Subscribe to eTOC


A celebration of sorts took place last fall, but the party was in print: Authors of a September 2002 Neurology supplement (Suppl 4) took note of the fifth anniversary of the Food and Drug Administration (FDA) approval of vagus nerve stimulation (VNS) to treat partial epilepsy in drug-resistant patients over age 12.They concluded that the treatment had fulfilled its early promise.

The VNS device, a pacemaker-like generator that is implanted in the upper left side of the patient's chest, has a bipolar lead that tunnels under the skin from the generator to the midcervical area where two helical stimulating electrodes are attached to the left vagus nerve. Doctors can change nerve stimulation settings with an external programming system, and patients can turn the device on and off by holding a magnet over it. Implanting the VNS device is costly, ranging between $10,000 and $20,000.

The FDA approved the VNS device largely on the basis of two large, multicenter, randomized, active-control trials in patients with intractable partial epilepsy (Epilepsia 1994:35:616–626; Neurology 1998:51:48–55). These trials included adults with frequent partial seizures that had not been controlled on antiepileptic drugs (AEDs). After implantation of the VNS, patients were randomized to receive either “low” or “high” stimulation. The results of both trials were similar – after three years, more than 40 percent of patients implanted with the VNS device achieved at least a 50 percent reduction in seizure frequency (Neurology 1999;53:1731–1735).


But despite evidence that VNS might be safe and effective, epilepsy researchers maintain that many questions about VNS remain unanswered. The mechanisms by which VNS reduces seizures are unclear and it is not known which patients are likely to respond.

Epileptologists interviewed for this Special Report said the device is promising, but stressed that VNS should be considered only when a patient plagued with frequent seizures has failed several trials of AEDs, including the newer, broad-spectrum AEDs, and is not a candidate for some type of resective surgery.

“The main mistake with VNS is in placing the devices in patients who are actually good candidates for surgery,” Gregory Krauss, MD, Assistant Professor of Neurology and Director of the Outpatient Epilepsy Center at Johns Hopkins in Baltimore, told Neurology Today. “VNS seldom cures epilepsy, whereas surgery usually does.”


Dr. Gregory Krauss

An exception to this are those patients – particularly children –-with disabling seizures such as drop attacks who are candidates for the more “drastic” surgical procedures like corpus callosotomy. VNS may be worth trying first because of the high morbidity of such surgery. It also may be used in patients in whom corpus callosotomy has failed.

Paul Mullin, MD, Assistant Professor of Neurology at Columbia University in New York, views VNS as a promising complementary therapy – to AEDs, because most patients treated by VNS continue to require AEDs, but at doses low enough to reduce adverse effects, and to surgery, because it can help either nonsurgical candidates or those for whom surgery would pose unacceptable adverse effects. “I think this is the beginning of a new era of epilepsy devices,” Dr. Mullin said.


Working with dogs, physiologist Jacob Zabara, PhD, is credited with first proposing that VNS might antagonize seizures by desynchronizing electrocerebral activities (Electroencephalogr Clin Neurophysiol 1985; 61:162). But Dr. Zabara, who was a co-founder of the private company, Cyberonics in 1987, was not the first to do animal experiments in this area.

In 1938, P. Bailey and F. Bremer reported that VNS in cats elicited synchronized activity in the orbital cortex (J Neurophysiol 1938;1:405–412). A. Zanchetti and colleagues published data on the effect of vagal afferent stimulation on EEG patterns in cats whose seizures were induced by topical strychnine in 1952 (Electroencephalogr Clin Neurophysiol 1952;4:357–361). And Dr. Mullin noted that external mechanical devices made for VNS in patients with epilepsy have been studied since the turn of the last century.


In contrast to the central nervous system toxicity often seen with AEDs, adverse effects of VNS include mainly hoarseness, as well as throat pain and coughing when the stimulator is on. Dr. Mullin said that other side effects include dyspepsia and headache, and rarely, there are wound infections and broken leads.

A vexing problem, however, is the inability to predict which patients will benefit from VNS. Selim Benbadis, MD, Associate Professor of Neurology and Director of the Comprehensive Epilepsy program at the University of South Florida in Tampa, said that before considering VNS, neurologists should first do EEG-video monitoring or other tests to rule out misdiagnosis,especially psychogenic seizures.


Inasmuch as the mechanism for VNS is not well understood, could VNS implantation have a placebo effect? Trying to rule out the placebo effect proved tricky in one of the original trials of VNS. A group of severely affected epilepsy patients was implanted with the VNS device and then randomized to control or treatment groups. Patients in the treatment group could tell when the stimulator was turned on, however, because they experienced a tingling sensation or hoarseness.

The investigators then randomized patients instead to high and low stimulation groups. The 31 patients in the high stimulation group had a mean seizure reduction of 31 percent, with 39 percent of patients having a 50 percent or greater seizure reduction. The 36 patients in the low group had a mean seizure reduction of 11 percent compared with baseline, and only 19 percent of patients had at least a 50 percent seizure reduction (Epilepsia 1994;35:637–643.) However, a 1997 report by the AAN Therapeutics and Technology Assessment Subcommittee raised this question: “Was the blinding distorted by the patients' ability to distinguish whether they were in the high versus low treatment group?” (Neurology 1997;49:293–297).


Dr. Selim Benbadis


Experts differ somewhat regarding what types of epilepsy are helped most by VNS and at what stage of the disorder VNS should considered. Dr. Benbadis and colleagues have proposed an algorithm based on the type of epilepsy (Neurology 2000; 55:1780–1784), suggesting, for example, that nonlesional neocortical epilepsies are “probably best approached with” VNS. They also advocate consideration of nonpharmacologic options early – after the first few drug failures. In general, patients with partial epilepsies or symptomatic generalized seizures are considered good candidates for VNS. Whether patients with primary generalized seizures benefit is not clear, according to Dr. Mullin.

J. Ben Renfroe, MD, of the Child Neurology Center of Northwest Florida in Pensacola, and James W. Wheless, MD, Associate Professor of Neurology and Pediatrics at the University of Texas in Houston, argue for earlier use of VNS in patients refractory to medical therapy. They reported that patients who had had epilepsy for five years or less or who had failed four or fewer standard AEDs before having VNS were more likely to report having no seizures at three months than were patients who had VNS implantation after having had epilepsy for more than five years – 15 percent versus 5 percent (Neurology 2002; 59:S26–S30). The former group was also more likely to have improvement in postictal state and seizure clustering.

“We suggest future prospective studies evaluating VNS therapy versus best medical therapy after the first two to three AEDs have failed, which typically occurs within two years of seizure onset,” they wrote.


Dr. Prakash Kotagal


Although VNS was approved only for patients over age 12 and only for patients with partial seizures, there is considerable off-label use. Many specialists in pediatric epilepsy implant the device in appropriate patients under age 12, including those with generalized seizures and those with Lennox-Gastaut syndrome, a severe, medication-resistant epilepsy syndrome.

At the AAN Annual Meeting in Hawaii, Andreas V. Alexopoulos, MD, and colleagues at the Cleveland Clinic, including Prakash Kotagal, MD, reviewed their preliminary experience with VNS in 32 children under age 18. Twelve children were under age 12; the two youngest were aged 2 and 3. Generalized epilepsy was present in 21 of the 32 children. The median seizure frequency reduction was 86 percent in 27 patients followed for 3 months and 74 percent in 21 patients followed for 6 months. At three months, 18 of 27 patients had greater than a 50 percent reduction of seizures, with no difference in the number of AEDs used before and after VNS.

“VNS worked very well in the children with generalized epilepsy,” Dr. Kotagal told Neurology Today. “In general, our experience with VNS seizure reduction in children is that it is much better than in adults. Why? One reason is that children have had epilepsy for a shorter time than adults. Also, generalized seizures may respond better than partial seizures. Of course, these patients have a very high seizure frequency so going from 20 to 30 seizures per day to three to four per day seems impressive. Still, I think the reduction is robust and is maintained over time.” In addition to the usual hoarseness or voice changes, rare adverse effects in some children include drooling, dysphagia, and increased hyperactivity. However, quality of life often improves in terms of alertness, increased communication, mood, school performance, clustering of seizures, and postictal periods.


Despite many studies, the mechanism by which VNS reduces seizures remains unclear. The cortical desynchronization theory advanced by Dr. Zabara and others has never been completely accepted, partly because the VNS settings effective in treating epilepsy are sub-threshold for activating C-fibers. In 2001, Scott.E. Krahl, PhD, and colleagues showed that C-fibers are neither necessary nor sufficient for VNS to suppress seizures in rats (Epilepsia 2001;42:586–589).

The sensory afferent cell bodies of the vagus nerve reside in the nodose ganglion and relay information to the nucleus of the solitary tract (NTS), explained Mark S. George, MD, Professor of Psychiatry, Radiology and Neurology and Director of the Center for Advanced Imaging Research and the Brain Stimulation Laboratory at the Medical University of South Carolina, Charleston. The NTS then relays this incoming sensory information to other parts of the brain via three main pathways: an autonomic feedback loop, direct projections to the reticular formation in the medulla, and ascending projections to the forebrain largely through the parabrachial nucleus and the locus ceruleus, one of the primary areas in the brain containing norepinephrine (Biol Psychiatry 2000;47:287–295).

The parabrachial nucleus and locus ceruleus directly connect to every level of the forebrain, including the hypothalamus and several thalamic regions that control the insula, orbitofrontal and prefrontal cortex. There are also direct connections to the limbic system.

In 1998, Dr. Krahl and colleagues showed that lesions the locus ceruleus in rats eliminated the ability of VNS to suppress seizures (Epilepsia 1998;39:709–714). This indicated that seizure suppression by VNS may depend on the release of norepinephrine and that noradrenergic agonists might enhance this effect.


No caption available.


But, “questions remain about the chemistry of the neurotransmitters or other substances involved in the VNS mechanism of action in humans,” according to Thomas Henry, MD, Associate Professor of Neurology and Director of the Emory University Epilepsy Center in Atlanta,

Dr. Henry has conducted PET and fMRI studies since 1995 in efforts to sort out the mechanism of action. “Our studies gave the first evidence in humans that VNS has widespread effects over both cerebral hemispheres as well as subcortical structures,” he told Neurology Today. “We did a follow-up study on at the correlation between alterations in blood flow during VNS and subsequent improvement in seizure control. It turns out that the people who had the greatest activation in thalami bilaterally were those who ultimately had the greatest improvement in seizure control” (Neurology1999;52:1166–1173).

Dr. Mullin commented, however, that the importance of cerebral blood flow changes in epilepsy has yet tobe established.


No caption available.


Worldwide, approximately 15,000 patients with epilepsy have undergone implantation of the vagus nerve stimulation device, termed the NeuroCybernetic Prosthesis by its sole manufacturer, Cyberonics Inc., of Houston, Texas (


Although seizure reduction is the main aim of VNS for epilepsy, many neurologists have noted additional salutary effects in implanted patients such as improved mood, cognition and memory. These observations have led to investigation of other possible uses of VNS, such as depression and pain.

Approval has been granted for VNS use for depression in Europe and Canada. However, although some of the pilot work was encouraging, (Biol Psychiatry 2002; 51:280–287), a larger study of 235 nonpsychotic but severely depressed patients at multiple sites failed to show a statistically significant difference between treated and control groups at three months.

The stimulation parameters used in epilepsy may not be the best ones for depression, said Dr. George, who is both a neurologist and psychiatrist and participated in these studies. “The idea is a compelling one and needs more research.”

Article In Brief

  • ✓ Five years after the Food and Drug Administration approved vagus nerve stimulation for partial epilepsy in drug-resistant patients over age 12, researchers are trying to understand its underlying mechanisms.
  • ✓ Candidates for VNS are those with frequent seizures who have failed drug regimens – and generally, nearly half will achieve a 50 percent reduction in seizure frequency – but a majority of patients continue to need antiepileptic drugs (AEDs).
  • ✓ Research suggests that VNS in combination with AEDs might help reduce seizures.
  • ✓ Most VNS for children under 12 is used off-label, but a study presented at this year's AAN Annual Meeting found that VNS worked well in children with generalized epilepsy.