Sleep apnea in women — why it is missed so often

Sleep apnea in women is easy to miss because the daytime story often sounds like insomnia or fatigue, the events cluster in REM, and the AHI depends on which hypopnea rule the lab used. A 2025 scoping review in Sleep Science and Practice found that scoring hypopneas with a 3 percent desaturation or an arousal, rather than a 4 percent desaturation alone, increased OSA diagnoses in women by a range of 12.8 to 63 percent across the studies it included. Medicare still pays from the 4 percent rule. That mismatch is a coverage problem sitting on top of a biology problem.

Why is the diagnosis missed so often?

The brochure of OSA was written around men who snore loudly, stop breathing, and fall asleep in chairs. Women more often report insomnia, fatigue, and mood change, with less theatrical snoring. That pattern is in the scoping review’s background and in the clinical literature it cites. It is also why a woman gets a sleep-hygiene handout or an antidepressant trial while her partner, with the same AHI, gets a machine.

Referral habits add a second filter. Studies the review cites have found that women with sleep-disordered-breathing symptoms are less likely than men to be diagnosed and treated. Stigma around snoring does not help. None of that means women “get less apnea.” It means the pathway into a lab was built for a different presentation.

If your complaint is unrefreshing sleep and an insomnia label, and someone in the house hears irregular breathing, you still need a respiratory recording. Treating only the insomnia, when the nights are full of hypopneas, leaves both problems half-addressed. COMISA (insomnia plus OSA) is a recognized overlap. It is not a reason to skip the breathing test.

What does REM-predominant disease look like on a study?

Obstructive events often worsen in REM because muscle tone falls. In many women the overnight AHI looks mild because most of the night is NREM, while the REM AHI is moderate or severe. The scoping review notes that REM-predominant OSA is more common in women, with a lower overall AHI and shorter events, whereas men more often have similar severity across NREM and REM. Won and colleagues, included in that review, found that sex differences in AHI under 4 percent scoring were driven largely by higher NREM AHI in men, while REM AHIs were similar.

A single-number AHI averages those stages. A woman with REM AHI 35 and NREM AHI 4 can print as “mild.” She can still have fragmented REM, morning headaches, and a partner who only notices the problem toward morning. Ask for the stage-specific indices, not only the summary line.

Home tests make this worse. They do not score REM with EEG. They count events per hour of recording. A short REM period at 5 a.m. can be diluted across a seven-hour recording, or missed if the device came off. Kapur 2017 already warns that HSAT underestimates AHI relative to PSG and cannot catch arousal-only hypopneas. Women whose events are milder desaturations plus arousals in REM are the patients that warning was written for.

What is 3 percent versus 4 percent hypopnea scoring?

The AASM Scoring Manual’s recommended adult hypopnea rule (often called 1A or AHI3A) is a 30 percent drop in airflow lasting at least 10 seconds, plus either a 3 percent oxygen desaturation or an EEG arousal. The acceptable or optional rule (1B or AHI4) requires a 4 percent desaturation and does not count arousal-only hypopneas. CMS coverage for PAP still uses the 4 percent definition. Many commercial payers accept the recommended 3 percent/arousal rule. The same night of sleep can therefore be “no OSA” on a Medicare worksheet and “moderate OSA” on an AASM-accredited report.

Chou, Khan, and Singh’s 2025 scoping review pulled eight studies that compared those rules with sex-specific results. Seven were retrospective; one was cross-sectional. All used polysomnography (one also included home tests). Across that set, moving from 4 percent to 3 percent criteria increased OSA diagnosis rates in women by 12.8 percent to 63 percent, and it increased detection of moderate-to-severe disease in several cohorts. The spread is wide because the studies enrolled different populations. The direction is not.

Examples from that review, restated rather than copied: Campos-Rodriguez found 12.8 percent of a women-only cohort fell below AHI 5 on 4 percent scoring but not on 3 percent scoring. Haile and colleagues, in a 2019 clinic sample, reported that using AHI3A increased OSA diagnosis by 30.4 percent in women and 21.7 percent in men. Khalid reported that 48.9 percent of women in their cohort would have gone undiagnosed on 4 percent criteria versus 28.5 percent of men. Won reported a larger relative AHI increase in women (83 percent, mean 10.4 to 19.1) than in men (64 percent, 17.6 to 28.9) when 3 percent scoring was applied. Hirotsu’s population sample showed a larger rise in mild OSA among women than among men when the 3 percent rule was used.

A 2025 commentary on Haile (PMC11874093) makes the coverage point sharply: restrictive AHI4 rules used by payers worsen under-diagnosis in females, who more often have lower AHIs, shorter events, and less desaturation. That commentary is an opinion piece on Haile’s data, not a second scoping review. The numbers to remember are still Haile’s and Chou’s.

If the extra diagnoses are “mild,” do they matter?

Sometimes. The scoping review is cautious: data that treating mild OSA in women helps are limited, and more outcome research is needed. The MERGE trial, which it cites, is the randomized study that showed CPAP can improve quality of life in mild OSA scored with 3 percent criteria; a later analysis reported particular benefit in women with mild disease. That is not a mandate to put every AHI 6 on a machine. It is a reason not to discard those events as scoring trivia.

Johnson and colleagues, in the review’s table, looked at people already prescribed PAP for mild sleep-disordered breathing under 4 percent AHI under 15, including some with upper-airway resistance physiology. Under 3 percent scoring, 63 percent of the women who had a positive 3 percent AHI and a PAP prescription would have been called “no OSA” by 4 percent rules. PAP adherence in the below-threshold group was similar to those with AHI of 5 or more. Adherence is not a randomized outcome. It is a clue that those patients were using the device.

Medicare patients are stuck in the 4 percent lane for coverage. If your 3 percent AHI is 18 and your 4 percent AHI is 4, a Part B claim for PAP may fail even if the AASM-accredited lab diagnosed OSA. That is a payer rule, not a statement that the arousals are imaginary. Ask which definition is on your report. If only one AHI is printed, ask the lab to score both.

Should women skip the home test?

Not automatically. Kapur still allows HSAT in uncomplicated adults at increased risk of moderate to severe OSA. Many women are not that phenotype: lower pre-test probability on the classic Kapur triad, more insomnia, more REM-only disease, more arousal-predominant hypopneas. The scoping review notes that women are more likely to have a nondiagnostic home test even when OSA is suspected, and that some insurers force a home test first. A negative home test is not a clearance. Kapur’s strong recommendation is to do PSG when HSAT is negative, inconclusive, or technically inadequate.

If you have insomnia, take opioids, have heart failure, or had a stroke, the home test was already the wrong tool for other reasons. If you are a woman with fatigue, a normal BMI, and a partner who hears gasping only near morning, start the conversation at sleep testing with an in-lab night on the table, not as a prize you earn after a failed home study.

Do pregnancy and menopause change this?

They change risk and presentation. The scoping review flags pregnancy and menopause as periods when OSA risk and symptoms shift with hormones, weight, and sleep fragmentation. Menopause is a time snoring and apnea become more prevalent; perimenopausal insomnia can hide the respiratory part. Pregnancy OSA belongs in obstetric risk discussions (hypertension, gestational diabetes) and is not something to diagnose from a wearable. Those are reasons to have a lower threshold to test, not reasons to invent a female-specific AHI cutoff. ICSD still uses AHI of 5 or more, with symptoms or comorbidity, as the diagnostic frame.

We will not quote a years-in-practice figure or a “women are always underdiagnosed by X percent” statistic that is not in a paper we opened. The range Chou reported is already wide enough to make the point.

If this sounds like your nights, call Westerville Sleep Center at (614) 898-9340. Bring prior reports and ask us which hypopnea rule they used. The sleep apnea page covers general diagnosis. This page is the sex-specific scoring and phenotype problem sitting underneath it.