Clinical Professor of Pediatrics University of Michigan Ann Arbor, Michigan
This abstract will be presented during a Pediatric Highlights platform session
Rationale: People with epilepsy often have a history of neonatal seizures, but there are few prospective studies of risk factors for epilepsy after acute symptomatic neonatal seizures. Similarly, little is known about the range of epilepsies among survivors of acute neonatal seizures. We aimed to fill these gaps with a multicenter, prospective observational study. Methods: Infants with acute symptomatic neonatal seizures born July 2015 to March 2018 were enrolled at nine Neonatal Seizure Registry sites. Seizure etiology was determined by site investigators; infants with neonatal onset epilepsy syndromes were excluded. One-hour EEG was recorded at corrected age three months. Post-neonatal epilepsy and functional development (Warner Initial Developmental Evaluation of Adaptive and Functional Skills - WIDEA-FS) were documented serially at ages 12, 18, and 24 months. Unadjusted associations were assessed by chi square and Cox proportional hazards. Adjusted models used backwards, stepwise multivariable logistic regression. Results: Among 282 infants, 37 (13%) had post-neonatal epilepsy by age 24 months [median age of onset eight months (IQR 3, 16)]. Epilepsy types were: focal epilepsy (20/37, 54%), mixed epilepsy (both focal and generalized features; 7/37, 19%), and uncertain epilepsy type (10/37, 27%); 13/37 (35%) had infantile spasms. At last follow-up, 20/37(54%) children with epilepsy had been seizure-free for ≥6 months, and 12/37(32%) had < 1 seizure per month, but 4/37(11%) had daily seizures. Nine (24%) had treatment-resistant epilepsy ( >2 antiseizure medications -ASM- prescribed after NICU discharge). Most children with post-neonatal epilepsy had abnormal neurodevelopment (WIDEA-FS >2SD below normal population mean for 81% of children with epilepsy vs 27% without epilepsy, RR 7.9, 95% CI 3.6-17.3, p< 0.0005). Neonatal risk factors for subsequent epilepsy included: severely abnormal EEG background, higher number of days with EEG-confirmed seizures, need for ≥2 ASMs for neonatal seizure control, and abnormal discharge neurologic exam (Table). Risk of epilepsy for infants whose ASM was discontinued prior to versus after discharge from the neonatal ICU was similar (11% vs 14%, p=0.5, propensity adjusted OR 1.3, 95% CI 0.6-2.9, p=0.6).
In an adjusted model, days of neonatal EEG-confirmed seizures (OR 1.5 per day of seizures, 95% CI 1.1-2.0, p=0.007) and abnormal discharge exam (OR 5.1, 95% CI 2.3-11.0, p< 0.04) were independently associated with epilepsy risk. Among 122 infants with three-month EEG, four already had hypsarhythmia and infantile spasms. Excluding these four, infants with abnormal three-month EEG developed epilepsy sooner than those with a normal EEG (hazard ratio 3.3, 95% CI 0.7-14.5, p=0.11; Figure), but three-month EEG did not improve the multivariable epilepsy risk model. Conclusions: In this multicenter study of infants who survived acute symptomatic neonatal seizures, 13% had post-neonatal epilepsy by age 24 months. Though half were seizure-free for ≥6 months at last follow-up, 80% had clearly abnormal development. Days of EEG-confirmed neonatal seizures was a potentially modifiable risk factor for post-neonatal epilepsy, but follow-up EEG did not improve risk prediction. These findings have important implications for family counseling and future studies designed to prevent post-neonatal epilepsy and optimize neurodevelopment. Funding: Please list any funding that was received in support of this abstract.: PCORI CER-1507-31187 and Pediatric Epilepsy Research Foundation Click here to view image/table