Abstracts

Rationale: Vagus nerve stimulation (VNS) has been shown to be a safe and effective adjunctive treatment option for patients with intractable partial onset seizures. Currently, more than 30,000 patients have been implanted with the NeuroCybernetics Prosthesis (NCP) worldwide. However, the clinical events that indicate the need for battery replacement are poorly defined. Older models (NCP model 100 and NCP model 101 with serial number below 10,000) rely on cumbersome chart calculations. Newer models (NCP model 101 with serial numbers above 10,000 and NCP model 102) have an elective replacement indicator (ERI). In our experience, seizure control may be lost due to the VNS battery approaching end of service (EOS) even when the ERI is not triggered. Methods: A retrospective chart review was performed of the 70 patients who underwent VNS battery replacements at the University of Southern California Medical Center between 1998 and 2008. The average monthly seizure frequency was calculated for a six month period of best seizure control while on VNS (VNS baseline) compared to the three month period immediately preceding VNS battery change. Patients were divided into 3 groups: ‘no change prior’ if no change in average monthly seizure frequency occurred or less than 25% increase over VNS baseline. A 25 to 74% increase was ‘moderate increase prior’ and a >75% increase was ‘clinically significant increase prior’. After replacement, average monthly seizure frequency for the 3 months prior to battery change was compared to the best three months post battery change. The same 3 groups were used: ‘no change post’ for less than 25% reduction in average monthly seizure frequency, ‘moderate improvement post’ for a 25 to 74% reduction and ‘clinically significant improvement post’ for a >75% reduction in seizure. Results: A total of 70 patients underwent 85 VNS battery replacements between 1998 and 2008. We analyzed 56 pre replacement devices and 55 post replacement devices in 48 patients. Patients were mainly excluded due to loss to follow up or inadequate seizure calendars. The average number of months to replace model 100 was 46 (n=31), 55 for model 101 (n=21), and 51 for model 102 (n=4). Only 2/11 (18.2%) devices with the ERI function indicated EOS. Patients who had ‘no change’ prior to battery replacement were likely to remain in the ‘no change’ category following battery replacement (69.6%, n = 16/23, p<0.01). Patients who had >25% increases in seizure frequency prior to battery replacement (>25%) were likely to have >25% decreases in seizure frequency following battery replacement (75%, n = 24/32, p<0.01). Conclusions: VNS therapy becomes clinically ineffective prior to complete battery depletion. Replacing the VNS battery when seizures worsen can result in improved seizure control even if the ERI is not triggered. Battery depletion may be poorly indicated by ERI. In our study, VNS battery replacement occurred months to years prior to published expected battery life. Clinicians should have a low threshold for VNS battery change when seizures frequency increases >25% over VNS baseline.