NeuroPace | Resources | Publications

EARLY SAFETY EXPERIENCE WITH A FULLY IMPLANTED INTRACRANIAL RESPONSIVE NEUROSTIMULATOR FOR EPILEPSY

David Vossler, MD ¹; Michael Doherty, MD ¹; Robert Goodman, MD, PhD ²; Lawrence Hirsch, MD ², Jacob Young, MD ¹ and Diana Kraemer ¹

¹ Epilepsy Center, Swedish Neuroscience Inst, Seattle, WA, United States, 98122 and ² Comprehensive Epilepsy Center, Columbia University, NYC, NY, United States, 10032

RATIONALE: As part of a multicenter feasibility study, NeuroPace responsive neurostimulators (RNSTM) have been implanted in 3 patients with refractory epilepsy. The RNS system (device and leads) is implanted in the skull to continuously monitor and record ECoG from depth or subdural strip electrodes. When electrical seizure discharges are detected, the RNS system delivers electrical stimulation to suppress ictal discharges before a clinical seizure occurs.

METHODS: Subjects (age 18-65) with >=4 disabling, partial seizures/month for 3 consecutive months were implanted with the RNS system. Responsive stimulation was enabled 1 week or later after surgery. Subjects were not blinded to stimulation. Seizure frequency, severity, and adverse events (AEs) were assessed throughout the study.

RESULTS: Three adults were implanted with the RNS system. Two had previous invasive EEG monitoring and cortical resection. The RNS device was secured within the right (R) parietal bone in 2 subjects and the left (L) parietal bone in the other. Two 4-electrode subdural strips were connected to each RNS device (Table 1).

Table 1

	Subject 1	Subject 2	Subject 3
Surgery Duration (Hrs)	4.3	3.6	2.5
Leads	2 R Parietal	L Parietal, L Temporal	R Parietal, Inter-hemispheric, R Parietal, R Temporal
Days Implanted	85	36	30
Days Therapy Enabled	76	34	0

Automatic detection of ictal discharges was enabled immediately after surgery. It correctly detected seizures in the 2 subjects that have had seizures since implantation. In both cases seizure onsets were characterized by beta-gamma frequency activity. Four AEs have been reported, all of which occurred in the first month postop (none during surgery) and all classified as mild. The AEs reported were edema, headache, diarrhea (Subject 1) and possible incision infection (Subject 3). The edema in Subject 1 was further classified as an adverse device effect: post-operative edema around the leads with quick recovery and no sequelae. Product relation was uncertain for one AE: Subject 3 was treated with IV antibiotic for a question of immediate postop wound infection; culture results were negative and the wound healed well. No product complaints have been reported.

CONCLUSION: Among the 3 subjects implanted with the RNS system as of 4/30/04, surgery has been safe. Responsive stimulation has been safe and well tolerated. The device has performed as expected. All AEs have been classified as mild. This is the first report ever of an implanted responsive neurostimulator that continuously monitors ECoG in order to automatically deliver therapy in response to ictal activity. Further implantations are scheduled prior to the Dec 2004 AES meeting.

Funding supported by: NeuroPace, Mt. View, CA

SUPPRESSION OF EPILEPTIFORM ACTIVITY BY RESPONSIVE ELECTRICAL STIMULATION IN EPILEPTIC PATIENTS

T. Tcheng, R. Esteller, J. Echauz
NeuroPace, Inc., Mountain View, California, USA

RATIONALE: Closed-loop responsive electrical stimulation is an emerging therapy for the treatment of intractable epilepsy. This therapy is being evaluated in clinical trials of the Responsive Neurostimulator (RNS™) system (NeuroPace, Inc.). During these trials, the RNS device was either 1) externalized and connected to cortical strip and depth electrodes used for pre-surgical monitoring, or 2) fully implanted and connected to permanent cortical strip and depth electrodes in ambulatory patients. Data from 35 epilepsy patients were evaluated. Electrocorticograms (ECoGs) containing stimulations delivered in response to detected events as well as non-treated detected events were stored by the RNS device and uploaded to a central database for subsequent analysis. Since there may be a correlation between acute electrographic responses to responsive stimulation and clinical effects, these electrographic responses were quantified and analyzed.

METHODS: Two event classes were identified for this analysis: 1) detection with stimulation ("stim"), and 2) detections without stimulation ("non-stim"). Electrographic activity was analyzed within 2-sec windows before and after each event, with an intervening 2-sec window ignored to avoid stimulation artifacts. Several features were quantified within these windows and the post-event vs. pre-event ratios and/or differences were calculated for each event (except those with stimulations occurring within either the pre-event or post-event window) and on each ECoG channel. The features analyzed include power spectral density (PSD), Line Length, and Area. Stimulated response ratios were compared between the "stim" and "non-stim" event classes.

RESULTS: he PSD ratio responses, as well as the normalized PSD difference between pre- and post-event windows, averaged across all "stim" events and across patients, showed a linear energy decrease, decreasing in magnitude with increasing frequency. In contrast, the "non-stim" events showed a linear energy increase, increasing in magnitude with frequency, for frequencies above 20 Hz; and between 0.5 and 20 Hz they have a small energy decrease, increasing in magnitude with frequency.

CONCLUSION: These observations suggest that closed-loop responsive electrical stimulation may acutely suppress epileptiform activity. The question of whether acute electrographic responses and suppression of epileptiform activity correlate with the clinical efficacy of responsive stimulation is currently under investigation.