The question
Tirzepatide has evidence for reducing obstructive sleep apnea severity in adults with obesity. The useful question is what “reducing” means: fewer breathing events during an hour of sleep is a measurable outcome, but it does not automatically mean every participant was cured or could stop using a breathing device.
The US FDA approved the finished medicine Zepbound for moderate-to-severe obstructive sleep apnea in adults with obesity in December 2024, alongside a reduced-calorie diet and increased physical activity. That is a specific indication, not an approval of every tirzepatide-containing product or a treatment claim for all causes of poor sleep. FDA approval announcement.
Two trials, two different starting situations
SURMOUNT-OSA comprised two randomized, double-blind, placebo-controlled trials lasting 52 weeks. One enrolled people not receiving positive airway pressure, or PAP, at baseline; the other enrolled people already using it. Participants had obesity and moderate-to-severe obstructive sleep apnea. Eli Lilly funded the program. Original trial publication.
Those two starting situations matter when reading a headline. A result among people who were not using PAP is not a direct comparison between medication and PAP. Nor does a medication-versus-placebo comparison among PAP users establish that the device is unnecessary. In each trial, the randomized comparison answers a narrower question about the effect of assigning tirzepatide within that study's design.
The official NCT05412004 trial record provides the protocol context and identifies the study as completed. “Completed” describes trial status; it does not mean every possible question about long-term treatment has been answered.
Read the unit before the percentage
The primary outcome was change in the apnea-hypopnea index, abbreviated AHI: the number of apneas and hypopneas per hour of sleep. In trial 1, the reported average change was −25.3 events per hour with tirzepatide and −5.3 with placebo. The estimated between-group difference was −20.0, with a 95% confidence interval from −25.8 to −14.2. Trial 2 reported changes of −29.3 and −5.5, respectively, for a difference of −23.8 events per hour. Study results.
A simple reading example makes the arithmetic clearer. In trial 1, subtracting the placebo change from the medication change gives −25.3 − (−5.3) = −20.0. The larger within-group reduction describes what happened in the medication group. The between-group difference accounts for the improvement also seen in the comparison group. They answer different questions, and both should keep their unit: events per hour.
An average change also leaves individual variation hidden. Imagine two hypothetical participants whose AHI each falls by 20 events per hour: one starts at 25, the other at 60. The same numerical reduction leaves them in very different places. These invented numbers illustrate the interpretation; they are not patient stories or reconstructed trial outcomes.
Why the PAP detail changes the conclusion
The FDA label explains that participants in the PAP trial suspended PAP for seven days before the primary endpoint assessment. It also explicitly says the trials did not evaluate the timing or appropriateness of discontinuing PAP in previously compliant users. Zepbound prescribing information, section 14.2.
That temporary study procedure is easy to misread. It allowed investigators to assess breathing severity under a specified testing condition. It was not a clinical instruction for readers to interrupt treatment, and it did not create a general rule for when to stop a device. Anyone considering a change to established sleep-apnea care needs an individualized reassessment with their treating clinician.
For an article or advertisement claiming a medication “replaces CPAP,” ask where the replacement decision itself was studied. A lower AHI and a safe decision to withdraw another therapy are separate propositions. The label's limitation is particularly useful because it identifies that missing step directly.
Benefits still need a safety and population context
Gastrointestinal adverse events were the most commonly reported events in the original publication and were generally mild or moderate. The approved product also has contraindications and warnings, including a boxed warning about thyroid C-cell tumors seen in rodents; the relevance to humans is unknown. These details belong alongside efficacy results, not below an assumed claim that a peptide is inherently gentle. Trial safety findings and prescribing information.
The trial is informative for its defined population and follow-up. It does not establish a treatment for ordinary snoring, insomnia, central sleep apnea or an unspecified sleep complaint. Those are broader claims than the approval described above. For context on another comparison that is frequently oversimplified, see our tirzepatide–semaglutide trial explainer.
A careful reading of SURMOUNT-OSA therefore ends with three separate pieces of information: the participants' starting situation, the measured change in breathing events, and the limits on translating that result into a care decision. Keeping those pieces visible makes the finding more useful than a blanket promise of better sleep.
Sources & further reading
Research context last reviewed October 5, 2026. Last updated October 5, 2026. Educational reading, not individual medical advice. How this article was prepared.
