Long‐term Deep Brain Stimulation for Refractory Temporal Lobe Epilepsy

Abstract

The amygdala and hippocampus are structures that often show specific initial electroencephalographic epileptiform discharges as a reflection of seizure onset in human temporal lobe epilepsy (1). These findings have been recorded with implanted depth electrodes in patients with refractory epilepsy in whom the ictal onset zone could not be identified on the basis of noninvasive evaluations. About 10% of patients with refractory epilepsy are scheduled for invasive video-EEG monitoring to localize the ictal onset zone during presurgical evaluation (2). We have previously shown that chronic deep brain stimulation (DBS) electrodes are suitable for intracranial ictal onset localization in the medial temporal lobe (3). Using DBS electrodes, we evaluated the efficacy and safety of amygdalo-hippocampal DBS (AHDBS), following invasive localization of the ictal onset zone, for reducing seizures in patients with refractory epilepsy. Seven patients (7 M) with refractory epilepsy were implanted with bilateral AHDBS electrodes and/or subdural grids for ictal onset localization and subsequent stimulation. In patients with ictal onset in the temporal lobe, AHDBS was initiated according to the side of ictal onset during a trial period with an external pulse generator. Patients in whom a significant reduction of interictal spikes and/or seizures was shown during the initial trial period were implanted with an abdominally located pulse generator. Patients were followed up at the epilepsy clinic every 2–4 weeks. Two patients had a left-sided focal medial temporal lobe onset. Three patients had a right-sided regional medial temporal lobe onset. One patient had a bilateral regional temporal lobe onset with predominant involvement of the left side. One patient had a left-sided regional medial temporal lobe onset. Six out of the seven patients had a > 50% reduction of interictal spikes during the initial AHDBS trial period. In one patient who showed very infrequent spiking, seizure frequency that had significantly decreased was used as a criterion for implantation. All patients were implanted with an internal generator. The mean follow up in these patients was 14 months (range: 5.5–21 months). One patient has been free of complex partial seizures (CPS) for 1.5 years and has been tapered off two antiepileptic drugs (AEDs). Three of the seven patients have a ≥50% reduction in seizure frequency: one of the three patients is seizure free during the day, and two of the three patients have been tapered off one AED. Two of the seven patients have a reduction of 25% of seizure frequency but have still been tapered off two AEDs; one of these patients had bilateral seizure onset and is currently being revised for bilateral AHDBS. In one patien,t no change in seizure frequency occurred. None of the patients report side effects. The development of neurostimulation for neurological indications is enhanced by two major concerns and limitations related to standard available treatments. First, there is a general tendency to look for treatments that are minimally invasive and harmful to the patient. Second, the refractoriness of some neurological diseases and the inability to treat them with the available means provides an impetus to search for novel treatments. Two approaches for DBS in refractory epilepsy can be followed. One is to target crucial central nervous systems structures that are considered to have a “pacemaker” or essential role in epileptogenic networks that have been identified, such as the thalamus or the subthalamic nucleus. Another approach is to interfere with the area of ictal onset itself. Our study aimed at evaluating the efficacy of DBS in the medial temporal lobe after the ictal onset zone had been identified in this region. Animal studies have shown positive effects on epileptic activity when electrical fields were applied to in vitro hippocampal slices (4). Also, in vivo studies in rats have shown that electrical stimuli (quenching) applied following a kindling stimulus can delay the development of the kindling process (5,6). Bragin et al. (7) found that repeated stimulation of the hippocampal perforant path in the kainate rat model significantly reduced seizures. In humans, preliminary short-term stimulation of hippocampal structures showed promising results, with significant reduction of interictal epileptiform activity and seizure frequency (8). More than half of the patients treated with AHDBS in this study had a reduction of seizure frequency of ≥50%, allowing tapering of one or more AEDs. None of the patients reported side effects or showed changes in bedside neurological and neuropsychological testing. Extensive neuropsychological testing comparing pre- and post-DBS results is currently being undertaken. The mechanism of action (MOA) of DBS in reducing seizures remains unclear. Some reports in the literature support the hypothesis that actual stimulation is not necessary to achieve efficacy; these reports claim that efficacy is based on the lesion provoked by the insertion of the electrode (9). This observation was described for electrode insertion in the anterior nucleus of the thalamus and referred to as “microthalamotomy.” Furthermore, reports exist describing prolonged seizure control in patients who underwent invasive recording with conventional electrodes (10). Blinded randomization of patient to “on” and “off” stimulation paradigms following implantation for substantial amounts of time (e.g., 6 months) may clarify this and may also simultaneously clarify the potential effect of sham stimulation due to an implanted device. A second hypothesis is that DBS acts through local inhibition induced by current applied to a specific CNS structure. This is the hypothesis of the so-called reversible functional lesion, where in the case of targeting crucial structures in a network, nuclei that are involved in propagating, sustaining, or triggering epileptic activity are inhibited. In the case of targeting an ictal focus, a similar reasoning may hold true, suggesting that applied current inhibits overexcitable tissue. Apart from this “local” inhibition, the MOA of DBS may be based on the effect on projections leaving from the area of stimulation to other central nervous structures. This may be the most likely hypothesis when crucial structures in epileptogenic networks are involved. However, considering that the medial temporal lobe structures are also potentially involved in these networks, it may be that targeting the ictal focus may also affect the epileptogenic network. When projections from one structure to another are involved, this may be through the activation of inhibitory projections or through the inhibition of (over)excitatory projections. In this open pilot trial, AHDBS significantly reduced seizure frequency during long-term follow up without side effects. For patients who are less suitable candidates for epilepsy surgery, AHDBS may become a valuable alternative. Controlled studies in larger patient series are mandatory to identify the potential treatment population and optimal stimulation paradigms.