erator, which may abort seizures in some patients if applied at the seizure onset.

As patients with refractory partial epilepsy must often undergo magnetic resonance imaging (MRI), it is critical that injury to the patient from the generator or electrode lead does not occur during the scan and, furthermore, that the quality of the MRI scan is not impaired. Studies showed that the NCP system is safe and the quality of the diagnostic information is not affected when a head coil is used for brain MRI (5). VNS is not affected by cellular phones, whether analog or digital (5).

MECHANISM OF ACTION

Explanations for the molecular mechanisms of pharmacologic antiseizure interventions, especially partial seizures, are usually classified into five categories: (i) blocking ion currents (sodium, potassium, calcium) across neuronal membranes; (ii) increasing brain inhibition by enhancing GABAergic neurotransmission; (iii) attenuation of glutamatergic excitatory neurotransmission; (iv) modifying the monoaminergic regulation of seizure control; and (v) unknown as of now.

The mechanism by which VNS modulates seizure control has not been fully elucidated. Traditionally, the vagus nerve has been considered to be a parasympathetic efferent nerve regulating autonomic functions such as heart rate and gastric tone. However, the vagus nerve (cranial nerve X) is actually a mixed nerve composed of approximately 80% afferent sensory fibers that carries information to the brain from the head, neck, and abdomen. The sensory afferent cell bodies reside in the nodose ganglion and relay information to the nucleus tractus solitarius (NTS) and, from there, onward to many brain areas, including the frontal lobe and the limbic system (2).

Recently, Walker and co-workers (6) found that microinjections of GABA or a glutamate antagonist into the NTS block experimental seizures in rats. As the NTS has direct connections to the locus coeruleus (LC) and the forebrain (in rats), the observation by Krahl et al. (7) is relevant. They reported that destroying the LC eliminated the ability of VNS to suppress seizures in rats. The LC is the site of many norepinephrine-containing neurons that have important connections to the limbic system, including the amygdala, hypothalamus, and orbitofrontal cortex, areas linked to the modulation of seizure control, mood, and anxiety. These findings recall the monoamine theory of epi

lepsy, which was popular before the current GABA and glutamate theories.

Recent experiments in amygdala-kindled cats show that VNS may, at least partly, delay the kindling process. Amygdala kindling in cats is one of several models used to study the effect of interventions on the progression of partial epilepsy (8). However, it would be premature to extrapolate these findings in laboratory animals to human epilepsy. Nevertheless, the results are encouraging and may prompt in-depth evaluation of VNS on the progressive aspects of partial epilepsy, such as seizure intractability and memory disturbances.

EFFICACY AND CLINICAL UTILITY

In patients with epilepsy, VNS reduces seizure frequency, attenuates seizure severity, and is associated with positive changes in alertness and mood. Further, some patients are able to abort seizures with the magnet. All of these factors contribute to the clinical utility of VNS (9–11).

Seizure reduction has been shown in a total of five controlled trials, including two double-blind, activecontrolled trials: the E03 and E05 trials (12, 13). In both trials, the primary efficacy analysis was percentage change in total seizure frequency during treatment with VNS versus baseline, comparing the two treatment groups: high stimulation (30 Hz, 30 s on, 5 min off, 500-microsecond pulse width) and low stimulation (1 Hz, 30 s on, 90–180 min off, 130-microsecond pulse width). In the E03 study, the high-stimulation group had a mean reduction in seizure frequency of 24.5% versus 6.1% for the low-stimulation group (P 5 0.01). In the E05 study, the reductions in seizure frequency were 28 and 15%, respectively (P 5 0.039).

Although the reasons remain elusive, the seizure prophylactic effects of VNS take time to work. In a prospective, 12-month, open-label assessment of 195 patients who had participated in the E05 study, all patients who had received high stimulation during the study were continued within recommended settings (0.25–3.5 mA, 7–60 s on, 1.1–180 min off, 20- to 30-Hz frequency, 500- to 750-microsecond pulse width) and all those who were on low stimulation during the double-blind trial were transferred to high stimulation (as defined for the E05 study) during the 12-month, open-label evaluation period (10). The primary efficacy outcome of the continuation study was the percentage change in total seizure frequency at 3 and 12 months compared with the 3-month preimplantation

baseline. At 3 months, the median reduction was 34%; at 12 months, it was 45% (P = 0.0001, 12 months vs 3 months). Thus, the efficacy of VNS improved in this study over 12 months of treatment, and 20% of patients achieved 75% or greater reduction in seizure frequency during the first year of treatment.

Additional supportive evidence for progressive improvement in seizure control over time comes from the company-sponsored voluntary patient registry program, which can be accessed through a website by participating physicians. Among 125 patients from the registry who were assessed by their physicians 12 months after implantation, the median seizure reduction was 58%, as compared with 33% at 3 months. Further, 22% of these patients had a 90% or more seizure reduction compared with baseline at 12 months, which was more than double the effect at 3 months (10%).

The long-term outcome of VNS was studied in 15 patients with refractory partial epilepsy who were followed for 29 months (11). The mean stimulation output was 2.25 mA. Four patients (27%) were completely seizure-free for $12 months. In 1 patient, one antiepileptic drug (AED) was tapered, in 10 patients drug treatment was unchanged, and in 4 patients one AED was added. While adjustment of AEDs may confound the specific contribution of VNS, these results suggest that VNS remains effective during long-term treatment of refractory partial epilepsy.

The E03 and E05 trials did not include formal evaluations of mood, alertness, or memory, though several investigators noted patient-perceived mood improvements (12, 13). More recently, an analysis of 636 patients entered into the patient registry showed patientperceived improved alertness in 59% of patients, 43% with improved mood, and 31% with better memory 12 months after implantation. Recently published retrospective data from Germany (14) and a prospective study in the United States (15) suggest that VNS reduces depressive symptoms in patients with epilepsy unrelated to changes in seizure control.

PLACE OF VNS IN THE TREATMENT OF EPILEPSY

Even though a proportion of patients become seizure- free with VNS, the vast majority need to continue their AED(s). However, with more experience in the effects of VNS in easier-to-treat epilepsies and ad-

vances in our understanding of how to individualize treatment, VNS may eventually play a future role as monotherapy first-line treatment.

If no improvement is seen after 2 years of adequate use of VNS, the stimulation current should be set at 0 mA for several months to confirm that no improvement has occurred. If this is the case, the generator can be safely removed through local anesthesia (as for any cardiac pacemaker). While the electrodes are most often left in place, successful removal of the electrodes has been recently reported (16).

A group of U.S. epilepsy experts recently wrote "The degree of improvement in seizure control from VNS remains comparable to that of new AEDs, but is lower than that of mesial temporal lobectomy in suitable surgical resection candidates. Efficacy of VNS in less severely affected populations remains to be evaluated. Nevertheless, sufficient evidence exists to rank VNS for epilepsy as effective and safe" (17). Consistent with these statements, most physicians recommend and use VNS after failure of three or more AEDs that are appropriate for partial seizures and that are pushed to maximally tolerated doses in patients with a verified diagnosis of partial epilepsy when resective temporal lobectomy for mesial temporal lobe epilepsy has been excluded as an option or previous resective surgery was not successful (Table 1).

Similarly, a group of European epilepsy experts recently concluded, based on controlled trials in partial epilepsy and clinical observations in refractory Lennox–Gastaut syndrome, that VNS is a palliative surgical procedure similar in efficacy to the newer AEDs for patients who cannot be treated sufficiently with existing AEDs or resective epilepsy surgery (18). Thus, although its value for treating patients with refractory generalized epilepsy is less well documented, VNS may be an option in refractory Lennox– Gastaut syndrome and related syndromes of catastrophic epilepsies in childhood and adolescence if standard treatment with three or more adequate AEDs does not achieve sufficient seizure control or proves to be intolerable (19).

In a subgroup of 24 patients from study E04 who had exclusively generalized seizures and only generalized epileptiform activity or generalized slowing in the electroencephalogram, 3-month treatment with VNS was compared with a 1-month baseline before implantation (19). Seven patients had an idiopathic etiology and 17 had a symptomatic etiology. Median seizure reduction was 46% and 11 of the 24 patients had at least a 50% reduction in seizure frequency. Side effects were mild. The open data from this short-term

1. No craniotomy is required.
2. The number of seizures is reduced by at least half in approximately 50% of patients with refractory partial epilepsy and Lennox–Gastaut syndrome.
3. Patient may have self-control over severe seizures by magnet.
4. Treatment compliance is ensured.
5. There are no interactions with AEDs.
6. VNS is well tolerated and well accepted by patients.

1. Only a small minority of patients become seizure-free.
2. Efficacy is difficult to predict prior to implantation.
3. Battery changes are required.
4. Voice changes and dyspnea on exertion may occur with stimulation.

clinical study suggest that VNS may have a beneficial effect on seizure frequency in children (aged 4 and older), adolescents, and adults with refractory generalized epilepsy.

Regardless of seizure type or epilepsy syndrome, written materials that provide information for patients, relatives, and other caregivers about VNS are helpful when implementing this therapy (20).

TOLERABILITY AND SAFETY

Complications of the surgical implantation are rare and transient, including (in declining incidence) infection (1.5%), vocal cord paresis (1%), and unilateral facial weakness (1%) (see DeGiorgio et al. (22) for a detailed discussion of the surgical technique). Perioperative adverse events reported by at least 10% of patients in the E03 and E05 trials were pain (29%), coughing (14%), voice changes (13%), chest pain (12%), and nausea (10%) (12, 13). Although a number of stimulation-related adverse events were reported during the E03 and E05 trials, the only adverse events that occurred significantly more often in the highstimulation treatment group than in the low-stimulation group were dyspnea and voice alteration. Adverse events were judged to be mild and transient in almost all patients (Table 2). No cognitive, sedative, visual, affective, or neurologic deficits were reported. There were no relevant changes seen in hematology or routine chemistry testing (4). No deaths occurred.

Given that the vagus nerve modulates cardiac function, the effects of VNS on cardiac function were studied extensively in more than 250 patients, mainly with 24-hour Holter monitors. The results showed no measurable effect of VNS on heart rate compared with baseline (13). Similarly, no cardiac problems have been encountered when VNS is turned on for the first

time in the outpatient setting, typically 2 weeks after implantation (22, 23). However, transient asystole lasting 10 to 20 s has been reported during the implantation procedure in 8 patients out of the estimated 7000 patients who have been implanted (1.1 per 1000) (24). Each of these instances occurred in the operating room during the first diagnostic testing of the lead—the so-called lead test. The lead test consists of using the generator to stimulate the left vagus nerve for approximately 15 s at 1.0 mA, 500 microseconds, and 20 Hz. Four of the eight patients elected to go ahead with the implant, while in the other four patients, the generator and the electrodes were removed. All of the patients recovered without sequelae. The reason(s) for this rare but potentially serious effect is unclear.

With regard to pulmonary function, no changes on routine measures were seen in clinical studies (4). More recently, the effects of VNS on four patients with sleep-related breathing were reported (25), suggesting that VNS should be administered with care in patients with preexisting obstructive sleep apnea (OSA) and that lowering the stimulus frequency or prolonging the off-time in these patients may prevent exacerbation of OSA. Because OSA may occur unrecognized in one of three patients with medically refractory partial

epilepsy (26), it would seem prudent to diagnose preexisting sleep apnea in candidates for VNS and to take appropriate action to avoid exacerbation during VNS treatment.

There are limited data on the effects of VNS on pregnancy. Of eight women treated with adjunctive VNS during pregnancy, five had normal babies, including one set of twins (27). One unplanned pregnancy was terminated with an elective abortion and one pregnancy was aborted based on abnormal in utero fetal development. One additional patient reported a spontaneous abortion, though the actual pregnancy was not confirmed.

CONCLUSION

VNS is a well-tolerated, effective, and safe form of neurostimulation similar in efficacy for partial seizures to the newer AEDs. This therapy is appropriate for patients who cannot be treated sufficiently with existing AEDs or resective surgery. In addition, clinical observations suggest that VNS treatment may be useful in patients with refractory Lennox–Gastaut syndrome and related syndromes.

REFERENCES

1. Schachter SC, Schmidt D, editors. Vagus nerve stimulation. London: Martin Dunitz, 2000.
2. Henry TR. Anatomical, experimental, and mechanistic investigations. In: Schachter SC, Schmidt D, editors. Vagus nerve stimulation. London: Martin Dunitz, 2000:1–30.
3. Henry TR. 10 most commonly asked questions about vagus nerve stimulation for epilepsy. Neurologist 1998;4:284–9.
4. Schachter SC, Saper CB. Vagus nerve stimulation. Epilepsia 1998;39:677–86.
5. Nyenhuis JA, Bourland JD, Foster KS, Graber GP, Terry RS, Adkins RA. Testing of MRI compatibility of the Cyberonics Modell 100 NCP® generator and Model 300 series lead. Epilepsia 1997;38(suppl 8):S140.
6. Walker BR, Easton A, Gale K. Regulation of limbic motor seizures by GABA and glutamate transmission in nucleus tractus solitarius. Epilepsia 1999;40:1051–7.
7. Krahl SE, Clark KB, Smith DC, Browning RA. Locus coeruleus lesions suppress the seizure-attenuating effects of vagus nerve stimulation. Epilepsia 1998;39:709–14.
8. Fernandez-Guardiola A, Martinez A, Valdes-Cruz A, Magdaleno-Madrigal VM, Martinez D, Fernandez-Mas R. Vagus nerve prolonged stimulation in cats: effects on epileptogenesis (amygdala electrical kindling): behavioral and electrographic changes. Epilepsia 1999;40:822–9.
9. Schachter SC. Efficacy, safety, and tolerability. In: Schachter SC, Schmidt D, editors. Vagus nerve stimulation. London: Martin Dunitz, 2000:51–64.

10. DeGiorgio CM, Schachter SC, Handforth A, et al. Prospective long-term study of vagus nerve stimulation for the treatment of refractory seizures. Epilepsia 2000;41:1195–200.
11. Vonck K, Boon P, D’Have M, Vandekerckhove T, O’Connor S, De Reuck J. Long-term results of vagus nerve stimulation in refractory epilepsy. Seizure 1999;8:328 –34.
12. Ben-Menachem E, Manon-Espaillat R, Ristanovic R, et al. Vagus nerve stimulation for treatment of partial seizures. 1. A controlled study of effect on seizures. First International Vagus Nerve Study Group. Epilepsia 1994;35:616–26.
13. Handforth A, DeGiorgio CM, Schachter SC, et al. Vagus nerve stimulation therapy for partial-onset seizures: a randomized active-control trial. Neurology 1998;51:48–55.
14. Elger G, Hoppe C, Falkai P, Rush AJ, Elger CE. Vagus nerve stimulation is associated with mood improvements in epilepsy patients. Epilepsy Res 2000;42:203–10.
15. Harden CL, Pulver MC, Ravdin LD, Nikolov B, Halper JP, Labar DR. A pilot study of mood in epilepsy patients treated with vagus nerve stimulation. Epilepsy Behav 2000;1:93–9.
16. Espinosa J, Aiello MT, Naritoku DK. Revision and removal of stimulating electrodes following long-term therapy with the vagus nerve stimulator. Surg Neurol 1999;51:659–64.
17. Fisher RS, Handforth A. Reassessment: vagus nerve stimulation for epilepsy: A report of the Therapeutics and Technology Assessment Subcommittee of the American Academy of Neurology. Neurology 1999;53:666–9.
18. Schmidt D, Elger CE, Stefan H, et al. The place of vagus nerve stimulation in the treatment of epilepsy [in German]. Nervenheilkunde 1999;18:558–61.
19. Labar D, Murphy J, Tecoma E. Vagus nerve stimulation for medication-resistant generalized epilepsy. E04 VNS Study Group. Neurology 1999;52:1510–2.
20. Schmidt D, Bourgeois B. A risk–benefit assessment of therapies for Lennox–Gastaut syndrome. Drug Saf 2000;22:467–77.
21. Schmidt D. Twenty-one questions you may have about vagus nerve stimulation. In: Schachter SC, Schmidt D, editors. Vagus nerve stimulation. London: Martin Dunitz, 2000:73– 84.
22. DeGiorgio CM, Amar A, Apuzzo MLJ. Surgical anatomy, implantation technique, and operative complications. In: Schachter SC, Schmidt D, editors. Vagus nerve stimulation. London: Martin Dunitz, 1999;31–50.
23. Asconape JJ, Moore DD, Zipes DP, Hartman LM, Duffell WH. Bradycardia and asystole with the use of vagus nerve stimulation for the treatment of epilepsy: a rare complication of intraoperative testing. Epilepsia 1999;40:1452–4.
24. Tatum WO, Moore DB, Stecker MM, et al. Ventricular asystole during vagus nerve stimulation for epilepsy in humans. Neurology 1999;52:1267–9.
25. Andriola MR, Rosenzweig T, Vlay S. Vagus nerve stimulator (VNS): induction of asystole during implantation with subsequent successful stimulation. Epilepsia 2000;41(suppl 7):223.
26. Malow BA, Edwards J, Marzec M, Sagher O, Fromes G. Effects of vagus nerve stimulation on respiration during sleep: a pilot study. Neurology 2000;55:1450–4.
27. Malow BA, Levy K, Maturen K, Bowes R. Obstructive sleep apnea is common in medically refractory epilepsy patients. Neurology 2000;55:1002–7.
28. Ben-Menachem E, Ristanovic R, Murphy J. Gestational outcomes in epilepsy patients receiving vagus nerve stimulation. Epilepsia 1998;39(suppl 6):S180.