“REM-related” Obstructive Sleep Apnea: An Epiphenomenon or a Clinically Important Entity?

Obstructive sleep apnea (OSA) is a prevalent sleep disorder characterized by recurrent episodes of complete or partial collapse of the upper airway that are associated with intermittent hypoxemia and recurrent arousals from sleep. Although upper airway collapse can occur in rapid eye movement (REM) and non-REM (NREM) sleep, the withdrawal of excitatory noradrenergic and serotonergic inputs to upper airway motor neurons during REM sleep further reduces pharyngeal muscle activity and substantially increases the propensity for upper airway collapse.¹ Thus, REM sleep in patients with OSA is typically associated with increased frequency of obstructive events that are often prolonged and accompanied with severe oxyhemoglobin desaturation. The term “REM-related” OSA has been coined to denote the occurrence of disordered breathing events that predominantly or exclusively manifest during REM sleep—an observation that is common in clinical practice and well recognized by healthcare professionals in sleep medicine. The prevalence of REM-related OSA in clinical samples is highly variable, ranging from 10% to as much as 36%.^2–5 The variability in reported prevalence is due, in part, to the differences in sample characteristics and the definitions used for REM-related OSA.⁶

A commonly used diagnostic criterion is based on the ratio of the apnea-hypopnea index (AHI) during REM sleep to the AHI during NREM sleep (AHI_REM/AHI_NREM), with a value of at least 2 indicating a predominance of disordered breathing during REM sleep. Although using a ratio of rates (i.e., events/h in REM sleep divided by events/h in NREM sleep) may have face validity, there are several inherent limitations. First, classifying patients with OSA based on the AHI_REM/AHI_NREM ratio by itself is problematic because it will undoubtedly designate some patients as having REM-related OSA when, in fact, there is also substantial disease during NREM sleep. For example, a patient with an AHI_REM of 70 events/h and an AHI_NREM of 35 events/h would be classified as having REM-related OSA when, in fact, the frequency of disordered breathing events during NREM sleep is not insignificant. In essence, the AHI_REM/AHI_NREM ratio does not accurately depict the occurrence of disordered breathing events predominantly during REM sleep because it can be high either with (1) a high AHI_REM, (2) a low AHI_NREM, or (3) a combination of both. Second, estimates of the AHI_REM/AHI_NREM ratio can be highly imprecise, particularly if no consideration is given to the amount of REM sleep. For example, a patient with 6 minutes of REM sleep and 3 obstructive events during that period (AHI_REM = 30 events/h) would be classified as having REM-related OSA, assuming that the AHI_NREM is 15 or fewer events/h. However, 6 minutes of REM sleep is not sufficient to precisely estimate the severity of disordered breathing during REM sleep. Thus, having a minimal requirement of REM-sleep duration is critical in assessing the presence of REM-related OSA. A third limitation with the use of the proposed ratio is that its numerator (AHI_REM) and denominator (AHI_NREM) are highly correlated (r = 0.5-0.6). The strong correlation is not surprising given that disordered breathing events often manifest during both NREM and REM sleep. Defining a clinical entity based on a ratio of 2 rates (i.e., AHI_REM and AHI_NREM) that are interdependent and correlated may not be possible because of the difficulty in separating out the NREM component of the underlying disease. In light of the aforementioned pitfalls, we propose the following criteria to designate the predominance of disordered breathing during REM sleep: (a) an AHI_NREM of fewer than 5 events/h and (b) an AHI_REM of at least 5 events/h with at least 30 minutes of REM sleep. If the amount of REM sleep is less than 30 minutes, estimates of AHI_REM should be considered imprecise, and decisions regarding disease classification should be made with caution. The threshold value of 30 minutes was based on the observation from the authors that 75% of all patients referred to the sleep laboratories for full-night polysomnography at the University of Chicago and Johns Hopkins University have had REM sleep time of at least 30 minutes.

Given the significant heterogeneity in defining REM-related OSA, it is not surprising that its epidemiology, natural history, and clinical significance are not well defined. A number of studies have shown that REM-related OSA is common in women and in younger patients.^3–5,7,8 Aside from such basic descriptions, further characterization of its clinical attributes remains unclear, particularly regarding the possibility that patients with OSA only during REM sleep may differ in health outcomes when compared with those with disordered breathing events that are not strictly confined to REM sleep. To consider REM-related OSA as an entity with clinical significance, it is paramount that empiric evidence be collected to determine if it operates as a prescriptive element for adverse neurocognitive, cardiovascular, and metabolic outcomes.

A number of clinic- and community-based studies have suggested that only obstructive events during NREM sleep are associated with excessive daytime sleepiness or impaired quality of life.^9–12 In contrast, although some studies^2,13 have reported a significant association between indices of REM-related disease and objective and subjective sleepiness, an equal number of other studies^10–12 have not found such associations. Data from the community-based Sleep Heart Health Study have recently shown that AHI_REM is not associated with daytime sleepiness or quality of life after adjusting for AHI_NREM.⁹ Collectively, these observations would suggest that perhaps REM-related OSA does not have associated impairments in daytime sleep tendency or quality of life. Alternatively, it is also possible that the instruments used in the available studies are crude and may not accurately appraise the associated impairments of sleepiness or quality of life. More importantly, it is important to recognize that neither the cardiovascular nor the metabolic implications of REM-related OSA have been systematically evaluated.

If REM-related OSA is of clinical significance, then the oligosymptomatic and milder nature of this condition, by virtue of the disproportionately lower percentage of REM sleep as a function of total sleep duration, increases the likelihood that these patients will often remain undiagnosed and untreated. It may also go underdiagnosed because sleep testing in clinical laboratories is often terminated earlier than the usual wake-up times, leading to REM-sleep curtailment. Although the phenomenon of REM-related OSA has been clinically recognized for years, a fundamental question remains unanswered: do obstructive disordered breathing events during REM sleep impose the range of consequences that are of similar or greater magnitude than those elicited by events occurring during NREM sleep? Specifically, what is the cardiovascular, metabolic, and cognitive burden of obstructive events that cluster in REM sleep when there is no significant disease during NREM sleep? It is now well established that, compared with NREM sleep, REM sleep is associated with greater sympathetic activity and cardiovascular instability in normal subjects and in patients with OSA.^14,15 Indeed, obstructive apneas and hypopneas during REM sleep lead to greater degrees of hypoxemia and higher levels of sympathetic activity, compared with events in NREM sleep.¹⁶ Thus, REM-related events could impose greater cardiometabolic risk than events during NREM sleep. It is also important to note that OSA in nonobese children is primarily a REM-related phenomena.⁷ Studies of pediatric patients with OSA have shown a multitude of end-organ morbidities, including metabolic dysfunction,¹⁷ vascular endothelial dysfunction,^18,19 hypertension,²⁰ systemic inflammation, and increase in oxidative stress.^21,22 Such observations regarding the health implications in children with OSA provide strong incentives for delineating the clinical importance of REM-related events in adults.

The question of whether REM-related OSA is a clinically important entity in adults has implications for practice and policy. In the absence of empirical evidence, there is lack of consensus as to whether patients with disordered breathing events predominantly during REM sleep should be treated with continuous positive airway pressure (CPAP). Furthermore, patients with REM-related OSA who are willing to assess whether nightly CPAP use improves their daytime sleepiness or fatigue may be denied therapy if their overall AHI does not meet a threshold for the diagnosis of OSA (e.g., ≥ 5 events/h). Importantly, if REM-related OSA is indeed part of the spectrum of disordered breathing that, over time, progresses to disordered breathing in all stages of sleep, lack of recognition and institution of therapy could lead to an increase in cardiometabolic risk. Thus, before concluding that REM-related OSA is just an epiphenomenon or does not warrant treatment, it is imperative that a systematic assessment of its effects be undertaken through rigorously conducted clinical trials and observational studies. Without doubt, numerous questions remain unanswered regarding REM-related OSA. For instance, should the definition of this clinical entity include criterion regarding absence of airflow limitation during NREM sleep? It could be easily argued that, to unequivocally define the effects of REM-related OSA, NREM sleep should be with minimal or no evidence of disordered breathing. Other questions of clinical significance include the duration of therapy that is necessary prior to reaching inferences regarding efficacy or effectiveness. Furthermore, given that REM-related OSA occurs in specific patient subsets, should intervention studies focus primarily on these subgroups? Only with rigorously designed studies that carefully incorporate the aforementioned considerations can the clinical relevance of REM-related OSA be adequately defined. The field of sleep medicine should not struggle with the issue of whether REM-related OSA is a distinct entity or an epiphenomenon. Rather the focus of future research should be on careful and continued probing of the health consequences related to OSA—a chronic disorder that manifests as a heterogeneous spectrum from airflow limitation (i.e., snoring) to predominance of REM-related events to the occurrence of apneas and hypopneas throughout all of sleep.

DISCLOSURE STATEMENT

This was not an industry supported study. Dr. Punjabi has received research grant support from RESMED Inc. for an ongoing multi-center clinical trial on the effects of positive pressure therapy in patients with obstructive sleep apnea. Dr. Mokhlesi has served as a consultant for Philips-Respironics.

CITATION

Mokhlesi B; Punjabi NM. “REM-related” obstructive sleep apnea: an epiphenomenon or a clinically important entity? SLEEP 2012;35(1):5-7.