Abstracts

Rationale: Traumatic brain injury (TBI) is diagnosed in 475,000 children in the United States per year. The incidence rates of post-traumatic epilepsy (PTE) in pediatric cohorts have varied across studies, in part due to small sample sizes, reliance on referral populations, and variable duration of follow-up. We aimed to determine the incidence rates and predictors of PTE within a Northern California population-based cohort of children who sought medical care after TBI.Methods: We conducted a retrospective population-based study of children enrolled within the Kaiser Permanente Northern California Medical Group. A cohort with TBI was identified through electronic searches and confirmed by retrospective chart review. PTE, defined as epilepsy onset more than 7 days after the index trauma, was identified within this cohort through electronic searches and diagnosis confirmed by chart review. Demographics, trauma history, clinical findings, EEG, and neuroimaging data were then abstracted from medical records. PTE incidence was determined by survival analyses and generalized linear modeling. For each case with PTE, one control without PTE was randomly selected from the TBI cohort. Odds ratios (OR) and 95% confidence intervals (CI) for predictors of PTE were calculated using conditional logistic regression.Results: From a study base of 1.3 million children, we identified a cohort of 29,589 children with TBI (at a median age of 12.0 years, interquartile range 5-16) who were followed for a total of 228,755 person-years. PTE developed in 186 children, for a cumulative incidence rate at 1, 5 and 10 years of 0.2%, 0.5%, and 0.8%, respectively. PTE followed a binomial age distribution, with relative risk highest amongst children who suffered a TBI under 5 years of age and over 15 years of age (p<0.05). Injury severity, determined by abbreviated injury score (AIS) and Glasgow coma score (GCS) did not vary significantly with age at TBI. Acute symptomatic seizures were diagnosed in 57 patients, among whom 26 developed PTE (46%). Patients with acute symptomatic seizures developed PTE more precipitously than those without early seizures, with a cumulative incidence at 1-year of 77% compared to 28% (p<0.001). Significant predictors of PTE on multivariate analysis included loss of consciousness of any duration (OR 36, CI 10-123), acute symptomatic seizures (OR 10, CI 1.1-96), AIS (OR 4, CI 2.3-507), ICH (OR 4, CI 1.1-25), prior TBI (OR 44, CI 5.4-356), history of ADHD (OR 22, CI 2.6-196) and a family history of epilepsy (OR 20, CI 2.2-179).Conclusions: Acute symptomatic seizures can provide a reliable marker for development of PTE. Development of epilepsy after TBI is also age-dependent, with infants and adolescents at highest risk. Injury severity alone did not account for the binomial age distribution in this cohort, suggesting epileptogenesis after trauma may be amplified during particular windows of brain development. These data will help in counseling families of affected children, and will inform the inclusion criteria in clinical trials investigating therapies to modify epileptogenesis after TBI.