Abstracts

Rationale: Vagal nerve stimulation (VNS) has been used as an effective treatment for medication-refractory epilepsy for decades, though a new generation of heart rate sensing VNS devices (HRSVNS) seeks to improve on existing outcomes by using closed-loop, pre-ictal autonomic changes to target stimulation. These devices (Aspire 106, Sentiva 1000) sense rapid changes in heart rate and possess nocturnal programming methods to discharge in response to abrupt fluctuations in heart rate. Though prior studies have demonstrated outcomes associated with HRSVNS devices independently (Boon, 2015; Fisher, 2016), it is unclear how HRSVNS compare with older models.We provide a single-center comparison of seizure reduction and associated factors at approximately 9 months follow-up for recipients of HRSVNS compared to recipients of older VNS models. Methods: To determine the effects of HRSVNS on seizure reduction compared with older models, we performed a retrospective chart review of all first-time VNS implants at our institution since 2004. Patients were included if they had at least 1 month of follow-up. Demographic and clinical data were collected pre-operatively, post-operatively, and at 9-months after implantation. The primary endpoint was change in seizure frequency. Secondary endpoints included surgical morbidity, anti-epileptic medications, change in seizure types, and Engel classification. Results: Twenty-four recipients of HRSVNS were followed for a median 252 days [IQR 290d – 201d] and 69 recipients of older VNS had a median follow-up of 256 days [IQR 281d – 225d]. Recipients of HRSVNS had a median decrease in seizure frequency of 73% [IQR 0-84%], compared to a median decrease of 50% [IQR 0-78] for recipients with older devices. This was a potentially clinically significant finding that did not reach statistical significance (Wilcox-test p-value = 0.18).At follow-up, a greater proportion of HRSVNS than non-HRSVNS patients experienced reductions in seizure frequency at different levels: > 25% reduction HRSVNS 64% vs older VNS 53% (p = 0.7); > 50%: 54% vs 40% (p = 0.4); > 75%: 39% vs 25% (p = 0.3); and > 90%: 14% vs 8% (p = 0.5). Patients with newer models were also less likely to experience post-implant seizure frequency increases > 25%: 14% vs 21% (p = 0.4); > 50%: 7% vs 20% (p = 0.1); > 75%: 4% vs. 19% (p = 0.06); and > 90%: 4% vs. 17% (p = 0.06).Patients receiving HRSVNS were also less likely to have Engle Class IV (“No worthwhile improvement”) outcomes (18% vs 33%; p=0.2). Other outcomes were similar: Engle Class I (11% vs 7%; p = 0.7), Engle Class II (22% vs 19%; p = 0.8), and Engle class III (48% vs 40%; p = 0.5).We found complication rates to be similar between devices, with the most common complications being hematoma (0% vs. 3%; p = 1.0) and hardware removal (4% in both groups; p = 1.0). The most common surgical morbidity was change in voice observed in a slightly higher percentage of HRSVNS patients compared with older devices (29% vs 20%; p = 0.1). Conclusions: We demonstrate that a higher proportion of patients with HRSVNS may have clinically meaningful reductions in seizure frequency and lower Engle IV outcomes than patients with older VNS devices, though our current sample is under-powered to determine if these differences are statistically significant. Furthermore, the results of this study suggest that older models may be more likely than HRSVNS to incur increases in post-implant seizure frequency greater than 50%. Our preliminary data are valuable for sample size estimation for larger studies. Studies with longer follow-up and a larger sample are warranted. Funding: No funding