Abstracts

Rationale: As many as 30-50% of seizures are not reported by patients, highlighting a need for objective measures of seizure burden and epilepsy severity. The RNS® System (NeuroPace, Inc.) is FDA approved as an adjunctive treatment for partial onset seizures in adults. In addition to providing responsive stimulation, the RNS System also records ambulatory electrocorticographic (ECoG) data and a continuous hourly count of detected ECoG events. We analyzed the correlation between the frequency of episodes detected by the device and the frequency of clinical seizures (CS) reported by patients. Methods: The relationship between ECoG events detected by the RNS® Neurostimulator and self-reported CS was examined in patients who participated in the RNS System Pivotal trial (n=191). The RNS Neurostimulator can count and record Long Episodes (LEs), which are electrographic events detected by the neurostimulator that exceed a physician-specified minimum duration. The approach for determining whether LE-triggered ECoG records were electrographic seizures (ES) has been described elsewhere (Quigg et al., Epilepsia 2015; 56(6):968-71). Based on that analysis, a subset of subjects with ≥ 4 LE-ECoGs and for whom ≥75% of their stored LE-ECoG events represented ES were identified. In this group, the LE rate per day was calculated for each subject. The relationship between LE and CS rates within a sliding 28-day window was assessed using Pearson correlations. Results: There were 54 subjects for whom, during an 84-day period with stable neurostimulator detection settings, ≥75% of stored LE-ECoGs were determined to contain ES. Within these subjects, 61% (33/54) had mesial temporal lobe epilepsy (MTLE), 32% (17/54) had seizures arising from the neocortex, and 7% (4/54) had seizures arising from both the MTL and the neocortex. The 28-day average LE rates ranged from 0 to 210.1 per day and the CS rates ranged from 0 to 24.7 per day.  For the majority of patients LE rates exceeded CS rates. Significant correlations (p < 0.05) between LE and CS rates were present within 72% (39/54) of subjects; 77% (30/39) were positive and 23% (9/39) were negative. 58% of subjects with MTLE had positive correlations; 24% had negative correlations; and 18% had no significant correlations. In comparison, 53% of subjects with neocortical seizures had positive correlations; 6% had negative correlations; and 41% had no significant correlations. The majority of subjects with negative correlations had MTLE (8/9). Conclusions: In this analysis, the LE and CS rates were often positively correlated, suggesting that, for some patients, the RNS System may provide an objective assessment of seizure burden. However, in most patients the self-reported CS may not represent the extent to which ES occur, perhaps because ES are subclinical, occur during sleep or are forgotten. The results for MTLE patients with negative correlations may suggest that these patients are more likely to report CS when they are having fewer ES, Funding: None.