Tirzepatide and sleep apnoea: the SURMOUNT-OSA results

Obstructive sleep apnoea (OSA) is the airway collapsing repeatedly during sleep. Each collapse cuts airflow, drops oxygen and wakes the brain briefly, usually without the person knowing. It is linked to daytime sleepiness, high blood pressure, heart rhythm problems and stroke. Excess weight is one of the strongest risk factors, because fat around the neck and tongue narrows the airway.

Severity is measured with the apnoea-hypopnoea index (AHI), the number of complete or partial breathing stops per hour of sleep, recorded in a sleep study. By convention, 5 to 15 an hour is mild, 15 to 30 moderate and above 30 severe. The usual treatment is positive airway pressure (PAP), a mask that splints the airway open. It works when it is worn. Many people do not wear it.

The trials at a glance

Who was studied

SURMOUNT-OSA was two phase 3, double-blind, placebo-controlled trials run side by side, 469 adults in total, all with moderate to severe OSA and obesity, and none with type 2 diabetes. Trial 1 took people who were not using PAP at the start. Trial 2 took people who were using PAP and kept using it. Participants were randomised one to one to the highest dose of tirzepatide they could tolerate, 10 mg or 15 mg weekly, or placebo, for 52 weeks.

These were people with severe disease and high weight. Average AHI at baseline was 51.5 events an hour in trial 1 and 49.5 in trial 2. Average BMI was 39.1 and 38.7.

What changed

In trial 1, AHI fell by an average of 25.3 events an hour on tirzepatide against 5.3 on placebo. In trial 2, it fell by 29.3 against 5.5. The estimated treatment differences were 20.0 and 23.8 events an hour, both highly significant.

To put it on the scale: a person starting at around 50 events an hour who lost the average 25 to 29 would end near the boundary between moderate and severe, or in the moderate band. Some went much further. The FDA noted that more people on Zepbound reached remission or mild OSA without symptoms than on placebo. Others moved less. An average describes the group, not you.

Every key secondary end point also favoured tirzepatide: percent change in AHI, body weight, hypoxic burden (a measure of how much oxygen is lost overnight), patient-reported sleep impairment and disturbance, hsCRP (an inflammation marker) and systolic blood pressure. Side effects were mostly gastrointestinal and mostly mild to moderate, the same pattern as the weight trials. The SURMOUNT trials page has the weight data in people without OSA.

Why weight is not the whole story

The trials reported more than the AHI because the AHI misses things. Two people with the same count of events can lose very different amounts of oxygen, depending on how long each event lasts. Hypoxic burden captures that, and it improved on tirzepatide. So did hsCRP and systolic blood pressure, two markers that connect sleep apnoea to heart disease. Patient-reported sleep impairment improved as well, which is the change most people notice first. None of these were tested as stand-alone primary end points, but all were multiplicity-controlled secondary end points, which is a stronger standard than an exploratory analysis.

The FDA approval, December 2024

On 20 December 2024 the FDA approved Zepbound for moderate to severe obstructive sleep apnoea in adults with obesity, alongside a reduced-calorie diet and more physical activity. The agency called it the first drug treatment option for certain patients with OSA. The indication is narrow on purpose: obesity is part of it, because the trials only studied people with obesity, and the evidence that the benefit comes largely through weight loss is strong.

What it does not mean

• It is not a replacement for PAP by default. Trial 2 tested tirzepatide added to PAP, not instead of it. Whether someone can stop PAP after losing weight is a decision for a repeat sleep study and the sleep physician, not the scales.
• It does not cover OSA without obesity. Sleep apnoea in people of normal weight, caused by jaw shape or airway anatomy, was not studied.
• Semaglutide does not have this indication. There is no equivalent dedicated OSA trial of semaglutide, and its label does not include OSA.
• It probably needs continued treatment. The trials ran 52 weeks on drug. Weight tends to return after stopping, and the weight regain page has the numbers; there is no reason to expect the airway to behave differently.

Things worth knowing before starting

Tell the sleep clinic as well as the prescriber. If you use PAP, auto-titrating machines record residual AHI each night, which gives a free trend line as the weight changes. If the mask starts to fit differently as the face slims, the clinic can refit it. And because the GLP-1 drugs slow stomach emptying, any sedation for a sleep procedure or surgery needs the team to know the date of your last dose. The side effects page covers the anaesthesia warning, and the tirzepatide titration page covers how doses step up to 10 or 15 mg.

The OSA log that helps a review is the dose step, weekly weight, the nightly AHI from the PAP machine if you have one, and a one-line note on daytime sleepiness. The free GLP-1 tracker keeps the dose, the weight, side effects and, on Premium, lab results on one timeline, so the numbers your prescriber asks about are in one place rather than spread across portals and notes. It is free to download.