In 2000, in a lab on Long Island, a surgeon named Kevin Tracey did something that sounds like it shouldn’t have worked. He took a rat, ran a mild electrical current through its vagus nerve, the long wandering cable that connects the brainstem to nearly every organ below the neck, and then hit the animal with a dose of bacterial toxin that would normally send its immune system into a lethal cytokine storm. The storm didn’t come. The current had told the body to keep its own inflammation in check. Tracey published the result in Nature and gave the circuit a name: the inflammatory reflex.

Twenty-five years later, that experiment is a shipping medical device. In August 2025 the FDA granted premarket approval to the SetPoint System, a small implant that sits on the left vagus nerve in the neck and fires for about sixty seconds a day. It is the first bioelectronic medicine ever approved to treat an autoimmune disease. And when you read the patents behind it, you realize it is doing, with a wire, exactly what the best-selling drug in the history of the pharmaceutical industry does with a molecule.

The molecule everyone was chasing

The drug is Humira. For most of the 2010s it was the single largest product in medicine, peaking near $20 billion a year in global sales before biosimilars ate into it, according to industry reporting compiled by MDedge and GlobalData. Humira is an antibody. You inject it, and it circulates through your body physically grabbing molecules of tumor necrosis factor (TNF, the master switch of inflammation) and neutralizing them before they can inflame a joint. Every anti-TNF biologic that treats rheumatoid arthritis, Crohn’s disease, and psoriasis works on the same logic: flood the bloodstream with something that soaks up TNF.

Tracey’s reflex offered a different route to the same molecule. Stimulate the vagus, and the signal travels to the spleen, where a specialized population of T cells releases acetylcholine. The acetylcholine lands on receptors on macrophages — the immune cells that manufacture TNF — and tells them to stop. Same target. Opposite tool. Humira removes TNF after the body makes it; the nerve pulse tells the body to make less in the first place. One is a $20-billion-a-year antibody you inject for life. The other is a one-minute electrical dose from an implant the size of a multivitamin.

This is the hidden twin. Two therapies, built by two entirely different industries, one wet-chemistry pharmaceutical and one implantable-electronics, that converge on the same node in the same biological circuit. Nobody designing Humira was thinking about nerves. Nobody wiring a nerve stimulator in the 1990s was thinking about arthritis.

What’s actually in the patents

SetPoint Medical, the company Tracey helped found, holds 49 US patent grants dating back to 2000. Twenty-four of them explicitly invoke inflammation, and the earliest of those was granted in 2013 — a decade before the device reached patients. Read them in sequence and you watch the engineering problem get solved one constraint at a time.

The early grants are about aim. A 2013 patent covers “devices and methods for optimizing electrode placement for anti-inflammatory stimulation” — figuring out exactly where on the vagus to put the current so you trigger the immune reflex without hitting the fibers that control heart rate. A 2014 grant then claims “single-pulse activation of the cholinergic anti-inflammatory pathway,” the discovery that you don’t need to zap the nerve continuously; one brief pulse a day is enough to keep TNF suppressed. That single insight is what turns a bulky epilepsy-style stimulator into something you can forget is inside you.

The recent grants are about disappearing the hardware. SetPoint’s 2025 patents describe a “leadless” and “batteryless” microstimulator with a total volume under 1.5 cubic centimeters, held against the nerve by a cuff with a pocket sewn into it so a surgeon can slide the device in and out. There is no wire running down to a battery pack in the chest, which is how legacy nerve stimulators are built. Instead, one patent describes a wearable charger worn “around the subject’s neck” that tops the implant up in under ten minutes a week, and an external controller the patent literally calls a “prescription pad,” so the doctor sets the dose the way they’d write a script. Another 2025 filing adds closed-loop control: the device reads a marker of inflammation and adjusts the gap between pulses on its own.

None of these read like a keyword pile. Delete the phrase “vagus nerve” from every one and they still describe the same machine getting smaller, smarter, and more autonomous around a single mechanism. That is the difference between a technology and a label.

The number that didn’t move — and the one that did

Here’s the counterintuitive part. If you count US patent grants that mention vagus nerve stimulation, the line is almost flat: roughly 30 to 38 a year, every year, for the past decade. By the crude metric of patent volume, nothing is happening. Vagus stimulation has been FDA-approved hardware since Cyberonics won clearance for epilepsy in 1997 and for depression in 2005. The count is old news.

What changed is what the nerve is being aimed at. For thirty years, VNS was a brain treatment: you stimulated the vagus to quiet seizures or lift depression. The inflammatory reflex reroutes the exact same anatomy at the immune system. The Feinstein Institutes, where Tracey is now president and CEO, hold 16 vagus-related grants of their own, and the clinical pipeline has caught up: in ClinicalTrials.gov, studies pairing vagus stimulation with inflammatory or autoimmune disease now number a dozen completed and seven actively recruiting, with targets running well past arthritis into Crohn’s, lupus, and colitis. SetPoint’s newest grants, from 2024 and 2025, push the same reflex toward neurodegenerative disease. The volume of patents was a red herring. The vector was the story.

Does it work, and who cares

Modestly, and a specific person. The approval rests on RESET-RA, a 242-patient, sham-controlled trial. At three months, 35.2 percent of patients getting real stimulation hit the standard ACR20 improvement threshold, versus 24.2 percent of patients whose implants were switched off, according to the peer-reviewed writeup of the approval in the journal Rheumatology. That is not a miracle margin. It is an eleven-point edge over a convincing placebo.

But look at who the device is for: rheumatoid arthritis patients who have already failed the biologics, the people for whom Humira and its successors stopped working or never did. For that group the current menu is to keep cycling through immunosuppressant drugs that carry infection risk and cost tens of thousands of dollars a year, indefinitely. A one-time implant that runs for a minute a day changes the shape of that bill and that risk profile, even at an eleven-point response edge. The open question, which the Rheumatology authors flag directly, is reimbursement: nobody has published the head-to-head economics against a lifetime of biologics. That number, when it lands, will decide whether this stays a niche for the drug-refractory or becomes something a payer offers earlier.

The deeper lesson is the one Steven Johnson keeps pointing at: the breakthrough was already wired in. The body has had a nerve-controlled brake on its own inflammation for as long as there have been vertebrates. The hardware to press that brake was FDA-approved and sitting in operating rooms since the Clinton administration. The drug industry spent $20 billion a year chasing the same molecule from the other direction. All it took was one surgeon on Long Island to notice that the two halves of the puzzle had been built in different buildings by people who never spoke.

Method note. Patent figures come from roughly 9.3 million US utility grants sourced from USPTO bulk grant XML, searched by full text and grouped by assignee, with variant spellings of each company combined; patent counts and the SetPoint portfolio timeline were pulled fresh for this piece. Clinical-trial counts come from ClinicalTrials.gov. The RESET-RA results, device description, and August 2025 approval date are drawn from the peer-reviewed approval summary in Rheumatology (PMC12889326) and SetPoint Medical’s own announcement; Humira sales figures from MDedge and GlobalData reporting; the founding Nature work and Tracey’s role from the Feinstein Institutes. A caveat for skeptics: patent volume in vagus stimulation has been flat for a decade, so this is a story about where an existing technique is being pointed, not about a filing surge. And RESET-RA’s effect size is real but modest — an eleven-point response edge over sham at three months, in patients who had already exhausted drug options.