Q: How do you approach patient selection for interventions such as neurostimulation (ie, device therapy)?

Antiepileptic drugs (AEDs) remain the mainstay of epilepsy treatment; however, approximately 30% of patients have what is known as drug-resistant epilepsy. It is usually at that point, after at least 2 trials of appropriately selected AED therapy, that we discuss all of the other categories of treatment options, which include devices, surgery, and, in rare cases, dietary therapy. Of the nondrug treatments, only resective surgery can be potentially curative (ie, seizure freedom, on or off medications); however, a large proportion of patients are ineligible for resection. With respect to dietary changes, many patients simply do not want to endure such severe lifestyle changes for what is likely to be an indefinite period of time. 

Three devices have earned US Food and Drug Administration (FDA) approval to target epileptogenic foci in conjunction with AED therapy. Whether you are considering DBS, VNS, or even RNS, consideration of the whole patient, including their goals and state in life, is key. VNS is a peripheral device, and DBS and RNS are placed into the substance of the brain (or some RNS candidates may only have electrodes on the brain surface). When considering these interventions, an initial discussion often centers on which device may be best for an individual patient. The evidence for DBS in generalized epilepsies is limited at this point. For certain focal epilepsies, DBS can be a very robust therapy that, when combined with AED therapy, can improve seizure control. A relative advantage with all of these devices is that, when we combine them with pharmacotherapy, we are not adding the side effects of another AED. So, you are not adding another AED to women of childbearing age and you are not introducing the potential for drug-drug interactions in patients being treated for high cholesterol or hypertension, for example. Like resective surgery, neuromodulation with implantable neurostimulation devices is considered to be relatively safe. 

Practical and cosmetic issues do come into play, however. The cranium must be thick enough to support the device (RNS) to be implanted in or on the brain, which generally does not occur until adulthood. Cosmetic-related issues also must be considered. Currently, the generators for DBS devices are bulky enough to be noticeable when implanted in the chest wall, which could present issues, particularly for younger patients who may be more concerned about their cosmetic appearance.

DBS is no longer considered an experimental therapy for epilepsy. There has been a randomized clinical trial with long-term follow-up to prove the safety and efficacy of this type of intervention, and the FDA has approved its use in adults with medically intractable focal onset seizures. Because it is unlikely to be curative, DBS is reserved for patients who are either not candidates for resective epilepsy surgery or who, unfortunately, may still not be completely seizure free after surgery. VNS and RNS are also certainly worth discussing after a patient with epilepsy is found to not be a candidate for resective epilepsy surgery. I offer DBS and VNS as an option after a medically intractable patient has tried 2 appropriately selected AEDs that have not worked, and only after a complete workup has determined that the patient is not a candidate for surgical resection. 

Regarding the question of which device to choose based on efficacy, this is an ongoing question. No head-to-head comparisons of DBS, VNS, and RNS have been performed. The only device approved for children with refractory epilepsy is VNS, which can be used in those as young as 4 years of age. RNS requires the creation of a cavity in the skull to house the battery and the recording unit, which would present significant difficulty in very young patients whose skulls have not yet fully formed. RNS may also have a positive effect on mood and cognition; these benefits are less clear for DBS at this time. Some studies suggest that VNS may have beneficial effects on mood. VNS has a separate FDA indication for severe, recurrent unipolar and bipolar depression; that is something to take into account when treating a patient with refractory seizures and comorbid depression.

DBS, RNS, and VNS are all acceptable therapies for medication-resistant focal epilepsy. However, no head-to-head data are available to compare the devices to one another. This sometimes leads to difficulty in deciding which treatment would potentially be best for a patient. Slight differences in indication can sometimes help in this decision making, but all 3 devices could potentially be reasonable options for an individual patient. In general, neurostimulation efficacy does not decrease over time and may, in fact, increase.

The seizure type may be most relevant. Some evidence suggests that VNS has efficacy for generalized onset seizures, while there is no such evidence for RNS and the efficacy with DBS in this setting is still being investigated, along with consideration of stimulation targets other than the  FDA-approved target of the anterior nucleus of the thalamus. DBS and RNS are both indicated for focal seizures but differ in the certainty of the localization expected. For RNS, a specific target for electrode placement is ideally needed, whereas with DBS, placement is predetermined and not dependent on seizure localization. Although these and other factors could influence decision making, there is still active discussion about which devices might be best used in any given case. For instance, Benbadis et al recently published a helpful algorithm on the use of epilepsy surgery and different forms of neurostimulation. Epileptologists are still learning how best to use each device; however, DBS, VNS, and RNS are all exciting due to the benefits that they can offer in the treatment of medication-resistant focal epilepsies. As with medication, where determining the best dose to use and the timing to apply requires structuring the trials to minimize the variables, the same issues exist for these devices (ie, optimal electrode placement and treatment parameters must be defined and characterized).