Vagus nerve stimulation: what is approved, what is promising and what is still unproven

This article is general information, not medical advice.

Vagus nerve stimulation (VNS) sends small electrical pulses to the vagus nerve, which runs from the brain through the neck to the internal organs (WashU). In the implanted depression and stroke systems, a device sits under the skin of the chest with a wire to the vagus nerve in the neck. One handheld device stimulates the nerve through the skin instead.

This article keeps four things apart: what regulators have approved, what we have read in peer-reviewed sources, what has been reported in press releases and trade press, and ideas that remain speculative. The regulatory details are for the United States, where the FDA reviews these devices. This article does not cover the UK position.

Approved uses (United States)

Epilepsy. The FDA's premarket approval record for the VNS Therapy System (PMA P970003) gives a decision date of 16 July 1997 and the generic device name "Stimulator, autonomic nerve, implanted for epilepsy". The record's approval statement names the device, then called the NeuroCybernetic Prosthesis (NCP) System, but does not set out the indication in full.

Depression. A later FDA supplement, decided on 15 July 2005, states that the device is indicated "for the adjunctive long-term treatment of chronic or recurrent depression for patients 18 years of age or older who are experiencing a major depressive episode and have not had an adequate response to four or more adequate antidepressant treatments".

Stroke rehabilitation. The FDA record for the Vivistim System (PMA P210007) gives a decision date of 27 August 2021. It states that the system "is intended to be used to stimulate the vagus nerve during rehabilitation therapy in order to reduce upper extremity motor deficits and improve motor function in chronic ischemic stroke patients with moderate to severe arm impairment".

Rheumatoid arthritis. The FDA record for the SetPoint System (PMA P240039) gives a decision date of 30 July 2025. It states that the system "is indicated for use in the treatment of adult patients with moderately to severely active rheumatoid arthritis (RA) who have had an inadequate response, loss of response, or intolerance to one or more biological or targeted synthetic disease modifying antirheumatic drugs (b/tsDMARDs)".

Migraine (non-invasive). The maker of the handheld gammaCore device announced on 29 January 2018 that it had received 510(k) clearance from the FDA "as an acute treatment of pain associated with migraine in adult patients". This comes from the manufacturer's announcement, not the FDA database. The same page says gammaCore "has not been shown to be effective for the prophylactic treatment" of cluster headache or migraine, so the clearance covers treating pain during an attack, not preventing attacks.

Each approval or clearance covers only the use named in it. None of them shows that the device works for other conditions.

Shown in peer-reviewed sources we read

Epilepsy. A Cochrane review (2022 update, evidence current to 3 March 2022) included five randomised controlled trials with 439 participants with drug-resistant focal seizures. The trials compared different levels or frequencies of stimulation, not stimulation against no device. The authors conclude that "VNS for focal seizures appears to be an effective and well-tolerated treatment", but say the evidence is "of moderate to low certainty".

Migraine (PRESTO). PRESTO was a double-blind, sham-controlled trial at 10 Italian sites that randomised 248 patients with episodic migraine. A later post hoc analysis (Journal of Headache and Pain, 2018, PMC) found that 46.8% of patients using the device had at least a one-point drop in pain on a four-point scale two hours after treating their first attack, against 26.2% on sham. The authors state that these were not the trial's "previously reported primary or key secondary end points". They also say the sham device "delivered an appreciable electrical signal", which may have masked some differences between the groups. We have not read the main PRESTO paper, so we do not report its primary result here.

Reported in press releases and trade press

We have not read the full trial papers behind the results in this section. The figures come from the press and trade reports linked.

Stroke (the Vivistim pivotal trial). News-Medical's report of the FDA approval describes a study of 108 patients at 19 sites in the US and UK. All were implanted. Both groups did the same rehabilitation exercises for six weeks; the control group received a very low level of stimulation for only the first five exercises of each session. On the Upper Extremity Fugl-Meyer Assessment, a stroke-specific measure of arm impairment, the treatment group's score rose by an average of 5 points after six weeks, against 2.4 points for controls. At 90 days after therapy, 47.2% of the treatment group had improved by 6 or more points, against 23.6% of controls. RAPS adds that the full results were published in The Lancet. The sources we read do not report outcomes beyond 90 days.

Depression (RECOVER). Washington University reported in December 2024 on the treatment-resistant depression arm of the RECOVER trial, published in Brain Stimulation. All 493 participants at 84 US sites were implanted, and about half (249) had their devices switched on for the 12-month randomised period. The primary measure showed no significant difference between the groups, while many other measures showed benefits for the active group. Complete remission was rare and did not differ between groups. The report says the "off" group "reported more improvement than expected", and that all participants knew the devices would be switched on at the end of the year, which the principal investigator said may have affected the control group. The device maker, LivaNova, sponsored the trial with the US Centers for Medicare and Medicaid Services.

Rheumatoid arthritis (RESET-RA). Healio reports that this double-blind, randomised, sham-controlled trial of 242 patients was published in Nature Medicine. The primary endpoint was an ACR20 response (a 20% improvement on a standard arthritis measure) at three months: 35.2% of patients on active stimulation reached it, against 24.2% on sham. The 6-month and 12-month figures of 50% and 52.8% come from an open-label period with no control group, so they cannot show what the device caused. Healio reports that the investigator it interviewed, John Tesser, has received consulting fees from SetPoint Medical, the device maker.

Limits and open problems

• Surgery. Implanted systems need an operation. In RESET-RA, Healio reports serious adverse events in 1.6% of patients, all around the time of the operation and all resolved. In the stroke trial, one control participant developed vocal cord palsy after implantation, which resolved five weeks after surgery, according to RAPS.
• Side effects. The Cochrane review found that shortness of breath, voice alteration and hoarseness were more common with high-level than with low-level stimulation.
• Variable response. In RESET-RA, about two thirds of patients on active stimulation had not reached an ACR20 response at three months.
• Control groups. Sham devices may not be fully inactive, as the PRESTO authors noted. The Cochrane authors say effective blinding is difficult because stimulation causes side effects such as voice changes. In RECOVER, knowing that devices would be switched on later may have affected the control group.
• Small and sponsor-led studies. Most of the newer results come from single trials, several with modest numbers, short controlled periods or device makers involved in the research. Results from one device do not transfer automatically to another, or to a different condition.

Speculative

Further uses have been discussed, including other inflammatory diseases and devices that switch stimulation on only when a sensor detects a change in symptoms. The sources we read contain no trial evidence for these, and each would need controlled testing before any clinical use.

What to watch next

• Longer-term results from RESET-RA. Healio reports a four-year open-label period is under way, but it has no control group.
• A larger or independent trial that repeats the stroke rehabilitation benefit, with longer follow-up.
• Further RECOVER results. Washington University reports that participants will be followed for another four years, and that the bipolar depression arm is still running.
• Non-invasive device trials that test a pre-specified preventive endpoint against sham.
• Controlled trials of stimulation in another inflammatory disease.