A woman in her late forties comes in tired. Not dramatic-tired — functional, foggy, the kind that gets filed under stress or perimenopause or "you have three kids." A sleep study gets ordered, mostly to rule things out. The report comes back with an apnea-hypopnea index of 11. Mild. She's told it's borderline, handed a pamphlet about a CPAP machine, and sent home with the quiet implication that her exhaustion is not quite a medical problem.
That number — and the way obstructive sleep apnea gets sorted into mild, moderate, and severe — was largely calibrated on men. This is not a conspiracy. It's a sampling problem that hardened into a definition, and it has shaped who gets diagnosed, who gets treated, and which treatments anyone bothered to design.
Where the "typical" apnea patient came from
The mental image most clinicians carry — heavyset man, loud snorer, witnessed pauses in breathing, falls asleep at red lights — is not invented. It's the picture that emerged from the foundational epidemiology of the field.
The most-cited origin point is the Wisconsin Sleep Cohort. Young et al. (1993), in the New England Journal of Medicine, recruited employed adults aged 30 to 60 and reported that roughly 24 percent of men and 9 percent of women had an AHI of 5 or more. The headline that traveled was the ratio: apnea looked like a roughly two-to-one or worse male disease. That ratio did real work. It shaped clinical suspicion, referral patterns, and the questions on intake forms — Do you snore? Has anyone seen you stop breathing? — which describe the male presentation more faithfully than the female one.
Here's the part that's thinner than the belief it produced. Later analyses suggested the clinic-referred male-to-female ratio (which ran as high as eight to one for years) reflected referral bias, not true prevalence. Women with the same physiology weren't getting sent for studies, because they didn't match the picture the early data had drawn. The picture was real. It was also partial. And a partial picture, repeated often enough, becomes the definition of the whole.
Is sleep apnea different in women?
Yes — in how it presents, where the events cluster during the night, and how it scores on the standard severity scale. Women with obstructive sleep apnea more often report fatigue, insomnia, morning headache, and low mood rather than the classic loud snoring and observed gasping. Their breathing events are more likely to be concentrated in REM sleep and in specific body positions, and more likely to be hypopneas (partial airway narrowing) than full apneas (complete collapse). Because the AHI averages events across the whole night, a woman whose apnea is severe during REM can post a low overall number and be labeled mild.
That gap between the lived problem and the scored number is the central injustice here, and it's mechanical, not vague.
What actually happens in the airway
Walk it through in the order it occurs. As sleep deepens and then shifts into REM, the body deliberately paralyzes most skeletal muscle — a safety feature that keeps you from acting out dreams. Some of that atonia reaches the muscles that hold the upper airway open, including the genioglossus, the tongue's main protractor. The airway gets floppier.
In a woman, the collapse often doesn't go all the way to a full stop. The airway narrows. Airflow drops, oxygen dips modestly, the brain registers the strain and produces a brief arousal — not always a full waking, just a shallowing of sleep — to restore muscle tone and reopen the passage. Then it happens again. And again, clustered in REM, which itself clusters in the back half of the night.
The cost isn't necessarily a dramatic oxygen plummet. It's the fragmentation. Sleep that never consolidates, an autonomic nervous system that keeps getting pinged, a person who slept eight hours and feels like she fought through four. The AHI, by averaging, can quietly launder all of this into a reassuring single digit.
Why the standard fix doesn't always fit
CPAP — continuous positive airway pressure — is genuinely the most effective treatment we have for moderate-to-severe obstructive sleep apnea. It splints the airway open with a steady stream of air. For the right patient it is close to curative on a nightly basis. None of what follows is an argument against it.
But CPAP was optimized for the patient the field knew best. Adherence is the field's oldest open wound: a meaningful fraction of patients — estimates commonly land around a third to a half over the long term — don't use the machine enough to benefit. The barriers women report skew toward discomfort, pressure intolerance, claustrophobia, and mask fit problems on smaller facial structures. When apnea is milder on paper, the calculus gets worse: harder to justify a difficult therapy for a number that reads "borderline," even when the daytime cost is real.
This is where the gentler, non-invasive end of the toolkit matters more than it usually gets credit for.
Positional therapy. Because female apnea is frequently positional — worse when supine, better on the side — keeping a person off her back can substantially cut events. The evidence is decent for the right candidate: a study by van Maanen and de Vries (2014) in SLEEP tested a vibrating neck-worn positional device in 31 patients and found a meaningful drop in supine sleep time and AHI. Small sample, short follow-up, but mechanistically clean.
Mandibular advancement devices. A custom oral appliance pulls the lower jaw slightly forward, enlarging the airway behind the tongue. For mild-to-moderate disease, well-fitted devices reduce AHI and, importantly, tend to win on adherence — people wear them. They rarely normalize severe apnea, and they require dental follow-up, but the comfort tradeoff often favors them in exactly the population most likely to abandon a mask.
Lower-pressure and auto-titrating PAP. Newer auto-adjusting machines deliver only as much pressure as the moment requires, which for milder, REM-predominant apnea can mean meaningfully lower average pressures — and lower pressure is the single most cited reason people tolerate the therapy at all.
I want to be honest about the gradient of evidence. CPAP's efficacy is well-established. Oral appliances for mild-to-moderate disease are well-supported, though less powerful. Positional therapy is plausible and mechanistically sound but tested in small studies. The idea that any of these should be preferentially selected by sex is, right now, reasoned extrapolation more than proven protocol. The biology of the disparity is solid. The treatment-matching is where the data thins.
An honest rule of thumb
If you're a woman with a "mild" result and real daytime symptoms, the number alone is not a reason to do nothing. Tonight, two things you can act on without a prescription: note whether your worst nights track with sleeping on your back, and bring the position-specific and REM-specific breakdown of your sleep study to your next appointment — not just the headline AHI.
| Therapy | Best fit | Evidence strength | Main barrier |
|---|---|---|---|
| CPAP | Moderate–severe | Strong | Comfort, adherence |
| Auto-PAP (lower pressure) | Mild–moderate | Good | Cost, fit |
| Oral appliance | Mild–moderate | Good | Dental follow-up |
| Positional therapy | Positional/supine-dominant | Moderate, small studies | Habit, comfort |
Ask specifically: What was my AHI during REM, and during supine sleep, separately? If those numbers are much higher than the average, the average was hiding something.
What this piece can't tell you
It can't tell you the true prevalence of obstructive sleep apnea in women, because we are still correcting for decades of who got studied — and the threshold of AHI 5 was never validated as the point where female symptoms begin. It can't tell you whether a sex-specific severity scale would change outcomes, because the trials that would settle that haven't been run at scale. And it can't promise that matching therapy to presentation beats the current one-size approach, only that the mechanism makes the question worth asking.
Where to look next: the literature on REM-predominant OSA, the ongoing debate over whether hypopnea scoring criteria undercount female disease, and any sleep study report that bothers to separate REM and supine events. The disease was named from a partial sample. The fix starts with reading your own data more closely than the summary line allows.
