How Studies of Sleep-Disordered Breathing Are Shedding Light on Atrial Fibrillation

Strong epidemiologic findings are put to the test in ongoing clinical studies


By Reena Mehra, MD, MS, FCCP, FAASM

Sleep-disordered breathing (SDB) exposes patients to chronic intermittent hypoxemia and broad swings in intrathoracic pressure. These effects alter autonomic balance, they have untoward impacts on cardiac preload and afterload, and they enhance inflammatory and oxidative stresses, all of which produce a pro-arrhythmogenic milieu (Figure 1).


Figure 1. Atrial fibrillation is demonstrated in the ECG channel in the context of severe repetitive apnea/hypopnea associated with oxygen desaturations during REM sleep.


Animal and human studies have identified potential mechanisms by which SDB directly and indirectly alters the functional and cardiac structural substrate for arrhythmogenesis in atrial fibrillation (AF) (Figure 2).


Figure 2. Schematic showing potential mechanisms by which sleep-disordered breathing (SDB) may contribute to atrial fibrillation (AF). Boxes with dashed lines represent the pathophysiologic indices that link SDB and AF (HRV = heart rate variability; HRT = heart rate turbulence; PAC= premature atrial contraction).

Epidemiologic evidence of a role for SDB in atrial fibrillation

Our group has performed several epidemiologic observational studies that have demonstrated statistically significant associations between SDB and AF (odds ratio point estimates, 2 to 4).

These associations remained even after we took into account a host of potential confounding factors such as age, sex, race, body mass index and self-reported comorbidities such as hypertension, diabetes mellitus, cardiovascular disease and heart failure. For example:

  • In a study of approximately 600 patients who participated in the Sleep Heart Health Study, we found that those with moderate to severe SDB on overnight polysomnography (PSG) had a fourfold higher odds of AF than did patients without SDB.
  • In a cohort of almost 3,000 older men, we found a stronger association with AF among those with central sleep apnea than those with obstructive sleep apnea, even after controlling for confounding factors. In this study, we noted a threshold effect in which patients with moderate to severe SDB (apnea-hypopnea index ≥ 24) had the highest incidence of AF, independent of any self-reported heart failure and cardiovascular disease. Patients who had a central apnea index greater than 3 had a threefold higher incidence of AF, and those with Hunter Cheyne Stokes breathing had an almost fivefold higher incidence.
  • We also examined the temporal relationships between discrete respiratory events and paroxysms of AF. This investigation involved a novel application of a case-crossover study design, which is well suited for studying short-lived exposures and outcomes. We found a strong temporal relationship between apneas/hypopneas and paroxysms of AF. Indeed, we noted a 17-fold higher odds of episodic AF during the 90 seconds following an apnea/hypopnea event than after a period of nonobstructed breathing. This finding supports the premise that SDB plays a role in the etiology of atrial arrhythmias.

Implications for future research

Now that compelling data have been accumulated regarding aspects of the SDB-AF relationship, future clinical and epidemiologic research should focus on specific areas, including:

  • Collection of objective data on cardiac function
  • Measurement and analysis of markers of autonomic function, systemic inflammation and oxidative stress
  • Examination of both daytime and nocturnal ECGs in an effort to further elucidate pathophysiologic underpinnings
  • Reversal of SDB pathophysiology in order to alleviate AF and its associated morbidity and mortality

To overcome knowledge gaps in these areas, our group is conducting two NIH-funded research studies, detailed below.

The SAFEBEAT study

The Sleep Apnea and Atrial Fibrillation Electrophysiology: Biomarkers and Evaluating Atrial Triggers (SAFEBEAT) study involves examination of paroxysmal AF because it provides an ideal milieu in which to investigate the immediate influences of SDB and to examine its temporal patterns in view of its intermittent nature.

In this case-control study, we are comparing 150 patients with paroxysmal AF with 150 controls without paroxysmal AF. Participants are being matched for important confounders such as age, sex, race and body mass index. They will be characterized on the basis of detailed collections of overnight sleep study data, echocardiographic measures, biomarkers and continuous ECG monitoring. With these data in hand, we will have the opportunity to explore the relationships between paroxysmal AF and both obstructive and central apnea.

Another goal is to investigate the associations between paroxysmal AF and age in SDB; our earlier work showed that the association of SDB and arrhythmia is stronger in younger patients.

Moreover, we are examining diurnal variations in paroxysmal AF in patients with SDB in terms of immediate and chronic SDB-related physiologic stresses (i.e., intermittent hypoxia, intrathoracic pressure alterations and autonomic influences).

Finally, we are assessing the effects of SDB treatment on paroxysmal AF to inform future randomized controlled trials in this area.

A trial to identify SDB phenotypes that predict atrial fibrillation

The primary goal of our second NIH-funded study, Sleep-Related Respiratory and Electrophysiological Atrial Fibrillation Predictors, is to identify PSG-based SDB phenotypes that predict incident AF. We will be investigating the relative contributions of central apnea and periodic breathing vs. obstructive apnea, as well as mediation by inflammation and oxidative stress.

Another aim is to identify PSG-derived ECG markers of atrial ectopy, conduction delay and autonomic imbalance, and then to evaluate the markers’ utility as predictors of incident AF.

For parts of this study, we will be collaborating with engineers at Case Western Reserve University.

Depending on what we find, the results of this study may lead to a shift in the current clinical paradigm by identifying which PSG-based physiologic indices should be included in standard PSG monitoring to forecast arrhythmias such as AF.


Dr. Mehra is Associate Professor of Medicine at Cleveland Clinic Lerner College of Medicine and Director of Sleep Disorders Research in Cleveland Clinic’s Sleep Disorders Center.