Annual Meeting Abstracts: View
(Abst. 3.235), 2014
ELECTRODERMAL SLEEP STORM ACTIVITY AS A BIOMARKER IN EPILEPSY PATIENT
Authors: Kush Kapur, Sigride Thome-Souza, Jacquelyn Klehm, Rani Sarkis, Elanagan Nagarajan, Michele Jackson, Rosalind Picard, Chiran Doshi, Christos Papadelis, Barbara Dworetzky, Claus Reinsberger and Tobias Loddenkemper
Content: Rationale: Use of spontaneous electrodermal activity (EDA) sleep storms in epilepsy has been raised as a new tool to guide clinicians regarding to seizure lateralization. A storm is defined as clusters of EDA during sleep with 5 EDA/minute over 10 consecutive minutes. There is no data on EDA sleep storms in patients with epilepsy. We believe that EDA sleep storms could be a useful biomarker to correlate epilepsy characteristics with sympathetic nervous system activity. Thus, our objective was to analyze EDA sleep storms in epilepsy patients and correlate with patient and seizure demographics, MRI lesion, and EEG findings to delineate a better profile of storms in this population. Methods: We prospectively enrolled 284 patients during epilepsy monitoring. EDA and accelerometry were recorded with portable wristband devices. We excluded patients who had a vagus nerve stimulator or a history of autonomic and/or skin disorders. Only spontaneous EDA sleep storms (not associated with seizures) recorded during the first night with storms were analyzed. We used the following criteria to define a storm: (i) slope exceeding 0.09 μS/sec and, (ii) > 3 EDA increases/30 seconds (Sano & Picard, 2011). We analyzed the EDA activity and accelerometry using a MATLAB algorithm. Results: We evaluated 74 patients with a median age of 16.7 (range: 0.8-67) years, including 34 (45.9%) girls. A median of 0.75 (SD: 0.4) storms were observed on the first night. There was an inverse correlation between age and area under the curve (AUC) (rs= -0.28, p=0.036). Patients with EDA storms were treated with fewer antiepileptic drugs (Median 2 vs 2.5, p=0.032). The presence of a storm was related to seizure onset ictal EEG localization (p=0.019), and most frequently generalized (10/11, 91%) and frontal (7/8, 87.5%) ictal EEGs onset (Table 1). The odds ratio (OR) of having EDA sleep storms in patients with generalized ictal EEG onset, compared to other EEG onsets (Temporal, Hemispheric and Posterior Quadrant), was 8.89 (p=0.058, 95%CI [0.922, 85.655]). The OR for storms in patients with frontal EEG onset, compared to other ictal EEG onsets, was 6.22 (p=0.122, 95%CI [0.623, 62.15]). The log area under the curve (AUC) of the EDA signal was marginally higher for generalized seizures in comparison to partial seizures (p <0.052) [Figure 1]. We also found an association between storms and interictal EEG lateralization (p=0.046). The OR of storms with right sided interictal EEG in comparison to left was 6.33 (p=0.043, 95% CI [1.061, 37.769]). Conclusions: Generalized ictal EEG onset and right-sided interictal epileptiform activity are more frequently seen in patients with EDA sleep storms. Regardless of seizure type, younger children had stronger EDA sleep storms. These results provide a potential novel biomarker for the assessment of epilepsy patients, and information on the interaction between and the sympathetic autonomic nervous system (Supported by Danny Did Foundation).