The 30-second summary
- Tirzepatide is the first medicine licensed for moderate to severe obstructive sleep apnoea in adults with obesity, approved in the US in December 2024.
- The effect is large. In SURMOUNT-OSA the apnoea-hypopnea index fell by 20 to 24 events an hour more than placebo, and between 42% and 50% of people reached remission.
- It is still not as good as CPAP, at roughly 22 events an hour against about 31. What it has over CPAP is that people actually keep doing it.
- The mechanism is mostly fat, including fat inside the tongue. Which means the improvement travels with the weight, in both directions.
Why obesity and apnoea are so tangled
Obstructive sleep apnoea affects close to a billion adults worldwide, and it is not a snoring problem. It is repeated collapse of the upper airway during sleep, each episode ending in a partial waking that you will not remember. It carries independent associations with cardiovascular disease, stroke and higher mortality.
Obesity is present in 60% to 70% of people with the condition and is the main modifiable risk factor. The reason is mechanical. Fat sits in the soft tissue beside the throat and, less obviously, inside the tongue itself. Awake, muscle tone holds the airway open. Asleep, that tone drops, and a narrower tube is easier to close.
CPAP, a mask that splints the airway open with air pressure, is the standard treatment and it works. The difficulty is that between 30% and 60% of people stick with it, which caps how much good it does across a population however well it performs in a clinic.
What SURMOUNT-OSA found
Two phase 3 randomised trials, reported in the New England Journal of Medicine in 2024, enrolled adults with moderate to severe obstructive sleep apnoea and obesity. Trial 1 enrolled people not on positive airway pressure; trial 2 enrolled people who were already using it. Everyone received the maximum tolerated dose of tirzepatide, 10 or 15 mg, or placebo, for 52 weeks.
At the start, the average apnoea-hypopnea index was 51.5 events an hour in trial 1 and 49.5 in trial 2. Severe, by any definition.
| Tirzepatide | Placebo | Difference | |
|---|---|---|---|
| Trial 1, no PAP | -25.3 events/hour | -5.3 | -20.0 |
| Trial 2, on PAP | -29.3 events/hour | -5.5 | -23.8 |
Both results were highly significant. Every prespecified key secondary endpoint improved as well: body weight, hypoxic burden, C-reactive protein, systolic blood pressure, and patient-reported sleep. A 2026 review in JAMA Otolaryngology put the remission rate at 42% to 50%, meaning an index below 5 events an hour, or below 15 without symptoms.
The most frequent adverse events were gastrointestinal and mostly mild to moderate, which is the familiar pattern.
How it compares with CPAP, honestly
The same review sets the two side by side: roughly 22 events an hour of improvement with a GLP-1, against about 31 with CPAP. A mask is still the more effective treatment for the airway.
Two things sit against that. The first is adherence, where CPAP's real-world performance is far below its clinical performance and a weekly injection's is not. The second is that a GLP-1 treats a cause and CPAP treats the consequence: the mask holds the airway open every night for as long as you wear it, while weight loss narrows the problem itself.
The review is careful about what is not yet known. Cardiovascular outcome benefits from treating apnoea this way remain unproven. And weight regain commonly occurs after stopping, which matters more here than in most contexts, because if the airway improved because the tongue got smaller, the airway will change back when it does not stay that way.
Six meta-analyses have found apnoea-hypopnea reductions with GLP-1s ranging from 5.7 to 21.9 events an hour, a wide spread that reflects how different the drugs, doses and populations have been.
What this means for how you eat
Two things, and they pull in the same direction.
The weight is doing the work, so protect the part of it that should not go. Airway fat falls with total fat, and there is no way to aim it. What you can influence is the composition of what comes off. Weight lost with too little protein and no resistance work takes more muscle with it, and muscle is not what is narrowing your airway. Protein floor first, load twice a week.
Treating the apnoea makes the eating easier. Short, fragmented sleep raises appetite hormones and pushes choices towards fast carbohydrate, which is a headwind you have been eating into for years without seeing it. People often describe the first month of properly treated apnoea as the first time in a long while that a morning did not start in deficit.
There is a practical overlap too. Eating late and lying down soon after is worse for reflux and worse for apnoea, and reflux is already more likely on a medicine that slows the stomach. Finishing your last real meal two to three hours before bed helps all three problems at once.
Questions for your clinic
Ask whether a repeat sleep study is planned, and when. Improvement you can feel is not the same as improvement that has been measured, and any change to CPAP should follow the second, not the first.
Ask what the plan is if you stop the medicine, or if supply is interrupted. That question is more important here than in most conversations about these drugs.
And say if you are on a GLP-1 already. Your sleep clinic and your prescriber are often not the same person, and neither of them can see the other's notes.
What to do with this
Whatever the airway is doing, the thing you control is what the weight loss is made of. The calculator sets a protein floor from your weight and your dose, and builds three days of food around it.
Sources
- Malhotra A, et al. Tirzepatide for the Treatment of Obstructive Sleep Apnea and Obesity. N Engl J Med 2024;391:1193-1205. PubMed
- Harris A, et al. Glucagon-Like Peptide-1 Receptor Agonists for Obstructive Sleep Apnea: A Review. JAMA Otolaryngol Head Neck Surg 2026. PubMed
- Sievenpiper JL, et al. Nutritional and lifestyle supportive care recommendations for management of obesity with GLP-1 based therapies. Obes Pillars 2026. PubMed
Medical disclaimer: Articles in the foodose research library are educational, not medical advice. Follow the instructions from your prescriber and the leaflet in your pack. See our full medical disclaimer.