Abstracts

Rationale: Abusive head trauma is a leading cause of disability and death in children. We evaluated the prevalence of non-convulsive seizures (NCS) in children with abusive head trauma, and hypothesized that NCS would occur in at least 50% of children. We explored indications for continuous EEG monitoring (cEEG), clinical and radiographic risk factors for seizures, and clinical outcomes. Methods: A retrospective study was conducted at a tertiary care hospital. Child Protection Team consult lists from 7/1/09 to 1/9/12 were screened to identify infants ≤2 years of age, evidence of head trauma by imaging or physical exam, determination of abuse by the Child Protection Team, and admission to the pediatric ICU (PICU). Clinical data was obtained from electronic medical records including presenting symptoms/signs, Glasgow Coma Scale (GCS), intubation status, EEG features including seizure occurrence and characteristics, imaging findings, PICU and hospital length of stay, mortality, and follow-up outcome. Data analysis was performed using Stata 10. Results: 32 abused children were identified with a median age of 4 months (interquartile range 3-5.5 months), and 21 underwent cEEG. Those monitored were more likely to have a lower admission GCS (8 vs. 15, p=0.053), more likely to be intubated (16 vs. 2, p=0.002), and have longer PICU (6 days vs. 1 day, p=0.001) and hospital stays (15 days vs. 6 days, p=0.002). Electrographic seizures occurred in 12 of 21 (57%). Status epilepticus occurred in 8 of 12 (67%) and seizures were entirely non-convulsive in 8 of 12 (67%). EEG background categories (discontinuous and slow/sedated sleep) were associated with presence of seizures (p=0.034). Subjects with seizures on EEG monitoring were not more likely to have clinically evident seizures on admission (67% with seizures vs. 33% without seizures, p=0.6), parenchymal imaging abnormalities (61% with seizures vs. 39% without seizures, p=0.4), or extra-axial imaging abnormalities (56% with seizures vs. 44% without seizures, p=0.72). Four of 21 (19%) patients died prior to discharge; none had seizures, but all had an attenuated or flat background. Among the 17 survivors, outcome data was available for 15 at a median duration of 10 months following PICU admission (interquartile range 6-16 months). For survivors, the presence of NCS was not associated with the Glasgow Outcome Score at follow-up (p=0.1). Conclusions: NCS are common in infants with abusive head trauma. Most seizures had no clinical correlate and thus would not be identified without cEEG. The associations between seizures and a higher GCS and more moderate EEG background abnormalities likely reflects that infants with the most severe brain injury and worse outcomes do not seize. In this small sample, neither clinically evident seizures on admission or parenchymal imaging abnormalities predict seizure occurrence during monitoring, although further study of seizure risk factors is needed. Further study is also needed to determine whether seizure identification and management improves outcome.