MANUSCRIPT CITATION

Shellhaas RA, Wusthoff CJ, Tsuchida TN, Glass HC, Chu CJ, Massey SL, Soul JS, Wiwattanadittakun N, Abend NS, Cilio MR; Neonatal Seizure Registry. Profile of neonatal epilepsies: Characteristics of a prospective US cohort. Neurology. 2017 Aug 29;89(9):893-899. PMID 28733343

REVIEWED BY

Amanda G Sandoval Karamian, MD
Department of Neurology, Division of Child Neurology
Stanford University School of Medicine, Palo Alto, CA

Fiona M Baumer, MD
Department of Neurology, Division of Child Neurology
Stanford University School of Medicine, Palo Alto, CA

TYPE OF INVESTIGATION

Etiology

QUESTION

What proportion of newborns with seizures have neonatal onset epilepsy, and what are the clinical characteristics, etiologies, and initial treatments prescribed for neonatal epilepsy?

METHODS

• Design: multicenter prospective cohort study enrolling all consecutive newborns diagnosed with seizures
• Allocation: not applicable
• Blinding: not applicable
• Follow-up period: Enrollment of neonates occurred between January 2013 and November 2015 with the clinical course of the neonates followed during their hospitalization through discharge or transfer.
• Setting: Seven neonatal neurology programs with level IV neonatal intensive care units in the United States form the Neonatal Seizure Registry. Sites were selected based on their adherence to the American Clinical Neurophysiology Society (ACNS) guideline on neonatal EEG monitoring.
• Patients: The authors reviewed all 611 neonates enrolled in the Neonatal Seizure Registry and compared neonates with acute symptomatic seizures (n=532) with those with seizures due to neonatal epilepsy (n=79). Neonates were considered to have epilepsy if they met any of the following: 1) a clinical diagnosis of neonatal epileptic encephalopathy (no developmental brain abnormality, history incompatible with benign neonatal seizures or BFNE, and lack of primary acute symptomatic etiology); 2) a congenital brain malformation as the seizure etiology; 3) a clinical diagnosis of BFNE or benign neonatal seizures. Neonates with a diagnosis of epilepsy but comorbid illness that predisposed to acute symptomatic seizures were included in the neonatal epilepsy group. The authors then provided a more detailed clinical description of the neonates with epilepsy.
• Intervention: There were no study interventions as this was an observational cohort. Evaluation and treatment of neonates with epilepsy was at the discretion of the clinical teams and not mandated by a research protocol.
• Outcomes: Clinical characteristics, etiologies, treatments, and hospital outcomes
  • Primary outcome: Clinical characteristics of neonates with epilepsy compared with acute symptomatic seizures
  • Secondary outcomes: Etiologies of neonatal epilepsies, including genetic testing results; number and type of anti-seizure medications at time of leaving hospital; hospital outcome (died, discharged home, transferred to another hospital, hospice)
• Analysis and Sample Size: 611 neonates were consecutively enrolled, 79 (13%) of whom had neonatal epilepsies and 532 of whom had acute symptomatic seizures.
  • Comparative statistics: Chi squared, Fisher exact tests, t-tests and Kruskal-Wallis tests were performed to compare the clinical characteristics between two groups, including:
    • Neonates with epilepsy versus neonates with acute symptomatic seizures (sex, gestational age, indication for EEG, EEG duration, age at first clinical and EEG seizure, seizure burden, presence of subclinical seizures, number of antiseizure medications at discharge, age at discharge or death, hospital outcome)
    • Neonates with epileptic encephalopathy versus neonates with congenital brain malformations (comorbid acute symptomatic seizure etiology, hospital outcome)
    • BFNE versus neonates with epileptic encephalopathy or neonates with brain malformations (seizure burden)
  • Descriptive statistics for the three groups of neonatal epilepsies:
    • All groups
      • Number with genetic testing performed
      • Genetic testing results/etiologies
      • Type of genetic testing ordered
      • Antiseizure medications at time of discharge
      • EEG background classification
      • Seizure semiologies
      • Hospital outcome (death, home, transfer to another hospital, hospice, long-term care facility)
    • Epileptic encephalopathy
      • Clinical diagnosis
      • Trial of vitamin (pyridoxine, folinic acid, or pyridoxal-5-phosphate
      • Comorbid acute symptomatic seizure etiology
    • Congenital brain malformations
      • Type of malformation
      • Presence and type of extracerebral malformation
      • Comorbid acute symptomatic seizure etiology
    • Benign familial neonatal epilepsy (including the single case of benign neonatal seizures)
      • Clinical diagnosis
    • Patient follow-up: All 79 newborns with neonatal epilepsy as determined by the study site investigator were included in the above analyses and followed through hospital discharge or transfer.

MAIN RESULTS

The authors consecutively enrolled 611 neonates with seizures from seven level IV NICUs in the United States. The majority of these infants (n=523, 87%) had acute symptomatic seizures (i.e. HIE, infection, hemorrhage), but a sizeable minority (n=79, 13%) had infantile epilepsy, defined as an epileptic encephalopathy (n=35, 44%), congenital brain malformation (n=32%, 41), benign familial neonatal epilepsy (BFNE) (n=11, 14%), or benign neonatal epilepsy (n=1, 1%). The authors first compared infants with seizures due to an acute symptomatic etiology with those with seizures due to neonatal epilepsy. There were no significant differences between the sex, gestational age, seizure burden, and hospital disposition between neonates with epilepsy versus those with acute symptomatic seizures. EEG monitoring was most often initiated for clinical events in neonates with epilepsy versus for possible subclinical seizures in neonates with acute symptomatic seizures. Neonates with epilepsy were also less likely to have subclinical seizures, had a shorter duration of EEG monitoring, were older at first seizure, and were older at discharge/death with more AEDs at discharge.

The authors then delved specifically into the infants with neonatal epilepsy syndromes, both describing each of these conditions separately as outlined in the descriptive statistics above and in Table 1, and then comparing seizure burden, hospital outcome, and comorbid acute symptomatic seizure etiology between the syndromes. They found that 58 (73%) of the 79 newborns with neonatal epilepsy had genetic testing performed, with 34 (59%) of those tested with a putative genetic etiology.

Of the 35 neonates with epileptic encephalopathy, 29 (83%) had genetic testing performed and 24 (83%) of these had a genetic etiology identified, with KCNQ2 variants being the most common (n=10). An epilepsy gene panel was the most frequently used genetic test (n=23, 79%). The most common seizure semiology was subclinical (n=18, 56%) followed by focal clonic (n=11, 34%). The anti-seizure medications most commonly prescribed at time of discharge home or transfer to another hospital were phenobarbital (n=23, 73%) and levetiracetam (n=14, 44%).

Of the 32 neonates with congenital brain malformations, the most common malformations were neuronal migration abnormalities in 15 (47%). Seventeen (53%) also had extracerebral congenital abnormalities. Twenty-three (72%) had genetic testing performed, 6 (26%) of which revealed a genetic diagnosis. Chromosomal microarray was the most commonly sent genetic test in this group (20, 87%), though gene panels were more often diagnostic (3 out of 4 infants). Seizure semiology was most commonly focal tonic (n=20, 57%), followed by subclinical in 15 (43%). Phenobarbital (n=17, 81%) and levetiracetam (n=9, 43%) were again the most common anti-seizure medications prescribed at time of discharge home or transfer to another hospital.

Six (55%) of the 11 neonates with benign familial neonatal epilepsy (BFNE) had genetic testing, with 3 of these (50%) revealing a pathogenic KCNQ2 variant. Gene panels were again the most commonly sent test. The most common seizure semiologies included multifocal clonic in 4 (36%) and focal clonic in 3 (27%). Phenobarbital was prescribed to 8 (73%) and carbamazepine to 3 (27%) upon hospital discharge.  One additional neonate had benign neonatal seizures with no genetic testing performed.

In comparing the groups, the authors found that neonates with brain malformations were more likely to have a comorbid acute symptomatic seizure etiology than neonates with epileptic encephalopathy (10/32 vs 3/35, p=0.02). Neonates with brain malformations were also more likely to die (n=7) or be discharged to hospice care (n=4) than those with epileptic encephalopathies (3 died and 0 discharged to hospice care, p=0.01). Neonates with BFNE had lower seizure burden compared with those with brain malformations or epileptic encephalopathy (p<0.0001).

Table 1. Descriptive statistics of the three subgroups of neonatal epilepsy

*See table e-1 of original manuscript

CONCLUSION

The authors conclude that neonatal epilepsies are often caused by genetic etiologies and typically present with clinically apparent seizure events rather than the subclinical events frequently seen in acute symptomatic seizures. Increased use of genetic testing in neonates with epilepsy may reveal an even greater proportion of underlying genetic etiologies for neonatal epilepsy.

COMMENTARY

Neonatal seizures affect 1-5 per 1,000 live births. While most neonates have seizures secondary to an acute insult, some have a neonatal epilepsy syndrome, typically caused by a genetic mutation and/or a brain malformation (1,2,3). The evaluation and management of these patients is evolving, complicated by the rarity of these disorders.  Shellhas et al (2017) have made a significant contribution to the field by assembling one of the largest prospective cohorts of neonates with seizures, 79 of whom had epilepsy. The cohort was rigorously collected using consecutive enrollment and following ACNS guidelines for EEG-confirmation of seizures to standardize diagnosis (4). Practice variations across centers can make assessment of seizure burden challenging. The authors successfully quantified seizure burden with a novel categorical scale, though the clinical significance of these categories is not explained in the study.  Despite the impressive collaboration, certain neonatal epilepsy syndromes (i.e. vitamin-responsive epilepsies) were not captured, highlighting the need for larger studies of longer duration.

The study provides an interesting comparison between children with acute symptomatic seizures and those with neonatal epilepsy. The groups were similar in most ways, though neonates with acute symptomatic seizures had a longer duration of EEG monitoring and more subclinical seizures than those with epilepsy. These differences may have been driven by clinical guidelines that dictate all neonates with HIE undergo 72-96 hours of EEG monitoring. As different standards of monitoring were applied, there may have been underestimation of subclinical seizures in the epilepsy group due to this information bias.

The study provides a careful description of the hospital courses of neonates with epilepsy, including their seizure characteristics, work-up and treatment. Most interestingly, 59% of children with neonatal epilepsy who underwent testing had an identifiable genetic cause, including 83% of those with an epileptic encephalopathy. The detailed genetic testing results in table e-1 further highlight the diagnostic yield of such testing, with variants of uncertain significance in only 5/35 tested. However, even in this cohort taken from level IV NICUs, only 73% underwent genetic testing. The authors conclude that genetic testing is warranted to guide management and prognosis. In considering this conclusion, clinicians must weigh the limitations of this study, which include issues inherent to the observational study design. The clinical rationale behind genetic testing is not detailed in the paper, and it therefore cannot be determined if the untested and tested groups were comparable. Finally, this study does not address whether genetic information influences clinical decision making; for many genetic mutations, there is still little clinical consensus on management.

Despite these caveats, this work illustrates that genetic testing identifies the etiology of neonatal epilepsy in a substantial subset of the population. Diagnostic yield will increase as epilepsy gene panels and whole exome sequencing techniques evolve. Understanding the genetic basis for neonatal epilepsy may change the course of these illnesses, as there is growing evidence that specific genetic conditions benefit from specific medications (i.e. carbamazepine for KCNQ2 mutations) (5,6). Improved early seizure control may in turn lead to improved cognitive outcomes (3,7).

REFERENCES

1. Vasudevan C, Levene M. Epidemiology and aetiology of neonatal seizures. Semin Fetal Neonatal Med. 2013; 18(4):185-91.
2. Ramantani G. Neonatal epilepsy and underlying aetiology: to what extent do seizures and EEG abnormalities influence outcome? Epileptic Disord. 2013; 15(4):365-75.
3. Glass HC. Neonatal seizures: advances in mechanisms and management. Clin Perinatol. 2014; 41(1):177-90.
4. Shellhaas RA, Chang T, Tsuchida T, Scher MS, Riviello JJ, Abend NS, et al. The American Clinical Neurophysiology Society’s guideline on continuous electroencephalography monitoring in neonates. J Clin Neurophysiol. 2011; 28:611–617.
5. Hani AJ, Mikati HM, Mikati MA. Genetics of pediatric epilepsy. Pediatr Clin North Am. 2015; 62(3):703-22.
6. Sands TT, Balestri M, Bellini G, Mulkey SB, Danhaive O, Bakken EH, et al. Rapid and safe response to low-dose carbamazepine in neonatal epilepsy. Epilepsia. 2016; 57(12):2019-2030.
7. Kang SK, Kadam SD. Neonatal Seizures: Impact on Neurodevelopmental Outcomes. Front Pediatr. 2015; 3:101.