Abstracts

Rationale: Patients with extra-temporal epilepsy and normal MRI represent a challenge for the surgical treatment. Up to 60% of them are not made seizure-free after the procedure; this is also true for 10-15% of the patients with temporal lobe epilepsy who are usually the best candidates for surgery. These surgical failures, which were medically refractory beforehand, have few therapeutic options. Neuromodulatory techniques, especially vagus nerve stimulation, have been tried over the last years. Deep brain stimulation emerged as a potentially new therapeutic option for these patients. Methods: Six patients who failed extratemporal (n=3) or temporal (n=3) cortical resection were studied. One patient had rolandic (previously submitted to sensitive and motor giri resection), 2 had frontal (previously submitted to frontal resection and 3 had temporal (previously submitted to cortico-amygdalo-hippocampectomy) lobe epilepsy; all had unremarkable MRI findings before cortical resection. All patients were implanted bilaterally at the anterior nucleus of the thalamus (AN) using a Kinetra device. Continuous bilateral stimulation was carried out using 130Hz, 300usec, 4-6V pulses; stimulus intensity was increased every 2 weeks (0.5V increments). CT scanning checked the electrode s position postoperatively. Results: A recruiting stimuli-driven response could be seen bilaterally over the fronto-temporal convexity after low-frequency AN intra-operative stimulation. A DC-shift was noted when high-frequency stimulation was used. The patient with rolandic epilepsy had 30% reduction in seizure frequency; those with frontal lobe epilepsy had 50 and 55% reduction in seizure frequency. Patients with temporal lobe epilepsy had 80, 80 and 90% seizure frequency reduction, respectively. Three patients presented transient (30-45 minutes) paresthesia during the programming sessions during which stimulus intensity was increased. There was no neuropsychological decline or somnolence. Conclusions: AN-DBS proved to be safe and effective in this patient population. Although the number of patients is small for statistical analysis, those with failed temporal lobe surgery did much better than those with extra-temporal epilepsy. These findings might correlate with the known direct projections from the AN to other Papez circuit targets (including the temporal lobe). There is no efficient direct projections from the AN to rolandic or lateral frontal cortex. Pre-study worries regarding side effects due to continuous bilateral AN stimulation did not materialized during follow-up.