A modeling study in The Lancet Respiratory Medicine projects that by 2050, obstructive sleep apnea will affect nearly 77 million US adults aged 30 to 69 — about 46% of that age group, up roughly 35% from 2020. Buried inside the same model is a smaller number that got a lot less attention: even with widespread GLP-1 use factored in, the researchers project it would reduce that 2050 total by less than 5%.
If your mental model is “GLP-1s are fixing obesity, and obesity causes sleep apnea, so this problem is already being solved,” the study’s own numbers say otherwise. Here’s why, and what the bigger issue actually is.
What the model projects
Researchers, led by ResMed’s Elroy Boers with UC San Diego’s Atul Malhotra as senior author, built an open-cohort dynamic population model — a simulation that ages a population forward year by year, tracking shifts in age, sex, and BMI — to estimate OSA prevalence from 2020 through 2050. It’s described as the first study of its kind to project US OSA prevalence over that horizon.
The headline numbers:
- Overall: 76.6 million adults aged 30–69, up from roughly 57 million in 2020 — a 35% relative increase.
- Women: projected to rise from 22.8% to 37.7% prevalence, a 65% relative increase, reaching 30.4 million cases.
- Men: projected to rise from 45.6% to 54.4% prevalence, a more modest 19% relative increase, reaching 45.9 million cases.
The driving forces are an aging population and rising BMI trends — plus, the authors note, a female population that has historically been underdiagnosed and under-recognized relative to actual risk, which pulls the sex gap closer together over time as detection improves.
Why GLP-1s barely register in the model
Obesity is a major, well-established OSA risk factor — excess tissue around the airway makes it more likely to collapse during sleep. So it’s a reasonable guess that a class of drugs producing double-digit average weight loss would meaningfully bend this curve down.
The model says otherwise, and the reason is worth sitting with: obesity is only one of several forces driving the projection, and it’s not even the fastest-moving one. Age is. Airway anatomy changes with age independent of weight, and the US population is simply getting older across this window. BMI trends matter too, but a population-level model has to weigh a drug’s plausible uptake rate, adherence over decades, and effect size against demographic momentum that doesn’t wait for anyone’s prescription to fill. Modeled against that backdrop, even optimistic GLP-1 uptake barely dents the total.
That’s a model’s estimate, not a fact carved in stone — model assumptions about future drug uptake and long-term adherence are exactly the kind of input that can turn out wrong in either direction over 25 years. One columnist covering the study speculated the real number could land lower than projected, as more people on GLP-1 therapy find their sleep improves without a device. That’s a plausible guess, not a finding — nobody has run the tracking study that would confirm it, and it cuts against what the model’s own authors built into their numbers.
Worth naming plainly: ResMed manufactures CPAP machines and other sleep-apnea treatment devices, and led this research through its own academic-industry collaborative. That doesn’t make the modeling wrong — the underlying demographic and BMI trend data are independently verifiable — but a company that sells more devices when more people are diagnosed has an interest in a larger projected patient population, and that’s worth knowing when you read the headline number.
The bigger problem the projection doesn’t fix
Here’s what matters more than any 2050 estimate: separate research puts current OSA prevalence at roughly 83.7 million American adults today, and an estimated 80–90% of those cases are undiagnosed.
That gap exists for ordinary, human reasons. Sleep apnea’s core symptom — repeatedly stopping breathing — happens while you’re unconscious. Most people only find out because a partner mentions it, if they have a partner sleeping next to them at all. The downstream symptoms — fatigue, brain fog, high blood pressure — get chalked up to age, stress, or “just how everyone feels” instead of investigated, especially in older or higher-BMI patients, which is backwards, since those are the groups at highest risk. Add limited access to sleep specialists, insurance friction around testing, and general unfamiliarity with who’s actually susceptible, and you get a condition that’s both extremely common and mostly invisible to the people who have it.
That gap is the same size whether the 2050 total lands at 77 million or something else. Rising future prevalence is a public-health planning problem. An 80-90% current diagnosis gap is a today problem, for anyone reading this who has never been tested.
Why this is the piece we keep coming back to
We don’t have a way to make GLP-1 drugs single-handedly solve population-level sleep apnea trends, and neither does anyone else’s model. What we can do is close the individual half of that gap: a 3–5 night at-home study using an FDA-cleared ring, read by a board-certified sleep clinician, turns “I probably don’t have this” into an actual apnea-hypopnea index number for your own body. If it’s clinically appropriate, treatment gets coordinated with a licensed provider. If it’s not, you’ve ruled it out instead of wondering.
A projection about 2050 can’t tell you anything about your own airway tonight. Only a measurement can.
This post is general information, not medical advice. It is not a claim that any medication or program prevents, treats, or cures sleep apnea, and it does not describe a guaranteed outcome of any NoBsRx program. Compounded medications are not FDA-approved. Treatment is never guaranteed — eligibility, diagnosis, and every treatment decision are made by independent licensed providers based on an individual clinical assessment, and availability varies by state. If you think you may have sleep apnea, talk to a licensed clinician.