The sleep apnea approval that quietly rewrote the coverage math

There is a sentence in American health policy that has cost more people more money than almost any other: drugs used for weight loss are excluded.

It sits in the Medicare statute. It sits, in various phrasings, in a very large number of commercial plan documents. And for two decades it did not matter much, because the drugs it excluded did not work well enough for anyone to fight about.

Then the FDA approved a GLP-1 for something that is not weight loss.

What the trial actually found

SURMOUNT-OSA enrolled adults with obesity and moderate-to-severe obstructive sleep apnea, in two arms — people using CPAP and people who were not. Participants received tirzepatide or placebo for a year.

The apnea-hypopnea index, which counts breathing interruptions per hour of sleep, fell substantially more on tirzepatide than on placebo in both arms. A meaningful proportion of participants improved enough to no longer meet the criteria for moderate-to-severe disease.

The mechanism is not exotic. Fat deposition around the upper airway and in the neck contributes to airway collapse during sleep; reduce it and the airway behaves better. What is notable is not that it worked but that anybody ran the trial — the commercial logic of proving a weight-loss drug does something other than weight loss is a direct response to how American coverage is written.

Why the label change matters more than the effect size

Medicare Part D cannot cover “agents when used for weight loss.” That language dates to 2003, and it has survived every attempt to remove it.

It does not, however, say that Part D cannot cover a drug that happens to cause weight loss when that drug is prescribed for a different approved indication. This is the same door that semaglutide walked through after SELECT, which established cardiovascular benefit in people with overweight or obesity and established cardiovascular disease.

The result is a system in which the same molecule, at the same dose, in the same body, is covered or excluded depending on which sentence your physician writes.

What this means if you have sleep apnea

If you have obesity and diagnosed moderate-to-severe OSA, and your plan has refused a GLP-1 as a weight-loss drug, the OSA indication is a genuinely different application — not a loophole, an approved use.

What payers will usually want:

1. A documented sleep study with a recorded apnea-hypopnea index in the moderate-to-severe range. A prior diagnosis from years ago may need repeating.
2. A documented BMI meeting the obesity criterion.
3. The correct diagnosis code on the claim, which sounds trivial and is the single most common reason these are denied.
4. Evidence of CPAP trial or intolerance, on some plans, though the approval does not require it.

The part that should bother you

I have a patient — a 58-year-old bus driver, badly under-slept for a decade — who had been denied twice for obesity. He was tested for sleep apnea because his wife recorded him on her phone and made him take it to a doctor. His AHI came back severe.

The medication that his insurer would not pay for in March was covered in September. Same man, same weight, same drug. What changed was that somebody finally ordered the right test.

That is not a story about a clever appeal. It is a story about a system where clinical need and coverage eligibility are two different things, and where the gap between them is closed by luck and persistence rather than by policy.

Persistence is easier when it is documented. For an apnea claim specifically, what helps is the study result, the dated weights, the doses, and — if you use CPAP — some record of how the two have interacted over time. He had none of that and won anyway; most people are not that fortunate. Keeping it in Zenday costs nothing and turns the next renewal into paperwork rather than another year of luck.

What to watch next

The strategic logic here is now obvious to both manufacturers, and the trial pipelines reflect it: cardiovascular outcomes, kidney disease, heart failure with preserved ejection fraction, liver disease, knee osteoarthritis. Each new indication is a clinical finding and, simultaneously, a coverage key cut for a different lock.

There is something faintly absurd about a system that requires a company to prove a drug treats five separate conditions in order to pay for the one everybody knows it treats. But while that is the system, the practical advice follows from it: find out what else you have.

The wider picture of what happens to sleep on this drug class — including the weeks when it briefly gets worse — is in sleep on a GLP-1.