The REM apnea-hypopnea index is the number of apneas and hypopneas per hour of sleep counted only during rapid eye movement (REM) stages, rather than across the entire night. A standard sleep study produces one overall AHI that averages events from all sleep stages, but that single number can hide a pattern in which breathing is relatively normal during lighter sleep and collapses repeatedly during REM. The REM AHI isolates that pattern, and a growing body of research suggests it carries independent health risks that the overall score may understate.
How It Differs From the Overall AHI
During a standard overnight sleep study, technicians tally every pause in breathing (apnea) and every episode of shallow breathing (hypopnea) across all sleep stages, then divide by total sleep time. That gives the overall AHI. If someone sleeps seven hours and has 35 events, the overall AHI is 5 events per hour, which sits right at the conventional cutoff for mild obstructive sleep apnea.
The REM AHI uses the same counting method but restricts the denominator to just the time spent in REM sleep. Because REM accounts for roughly 20 to 25 percent of a normal night, the math can look very different. A person might have almost no events during non-REM stages and dozens during REM, yet still register a modest overall AHI that does not trip the alarm. The REM AHI captures that concentration of events and gives clinicians a much sharper picture of what is happening during the most physiologically vulnerable phase of sleep.
Why Breathing Gets Worse During REM
REM sleep comes with a near-complete loss of skeletal muscle tone, a feature called REM atonia. The muscles that hold the upper airway open are not spared. Animal research has shown that the drop in upper airway muscle activity during REM is driven primarily by the withdrawal of excitatory signals from serotonin and other chemical messengers, rather than by active inhibition clamping the muscles shut.1PubMed. Control of Upper Airway Motoneurons During REM Sleep In practical terms, the tongue and soft palate become floppy, and the airway is far more likely to narrow or close entirely.
The consequences go beyond just more frequent events. Apneas during REM tend to last longer and produce deeper drops in blood oxygen than events during non-REM sleep. One study found this pattern held across every patient examined, with more pronounced oxygen desaturation per episode in REM compared to non-REM stages.2PubMed. Apnea duration and hypoxemia during REM sleep in patients with obstructive sleep apnea A comprehensive review confirmed the finding broadly: respiratory events are usually longer and more desaturating in REM than in non-REM sleep.3European Respiratory Review. REM sleep obstructive sleep apnoea This means a REM AHI of 30 is not physiologically equivalent to an overall AHI of 30. Each event counted in the REM number tends to hit the body harder.
Who Gets REM-Predominant Sleep Apnea
REM-predominant obstructive sleep apnea, where breathing events cluster heavily or exclusively in REM, is not evenly distributed across the population. Women are far more likely to have this pattern than men. One study comparing men and women with OSA found that women had a significantly lower non-REM AHI than men but a nearly identical REM AHI, meaning their disease was much more concentrated in REM. The difference was stark: REM-predominant OSA occurred in about 62 percent of women with OSA versus 24 percent of men.4PubMed. Gender differences in the polysomnographic features of obstructive sleep apnea
This sex difference has real diagnostic consequences. Because the overall AHI averages together REM and non-REM stages, and because women often breathe normally during non-REM sleep, their overall AHI may come back low enough to be considered normal or borderline. The disease is there, concentrated in the hours of REM, but the standard scoring method dilutes it. REM-predominant OSA generally shows up in people with mild-to-moderate overall scores, which is exactly the range where a clinician might decide treatment is optional.3European Respiratory Review. REM sleep obstructive sleep apnoea
The Hypertension Connection
The most studied health risk tied specifically to the REM AHI is high blood pressure. Data from the Wisconsin Sleep Cohort, a long-running population-based study, showed a dose-response relationship between a rising REM AHI and the likelihood of having hypertension. The association was clearest when the REM AHI reached 15 or higher. In people whose non-REM AHI was very low (5 or below), each doubling of the REM AHI was associated with about 24 percent higher odds of hypertension. The longitudinal arm of the same study also found that REM AHI predicted the future development of hypertension, while non-REM AHI did not.5PubMed Central. Obstructive Sleep Apnea during REM Sleep and Hypertension. Results of the Wisconsin Sleep Cohort
The mechanism connecting sleep apnea to blood pressure involves surges in sympathetic nervous system activity. Each apnea triggers a burst of adrenaline-like signaling, a spike in heart rate, and a sharp rise in blood pressure as the body fights to reopen the airway. Over time, these repeated surges alter the balance between the body’s “fight or flight” and “rest and digest” nervous systems, amplifying blood pressure swings not just at night but into the following day.6PubMed Central. Sleep Apnea, Autonomic Disturbances, and Blood Pressure Variability Because REM events tend to be longer and produce worse oxygen drops, each event may deliver a more potent sympathetic punch.
The picture is not perfectly simple, however. A study of a Chinese clinical cohort found that when overall AHI was accounted for, patients meeting the traditional definition of REM-predominant OSA actually had lower blood pressure and milder disease overall. The researchers concluded that total AHI appeared more influential on hypertension than the predominance of REM events alone, suggesting that REM concentration of events matters most when non-REM breathing is relatively normal.7PubMed. Rapid eye movement (REM)-related obstructive sleep apnea and hypertension: insights from the clinical spectrum of apnea-hypopnea index ratios across REM and non-REM sleep stages in a Chinese cohort In other words, a high REM AHI seems to pose the most independent risk when it exists in someone whose overall sleep study otherwise looks reassuring.
Beyond Blood Pressure
Cardiovascular risk from REM OSA extends past hypertension. Analysis of the Sleep Heart Health Study found that severe REM OSA, defined as a REM AHI of 30 or higher, was associated with a higher incidence of composite cardiovascular events. The association was especially strong in people who already had cardiovascular disease, where the adjusted hazard ratio reached about 2.5 compared to those with a REM AHI under 5.8PubMed Central. Obstructive Sleep Apnea during REM Sleep and Cardiovascular Disease This suggests that REM-specific apnea may act as an accelerant in people already carrying cardiovascular risk, rather than initiating disease from scratch in otherwise healthy individuals.
Metabolic effects have also been linked to REM-stage breathing disturbances. In patients with OSA who had not yet developed diabetes, the apnea index during REM sleep was independently associated with insulin resistance. Intermittent oxygen drops and disrupted REM sleep emerged as the main contributors to worsening insulin sensitivity in that group.9PubMed. Characteristics of obstructive sleep apnea related to insulin resistance The connection makes biological sense: REM sleep is a metabolically active brain state, and repeated interruptions may interfere with hormonal regulation that normally happens during those hours.
Memory, Mood, and Cognitive Effects
REM sleep is when much of the brain’s memory consolidation work happens, particularly for emotional and procedural memories. Disrupting REM with repeated apneas appears to carry cognitive costs. A review of the clinical literature links disrupted REM sleep to impaired memory consolidation, reduced attention, and mood disturbances.10Current Sleep Medicine Reports. REM-Obstructive Sleep Apnea: Clinical Characteristics and Cardiovascular, Cognitive, Neurobehavioral, and Metabolic Implications
Research into Alzheimer’s disease risk has zeroed in on REM-specific breathing disturbances as a potential contributor. A study of older adults found that higher REM AHI and greater oxygen desaturation during REM were associated with worse performance on verbal learning and memory tests. The relationship was strongest in adults over 60 and in carriers of the APOE4 gene variant, which is the most well-established genetic risk factor for Alzheimer’s.11PubMed Central. Older adults at greater risk for Alzheimer’s disease show stronger associations between sleep apnea severity in REM sleep and verbal memory This does not prove that REM apnea causes Alzheimer’s, but it raises the possibility that treating REM-specific breathing problems could be especially relevant in people already at elevated dementia risk.
Why There Is No Agreed-Upon Definition
One of the most frustrating aspects of REM-predominant OSA, for clinicians and patients alike, is that there is no standardized definition. Different studies use different cutoffs to decide when someone “has” REM OSA. Some require the REM AHI to be at least twice the non-REM AHI. Others require the overall AHI to be under a certain threshold while the REM AHI exceeds it. Still others use a ratio-based approach. This lack of consensus makes it difficult to compare studies, estimate how common the condition is, or write clear treatment guidelines.12PubMed. Rapid eye movement related obstructive sleep apnea: Where do we stand?
The uncertainty extends to treatment decisions. When someone has a REM AHI of 25 but an overall AHI of 4, do they have sleep apnea that warrants treatment? There is no consensus.13Sleep. 1018 Finally Awake: Treatment of Severe REM-Isolated Obstructive Sleep Apnea Insurance companies typically use the overall AHI to determine coverage for CPAP devices, which means a person whose breathing events are packed into REM may not qualify for treatment even though their REM-specific severity would warrant it. The person experiences all the consequences of severe apnea during a critical sleep phase, but on paper they look mild.
Treatment Considerations
When REM-predominant OSA is treated, the therapy is generally the same as for other forms of OSA: continuous positive airway pressure, or CPAP. But there are some practical wrinkles. Research shows that the effective pressure needed to keep the airway open is modestly higher during REM than during non-REM sleep.14PubMed Central. Effective Continuous Positive Airway Pressure Changes Related to Sleep Stage and Body Position in Obstructive Sleep Apnea during Upward and Downward Titration: An Experimental Study Auto-adjusting CPAP machines, which ramp pressure up and down throughout the night in response to detected airway obstruction, may handle this automatically. Fixed-pressure machines set too low could leave REM events undertreated while adequately addressing non-REM events.
REM sleep tends to occur in longer bouts toward the end of the night, which creates a timing problem. If someone removes their CPAP mask partway through the night, as many users do, they may be sleeping unprotected during exactly the hours when their breathing is worst. This is one reason that overall CPAP adherence matters more for people with REM-predominant disease: partial use may leave the most dangerous sleep phase uncovered.
Positional therapy is another option worth mentioning. People are more likely to sleep on their backs during REM, and the supine position worsens airway collapse. Some patients find that sleeping on their side reduces events enough to make a meaningful difference, particularly during REM. However, positional changes alone are unlikely to resolve severe REM-stage apnea, and the data on combining positional strategies with CPAP for this specific subtype remain limited.
REM-Predominant OSA in Children
REM-predominant OSA is not an adults-only phenomenon. Children can show the same clustering of breathing events in REM sleep, and it appears to come with its own set of consequences. Data from the Childhood Adenotonsillectomy Trial found that children with REM-predominant OSA had increased daytime sleepiness, higher overall AHI, and lower oxygen saturation nadirs compared to children whose OSA was more evenly distributed across sleep stages.15PubMed Central. Risk Factors, Symptom Burden, and Treatment Response in Pediatric REM-Predominant OSA: Evidence From the Childhood Adenotonsillectomy Trial In pediatric sleep medicine, where the threshold for diagnosing OSA is much lower (an AHI of just 1 to 2 events per hour can be considered abnormal), recognizing a REM-specific pattern may change whether a child is referred for adenotonsillectomy or monitored conservatively.
The daytime sleepiness finding is worth noting because parents and pediatricians sometimes assume that a child’s overall AHI is too low to explain their symptoms. A child who is exhausted and struggling in school despite an overall AHI of 3 might make much more sense once you see that their REM AHI is 15 and their non-REM AHI is close to zero. The overall number is technically below most surgical thresholds, but the child is losing most of the restorative benefit of REM sleep every night.
What a Sleep Study Report Tells You
If you have had a polysomnography (overnight sleep study in a lab), your report likely includes the overall AHI. Whether it also breaks out REM and non-REM AHI separately depends on the sleep lab and the interpreting physician. Many commercial reports now include stage-specific data, but it is not universal. Home sleep tests, which are increasingly common, generally cannot distinguish REM from non-REM sleep because they lack the EEG sensors needed to identify sleep stages. This means REM-predominant OSA is essentially invisible on a home study.
If your overall AHI is low but you still feel unrefreshed, excessively sleepy, or have trouble with memory and mood, it may be worth asking whether your breathing events were concentrated in REM. A lab-based study is needed for that determination. Likewise, if you have been told your sleep apnea is “mild” based on the overall AHI but you are being treated for resistant hypertension or are at elevated risk for Alzheimer’s, the REM AHI could add relevant information to your clinical picture.
Sleep medicine is gradually moving toward a more granular approach that considers not just how many events happen per hour on average, but when they happen, how long they last, and how much oxygen drops with each one. The REM AHI is part of that shift. It is not yet embedded in formal diagnostic criteria or insurance reimbursement algorithms, which means clinicians who want to act on it often have to make their case individually. But the evidence supporting its relevance, particularly for blood pressure, cardiovascular risk, metabolic health, and cognition, has been accumulating steadily enough that ignoring it feels increasingly hard to justify.