Abstract
Seizures in newborns are the most frequent manifestation of an acute neurologic dysfunction, but sometimes they can be the early onset of an epileptic encephalopathy, mainly of genetic origin. When neonatal seizures are suspected, a video-encephalographic (EEG) recording is fundamental for correct interpretation of the clinical phenomena to avoid unnecessary anticonvulsant therapy. After a diagnosis of seizures, continuous monitoring is necessary to evaluate seizure burden, to diagnose status epilepticus, to monitor efficacy of anticonvulsant medications, and to form a differential diagnosis with the uncoupling phenomenon. Although the outcome of newborns depends mostly on the underlying etiology, the presence of seizures seems to have a negative prognostic value, especially if the seizures are prolonged. The most common antiepileptic drugs are only partially effective. Phenobarbital is the most used first-line drug with no differences between preterm and full-term newborns. Moderate to severe abnormal background EEG activity and electrographic-only seizures seem to be associated with a lack of response to phenobarbital. Moreover, seizures management in newborns has not significantly changed if the etiology is acute symptomatic. However, evidence of promising therapeutic approaches for neonatal epilepsy encephalopathies is growing. An accurate and prompt diagnosis of seizures is required for early management to avoid the potential detrimental effects of seizures on the immature brain.
Keywords
anticonvulsant drugs, epilepsy, newborn
- •
Seizures are the most common manifestation of acute neurologic dysfunction in newborns. Neurophysiological monitoring (with cEEG or, if unavailable, aEEG) is recommended for correct diagnosis.
- •
After EEG confirmation, management is based upon diagnostic work-up and the choice of the optimal therapy according to the clinical scenario (acute symptomatic origin versus suspected genetic origin).
- •
Prognosis seems to be mostly related to etiology, and whether seizures may aggravate the underlying brain injury/disfunction is still debated.
- •
The introduction of new, more effective, anti-seizure drugs and their combined use with neuroprotective agents may improve the outcome of these newborns in the next future.
Neonatal seizures represent a frequent clinical sign of a central nervous system (CNS) disorder in newborns. The vast majority (85%) of seizures are of acute symptomatic origin and need to be distinguished from neonatal-onset epilepsies and neonatal-onset epileptic encephalopathy, mainly of genetic origin (as discussed later in the chapter). Acute symptomatic neonatal seizures may indicate a potentially treatable etiology, such as metabolic derangements (i.e., glucose or electrolyte abnormalities) or the presence of an acquired brain injury or dysfunction, or they can be secondary to neonatal encephalopathy and hypoxic-ischemic encephalopathy, structural brain injuries (i.e., ischemic/hemorrhagic stroke), or CNS/systemic infections, with prognostic implications for both mortality and subsequent neurologic outcome.
The importance of acute symptomatic seizures also derives from their epidemiologic relevance (the estimated incidence being between 1% and 3.5% in full-term newborns and even higher in preterm infants). Moreover, they require complex acute medical decision making regarding treatment strategies, which are mandatory because of a potential risk of additive detrimental effects of seizures on the neonatal brain, as suggested by experimental models yet not clearly elucidated in humans.
Diagnosis
Classifications
Various clinical classifications of seizures have been proposed in the literature ( Table 7.1 ). Clinical classifications of seizures are based mainly on the description of their motor manifestations (focal clonic, multifocal clonic, tonic, myoclonic, and subtle). “Subtle” seizures are described as such because their signs may be difficult to recognize; the signs include abnormal eye movements, automatisms (e.g., tongue protrusion or lip smacking, swimming/pedaling/boxing movements), or autonomic phenomena (e.g., apnea, tachycardia, tachypnea, flushing).
| Author (Year) | Type | Characterization | Ictal EEG Discharges | |
|---|---|---|---|---|
| Volpe (1989) | Subtle | Ocular, oro-buccal-lingual movements, autonomic, repetitive stereotyped movements (i.e., pedaling, boxing, swimming) | Common | Uncommon |
| Variable/inconsistent | ||||
Clonic
| Repetitive rhythmic jerking | + + | ||
Tonic
| Stiffening, decerebrate posturing | + | + | |
Myoclonic
| Rapid, isolated jerks | + | + | |
| Lombroso (1996) | Subtle, minimal, fragmentary | Paroxysmal, stereotyped, periodic Mostly in newborns with severe CNS insult | Good or variable ictal correlation Interictal background patterns usually abnormal Absence of EEG ictal discharge does not necessarily rule out seizures | |
Clonic
| Migrating form limb to limb, alternate side Rare. Jacksonian march exceptional. | Consistent | ||
Tonic
| Stereotyped, often accompanied by autonomic changes Symmetric tonic postures Abrupt tonic limb abduction/adduction or extension/flexion → epileptic spasms | Frequently abnormal background δ/α/β discharges No ictal EEG correlate, abnormal background EEG discharges usually (not invariably) present | ||
| Myoclonic | Erratic, fragmentary or more generalized myoclonic jerks often associated with tonic spasms/multifocal tonic or clonic patterns/mixed seizure types | Burst-suppression pattern | ||
| Mizrahi and Kellaway (1987, 1998) | Focal clonic | Rhythmic muscle contractions. Unifocal/multifocal, Synchronous/asynchronous, Not suppressed by restraint. | Pathophysiology | |
| Epileptic | ||||
| Focal tonic | Sustained posturing, tonic eye deviation Not provoked, not suppressed by restraint | Epileptic | ||
| Myoclonic | Random, single rapid muscle contraction
| Epileptic/nonepileptic | ||
| Spasms | Flexor/extensor May occur in clusters Not provoked, not suppressed by restraint | Epileptic | ||
| Generalized tonic | Sustained symmetric posturing, may be stimulus-sensitive, may be suppressed by restraint | Nonepileptic | ||
| Motor automatisms | Ocular, oro-bucco-lingual, progression movements of limbs May be provoked or intensified by external stimuli and suppressed by restraint | Nonepileptic | ||
| Electrographic | None | Epileptic | ||
Focal Clonic Seizures
Focal clonic seizures are repetitive, rhythmic contractions of specific muscle groups. Clonic movements typically show a rate of repetition of 1 to 3 Hz with decreasing frequency and increasing amplitude with the passing of time and with increasing size of the involved muscle groups. They have a consistent electrocardiographic (EEG) correlate. Compared with paroxysmal nonepileptic phenomena such as clonus or tremor, they tend to be slower and more rhythmic. Most importantly, they cannot be stopped by restraint of the involved body part. They can be unifocal or multifocal, alternate between involved sites, or be simultaneous but asynchronous, different from what is usually observed in older age groups.
Focal Tonic Seizures
Focal tonic seizures are characterized by sustained asymmetric posturing of the limbs or trunk or by tonic deviation of the eyes. They are typically associated with focal EEG discharges.
Myoclonic Jerks
Myoclonic jerks in newborns can be either epileptic or nonepileptic in origin. The movements have a brief, jerky, shocklike appearance. Speed is influenced by the size of the muscle group involved. They can involve different districts and be fragmentary, focal, or multifocal. They can be isolated or repetitive, in this case with a slow, erratic, or irregular rate of recurrence (differentiating them from clonic events). Some forms of myoclonic jerks may occur with consistent EEG seizure discharges, although some do not. Finally, they can be spontaneous or provoked by stimulation.
Autonomic Signs
Autonomic changes such as alterations in heart rate, respiration rate, and blood pressure; flushing’ salivation; or pupil dilatation have been described in seizures, although they are rarely isolated and typically associated with motor phenomena.
Spasms
Epileptic spasms may occur in neonates. Although they are not seen in neonates with seizure symptomatic of acute events such as hypoxic-ischemic encephalopathy (HIE) or stroke, they are the main seizure type in Ohtahara syndrome (see later text). They mainly involve truncal and upper limb muscle groups and can be divided into extensor, flexor, or mixed types.
Subclinical Seizures
Seizures may present without overt clinical manifestations, especially in critically ill infants. This phenomenon is possibly linked to the involvement of non-eloquent areas or rather to the involvement of nonmotor areas. It might also be that, in some instances, the high figures of electrographic-only seizures are secondary to ascertainment methods such as the duration of monitoring; the use of a multiple-camera approach compared with a single camera synchronized to the EEG was associated with improved detection of subtle facial motor components to seizures. On the other hand, it is interesting to note that exclusively electrographic neonatal status epilepticus (NSE) seems to be exceptional, as we found that in our cohort of preterm newborns, in all cases NSE became clinically evident at some point.
Definition of Seizures
Based on the high rate of subclinical seizures, the need for differential diagnosis with various paroxysmal nonepileptic phenomena and the unreliability of clinical diagnosis of seizures even by experienced personnel, a definition of seizures based only on clinical grounds is considered inaccurate. Therefore the current definition of neonatal seizures is based on EEG criteria. A seizure is defined as a sudden, repetitive, stereotyped episode of abnormal electrographic activity with peak-to-peak amplitude of at least 2 μV, a minimum duration of 10 seconds, and evolution with a clear beginning, middle, and end, with or without a clinical correlate. Conventionally, seizures need to be separated by at least 10 seconds to be considered as distinct, even if this cutoff has not been based on physiologic considerations.
Definition of Neonatal Status Epilepticus
NSE has been variably defined in the literature. Its most widely accepted definition is continuous seizures lasting for more than 30 minutes or seizure present for at least 50% of the recording time, with no return to the baseline neurologic condition. Alternative operational definitions with a shorter time cutoff have also been proposed, even if a specific threshold has yet to be found. For this reason, the literature has also referred to “seizure burden” to indicate the percentage of the EEG recording with ictal activity.
Controversies in Definition and Classification of Seizures
The current EEG definition of seizures poses some issues related to its applications to seizure types characteristically shorter than 10 seconds in duration, such as myoclonic seizures and spasms.
Myoclonic seizures are considered truly epileptic in origin if they are time-locked to an epileptic discharge on EEG that precedes the motor event (as detected on the electromyogram trace) by typically 20 to 40 msec. The gold standard for the confirmation of the epileptic nature of myoclonic phenomena is represented by back-averaging.
Similarly, in the case of epileptic spasms, which are rare but possible in newborns, the “diamond-shaped” potential on the electromyograph (typical duration of 1–2 seconds) correlates with typical, although variable, EEG discharges. In addition, newborns typically have more than one seizure type.
Differential Diagnosis
For electrographic discharges, the two main differential diagnoses are seizures with artifacts and those with brief rhythmic discharges. A wide range of artifacts of both biologic and environmental origin can be detected, especially in the high–background noise situation of neonatal intensive care units (NICUs). A description of such artifacts and the methods used to discriminate them from real electrographic discharges is beyond the scope of this chapter. Brief rhythmic discharges are, by definition, electrographic discharges shorter than the duration required for an EEG definition of seizures. They have been associated with the occurrence of seizures and with background EEG abnormalities.
Clinically the main differential diagnosis for seizures is with paroxysmal nonepileptic motor phenomena, which can be either physiologic or abnormal. There is a wide range of paroxysmal nonepileptic motor phenomena that occur in newborns, including tremor and jitteriness, benign neonatal sleep myoclonus, startle reflex, ocular movement disorders (i.e., paroxysmal tonic upgaze and downgaze, opsoclonus), paroxysmal dystonia, bilateral tonic stiffening, and hyperekplexia. The occurrence of these events is facilitated by the immaturity of corticospinal tracts in newborns.
A number of features can assist in the differential diagnosis versus epileptic events in some, although not all, of these phenomena. These include stimulus sensitivity (e.g., different stimuli, especially auditory, for startle reflex; crying and stress for tremors and jitteriness; sudden visual stimuli or movement for paroxysmal tonic up/downgaze; rocking or repetitive auditory stimuli for benign neonatal sleep myoclonus), habituation (present in startle reflex but absent in hyperekplexia), and association with behavioral states (benign neonatal sleep myoclonus occurs only during sleep, mainly in quiet sleep, and stops with arousal). Holding the affected limbs can differentiate some paroxysmal motor phenomena from epileptic ones: gentle restraint stops physiologic tremor and may exacerbate benign neonatal sleep myoclonus, while it does not modify epileptic events. Paroxysmal nonepileptic events can be situational—for example, paroxysmal dystonic events in Sandifer syndrome are brought on by feeding. The use of polygraphic video-EEG recording is often crucial for correct diagnosis in many of these conditions and is considered a cornerstone for the discrimination between true epileptic seizures and, for example, brainstem release phenomena. However, seizures originating from subcortical regions or from deep, mesial cortical areas and not propagating to the cortical surface may not be detected by surface EEG electrodes and remain an open issue.
Diagnostic Tools and Monitoring
When seizures are suspected because of the presence of risk factors (i.e., fetal distress, CNS infection, HIE, preterm birth, intracranial hemorrhage, cardiac surgery) or because of suspicious clinical events, a video-EEG recording is necessary for correct interpretation of clinical phenomena to avoid the prescription of unnecessary anticonvulsant therapy. Based on the high rate of electrographic-only seizures (which would otherwise go unrecognized), initiation of video-EEG monitoring is recommended and should be continued for a period of 24 hours after seizure cessation according to the American Clinical Neurophysiology Society. When the indication for EEG is the differential diagnosis of a clinical event, its duration should primarily depend on the ability to record multiple typical clinical events.
Once seizures have been diagnosed, prolonged or continuous monitoring is necessary to correctly estimate the seizure burden, to monitor the efficacy of anticonvulsant medications, and to assist in differential diagnosis with the uncoupling phenomenon. Monitoring is also recommended during and after drug discontinuation and during therapeutic hypothermia and rewarming in newborns with HIE, given the risk of seizure occurrence in this population.
Furthermore, conventional EEG provides relevant information for prognosis based on background activity. Correct evaluation of background activity requires the recording of wakefulness and active and quiet sleep, usually corresponding to at least 1 hour of recording, which might need to be prolonged in case of disruption of sleep rhythms—for example, after the acute phase of encephalopathy.
An alternative strategy for brain function monitoring is represented by amplitude-integrated EEG (aEEG). aEEG is an EEG-based bedside brain monitoring instrument commonly used in NICUs. Current practice implies use mainly by neonatologists and nurses. It displays one or two channels of EEG data after filtering, rectification, and smoothing on a semilogarithmic scale, enabling a reduction in the time needed both for electrode placement and diagnostic interpretation. It is especially used in the context of neonatal HIE for infants undergoing therapeutic hypothermia. Studies confirm the reliability in assessing background activity and degree of HIE with good correlation with outcome. Nevertheless, aEEG is characterized by poor reliability if reported by unexperienced personnel compared with continuous EEG, especially for seizures monitoring. aEEG filters out most of the slow activity and uses a time-compressed scale and a limited set of electrodes, with a potential risk of underdiagnosis, especially for seizures arising outside the centrotemporal regions or seizures of short duration. This might be a relevant issue in preterm newborns, as some studies report a higher prevalence of seizures from the frontal or occipital regions, especially under 29 weeks of gestational age (WGA). Whereas conventional EEG is characterized by high sensitivity, reliance on aEEG alone might be misleading in preterm infants because seizures tend to propagate less. A risk of overdiagnosis also exists, because the long monitoring periods and the high background noise typical of ICUs might facilitate the occurrence of false-positive findings secondary to artifacts. Therefore for the diagnosis of seizures, aEEG is considered a screening tool. Consequently, evaluation of the raw EEG trace on the digital aEEG devices is often required and recommended to recognize and quantify seizures as it allows higher sensitivity and specificity. Table 7.2 briefly summarizes the main aims of cEEG monitoring in newborns with suspected seizures or at risk of seizures.
| Neonatal Seizures (Suspected/Confirmed) | High-Risk Newborn |
|---|---|
| Differential diagnosis (including verification of aEEG patterns) | Evaluation of background activity |
| Diagnosis of NSE | Diagnosis of associated seizures or NSE (especially electrographic-only) |
| Treatment assessment | Prognostication (with serial recordings) |
| Diagnosis of uncoupling | Detection of focal cerebral injury |
EEG Characteristics of Neonatal Seizures
Seizures typically last between 1 and 5 minutes and most last less than 3 minutes. Seizure duration has been reported to be longer in full-term newborns owing to a higher incidence of NSE. These data, based on the use of standard cEEG, were later confirmed by continuous EEG monitoring, although NSE (seizures lasting 30–40 minutes) were also frequent.
Considering the onset focus and the region involved at the time of the maximal spread, seizures are divided into localized or lateralized or unifocal, multifocal, and bilateral independent. A focal onset is present in the majority of cases, especially in full-term newborns, whereas the onset in preterm newborns can be either regional or focal. In the majority of cases, seizure onset is from the centrotemporal or vertex region. However, the presence of multiple independent areas of discharge at seizure onset is not unusual in the newborn period. Typical features of seizures in newborns include migration (defined as a change of seizure focus within one hemisphere) and shifting (defined as a change of focus from one hemisphere to the other, termed “flip-flop” when discharges finally shift back to the original hemisphere of ictal discharges). According to some authors the spread of discharges is mainly ipsilateral for seizures with a focal onset and contralateral or bilateral for regional-onset seizures. Description of the seizure spread might enable improved correlation with evolving clinical signs in term infants, even though previous reports failed to clearly describe such correlation in preterm newborns. In addition, the spread to the contralateral hemisphere has been associated with outcome. Ictal discharges in newborns tend to be characterized as follows: focal spike/sharp wave discharges, focal low-frequency discharges, focal rhythmic discharges, and multifocal discharges. Repetitive spikes, sharp waves, stereotyped wave complexes, and α, θ, or δ discharges are also frequent. Different patterns can be associated in single patients in one-third of cases. It is also possible to record different seizures occurring simultaneously over different areas of both hemispheres. Spike potentials often show slower rate and longer duration in neonates (“burn-out spikes”) compared with older patients. During the ictal discharge, delta and sharp waves are among the most frequent findings in both full-term and preterm infants. Evolution of discharges during the seizure usually consists of a decrease in frequency and an increase in amplitude. The ability to generate well-formed spikes at high discharge rates tends to improve with increasing conceptional age (CA) but is impaired in case of severe brain injury. Nonetheless, in severely affected neonates seizures might lack a clear field and pattern of evolution, with a prevalence of low amplitude and slow frequency. In these patients, seizures also tend to be prolonged and to remain confined to one single EEG channel or to alternatively change hemispheres.
The impact on brain physiology or on subsequent outcome of different ictal patterns is still unknown, although α discharges have been associated with worse prognosis. No correlation between seizure characteristics and etiology has been reported. Furthermore, there is no clear association between clinical semiology and characteristics of EEG discharges. However, in a cohort of preterm newborns Okumura and colleagues found that the duration of seizures was relatively shorter when motor phenomena were present compared with apneic or electrographic-only seizures. There is a tendency for neonates with severely abnormal background to present with electrographic-only seizures.
Standard EEG Versus Long-Term or Continuous Monitoring
Continuous conventional long-term EEG monitoring (cEEG) is now considered the gold standard for seizures diagnosis.
Controversies in Monitoring
- •
Electrographic-only seizures: Owing to the high incidence of subclinical seizures and the caveats for clinically based differential diagnosis of seizures, continuous electroencephalographic (cEEG) monitoring is required.
- •
Identification of high-risk newborns: Infants with acute brain insult, including HIE and stroke, and/or suspected seizures on clinical observation, or neonates with encephalopathy of unknown origin.
- •
Higher seizure burden period in full-term versus preterm newborns: The onset of electrographically confirmed seizures tends to be earlier in full-term than preterm newborns. As recently shown by Lynch and colleagues, in full-terms infants with HIE the seizure burden tends to be maximal at a mean of 22.7 hours of life, and the last seizure is recorded at a mean of 55.5 hours.
- •
Seizures in preterm newborns: Among preterm infants, seizures occur in the first 48 hours after birth in approximately 10% of infants born below 29 weeks CA and in 50% of infants born at 30 weeks or later.
Prolonged EEG monitoring in full-term newborns is recommended in the context of encephalopathy, including HIE, during hypothermia and rewarming, and in neonates with suspected seizures, while in preterm infants it is based on gestational age and on the existence of specific risk factors, based mainly on the major etiologic role of intraventricular hemorrhage and its complications. The optimal duration and timing of monitoring in other etiologic groups are less clearly defined.
Management
Investigations and Etiologies
Seizures may have many different underlying etiological factors. A list of the main diagnostic investigations that might be required in the diagnostic workup of the newborn with seizures is reported in Table 7.3 . Some causes of seizures have been diminishing—for instance, transient metabolic disorders and neonatal traumatic lesions, infections—owing to the marked advances in neonatal care, whereas other etiologies, such as stroke and genetic etiologies, are growing in number and are likely due to improved early recognition with widespread use of prolonged video-EEG in the NICU and more sophisticated testing. Overall, the past few years have seen a major improvement in recognition and definite diagnosis and a reduction of the number of newborns with seizures of unknown etiology.
| History (identification of newborn at high risk for neonatal seizures) | Family history, pregnancy history; delivery history; feeding history; clinical characteristics of the events |
| Physical examination | Physical examination (dysmorphic features) Neurologic examination (signs of encephalopathy) Fundus oculi examination. |
| Investigations | First-line investigations: complete blood cell count; glycemia; arterial pH; calcemia, electrolytes, urea and creatinine; urine microscopy and culture; blood lactate and pyruvate; ammonemia; blood culture; lumbar puncture including CSF lactate, CSF cells and protein, prolonged video-EEG, head ultrasounds Second-line investigations: screen for congenital infections; blood, urinary and cerebrospinal amino acids evaluation; urine organic acids; plasma acyl-carnitines; urine sulfite; Very long chain fatty acids; urinary organic acids, MRI, genetic testing. |
HIE still represents the major cause of seizures in neonates, especially in term infants. Usually seizure onset is between 6 and 24 hours after birth. Seizures are usually short and isolated at onset and progressively become more frequent and prolonged, mainly between the 12th and the 24th hours of life, and are often resistant to antiseizure therapy. The advent of hypothermia as the standard of care for newborns with HIE has partially changed the profile of seizures in these newborns. First, the typical time evolution found in normothermic newborns with HIE seems to be absent in newborns treated with hypothermia, where no consistent pattern can be recognized. The postnatal age of first recorded electrographic seizure is found to be similar in cooled and normothermic newborns by some authors and to be later by others (in the first 48 hours in 76% of cases). The use of hypothermia as standard of care has led to a significant reduction of seizure burden in cooled neonates with moderate HIE, possibly contributing to improved outcome. Glass et al., who used continuous EEG monitoring, detected seizures in approximately 50% of newborns in the majority of cases during the cooling period. Interestingly, early EEG background was the best predictor of seizure occurrence, even superior to the clinical signs of encephalopathy. Therapeutic hypothermia has been found to be associated with a lower incidence of seizures in the first 6 months of life after discharge (16% vs. 53%), although the short follow-up likely limits the significance of these findings. The rewarming period is also considered at risk of breakthrough seizures in this population, although there are few data to support this assumption. The association between the occurrence of seizures and the presence of brain injury, already found in noncooled neonates with HIE, has been confirmed in cooled neonates. Moderate to severe brain injury (as detected by magnetic resonance imaging [MRI] performed at a median of 5 days of age) is more common in newborns with seizures than in those without (irrespective of the presence or absence of a clinical correlate to discharges) and is universal in the subgroup with status epilepticus. Seizures are also more likely to occur later and to be resistant to phenobarbital load in the more severely affected newborns. The transient metabolic disorders that can be associated (hypoglycemia, hypocalcemia, hyponatremia) can be responsible for the refractoriness of the seizures. Novel therapeutic neuroprotective strategies (preventive, rescue, and repair) in association with hypothermia are being developed, aiming to obtain a better outcome, particularly in the severe cases in which rescue cannot currently be achieved.
Intracranial hemorrhages represent another important cause of seizures in neonatal age, and intraventricular hemorrhage of grade III or IV is the most frequent etiology in preterm newborns. Almost 90% of intracranial bleeding develops within the first 3 days of life, with an incidence higher in the early gestational ages. Strokes are frequently reported in newborns with seizures (10%–18% of seizures), especially in recent years, probably owing to our better ability to diagnose them with improved neuroimaging techniques. Strokes should be suspected in newborns with (1) focal or multifocal clonic seizures that begin at 12 hours to 3 days of life, (2) normal or almost normal interictal neurologic examination findings, (3) normal Apgar scores at birth, and (4) with no evidence of encephalopathy. Subarachnoid and subdural hemorrhages are rare and are often due to obstetric trauma and occasionally can lead to seizures; however, they have a good prognosis. CNS infections such as meningitis and/or encephalitis are responsible for 3% to 9% of neonatal seizures, particularly from viruses of the TORCH (toxoplasmosis; other [congenital syphilis and viruses], rubella, cytomegalovirus, and herpes) complex, whereas bacterial agents are more likely to be involved are group B Streptococcus, Listeria, and Escherichia coli . Herpes simplex viruses usually determine a diffuse brain damage with a specific EEG pattern (diffuse slow background with periodic complexes, sharp and slow waves over the frontal and temporal regions, which, nevertheless, are only present in 50% of cases). A lumbar puncture should be performed in any neonate with seizures and signs of infection. Cerebral malformations can be responsible for 3% to 9% of the epileptic seizures in the first days of life. Most have a genetic basis and are associated with an adverse outcome. Seizures are also observed in 2% to 11% of infants with neonatal abstinence syndrome, together with tremors, irritability, excessive crying, and diarrhea. The spectrum of abstinence syndrome has changed because in the 1970s it was generally secondary to the use of opioids, whereas more recently an increased use of multiple opioids has been observed to be complicated by the abuse of several other licit and illicit substances.
Among the transient metabolic alterations, hypoglycemia, particularly when glycemia is less than 40 mg/dL, is one of the most common causes of seizures; it can be either isolated or more frequently is associated with congenital metabolic defects of glucose metabolism or fatty acid β-oxidation or with pancreatic lesions, causing early-onset, prolonged and severe convulsions, and high mortality and morbidity. Another common condition is early hypocalcemia occurring between the second and the third day of life, or the late form at the end of the first week or the beginning of the second week. This condition can be controlled by calcium and today is easy to cure. Hyponatremia/hypernatremia are uncommon etiologies that can result from inappropriate secretion of antidiuretic hormone, congenital adrenal dysfunction, or inappropriate management of intravenous fluids. Neonatal hypernatremia can be seen in breastfed infants with excessive weight loss, dehydration, and poor feeding. Electrolyte abnormalities may be caused by HIE, intracranial hemorrhage, or meningitis.
Seizures during the neonatal period caused by inherited metabolic diseases (IMDs) are rare and they can provoked by several mechanisms: cerebral energy failure, toxic effects of metabolites, neurotransmitter modification–associated brain malformations, and vitamin defects. A suspicion of IMD must arise when the newborn is severely encephalopathic and shows a severe disorganization of the EEG background activity, early onset of several seizure types (sometimes within the first hour of life, sometimes beginning antenatally), myoclonic seizures that are usually drug-resistant, and sometimes with suggestive MRI findings (such as signs of brain atrophy or of hypoxic-ischemic injury in the absence of a hypoxic insult at delivery). Other neurologic signs such as abnormal movements, tremor, hyperexcitability, hypotonia, and chronic hiccup (nonketotic hyperglycinemia) can also be associated. This clinical scenario should alert the physician to an IMD, and the infant should be immediately evaluated with analysis of known biomarkers ; an early dietary restriction or supplementation of vitamin trials should be started, even before establishing a definitive diagnosis that might require time, because it is important in those treatable conditions to avoid permanent brain damage. The causes more often identified are nonketotic hyperglycinemia, pyridoxine-dependent epilepsy, pyridoxamine 5′-phosphate oxidase deficiency, sulfite oxidase deficiency, molybdenum cofactor deficiency, glucose transporter type 1 deficiency, congenital lack of glutamate transporter, and mitochondrial disorders.
Infants with brain malformations can rarely present with neonatal seizures because they usually will develop seizures in the postneonatal period. Most are related to cortical migration disorders such as lissencephaly, pachygyria, polymicrogyria, cortical dysplasia, schizencephaly, and heterotopia. Some of these disorders may be seen in congenital infection or in IMD and there is increasing evidence of a genetic basis in the majority. The prognosis is generally poor.
Neonatal epilepsy syndromes are rare and their genetically molecular mechanisms are being better delineated owing to the recent advances in genetics. They may vary in their clinical findings from a benign condition to neonatal-onset epileptic encephalopathies characterized by the association of epilepsy and cognitive deficit.
There are also epileptic syndromes with onset during the neonatal period. These are benign familial neonatal epilepsy, early myoclonic encephalopathy, and early epileptic neonatal encephalopathy, although more etiology-specific syndromes have being recognized in the past few years.
Neonatal seizures are a critical clinical condition often sustained by an acute brain injury; the most common etiologies are HIE, intracranial hemorrhage, and ischemic stroke. However, in a few cases the etiology still remains unknown even despite progresses in neuroimaging techniques and in the genetic analysis methodologies. In the most severe clinical scenarios, including the early-onset neonatal epilepsies and epileptic encephalopathies, repeated genetic testing may identify new mutations even later in life, possibly suggesting interesting diagnostic and therapeutic perspective. In other cases, a second brain MRI later in life with higher resolution properties may discover small causative structural lesions.
Therapy
Based on the unavailability of strong evidence-based data from clinical studies, existing guidance for the pharmacologic management of seizures in the neonate is currently scanty. In fact, the majority of papers on the management of seizures are retrospective chart reviews or case reports, and the low number of participants reduces the generalizability of results. Only three randomized controlled trials (RCTs) have been published so far.
The World Health Organization elaborated a guideline for neonatal seizures in 2011, and a proposal for seizure management in low-income countries was published in 2007. Flow-charts and protocols have been published in various review articles.
The strong recommendation of the use of phenobarbital as the first-line drug in seizures, despite very low-quality evidence, derives from the availability of RCT studies and from longstanding, widespread use in clinical practice. Although the reported efficacy of phenytoin is not significantly different from that of phenobarbital, the former is usually administered as a second-line treatment rather than as a first-line one owing to the risk of side effects and unease of use, its link to poorly predictable blood concentrations, and difficulties in titration. The suggested third-line treatment is usually represented by benzodiazepines, notably midazolam. However, advocated treatment strategies after phenobarbital or phenytoin failure are based on very low-quality evidence. Adjusted dosing regimens for lidocaine have been reported to be associated with a lower rate of cardiac adverse events than previously reported. Lidocaine has been included in a small RCT comparing it with benzodiazepines.
Recommendations from treatment guidelines are partially reflected by clinical practice. According to hospital-based studies, phenobarbital is the most common initial loading dose and the most used medication, with no differences between preterm and full-term newborns and no dosage changes based on gestational age. However, the second most commonly prescribed medication was levetiracetam, while phenytoin was the least likely to be prescribed. The increasing use of levetiracetam is motivated by the availability of intravenous and enteral preparations and reassuring data on its safety and efficacy, even though it represents an off-label prescription and data on its effectiveness are still limited and of poor quality (RCTs are not available).
Efficacy and Safety of Antiseizure Medications for Seizures
The response rate to first-line anticonvulsant medications used for the treatment of acute symptomatic seizures is suboptimal. No clear-cut differences in response seem to exist between preterm and full-term newborns, even if some age or etiology dependency might exist. For example, the administration of lidocaine determined a lower response rate in preterm newborns compared with full-term newborns.
Concern has been raised about the possible long-term detrimental side effects of anticonvulsant medications, especially phenobarbital. However, the only clinical data are translated from the negative effects on intelligence quotient (IQ) reported in cohorts of children with febrile seizures, although experimental data on animal models seem to confirm an association between phenobarbital administration in the neonatal period and neuronal apoptosis in the developing rat brain and an interference with maturation of synaptic connections. Described long-term effects in animals (exposed to very high loading doses of phenobarbital) include schizophrenia-like behavioral abnormalities and impaired learning, memory, and social interaction. A retrospective clinical study of infants receiving phenobarbital for seizures (adjusted for the number of electrographic seizures and gestational age) and evaluating the effect of cumulative exposure to levetiracetam and phenobarbital found a greater association with worse Bayley Scales of Infant Development cognitive and motor scores with phenobarbital compared with levetiracetam and a higher rate of cerebral palsy in the group treated with phenobarbital but not with levetiracetam. In immature animal models, doses of phenytoin higher than the usual loading dose in humans have been experimentally associated with apoptosis and synaptic disruption. Data for levetiracetam are controversial, showing both a possible neuroprotective effect and detrimental effects (apoptosis, increased brain injury).
Phenobarbital and Phenytoin
The two RCTs comparing phenobarbital and phenytoin report the following percentages of response: 43% of responders to phenobarbital as first-line treatment compared with 45% with phenytoin; nonresponders had similar results once switched to the alternative medication (total response of 57% in those receiving phenytoin as second-line treatment versus 62% in those receiving phenobarbital as second-line treatment). The second open-label, cross-over study found a 14.5% response with phenytoin as first-line treatment, increasing to 80% when switched to phenobarbital as second-line treatment, while the second arm had a 72.2% response with phenobarbital, increasing to 96.3% with phenytoin. However, these studies have relevant limitations, including a clinically based diagnosis of seizures and the absence of EEG monitoring, that might have resulted in underestimation of the seizure burden or even a potential for misdiagnosis.
Pharmacokinetic data in newborns are better known for phenobarbital compared with other medications, as they have been investigated in various papers. Neither weight nor gestational age seems to significantly influence dose-related blood levels. However, infants of less than 30 weeks’ gestational age or weighing less than 1500 g might need lower doses to achieve the same blood levels. As protein levels are lower in newborns than in older ages, the free fraction of drugs is expected to be higher and might be altered by organ dysfunction, leading to toxic side effects, especially respiratory and cardiovascular.
Difficulties in the management of phenytoin (independent of gestational age) consist of the need for frequent monitoring of blood levels, erratic absorption when administered enterally, the need for frequent dose administration (every 6–8 hours), and ongoing metabolism changes in infancy.
Midazolam
Low-quality evidence from clinical studies limited by methodologic issues, including study design and sample size, report a highly variable efficacy of midazolam used as a second- or third-line agent (0%–100%).
Lidocaine
The reported efficacy of lidocaine varies between approximately 10% and 78%. Interestingly, a recent retrospective cohort study of 413 newborns found an increasing response rate when given as third-line as opposed to second-line treatment (67.6% vs. 21.4% in term newborns and 40.7% vs. 16.4% in preterm newborns), possibly highlighting the importance of the timing of treatment.
In the past, cardiotoxic concentrations were found in the majority of neonates receiving lidocaine with standard lidocaine infusion, with the risk of cardiac adverse events increasing with concentrations higher than 9 mg/L. Therefore new dosing regimens were proposed, with resultant improved tolerability.
Levetiracetam
Available clinical studies on levetiracetam (mainly as second- or third-line treatment) have a weak design, limiting the value of data on efficacy. There are no RCTs; studies are mainly retrospective (only two prospective studies have been published to date ) or case reports. The reported effectiveness is variable, ranging from 35% to 86%, although differences (and potential biases) in study design, timing, and method of ascertainment of response limit the quality of the evidence. Additionally, the mechanism of action is not fully known. Because of the increased volume of distribution in neonates, it might be possible that newborns require higher loading doses; however, elimination half-life seems to be longer than in older patients owing to the immaturity of renal function. Clearance of levetiracetam in preterm infants is lower than in older children (half-life: 9 hours compared with 5–7 hours). Safety-wise, its use has not been associated with severe or life-threatening adverse events.
Topiramate
There are few studies on the use of topiramate in newborns. Its use is limited by the unavailability of a formulation for intravenous use. Hypothermia increases its half-life and concentrations. The data on efficacy and safety of the main anticonvulsant medications used in newborns are summarized in Table 7.4 .
Related posts:
Cerebral Circulation and Hypotension in the Premature Infant: Diagnosis and Treatment
Hypothermia for Neonatal Hypoxic-Ischemic Encephalopathy: Different Cooling Regimens and Infants Not Included in Prior Trials
General Supportive Management of the Term Infant With Neonatal Encephalopathy Following Intrapartum Hypoxia-Ischemia
Neonatal Meningitis: Current Treatment Options
Hyperbilirubinemia and the Risk for Brain Injury
Amplitude-Integrated EEG and Its Potential Role in Augmenting Management Within the NICU
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
