Abstracts

Rationale: Poorly controlled epilepsy is associated with higher mortality compared to the general population. A significant percentage of deaths in epilepsy are due to sudden unexpected death or SUDEP. Although the mechanisms that underlie SUDEP are not clear, possible causes include seizure-related hypoxia and/or cardiac arrhythmia. To investigate whether cardiac autonomic modulation is dysfunctional during and after seizures, we measured heart rate (HR) and heart rate variability (HRV) before, during and after seizures to better understand mechanisms that may underlie SUDEP. Methods: Twenty-five patients with pharmacoresistant epilepsy were admitted to the UCLA epilepsy monitoring unit for video-EEG recording. Simultaneous EEG and ECG was carried out during complex partial (n=13) and generalized tonic-clonic seizures (n= 12). EEG was manually reviewed to determine the beginning and end of ictal discharges. A semi-automated computer algorithm detected successive QRS complexes in the ECG signal. Time series corresponding to R-waves were manually reviewed to remove artifacts, and linear interpolation was used to correct R-R intervals that deviated more than 20% from the preceding R-R interval. Automated software (Kubios HRV, Kuopio, Finland) was used to compute time- and frequency-domain, and non-linear measures of HRV for each R-R series during the following epochs: immediately preceding seizure onset (preictal), from beginning to end of EEG ictal discharges (ictal), and two consecutive episodes immediately following the end of the seizure (labeled early & late postictal). Repeat measures analysis of variances was used to evaluate HR and HRV in relation to periictal episodes. Results: The mean duration of ECG analyzed for each episode was as follows (minutes: seconds): preictal: 2:32 0:29; ictal 1:38 0:44; early postictal 2:00; and, late postictal 3:15 0:54. Compared with preictal epochs, the following significant changes were detected during ictal, early and late postictal periods (% change from preictal): higher HR (61, 34 & 22), reduced RMSSD (-17, -10 & -10), lower sample entropy (-58, -30 & -13), and reduced short-term variability on Poincar plots (-18, -10 & -11). An increase in long-term variability on Poincar plots was observed during ictal and early postictal periods (17 & 8%) in relation to preictal periods. Furthermore, a significant reduction in high frequency power (-25%) and higher ratio of low frequency to high frequency power (65%) was observed during ictal episodes compared to preictal periods. No significant differences were detected in SDNN, coefficient of variation R-R intervals, or low frequency power. Conclusions: These data confirm results from previous experimental animal studies and indicate a significant and robust reduction in vagal-mediated regulation of HR and HRV during ictal and immediate postictal periods. The present study and results from previous patient studies suggest that ictal and postictal dysautonomia and deranged HRV may be critical factors in the lethal events associated with SUDEP.