Abstracts

Rationale: Persons with epilepsy are at risk for Sudden Unexplained Death in Epilepsy (SUDEP). The underlying pathophysiology is believed to be acute cardiac and/or pulmonary failure, perhaps related to a seizure. However, there are few cases where direct electrocorticographic (ECoG) data has been recorded during a SUDEP event. Data from 5 SUDEP cases while undergoing ECoG sensing and recording are presented to determine if ECoG offers a better understanding of SUDEP.Methods: Data was collected from 256 subjects and over 880 implant years and 801 stimulation years during investigational trials of the RNS System (NeuroPace, Inc). Subjects had medically intractable partial onset seizures and were implanted with a responsive neurostimulator and 2 depth and/or subdural leads placed at the seizure focus. The neurostimulator was programmed to detect ECoG activity identified by the physician as significant, and, for some subjects, to provide stimulation in response to the detected event. The neurostimulator records the time of all detections, and stores samples of the ECoG according to storage triggers programmed by the physician. An independent SUDEP Analysis Committee reviewed all subject deaths to determine whether the death was a SUDEP. Results: Six subjects (4 with stimulation on) had possible or definite SUDEP. ECoG data was available for 5 of the 6 subjects; one subject had an unwitnessed death and was buried before neurostimulator data could be obtained. One subject had detections of epileptiform activity lasting several minutes followed by no subsequent detections, and was found dead 1/2 hour later. Another subject had frequent detections of epileptiform activity, which ended abruptly. An ECoG containing high amplitude spike activity was stored within the last hour of detections. This subject was found dead 11 hours later. The next subject had frequent detections of epileptiform activity, which lasted over 20 minutes, followed by an electrographic seizure, after which there were no detections. This subject was found dead 4 hours later. Another subject had a 1.5 minute electrographic seizure that corresponded to a witnessed GTC, after which there were no further detections. At the conclusion of the clinical seizure, the subject was apneic, in atrial fibrillation and pulseless. Although cardiac activity was restored, the subject remained unconscious with no sign of brainstem function and life-support was withdrawn the following day. The final subject was found dead about 5 hours after her last contact with a person; there were no ECoG detections or recordings over the time span in which the SUDEP occurred. Conclusions: These cases suggest that SUDEP may occur during periods of frequent epileptiform discharges as well as at the time of a seizure. Four of the 5 subjects reported here had epileptiform activity and/or electrographic seizures preceding the SUDEP event. The abrupt absence of detections may correspond to the time of death. The RNS System s ability to store information regarding electrocorticographic activity may provide information to help better understand the mechanism of SUDEP.