Genio: The Battery-Free Sleep Apnea Implant

How I explain Genio to my patients

Genio is a bilateral hypoglossal nerve stimulator placed through a single incision under the chin. Like Inspire, it sends gentle pulses to the nerve that controls your tongue, so that during sleep your tongue moves forward, your airway opens, and you breathe normally. That much is the same.

What is different is where it sits and how it is powered. The muscles that pull your tongue forward sit side by side in the midline, just behind the chin, close enough together that one small implant placed between them can reach the nerve branches on both sides at once.^[9] That is the whole idea behind the device: one implant under the chin working on both sides, rather than reaching a single nerve on one side of the neck.

Independent academic anatomy study. Not funded by, and not about, Genio.

The second difference is the power source. Genio has no implanted battery. It is powered from outside, every night, by a small chip you clip onto an adhesive patch under your chin. In the morning you take it off and put the chip on its charger.

The third difference is the timing of the stimulation. Inspire senses your breathing and fires in step with each breath. Genio does not sense your breathing at all — it runs a fixed on-and-off cycle. There are reasons to think synchronising stimulation to the breath helps: in one small study, deliberately inverting the breathing sensor on the unilateral device, and then blunting it, made patients’ apnea measurably worse both times.^[16] Whether that translates into a difference in outcomes between the two devices is a separate question, and I review Genio’s actual success rates in depth below.

That sensing study involved three patients, and its authors are Inspire investigators and consultants. It is a direct experiment rather than an observation, which is what makes three patients worth citing at all — but it is three patients.

Some of these differences may be considered advantages or disadvantages, but ultimately it is up to you to decide what is the best option. If what you want is the two devices set side by side, I have written that comparison separately: Inspire vs Genio — the full side-by-side comparison.

What’s actually being implanted

The implant itself

A single stimulator, weighing under three grams, positioned in the floor of the mouth beneath the chin so that it straddles the nerve branches on both sides. That is the entire implant. No battery. No leads. No wires tunnelled to the chest. No separate breathing sensor.^[7]

The three pieces you keep outside your body

An activation chip, which is reusable and holds its own battery. A disposable patch, a single-use adhesive pad that goes under the chin each night and carries the energy from the chip to the implant. And a charging unit that sits by the bed and recharges the chip during the day.^[11]

The nightly routine is: patch on, chip clipped to the patch, start therapy from the smartphone app, sleep. In the morning, chip off, patch off, chip onto the charger. If you need to get up in the night, you can pause therapy from the app, or simply take the chip off.

Why “battery-free” is not the same as “never needs another operation”

This is the inference I most often have to correct, and I would rather correct it here than in clinic after you have already decided.

It is true that Genio has no implanted battery, and it is fair to say that this removes the scheduled generator-replacement operation that battery-powered stimulators eventually require. It does not mean the implant never needs another operation. In the pivotal trial, roughly one patient in fourteen had a second device operation within the first year, and through twenty-four months 11.3% — about one in nine — had a repositioning, a revision, or a removal.^[1] Those were not battery operations. They were migration, extrusion, malfunction, and loss of stimulation. I come back to this in detail under risks.

What the surgery and recovery look like

Before surgery: the sleep study and the sleep endoscopy

You will need a recent sleep study, and you will need a drug-induced sleep endoscopy — a brief look at your airway under sedation to see where and how it collapses. In the trial that led to approval, every candidate had one, and the recordings were reviewed by an independent central laboratory rather than by the operating surgeon.^[2] If you want to know what that exam involves, I have written it up separately: What is DISE?

The operation itself

General anesthesia, one incision under the chin, and in most cases home the same day. The implant is positioned over the tongue muscle, tested during surgery to confirm it moves the tongue the right way, and secured. Nothing else is opened; there is no chest incision.^[7] If you would rather see it than read about it, there is step-by-step operative footage further down this page — kept behind a warning, because it is real surgical video.

On operative time: the pivotal trial reported a mean of 2.37 hours, give or take about 50 minutes.^[1] Treat that number carefully. Published operative times are not all measured the same way — some count only the time actually spent operating, others count the whole time you are in the operating room, and the two differ by roughly an hour. Trial figures also come from teaching hospitals with trainees in the room. In my own practice both this operation and the unilateral one usually take about an hour of actual operating.

What is fair to say is that surgeons in general have far less experience with Genio, simply because it is newer. That will change with time, and it is worth asking your own surgeon directly how many they have done.

The first few weeks

Expect soreness and swelling under the jaw, worst in the first three days and improving steadily after that. Most people are back to normal activity and diet within two to four weeks. The device stays off during this period.

Turning it on — and the months of fine-tuning that follow

The manufacturer’s labeling says the implant should not be activated for at least six weeks after surgery, to allow proper healing;^[11] in the pivotal trial, activation was done at two months.^[2] After activation you begin using it right away and can adjust the strength yourself at home.

What the brochures do not tell you is how much fine-tuning the trial actually involved. Participants had an in-lab sleep study with the device adjusted while they slept at four months and again at six, and then a further round at eight, nine or ten months before the twelve-month measurement. The protocol was explicit that patients whose therapy was not yet optimized got more of these visits, not fewer.^[1] An independent commentary on the trial made the point that without that kind of structured follow-up, many patients on any nerve stimulator end up under-treated.^[5] Plan on this being a process measured in months, not a switch that gets flipped once.

1. Surgery dayOutpatient, general anesthesia, home the same day in most cases.
2. Days 1–3Most discomfort. Swelling under the jaw is expected.
3. Weeks 1–2Wound check. Steady improvement. Device still off.
4. Weeks 2–6Back to normal activity and diet for most people. Device still off.
5. At least 6 weeksActivation, equipment, and learning the nightly routine.
6. Months 2–10Titration, follow-up visits, and a sleep study to check how well it is working.

Watching the operation, if you want to

Some people want to see exactly what is going to be done to them, and some people very much do not. Both are normal, and neither is the better patient. The footage below is a real Genio implant — my own operative video, shared for teaching. It stays hidden until you choose to open it. Nothing on this page requires you to watch it, and you can understand the operation completely from the text above.

Content warning · real surgery This shows an actual operation, including an incision, blood, and exposed tissue. If you are squeamish, or you are about to have this surgery yourself and think it might unsettle you, it is completely reasonable to skip it. Nothing is lost by doing so. Show me the video Hide the video

The full procedure — about seven minutes

The video is annotated for teaching and follows the operation in order:

• The submental incision and exposure of the anatomy beneath the chin
• Dissection to identify both hypoglossal nerves
• Positioning of the stimulation paddles on each nerve
• Confirmation of bilateral tongue motion before closure

A second case — about five minutes

A second example, so you can see that the anatomy is not identical from one person to the next.

Surgeons, trainees and sleep clinicians who want the instrument set and the rest of the teaching material will find it on the page for health care professionals.

Am I a candidate?

What the FDA actually approved

Two things in that wording are worth noticing. First, read the verb carefully: the label says you must have been confirmed to fail, cannot tolerate, or be ineligible for the standard treatments it lists. That last option matters. It does not mean you have to work through an oral appliance and a weight-loss medication before you can be considered — it means each of those has to have been properly accounted for, and “you are not a candidate for it” counts. In practice the conversation almost always centers on CPAP. (What does it actually mean to fail CPAP?)

Second, and less obviously: the approved indication contains no BMI limit, no upper age limit, and no requirement for a sleep endoscopy, even though the trial that supported it had all three.

What the trial actually required — and why that matters more

The people studied were aged 22 to 75, with a BMI of 32 or under, an AHI between 15 and 65, fewer than 25% central or mixed events, and no complete concentric collapse on sleep endoscopy.^[3] Those are the conditions under which the results below were produced. Outside them, nobody knows. That is why I hold to the trial’s limits in my own practice even where the label is broader.

If you want to check where you stand on two of those: work out your BMI with the checker here, and run your AHI, BMI and CPAP history through the coverage checker to see what your own insurer requires — those criteria are often stricter than the FDA label.

The screening reality

It is worth knowing how selective this was. 687 people consented to be screened; 115 were implanted. Most were excluded because their sleep study did not qualify; 64 were excluded specifically because their sleep endoscopy showed complete concentric collapse.^[1] If you are told you are not a candidate, you are in ordinary company.

What the data shows

Where Genio’s evidence comes from — and how young it is

I want you to see the shape of the evidence before you see any number from it. The entire published record for this device is one pivotal trial,^[2] one early feasibility trial,^[4] one small non-randomized comparison,^[6] a handful of single-patient case reports, and the FDA’s review document — roughly 140 patients in the published primary literature.

It is worth separating two things that often get run together. Published effectiveness results stop at twelve months: that is the reporting point of the pivotal trial, and the earlier feasibility trial reported at six.^[2][4] Safety follow-up runs considerably longer. The pivotal trial is designed to run five years, and by the time FDA reviewed it, 74 patients had passed two years and 30 had passed three — none had reached four.^[1] Some of the numbers further down this page, including the two-year reoperation rate, come from that longer safety record rather than from the twelve-month paper.

Genio has also been in patients longer than its US approval date suggests. The first-in-human trial enrolled between April 2017 and February 2018, and the device has been in European use since,^[4] which is where the European follow-up data further down comes from. What arrived in August 2025 was FDA approval, not the first patient.

The gap with the unilateral device is still there, and I would not want to talk you out of it: that device has been implanted in the United States since 2014, with published five-year outcomes and a registry of thousands of patients. Genio’s published record is a fraction of that. But the accurate framing is “a much shorter and thinner track record,” not “a year of data.”

One more thing belongs here, because it is easy to miss: the implant used in the pivotal trial was an earlier version than the one FDA approved. The approved model had not been marketed anywhere in the world before, and FDA bridged the gap using bench and engineering testing rather than new patient data.^[1]

The two main results, with their denominators

Out of every 100 people who had the operation, about 64 met the standard definition of surgical success one year later.

The 82% figure is correct as far as it goes: it counts only the 88 people who finished the study exactly as planned and leaves out the 27 who did not. Whenever you see a Genio success rate quoted in the eighties, that is the number being used. The figure to plan around is the first row.^[1][2]

One more caveat on the oxygen endpoint: the bar for “success” there was a 25% reduction in oxygen-drop events, with no ceiling on what the remaining number could be. That is a lower bar than most people would assume from the word success.

Nyxoah designed, funded, and conducted this trial and collected the data. Independent monitors verified the data; independent statisticians analyzed it.

What the average patient’s sleep study looked like

Among the 89 people who completed a twelve-month sleep study, the average AHI fell from 28.0 to 9.5 events per hour — a mean reduction of 18.3 events. Oxygen-drop events fell from 27.0 to 9.2. Time spent with oxygen below 90% fell from 12.4% to 5.0% of the night.^[1]

In that same group, 82% finished with an AHI under 15 and about 35% with an AHI under 5. Counted against all 115 people who had surgery, the under-5 figure is 27%. And the headline that matters is this: the average AHI at one year was still 9.5. Improved, substantially — not eliminated.

Sleepiness, quality of life, and snoring

Epworth sleepiness scores improved from 9.6 to 6.2. That is a real change, but note the starting point: the group’s average was already below 10, the usual threshold for abnormal daytime sleepiness. Functional outcomes (FOSQ-10) improved from 16.0 to 18.2 out of 20.^[1]

Snoring is where partners notice the difference: the share of bed partners reporting no snoring or soft snoring rose from 16.5% to 69.6%.^[1]

Satisfaction was 89.8% extremely or somewhat satisfied — which also means about one patient in ten was dissatisfied, most often because of residual snoring or discomfort.^[1]

Reported use averaged 6.24 hours a night. That figure deserves an asterisk: it came from patient diaries kept by roughly 70 of the 115 participants, not from the device itself. The independent editorial on the trial noted that the current software can record usage objectively and that it was not used for this report.^[5]

What the trial did not show

• There was no control group, no sham, and no comparison with CPAP, with surgery, or with any other stimulator. It was a single-arm, open-label study.^[2]
• No effectiveness outcomes past twelve months have been published yet. Safety follow-up continues within the five-year trial and long-term results are still to come.^[2]
• Patients with complete concentric collapse were excluded.
• BMI above 32, age over 75, AHI above 65, and predominantly central apnea were not studied.
• The cohort was 70.4% male, 93.9% white, with a mean BMI of 28.5. FDA itself cautions that results by race should be interpreted cautiously.^[1]

The European follow-up numbers, which are worse

I would rather you heard this from me than found it yourself. Alongside the trial, FDA reviewed an ongoing European study called EliSA, in which the manufacturer has been following patients implanted after Genio was approved in Europe.^[14] Of 101 patients implanted, 63 had reached twelve months when the data were cut in December 2024. Their average AHI fell from 36.1 to 24.7, and 38.1% met the responder definition, against 63.5% in the trial.^[1] With sixty-three patients, that 38.1% could reasonably sit anywhere between about 27% and 50%.

Those two figures are not counted the same way, and I want to be straight about which direction that cuts. The 38.1% counts only the 63 patients who had reached a year. The 63.5% counts all 115 trial patients, including everyone who dropped out or was never measured, every one of them scored as a failure. Counted consistently — either both by completers, or both by everyone implanted — the gap widens rather than narrows. What I have shown you is the comparison most favorable to Genio. And what both averages hide is the spread: in the European group, individual results were spread more widely at one year than before surgery, while in the trial they were spread less.^[1] That is not a group who all did moderately well. It is a group in which some did well and some did not.

EliSA is not a look at ordinary practice, whatever it gets called. It is a company-run study with a protocol, named investigators and scheduled visits.^[14] What it is, is a less intensively titrated protocol, and that difference is large. Trial patients had an in-lab sleep study with the device adjusted while they slept at four months, at six months, and again at eight, nine or ten months — and the protocol deliberately gave the patients who were not yet doing well more of these, not fewer, before the twelve-month measurement.^[1] The European patients had one required sleep study. In one report where the same patients were measured both ways, that choice alone moved the measured success rate from 80% to 46%^[13] — a swing larger than the entire gap you are looking at here. That work was done in patients with the unilateral device, so the same caution applies to its registry figures, including the ones quoted elsewhere on this site. I have written that measurement problem up in full, with the Inspire numbers and my own results: what “success” actually means with an implant.

Two more things belong here. Those 63 were the earliest patients implanted, and the programming approach has been revised since — FDA prints that explanation, and the others, without saying whose it is; it neither attributes them to the manufacturer nor endorses them.^[1] And the same FDA document describes a second, much smaller European group — thirteen patients with a collapse pattern Genio is not approved to treat in the United States — in which about two-thirds met the same benchmark.^[1] Thirteen patients cannot outweigh sixty-three, but it sits a page away in the document I am citing and you should know it is there.

So: I do not read 38.1% as the number you should expect, and I will not pretend otherwise. I also cannot tell you it is wrong. What I take from it is narrower and more useful to you — the trial result was produced with a great deal of titration, and how thoroughly your device gets tuned in the year after surgery is part of your outcome, not a formality. Ask about that. No completed study of Genio in ordinary practice exists yet, and the US post-approval study that will answer this properly does not report until 2028.^[15] I do not expect real-world results to mirror this number, but we need more data outside of clinical trials before anyone, including me, can tell you what the success rate will be in a real-world setting.

Does stimulating both sides work better than one?

It is a reasonable idea, and it is the idea Genio was built around. Stimulating the nerve branches on both sides at once should, in principle, pull the tongue forward symmetrically rather than to one side. FDA lists bilateral stimulation among the device’s potential benefits.^[1]

It has not been demonstrated. There has never been a randomized trial comparing bilateral against unilateral stimulation. The one published head-to-head comparison enrolled nineteen patients in total — ten with the unilateral device, reviewed retrospectively, and nine with the bilateral device, followed prospectively. Every comparison in it was statistically non-significant, and on the raw twelve-week numbers the unilateral group was slightly lower on apnea index, oxygen-drop index, and overall AHI. With nine and ten patients per group, that study cannot show a difference and cannot show equivalence. The correct reading of it is we cannot tell.^[6]

A 2026 meta-analysis of hypoglossal nerve stimulation included only three bilateral cohorts out of thirty-nine, and its analysis by laterality did not favor bilateral stimulation — if anything it trended the other way. Its senior author is an employee of the bilateral device’s manufacturer, which makes that result harder to dismiss rather than easier.^[12]

So when you read that stimulating both sides is better, understand what is behind that claim: a sound anatomical rationale and no comparative outcome data. I implant both devices. I do not choose between them on this basis, and I would be sceptical of anyone who does.

For referring physicians: the numbers with their denominators

Metric	Value	Analysis set	Source
AHI responder (Sher)	63.5% / 66.4% / 81.8%	115 implanted / 110 FAS / 88 PP	SSED Tables 36, 34, 35
ODI responder (≥25%)	71.3% / 74.5% / 92.0%	115 / 110 / 88	SSED Tables 36, 34, 35
AHI, baseline → 12 mo	28.0 → 9.5 (−18.3; 95% CI −20.8 to −15.8)	89	SSED Table 37
ODI, baseline → 12 mo	27.0 → 9.2	89	SSED Table 37
T90, baseline → 12 mo	12.4% → 5.0% of night	89	SSED Table 37
ESS / FOSQ-10	9.6 → 6.2 / 16.0 → 18.2	89	SSED Table 37
Device/procedure-related SAE, 12 mo	11.3% (13/115); 7% excluding reoperations FDA did not classify as SAEs	115	SSED Tables 21–25
Non-serious device/procedure AE	73.9% (85/115); 252 events	115	SSED Tables 26, 30
Repositioning / revision / explant, 24 mo	11.3% (13)	115	SSED §XIV.A
Mean operative time	2.37 ± 0.83 h	115	SSED §XII
Screening yield	687 consented → 115 implanted (16.7%)	—	SSED §X.B
Follow-up reached	74 at 24 mo · 30 at 36 mo · 0 at 48 mo	115	SSED Tables 32–33
EliSA European follow-up study (interim)	AHI 36.1 → 24.7; Sher responder 38.1% (95% CI 27–50); AHI−50% responder 41.3%	63 completers of 101 implanted; no imputation — not comparable to the imputed 63.5% above	SSED §XII, Tables 43–44

Sourcing notes: where the SSED’s conclusions section disagrees with its own tables (device-related SAE 6.1% vs 5.2%; dysphagia 11.7% / 17.4% / 18.3%), this page cites the tables. Where the JCSM narrative disagrees with its Table 2 (REM AHI, supine AHI baseline, FOSQ baseline SD), this page omits those variables. The T90 row uses the SSED value because the journal’s row is labeled “minutes” while its own footnote defines it as a percent. The DREAM protocol paper is a plan, not a result; its 65%/64% figures are sample-size assumptions and are not reported here.

Living with Genio

The nightly patch and chip — what to expect

This is the part of Genio that people underestimate, and it is one of the drawbacks, or one of the key differences, you have to consider. Skin irritation from the adhesive patch was the single most common side effect in the trial, affecting 24.3% of patients — nearly one in four.^[1] The manufacturer’s own labeling classes temporary local skin irritation as “very common,” meaning more than one patient in ten.^[11]

That is why it is important that we have you trial the stickers before you commit. Some people react to the adhesive. Some find it uncomfortable to wear all night. And for men with beards, facial hair can stop the patch adhering — in the trial, participants with beards were counselled to trim them so the patch would stick.^[2] Some shaving may be required. None of this is a reason to rule the device out, but all of it is worth finding out before surgery rather than after.

MRI

The implant is MR Conditional at both 1.5 tesla and 3 tesla, and no body region is off limits, provided the scan follows the manufacturer’s conditions.^[11] Practically, this may not matter for many patients — but in certain circumstances, such as active monitoring for prostate cancer or some cardiac conditions, it can matter a great deal, and it is worth raising with your surgeon before you decide.

The conditions are specific: at least eight weeks after surgery; the external patch and chip must never be brought into the scanner room; and there are limits on scan mode and duration. Radiologists should work from the manufacturer’s current MRI guidelines rather than from this page — the labeling documents have been revised, and one detail differs between versions.

Who tends to choose Genio

After all the caveats, it is fair to say who actually picks this device in my clinic. Patients lean toward Genio when they want nothing implanted in the chest and no scheduled battery-replacement operation; when wider MRI access matters to them — active prostate or cardiac surveillance is the common example; or when the single incision is what they care about. Patients lean away when they do not want a nightly patch-and-chip routine, or when the longest track record is the deciding factor. Both are sensible positions, and I have implanted for both kinds of patient.

Dental work, security screening, and other implants

Tell your dentist you have the implant before any procedure. This is not a formality: in the trial, one device eroded through the lining of the floor of the mouth roughly 310 days after surgery, following an unrelated dental implant procedure.^[1] The stimulator sits directly beneath that lining, which is why injections and retractors pressing on the floor of the mouth are best avoided.

The implant contains metal and may trigger security screening, so carry your implant card. And Genio is not compatible with certain other active implanted devices — review your full medical history with your surgeon before proceeding.

Risks and complications — the full picture

Through twelve months, 13 of 115 patients (11.3%) had a serious adverse event related to the device or the procedure. FDA notes that if the one repositioning and four replacement surgeries are excluded — neither the investigators nor the independent events committee classified those as serious — the rate is 7%. I give you both framings because the difference between them is a judgment call, not a fact.^[1]

The specific serious events were: device dislocation in 2 patients, difficulty swallowing in 2, device extrusion in 1, a new left bundle branch block in 1, a nosebleed in 1, a hematoma at the incision in 1, plus the repositioning and four replacement surgeries. Of those four replacements, two succeeded; two failed because of scar tissue and became removals. Three devices in total were removed for migration or extrusion.^[1]

Non-serious side effects were common: 73.9% of patients had at least one. The most frequent were skin irritation from the patch (24.3%), difficulty swallowing (17.4%), swelling at the incision (15.7%), and discomfort from the stimulation itself (14.8%). Most of these settle during recovery.^[1]

What we don’t know yet — and what FDA did about it

When Genio was approved in August 2025, not one patient in the pivotal trial had been followed for four years. Thirty had reached three years; 74 had reached two. Those two- and three-year data are safety data, reviewed by FDA but not yet published as effectiveness outcomes. I cannot tell you what this device looks like at year five, because nobody can yet.^[1]

FDA reached a similar conclusion and acted on it. As a condition of approval it required two long-term studies — five-year follow-up of the original 115 patients, and a new 229-patient study also followed for five years — and it set explicit ceilings on what it will accept: migration no more than 5%, explant no more than 4%, revision no more than 6.6%.^[1] Regulators do not impose numerical ceilings on a device whose long-term behavior they consider settled.

An independent commentary in the Journal of Clinical Sleep Medicine, written by two physicians with no relationship to the manufacturer, called the pivotal trial “a cautious step forward on a promising path.”^[5] I think that is the right register. Not a warning. Not an endorsement. A young device with real results and an unfinished record.

Independent editorial. Neither author reports a relationship with Nyxoah; not industry supported.