FDA-Authorized Vagus Nerve Stimulation Devices: Indications, Regulatory Pathways, and Pivotal Evidence

FDA-authorized vagus nerve stimulation devices: indications, pathways and pivotal evidence

FDA has authorized vagus nerve stimulation (VNS) products through all three main device pathways. Implanted cervical stimulators went through premarket approval (PMA) for epilepsy, depression, stroke rehabilitation and rheumatoid arthritis. Obesity was a separate PMA for an implanted abdominal vagal-blocking system, not a cervical stimulator. Non-invasive cervical stimulation for headache started with a De Novo and grew through a chain of 510(k)s. Several ear-worn or percutaneous nerve field stimulators whose labeled targets include branches of cranial nerve X were authorized by De Novo or 510(k) for opioid withdrawal and pediatric IBS pain. The pattern is consistent: novel implanted indications needed a randomized, blinded pivotal trial with an active or sham control, while follow-on 510(k)s often added indications or populations based on previously submitted trials, extrapolation, or literature.

At-a-glance summary

Implanted cervical VNS: VNS Therapy System (P970003)

Epilepsy (original PMA, 1997)

FDA approved the LivaNova (originally Cyberonics) VNS Therapy System under P970003 on July 16, 1997, under product code LYJ ("stimulator, autonomic nerve, implanted for epilepsy") ²⁵⁰. The original indication was adjunctive therapy for adults and adolescents over 12 years of age with refractory partial-onset seizures ⁷¹⁴.

The effectiveness evidence centered on two randomized, parallel-group studies, E03 (126 patients) and E05 (262 patients), comparing HIGH and LOW stimulation ⁵⁸⁰. The design used a 12-week baseline, randomization, a 12-week randomized phase and an open-label extension ⁵⁷⁵. Notably, the comparator was low-dose stimulation, not sham or no stimulation. Acute-phase seizure-rate results were:

Safety findings were mostly mild or moderate events. Across the randomized and extension phases, the most common non-serious events were voice alteration (50%) and increased cough (41%). Serious surgery-related events occurred in 4.1% of patients and serious stimulation-related events in 1.2% ⁵⁸³⁵⁸⁴. The labeling also noted at least two unequivocal episodes of status epilepticus among 441 adults, with no reliable incidence estimate possible ⁵⁸⁵.

Treatment-resistant depression (S050, 2005)

Supplement S050, approved July 15, 2005 under product code MUZ ⁷⁴, added adjunctive long-term treatment of chronic or recurrent depression in adults 18 or older experiencing a major depressive episode who have not responded adequately to four or more adequate antidepressant treatments ⁷¹⁸. The available S050 approval record does not set out the depression study names, designs, endpoints or FDA's benefit-risk reasoning ⁷¹⁷. Readers should not assume the epilepsy E01 to E05 data apply here, since those studies addressed seizure reduction only ⁹⁶⁴.

Pediatric epilepsy expansion (S207, 2017)

S207, approved June 23, 2017, extended the epilepsy indication to patients 4 years of age and older with refractory partial-onset seizures, keeping the adjunctive-therapy wording ^711714715. The available record does not identify the data sources supporting the age expansion ⁷¹¹.

The PMA remains active. The supplement series runs past S250, with the most recent listed supplement (S254) decided October 2, 2026 ²⁵¹.

Vivistim Paired VNS System (P210007): stroke rehabilitation

FDA approved Vivistim on August 27, 2021, under the new product code QPY ("stimulator, autonomic nerve, implanted for stroke rehabilitation") ³⁰³. The device stimulates the vagus nerve during rehabilitation therapy to reduce upper-extremity motor deficits and improve motor function in chronic ischemic stroke patients with moderate to severe arm impairment ⁵⁹⁹. What makes it new is the paired stimulation paradigm: stimulation is timed to rehabilitation movements rather than delivered continuously.

• Pivotal trial: VNS-REHAB (MT-St-03) was a blinded, randomized, controlled study in 108 implanted participants. Controls received the same rehabilitation with minimal VNS to maintain blinding and crossed over after the 90-day assessment ⁶⁰⁰.
• Primary endpoint: Change in Upper Extremity Fugl-Meyer (FMA-UE) score from baseline to one day after six weeks of in-clinic therapy. Scores improved 5.0 ± 4.4 points with VNS (n=53) vs 2.4 ± 3.8 with control (n=55), p=0.0014 ^601602603.
• Supporting studies: The MT-St-01 feasibility study showed a 9.3-point FMA-UE gain vs 3.0 in non-implanted controls ⁸⁰³. In the MT-St-02 pilot, 88% of VNS subjects vs 33% of controls had a clinically meaningful FMA-UE response at three months ⁸⁰⁴.
• Safety and benefit-risk: Adverse events were generally mild to moderate and transient (hoarseness, cough, throat irritation, pain). There was one serious surgery-related event, vocal-cord paresis with dysphonia, from which the patient fully recovered ⁶⁰⁴⁶⁰⁵. FDA concluded that probable benefits outweigh probable risks, citing a low serious-AE rate and a durable effect to at least 12 months ⁸⁰⁵.

SetPoint System (P240039): rheumatoid arthritis

The SetPoint System was approved July 30, 2025, under product code SFJ ("vagus nerve stimulator for rheumatoid arthritis") ¹³. It is the first FDA-approved VNS indication outside neurology, psychiatry and metabolism. FDA had granted Breakthrough Device status on March 15, 2024 ⁸²⁰.

• Indication: Adults with moderately to severely active RA who have had an inadequate response, loss of response, or intolerance to one or more biologic or targeted synthetic DMARDs ⁸¹⁷.
• Pivotal trial: RESET-RA (SPM-020), a prospective, multicenter, randomized, double-blind, sham-controlled trial that implanted and randomized 242 patients (122 active, 120 sham). The blinded primary assessment was at Week 12, followed by crossover and open-label follow-up ^705706707.
• Primary endpoint: ACR20 response at Week 12. In the ITT analysis, 35.2% (43/122) responded with active stimulation vs 24.2% (29/120) with sham. The reported treatment difference was 11.8 percentage points (95% CI 0.6 to 23.1; p=0.0209), slightly larger than the crude gap between the two response rates ⁷⁰⁸⁷⁰⁹.
• Safety: Device-related serious adverse events occurred in 1.6% of patients, all implant-related and resolved. There were no deaths and no unanticipated adverse device effects. FDA also noted charger-connection problems and device-related protocol deviations ⁷¹⁰.
• Dosing: The pilot study (SPM-008) used one minute of stimulation once daily, and four-times-daily dosing is described as not indicated ⁸¹⁹.

Post-approval activity is already under way, with supplements S001 through S003 (the latter covering the charger) decided between January and October 2026 ⁷⁶⁵.

MAESTRO Rechargeable System (P130019): obesity via vagal blocking

The MAESTRO system, approved January 14, 2015 under product code PIM ("neuromodulator for obesity") ³⁰⁹, delivers vagal blocking (VBLOC) therapy ⁶⁷⁸. Its indication covers weight reduction in adults 18 and older with a BMI of 40 to 45 kg/m², or 35 to 39.9 kg/m² with at least one obesity-related comorbidity, who have failed a supervised weight-management program within the prior five years ⁶⁷⁷.

The approval matters for regulatory strategy because FDA approved the device even though it missed its prespecified pivotal endpoints:

• ReCharge trial: Prospective, multicenter, randomized (2:1), double-blind, sham-controlled, with 239 participants randomized (162 VBLOC, 77 sham) and 233 implanted ⁶⁷⁸⁶⁷⁹.
• Co-primary endpoints, neither met: Mean percent excess weight loss (%EWL) was 24.4% with VBLOC vs 15.9% with sham. The 8.5-point difference (95% CI 3.1 to 13.9) fell short of the 10-point superiority margin, and the responder thresholds were also missed ^680681682.
• Advisory panel: The panel voted 8 to 1 on safety, 4 to 5 on effectiveness, and 6 to 2 (one abstention) that benefits outweigh risks ⁸²¹.
• FDA rationale: FDA cited the absolute weight loss, the more than half of VBLOC patients reaching 20% or more EWL, weight loss sustained to 12 to 18 months, and few serious adverse events. It also noted that the unexpectedly high sham response compressed the between-group margin ⁸²².

Non-invasive cervical VNS: gammaCore (De Novo followed by 510(k) iterations)

De Novo DEN150048 (2017)

The gammaCore Non-invasive Vagus Nerve Stimulator received De Novo authorization on April 14, 2017 (product code PKR) ³⁴³. It is indicated for non-invasive vagus nerve stimulation on the side of the neck for the acute treatment of pain associated with episodic cluster headache in adults. Effectiveness in chronic cluster headache was not established ⁵⁵⁶. The device type is "external vagal nerve stimulator for headache," Class II. The corrected classification order cites 21 CFR 882.5892, while the De Novo decision summary prints 882.8592 ^571572573.

The evidence is a useful case study in subgroup-driven labeling:

• ACT1 was a multicenter, double-blind, randomized, sham-controlled pivotal trial (150 randomized; 133 in the mITT analysis) ^557558559. The primary endpoint (mild or no pain at 15 minutes for the first attack) was not significant overall: 26.7% vs 15.1%, P=0.10. It was significant in the episodic subgroup: 34.2% vs 10.6%, P=0.008 ^562563564.
• ACT2 reported attack-level pain freedom at 15 minutes in episodic patients of 35.2% vs 7.4% (P=0.04) ⁵⁶⁵⁵⁶⁶.
• No device-related serious adverse events were reported. The most common effects were muscle twitching, neck discomfort and application-site irritation ^567568569570.

Those subgroup results line up with the final indication: episodic cluster headache only, with chronic cluster headache excluded ⁵⁵⁶.

510(k) expansion chain

Two points stand out for 510(k) strategy. First, FDA accepted new indications within the same device type through 510(k) where the technology was unchanged and the treatment protocol differed ^850753866. Second, the migraine-prevention clearance rested on a sham-controlled program whose primary ITT analyses were not shown to be positive in the cleared summary, with support coming from a post hoc adherence-based analysis ⁷⁵⁹⁷⁶².

Auricular and percutaneous nerve field stimulators with vagal targets

A related group of devices stimulates the external ear region. Their labeled targets include branches of cranial nerve X, but they are not classified as vagus nerve stimulators.

• NSS-2 BRIDGE (DEN170018, 2017-11-15) ⁶¹¹: A percutaneous nerve field stimulator applied to branches of cranial nerves V, VII, IX and X and the occipital nerves, as an aid to reduce opioid withdrawal symptoms ⁸²⁹. Evidence was a single-arm, open-label, retrospective study (n=73). COWS scores fell 62.7% at 20 minutes and 84.6% at 60 minutes ^833834835, supported by a 1,207-patient retrospective safety chart review ⁸³⁷.
• IB-Stim (DEN180057, 2019-06-07) ⁹⁶⁷: Same nerve targets, indicated for functional abdominal pain associated with IBS in patients aged 11 to 18 ⁹⁸². The De Novo created 21 CFR 876.5340, Class II, "non-implanted nerve stimulator for functional abdominal pain relief" (QHH) ⁹⁸⁸. The pivotal double-blind, sham-controlled trial enrolled 115 adolescents with mixed functional GI disorders ⁹⁸³⁹⁸⁵. FDA narrowed the indication to IBS based on a post hoc subgroup analysis ⁹⁸⁶⁹⁸⁷. Later 510(k)s by NeurAxis followed in 2024 and 2025 ^970969968.
• Sparrow Therapy System (K201873): A transcutaneous earpiece cleared against NSS-2 BRIDGE for opioid withdrawal ⁷⁸⁹⁷⁹⁰. It was supported by a 26-patient COWS study (50.4% mean reduction at 60 minutes) and a patient-preference analysis ⁷⁹⁵⁷⁹⁶.
• Sparrow Ascent (K230796, K251246): Cleared for the same indication ⁷²⁵⁶³⁶. K251246 states that the device stimulates the auricular branch of the vagus nerve (ABVN) at the mastoid, rather than at the cymba concha as the predicate does. This site change was supported by a sham-controlled crossover fMRI study in 24 healthy adults showing correlated brain activation (r=0.61, p<0.001) ⁶³⁷⁶⁴⁰.

Other devices in the same product codes (for example Drug Relief, NET Device, First Relief and TEA Device) were not individually reviewed here ^302606965966.

Regulatory takeaways

• The PMA is the default for implanted VNS in a new indication. Each new implanted indication (epilepsy, stroke, RA, obesity) received its own product code and a randomized pivotal trial ^2503033309. Controls ranged from low-dose stimulation (E03/E05) and minimal-stimulation active control (VNS-REHAB) to full sham (RESET-RA, ReCharge) ^580600706678.
• A missed primary endpoint does not automatically mean denial. MAESTRO's approval shows FDA weighing absolute clinical benefit, sham-response inflation and safety in a benefit-risk framework after a split advisory panel ⁸²¹⁸²².
• Non-invasive VNS shows the De Novo-then-510(k) lifecycle. One De Novo created the classification, and later 510(k)s added acute migraine, cluster and migraine prevention, adolescent use and hemicrania indications, often without new device-specific clinical testing ^343773922864.
• Subgroup and post hoc analyses have shaped labeling. Episodic-only cluster headache (gammaCore), IBS-only pediatric abdominal pain (IB-Stim) and the PREMIUM adherence analysis all show FDA matching the cleared indication to the population where evidence was strongest ^564987759.
• Gaps in the public record. The available S050 and S207 approval records do not set out the clinical evidence behind the depression and pediatric epilepsy expansions ⁷¹⁷⁷¹¹. PRESTO's sample size and numerical results are not stated in the K173442 summary ⁸³⁸. These are good candidates for a follow-up review of SSEDs, panel transcripts or published literature.