The SURMOUNT-OSA results were clear but narrow. After 52 weeks, adults with obesity and moderate-to-severe obstructive sleep apnea had about 20 to 24 fewer breathing interruptions per hour with tirzepatide than with placebo. The study did not show that tirzepatide replaces a breathing machine, works for every kind of sleep apnea, or produces the same result in every person.
What question did SURMOUNT-OSA ask?
SURMOUNT-OSA was really two late-stage trials run under one registered study plan. Both asked whether tirzepatide changed the apnea-hypopnea index (AHI, the number of times breathing stops or becomes shallow in an hour of sleep) more than placebo.
Trial 1 studied people who were unable or unwilling to use positive airway pressure (PAP, a machine that keeps the airway open). Trial 2 studied people who had used PAP for at least three months and planned to continue it. Everyone was randomly assigned, and neither participants nor investigators knew who received tirzepatide or placebo.
The study doses were the maximum each participant could tolerate, either 10 mg or 15 mg once weekly. That is a report of the protocol, not a dosing recommendation.
Who was enrolled, and who was not?
The 2024 New England Journal of Medicine paper reported 234 adults in trial 1 and 235 in trial 2. Mean age was about 48 and 52 years, respectively. Roughly seven in ten participants were men.
Everyone had obesity and an AHI of at least 15, the threshold for moderate sleep apnea. The average starting AHI was about 50 events an hour in both trials, which is severe. People with type 1 or type 2 diabetes, central or mixed sleep apnea, recent large weight changes, or planned sleep-apnea or obesity surgery were excluded.
That matters. These results do not answer the broader question of whether tirzepatide helps mild sleep apnea, sleep apnea without obesity, or central sleep apnea. Our separate guide to tirzepatide for sleep apnea explains who the branded indication covers.
What was measured?
The primary endpoint, meaning the result chosen in advance as the main test, was the change in AHI at week 52. An overnight sleep study counted the breathing events.
The researchers also measured body weight, blood pressure and hypoxic burden, which captures how deep and how long oxygen drops during breathing events. Another target combined an AHI below 5 with an AHI of 5 to 14 and an Epworth Sleepiness Scale score of 10 or less. That scale is a short questionnaire about daytime sleepiness.
SURMOUNT-OSA results at 52 weeks
The headline result was the same in both settings: AHI fell more with tirzepatide than with placebo.
| Result | Trial 1: no PAP | Trial 2: continuing PAP |
|---|---|---|
| Participants randomized | 234 | 235 |
| AHI change, tirzepatide | −25.3 events/hour | −29.3 events/hour |
| AHI change, placebo | −5.3 events/hour | −5.5 events/hour |
| Difference versus placebo | −20.0 events/hour | −23.8 events/hour |
| AHI cut by at least half, tirzepatide vs placebo | 61.2% vs 19.0% | 72.4% vs 23.3% |
| AHI below 5, or 5–14 without marked sleepiness | 42.2% vs 15.9% | 50.2% vs 14.3% |
| Body-weight change, tirzepatide vs placebo | −17.7% vs −1.6% | −19.6% vs −2.3% |
The placebo-adjusted AHI differences had 95% confidence intervals of −25.8 to −14.2 in trial 1 and −29.6 to −17.9 in trial 2. A confidence interval is the range the analysis says is reasonably compatible with the data. Neither interval came close to zero.
These are group averages, not a forecast. The breathing result also arrived alongside substantial weight change, so the trial cannot tell us how much came from weight reduction and how much, if any, came from another effect of the drug.
What the side-effect table showed
Stomach and bowel symptoms were the most common problems. Most were mild to moderate and clustered while doses were being increased.
| Adverse event | Trial 1: tirzepatide vs placebo | Trial 2: tirzepatide vs placebo |
|---|---|---|
| Diarrhea | 26.3% vs 12.5% | 21.8% vs 8.8% |
| Nausea | 25.4% vs 10.0% | 21.8% vs 5.3% |
| Vomiting | 17.5% vs 4.2% | 9.2% vs 0.9% |
| Constipation | 15.8% vs 2.5% | 15.1% vs 4.4% |
| Stopped treatment because of an adverse event | 4.4% vs 1.7% | 3.4% vs 7.0% |
Serious adverse events occurred in 7.9% versus 5.8% in trial 1 and 5.9% versus 10.5% in trial 2. Two confirmed pancreatitis cases occurred in the trial 2 tirzepatide group. There were no deaths. The wider tirzepatide side-effects guide covers label warnings beyond this one study.
The journal later corrected two cells in the safety table: severe low-blood-sugar events in both placebo groups were zero, not one. The correction did not change the AHI findings.
What newer analyses add, and what they cannot fix
As of September 18, 2026, the day this article was written, two new post-hoc analyses had examined the same trial data. Post-hoc means the question was asked after the main study was planned, so it can add context but is weaker than a new randomized trial.
A July 7, 2026 analysis in Sleep found that participants who began with worse fatigue, sleepiness or sleep quality often reported larger gains on matching questionnaires. Objective changes such as AHI and weight were generally similar across those symptom groups. An August 31, 2026 analysis in the Journal of Clinical Sleep Medicine reported AHI improvement across subgroups defined by age, sex, starting BMI, starting AHI and neck size. Both analyses were exploratory, and several authors were Lilly employees.
The original trials were also sponsored, designed and analyzed with Lilly involvement. They lasted one year, did not measure heart attacks or survival, and did not compare tirzepatide directly with another medicine. Trial 2 was not a test of replacing PAP: participants continued it, then paused it for seven days before the main sleep-study assessment.
Semaglutide is a reasonable alternative to discuss because it is another prescribed incretin medicine, but SURMOUNT-OSA did not test it. This study cannot say which drug changes AHI more.
What changed after the trial
On December 20, 2024, the agency added an obstructive-sleep-apnea indication for Zepbound. The current Zepbound label, revised 8/2026, says the brand is indicated "to treat moderate to severe obstructive sleep apnea (OSA) in adults with obesity." The FDA said the improvement was likely related to weight reduction in its approval announcement.
Through Promise, tirzepatide is dispensed as a compounded medication, which is different from an FDA-approved product: the formulation offered here is not FDA-approved. SURMOUNT-OSA tested the branded study product under a fixed protocol; it did not establish that a compounded preparation is equivalent or carries the branded indication. Compounded tirzepatide follows a different regulatory path.
A licensed provider may still prescribe a compounded formulation; that decision is between you and your doctor. At Promise, a licensed provider reviews every request and prescribes only when medically appropriate. Not everyone qualifies.
