Abstracts

Rationale: Sudden unexpected death (SUDEP) is a significant epilepsy-associated mortality. Childhood-onset epilepsy, lack of seizure control, and history of status epilepticus (SE) are known risk factors. In childhood-onset epilepsy, the cumulative SUDEP risk increases with epilepsy duration, suggesting that pathological changes may develop over time, ultimately contribute to SUDEP. Because cardiac changes are a comorbidity associated with epilepsy, we sought to investigate whether cardiac electrical alterations develop over time in children with epilepsy who may be at risk for SUDEP. Methods: We prospectively identified children admitted to the pediatric intensive care unit for seizures or SE between 4/2014 and 7/2017. Children were included if they had at least one 12-lead ECG study in the electronic medical record (EMR) database and did not have pre-existing cardiac conditions or known ion channelopathies. Children were classified into 2 cohorts (control, epilepsy). For the epilepsy group, we determined the age of epilepsy onset based on the documentation from neurology. For each ECG, we obtained the age of the subject and the indication for the study. Only the first ECG study was included in the analyses if there were multiple ECG studies in the same month. The primary outcome was altered cardiac electrophysiology by the official cardiology interpretation, which was defined as QRS axis deviation, altered PR, QRS and corrected QT intervals, or ST segment and T wave changes. We performed Student t or Mann-Whitney test for continuous variables; Pearson χ2 for categorical variables; and logistic regression to identify factors associated with ECG alterations. Data are expressed as median [25%-75%ile], mean ± SEM, or odds ratio [95% confidence interval]. Results: We identified 427 children during the study period and of these 229 met the study criteria. 96 children had no history of epilepsy (controls, 220 ECGs), while 133 children had epilepsy (334 EKGs). In children with multiple ECG studies in the EMR, the median time between consecutive ECGs was 3 months [1-11 months] in controls and 9 months [3-21 months] in the epilepsy group (p < 0.01). The median age of the subject at the first ECG was 18 months [8-79 months] in controls and 35 months [10-124 months] in the epilepsy group (p = 0.18); and at the last ECG was 45 months [25-99 months] in controls and 100 months [53-185 months] in the epilepsy group (p < 0.01). Compared with controls, the epilepsy group was more likely to have ECG abnormalities (OR 1.6 [1.13-2.25], p < 0.01). Within the epilepsy group, ECG alterations occurred at an older age (85.8 ± 5.0 vs. 112.7 ± 5.7 months, normal vs. altered EKG, n = 157 vs. n = 177, p < 0.01). Similarly, ECG alterations were associated with longer epilepsy duration (49.0 ± 3.8 vs. 74.4 ± 4.6 months, normal vs. altered EKG, n = 154 vs. n = 168, p < 0.01). Logistic regression revealed that for each year of epilepsy, there was an increased likelihood of having ECG alterations (OR 1.11 [1.06-1.17], n = 322, p < 0.01). In contrast, age was not a factor associated with ECG abnormalities in the control subjects. Conclusions: Our findings suggest that there are chronic cardiac electrical alterations paralleling the course of epilepsy in the cohort studied. The observation that increasing epilepsy duration is associated with an increased likelihood of ECG alterations suggests that cardiac surveillance may be warranted in select groups of patients with epilepsy. Studies are ongoing to identify additional factors contributing to chronic cardiac alterations associated with epilepsy. Funding: No funding