Brief History

Experimental use of VNS to treat epilepsy can be traced to the 1880s.² In 1938, Bailey and Bremmer demonstrated desynchronization of orbital cortex activity with the use of VNS in a cat model.³ Zanchetti and colleagues showed that intermittent VNS reduced or eliminated interictal epileptic events that were chemically induced in the frontal cortex of cats.⁴ In 1980, Radna and MacLean found that VNS caused changes in single-unit activity within the basal limbic structures of squirrel monkeys.⁵ Based on these experiments, Zabara in 1985 proposed that if VNS could desynchronize electroencephalographic activity, it might be effective in attenuating epileptic seizures.⁶ Subsequent animal work by Zabara⁷ and others^8–10 supported Zabara’s hypothesis and allowed clinical trials to be performed in humans.

In 1987, a company—Cyberonics, Inc. (Houston, Tex)—was founded to develop VNS therapy in humans. In 1988, the first epileptic patient to undergo implantation of a VNS therapy device became seizure free.¹¹ Five acute-phase clinical studies analyzing the safety and effectiveness of VNS therapy followed^12–16 and culminated in FDA approval of VNS therapy “for use as an adjunctive therapy in reducing the frequency of seizures in adults and adolescents over 12 years of age with partial onset seizures that are refractory to antiepileptic medications.”¹⁷

Cyberonics, Inc., created a long-term outcome registry that compiled information on patients receiving VNS therapy. The registry opened on November 7, 1997, and closed on April 1, 2003. Participation was voluntary and data were provided by participating physicians. Data were not available for all patients at all time intervals. Median reductions in seizure activity were 46% (n = 4448) at 3 months, 57% (n = 2696) at 1 year, and 63% (n = 1114) after 2 years of VNS therapy.¹⁸ Overall, studies report that in responding patients, the effectiveness of VNS steadily improves over the first 3 to 12 months of stimulation. Many subsequent long-term studies have been published supporting the efficacy, durability, and cost-effectiveness of VNS therapy for epilepsy.^19–33 In addition to refractory partial-onset seizures, VNS is used, off-label, to treat children younger than 12 years with generalized epilepsy and as an adjunct to other surgical procedures (when they are insufficient to control seizure activity).^34–37

Patient Selection

Before considering VNS therapy, patients with medically refractory epilepsy should undergo evaluation at a comprehensive epilepsy center. The evaluation usually includes a complete history and physical examination, video electroencephalographic monitoring to obtain ictal and interictal data, neuropsychological testing, and anatomic and functional neuroimaging. Antiepileptic medications are optimized by an epileptologist, and surgery is considered only after failure of two or more adequate antiepileptic drug trials and the completion of a phase I evaluation. VNS is generally considered palliative because it is rarely curative. Surgical resection of the epileptogenic zone, when indicated, typically achieves higher rates of seizure control. Therefore, VNS is usually offered to those who are not candidates for potentially more effective resective surgical procedures or for patients who refuse resective surgery. In addition, with current clinical data, it is not possible to predict which patients will achieve a favorable response to VNS therapy or in what way VNS therapy will affect their seizure control. VNS therapy, when effective, can cause decreases in seizure duration, frequency, or intensity and may make it feasible to decrease the amount of antiepileptic medications needed to achieve adequate seizure control.

Once identified as candidates for the device, patients should be counseled about the potential risks and benefits of VNS therapy, as well as the risks and benefits of other therapeutic options, including resective or other surgeries, further medication trials, and a ketogenic diet (in the case of eligible children). It is also important to describe the long-term outcome and seizure control data that are available, including the possible length of time needed to achieve full efficacy of the device. This approach will result in more realistic expectations and better compliance.

Anatomic Considerations

The majority of vagal nerve fibers are general somatic and special visceral afferents projecting to the brain, along with efferent projections to the larynx and parasympathetic projections to the heart, lungs, and gastrointestinal tract. The VNS electrode is applied to the midcervical portion of the vagal nerve, which is relatively free of branches. The upper cervical vagal nerve gives off branches to the pharynx, carotid sinus, and superior and inferior cardiac branches leading to the cardiac plexus. Studies in dogs suggest that the right vagal nerve preferentially innervates the sinoatrial node of the heart whereas the left vagal nerve projects to the atrioventricular node. Accordingly, the NCP electrode is usually applied to the left vagal nerve to avoid stimulation-related asystole or bradycardia.³⁸ A small series of right-sided VNS implants in children, which included Holter monitoring of patients after surgery, failed to demonstrate any changes in heart rate with stimulation.³⁹

The recurrent laryngeal nerve travels with the main vagal nerve trunk and then branches caudally at the aortic arch before ascending in the tracheoesophageal groove. As a result, changes in the character and quality of voice are common during nerve stimulation in the early period after NCP implantation. The voice changes usually resolve within weeks of surgery.

Other nerves in the region of the vagal nerve can be affected at surgery. The phrenic nerve lies deep to the carotid sheath, and unilateral paralysis of the left hemidiaphragm has been reported during periods of VNS. Hypoglossal and facial nerve fibers are found well above the midcervical trunk, but injuries to both have been reported after VNS implantation. The sympathetic trunk runs deep to the common carotid artery and provides fibers that ascend with the internal carotid artery. There is a report of Horner’s syndrome developing after VNS implantation, presumably caused by injury to the sympathetic plexus or the fibers along the internal carotid.⁴⁰

Neurocybernetic Prosthesis

The NCP has two implantable components: a generator and a stimulating electrode (Figs. 67-3 to 67-6; also see Figs. 67-1 and 67-2). The generator consists of an epoxy resin header with a receptacle for the connector pin or pins from the electrode and a titanium module containing a lithium battery and the generator. The electrode is secured to the connector pin receptacle with a set screw or screws tightened with a hexagonal torque wrench included with the generator packaging. The generator contains an antenna that receives radiofrequency signals from the programming telemetry wand and transfers them to a microprocessor that regulates the electrical output of the pulse generator. The generator delivers a charge-balanced waveform characterized by five programmable parameters: output current, signal frequency, pulse width, signal-on time, and signal-off time. Higher stimulation frequencies and longer signal-on times result in a shorter duration of battery service life. The NCP electrode is insulated with a silicone elastomer and can be implanted safely in patients with latex allergies. One end of the lead has a connector pin or pins that insert directly into the generator (see Figs. 67-3 and 67-4); the other end has an electrode array consisting of three discrete helical coils that are placed around the vagal nerve (see Figs. 67-5 and 67-6). The middle and distal coils are the positive and negative electrodes, respectively, and the most proximal coil serves as an anchoring tether to prevent excessive force from being transmitted to the electrodes when patients turn their neck. Each electrode helix contains three loops. Embedded inside the middle turn is a platinum coil that is welded to the lead wire. Suture tails extending from either end of the helix allow manipulation of the coils without injuring the platinum contacts. A silicone electrode collar is included with the electrode and is used to anchor the electrode to the soft tissue of the neck, proximal to the helical coils. The portion of the electrode between the electrode collar and the inferior helix creates a “strain release loop” that further protects the vagal nerve from unwanted traction. A handheld NCP magnet performs several functions. When passed over the chest wall overlying the generator, it triggers stimulation superimposed on the baseline output. This on-demand stimulation can be performed by a patient or caregiver at the onset of an aura and can sometimes diminish or abort an impending seizure. In addition, if the NCP appears to be malfunctioning or if the patient wishes to terminate stimulation for any other reason, the system can be turned off by placing the magnet over the generator site continuously.

FIGURE 67-3 Diagram demonstrating connection of a monopolar electrode to a monopolar vagal nerve stimulation generator.

FIGURE 67-4 Diagram demonstrating connection of a bipolar electrode to a bipolar vagal nerve stimulation generator. Note that the positive lead is inserted into the inferior pin site. The positive electrode pin has a white mark.

FIGURE 67-5 Monopolar vagal nerve stimulation electrode.

FIGURE 67-6 Bipolar vagal nerve stimulation electrode.

The NCP has undergone a series of revisions since introduction of the model 100 (see Fig. 67-1). The original model 100 and the second-generation model 101 were used with a bipolar helical lead. The third- and fourth-generation models (102 and 103) incorporated a monopolar lead. Generators 102R and 104 have bipolar lead acceptors, so revision of models 100 and 101 (with bipolar electrodes) can be performed without replacing the electrodes (see Figs. 67-3 and 67-4). The original programming hardware included a programming wand attached to a laptop computer. The laptop has been replaced with a personal digital assistant (PDA) and a similar programming wand (Fig. 67-7