Abstracts

Rationale: Palliative procedures for drug resistant epilepsy can be helpful, however, selecting the best approach for an individual remains a challenge. We compared the long-term outcomes of palliative disconnection procedures and vagal nerve stimulation (VNS) at a large academic epilepsy center.Methods: Subjects were selected from surgical or implantation cases treated at the MGH Epilepsy Service between 1993 and 2012 and included patients that underwent disconnection procedures such as corpus callosotomy or multiple subpial transections (CC/MST) or implantation of a VNS. All patients were followed for at least 1 year. Demographic data and the pertinent medical information including assessment of seizure outcome (Engel classification system) at 1, 3 and 5 years were gathered from the medical record.Results: 32 patients were identified (CC/MST: 10, VNS: 22). The average age at operation was 44 (+/-15), and 50% were female. Groups were comparable with regard to etiologic diagnosis, type of seizures and date of procedure. Mean follow-up was 77 months (23 to 131 months). At one, three and five years, disconnection surgery was superior to stimulation device implantation with respect to seizure reduction. At the one year, 10% of CMT/MST patients were class I, 40% II, 50% III and 0% IV. VNS patients outcome was 14% class I, 5% in II, 41% III and 41% IV. At 3 years (n=24, CC/MST = 8, VNS = 16) CC/MST patients were Engel class I 13%, II 38%, III 50%, IV 0% while VNS patients were class I 6%, II 0%, III 56%, IV 38%. At 5 years after surgery outcome was similar: CC/MST Group: I 17%, II 33%, III 50%, IV 0%. VNS Group I 8%, II 0%, III 54%, IV 38%. A non-parametric two-group analysis using the Wilcoxon (Mann Whitney) test showed that the Engel class in the two groups consistently differed (p = 0.02 at 1 year, p = 0.009 at 3 years and p = 0.02 at 5 years). Post-surgical complications were present in 30% of disconnection surgery patients (20% after subpial transections and 10% after callostomy). There were no complications in the stimulation device group.Conclusions: We found that seizure control was better after surgical disconnection than after implantation of a VNS, and this advantage persisted over a prolonged follow-up period. Disconnection procedures carry a higher complication rate in our series that must be considered in guiding patients. Additional research of the influence of pre-surgical variables will help us understand what outcomes patients value the most when palliation is the objective and how best to assess patients' health status outcomes over time.