Does Inspire Actually Work? Reading the Trial Data Properly
The implant has real evidence behind it and a 63% response rate at five years. It also has entry criteria so specific that most people with sleep apnea were never eligible for the trial.
Inspire is the implanted device that stimulates the hypoglossal nerve during sleep, pushing the tongue forward on each breath. If you have failed CPAP and someone has raised it, the question you actually want answered is whether it works.
It does, for the people it was tested on. The whole story is in that second clause, and it is the part the marketing does not dwell on. What the device is and how it is implanted is covered separately — this is about the evidence.
The pivotal trial
The STAR trial was published in the New England Journal of Medicine in 2014 and is the study the approval rests on. It enrolled 126 people with moderate-to-severe obstructive sleep apnea who could not accept or adhere to CPAP.
At twelve months, median AHI fell 68%, from 29.3 events per hour. Oxygen desaturation index improved alongside it.
That is a large effect and it is real. Two things about the design matter for how much weight to put on it.
It was a single-group cohort, not a randomised trial. Everyone got the implant; there was no control arm receiving a sham or an alternative. The only randomisation was a sub-study in which people who had already responded were randomly assigned to have their device switched off — which tests whether the benefit depends on the stimulation, and it does, but it tells you nothing about how many people respond in the first place.
The population was narrow. 83% men, mean age 54.5, and mean BMI 28.4 — which is overweight, not obese, in a condition where obesity is the dominant risk factor.
The five-year follow-up, and its arithmetic
Five-year outcomes were published in 2018, and this is where the numbers need care.
Of the original 126, 97 completed the protocol and 71 consented to a voluntary sleep study at five years.
Among those 71, the response rate — defined as AHI under 20 and at least a 50% reduction — was 75%. Apply last-observation-carried-forward across the fuller cohort and it becomes 63% at five years.
Sleepiness and quality of life moved substantially: Epworth normalisation rose from 33% to 78%, quality-of-life normalisation from 15% to 67%. Serious device-related adverse events, all involving lead or device adjustment, occurred in 6%.
Three caveats worth holding onto:
The 75% is measured in volunteers. The five-year polysomnogram was optional and 71 of 126 took it. People for whom a device is working are more likely to return for a sleep study to prove it. The 63% figure is the more conservative read, and it is the one to quote.
"Response" does not mean cured. The threshold is an AHI under 20. Someone entering at 35 and finishing at 18 is a responder and still has moderate sleep apnea.
The entry criteria were restrictive, and this is the single most important thing on this page.
Who the trial excluded
To enrol in the five-year cohort you needed CPAP failure, moderate-to-severe OSA, a BMI under 32, and no unfavourable collapse pattern on drug-induced sleep endoscopy.
That last one does real work. Concentric collapse at the palate predicts non-response, so those people were screened out before implantation. The trial therefore reports outcomes in patients pre-selected for a favourable airway — which is appropriate trial design and also means the response rate does not transfer to whoever walks into a clinic. What sleep endoscopy is and what it decides.
Put the criteria together and the tested population is: a not-very-obese person, usually male, with a specific collapse pattern, who has genuinely tried CPAP. If you have a BMI of 38, this evidence is not about you. That is not a reason the device cannot help — it is a reason nobody can quote you a success rate from it.
What has been added since
Later work has extended the picture rather than overturned it. Registry and cohort analyses have examined how sleep endoscopy findings relate to outcomes, and how post-operative weight change affects success — weight gain after implantation works against the therapy, as you would expect from the BMI ceiling in the trials.
There is also interest in whether stimulation improves cardiovascular outcomes rather than only the AHI, which is being examined but is not settled. Treat cardiovascular benefit as plausible and unproven.
Evidence tier 1 for AHI reduction in the selected population. Tier 2 for anyone outside those criteria, where the mechanism applies but the numbers do not.
The honest summary
If you have moderate-to-severe apnea, have genuinely failed CPAP, have a BMI in the twenties or low thirties, and your sleep endoscopy shows a favourable collapse pattern, this is a treatment with five-year data and a roughly two-in-three chance of meaningful response. For an implanted device in a field with few options, that is a strong position.
If you fall outside those criteria — and most people with sleep apnea do — then the correct thing to be told is that the evidence does not cover you, not that it is 75% effective.
And it is major surgery for a condition with non-surgical options, most of which should be exhausted first. Most CPAP failure is mask fit, pressure and mouth leak, and is fixable. Fix CPAP before replacing it, and the full range of alternatives.
If you have not been diagnosed and are reading this because you snore, none of it applies yet. Witnessed breathing pauses, gasping arousals, unrefreshing sleep or daytime sleepiness that more sleep does not fix are the signs that warrant testing. Snoring versus apnea.