February 08, 2016
12 min read

Sleep disorders and heart disease: Complicated relationship needs more research

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

In recent years, strong evidence has emerged linking sleep disorders to various forms of heart disease — in particular, HF and hypertension. What is less certain, however, is whether treatment of sleep disorders before the onset of heart disease can prevent heart disease, and how to incorporate evaluation of sleep disorders such as sleep apnea into a cardiologist’s general clinical practice.

“Untreated, sleep apnea is associated with increased CHD events, such as MI, atrial fibrillation and other arrhythmias, stroke, sudden death and progression to HF,” Rami Khayat, MD, associate professor of medicine at Ohio State University and director of its Sleep Heart Program, told Cardiology Today.

Although evidence continues to accumulate on the mechanisms of sleep apnea and its association with different forms of heart disease, there is still much uncertainty about how the two relate to each other and the optimal treatments for patients with both.

Virend K. Somers, MBChB, DPhil, FAHA, FACC, of Mayo Clinic, discusses the challenges of treating patients who have sleep apnea and some form of heart disease.
Virend K. Somers, MBChB, DPhil, FAHA, FACC, of Mayo Clinic, discusses the challenges of treating patients who have sleep apnea and some form of heart disease.

Image: Matthew Meyer, Mayo Clinic; printed with permission.

“It is difficult to tease out whether it is the CVD or the obesity or the diabetes that causes obstructive sleep apnea, or whether obstructive sleep apnea is there as a bystander,” Martin R. Cowie, MD, MSc, FRCP, FRCP (Ed), FESC, professor of cardiology at Imperial College London and Royal Brompton Hospital, said in an interview. “But if clinicians and cardiologists screen for sleep-disordered breathing, they will find a substantial number of people that have this disorder.”

Further complicating the issue were the unexpected results of the SERVE-HF study presented at the 2015 European Society of Cardiology Congress and published in The New England Journal of Medicine. Cowie and fellow investigators for the SERVE-HF study found that treating central sleep apnea (CSA) with adaptive servo-ventilation in patients with HF led to worse outcomes.

Wide prevalence of sleep apnea

The most common form of sleep apnea, obstructive sleep apnea (OSA), is widely prevalent in people with a variety of heart diseases.

It is estimated that “40% of patients with OSA have hypertension, and about 40% of patients with hypertension will have OSA,” Virend K. Somers, MBChB, DPhil, FAHA, FACC, the Alice Sheets Marriott Professor of Cardiovascular Diseases at Mayo Clinic, told Cardiology Today. However, he noted that “these numbers depend on demographics.” For example, “if you have more women [in a specific demographic], there might be a lower likelihood of OSA, and if you are not obese, you’ll have less sleep apnea. For AF, if you take all the patients with AF coming for cardioversion, 50% will have a high likelihood of significant sleep apnea. If we look at people who come into the hospital with MI, the level of sleep apnea tends to be around 50% to 60%. That has been consistent across different studies in different environments.”

Much of the research on the link between heart disease and sleep apnea has been conducted in patients with HF. It is estimated that between 50% and 70% of patients with HF have some form of sleep apnea. “In the U.S. population, if you [look at] patients with HF, you’d find about 35% will have OSA and maybe 30% will have CSA,” Somers said.

Ajay V. Srivastava, MD, FACC, advanced HF cardiologist at Scripps Clinic, La Jolla, California, said patients with HF or hypertension are more likely to have sleep apnea based on an increasing number of the following characteristics: significant snoring, fatigue during the day, history of high BP, partner noticing that they stop breathing at night, BMI greater than 35 kg/m2, neck circumference greater than 14 inches and male sex.

Ajay V. Srivastava, MD, FACC
Ajay V. Srivastava

Various studies have reported that, compared with age-, sex- and BMI-matched controls, those with OSA may have a twofold risk for onset of HF, threefold risk for onset of stroke and fourfold risk for onset of hypertension, experts told Cardiology Today.

CSA is a condition found almost exclusively in patients with HF and “people who take a lot of opiates for good or bad reasons,” Cowie said. Unlike OSA, which often arises before the onset of HF, hypertension or other heart disease, CSA typically arises after the onset of HF, which means that the relationship between OSA and heart disease may be quite different than the one between CSA and heart disease.


Mechanisms of risk

OSA appears to predispose people to heart disease because it puts pressure on the heart and adversely affects the autonomic nervous system, experts told Cardiology Today.

“In [OSA], there are big swings in BP and heart rate and big changes in the pressure in the large arteries and veins in the heart,” Cowie said. “It affects the autonomic nervous system with big bursts of sympathetic activity. It’s not just a risk marker in these patients, but a risk factor.”

Specifically, Khayat said, the airway closure occurring during OSA “results in a pattern of hypoxia and is usually terminated by an arousal from sleep. The patient experiences recurrent episodes of hypoxia and reoxygenation through the night. For every episode, each time you have a dip in oxygen level, the body has a response in increased sympathetic activity like adrenaline, which results in vasoconstriction, which results in vascular wall damage, increased BP decreased perfusion to the heart and endothelial dysfunction. This increases the work on the heart and can result in left ventricular hypertrophy.”

Other consequences of this activity, he said, can include putting the heart into arrhythmia, increasing inflammation and increasing the level of renin-angiotensin, which is linked to increased risk for hypertension and HF. Somers cited left atrial enlargement as another likely consequence.

“The mechanism by which [OSA] contributes to CVD is multifactorial and is involved in most of the pathogeneses of CVD itself,” Khayat said.

The causal arrow does not always point one way, Somers noted. Patients with HF appear to be predisposed to later developing OSA, he said.

“When you get HF, you develop fluid retention and edema, and when you sleep, you may develop edema of the upper airway,” he said. “In which case, the resistance to inspiration is greater and the likelihood of collapse is heightened, which could itself contribute to apnea.”

Strangely, while excessive daytime sleepiness is common in many patients with OSA, it usually does not occur in patients with OSA and HF or AF, Somers said. “We really don’t know the reason for this,” he said. “The caution is that just because your patient is not sleepy, if he or she has bad HF or bad AF, it doesn’t mean they don’t have severe OSA.”

Available research

Although researchers have developed a better understanding of the relationship between sleep apnea and heart disease in recent years, most studies conducted in this area have been of mechanistic or observational design, and there has been little in the way of randomized controlled trials to guide clinical practice.

“While we have consistent and striking data that treatment of OSA improves the outcomes of CVDs, we do not have the randomized controlled studies that we can definitively use as a basis for mandating the treatment of [OSA] as part of standard care in patients with heart disease,” Somers said.


A number of studies have found that the standard treatment for OSA, continuous positive airway pressure (CPAP), benefits patients with HF and OSA, Srivastava said.

“CPAP improves BP and heart function in these patients, and the studies have shown that these patients can live longer just by doing this,” he said. “One of the challenges we face is that in any study of a patient population with HF, because they have so many comorbidities, including hypertension, diabetes and CAD, it’s tough to demonstrate that a single therapy has mortality benefit. But because their heart function improves, because patients’ quality of life is better because they sleep better, and because they have less arrhythmia from being on [CPAP], we believe that this is a good therapy to offer.”

Specifically, he said, studies have shown that with CPAP, patients with OSA and HF “can exercise better and their oxygen levels at night are higher. When [researchers] did blood tests to check for biomechanical markers of stress and an activated sympathetic system, they showed that patients on [CPAP] have less of these stress chemicals in their blood.”

The effect of CPAP on CSA is less clear, however. In the CANPAP study of 258 patients with HF and CSA, treatment with CPAP had no effect on transplant-free survival and, while it reduced mean apnea–hypopnea index, it did not reduce it lower than the threshold for inclusion in the study. However, in a post hoc analysis, Michael Arzt, MD, from the University of Toronto, and colleagues found that, at 3 months, patients assigned CPAP who had apnea–hypopnea index reduced below 15 had greater LV ejection fraction (P = .001) and better transplant-free survival (HR = 0.371; 95% CI, 0.142-0.967) compared with controls; similar findings were not observed for patients assigned CPAP whose apnea–hypopnea index was not reduced below 15 at 3 months.

SERVE-HF surprises

With the benefits of CPAP in patients with CSA uncertain, researchers conducted the SERVE-HF study to evaluate whether adding adaptive servo-ventilation (AutoSet CS, ResMed) to guideline-directed medical therapy would improve survival and CV outcomes in patients with HF with reduced ejection fraction and CSA.

Martin R. Cowie, MD, MSc, FRCP, FRCP (Ed), FESC
Martin R. Cowie

“We expected that we would show that our HF patients with CSA would have better results if we treated it with [adaptive servo-ventilation] technology, which is a form of pressure support [that] gets rid of the CSA,” said Cowie, first author on the SERVE-HF study. “And yet, it showed no improvement in the outcomes of our patients. What we did see, which was completely unexpected, was a 34% increase in CV mortality, with a P value very unlikely to be due to chance. We saw the effect occurring early and continuing throughout the study.”

The CV mortality results were driven by sudden cardiac death, Cowie said. “Many patients felt they were better on the therapy, and then they just died suddenly,” he said. “If they had a defibrillator, they were partially protected from that increased sudden death risk, but not completely, suggesting that it was arrhythmic in origin.”

He said a soon-to-be-published substudy indicates that the therapy had no effect on the HF syndrome, but increased sudden cardiac death anyway. “Presumably, it’s an autonomic phenomenon, and it’s led to the thought that, possibly, CSA is an adaptive response by a sick heart. If you have HF, you develop Cheyne-Stokes respiration, and by some mechanism that helps reduce the impact of the syndrome on your arrhythmia. The other theory is that perhaps pressure support, or at least [adaptive servo-ventilation]-type pressure support, is harmful to very poor LVs by direct mechanism, ie, by the pressure being applied, particularly if the patient has been diuresed too much and has low cardiac pressures. I’m less enthusiastic about that interpretation because we don’t see an effect initially which then disappears. We don’t see any worsening in LV function or structure. We don’t see any increase in HF hospitalizations. So it’s much more likely to be a neurohormonal autonomic nervous system type of effect.”


Somers, who was on the SERVE-HF steering committee, said it has yet to be determined how the results will affect clinical practice. “Is it the positive airway pressure that’s the problem, or should we maybe not be treating [CSA]?” he asked. “Or is it Cheyne-Stokes that should not be treated? A nuanced approach to the data suggests that it was mainly in patients with Cheyne-Stokes as opposed to less organized CSA where the increase in mortality was evident. Right now, we are trying to regroup and figure out a logical, effective approach on how to move forward from here.”

Patients with LVEF greater than 45% and those with HF with preserved EF were not included in SERVE-HF, so for them, “it would still be reasonable to consider treating central apnea, particularly if it’s symptomatic,” Somers said.

Implications for treatment

Because there are so few conclusive randomized controlled trials in this area, there is not much in the way of guidelines, either.

The only document devoted specifically to sleep apnea and CVD is a scientific statement from the American Heart Association and American College of Cardiology in collaboration with the NIH that was published in 2008.

Somers, chair of the writing group, said the authors concluded that “in the absence of randomized trials, treatment should be individualized based on the needs of the patient. In general, it’s probably better to treat until we know better.”

He said his own approach is to treat the OSA unless there is a compelling reason not to. “If the patient has a significant cardiac comorbidity, whether it’s HF, recently cardioverted AF, severe hypertension, recent MI, risk for sudden cardiac death or low ejection fraction, I will be even more inclined to strongly recommend treatment for the OSA, especially if the patient is sleepy [during the daytime].”

For optimal treatment to occur, patients must recognize potential OSA symptoms and discuss them with their doctor, and doctors must refer patients they suspect have sleep apnea for a sleep study, Srivastava said.

“The biggest problem we have that limits people from being treated for this condition is underdiagnosis,” he said.

OSA should be treated with CPAP, regardless of which cardiac comorbidities are present, Srivastava said, although “most patients will report that it’s not comfortable and they don’t want to do it. It requires the health care provider to spend time with the patient because it takes some convincing. The newer forms of CPAP have gotten smaller and less cumbersome when at sleep, so they’re more effective as a therapy than in years before.”

Rami Khayat, MD
Rami Khayat

For patients with OSA who cannot tolerate CPAP, there are alternative therapies such as pharmacological therapies, dental appliances or implantable stimulators, but the literature on these is sparse, Khayat said.

Somers said that the best approach for OSA and HF is “to treat the HF aggressively and, if you can, simultaneously treat the apnea, and that will improve BP control and diuresis, may help decrease the likelihood of developing AF or could help you address the AF that may be causing HF.”

CSA is a much tougher diagnosis to make because it shares many of the same symptoms as narcolepsy and certain lung diseases, Srivastava said. There is no FDA-approved therapy specifically for CSA, so patients should be treated for their HF or other cardiac comorbidities as they would otherwise, and “trying different sleep-inducing regimens and seeing if you can overcome the sleep apnea in these patients and improve their quality of life,” he said.

Cowie said: “The data so far suggest that if it’s [CSA], leave it well alone. Treat the HF because that will help the sleep apnea get better ... and don’t treat it specifically with pressure support therapy.”


Given that obesity confers increased risk for sleep apnea and “there is evidence that sufficient weight loss contributes to decreased severity of sleep apnea,” encouraging patients to lose weight is also a good strategy, “but it is not often feasible in patients with CVD,” Khayat said.

A recently published study in Science Translational Medicine indicates that statin therapy could help mitigate the increased risk for heart disease associated with OSA. Memet Emin, MD, from Columbia University College of Physicians and Surgeons, and colleagues discovered that patients with OSA have an unusual cellular distribution of the protein CD59, which appears to contribute to vascular injury. “Increased internalization of CD59 appeared dependent on cholesterol-enriched lipid raft formation in [endothelial cells] and was blocked by statins, suggesting an additional protective effect of statins in OSA patients,” Emin and colleagues wrote.

Future investigation

After the surprise results of SERVE-HF, more research is necessary to optimize treatment of people with sleep apnea and some form of heart disease.

“It has certainly raised awareness. Cardiologists are actually speaking to respiratory and ventilation specialists about it,” Cowie said. “It shows that things can look nice on paper and diagrams, but until we actually do the outcome studies, we never quite know what an intervention might do.”

Also underway, the ADVENT-HF study is evaluating adaptive servo-ventilation in patients with HF with reduced EF and sleep apnea, mostly OSA, and should provide valuable information on who can benefit from the technology, Cowie said.

Somers said the “highest priority” is conducting “randomized controlled trials to identify whether treatment of OSA prevents the development of heart disease or improves hard outcomes of heart disease. We have data suggesting that ejection fraction gets better, patients feel better, exercise capacity gets better, but the question is, do they live longer? Do they have fewer hospitalizations and MIs?”

In addition, he said, more research is needed on “what sleep deprivation is doing to the CV system.”

Srivastava called for “a registry to see which patients are getting put on [sleep apnea] therapies. There is no database as of now that tracks patients with HF and sleep apnea to see how their quality of life is improving, and whether [sleep apnea treatment] has any effects on long-term outcomes.”

Khayat said he would like to see more mechanistic studies on how positive pressure affects cardiac function in patients with HF, and on how sleep apnea contributes to CVD.

“We have made a lot of progress in the last 10 to 15 years, but there is still a lot of work to be done to understand how sleep apnea affects or produces HF and hypertension,” Khayat said. – by Erik Swain

Disclosures: Cowie reports receiving institutional grant support from Bayer and ResMed and consulting for Boston Scientific, Medtronic, Novartis, Pfizer, ResMed, Respicardia, Servier, Sorin and St. Jude Medical. Khayat reports receiving research support from Philips Respironics. Somers reports consulting for GlaxoSmithKline, Philips Respironics, ResMed, Respicardia, Ronda Grey, Sorin and U-Health. Srivastava reports no relevant financial disclosures.