from those in infancy and childhood. Neonatal seizures do not easily fit into traditional classification schemes.^{27,28,32,33,34} While this continues to be a widely accepted tenant, the most recently proposed ILAE (The International League Against Epilepsy) classification suggested that neonatal seizures should not be considered a distinct seizure type, but could be classified within its more general, universal scheme—which would classify all neonatal seizures as either “focal seizures” or “other.”³⁵

There has been an evolution over several decades of the characterization and classification of neonatal seizures beginning with the pioneering work of French investigators utilizing clinical observation and EEG through the most recent work utilizing EEG-video monitoring. There was initial recognition that neonatal seizures had clinically distinct features,³² then the development of standardized classification systems;^28,33,36 followed by more detailed characterization through initial EEG-video monitoring;^27,31 application of computer-assisted neurophysiological monitoring;^37,38,39 and most recently the use of EEG-video monitoring to refine techniques of clinical recognition and diagnosis.²⁹

Neonatal seizures are classified by several different methods. They may be classified according to clinical features only: based upon either the most predominant clinical feature of the seizure or a more detailed description of the sequence of clinical events. A clinical classification includes: focal clonic, focal tonic, myoclonic, spasms, tonic, and motor automatisms (Table 9–2).²⁷ This scheme utilized the term “motor automatism” to refer to behaviors such as oral–buccal–lingual movements, ocular signs, and movements of progression including stepping, bicycling, and rotary limb movements. These events have also been referred to by as “subtle seizures” by Volpe.^28,33 In addition to these motor manifestations, some neonatal seizures may also have clinical features related to activation of the autonomic nervous system such as changes in heart rate, systemic blood pressure, and respirations,²⁸ although these changes rarely occur in the absence of accompanying motor features.²⁷

As discussed earlier, seizures may also be classified according to their electroclinical findings: the relationship between clinical and electrographic findings. When there is a temporal overlap of clinical and EEG seizures, the events are referred to as “electroclinical” seizures. Those events with only clinical manifestations are “clinical only” seizures, and those manifested only by EEG seizure activity are “electrical only” or subclinical.

Less frequently, seizures have also been classified according to their pathophysiology: epileptic or nonepileptic origin.²⁷ Seizures that are most clearly of epileptic origin are: focal clonic, focal tonic, some myoclonic, and spasms. Those that are more likely to be of nonepileptic origin are: generalized tonic, some myoclonic, and motor automatisms.

Traditionally, the ILAE has also included some neonatal epileptic syndromes in its classification.^34,40 Most recently, the ILAE proposed a revised syndromic classification that now includes: benign neonatal familial seizures, early myoclonic encephalopathy (EME), and Ohtahara syndrome (early infantile epileptic encephalopathy [EIEE]).³⁵ These syndromes will be discussed later.

Electrographically, all neonatal seizures have focal origin (except for epileptic spasms that are typically characterized by a generalized electrodetrimental event and some generalized myoclonic events, which may be generated at a subcortical level). The focal discharges may remain confined to one region or may spread to involve wider areas or the hemispheres opposite to the site of origin. One of the most common sites of seizure onset is the temporal lobe (Fig. 9–1), although they may arise from the frontal, occipital, central, or midline regions. Electrographic events may be brief, with 50% lasting less than a minute⁴¹ and may tend to be shorter in preterm babies.⁴² However, electrographic seizures may also be sustained over several minutes, with some even longer consistent with status epilepticus.

Clinical semiology is, for the most part, determined by site of origin and spread of electrographic seizures. When electrographic seizures remain confined to the neocortex, the clinical manifestations of seizures are predominantly motor; and most of neonatal seizure semiology is based upon these findings. However, in the neonate, the development within the limbic system with its connections to midbrain and brain stem is more advanced than the cerebral cortical organization, leading to a higher frequency of mouthing, ocular changes, apnea, and other clinical features related to the autonomic nervous system in neonates than in older children.

Electrographic seizures in the absence of clinical events may occur in a number of settings. The most obvious occurrence is in infants who have been pharmacologically paralyzed for respiratory care. These electrographic events may also occur in infants with severe encephalopathies and are characterized as “seizure discharges of the depressed brain”⁴³ and “alpha seizure discharges”^44,45 (Fig. 9–1).

In addition, the initial response to the therapy of electroclinical seizures with AEDs is characterized by control of the clinical events with the persistence of electrographic seizures—the clinical events is uncoupled from the electrographic events (see Section Treatment); also called electroclinical dissociation.^27,46 More recently, it has been suggested that hypothermia therapy for HIE may also uncouple electroclinical seizures, with a high incident of subclinical seizures in these cooled infants.^47,48

Although by far the most neonatal seizures are acute and symptomatic, there are several well-defined syndromes which have their onset in the in the neonatal period including neonatal encephalopathies and errors of metabolism.⁴⁰

BENIGN NEONATAL SEIZURES

Benign neonatal seizures (BNS) are also referred to as benign idiopathic neonatal convulsions and, historically, “fifth day fits.” As the designation suggests, seizure onset occurs around the fifth day of life (day 1 to day 7, with 90% between day 4 and day 6) in otherwise healthy neonates. At present the etiology is unknown. Clinically, seizures are characterized as focal clonic variably accompanied by apneic.⁴⁹ The interictal EEG shows theta pointu alternant in approximately 60% of cases—although this finding is not unique to this disorder or diagnostic. In the remaining neonates, the background activity may be either discontinuous, with focal or multifocal abnormalities, or normal. Ictal recordings show unilateral or bilateral rhythmic spikes or slow waves. Treatment with AEDs may not be necessary, because of the self-limiting clinical course of the seizures. However, the diagnosis is one of exclusion. Seizures usually resolve within days. The outcome is good, but an increased risk of minor neurological impairment has been reported.¹ The incidence of BNS has diminished over the past several years to the point that ILAE in its recent proposed classification revision of epileptic syndromes did include BNS as a recognized syndrome.³⁵

BENIGN FAMILIAL NEONATAL SEIZURES

Benign familial neonatal seizures (also referred to as benign familial neonatal convulsions [BFNC])⁴⁹ constitute a rare disorder with autosomal dominant inheritance with a mutation in the voltage-gated potassium channels genes (in most cases at 20q13.3, in a few families at 8q24).⁵⁰ Seizures occur mostly on the second or third day of life in otherwise healthy neonates and tend to persist longer than in benign idiopathic neonatal convulsions. They are mainly focal clonic, sometimes with apneic spells; tonic seizures have rarely been described. The background activity is normal with no specific pattern, although theta pointu alternant has also been associated with this disorder. Therapy is controversial and seizures usually resolve within weeks. The outcome is favorable, but secondary epilepsy may occur.

EARLY MYOCLONIC ENCEPHALOPATHY

EME is a syndrome often associated with inborn errors of metabolism, but cerebral malformations have also been reported.⁵¹ Onset is nearly always in the first month of life. The clinical ictal manifestations include: (1) partial or fragmentary myoclonus, (2) massive myoclonias, (3) partial motor seizures, and (4) tonic spasms. Background EEG activity is characterized by a suppression–burst pattern—bursting consisting of complex bursts of spikes and sharp waves lasting for 1–5 seconds and periods of suppression lasting from 3 to 10 seconds in both waking and sleep. The EEG may later evolve toward atypical hypsarrythmia. Seizures are typically resistant to AED treatment, although hormonal therapy such as ACTH has been reported to may have some temporary effect. All infants are severely neurologically abnormal and half of them die before the age of 1 year.

EARLY INFANTILE EPILEPTIC ENCEPHALOPATHY WITH BURST–SUPPRESSION PATTERN (OHTAHARA SYNDROME)

Age of onset of EIEE is in the first 3 months of life with frequent tonic spasms (100–300 per day), often occurring in clusters. Partial seizures may also occur.⁵¹ The EEG is characterized by a burst-suppression pattern, both in sleep and waking. It may be asymmetric, in part reflecting underlying etiologic factors. This syndrome is usually associated with cerebral malformations. Seizures are resistant to AED treatment, although it has been reported that ACTH may have some temporary effect. The prognosis is serious, but may be somewhat better than for EME. Evolution into infantile spasms is common. Both EME and EIEE have clinically and electrographically distinct features. However, there are also similarities, which have prompted some to suggest that they are not two syndromes, but rather part of a spectrum of a single disorder.⁵¹