VAGUS NERVE STIMULATION

DEPRESSION

0193-953X/OO $15.00

+ .OO

VAGUS NERVE STIMULATION A Potential Therapyfor Resistant Depression? Mark S. George, MD, Harold A. Sackeim, PhD, Lauren B. Marangell, MD, Mustafa M. Husain, MD, Ziad Nahas, MD, Sarah H. Lisanby, MD, James C. Ballenger, MD, and A. John Rush, MD

OVERVIEW

For more than a century, neuroscientists have tried to develop noninvasive methods to stimulate the brain, thereby avoiding brain surgery and open craniotomy. As early as 1938, some investigators conjectured that vagus nerve stimulation(VNS) would alter higher brain activity. Within the past 20 years, VNS in the neck has become routine, and a specific type of VNS has been approved by the US Food and Drug Administration for treatment-resistant partial onset seizure disorder,

Several authors holdresearch contracts(MSG, AJR, HAS,LBM) or grants (MSG, AJR) from Cyberonics, Inc., the manufacturer of the NCP System to deliver vagus nerve stimulation (VNS). No author hasa direct financial interestin Cyberonics (e.g., stocks or consulting boards),and nocompensation was given for writing this article. NCP and VNS are trademarks of Cyberonics, Inc.

From the Departments of Psychiatry (MSG, ZN, JCB), Radiology (MSG), and Neurology (MSG), Medical University of South Carolina; the Ralph H.Johnson Veterans Hospital (MSG), Charleston, South Carolina;the Department of Biological Psychiatry, New York State Psychiatric Institute; the Departments of Psychiatry (HAS, SHL) and Radiology (HAS), College of Physicians and Surgeons, Columbia University, New York, New York; the Department of Psychiatry, Universityof Texas southwestern Medical Center at Dallas, Dallas (AJR, MMH); and the Department of Psychiatry, Baylor College of Medicine, Houston, Texas (LBM)

THE PSYCHIATRIC CLINICS OF NORTH AMERICA VOLUME 23 * NUMBER 4 DECEMBER 2000

757

758

GEORGE et a1

with 9000 patients receiving implants worldwide. This year, a pilot study51 using VNS in subjects with treatment-resistant depression showed encouraging results. This article summarizes the theory and practice of VNS and addresses many of the issues that must considered in the development of VNS as a potential antidepressant therapy and neuroscience probe. VNS seems to be a promising new somatic intervention that also may improve the understanding of brain function. History of Vagus Nerve Stimulation

Interest in whether and how autonomic functions modulate activity in the limbic system and higher cortex has been longstanding. Numerous studies have identified extensive projections of the vagus nerve by its sensory afferent connections in the nucleus tractus solitarius (NTS) to diverse brain regions, including: 1938-Bailey and Bremer: In cats, VNS elicited synchronous orbital cortex activity. 1949-MacLean and Pribram: In monkeys, VNS caused lateral frontal changes. 1951-Dell and Olson: In awake cats with high cervical spinal section, VNS evoked a slow wave response in the anterior rhinal sulcus and amygdala. 1980-MacLean: In monkeys VNS, caused marked single unit effects on cingulate, thalamus, and basal limbic structures. 1985-Zabara: In dogs, VNS caused cortical EEG changes and stopped seizures. 1988-Kiffin Penry (Bowman-Gray): First human implant, resistant epilepsy patient. 1998-MUSC, UTSW, New York State Psychiatric Institute, Baylor: Resistant depression. In 1938, Bailey and Bremer3 reported that in cats, VNS elicited synchronized activity in orbital cortex. In 1949, M a ~ L e a nstimulated ~~ the vagus nerve, recorded EEG from the cortical surface of anesthetized monkeys, and found inconsistent slow waves generated from the lateral frontal cortex. In 1951, Dell and Ols0n14found that VNS evoked a slow wave response in the anterior rhinal sulcus and amygdala in awake cats with high cervical spinal section. Reasoning from this body of literature, Zabara” 73 demonstrated an anticonvulsant action of VNS on experimental seizures in dogs. Although the vagus nerve is an autonomic nerve, based on known anatomy, Zabara hypothesized that VNS could prevent or control the motor and autonomic components of epilepsy. Z a b a ~ ainitially ~~ hypothesized

VAGUS NERVE STIMULATION

759

that VNS had two distinct antiepileptic mechanisms of action: (1) a direct inhibition terminating the beginning or ongoing seizure and (2) a long-lasting inhibition, which increased with continued periods of stimulation. He observed that the inhibitory effect on seizures outlasted the VNS period by approximately four-fold in the acute model and speculated that the anticonvulsant effect would be longer in a chronic model. Penry et a147,52 ushered in the modern use of VNS with the first human implant for the treatment of epilepsy in 1988. Vagus Nerve Anatomy

Traditionally, the vagus nerve has been considered a parasympathetic efferent nerve (controlling and regulating autonomic functions e.g. heart rate and gastric tone); however, the vagus nerve is a mixed nerve composed of approximately 80% afferent sensory fibers carrying information to the brain from the head, neck, thorax, and abdomen.18 The sensory afferent cell bodies of the vagus nerve reside in the nodose ganglion and relay information to the NTS. These fibers are different from those that go to the other motor nuclei of the vagus nerve (Fig. 1). The NTS relays this incoming sensory information to the rest of the brain through three primary pathways: (1) an autonomic feedback loop, ( 2 ) direct projections to the reticular formation in the medulla, and (3) ascending projections to the forebrain largely through the parabrachial nucleus (PB) and the locus ceruleus (LC). The PB sits adjacent to the LC (one of the primary norepinephrine-containing areas of the brain; Fig. 2). Lesioning the LC in rats eliminates the ability of VNS to suppress seizures%associating LC stimulation and modulation of noradrenergic transmission with the anticonvulsant action of VNS (Fig. 2). The PB and LC directly connect with various forebrain structures, including the hypothalamus and several thalamic regions that control the insula, orbitofrontal cortex and prefrontal cortex. Perhaps important for mood regulation, the PB and LC have direct connections to the amygdala and the bed nucleus of the stria terminalis-structures that are associated with emotion recognition and mood regulation66 (for reviews of the functional neuroanatomy of depression, see references 21 and 34). These brain stem and limbic anatomic connections seem to have functional consequences. The oncogene C-fos is a general marker for cellular activity. C-fos studies in rats during VNS reveal increased activity in the amygdala, cingulate, LC, and hypothalamus.43Walker et a16*have outlined a possible role of the NTS in how VNS reduces seizures. By microinjecting the NTS with y-aminobutyric acid (GABA) agonists or glutamate antagonists, they found that increased GABA or decreased

0

m

U

Figure 1. A, Cross-sectional view of the brain stem, illustrating the origins of several components of the vagus nerve (X), which is a mixed sensory and motor nerve. Efferent motor fibers originate in the dorsal nucleus of the vagus, whereas afferent motor fibers go to the nucleus ambiguus. Afferent sensory fibers, which make up 80% of the left vagus, terminate in the solitary tract, which projects to the midline raphe and locus and likely is the path by which vagus nerve stimulation has antiseizure and other neuropsychiatric effects. B,Demonstration of the other connections of the vagus. C, A transverse slice through the brain stem at the level of the nucleus tractus solitarius. NUC = nucleus; V = trigeminal.

VAGUS NERVE STIMULATION

761

Figure 2. Known connections of the nucleus tractus solitarius to the parabrachial nucleus and the locus ceruleus. Lesioning the locus ceruleus in rats eliminates the antiepileptic properties of vagus nerve stimulation. The locus ceruleus is the site of many norepinephrine-containing neurons, which have important connections to the amygdala, hypothalamus, insula, thalamus, orbitofrontal cortex, and other limbic regions linked to mood regulation. (from George MS, Sackeim HA, Rush AG, et al: Vagus nerve stimulation: A new tool for brain research and therapy. Biol Psychiatry 47:287-295, 2000; with permission.)

glutamate in the NTS blocked seizures. Indirectly, these findings suggest that VNS may change NTS GABA and glutamate concentrations, with secondary changes in the function of specific limbic structures mentioned earlier (e.g., the amygdala, cingulate, insula, and hippocampus). Incoming sensory (i.e., afferent) connections of the vagus nerve provide direct projections to many of the brain regions associated with neuropsychiatric disorders. These connections provide a basis for understanding how VNS might be a portal to the brain stem and connected limbic and cortical regions. These pathways likely account for the neuropsychiatric effects of VNS, and they invite additional theoretic considerations for potential research and clinical applications. How Is Vagus Nerve Stimulation Performed?

The term vugus nerve stimulation generally refers to several different techniques used to stimulate the vagus nerve, including studies of animals in which the vagus nerve was accessed through the abdomen

762

GEORGE et a1

and diaphragm. For practically all studies in humans, VNS refers to stimulation of the left cervical vagus nerve using the NeuroCybernetic Prosthesis (NCP) Systems9(Cyberonics, Houston, TX; Fig. 3). VNS has been commercially available for the treatment of resistant partial onset seizures in patients with epilepsy in Europe since 1994 and in the US since 1997. Approximately 9000 people worldwide have had these stimulators implanted. Typically, epileptic patients considering VNS have had unsatisfactory seizure control, commonly with several medications, and VNS is an option before brain surgery for some patients. In many ways, VNS delivered through the NCP System is much like cardiac pacemakers. In both cases, a subcutaneous generator sends an electric signal to an organ through an implanted electrode. The surgery for implanting the NCP pulse generator in the chest is much like inserting a cardiac pacemaker.’ Of course, the two techniques differ in site of stimulation. VNS is delivered through the NCP pulse generator, an implantable, multiprogrammable, bipolar pulse generator (about the size of a pocket watch) that is implanted in the left chest wall to deliver electric signals to the left vagus nerve through a bipolar lead. The electrode is wrapped around the vagus nerve in the neck, near the carotid artery, using a separate incision and is connected to the generator subcutaneously. Although VNS implantation surgery was once initially done primarily by neurosurgeons in inpatients under general anesthesia, the implantation procedure has been undertaken by other surgical subspecialties, sometimes in outpatients under regional anesthesia. The NCP Programming Wand and Software, together with a personal computer, provide telemetric communication with the pulse generator that enables noninvasive programming, functional assessments (device diagnostics and interrogation), and data retrieval. The NCP System includes mechanical and electric safety features that minimize the possibility of high-frequency stimulation that could damage tissue. Each patient also is given a magnet that, when held over the pulse generator, turns off the stimulation. When the magnet is removed, normal programmed stimulation resumes, which allows patients to control and temporarily eliminate stimulation-related side effects during key behaviors, such as public speaking (voice tremor) or heavy exercising (mild shortness of breath). Vagus Nerve Stimulation in Relation to Other Somatic Interventions

Although advances in neuropsychopharmacology have captured much interest, recent developments in physical or somatic interventions

VAGUS NERVE STIMULATION

763

Figure 3. Stimulation of the vagus nerve using the NCP system (Cyberonics, Houston, TX). A generator approximately the size of a pocket watch (A) is implated into the left chest wall (19).Bipolar electrodes are wrapped around the left vagus (C) higher in the neck through a separate incision and tunneled under the skin to the generator. A physician can control the intensity and rate of generator firing by holding over the chest wall (and generator) a telemetric wand connected to a portable computer (not shown). (Courtesy of Cyberonics, Houston, TX.)

764

GEORGE et a1

(e.g., nonpharmacologic) are causing a resurgence of interest in alternative methods that directly affect brain function. Thus, the development of VNS is occurring at a time when there is also great interest in other ways of altering brain function-electroconvulsive therapy (ECT), transcranial magnetic stimulation, and deep brain stimulation (DBS). For example, the field of ECT continues to advance, with recent demonstrations of the need for dosage t i t r a t i ~ n ~ regional ~; specificity of ECT eff e c t ~55;~ ~and , production of ECT-like seizures using magnetic, rather than electric, currenP7 (called by these authors magnetic seizure therapy). Interest in transcranial magnetic stimulation (without seizure production), which holds promise as a research tool with potential clinical applications, also is considerable.19,*O The most anatomically discrete, and most invasive, method of stimulating deep brain structures is DBS, in which a thin electrode is inserted directly into the brain and different currents are applied at varying depths until the desired effects are achieved. High-frequency (> 80 Hz) electric stimulation of the middle thalamus or subthalamic nucleus has been found effective in patients with Parkinson’s disease.13,36 Although DBS has the advantage over brain surgery (pallidotomy) of being reversible, the implant procedure is associated with significant morbidity and mortality rates. Although this technique has not been used to treat major depression, mood effects of the stimulation have been reported, and clinical trials for depression are underway? A recent open-labeled report of the beneficial effects of internal capsule DBS as therapy for treatmentresistant obsessive-compulsive disorder hints at the promise of this technique,& although much more research is needed.

Summary of Vagus Nerve Stimulation in Patients with Epilepsy Two double-blind studies (labeled E038 and E0528)have been conducted in patients with epilepsy, with a total of 313 treatment-resistant completers. In this difficult-to-treat group, the average decrease in seizure frequency was approximately 25% to 30% compared with baseline. Data from uncontrolled observations suggest that, contrary to a tolerance effect, improvement in seizure control is maintained or may improve over time.23* 58 Little or no tolerance over time to the anticonvulsant actions of VNS seems to exist. In the second controlled study (E05) of VNS in patients with epilepsy, no severe adverse events occurred that were judged by investigators to be probably or definitely related to VNS.28,59 In 3 of 199 patients, infection following surgery led to device removal. Other surgery-related

VAGUS NERVE STIMULATION

765

adverse events, all of which dissipated over time, included left vocalcord paralysis (2 of 199 [1%] patients); lower facial-muscle paresis (2 of 199 [1%] patients); and pain and accumulation of fluid over the pulse generator, mandating aspiration (1of 199 [0.5%] patients). The perioperative adverse events reported by at least 10% of patients were pain (29%), coughing (14%), voice alteration (13%), chest pain (12%), and nausea (10%). The adverse events reported in patients in the treatment group at some time during treatment that were significantly increased from baseline were voice alteration or hoarseness, cough, throat pain, nonspecific pain, dyspnea, paresthesia, dyspepsia, vomiting, and infection. The only adverse events that occurred significantly more often in the treatment group than in the control group were dyspnea and voice alteration. Adverse events were judged to be mild or moderate in 99% of cases. No cognitive, sedative, visual, affective, or coordination side effects were reported. No significant changes occurred in Holter monitoring, in the results of pulmonary function tests, or in subjects’ hematology values or common chemistry ~ a l u e s . 5No ~ subjects died during the E03 or E05 controlled studies (N = 313). Short-term adverse events that were surgery related were rare and usually resolved. Stimulation-related adverse events (i.e., those that occurred only when the vagus nerve was stimulated), were reduced by decreasing the intensity of the electric current. Stimulation-related adverse events can be aborted by the patient placing a hand-held magnet over the generator, which turns the device off until a physician decreases the stimulation intensity. These adverse events rarely lead to VNS therapy discontinuation in epileptic patients!” 59 Although neurologists were initially skeptical about VNS, a reassessment by the American Academy of Neurology has concluded that VNS for the treatment of epilepsy is “effective and safe.”I7 In clinical studies of epilepsy, the efferent peripheral effects of VNS are minimal, without significant gastrointestinal or cardiac side effects.8,58

Effects of Vagus Nerve Stimulation on Cardiac Function Potential cardiac and gastrointestinal effects of VNS need to be carefully examined. Cardiac evaluations have been performed on more than 250 epileptic patients receiving VNS in clinical trials. Holter monitoring results from clinical studies in epilepsy indicated no significant changes from baseline in cardiac function during stimulation.28 Some investigators have suggested that, because of the rotation of the body during embryonic development, the left and right vagus nerves carry different information. This theory implies that the right vagus

766

GEORGE et a1

nerve is closely associated with the cardiac atria, and the left vagus nerve, with cardiac ventricular function, perhaps explaining the lack of cardiac effects of left VNS. Atrial stimulation is much more likely to produce arrhythmia than is ventricular stimulation. Also, the NCP System stimulating electrode is intended to be positioned inferior to the cardiac branch of the vagus nerve; however, this line of thinking is not totally supported by early animal VNS research. Also, one patient was treated on the right side without complication (G. L. Morris, MD, personal communication, 1999).An alternative theory holds that the cardiac functional response is more a function of stimulation parameters than whether the right or left nerve is stimulated. The cardiac effects of VNS have been much less than might have been predicted given the altered parasympathetic tone that should result from VNS; however, one important exception exists. During the implantation procedure, six cases of 10 to 20 seconds of asystole have been reported in epileptic patients in postmarketing use. No cases were reported in the original epilepsy or depression clinical t r i a l ~ All . ~ ~six asystole events were encountered during the diagnostics test (i.e., lead test) stimulation-the first stimulation that a patient receives after implantation while in the surgical suite. The lead test entails approximately 15 seconds of stimulation at 1.0 mA, 500 microseconds, and 20 Hz of VNS. Of these six patients, three went on to implantation (in three others, the generator and leads were removed). No long-term sequelae have been reported from the asystolic events in these patients. Most importantly, no cardiac events have been reported when the device is turned on for the first time outside of the surgical suite.*,64

Longer-Term Anticonvulsant Efficacy and Tolerability In patients with epilepsy, the long-term efficacy of VNS is maintained or improved,42whereas the frequency of adverse events generally decreases as patients accommodate to s t i m ~ l a t i o nThe . ~ ~patient with the longest exposure to VNS has had the system operating for 10 years. The experience in epileptic patients provides important safety data on VNS used continuously for one year or more. Side effects tend to decrease over time as patients accommodate to the effects of stimulation. In one sample, two years after initial implantation, more than 80% of epilepsy study patients (142 of 172 patients) continued with VNS, suggesting acceptable tolerability over the long term.I2

Cost of Vagus Nerve Stimulation In the United States, VNS delivered with the NCP System costs roughly $9200 for the generator and electrode. The surgical and hospital

VAGUS NERVE STIMULATION

767

costs for the implantation surgery vary widely and are more difficult to quantify. They may become less expensive with the recent trend toward outpatient surgery and local anesthesia. Typically, the total charges for the NCP System plus implantation vary from $12,000 to $25,000. Most federal and private insurance companies reimburse this procedure for patients with treatment-resistant epilepsy. Following implantation, the cost of the device is minimal because the battery lasts several years, and no maintenance is required. VNS settings are adjusted by treating physicians in their offices using the personal computer and attached programming wand. Unfortunately, only a few patients with epilepsy achieve full seizure remission with VNS or are able to reduce substantially or withdraw antiepileptic medications. Most regimens combine VNS with medications. Thus, VNS, as now delivered, has not been shown to be a substitute for anticonvulsant medications. As the mode of action of VNS is better understood, it is hoped that refinements of VNS settings may improve the clinical outcomes, with better cost savings on concomitant medications. Rationale for Studying Vagus Nerve Stimulation in Mood Disorders In addition to the neuroanatomic considerations, several additional lines of evidence provided the background for initiating study of whether VNS has efficacy as therapy for treatment-resistant depression. These hints came from (1) mood effects of VNS observed in patients with epilepsy, (2) evidence from positron emission tomography imaging that VNS affects the function of important limbic structures, (3) the role of anticonvulsant medications in mood disorders, and (4) neurochemical studies in animals and humans that revealed that VNS alters concentrations of monoamines within the CNS. Mood Effects in Epileptic Patients Initial uncontrolled clinical observations and a prospectivez9and a retrospective (Elger et al, unpublished observation, 1998) study suggest that VNS improves mood in patients with epilepsy. These mood changes could not be entirely accounted for by reduced seizure activity. During the clinical trials of VNS for epilepsy, several investigators found mood improvement.8,28 Although decreased seizure frequency likely accounted for some of these changes, the clinical impression was that they went beyond khat attributable to improved seizure control alone. For example, some ptients with minimal or no improvement in seizure frequency

768

GEORGE et a1

also reported substantial improvements in mood. Improved quality of life also was seen, without seizure frequency changes in some patients. Thus, although no specific measures of depressive symptoms were obtained during the epilepsy trials, these reports suggested that VNS might be associated with significant mood improvement. This is reminiscent of the "psychotropic" effects reported in epileptic patients treated with carbamazepine, which led to the clinical trials in mood-disordered patients? In one prospective study of patients with epilepsy (N = 34), a trend toward mood improvements was seen in the 14 subjects who received VNS, based on the Cornell Dysthymia Rating Scale (P < 0.1).29 The significant improvement in seizure frequency found in the VNS group was not related to mood change. Position, Emission Tomography Studies of Limbic Activation

In 10 patients with epilepsy who received VNS, positron emission tomographic measurements were taken three times before and then during VNS.30The results demonstrated increased brain blood flow from rest to during VNS in the rostra1 medulla, thalamus, hypothalamus, insula, and postcentral gyms, with greater activation on the right (contralateral to the device) than left side. In contrast, blood flow was significantly decreased during stimulation bilaterally in the hippocampus, amygdala, and cingulate gyms. The cingulate gyms has been associated with imaging studies of depression pathophysiology, and a decrease in cingulate activity with antidepressant response has been seen in several studies (sleep deprivation,l5. 70 SSRI treatment,41 TMS,19 and symptom provocation with drugsg), although exceptions exist. VNS changes in activity of the brain stem, limbic system, and other CNS areas may be compatible with antidepressant activity. Anticonvulsants as Mood Stabilizers

Another reason for considering VNS in the treatment of mood disorders rests on the substantial evidence that some anticon vulsant medications have mood-stabilizing and possibly antidepressant effects.25, Several anticonvulsants have found a role in mood stabilization (carbamazepine" 45 or valproic acidlo,") or as an antidepressant in treating bipolar disorder (depressed phase; lamotrogine)*6).Also the most effective antidepressant treatment, ECT, has powerful anticonvulsant eff e c t ~ .55~Thus, ~ , it was logical to hypothesize that an effective antiepileptic device also might have antidepressant or mood-stabilizing effects.22

VAGUS NERVE STIMULATION

769

Neurochemical Changes

The anticonvulsant mechanisms of action of VNS are unknown; however, clinical and animal studies suggest that VNS results in changes in the transmission of serotonin (5-HT) norepinephrine (NE),35GABA, and glutamate,68neurotransmitters that are associated with the pathophysiology of major depression. VNS in animals activates the LC, the main source of CNS NE-containing neuronal cell bodies.43In patients with epilepsy VNS seems to increase cerebrospinal fluid 5-hydroxyindoleacetic acid-the major metabolite of 5-HT.7 Because many of the available therapies are believed to work using the same neurotransmitters (5-HT or NE), it was reasonable to hypothesize that VNS also might have antidepressant activity. Also autonomic nervous system dysfunction is well established in major depression (e.g., the link between major depression and cardiovascular illness; see reference 24), possibly implicating the abnormal vagus tone. Thus, if depressed patients have abnormalities in brain regions that control the vagus nerve (i.e., "top-down" regulation), then VNS theoretically might normalize activity in this dysfunctional circuit (i.e., a "bottom-up" approach).

REVIEW OF RECENT OPEN STUDY

Several lines of evidence pointed toward the possible benefit of VNS as an antidepressant or mood-stabilizing therapy. This evidence included clinical observations in epilepsy patients anatomic afferent connections of the left vagus nerve to CNS structures relevant to mood regulation, the anticonvulsant activity of VNS and the role of anticonvulsant medications or ECT in treating mood disorders, neurochemical studies indicating VNS effects on key neurotransmitters involved in mood regulation, and evidence that VNS changes the metabolic activity of key limbic-system structures. This evidence led to an initial openlabelled trial5I conducted at 4 sites (UTSW in Dallas, MUSC in Charleston, Columbia PI in NY, and Baylor in Houston). The study design involved selecting treatment-resistant patients with major depression and then adding VNS to a stable baseline of antidepressant medications. No stimulation was given for the first two weeks following implantation, creating a single-blind placebo phase and allowing for surgical recovery. All patients met eligibility criteria by failing at least 2 adequate treatment trials in the current major depressive episode (MDE) according to the Antidepressant Treatment History F0rm.4~

770

GEORGE et al

Subjects

A total of 38 patients were enrolled, of whom 30 underwent implantation (14, Dallas; 7, Charleston; 6, Houston; 3, New York). Data are available for these 30 patientsI5lall of whom completed the acute study and the available postacute phase outcomes (6-month f o l l o w - ~ p )The .~~ clinical and demographic features of this treatment-resistant sample are summarized in Tables 1 and 2. The patient population was 66% female. Most had nonpsychotic, nonatypical major depressive disorder (MDD) (70%),of whom nearly 50% had recurrent MDD. The other 30% consisted of subjects with bipolar disorder, depressed phase, nonrapid cycling, nonpsychotic, and nonatypical. The mean duration of the current MDE was 10.3 years (SD, 12.5) (range, 0.349.5). The median duration of current MDE was 4.3 years. Nearly two thirds (63%) of patients (n = 30) had been in the current MDE for 2 years or more. This group was severely ill, and over their lifetimes, patients averaged 18.4 (SD, 7.3; range, 6-39) antidepressant and mood-disorder therapies, of which 9.6 (SD, 3.5; range, 3-17) were antidepressant medication trials. As a group, the patients had prolonged, severe, and disabling MDEs and disabling current MDEs. In terms of treatment resistance, 30% had failed 2 treatments; 7% had failed 3; 20% had failed 4; and 43% had failed five or more well-documented treatments that met Antidepressant Treatment History Form criteria during the current MDE.

Concomitant Treatments During the Acute Study

During the acute study, patients received an average of 3.5 (SD, 1.7; median, 4; range, 0-8) other mood-disorder therapies during VNS therapy.

Table 1. CLINICAL AND DEMOGRAPHIC FEATURES OF DEPRESSED SUBJECTS IN THE INITIAL CLINICAL TRIAL (N = 30) Variable

%

Female MDD, recurrent* MDD, single episode' Bipolar I* Bipolar 11* In current MDE > 2 vearst

67 50

20 13 17 70

*According to DSM-IV (APA, 1998). tFor this study, recurrent required four or more MDEs in a lifetime. MDEs = major depressive episodes; MDD = major depressive disorder.

VAGUS NERVE STIMULATION

771

Table 2. DEMOGRAPHICS OF SUBJECTS IN THE STUDY BY RUSH ET AL5’ ( N = 30) Variable

Meank SD (y)

Median (y)

Range (Y)

Age Duration of current MDE Age at onset of current MDE Total duration of illness Age at onset of illness

47.5 ? 7.5 10.3f12.5 37.2? 12.4 19.3f 13.1 28.2 f12.5

49.7 4.7 40.4 19.6 27.9

28.M3.1 0.3-49.5 8.Cb57.6 0.3-49.5 8.0-53.5

MDE

=

major depressive episode.

Vagus Nerve Stimulation Therapy

All 30 patients who underwent implantation had the stimulation parameters set at 500-ks (0.5 ms) pulse width and 20-Hz (n = 25) or 30Hz (n = 5) frequency for 30 seconds and 5 minutes except for 1 patient who received 250-ks pulse width and 3 patients who received stimulation for 30 seconds, and 3 minutes of rest. Output current ranged from 0.25 mA to 3.0 mA, depending on patient tolerance (median, 0.75 mA). After the stimulation parameters were set at the end of a 2week stimulation-adjustment period, no patient required stimulation parameter adjustments during the acute study. Symptomatic Outcomes

Figure 4 presents the Hamilton Rating Scale for Depression (HRSD,,) total score at the exit visit for each of the 30 patients and the percentage of reduction on average (2 visits) baseline HRSD,, for each patient. Overall, a 40% response rate accurred using 50% or more reduction in baseline HRSD,, total score to define response (12 of 30 responders). At exit, according to the Clinical Global Impression-I scale, 3% of patients were minimally worse, 27% were unchanged, 30% were minimally improved, 20% were much improved, and 20% were very much improved at acute phase study exit. When remission is defined as exit HRSD,, score of 10 or loss, 17% of patients remitted at the end of the acute phase. Speed of Onset of Response

No one had significant depression score changes or responded (> 50% depression improvement from baseline) during the first 2 weeks when the device was off. The time to respond appeared gradual and

772

GEORGE et a1

h

*

*

E

U

c i -

-25

li

~__~_.__._.___.__.._I._..............

...... . ................

.

ly.-

.....__._.....I._.....

I

-50

Individual Patients

Figure 4. Percentage of change from baseline in HRSD,, for all 30 subjects in the acute pilot study. Note that bipolar and unipolar responders were found. *bipolar 1; "bipolar 2; solid bar = bipolar; open bar = unipolar. (From Rush AJ, George MS, Sackeim HA, et al: Vagus nerve stimulation (VNS) for treatment-resistant depressions: A multicenter study. Biol Psychiatry 47:277-286, 2000; with permission.)

analyzed separately, for responders (n = 12), early effects (i.e., during VNS dose adjustment) are suggested; however, more than half of the total reduction from 39.1 (average at baseline) to 11.5 (average at exit) for responders occurred over 6 of the 8 weeks of fixed-dose VNS. As of this writing, improvements obtained acutely have largely been sustained among the responders seen in longer-term follow-up (n = 10 of 12 responders). Two responders have lost response. On the other hand, seven of the 18 nonresponders have responded during the long-term follow-up (although sometimes in the context of medication changes). After 6 months of treatment, 17 of 30 (57%) of the treatment-resistant patients met criteria for response (50% reduction in HRSD,, scores). Response Predictors?

Because VNS involves a surgical implant and is relatively expensive, it is desirable to predict before surgery who likely will respond. The only potentially significant factor concerned ECT response; only 1 of 7 patients who did not respond at all (either partially or transiently) to ECT responded to VNS. The odds ratio was highest when patients who had completely failed to respond acutely to ECT were compared with

VAGUS NERVE STIMULATION

773

all other patients (i.e., those who never had ECT combined with those who had a transient or a partial ECT response); however, this relationship did not reach statistical significance ( P = 0.14). Also, patients who received more exhaustive and aggressive pharmacotherapy during the index episode seemed less likely to respond. The few patients studied at Columbia had especially robust pharmacologic and ECT treatment, and none responded. In this small study, subjects who were stimulated with a weaker current tended to have better responses. This paradoxic trend needs further investigation and may be causal or a consequence of more ill subjects receiving stronger currents. Adverse Events and Tolerability

All but one subject has elected to keep the device implanted, suggesting good tolerability. Interestingly, the very first patient implanted (BPAD type 1 depressed) developed a dysphoric hypomania, which subsided with a reduction in the stimulation output current. Two others cases of hypomania also have occurred in the ongoing follow-up extension of this study to include an additional 30 patients (M. George, MD, H. A. Sackeim, PhD, personal communication, 2000). Four severe adverse events occcrred: (1)infection; (2) leg pain; (3) agitation or panic; and (4) agitation, irritability, and dysphoria. Three patients (10%) reported abnormal wound healing, which involved a longer time for implant incisions to heal. All patients’ wounds healed without significant intervention. These events all occurred at one site; the surgeon has since modified his incision-closure technique. One additional patient (3%) had a wound infection that mandated hospitalization and intravenous antibiotic therapy. In the extension study, a few patients have required hospitalization because of worsening of depressive symptomatology (A. J. Rush, MD, H. A. Sackeim, PhD, personal communication, 2000). Some adverse events also have developed during the follow-up phase of the study. A 40-year-old woman who was a heavy smoker developed shortness of breath after the acute phase. She was diagnosed with asthma and treated with albuterol. Potentially relevant to this observation, a worsened forced expiratory volume at 1 second was reported following VNS in epileptic patients with a history of preexisting pulmonary i r n ~ a i r m e n t . ~ ~ Limitations and Discussion

Given the small sample size and open-labeled design, these findings in a treatment-resistant group are preliminary. Obviously, without a

774

GEORGE et a1

randomized, sham control group, one cannot draw definite conclusions about effectiveness in this patient population. A multisite, randomized study was begun this year; however, the severe, chronic, and treatmentresistant nature of the depressive episodes in this patient sample suggest that only 5% to 10% of these patients would have been expected to improve spontaneously during the 3 months after implantati~n.~~, 65 The response rate that was found (i.e., 40% acute, 57% at 6 mo) after implantation substantially exceeds these expectations. Most responders to date have sustained their acute improvements in the longer-term follow-up, suggesting that VNS provides ongoing benefit for those who respond. Relapse rates of 30% to 60% commonly have been reported in patients with far less severe, nonresistant major depression while in continuation or maintenance medication treatment studies. Following ECT, relapse rates are especially high among patients who are medication resistant.60 Several additional points argue against placebo effects, accounting for much of these results. First, no patient responded in the 2-week, postimplantation, single-blind, recovery period. Also, follow-up data suggest that patients who initially improved retained that improvement after acute study exit. A pattern of sustained benefit is unlikely to reflect placebo response.5Q 65 These pilot results need replication in a masked study before acceptance, but if a masked study shows efficacy greater than sham in a treatment-resistant group, with effect sizes this large that persist, then a host of important clinical and research issues will surface immediately. Future studies should likely give consideration to the following questions. 1. Can nonresponders to VNS during the acute study become responders after long-term treatment with changes in stimulation parameters? This area is being addressed partially by the multisite group involved in the initial study. In the open study, antidepressant response paradoxically occurred more commonly in those with lower intensity settings, although this, finding was insignificant. In ongoing treatment, acute-phase nonresponders are having their VNS intensity systematically decreased as a first step. Then, if still no response occurs, the pulse width is halved. Systematic work like this is needed for better understanding of the effect of stimulus parameters. 2. Does VNS continue to provide ongoing, sustained symptom relief over months or years after acute phase response? The preliminary follow-up data suggest that this is true in depression, much as it is in epilepsy. Current thinking in chaos theory may begin to explain this lack of development of tolerance to VNS in patients who have become tolerant of medications. The intermit-

VAGUS NERVE STIMULATION

775

tent nature of VNS (i.e., 30 s on, 5 min off) is more resistant to the development of tolerance than is constant VNS (i.e., always on). Similarly, some investigators argue that, with cardiac pacemakers, the pattern most resistant to developing tolerance is a constantly irregular pattern, in which stimulation length and the rest length constantly vary within limits (e.g., 20-30 s on, 4-5 min off). These theories are easily testable in animal models of epilepsy and also could be tested with interleaved functional imaging. Improved understanding of this issue of tolerance and how it does or does not develop as a function of the intermittance of stimulation would greatly advance many areas of therapeutics in neuropsychiatry. 3. Do "predictors" or "correlates" of response or of time to response exist? In the epilepsy literature, the degree of blood flow change in the thalamus when VNS is first started seems to correlate with later clinical response.31A similar study is underway, with depressed subjects receiving VNS, to determine whether changes in regional brain activity during VNS correlate with later clinical antidepressant response. 4. Can medications be reduced or eliminated after a stable, sustained response to VNS is obtained? That VNS does not allow most epileptic patients to discontinue anticonvulsant medications is disappointing, but, with improved understanding of neuronal effects, VNS might be made more effective with concurrent reduction in medications. These issues are untested in patients with depression. Patients in the pilot study were required not to alter their medication regimens to obtain a realistic view of the effects of VNS. Similarly, an entire pharmacology of adjunctive VNS may be developed. Assuming that the suggested antidepressant effects of VNS are mediated predominantly through NE-containing neurons in the LC, should one use augmenting medications that are within this class (e.g., NE reuptake blockers, such as tricyclic antidepresents or reboxetine) or in a different class (e.g., selective 5-HT reputable inhibitors? Animal studies have shown that altering the level of GABA or glutamate in the NTS markedly changes VNS anticonvulsant effects on seizures. Similar research at a clinical level is needed in terms of supplementation with GABAergic or glutamatergic mood stabilizers. 5. Where does VNS fit into treatment algorithms for managing patients with MDD or bipolar disorder? The implantation of a device is unlikely to be a first-line treatment for depression, but more research is needed to determine where in a treatment algorithm one might resort to VNS. This issue is most important with respect to the relationship to ECT Altogether, 19 of 30

776

GEORGE et a1

patients (63%) had received ECT in their lifetimes. Seventeen (57%)had received ECT during the current MDE, and ten (33%) had received ECT within 2 years of study entry. Based on all 19 patients who had ever received ECT, 7 had no response to ECT, 3 had partial responses, 8 had transient responses, and 1 had a sustained response. Only one of the (7) patients who had a prior nonresponse to ECT was a responder to VNS. Thus, preliminarily, clear nonresponse to ECT seems to be a relative negative response predictor for VNS. If the open results hold up in the planned double-blind study and VNS is found effective for treating depression, then determining where VNS would fit into the treatment of depression, and whether VNS should be administered before, after, or in conjunction with ECT, are important questions for research. Similarly, given the small sample treated to date, how the degree and extent of treatment resistance predicts response to VNS is unclear. Whether established resistance to any class of antidepressants (e.g., tricyclic antidepresants, monoamine oxidane inhibitors, and selective 5-HT reuptable inhibitors) has predictive value is yet to be examined. For example, evidence shows that tricyclic-antidepressant resistance predicts poorer response to ECT, whereas selective 5-HT reuptake inhibitor resistance has no predictive value.49Clearly, whether patients who respond to VNS overlap substantially with patients who may respond to a particular class or a specific antidepressant medication is important to determine. 6. How useful is VNS in bipolar disorder? Several anticonvulsants are effective for the treatment of bipolar disorder, including valproate, carbamazepine, ECT, and possibly some of the never anticonvulsants. Although hypomania was found in the pilot and following study, the small sample size and the lack of a randomized control group make determination of cause difficult, particularly because most bipolar patients also were receiving antidepressant medications. Given the limited options available for patients with rapid-cycling bipolar disorder, more research in this area seems warranted. OTHER AREAS OF POTENTIAL CLINICAL PROMISE WITH VAGUS NERVE STIMULATION Anxiety Disorders

VNS is an effective anticonvulsant. Evidence51 suggests that VNS also may have antidepressant properties. The vagus nerve is an im-

VAGUS NERVE STIMULATION

777

portant route of information into the CNS. The sensory afferent nerve fibers that VNS directly stimulates terminate in the NTS. These fibers are the primary means by which the brain receives information from the viscera. From there, information travels to the LC, the primary site of all NE fibers in the brain. NE has long been considered one of the primary neurotransmitters involved in the pathogenesis and regulation of anxiety. A device that directly stimulates this NE control site may have therapeutic effects on anxiety. The historical importance of this pathway that VNS modulates can be seen in the oldest theory about the brain origin of fear, called the James-Lang theory of emotions. James33argued that the experience of emotion resulted from visceral changes in the body, and it was the brain’s interpretation of these signals through the vagus nerve that caused anxiety and other affective states. He argued that rapid heartbeat and shortness of breath are not a consequence but a cause of anxiety. According to this theory, one thinks he or she is anxious because the heart beats fast, and then the brain gets this signal (through the vagus), and one experiences anxiety. This theory, although often challenged, has been difficult to disprove, and most modern anxiety researchers argue that the ultimate answer lies in central-peripheral loops; however, most investigators agree that information flowing through the vagus to higher brain regions is an important part of anxiety regulation. A device that directly affects this information flow may be a useful way of treating anxiety. Several theories of the anxiety disorders posit a faulty interpretation or erratic availability of this peripheral information into the CNS.26* 3 3 , 6 9 Thus, altering the flow of this information with VNS may have therapeutic potential in anxiety disorders (e.g., general anxiety disorder or panic disorder). In the initial pilot study in depression, VNS caused improvements in Hamilton anxiety scores that were as large as the changes in core depression ratings. Obesity, Sleep, and Pain Control

Similarly, the vagus nerve contains information about hunger, satiety, and pain. Potential studies in the areas of treatment-resistant obesity, addictions, or pain syndromes theoretically also are justified. Moreover, the NTS sends fibers into the dorsal raphe and areas that control levels of alertness. Thus, VNS might be considered as a potential therapy for some disorders of sleep or alertness, such as coma or narcolepsy. For example, a study of 10 epileptic patients found that high-intensity, highfrequency VNS reduced total time in rapid eye movement sleep, and rapid eye movement sleep was less fragmented,67similar to changes seen with antidepressant medications. Therefore, in addition to advancing the

778

GEORGE et a1

understanding about the pathophysiology of various neuropsychiatric disorders, VNS may have other therapeutic applications, which may be informed by the anatomy of vagus pathways. A fascinating study of bilateral VNS using bilateral supradiaphragmatic stimulation in dogs found that VNS caused marked weight loss over time, likely not caused by changes in metabolism but rather changes in vagus-mediated information about when the stomach was full or when satiety occurs (M. Roslin, MD, et al, NAASO, Charleston, SC, personal communication, October 1999). Additional Refinements in the Use of Vagus Nerve Stimulation

The delivery of VNS involves a surgical procedure that includes exposure and manipulation of the carotid artery and the cosmetic, inconvenience, and tolerability issue with implanting a generator in the chest wall. As a consequence, the NCP System typically is used in patients who have not responded to other therapies. Some patients with epilepsy or depression receive little to no benefit despite having had surgery. The development of less invasive ways of delivering VNS and predicting which patients will benefit are needed. TMS has been used in a few studies to stimulate the vagus and other cranial nerves," 40, but a reliable indicator of effective VNS is difficult to identify (H. Sackeim, PhD, S. Lisanby, MD, M. George, MD, unpublished observations) other than brain stem-mediated responses in the swallowing pathways, as described by Hamdy et aLZ7Also, transcutaneous magnetic stimulation of the vagus nerve would be impractical for long-term, 24 d treatment. Another possibility might be to develop a temporary percutaneous method of stimulation, but a different approach would be to identify the subset of patients who are most likely to benefit from VNS by using functional neuroimaging, clinical history of previous response to antidepressants, or other measures. 277

OTHER AREAS OF POTENTIAL RESEARCH PROMISE AND THEORETICAL ISSUES Effects on Memory

Theoretically, VNS might have effects on cognition in three time domains. First, testing while the stimulator is immediately on (for 5-10 s) may show differences from rest caused by activation of the LC and related cortical and limbic regions. Also, one could test for the effects of a cycle of stimulation over several hours, which would encompass

VAGUS NERVE STIMULATION

779

stimulation both on and off. Finally, as in the clinical trials, one can assess the effects of VNS administered constantly over several weeks or months. In terms of effects over time in clinical patients, comprehensive neuropsychological testing was performed in the initial open depression study (N = 27 who completed neuropsychological testing before and after stimulation over 8 weeks (H.A. Sackeim, PhD, unpublished observations, 2000). No evidence of deterioration in any neurocognitive measure was found. Relative to baseline, improvement in motor speed (finger tapping), psychomotor function (digit symbol test), language (verbal fluency), and executive functions (logical reasoning, working memory, response inhibition, or impulsiveness) occurred. For some measures, improved neurocognitive performance correlated with the extent of reduction in depressive symptoms. VNS output current was unrelated to changes in cognition. Thus, over the timeframe of the depression and epilepsy clinical experience, deleterious effects on cognition generally have not been found. A study by Clark et all1 hints at the potential of VNS to investigate the neurobiology of learning and memory. They examined word-recognition memory in 10 patients enrolled in a clinical study of VNS for treating epilepsy. VNS administered after learning and during memory consolidation caused intensity-dependent enhancement of word recognition relative to sham stimulation. Other work has shown that vagotomy attenuates the memory-enhancing properties of amphetamine. Investigation is needed in animals and then in humans to understand the immediate, acute, and chronic effects of VNS on cognition as a function of the VNS parameters (i.e., intensity, pulse, width, frequency, duty cycle, and timing of VNS relative to task). A Need for Understanding Neurobiology VNS can be delivered at different amplitudes and frequencies and with different pulse widths, all of which can be delivered at various duty cycles (ratio of on to off time). More basic studies would advance this area through understanding how varying combinations of these parameters affect different brain regions or influence different neuropsychiatric conditions. In the memory study by Clark et all1 and in animal studies, evidence suggests that VNS delivered with different parameters from those commonly used for epilepsy may produce different CNS effects that may, in turn, broaden the clinical indications. SUMMARY VNS builds on a long history of investigating the relationship of autonomic signals to limbic and cortical function and is one of the

780

GEORGE et a1

newest methods to physically alter brain function. VNS is a clinically useful anticonvulsant therapy in treatment resistant patients with epilepsy, and pilot data suggest that it has potential as an antidepressant therapy. The known anatomic projections of the vagus nerve suggest that VNS also might have other neuropsychiatric applications. Additional research is needed to clarify the mechanisms of action of VNS and the potential clinical utility of this intriguing new somatic portal into the CNS. ACKNOWLEDGMENTS Dr. George thanks Dr. Paul MacLean for helpful discussions about the relationship of the autonomic nervous system and the limbic system, with particular attention to the role of the vagus nerve and the nucleus solitary tract.

References 1. Amar AP, Heck CN, Levy ML, et al: An institutional experience with cervical vagus nerve trunk stimulation for medically refractory epilepsy: Rationale, technique and outcome. Neurosurgery 43:1265-1280, 1998 2. Asconape JJ, Moore DD, Zipes DP, et al: Bradycardia and asystole with the use of vagus nerve stimulation for the treatment of epilepsy: A rare complication of intraoperative device testing. Epilepsia 40:1452-1454, 1999 3. Bailey P, Bremer F: A sensory cortical representation of the vagus nerve. J Neurophysiol 405-412, 1938 4. Ballenger JC, Post RM: Carbamazepine (Tegretol) in manic-depressive illness: A new treatment. Am J Psychiatry 137782-790, 1980 5. Bejjani BP, Damier P, Arnulf I, et al: Transient acute depression induced by highfrequency deep brain stimulation. N Engl J Med 3401476-1480, 1999 6. Benecke R, Meyer BU, Schonle F, et al: Transcranial magnetic stimulation of the human brain: Responses in muscles supplied by cranial nerves. Exp Brain Res 7k623-632, 1988 7. Ben-Menachem E, Hamberger A, Hedner T, et al: Effects of vagus nerve stimulation on amino acids and other metabolites in the CSF of patients with partial seizures. Epilepsy Res 20:221-227, 1995 8. Ben-Menacham E, Manon-Espaillat R, Ristanovic R, et al: Vagus nerve stimulation for treatment of partial seizures: I. A controlled study of effect on seizures. Epilepsia 35:616626, 1994 9. Bremner JD, Innis RB, Salomon RM, et a1 PET measurement of cerebral metabolic correlates of depressive relapse. Arch Gen Psychiatry 54:364-374, 1997 10. Calabrese JR, Delucchi GA: Phenomenology of rapid cycling manic depression and its treatment with valporate. J Clin Psychiatry 50:3&34, 1989 11. Clark KB, Naritoku DK, Smith DC, et al: Enhanced recognition memory following vagus nerve stimulation in human subjects. Nat Neurosci 2:94-98, 1999 12. Cyberonics: Physician’s Manual for the NCP Pulse Generator. Houston, TX, 1998 13. Damier P: The stimulation of deep cerebral structures in the treatment of Parkinson’s disease [abstract]. European Neuropsychopharmacology 8(suppl):89-13, 1998 14. Dell P, Olson R Projections ’secondaires’ mesencephaliques, diencephaliques et amygdaliennes des afferences viscerales vagales. C R SOCBiol 145:108&1091, 1951 15. Ebert D, Feistel H, Barocka A, et al: Increased limbic flow and total sleep deprivation in major depression with melancholia. Psychiatry Res 55:lOl-109, 1994 16. Fatemi SH, Rapport DJ, Calabrese JR, et al: Lamotrigine in rapid-cycling bipolar disorder. J Clin Psychiatry 58:522-527, 1997

VAGUS NERVE STIMULATION

781

17. Fisher RS, Handforth A: Reassessment: Vagus nerve stimulation for epilepsy: A report of the Therapeutics and Technology Assessment Subcommittee for the American Academy of Neurology. Neurology 53:66&669, 1999 18. Foley JO, DuBois F: Quantitative studies of the vagus nerve in the cat The ratio of sensory and motor studies. J Comp Neurol 6749-67, 1937 21. George MS, Post Rh4, Ketter TA, et al: Neural mechanisms of mood disorders. Current Review of Mood and Anxiety Disorders 1:71-83, 1997 20. George MS, Lisanby SH, Sackeim H A Transcranial magnetic stimulation: Applications in neuropsychiatry. Arch Gen Psychiatry 56300-311, 1999 19. George MS, Belmaker RH: Transcranial Magnetic Stimulation in Neuropsychiatry. Washington, DC, American Psychiatric Press, 2000 22. George MS, Sackeim HA, Rush AJ, et al: Vagus nerve stimulation: A new tool for brain research and therapy. Biol Psychiatry 47287-295, 2000 23. George R, Salinsky M, Kuzniecky R, et al: Vagus nerve stimulation for treatment of partial seizures: 11. Long-term follow-up on first 67 patients exiting a controlled study. Epilepsia 35:637-643, 1994 24. Glassman A H Depression, cardiac death, and the central nervous system. Neuropsychobiology 3780-83, 1998 25. Goodwin FK, Jamison KR: Medical treatment of acute bipolar depression. In ManicDepressive Illness. Oxford, Oxford University Press, 1990, pp. 630-664 26. Gray JA: The Neuropsychology of Anxiety: An Enquiry into the Functions of the Septo-Hippocampal System. New York; Oxford University Press, 1982 27. Hamdy S, Rothwell J, Aziz Q et al: Long-term reorganization of human motor cortex driven by short-term sensory stimulation. Nat Neurosci 1364-68, 1998 28. Handforth A, DeGiorgio CM, Schachter SC, et al: Vagus nerve stimulation therapy for partial-onset seizures: A randomized active-control trial. Neurology 51:4&55, 1998 29. Harden CL, Pulver MC, Nikolov B, et al: Effect of vagus nerve stimulation on mood in adult epilepsy patients abstract. Neurology 52(suppl 2):238, 1999 30. Henry TR, Bakay RAE, Votaw JR, et al: Brain blood flow alterations induced by therapeutic vagus nerve stimulation in partial epilepsy: Acute effects at high and low levels of stimulation. Epilepsia 39:983-990, 1998 31. Henry TR, Votaw JR, Pennell PB, et al: Acute blood flow changes and efficacy of vagus nerve stimulation in partial epilepsy. Neurology 52:116&1173, 1999 32. Hu TW,Rush AJ: Depressive disorders: Treatment patterns and costs of treatment in the private sector of the United States. SOCPsychiatry Psychiatr Epidemiol 30:224.230, 1995 33. James W What is an emotion? Mind 9:188-205, 1884 34. Ketter TA, George MS, Kimbrell TA, et al: Functional brain imaging in mood and anxiety disorders. Current Review of Mood and Anxiety Disorders 1:9&112, 1997 35. Krahl SE, Clark KB, Smith DC, et al: Locus coeruleus lesions suppress the seizure attenuating effects of vagus nerve stimulation. Epilepsia 39:709-714, 1998 36. Limousin P, Krack P, Pollak P, et al: Electrical stimulation of the subthalamic nucleus in advanced Parkinson’s disease. N Engl D Med 339:11051111, 1998 37. Lisanby SH, Luber B, Finck AD, et al: Deliberate seizure induction with repetitive transcranial magnetic stimulation in nonhuman primates. Arch Gen Psychiatry, in press 38. Lotvall J, Lunde H, Augustinson LE, et al: Airway effects of direct left-sided cervical vagal stimulation in patients with complex partial seizures. Epilepsy Res 18:149-154, 1994 39. MacLean PD: The Triune Brain in Evolution: Role in Paleocerebral Functions. New York, Plenum Press, 1990 40. Mador MJ, Rodis A, Magalang UJ, et al: Comparison of cervical magnetic and transcutaneous phrenic nerve stimulation before and after threshold loading. Am J Respir Crit Care Med 154:448-453, 1996 41. Mayberg HS, Brannan SK, Mahurin RK, et al: Cingulate function in depression: A potential predictor of treatment response. NeuroReport 8:1057-1061, 1997 42. Morris GL, Mueller WM, and the Vagus Nerve Stimulation Study Group (E01-E05): Long-term treatment with vagus nerve stimulation in patients with refractory epilepsy. Neurology 53:1731-1735, 1999

782

GEORGE et a1

43. Naritoku DK, Terry WJ, Helfert RH: Regional induction of Fos immunoreactivity in the brain by anticonvulsant stimulation of the vagus nerve. Epilepsy Res 22:53-62,1995 44. Nuttin 8, Cosyns P, Demeulemeester H, et al: Electrical stimulation on anterior limbs of internal capsule in patients with obsessive-compulsive disorder. Lancet 354:1526, 1999 45. Okuma T, Kishimoto A, Inoue K, et al: Anti-manic and prophylactic effects of carbamazepine (Tegretol) on manic depressive psychosis. Folia Psychiatrica et Neurologica 27283-297, 1973 46. Olfson M, Marcus S, Sackeim HA, et al: Use of ECT for the inpatient treatment of recurrent major depression. Am J Psychiatry 155:22-29, 1998 47. Penry JK, Dean JC: Prevention of intractable partial seizures by intermittent vagal stimulation in humans: Preliminary results [abstract]. Epilepsy 31 (suppl):40-43, 1990 48. Post RM, Weiss SRB, Chuang DM: Mechanisms of action of anticonvulasants in affective disorders: Comparisons with lithium. J Clin Psychopharmacol, in press 49. Prudic J, Haskett RF, Mulsant B, et al: Resistance to antidepressant medications and short-term clinical response to ECT. Am J Psychiatry 53:985-992, 1996 50. Quitkin FM, McGrath PJ, Rabkin JG, et al: Different types of placebo response in patients receiving antidepressants. Am J Psychiatry 148:197-203, 1991 51. Rush AJ, George MS, Sackeim HA, et al: Vagus nerve stimulation (VNS) for treatmentresistant depressions: A multicenter study. Biol Psychiatry 47277-286, 2000 52. Rutecki P: Anatomical, physiological and theoretical basis for the antiepileptic effect of vagus nerve stimulation. Epilepsia 31 (suppl) 1 4 , 1990 54. Sackeim HA, Decina P, Kanzler M, et al: Effects of electrode placement on the efficacy of titrated, low-dose ECT. Am J Psychiatry 1441449-1455, 1987 55. Sackeim HA, Decina P, Malitz S, et al: Anticonvulsant and antidepressant properties of electroconvulsive therapy: A proposed mechanism of action. Biol Psychiatry 18:1301-1310, 1983 56. Sackeim HA, Decina P, Portnoy S, et al: Studies of dosage, seizure threshold, and seizure duration in ECT. Biol Psychiatry 22249-268, 1987 57. Sackeim HA, Prudic J, Devanand DP, et al: Effects of stimulus intensity and electrode placement on the efficacy and cognitive effects of electroconvulsive therapy. N Engl J Med 32M39-846, 1993 53. Sackeim HA: The anticonvulsant hypothesis of the mechanisms of action of ECT Current status. J ECT 15526, 1999 58. Salinsky MC, Uthman BM, Ristanovic RK, et a1 and the Vagus Verve Stimulation Study Group: Vagus nerve stimulation for the treatment of medically intractable seizures: Results of a 1-year open-extension trial. Arch Neurol 53376-1180, 1996 59. Schachter SC, Saper CB: Vagus nerve stimulation (progress in epilepsy research). Epilepsia 39:677-686, 1998 60. Shapira B, Gorfine M, Lerer B: A prospective study of lithium continuation therapy in depressed patients who have responded to electroconvulsive therapy. Convuls Ther 11:80-85, 1995 61. Shea MT, Elkin I, Imber SD, et al: Course of depressive symptoms over followup: Findings from the National Institute of Mental Health Treatment of Depression Collaborative Research Program. Arch Gen Psychiatry 49:782-787, 1992 62. Sirnilowski T, Catala M, Rancurel G, et al: Impairment of central motor conduction to the diaphragm in stroke. Am J Respir Crit Care Med 154436441, 1996 63. Swarm AC, Bowden CL, Morris D, et al: Depression during mania: Treatment response to lithium or divalproex. Arch Gen Psychiatry 5437-42, 1997 64. Tatum WO IV, Moore DB, Stecker MM, et al: Ventricular asystole during vagus nerve stimulation for epilepsy in humans. Neurology 52:1267-1269, 1999 65. Thase ME, Rush AJ. Treatment resistant depression. In Bloom FE, Kupfer DJ (eds): Psychopharmacology: Fourth Generation of Progress. New York, Raven Press, 1995; pp 1081-1097 66. Van Bockstaele EJ, Peoples J, Valentino RJ: Anatomic basis for differential regulation of the rostrolateral peri-locus coeruleus region by limbie afferents. Biol Psychiatry 46: 1352-1363, 1999 67. Vaugh BV, DCruz OF: Effect of vagal nerve stimulation on sleep [abstract]. Epilepsia 40:137, 1999

VAGUS NERVE STIMULATION

783

68. Walker BR, Easton A, Gale K: Regulation of limbic motor seizures by GABA and glutamate transmission in nucleus tractus solitarius. Epilepsia 40:1051-1057, 1999 69. Watkins LL, Grossman P, Krishnan R, et al: Anxiety and vagal control of heart rate. Psychosom Med 60498-502, 1998 70. Wu JC, Gillin JC, Buchsbaum MS, et al: Effect of sleep deprivation on brain metabolism of depressed patients. Am J Psychiatry 149:538-543, 1992 71. Zabara J: Inhibition of experimental seizures in canines by repetitive vagal stimulation. Epilepsia 33:1005-1012, 1992 72. Zabara J: Peripheral control of hypersynchronous discharge in epilepsy. Electroencephalogr Clin Neurophysiol 6l(suppl): 162, 1985 73. Zabara J: Time course of seizure control to brief, repetitive stimuli [abstract]. Epilepsia 26:518, 1985

Address reprint requests to Mark S. George, MD Department of Radiology Medical University of South Carolina 171 Ashley Avenue Charleston, SC 29425