REM Sleep Apnea: Causes, Symptoms, and Treatment Options

Obstructive sleep apnea that clusters heavily in REM sleep is a recognized clinical pattern in which breathing interruptions occur predominantly, or exclusively, during the rapid-eye-movement stage of sleep while the rest of the night stays relatively quiet. Despite being discussed in the sleep-medicine literature for decades, REM-related obstructive sleep apnea (often shortened to REM-OSA) still lacks a single agreed-upon definition, which makes it both underdiagnosed and undertreated. The condition has its own set of causes rooted in REM-specific physiology, can carry health consequences that rival more severe forms of general sleep apnea, and poses unique challenges when it comes to treatment adherence and diagnostic accuracy.

What Counts as REM-Related Sleep Apnea

There is no universally accepted cutoff that separates REM-related sleep apnea from ordinary obstructive sleep apnea. Researchers have used a variety of definitions over the years, but the most common one requires a normal or only mildly elevated breathing-disruption score during non-REM sleep combined with a markedly elevated score during REM sleep, along with a REM-to-non-REM ratio greater than two.¹ In practical terms, a person with this pattern might record an overall apnea-hypopnea index (a count of breathing pauses per hour) that looks mild or even normal, because the relatively calm non-REM hours dilute the severe REM hours. That averaging effect is a core reason REM-OSA slips under the radar.

The lack of a standardized definition also complicates research. Some studies set the non-REM threshold at fewer than five events per hour, others at fewer than fifteen. Some require only that the REM score be higher than the non-REM score; others demand a ratio of two or more. When you compare studies that use different thresholds, the prevalence estimates and demographic profiles shift, which is one reason experts have debated for years whether REM-OSA is truly a distinct condition or just a normal feature of mild sleep apnea.

Why Breathing Gets Worse During REM

REM sleep brings a natural loss of muscle tone across much of the body. The muscles that keep the upper airway open are not exempt. Research measuring airway compliance in healthy sleepers found that the upper airway became significantly more collapsible during both tonic and phasic REM sleep compared with non-REM sleep, where the airway was actually more compliant than during wakefulness.² In someone whose airway is already somewhat narrow due to anatomy, weight, or both, this REM-related loss of muscle support can tip a borderline airway into outright collapse.

Recent work has refined that picture further. Patients with REM-predominant sleep apnea appear especially prone to passive airway collapsibility, as opposed to the ventilatory control instability that drives non-REM-predominant patterns. In other words, the problem in REM-OSA is less about the brain’s breathing signals misfiring and more about the physical airway folding shut when the muscles relax.³ Evidence suggests that the withdrawal of the brain’s drive to breathe during REM, rather than any special reduction in individual muscle activity, largely explains why the pharynx collapses more readily in this sleep stage.³

The Low-Score, High-Risk Paradox

One of the most counterintuitive features of REM-related sleep apnea is that a mild-looking overall score can coexist with oxygen drops just as concerning as those seen in more severe disease. Apneas during REM sleep tend to last longer, and the oxygen desaturations that follow are deeper, compared with apneas during non-REM sleep.⁴ A recent study examining different sleep-apnea phenotypes confirmed this pattern: the REM-related group had the lowest overall apnea-hypopnea index (around six events per hour on average) yet carried a comorbidity burden indistinguishable from a group whose average index was above thirty, with comparable mean overnight oxygen saturation across the phenotypes.⁵

The REM-related group actually spent more time with their oxygen saturation below 90 percent than both the REM-aggravated and stage-independent groups, even after adjusting for other factors.⁵ This is the paradox that makes REM-OSA clinically important: the number that clinicians and insurers rely on to gauge severity can look reassuringly low while the oxygen stress that actually harms your body is anything but low.

Symptoms and Who Gets Them

Daytime sleepiness is the symptom most people associate with sleep apnea, and it does occur in REM-related OSA, but the picture is more nuanced. Because REM-stage breathing events can fragment the sleep stages most involved in dreaming and emotional regulation, people with REM-OSA sometimes describe poor sleep quality that feels out of proportion to their mild-sounding diagnosis. Research in a large clinical population found that depressive symptoms and higher body mass index were the strongest independent predictors of both sleepiness and reduced quality of life in the REM-related group, more so than the breathing-disruption numbers themselves.⁶

Snoring, gasping awakenings, and morning headaches are common across all forms of obstructive sleep apnea. What sometimes tips clinicians off to a REM-dominant pattern is a patient who says the snoring and choking happen mainly in the second half of the night, since REM periods grow longer as the night progresses. Bed partners may notice long stretches of quiet breathing early on, followed by loud, irregular episodes closer to morning.

Cardiovascular and Metabolic Consequences

The intermittent drops in oxygen caused by sleep apnea trigger surges in sympathetic nervous system activity, the body’s “fight or flight” wiring. In patients with moderate-to-severe obstructive sleep apnea, these breathing-event-related blood pressure surges correlate directly with sympathetic overactivity, contributing to a pattern of nocturnal hypertension that raises long-term cardiovascular risk.⁷ Because REM-stage apneas produce deeper oxygen drops, the sympathetic spikes during REM can be especially pronounced.

On the metabolic side, obstructive sleep apnea has been independently linked to insulin resistance and abnormal lipid metabolism.⁸ Evidence specifically implicating REM-stage events comes from a study that found the apnea index during REM sleep was independently associated with insulin resistance after adjusting for confounders, alongside the oxygen desaturation index and time spent below 90 percent saturation.⁹ The fact that REM-specific apnea events showed their own independent relationship with metabolic dysfunction strengthens the case that REM-OSA is more than an academic curiosity.

Links to Cognitive Health

REM sleep plays a well-established role in memory consolidation, emotional processing, and learning. When breathing disruptions fragment REM, those functions can suffer. Research on cognitively normal older adults found that those with untreated obstructive sleep apnea showed progressive deterioration in spatial navigational memory across morning testing sessions, while people without OSA improved over the same sessions.¹⁰ That pattern hints at impaired overnight memory consolidation even in people who do not yet show cognitive decline.

Looking further downstream, a study measuring blood biomarkers found that patients with obstructive sleep apnea had elevated serum levels of amyloid-beta and tau proteins, both associated with neurodegenerative disease. REM sleep duration was negatively correlated with total amyloid-beta levels, suggesting that the more REM is disrupted, the higher these markers climb.¹¹ The researchers noted that obstructive sleep apnea is a modifiable risk factor and that treating it may slow or reverse some of these neurodegenerative changes. For someone whose apnea is concentrated in REM, the cognitive stakes may be proportionally higher because the very stage being disrupted is the one most critical for brain maintenance.

Treatment With CPAP and the Adherence Problem

Continuous positive airway pressure remains the first-line treatment for obstructive sleep apnea regardless of which sleep stages are affected. The pressurized air splints the airway open and eliminates both REM and non-REM events. The challenge with REM-OSA is that the person’s overall numbers look mild, which can make the nightly inconvenience of wearing a mask feel disproportionate. One small study found that REM-related OSA was the single strongest predictor of dropping out of CPAP therapy, with none of the patients classified as having good adherence meeting the REM-related criteria.¹² However, a larger cohort study found no significant association between REM predominance and CPAP adherence, suggesting the relationship is not settled and may depend on how adherence is defined and how patients are supported.¹³

An interesting wrinkle is the phenomenon of REM rebound. When a person with OSA begins CPAP, the newly unobstructed sleep can unleash a surge of REM sleep that the brain had been missing. Patients who experienced this REM rebound on their first night of CPAP showed significantly greater early compliance with the device compared with those who did not rebound, possibly because the flood of REM-rich sleep felt subjectively restorative.¹⁴ Patients who noticed a subjective improvement in sleep quality on their first CPAP night showed both more deep sleep and more REM sleep.¹⁵ For REM-OSA patients in particular, that first-night experience could be either a powerful motivator or, if the rebound is extreme, a confusing event that needs clinical framing.

Alternatives to CPAP

For people who cannot tolerate CPAP, several alternatives deserve consideration, and some have particular relevance for REM-predominant patterns.

Positional therapy: REM-related OSA has been associated with the supine sleeping position.¹³ A Cochrane review of positional therapy for sleep apnea found that while CPAP produced a greater reduction in breathing events, patients using positional devices (such as bumper belts or electronic trainers that discourage back sleeping) used their device about two and a half more hours per night than CPAP users, indicating better real-world adherence.¹⁶ For someone whose REM events are already concentrated in the supine position, positional therapy could address a meaningful share of the problem.
Oral appliances: Mandibular advancement devices push the lower jaw forward to widen the airway. A case report documented complete resolution of REM-predominant OSA with a mandibular advancement device in a patient whose condition had also been causing abnormal sleep behaviors.¹⁷ Case reports are thin evidence on their own, but oral appliances are already well-established for mild-to-moderate OSA in general, and the REM-predominant pattern often falls into that severity range.
Hypoglossal nerve stimulation: This surgically implanted device stimulates the nerve controlling the tongue to keep the airway open during sleep. Early data looking specifically at REM-exacerbated OSA found no significant difference in response rates between REM-dominant and non-REM-dominant patients, suggesting the device works comparably across phenotypes.¹⁸

Weight loss also warrants mention, given that higher body mass is a consistent predictor of both REM-related OSA and the severity of symptoms within the REM-related group. No single intervention works for everyone, and many sleep physicians end up combining approaches, such as positional therapy plus an oral appliance, to maximize coverage across sleep stages and body positions.

Why Home Sleep Tests Can Miss It

Most at-home sleep apnea testing kits do not include the electroencephalography channels needed to tell REM from non-REM sleep. Without sleep staging, the device calculates a breathing-event index using total recording time rather than actual sleep time, which tends to dilute the score because it counts time spent lying awake as breathing-normal time.¹⁹ For someone with REM-OSA, this limitation is doubly problematic: the events are already concentrated in a small fraction of sleep time, and the home test further spreads them across the entire recording window. The result is a number that can look normal when it is not.

If you have daytime sleepiness, snoring, or fatigue that seems out of proportion to a “normal” home sleep test result, an in-laboratory polysomnography with full sleep staging can reveal what the home test missed. This is especially worth pursuing if you are taking antidepressants, which suppress REM sleep and can further mask a REM-predominant breathing pattern on any type of study.

How Antidepressants Complicate the Picture

Many commonly prescribed antidepressants, particularly SSRIs and SNRIs, reduce the total amount of REM sleep. For someone whose breathing problems live almost entirely in REM, less REM means fewer recorded events and a lower overall score. Clinicians have flagged this as a route to underdiagnosis: patients on these medications may present with classic symptoms of sleep-disordered breathing, yet their sleep study comes back looking normal or near-normal because the very sleep stage where their apnea would show itself has been pharmacologically compressed.²⁰

The flip side is equally important. If a patient abruptly stops an antidepressant, the REM-suppressive effect lifts and a REM rebound can follow. One dramatic case report documented a patient who stopped paroxetine three days before a CPAP titration study and spent 72 percent of total sleep time in REM, far above the normal 20 to 25 percent.²¹ For someone with REM-predominant apnea, an abrupt medication change around the time of a sleep study can produce wildly misleading results in either direction. Sleep physicians increasingly pay attention to medication timing when interpreting polysomnography data, but patients can help by flagging any recent changes in antidepressant use before a study.

Children and REM Sleep Apnea

Children spend a larger proportion of their sleep in REM than adults do, which makes REM-related breathing events especially relevant in pediatric populations. A study of 100 children undergoing adenotonsillectomy for obstructive sleep apnea found that even those classified as having mild disease overall could harbor pockets of severe apnea when the data were broken down by sleep stage and body position. After surgery, the children who had the lowest baseline amounts of REM sleep showed significant increases in REM duration, suggesting their preoperative REM had been truncated by the breathing disruptions.²² This matters for development: REM sleep is thought to play a role in brain maturation, and chronic disruption during childhood could have downstream effects on learning and behavior.

Pediatric sleep studies already include sleep staging as standard, which is an advantage over the adult home-test landscape. Still, the overall apnea-hypopnea index can obscure stage-specific severity in children just as it does in adults, so pediatric sleep specialists increasingly look at the stage-by-stage breakdown when deciding on treatment.