Abstracts

Rationale: Seizure-related cardiorespiratory changes are observed in sudden unexpected death in epilepsy (SUDEP) and near SUDEP. Yet, electro-clinical correlations of autonomic semiology in pharmaco-resistant focal epilepsy are limited. We studied cardiorespiratory changes with SEEG-recorded seizures using a multimodal system integrating neurophysiologic, cardiorespiratory and polysomnographic signals. We aimed to determine if seizures of insular origin differed on key cardiorespiratory variables from those arising from other areas.  Methods: Participants were adults with pharmaco-resistant focal epilepsy undergoing SEEG. Temperature and pressure-based airflow, effort (respiratory inductance plethysmography), EKG (R-R interval), pulse oximetry (SpO2), end tidal CO2 (EtCO2; Nihon Kohden Nomad) and transcutaneous CO2 (TcpCO2; Sentec) were recorded and analyzed at pre-ictal, ictal and post-ictal time points (figure 1). Vertex scalp needle electrodes identified postictal generalized EEG suppression (PGES). Linear mixed-effects models compared respiratory rate (RR), peak CO2, nadir SpO2 and >=4% oxygen desaturation and >=4% mmHg hypercapnea event duration between Insula and Non-insular onset/early spread seizures and between seizures with and without ictal respiratory events, and assessed the association with duration rate variables (total desaturation or hypercapnea duration/seizure duration). Generalized estimating equation accounted for repeated seizures on the same patient.  Results: Twenty seven adults (34.2±9.6 yr) contributed 60 seizures including 48 focal onset seizures (17 Insular and 31 Non-insular). Non-insular seizures arose from mesial temporal (35%), orbitofrontal/cingulate (35%), neocortical temporal (26%) and occipital (3%) regions. Insular and Non-insular groups were had similar seizure duration (87[68,194] vs. 80[34,173] sec, p=0.66), NREM sleep origin (58.8 vs. 48.4%, p=0.66), presence of PGES (17.6 vs. 16.1%, p=0.99), and presence of ictal respiratory events (35.3 vs. 45.2%, p=0.90). Central apnea/pause was the most common event in both groups.  RR change was greater in Insular seizures (29.1 vs. 23.9 bpm; p=0.28). Insular and Non-insular SpO2 nadir (89[74, 91] vs. 90[77, 93]%, p=0.28), peak CO2 (44±10 vs. 43±8 mmHg, p=0.74), HR change (17.6±22.8 vs. 14.9±19.0 bpm, p=0.60), peak HR (143.1±54.1 vs. 140.1±41.2 bpm, p=0.77), total desaturation duration/seizure duration (3.9[0.73,4.8] vs. 1.4[0.63,3.4], p=0.87) and total hypercapnea duration/seizure duration (3.7[0.68,5.8] vs. 3.8[1.2,6.0], p=0.85) were comparable. Regardless of localization, presence of ictal respiratory events was associated with larger RR increase (6.8 vs. 2.6 bpm, p=0.42) and postictal HR (102.9 [91.2, 114.6] vs. 87.4 [78.3, 96.5] bpm, p=0.42) and lower SpO2 nadir (79.6 [74.4, 84.9] vs. 89.8 [85.7, 93.8] %, p=0.005). Conclusions: Insular onset/early spread seizures had greater RR change than other seizures, while SpO2, CO2 and HR changes were comparable. In both groups, desaturation and hypercapnea duration exceeded seizure duration. Independent of seizure origin, presence of ictal respiratory events was associated with larger RR and HR increase and lower SpO2. Among multiple autonomic signs and symptoms emerging in the course of temporal and perisylvian seizures, RR change seems to be a specific marker of early insular involvement. This study demonstrates that anatomo-electroclinical correlations may differentiate autonomic manifestations of seizures arising from distinct central autonomic areas.  Funding: No funding