Among adult patients with medically refractory partial-onset seizures, treatment with responsive neurostimulation is not associated with cognitive decline over two years, according to data published in the November issue of Epilepsia. The therapy may in fact induce small but significant improvements in naming among patients with neocortical seizure onset and modest improvements in verbal learning among patients with seizure onset in the mesial temporal lobe.

“These findings require replication but suggest potential cognitive therapeutic benefit associated with responsive neurostimulation,” said David W. Loring, PhD, Professor of Neurology and Pediatrics and Director of Neuropsychology in the Neurology Department at Emory University in Atlanta. “It is unclear if these effects are due to a direct positive effect on the stimulated structures or to modulatory reductions in the adverse effects of seizures or interictal discharges.”

Blinded and Open-Label Stimulation

Neurostimulation has been associated with declines in verbal fluency among patients with Parkinson’s disease, and with a decreased rate of cognitive decline among patients with Alzheimer’s disease. Patients with intractable partial seizures who received deep brain stimulation (DBS) complained of poorer memory and mood, although they had no evidence of neuropsychologic declines.

Dr. Loring and colleagues sought to determine the effect of long-term responsive neurostimulation on neuropsychologic performance among patients with epilepsy. The researchers examined data from a double-blind, randomized, sham-stimulation controlled study of the RNS System, which is manufactured by NeuroPace. Participants were between ages 18 and 66 and had an average of three or more simple partial motor, complex partial, or secondarily generalized tonic–clonic seizures per month. All participants had inadequate seizure control with at least two antiepileptic drugs (AEDs).

After a three-month baseline period, patients received implantation of the neurostimulator and leads. The device was programmed immediately to detect specific patterns of brain activity. One month after implantation, patients were randomized in groups of equal size to receive active or sham stimulation in response to detections. Participants were followed for four months before they received open-label responsive neurostimulation for 18 months.

The researchers collected neuropsychologic data at baseline and during the open-label period at one and two years after implantation. Primary neuropsychologic tests included the Boston Naming Test (BNT) and the Rey Auditory Verbal Learning Test (AVLT). Additional neuropsychologic domains were analyzed as secondary outcomes.

Seizure-Onset Zone Affected Outcomes

Data for 175 patients were analyzed. The study population’s average age was approximately 35, and 18% of participants were female. In all, 86 patients had seizure onset in the mesial temporal lobe, and 76 had seizure onset in the neocortex. Thirteen patients had seizure onset in both regions. They were included in analyses of all participants, but not in analyses according to seizure-onset zone.

Dr. Loring and colleagues observed no significant group performance decline on any neuropsychologic outcome measure. They did, however, find significant improvements on the BNT in 23.5% of participants. Approximately 7% of participants had declines on the BNT. The investigators also noted statistically significant improvements on the AVLT in 6.9% of patients. Approximately 1% of patients had declines on the AVLT. Dr. Loring and colleagues found a trend toward significant improvement on delayed AVLT recall, but no performance change for delayed AVLT recognition.

When the researchers analyzed the data according to seizure-onset zone, they found a significant improvement on the BNT among patients with neocortical seizure onsets. This improvement was not observed in participants with seizure onset in the mesial temporal lobe. In contrast, Dr. Loring’s group noted a significant overall improvement in AVLT learning in patients with seizure onset in the mesial temporal lobe, but not in patients with neocortical seizure onset. Adjusting the data for prior epilepsy surgery and for change in seizure frequency did not alter the results.

Improvements in naming and verbal memory are “an important observation, given the cognitive side effects of some AEDs and the presence of cognitive decline with DBS stimulation of the subthalamic nuclei or of the globus pallidus for Parkinson’s disease,” said Dr. Loring. “The double dissociation observed serves primarily to provide empirical evidence that neurostimulation techniques may ultimately provide therapeutic benefit, and to conceptually support future studies directed explicitly at establishing and optimizing neurostimulation parameters to enhance cognition.”

The fact that naming and verbal memory performance varied according to the seizure-onset region indicates that performance improvements were unlikely to be practice effects. “Practice effects are expected to occur at equal magnitudes in all clinical groups with comparable overall status,” said Dr. Loring. Nor did cognitive outcomes at two years appear to be associated with changes in AEDs.

“This series provides evidence that adults with frequent, intractable, partial onset seizures treated with responsive neurostimulation are not at increased risk for developing cognitive dysfunction, and that some aspects of cognition may be improved,” Dr. Loring concluded.

—Erik Greb

Responsive Neurostimulation for Epilepsy Does Not Impair Cognition