Abstracts

Rationale: Traumatic brain injury is associated with clinical and non-convulsive seizures. However, few studies reported the epidemiology of seizures in children with abusive head trauma. This study aimed to describe patient characteristics and to determine seizure prevalence and outcome in children with abusive head trauma.Methods: The Trauma Registry at Children’s Healthcare of Atlanta (CHOA) is a prospectively-collected and state-mandated database with requisite data fields. Mechanism of trauma is subcategorized into accidental, non-accidental, undetermined. From the trauma registry, 271 children less than or equal to 35 months old were determined as abusive head trauma by physical examination, neuroimaging, and determination of abuse by the Child Protection Team at Children’s Healthcare of Atlanta between 2009 and 2014. The medical records were reviewed further with using a standardized data collection form.Results: Median age was 4 months (0-35) on admission. All had head injury such as skull fracture with intracranial hemorrhage (ICH) (151, 55.7%), ICH only (72, 26.6%) and skull fracture only (48, 17.7%). One hundred fifty-nine patients (58.7%) required intensive care and 101 patients needed ventilator care. Thirty-four patients (12.6%) expired. Among 271, electroencephalography was done in 112 (41.3%) including 69 cases with continuous ICU EEG (cEEG) monitoring and 43 cases with routine EEGs. cEEG was started within 1 day (0-2) of admission and median duration of monitoring was 3 days (1-18). Among 69 patients with cEEG monitoring, 40 (58.0%) had seizures (Group A) and 29 did not have seizure (Group B) during cEEG monitoring. In Group A, 30 (75%) had electrographic seizures, 10 (25%) had both clinical and electrographic seizures, and 21 (52.5%) had non-convulsive status epilepticus. The first seizure occurred commonly in the first day of monitoring (24, 60%). Median day to seizure remission was 3 days (1-15). The maintenance antiepileptic drugs (AEDs) were commonly high-dose phenobarbital (13, 32.5%) or ≥ 2 AEDs of phenobarbital, fosphenytoin and levetiracetam (9, 22.5%). To treat non-convulsive status epilepticus, coma was induced by midazolam (9), pentobarbital (11) and both (1). In group B, pentobarbital induced coma was applied to control increased intracranial pressure in 8 patients (27.6%). The patients in Group B were more likely to have a history of cardiac arrest (0 out of 40 vs 4 out of 29), lower Glasgow coma scale on admission (6.59 vs 8.74), more need for disposition to long-term care hospital (0 out of 40 vs 2 out of 29) and higher mortality rate (2 out of 40 vs 10 out of 29).Conclusions: Electrographic seizures and non-convulsive status epilepticus are common in children with abusive head trauma. The outcome was poorer in patients without seizure in this cohort. This observation warrants further study.