Neuroimmune Modulation for Drug-Refractory Rheumatoid Arthritis: Long-Term Safety and Efficacy in Patients Enrolled in a Pilot Vagus Nerve Stimulation Study

Abstract

Introduction

Vagus nerve stimulation activates neuroimmune reflexes that modulate systemic inflammation and may represent a novel non-pharmacologic treatment modality for autoimmune diseases like rheumatoid arthritis (RA). In a 3-month first-in-human, double-blind trial, 50% of patients with drug-refractory RA improved clinically, two patients achieved remission, and pro-inflammatory cytokines declined by 30–50% with daily stimulation. The current study is a 36-month extension of that trial, designed to assess the sustained safety and efficacy of the neuroimmune modulation device in patients with multidrug-refractory RA.

Methods

Patients (N = 14) with active RA and prior insufficient response to at least two different biological or targeted synthetic disease-modifying antirheumatic drugs (b/tsDMARDs) with at least two mechanisms of action were implanted with a novel neuroimmune modulation device that stimulates the cervical vagus nerve, treated for 12 weeks, and then assessed for safety and clinical effectiveness in an open-label 36-month extension.

Results

Eleven patients completed the extension study through month 36. Patients had previously failed an average of 4.8 different drugs, with 64% (9/14) having failed a tsDMARD before enrolling in the study. The median change in clinical disease activity index (CDAI) score from day 0 to month 36 was − 17.8 (SEM 4.9). At month 36, 64% (7/11) of patients achieved a CDAI response that met or exceeded the minimal clinically important difference. Two of these seven patients were treated with daily stimulation alone, while five patients combined stimulation with an adjunctive b/tsDMARD. No device-related infections, cardiac events, surgical revisions, or device explants were reported. Two adverse events related to the device occurred in a single patient: a mild sore throat and moderate tenderness near the implant site. These events were non-serious, anticipated, and resolved.

Conclusion

In this first-in-human long-term extension study, neuroimmune modulation was well tolerated among patients with multidrug-refractory RA, with reductions of clinical disease activity that were maintained through 36 months.

Key Summary Points

Why carry out this study?

Electrical stimulation of the vagus nerve to activate innate protective neuroimmune reflexes represents a novel means of treating inflammatory autoimmune diseases such as rheumatoid arthritis (RA).

We previously reported primary clinical outcomes from a first-in-human, double-blind study of a novel neuroimmune modulation system at 3 months. In that study, 50% (5/10) of patients with drug-refractory RA receiving active daily vagus nerve-targeted neuroimmune modulation met or exceeded the minimal clinically important difference in disease activity score 28 based on C-reactive protein (DAS28-CRP); two patients achieved DAS28 remission (DAS28-CRP < 2.6); and pro-inflammatory cytokines were reduced by 30–50%.

The current study is a long-term extension designed to assess the 36-month safety and efficacy of this device.

What was learned from the study?

In this 36-month extension study, 64% of patients, despite prior failure of an average of 4.8 therapies, achieved clinically meaningful improvement in clinical disease activity scores. No serious device-related adverse events occurred and only two mild, self-limiting stimulation-related events were reported in a single patient.

These findings provide further evidence of the long-term efficacy and safety of vagus nerve-targeted neuroimmune modulation in RA and support the rationale for evaluating this approach in a larger randomized controlled trial.

Introduction

Rheumatoid arthritis (RA) is a chronic, progressive, inflammatory autoimmune disease with a complex etiology [1]. Patients commonly report persistent and debilitating symptoms, including pain, fatigue, and loss of function [2]. Estimates place the number of adults with RA at over 1.3 million in the USA, 3 million in Europe, and 1.24 million in Japan [3–5]. RA drug treatment selection is dependent on severity, patient comorbidities, and initial response to therapy. Methotrexate (MTX), a conventional disease-modifying antirheumatic drug (DMARD), is typically first-line therapy [6–8]. Biological DMARDs (bDMARDs), such as tumor necrosis factor (TNF) inhibitors, and targeted synthetic DMARDs (tsDMARDs), such as Janus kinase (JAK) inhibitors, are more potent RA therapies used for patients who do not respond, or who are intolerant, to conventional DMARDs [6–8]. Although these therapies have improved disease management, a significant number of patients with RA do not achieve or sustain remission or low disease activity, which are the primary treat-to-target goals for management of RA [2, 5, 8–13]. Despite targeting critical pathogenic inflammatory pathways, lack or loss of efficacy, comorbidities, adverse effects, and drug intolerance further complicate care, and patients that fail to respond to at least two different drugs with different mechanism of actions (MOAs) are less likely to respond to subsequent treatment [14]. The heterogeneity of RA pathogenesis suggests that additional, unaddressed immunological pathways may be contributing to disease persistence in these cases, underscoring the need for continued innovation in therapeutic development.

The central nervous system regulates innate and adaptive immunity through several pathways, including the vagus nerve-mediated “inflammatory reflex” that senses inflammation and responds reflexively via its efferent arm, the cholinergic anti-inflammatory pathway [15, 16]. Numerous animal models have demonstrated that electrically stimulating the vagus nerve can modulate systemic inflammation, including in rodent models of RA (collagen- and zymosan-induced arthritis) [17–20]. It is therefore plausible that modulating these pathways by targeted stimulation of the vagus nerve might have therapeutic potential for the treatment of RA [21].

Vagus nerve-mediated neuroimmune modulation was first evaluated as a treatment for RA in a single-arm, open-label proof-of-concept study conducted using a reprogrammed, commercially available vagus nerve stimulation device approved for drug-refractory for epilepsy [22]. This study demonstrated significant inhibition of cytokine production and clinically meaningful improvements in disease activity, whereas withdrawal of the treatment resulted in worsening disease activity. A subsequent first-in-human study was conducted in a drug-refractory difficult-to-treat population to examine the safety and tolerability of a miniaturized integrated neurostimulation device designed for the treatment of RA [23]. We previously reported the primary clinical outcomes of this study, demonstrating that the device and treatment were safe and well tolerated. Despite the highly drug-refractory background of the study population, we observed that 50% (5 of 10) of stimulated patients with multidrug-refractory RA met or exceeded the minimal clinically important difference in disease activity score 28 based on C-reactive protein (DAS28-CRP) of − 1.2. Notably, two patients achieved remission (DAS28-CRP < 2.6); and levels of pro-inflammatory cytokines decreased by 30–50% following 12 weeks of neuroimmunomodulation treatment [23]. We now report on the long-term safety and efficacy of neuroimmune modulation from a 36-month open-label extension of this study in patients with multidrug-refractory RA.

Methods

Study Design and Participants

The study was a multisite, first-in-human study to assess the safety and efficacy of a miniaturized integrated neurostimulation device in adult patients with active moderate-to-severe RA (ClinicalTrials.gov Identifier NCT04862117) enrolled at four clinical research sites in the USA. Full study design details have been previously published [23]. Patients were aged 22–75, inclusive, with active, adult-onset, moderate-to-severe RA and an incomplete response or intolerability to at least two b/tsDMARDs having at least two different mechanisms of action (i.e., difficult-to-treat). Active moderate-to-severe RA was defined by presentation at screening with at least 4/28 tender and 4/28 swollen joints and clinical disease activity index (CDAI) score > 10. There was no CRP threshold level required for enrollment.

The primary study was conducted in two stages: stage 1 (N = 3) was open-label, and stage 2 (N = 11) was randomized, sham-controlled, and double-blinded [23]. All patients in stage 1 received active stimulation for 1 min once per day (QD). Following a safety review, stage 2 was initiated and the remaining 11 patients were randomized 1:1:1 into one of two active treatment groups and received either active stimulation for 1 min once per day (QD, N = 3) or for 1 min four times per day (QID, N = 4) or non-active (control) stimulation of 0 μA 1 min QD (N = 4). All patients who completed the week 12 (end of primary study) visit (N = 14) enrolled into an open-label, 36-month, long-term extension where patients continued to receive their assigned treatment: active stimulation for 1 min either QD (N = 6) or QID (N = 4). All control patients were crossed over to active stimulation for 1 min and randomized to either QD (N = 2) or QID (N = 2). Once crossed over to the open-label period, stimulation was titrated over 5 weeks by adjusting the output current of the implant to each patient’s upper comfort level. From week 5 onward, patients were then treated for 1 min QD or 1 min QID. During the long-term extension study, patients were permitted to be treated with adjunctive DMARD, including non-experimental conventional and b/tsDMARDs. Adjunctive treatment could be provided at any time if a patient experienced worsening RA symptoms or did not experience adequate clinical improvement. The decision about the need, type, and timing of adjunctive treatment was left to the investigators and the patient’s discretion.

This study complied with the Declaration of Helsinki and Good Clinical Practice Guidelines established by the International Conference on Harmonization. The institutional review board and independent ethics committees at each investigational center (Biomedical Research Alliance of NY and Western Copernicus Group) reviewed and approved the final protocol and informed consent documentation. All patients provided their written informed consent.

Description of Neuroimmune Modulation System

The neuroimmune modulation system (SetPoint Medical, Valencia, CA) consisted of two implanted components: a miniaturized neurostimulator with integrated electrodes and a silicone pod to position the pulse generator on the vagus nerve; and two external components: a wireless charger and an iPad application to program the pulse generator. The neurostimulator was implanted on the left cervical vagus nerve through a single incision under general anesthesia. Once implanted, the vagus nerve fits into a groove on the base of the neurostimulator, where the integrated electrodes are oriented in direct apposition to the nerve for efficient stimulation. The device is powered by a rechargeable lithium-ion battery with a usage life of at least 10 years. The neurostimulator is recharged using a proprietary radio frequency-based external wireless charger worn around the neck for 5–10 min weekly. The charger also provided wireless telemetry for transmission and receipt of information with an Apple^® iPad-based proprietary software application, which clinical site staff used to program and monitor implants. Details on the surgical implant procedure have been previously published [24].

Stimulation parameters were specifically designed to activate immune modulatory pathways that decrease systemic inflammation. These parameters, including output current, pulse frequency, pulse duration, and duty cycle (on time/off time), were based on extensive preclinical work in several animal models of inflammation [25]. The maximum stimulation parameters used in the study were 2.5 mA current, 250 μs (62.5 μs interphase interval) pulse width, and 10 Hz frequency. All pulse parameters were below or in the low range of what is used clinically for vagus nerve stimulation in epilepsy and depression.

Study Assessments

Patients were assessed for safety and durability of response throughout the follow-up period. In the primary study, demographics and baseline characteristics were collected 6 weeks prior to randomization (4 weeks prior to implantation). Follow-up assessment visits occurred 1 week, 4 weeks, 8 weeks, and 3 months following randomization (day 0). In the long-term, open-label extension, assessments occurred at months 6, 9, 12, 18, 24, and once every 3 months from month 30 through month 36. The primary objective of the extension study was to assess the long-term safety of the device and the neuroimmune modulation treatment, as well as the durability of clinical response. Safety and tolerability were assessed by examining incidence rates of treatment-emergent adverse events (TEAEs) and device effects, serious adverse events (SAEs), serious adverse device effects, and any unanticipated or serious adverse device effects. Safety was also assessed through the collection of device deficiencies, abnormalities in safety laboratory results, electrocardiogram (ECG) telemetry, or a paper ECG rhythm strip during in-clinic stimulations, and treatment-emergent changes in regularly scheduled 12-lead ECGs. Long-term clinical response was determined by assessing changes in CDAI response rates over 36 months.

Statistical Analysis

No formal hypothesis testing was established for this study. The primary safety dataset included all patients enrolled. The efficacy population included all patients implanted with a device and had at least one post-implantation data point available for analysis. A frequency table was generated to summarize the type and number of adverse events (AEs). For all efficacy endpoints, summary statistics and two-tailed 95% confidence intervals for each of the means were calculated.

Results

All patients (N = 14) from four clinical sites in the USA completed the last per-protocol visit (month 3) of the primary study and continued with participation in the long-term extension for assessment of safety and durability of response. These patients were originally assessed for eligibility between March 13 and August 8, 2018. Eleven of 14 patients enrolled in the extension study completed 36 months of follow-up as of March 21, 2022 (Fig. 1). Two patients withdrew consent, and one was lost to follow-up.

Fig. 1.

Patient disposition. *Treatment allocation for 1 patient changed from 1 min 4 times per day (QID) to 1 min once per day (QD) prior to month 12 per patient and investigator discretion in conjunction with ongoing adjunctive therapy

Baseline demographics were similar between treatment groups (Table 1). All but three patients were women, the mean age was 50.9 years, the mean duration of RA was 14.3 years. Patients had previously failed, on average, 4.8 different b/tsDMARDS with 9 out of 14 (64%) having failed a tsDMARD. The mean baseline CDAI and DAS28-CRP were 42.6 and 6.2, respectively.

Table 1.

Baseline demographics of patients at the screening visit of the primary study

	Overall (n = 14)	QD (n = 8)	QID (n = 6)
Age (year)	50.9 (3.4)	52.8 (5.4)	48.5 (3.9)
Female gender (%)	11 (79%)	7 (88%)	4 (67%)
RA disease duration (year)	14.3 (3.1)	17.1 (5.2)	10.5 (1.5)
Prior b/tsDMARD treatment	4.8 (0.5)	4.5 (0.8)	5.2 (0.7)
CDAI	42.6 (4.1)	39.8 (4.7)	46.3 (7.5)
DAS28-CRP	6.2 (0.4)	6.2 (0.5)	6.2 (0.6)
RF positive (%)	12 (86%)	6 (75%)	6 (100%)
ACPA positive (%)	13 (93%)	7 (88%)	6 (100%)

Data are n (%) or mean (standard error of the mean)

ACPA anti-citrullinated protein antibody, b/tsDMARDs biological or targeted synthetic disease-modifying antirheumatic drugs, CDAI clinical disease activity index, DAS28-CRP disease activity score 28 based on C-reactive protein, RA rheumatoid arthritis, RF rheumatoid factor, QD 1 min once per day, QID 1 min 4 times per day

Efficacy

Change in efficacy outcomes were assessed relative to baseline, defined as the initiation of stimulation. For the treatment group this was on day 0, approximately 14 days following implantation and immediately prior to randomization. For the control group, baseline was defined at month 3, just before crossover to active daily stimulation. The mean CDAI score improved from 36.8 (SEM 5.6) at baseline (N = 14) to 15.0 (SEM 3.5) at month 36 (N = 11). The median change in CDAI score from day 0 to month 36 was − 17.8 (SEM 4.9) (Fig. 2). At month 36, 64% (7/11) of patients had a CDAI response that met or exceeded the minimal clinically important difference (MCID). Out of these seven patients, two were treated with neuroimmune modulation without augmentation with a b/tsDMARD, while five combined neuroimmune modulation with adjunctive b/tsDMARD.

Fig. 2.

36-Month mean change (± standard error of the mean) from baseline in clinical disease activity index (CDAI). Number of patients with CDAI data: N = 14 from screening to day 0, N = 13 from month 3 to month 12, N = 11 from month 18 to month 36. PS primary study, LTE long-term extension

The proportion of patients in low disease activity increased from 7% (1/14) at the initiation of stimulation to 46% (5/11) at month 36, while the proportion of patients in high disease activity declined from 71% (10/14) at the initiation of stimulation to 36% (4/11) at month 36 (Table 2). Among the five patients that experienced reduced disease activity at month 36, three were on continuous neuroimmune modulation with adjunctive b/tsDMARD. Low disease activity was achieved by four of these five: two transitioned from high disease activity to low disease activity at month 36; two transitioned from moderate disease activity to low disease activity at month 36; and one transitioned from high disease activity to moderate disease activity at month 36. Three out of the eight patients successfully reduced or maintained their disease activity by incorporating a b/tsDMARD with a mechanism of action to which they had previously shown no response.

Table 2.

Clinical disease activity index (CDAI) clinical disease activity and number of patients with augmented therapy

CDAI	Definition	Day 0/month 3a (N = 14)	Month 6 (N = 13)	Month 9 (N = 13)	Month 12 (N = 13)	Month 18 (N = 11)	Month 24 (N = 11)	Month 30 (N = 11)	Month 36 (N = 11)
Remission	0 to ≤ 2.8	0 (0)	2 (15)	2 (15)	1 (8)	2 (18)	2 (18)	3 (27)	0 (0)
Low	> 2.8 to ≤ 10	1 (7)	1 (8)	2 (15)	1 (8)	1 (9)	2 (18)	2 (18)	5 (45)
Moderate	> 10 to ≤ 22	3 (21)	4 (31)	2 (15)	4 (31)	3 (27)	2 (18)	1 (9)	2 (18)
High	> 22	10 (71)	6 (46)	7 (54)	7 (54)	5 (45)	5 (45)	5 (45)	4 (36)
Augmented with b/tsDMARDs		N/A	3 (23)	2 (15)	3 (23)	6 (55)	8 (73)	9 (82)	9 (82)

Data are n (%)

b/tsDMARDs biological or targeted synthetic disease-modifying antirheumatic drugs, N/A not applicable

^aAt initiation of stimulation (day 0 for treatment groups and month 3 for control to active crossover group)

Safety

The long-term safety of the device and the neuroimmune modulation were maintained through 36 months of follow-up, with no new reports of related SAEs, deaths, serious infections, cardiovascular events, or unanticipated device-related adverse events (Table 3). One patient developed two anticipated device-related adverse events. The patient reported mild oropharyngeal pain and moderate implant site pain that resolved without additional sequelae with no evidence of site infection. Three SAEs that were all unrelated to the device or stimulation occurred in three patients. One patient developed severe pneumonia and another developed squamous cell carcinoma of the lung. The severe pneumonia was determined to be related to coronavirus disease 2019 (COVID-19) infection. Another patient was diagnosed with large granular lymphocytic leukemia (LGL) presenting as severe neutropenia. The patient’s LGL was in remission following treatment and was thought to be disease-related, as LGL is associated with long-standing, poorly controlled RA [26]. A total of 60 non-serious adverse events were reported across the 11 patients who completed the 36-month study with 60% being mild and 40% being moderate. These adverse events were unrelated to the study device, implant procedure, and/or stimulation. Unrelated AEs of infection were reported and included otitis externa, viral upper respiratory tract infection, skin infection, sinusitis, and COVID-19 infection. All these infections resolved without sequelae. No device deficiencies that affected patient safety or care were identified. A total of 59 protocol deviations were recorded, all determined to be minor and did not affect data validity. No patient required surgical revision or device explantation procedure at any point throughout the 36-month follow-up period. Overall, long-term experience with the device and therapy was well tolerated by patients with no new safety risks or concerns identified.

Table 3.

Summary of safety during the 36-month follow-up period

	# events, # patients
Adverse events (AE)	65, 14
Non-serious unrelated	60, 14
Mild	32, 7
Moderate	26, 10
Severe	2, 2
Non-serious related	2
Device/stimulation-related	2, 1
Mild oropharyngeal pain	1, 1
Moderate implant site pain	1, 1
Implantation/explantation-related	0
Serious unrelated AE	3, 3
Pneumonia	1, 1
Large granular leukemia	1, 1
Squamous cell carcinoma of lung	1, 1
Serious related AE	0
AEs leading to discontinuation	0
Deaths	0

Discussion

Despite the expanding therapeutic landscape for managing RA, a significant proportion of patients are unable to achieve or maintain adequate disease control. These patients are less likely to respond to additional, successive treatments [27]. This first-in-human trial specifically enrolled patients with active disease despite prior treatment with multiple b/tsDMARDs with multiple mechanisms of action. Although designed as a small pilot study, meaningful improvements in clinical disease activity were observed in this difficult-to-treat population.

Our findings offer the first documentation of the long-term safety and potential efficacy of vagus nerve-mediated neuroimmune modulation as a treatment for RA. Over 36 months of follow-up, sustained clinical benefit was observed, with a median change in CDAI score of − 17.8, indicating substantial long-term disease activity improvement. At month 36 of the study, 9 of 11 patients were treated with neuroimmune modulation therapy combined with b/tsDMARDs. Among these, five maintained or improved their disease activity levels long-term, while three lowered or maintained their disease activity with a previously ineffective b/tsDMARD mechanism of action. These findings suggest that neuroimmune modulation may lead to significant improvement of disease activity, either on its own or combined with b/tsDMARDs.

Long-term safety findings were unremarkable, with no new related serious adverse events or serious device-related or serious stimulation-related complications reported during the 36-months follow-up. The three serious adverse events were unrelated to the device or stimulation, all of which resolved without complications. The device performed as intended, with no need for surgical revisions or device explantations, and the device deficiencies noted did not result in adverse events or changes to the study device’s risk profile.

Vagus nerve-targeted neuroimmune modulation offers a promising novel approach to improving RA treatment by leveraging the body’s own central nervous system-mediated regulatory mechanisms that reflexively modulate systemic inflammation. Electrically stimulating the vagus nerve can activate anti-inflammatory pathways including the cholinergic anti-inflammatory pathway, which reduces the release of pro-inflammatory cytokines including TNF, interleukin-1, and interleukin-6. These and other cytokines drive the systemic inflammation characteristic of RA [21, 22]. Unlike broad-spectrum immunosuppressive drugs, neuroimmune modulation provides a targeted approach to reducing inflammation without broadly suppressing the immune system. By effectively modulating inflammation, neuroimmune modulation can potentially reduce the need for conventional DMARDs and b/tsDMARDs, minimizing their associated side effects and immunosuppressive risks. Neuroimmune modulation may also be used in combination with approved therapies to potentially enhance effectiveness. As a non-pharmaceutical therapy, neuroimmune modulation mitigates many of the adverse events associated with long-term immunosuppressive medication use, such as cytopenias due to bone marrow suppression, hepatotoxicity, and increased risk of serious infection. The principal risks inherent to this approach are implantation procedure-related or occur during stimulation delivery, which in this study was administered 1-min QD or QID. Adverse events typically associated with cervical vagus nerve stimulation devices include neck pain or discomfort at the implant site, and voice changes such as hoarseness and coughing, which are typically managed by adjusting stimulation settings [28]. Although rare, device malfunction or failure may require surgical revision or replacement [29]. During the initial 12 weeks of this study, as previously reported, one patient developed left vocal cord paresis within 24 h of the procedure, presenting with stridor and dysphonia; these symptoms resolved over time without permanent clinically significant sequelae, and the patient reported no additional issues during long-term follow-up [23]. Hemodymanic changes including hypotension or bradycardia were not observed during the study. Overall, the safety profile observed was consistent with prior studies of vagus nerve stimulation devices, demonstrating that this approach is generally well tolerated and suitable for long-term use [30]. The established success of vagus nerve stimulation in the treatment of drug-resistant epilepsy, where it has proven both effective and durable over extended periods, further supports its potential as a long-term management strategy for RA, including in patients with treatment-resistant disease [31].

First-in-human clinical studies come with several limitations that can impact their reliability and generalizability such as small sample size, patient heterogeneity, and lack of statistical power. Despite these limitations, our findings provide important preliminary evidence supporting the long-term therapeutic potential of vagus nerve-mediated neuroimmune modulation for the treatment of RA and served as the foundation for the larger randomized RESET-RA study, designed to generate registration-enabling evidence for a potential US Food and Drug Administration approval [32].

Conclusions

Neuroimmune modulation offers an important new potential treatment for managing RA, including for patients who do not respond to or cannot tolerate biological DMARDs. This novel, non-pharmacologic treatment option demonstrated sustained efficacy and a favorable safety over 3 years of follow up, suggesting its potential for durable RA management. Importantly, the ability to use neuroimmune modulation in combination with existing DMARDs without overlapping safety concerns broadens its clinical applicability. Overall, these findings support the development of this novel device-based treatment approach for RA.

Author Contributions

All authors (Norman B Gaylis, David Sikes, Alan Kivitz, Diane Lewis Horowitz, Melissa Evangelista, Yaakov A Levine, and David Chernoff) contributed to the study conception and design. Material preparation, data collection and analysis were performed by Melissa Evangelista, Yaakov A Levine, and David Chernoff. The first draft of the manuscript was written by David Chernoff and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

SetPoint Medical sponsored and provided financial support for the trial. The journal’s Rapid Service Fee was funded by SetPoint Medical.

Data Availability

The dataset generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Conflict of Interest

Norman B Gaylis reports personal fees from Arthritis and Rheumatic Disease Specialties, during the conduct of the study. Alan Kivitz is involved in a clinical research trial sponsored by Altoona Center for Clinical Research, during the conduct of the study; reports personal fees from AbbVie, Celgene, Horizon, Janssen, Merck, Genzyme, Regeneron, Sun Pharma Advanced Research, Boehringer Ingeleheim, and Flexion, personal fees, participation on an advisory committee or review panel, and stock ownership from Pfizer and Sanofi, personal and consultant fees, and stock ownership for Gilead, and stock ownership in Amgen and GlaxoSmithKline, outside of the submitted work. Melissa Evangelista reports personal fees and stock from SetPoint Medical, outside of the submitted work, and is an employee of SetPoint Medical. Yaakov A Levine reports personal fees and stock from SetPoint Medical, outside of the submitted work, and is an employee of SetPoint Medical. YAL also holds several patents around use of vagus nerve stimulation for inflammatory disease. David Chernoff reports personal fees, outside of the submitted work, and is an employee and stockholder at SetPoint Medical. David Sikes and Diane Lewis Horowitz declare no competing interests.

Ethical Approval

This study complied with the Declaration of Helsinki and Good Clinical Practice Guidelines established by the International Conference on Harmonization. The institutional review board and independent ethics committees at each investigational center (Biomedical Research Alliance of NY and Western Copernicus Group) reviewed and approved the final protocol and informed consent documentation. All patients provided their written informed consent.