Sleep Apnoea, GLP-1s, and Why Women Get Missed

Two separate things are true here

The first is straightforward and well evidenced: these drugs substantially improve obstructive sleep apnoea, and the trial data behind that is better than almost anything else on this site.

The second is the one worth your time: sleep apnoea in women is routinely missed, because it does not present the way the textbook picture says it should. Which means a woman can spend years being treated for something else entirely — and can be prescribed a GLP-1 for weight, unknowingly treat a condition nobody diagnosed, and never find out that was what changed.

What the trials found

SURMOUNT-OSA was not one study but two, reported together in the New England Journal of Medicine: both 52 weeks, both double-blind and placebo-controlled, both in adults with moderate-to-severe obstructive sleep apnoea and obesity, randomised 1:1 to the maximum tolerated dose of tirzepatide (10 mg or 15 mg) or placebo¹.

The split between them is the clever part. Trial 1 enrolled people not using positive airway pressure. Trial 2 enrolled people already on PAP therapy. So the question was not only "does this work" but "does it add anything for someone already treated".

The primary endpoint was the apnoea–hypopnoea index — the AHI, meaning the number of times per hour of sleep that breathing stops or becomes shallow.

Both differences were significant at p<0.001¹.

Read the baselines first. A mean AHI around 50 means breathing was interrupted roughly fifty times an hour, every hour, all night. Roughly halving that is not a marginal improvement in a lab number — that is the difference between sleep that restores someone and sleep that does not.

Improvements also reached significance across every prespecified key secondary endpoint: hypoxic burden, high-sensitivity CRP, systolic blood pressure, body weight, and patient-reported sleep disturbance and impairment¹. A later report in Nature Medicine examined the cardiometabolic outcomes in more detail. Those measures were prespecified; alongside them the authors ran post hoc analyses, including a mediation analysis asking how much of the benefit was attributable to weight loss as against the reduction in apnoea events and hypoxic burden themselves² — a question that matters, because it separates "this helped because you lost weight" from "this helped your breathing".

Adverse events were mostly gastrointestinal and mostly mild to moderate — the familiar profile.

Why this section is the important one

Here is the part almost nothing else covers.

Obstructive sleep apnoea has historically been treated as a male condition, and the diagnostic picture most clinicians carry reflects that: loud snoring, witnessed pauses in breathing, daytime sleepiness, a partner who reports it.

Women frequently do not present that way. A review of OSA in women in primary care describes them presenting instead with insomnia, depressive symptoms, fatigue, morning headache and nightmares — non-specific symptoms that lead directly to underdiagnosis and undertreatment relative to men³.

Look at that list again and consider what a busy mother in her forties gets told those symptoms are. Stress. Perimenopause. Depression. Not sleeping well because of the children. Being tired because everyone is tired.

A narrative review of sex and gender differences in sleep-disordered breathing reaches the same conclusion and adds a mechanical reason: the screening tools themselves lack sensitivity in women, because they were built around the typical male presentation⁴. So the instrument used to decide whether to refer someone for a sleep study is less likely to flag a woman who has the condition.

Menopause matters too. The male-female gap in prevalence narrows after menopause, which implicates menopausal status itself rather than sex alone — and points at sex hormones, fat distribution and upper-airway collapsibility as contributors⁵. So the years when a woman is most likely to be told her exhaustion is hormonal are also the years her actual risk of sleep apnoea is rising.

What this means practically

• If you are exhausted in a way that sleep does not fix, ask specifically about sleep apnoea. Not "am I tired", but that condition by name. The screening questionnaires may not surface it for you.
• Morning headache, insomnia and nightmares are on the list. If you have been assuming apnoea means snoring, those are the symptoms most likely to be attributed to something else.
• You do not need a partner's report to be referred. Many women have no one observing them sleep, and the absence of that report is not evidence of absence.
• If you already use PAP, this trial still applies to you. Trial 2 was specifically people on PAP, and it showed the larger absolute reduction of the two.
• If you start a GLP-1 and your sleep changes markedly, that is worth saying out loud to whoever manages your care. It may be information about a condition nobody has assessed.

Where this sits on the evidence scale

The treatment claim: high, and among the strongest on this site. Two independent 52-week randomised double-blind trials, a hard physiological primary endpoint measured in a sleep laboratory rather than reported by patients, consistent direction and magnitude across both, and concordant movement in secondary measures including blood pressure and inflammatory markers.

The women-are-missed claim: moderate but consistent. It rests on reviews and primary-care series rather than randomised evidence, because you cannot randomise someone's diagnostic pathway. Several independent reviews describe the same pattern, and the mechanism — screening instruments calibrated on men — is concrete rather than speculative.

Two things would sharpen this. SURMOUNT-OSA reported results for its trial populations, not broken out by sex, so the size of the benefit specifically in women remains an inference from the overall result. And nobody has tested whether a sex-specific screening instrument would close the referral gap, which is the study that would actually change practice.