CHAPTER 77: SLEEP-DISORDERED BREATHING AND CARDIAC DISEASE

Sleep-disordered breathing (SDB), or sleep apnea, is common in patients with cardiovascular disease (CVD), and its presence is associated with a poorer prognosis and high health-care costs.^1,2 Increasingly, international guidelines suggest it is worthwhile to screen for this condition and that appropriate treatment may not only improve the patient's quality of life, but may also reduce the risk of cardiovascular events. However, recent evidence suggests that the relationship between SDB and the underlying cardiovascular condition may be complex, particularly in heart failure (HF), and further research is required to clarify the most appropriate treatment pathway for the individual patient.

SDB includes obstructive sleep apnea (OSA), central sleep apnea (CSA), or a combination of both. In OSA (the most common form of SDB in the general population), there is collapse of the pharynx during sleep with consequent upper airway obstruction, often with snoring (Fig. 77–1).³ Predisposing factors include obesity, a short neck, and retrognathism. In HF, rostral fluid shift during sleep can lead to pharyngeal edema, which may exacerbate the tendency to obstruction.⁴

CSA is usually associated with HF, although it has also been observed in patients with stroke, especially in the acute phase, and in those with renal failure.⁵ In CSA, the underlying abnormality is in the regulation of breathing in the respiratory centers of the brainstem. In normal physiology, minute ventilation during sleep is primarily regulated by chemoreceptors in the brainstem and carotid bodies, which trigger an increase in respiratory drive in response to a rise in partial pressure of arterial carbon dioxide (PaCO₂), thus maintaining PaCO₂ within a narrow range. Patients with HF and CSA tend to have an exaggerated respiratory response to carbon dioxide (CO₂), associated with excess sympathetic nervous activity, so that modest rises in PaCO₂ that may occur during sleep result in inappropriate hyperventilation.^5,6,7 This drives the PaCO₂ below the “apneic threshold,” at which point the neural drive to respire is too low to stimulate effective inspiration and an apnea (complete pause in breathing) or hypopnea (partial reduction in airflow) ensues. PaCO₂ subsequently rises, and the cycle is repeated. This overshoot of the homeostatic feedback loop is exacerbated by the prolonged circulation time between the alveoli and the brainstem seen in more severe HF, so that the PaCO₂ sensed in the brainstem may not accurately reflect the PaCO₂ at the lung. CSA is associated with lower resting PaCO₂, increased sympathetic nervous activity, and more severe cardiac dysfunction.⁸ The more prolonged the circulation time, the longer the duration of the hyperpneic phase of CSA.^9,10 In addition, pulmonary congestion and edema lead to stimulation of J receptors in the lungs, triggering reflex hyperventilation. A particular form of CSA is a periodic pattern of hyperventilation followed by hypoventilation, termed Cheyne-Stokes respiration (CSR). The physiology of CSA is shown in Fig. 77–2.¹¹

A tendency to progress from OSA to CSA over the course of the night has been observed in patients with HF. This is thought to be secondary to progressive pulmonary congestion and deteriorating hemodynamics, which may be exacerbated by the SDB.⁸

Interestingly, periodic breathing (CSR) is not limited to sleep but can occur at rest or develop during exercise in patients with advanced HF.^12,13 Recently, it has been suggested that CSR may be a compensatory mechanism in patients with HF.^14,15

Apnea is currently defined as a reduction in airflow by at least 90% of pre-event baseline for at least 10 seconds; hypopnea is defined as a reduction in airflow by at least 30% from baseline for at least 10 seconds, associated with a fall in arterial oxygen saturation of at least 3% or an arousal from sleep.¹⁶ In OSA, there is evidence of ongoing respiratory effort throughout the apneic-hypopneic event, often with paradoxical movement of the chest and abdomen as breathing against a closed airway is attempted. In contrast, apneas and hypopneas in CSA are accompanied by a marked reduction or cessation of respiratory effort.

The severity of SDB is generally described by the average number of apneic and hypopneic events per hour of sleep, or the apnea-hypopnea index (AHI). Normal is defined as up to 5 events per hour, mild as 5 to 15 events per hour, moderate as 15 to 30 events per hour, and severe as > 30 events per hour. The number and severity of oxygen desaturations may also be used as a metric of the severity of SDB. Additionally, those in whom > 50% of events are obstructive are labeled as having predominantly OSA, and those in whom > 50% of events are central are labeled as having predominantly CSA.

In patients without CVD, typical symptoms of SDB include excessive daytime sleepiness including drowsy driving, insomnia, morning headaches, depression, cognitive dysfunction, nocturnal dyspnea, nocturia, and erectile dysfunction. However, there is wide interindividual variation in symptoms, especially between male and female patients.¹⁷

At least 50% of patients with severe OSA do not report symptoms of unrestful sleep. This proportion is even higher in patients with OSA and CVD, who often primarily report the symptoms of the underlying CVD, such as chest pain or breathlessness, rather than the typical symptoms of OSA.^18,19

Patients with HF and SDB do not tend to complain of daytime sleepiness, possibly because of the high sympathetic tone found in HF. Screening questionnaires that include questions about daytime sleepiness (such as the Epworth Sleepiness Scale used to screen for OSA in non-HF populations) are, therefore, not useful.²⁰

Attended in-hospital polysomnography (PSG), including assessment of respiratory movement, oxygen saturation, nasal and oral airflow, snoring, electroencephalography, electrocardiography, electromyography, and ocular movement, has long been considered the gold standard test for sleep disorders. This provides comprehensive data, but it is expensive, laborious, and not available in all centers. More limited, multichannel sleep polygraphy with oxygen saturation, nasal airflow, and chest and abdominal movement recorded is more widely available and can be set up by the patient at home. It is comfortable to wear and less expensive than in-hospital PSG, and the data produced are relatively simple to analyze (Fig. 77–3).²¹ Studies comparing the diagnostic accuracy of home polygraphy with gold standard PSG have shown that polygraphy has a sensitivity and specificity of 90% to 100% for the diagnosis of significant SDB, at least in patients with HF.^22,23 An advantage of PSG is that periods of wakefulness are easily identified and, therefore, AHI can be recorded during sleep only, whereas in polygraphy, AHI is calculated throughout the recording period regardless of sleep pattern, and this may lead to an underestimation of the severity of the SDB.

Even simpler screening for SDB may be performed by recording nocturnal oxygen saturation via a finger probe. For the diagnosis of moderate-to-severe SDB in patients with HF, a sensitivity of 93% and specificity of 73% compared to PSG when using a cutoff of 12.5 desaturations of ≥ 3% per hour for patients have been reported, implying that few patients with clinically important SDB would be missed by this simple first-stage approach.²⁴ Oxygen saturation monitoring is a simple and inexpensive method for screening for SDB, but it cannot determine the phenotype of SDB, and further investigation with (at least) polygraphy is mandatory in those who test positive and in anyone who tests negative but in whom clinical suspicion remains high.

Nocturnal heart rate variability, whether on a 24-hour tape or from an implanted device, reflects autonomic tone, but does not appear to be a useful approach to screening, although more sophisticated analytical algorithms show promise.²⁵

In the past 10 years, there has been increasing interest in whether algorithms could be developed in cardiac implantable electronic devices (such as pacemakers and defibrillators) to accurately detect and quantify SDB.²⁶ It is possible to continually measure thoracic impedance between the right ventricular lead tip and the generator. On inspiration, the increased volume of air in the chest increases thoracic impedance with the inverse occurring on expiration, with consequent proportional changes in detected potential difference. SDB diagnostic algorithms are now commercially available. The DREAM study reported a sensitivity of 89% and specificity of 85% for the diagnosis of moderate-to-severe SDB by a pacemaker algorithm using transthoracic impedance and minute ventilation sensors.²⁷ Studies on other device platforms are currently under way (including NCT02204865 [ClinicalTrials.gov identifier]). With the additional facility to download implanted device data remotely, it is hoped that changes in SDB severity might be useful as an early marker of HF decompensation, providing a window of opportunity to optimize treatment. Further research, including prospective validation studies, is required.

Age

The prevalence of SDB varies considerably by age group, with one peak occurring in children ages 4 to 8 years, related to adenotonsillar hypertrophy or craniofacial abnormalities, after which there is a declining prevalence until early adulthood, when the rates of SDB progressively increase into older age,²⁸ perhaps related to reduced pharyngeal caliber, increased pharyngeal length and parapharyngeal fat pad volume, or impairment of protective pharyngeal reflexes.^29,30,31 Among older individuals, the prevalence of SDB is at least 20%.^32,33,34 The high prevalence of SDB in older individuals, the population group at highest risk for CVD, underscores the importance of determining the link between this comorbidity and morbidity and mortality in cardiovascular practice.³³

Sex

Most population-based data indicate that the prevalence of SDB is two- to three-fold higher in men compared with women.^35,36 Men also experience a more rapid progression of the AHI with age than women,^28,37 although sex differences decrease after women reach menopause. Women tend to have less severe OSA than males as a result of milder episodes of upper airway resistance and flow limitation that do not meet the criteria for apnea. This is likely a result of differences in the upper airways geometry, parapharyngeal fat distribution, and respiratory stability.³⁵ These gender differences tend to decrease with age.¹⁷

Ethnicity

Differences in the prevalence of SDB among individuals of varied racial and ethnic backgrounds likely reflect genetic factors that influence craniofacial shape, body fat distribution, and physiologic risk factors for SDB, as well as the influence of environmental factors. In the United States, the prevalence of SDB among African Americans has been reported to be between two- and six-fold higher than European American populations.^38,39,40,41 Race-based differences in chemoreceptor and baroreceptor responses also may predispose African Americans to the development of SDB.⁴² Several studies of Asian populations, with lower rates of obesity than in many Western societies, also have identified a relatively high prevalence of SDB, which is likely attributable to race-based differences in craniofacial morphology.^43,44 In an elderly US cohort of men, the highest prevalence of SDB was among the Asian American sample, who, after considering the effects of obesity, had a two-fold higher SDB prevalence compared with whites.³³

Obesity

Approximately 40% to 60% of cases of SDB are attributable to being overweight.³⁵ It is estimated that a 1% increase in body mass index (BMI; kg/m²) is associated with a 3% increase in AHI.⁴⁵ Excess body weight increases susceptibility to upper airway collapsibility by increased mechanical loading of the upper airway, which causes the upper airway to assume an elliptical, relatively collapsible shape, and by reducing resting lung volume, resulting in a loss of caudal traction on upper airway structures. Additional effects may be related to release of adipokines such as leptin, which can affect ventilatory control.⁴⁶

Smoking and Alcohol Use

Although exposure to tobacco smoke may increase upper airway inflammation and snoring, smoking is not a firmly established risk factor for SDB. Alcohol use may exacerbate SDB, especially during the first 2 hours after ingestion, as a result of a decrease in upper airway muscle tone and perhaps a reduction in central respiratory drive.^47,48 In patients with SDB treated with positive airway pressure, alcohol use may increase the required pressure needed to maintain airway patency.⁴⁷

Heart Failure

There is likely a bidirectional relationship between congestive HF and SDB, particularly CSA. Significant risk factors identified for CSA in HF in patients referred to a sleep laboratory include male sex (odds ratio [OR] = 3.50), atrial fibrillation (AF; OR = 4.13), age > 60 years (OR = 2.37), and hypocapnia (partial pressure of carbon dioxide [PCO₂] < 38 mm Hg during wakefulness; OR = 4.33).⁴⁹ HF may contribute to SDB development by precipitating periodic breathing and upper airway neuromotor instability as well as through effects that include accumulation of fluid or edema of the upper airway, thereby enhancing collapsibility and thus the tendency to obstruction. A recent study of more than 6500 patients in Germany with systolic HF confirmed the strong association between obesity, male sex, AF, age, and poorer left ventricular systolic function and the presence of SDB.⁵⁰

Intermittent Hypoxemia

Cyclical episodes of hypoxemia-reoxygenation occur in patients with SDB and result in increased inflammation and production of oxygen-derived free radicals, with likely resultant tissue damage.⁵¹ The apnea-related recurrent hypoxemia-reoxygenation in SDB may be analogous to ischemia/reoxygenation injury.^52,53,54,55 The severity of oxygen desaturation is associated with levels of vascular endothelial growth factor, a stimulator of neoangiogenesis.⁵⁶ Intermittent hypoxia and reoxygenation may also result in activation of the proinflammatory transcription factor nuclear factor-κB.⁵⁷ Atherogenesis may result as a consequence of endothelial cell and leukocyte activation and increased expression of adhesion molecules.^{58,59,60,61,62} Endothelial damage may also result from intermittent blood pressure and heart rate surges associated with marked sympathetic activation.⁶³ Other adverse responses to hypoxia include activation of “stress” genes⁶⁴ that influence oxygen delivery, such as hypoxia-inducible factor-1.⁶⁵

Sympathetic Nervous System Activation

Sympathetic nervous system activity is heightened in SDB, with increased muscle sympathetic nervous system activity (measured by peroneal microneurography)^66,67,68 and elevated urinary norepinephrine concentrations.⁶⁸ Urinary norepinephrine increases in parallel with the severity of SDB, the number of arousals, and the degree of oxygen desaturation.⁶⁹ The mechanism for the increased sympathetic nervous system activation remains unclear, but appears to be related to upper airway closure, hypoxemia, hypercarbia, and arousals associated with the respiratory events.^70,71 SDB treatment with positive airway pressure (PAP) therapy lowers urinary norepinephrine excretion^68,72,73 and plasma norepinephrine, vanillylmandelic acid, and metanephrine levels.⁷⁴ Similarly to hypoxemia, enhanced sympathetic nervous system activity may stimulate the expression of inflammatory cytokines.⁷⁵

Alterations in Intrathoracic Pressure

The repetitive inspiratory efforts during the apneas and hypopneas in SDB lead to exaggerated negative intrathoracic pressure swings (up to -65 mm Hg intrathoracic pressure in OSA), causing a cardiac load unique to SDB,^76,77 including increased left ventricular transmural pressure, increased afterload and right ventricular venous return, and an abnormal leftward shift of the interventricular septum.⁷⁸ This leads to increased myocardial oxygen demand, impaired myocardial relaxation, and reduced cardiac output. Progressive increases in intra-atrial pressures lead to atrial myocardial overstretching and dilation, causing cardiac volumetric changes and electrical remodeling that may lead to AF.^79,80 Adverse effects on intrathoracic pressure may worsen the hemodynamic status in patients with HF. Several studies have reported that left ventricular diastolic function declines in association with SDB, especially with aging.^81,82,83

Cardiac Remodeling

Animal models have shown the development of hypertension, left ventricular hypertrophy, and reduced left ventricular ejection fraction as a result of experimentally manipulated long-term SDB.^84,85 In humans, a progressive increase in left ventricular mass index with AHI level, independent of BMI, has been reported in the Sleep Heart Health Study, an observational cross-sectional study investigating cardiovascular outcomes in SDB.⁸⁶ More severe SDB, as defined by higher AHI and more hypoxemia, was associated with greater left ventricular systolic dimensions and lower left ventricular ejection fraction. Left ventricular diastolic dysfunction also appears to be poorer in patients with more severe SDB, independently of obesity, diabetes mellitus, and hypertension.⁸⁷ SDB may more adversely affect myocardial function in patients with underlying coronary artery disease (CAD) than in those without CAD.⁸⁸ Mechanisms of such harm are likely to include marked swings in preload and afterload, sympathetic nervous system activity, increased myocardial oxygen demand and decreased myocardial oxygen supply, and direct effects of hypoxemia.⁸⁹

Sleep Reduction and Fragmentation

SDB likely exerts its negative physiologic effects in part as a result of reduced quantity of sleep and excessive sleep fragmentation caused by repetitive upper airway obstruction-induced sleep disruption. Reduced sleep duration has been associated with cardiovascular morbidity and all-cause mortality in several epidemiologic surveys.^{90,91,92,93,94} The proposed pathophysiologic mechanisms likely include increased inflammation as evidenced by research showing elevations in interleukin (IL)-6,⁹⁵ high-sensitivity C-reactive protein,⁹⁶ and leukocyte counts^97,98 occurring with sleep deprivation. Because sleep is a complex neurophysiologic process, reductions in specific stages of sleep, as may occur in SDB, may be particularly deleterious. For example, reduction in rapid eye movement sleep may contribute to activation of lipid peroxidation and inhibition of antioxidants,⁹⁹ and reduction in slow-wave sleep, a state highly associated with endocrine functions, may decrease insulin sensitivity, increase cortisol, and increase central body fat deposition.^100,101

Metabolic Dysregulation

Several observational studies have demonstrated associations between sleep apnea and insulin resistance that are independent of obesity.¹⁰² An uncontrolled interventional study demonstrated improved insulin sensitivity in patients with moderate-to-severe sleep apnea with treatment, with the best responses occurring in those with a BMI < 30 kg/m².¹⁰³ Using frequently sampled intravenous glucose tolerance testing, SDB was found to be associated with reduction in insulin sensitivity, glucose effectiveness, and pancreatic β-cell function.¹⁰⁴ Glucose intolerance and insulin resistance in SDB may be mediated by upregulation of inflammatory cytokines.¹⁰⁴ Intermittent hypoxia associated with SDB may lead to glucose intolerance and insulin resistance by promoting the release of proinflammatory cytokines, such as IL-6 and tumor necrosis factor (TNF)-α. IL-6 levels have been correlated with indices of insulin resistance, and higher levels have been associated with an increased risk of type 2 diabetes mellitus.^105,106,107 Increased levels of TNF-α also likely contribute to the development of insulin resistance.^{108,109,110,111} Two clinic-based studies have shown that plasma levels of IL-6 and TNF-α are higher in patients with SDB than in control subjects.^112,113 Animal studies showing an increase in insulin levels with exposure to hypoxemic conditions provide further evidence implicating SDB in the pathogenesis of metabolic dysfunction.^114,115,116

Increased sympathetic activity associated with SDB also may affect glucose homeostasis by increasing glycogen breakdown and gluconeogenesis.^67,117,118 Further predisposition toward metabolic dysfunction may occur through effects of SDB on the hypothalamic-pituitary-adrenal axis. Experimental partial or total sleep deprivation has been shown to increase levels of plasma cortisol on the following evening at a time when the circadian rhythm of the hypothalamic-pituitary-adrenal axis is at its nadir.¹¹⁹ The increase in evening cortisol can result in pronounced increases in serum glucose levels and insulin concentrations and increased insulin secretion.¹²⁰

Endothelial Dysfunction

Endothelial dysfunction may occur in SDB as a result of systemic inflammation, oxidative stress, and sympathetic nervous system activation. Individuals with SDB have impaired resistance-vessel endothelium-dependent dilation.¹²¹ A controlled study in normotensive men with SDB demonstrated that endothelium-dependent, flow-mediated dilation significantly increased in men who were compliant with 6 months of SDB treatment with PAP.¹²² Two studies suggested that the effects of SDB on endothelial function may be stronger in women than men.^123,124

Thrombosis

There is some evidence for a hypercoagulable state in sleep apnea, which may contribute to the increased risk of vascular events. Sleep apnea has been associated with increased levels of plasminogen activator inhibitor-1 (PAI-1),¹²⁵ fibrinogen,¹²⁶ activated coagulation factors XIIa and VIIa, thrombin/antithrombin III complexes, and soluble P selectin.¹²⁷ PAI-1, a molecule that inhibits fibrinolysis by inactivation of tissue plasminogen activator, has been shown to contribute to atherothrombotic events and an increased risk of recurrent myocardial infarction (MI).^128,129,130 PAI-1 levels have been shown to vary with measures of hypoxia^131,132 and sympathetic nervous system activation.¹³³ PAI-1 levels have been reported to be elevated in individuals with SDB and to improve with SDB treatment.¹³⁴ Fibrinogen plays a role in clot formation and has been identified as a predictor of CAD.¹³⁵ Several small studies have identified an association between fibrinogen and SDB^136,137; however, potential confounders, such as obesity, were not considered in these studies.

The potential links between SDB and cardiac disease are illustrated in Fig. 77–4.

Overall Mortality

Observational studies of clinic-based patients have reported that patients with SDB have a higher risk of death and that treatment with continuous positive airway pressure (CPAP) therapy may reduce this risk.^{32,138,139,140,141,142,143,144,145,146,147,148,149} Many of these studies were small and did not fully consider confounding factors such as obesity, hypertension, and CVD, and changes in other treatment. However, three larger population-based cohort studies have confirmed that after attempting to account for such confounders, SDB appears to be independently associated with all-cause mortality.^150,151,152

The Australian Busselton Study of 397 participants followed for a mean period of 13.4 years reported an adjusted hazard ratio for mortality of 4.4 to 6.2 for moderate-to-severe SDB (AHI ≥ 15).¹⁵⁰ The Wisconsin Sleep Cohort Study involving 1522 participants, with follow-up over a period of 18 years, demonstrated no significant increase in all-cause mortality in individuals with an AHI between 15 and 30, but demonstrated a significant hazard ratio of 2.7 to 3.8 for severe SDB (AHI ≥ 30).¹⁵² In the much larger Sleep Heart Health Study, which included 6441 middle-aged and older adults with an average follow-up time of 8.2 years from several US communities, SDB was associated with increased all-cause and CVD-related mortality (Fig. 77–5). The association was strongest in men aged 40 to 70 years with severe disease (AHI ≥ 30), with an adjusted hazard ratio of approximately 2.¹⁵¹

Hypertension

Animal studies suggest a central role of hypoxemia occurring with upper airway obstruction in the pathogenesis of hypertension,^84,153 whereas a human study suggests a critical role of upper airway-associated arousal.¹⁵⁴ Increases in blood pressure and heart rate occur approximately 5 to 7 seconds after the termination of an apnea or hypopnea, coincident with arousal, peak ventilation, and the nadir of oxygen saturation.^155,156 The hemodynamic profile of blood pressure patterns in SDB is characterized by a nondipping blood pressure pattern,¹⁵³ a profile also associated with an increase in CVD.¹⁵⁷

Repetitive upper airway obstruction-associated sympathetic nervous system activation, intermittent negative intrathoracic pressure swings, sleep deprivation, and recurrent episodes of hypoxemia and hypercapnia resulting in arterial chemoreceptor activation and increased sympathetic activity, which individually or interactively may result in acute blood pressure increases, may also contribute to sustained sympathetic activation and chronically elevated blood pressure even during nonsleep periods.

Abundant cross-sectional data support an association between SDB and hypertension independent of obesity,^{34,158,159,160,161,162,163,164,165} including data from large cohorts such as the Wisconsin Sleep Cohort Study and the Sleep Heart Health Study. In a cross-sectional analysis of 709 Wisconsin Sleep Cohort participants, an independent association between blood pressure and SDB (AHI ≥ 5) was observed.³⁴ In a Pennsylvania-based study involving 1741 participants, a dose-response pattern was observed for SDB and hypertension (ie, the ORs for hypertension were 6.8, 2.3, and 1.6 for moderate or severe SDB, mild SDB, and snoring, respectively). Additionally, this association was strongest among individuals who were young and of normal weight.¹⁶⁶ In a cross-sectional analysis of data from 6132 participants in the Sleep Heart Health Study, mean systolic and diastolic blood pressure and prevalence of hypertension increased significantly with increasing severity of SDB, measured by the AHI and percentage of sleep time spent below 90% oxygen saturation, and a 40% to 50% increased adjusted odds of hypertension was observed for those with an AHI ≥ 30 or upper quartile of hypoxemia.¹⁶⁵

In a sample of 2677 adults referred to a sleep clinic with suspected SDB, the mean blood pressure level and prevalence of hypertension were reported to increase linearly with severity of SDB; each additional apneic event per hour of sleep was estimated to increase the odds of hypertension by approximately 1%, and each 10% decrease in nocturnal oxygen saturation increased the odds by 13%.¹⁶³ Overall, these studies demonstrate associations between SDB and hypertension even with mild levels of SDB, show dose-response relationships between hypertension and increasing SDB severity, and demonstrate persistence of statistical significance of these associations even after taking into account measures of obesity.

A causal association between SDB and hypertension is supported by longitudinal data from the Wisconsin Sleep Cohort Study, which showed significant dose-response relationships, such that relative to the reference category (baseline AHI = 0), the ORs for the presence of hypertension at the 4-year follow-up were 1.42 for an AHI of 0.1 to 4.9, 2.03 for an AHI of 5.0 to 14.9, and 2.89 for an AHI of ≥ 15.0.¹⁶⁷ Recent findings from the Sleep Heart Health Study also demonstrated a significant longitudinal relationship between SDB and the development of hypertension among individuals who were normotensive at baseline examination; however, this relationship was attenuated to borderline significance after taking obesity into account.^168,169

Randomized clinical trials have assessed the effects of CPAP treatment on OSA in more than 1600 patients with hypertension.¹⁷⁰ In the majority of studies, the effect on blood pressure has been favorable. The results of two meta-analyses, including studies that were conducted largely in normotensive patients, suggest that the reduction in blood pressure during CPAP therapy is in the region of 1.5 to 2.5 mm Hg^171,172; the magnitude of effect varied from zero up to a 10-mm Hg reduction in blood pressure.^173,174 Regression analysis estimated that 24-hour mean blood pressure would decrease by 1.39 mm Hg for each 1-hour increase in effective nightly use of a CPAP device.¹⁷²

In the most recent meta-analysis of published data, which included 29 randomized controlled trials, mean reductions in daytime systolic and diastolic blood pressures with CPAP (vs no treatment) were 2.6 and 2.0 mm Hg, respectively; corresponding nighttime reductions were 3.8 and 1.8 mm Hg.¹⁷⁵ Another interesting finding of this systematic review was that there was an association between the severity of OSA at baseline and the beneficial effects of CPAP on blood pressure: the higher the baseline AHI, the greater the reduction in systolic BP, suggesting that patients with more severe OSA may benefit the most from CPAP therapy in terms of blood pressure reduction.

The greatest benefits of CPAP therapy for OSA have been seen in patients with resistant hypertension.^{164,176,177,178} In such patients, treatment of OSA with CPAP reduced daytime blood pressure by 6.5 mm Hg, compared with a 3.1-mm Hg increase in untreated patients over the study period.¹⁷⁷

It is important to mention that compliance to CPAP therapy by patients with resistant hypertension plays a key role in achieving optimal blood pressure reduction. Decreases in 24-hour blood pressure were documented in patients with resistant hypertension who used CPAP for at least 5.8 hours per night, but not in those with lower usage,¹⁷⁸ and another study of CPAP treatment for OSA in patients with resistant hypertension reported a significant correlation between hours of CPAP use and decreases in 24-hour systolic and diastolic blood pressure.¹⁷⁹

It may take 3 to 6 months to achieve the maximum reduction in blood pressure associated with CPAP therapy.^180,181 OSA is recognized as the most prevalent risk factor contributing to resistant hypertension,¹⁸² and an American Heart Association (AHA) scientific statement recommends that OSA in patients with resistant hypertension should be treated with CPAP.¹⁸³

Cardiac Arrhythmia

Bradycardia and Atrioventricular Block

Bradycardia and atrioventricular block are commonly observed in association with SDB and most likely are mediated by vagal stimulation accompanying apneas and hypopneas.^184,185 Atrioventricular block has been noted to be frequent in SDB and tends to occur during rapid eye movement sleep and improve with CPAP treatment.^186,187

Atrial Fibrillation

Sleep apnea is highly prevalent in patients with AF. More than 50% of those with persistent AF or with paroxysmal AF with a high AF burden have been shown to have clinically relevant SDB.¹⁸⁸ The prevalence of CSA (as opposed to OSA) in patients with AF is not well described. One study has reported a prevalence of 79% for CSA in pacemaker recipients with permanent AF¹⁸⁹; this was probably a result of a high prevalence of underlying HF and reduced systolic left ventricular function in this study population. A study in AF patients with normal left ventricular systolic function reported prevalence rates of 31% for CSA/CSR and 43% for OSA.¹⁹⁰

Among participants in the Sleep Heart Health Study, moderate-to-severe SDB, as compared with minimal SDB, was associated with a four-fold increased adjusted odds of AF on overnight PSG. Among elderly individuals, stronger associations have been reported between AF and CSA compared with OSA.¹⁹¹

Patients with AF with untreated SDB have a higher recurrence of AF after cardioversion than patients without a known SDB diagnosis or patients with treated SDB.¹⁹² Data from a meta-analysis reported that the risk ratio for recurrent AF after pulmonary vein isolation was 1.25 in patients with, versus without, OSA.¹⁹³ Prevention of obstructive respiratory events in OSA using CPAP reduces the risk of AF recurrence after ablation therapy.^194,195 In a study of 426 patients,¹⁹⁵ CPAP therapy in patients with OSA and AF undergoing pulmonary vein isolation was associated with a higher AF-free survival rate at 12 months (71.9% vs 36.7% in OSA patients not treated with CPAP), a rate only slightly higher than found in the group of patients without OSA. In AF management guidelines, OSA is mentioned as being associated with AF and as a factor contributing to a reduction in the success of ablation procedures.¹⁹⁶ Consideration of screening for sleep apnea in such patients is reasonable, and if OSA is diagnosed, treatment with CPAP may maximize the effectiveness of other rhythm control strategies.

An increased occurrence of AF after coronary artery bypass graft surgery also has been reported to occur in patients with SDB compared with those without SDB, suggesting the importance of untreated SDB in modifying post-cardiac surgery cardiovascular outcomes.¹⁹⁷

Several mechanisms may contribute to the development of arrhythmia (such as AF) in sleep apnea.¹⁹⁸ In animal studies and in clinical observation, the negative thoracic pressure during obstructive respiratory events in particular was identified as the most relevant factor for the perpetuation and initiation of AF. Negative thoracic pressure changes result in increased occurrence of atrial premature contractions, potentially triggering AF episodes.^199,200 In a pig model of OSA, application of negative tracheal pressure during tracheal occlusion, but not tracheal occlusion without applied negative tracheal pressure, reproducibly and reversibly shortened the atrial refractory period and strongly enhanced the inducibility of AF. These arrhythmogenic electrophysiologic changes were largely driven by combined sympathovagal activation, as it could be modulated by sympathetic as well as vagal inhibition.^199,200,201 Additionally, repetitive obstructive respiratory events resulted in an arrhythmogenic structural substrate for AF characterized by local conduction disturbances as a result of increased atrial fibrosis and distribution of connexins in a rat model of sleep apnea.²⁰²

Ventricular Arrhythmia

An increase in ventricular arrhythmias in association with SDB was demonstrated in the Sleep Heart Health Study, which reported that after adjusting for age, sex, BMI, and prevalent coronary heart disease, individuals with SDB had a three-fold increased odds of nonsustained ventricular tachycardia and almost twice the odds of complex ventricular ectopy based on electrocardiographic data from overnight PSG.²⁰³ An attenuated association of SDB and ventricular cardiac arrhythmias was observed with increasing age, with potential reasons including survivorship bias, competing risk factors, and/or age-related alterations in the cardiac conduction system altering risk profiles in older adults.²⁰³ A large community-based cohort of older men, the Outcomes of Sleep Disorders in Older Men Study, identified stronger associations of OSA and hypoxia with nocturnal ventricular arrhythmias compared with CSA.¹⁹¹ Several clinic-based studies also have examined the association between arrhythmias and SDB. In a study of 400 patients with severe SDB, almost half demonstrated cardiac arrhythmias after evaluation by PSG and 24-hour Holter monitoring.²⁰⁴ Two other studies, however, yielded conflicting results as a result of attenuation of associations after adjustment for confounders.^205,206 Improvement in ventricular ectopy with CPAP treatment has been shown in HF patients²⁰⁷ and in general SDB patients,²⁰⁸ with one study showing concomitant improvement in sympathetic activity.²⁰⁷

An AHI of > 20/h was a significant and independent risk factor for incident sudden cardiac death in a study of more than 10,000 patients with HF referred for PSG.²⁰⁹ Coexisting HF and SDB (both OSA and CSA) increase the risk of developing malignant ventricular arrhythmia in patients with implanted cardioverter-defibrillators²¹⁰ (Fig 77–6). Severe OSA also increases the risk of ventricular premature beats, nonsustained ventricular tachycardias, and nocturnal sudden cardiac death.^203,211 It has been shown that an episode of AF or nonsustained ventricular tachycardias was almost 18 times more likely to occur within 90 seconds of an apnea or hypopnea compared with normal breathing.²¹²

Registry data show that treatment of SDB with servo-assisted positive airway pressure (adaptive servo-ventilation [ASV]) in HF patients with implantable cardioverter-defibrillator devices is associated with less use of defibrillatory therapies and improvements in cardiac function and respiratory stability.²¹³ However, in light of the increased mortality in a large randomized trial of ASV in patients with HF and reduced ejection fraction and predominantly CSA (SERVE-HF study), discussed further below, mask-based therapy for CSA should be avoided at the present time.

Coronary Artery Disease

Animal studies support the occurrence of myocardial ischemia in settings of limited coronary artery flow and SDB, with SDB-associated physiologic stressors causing reductions in myocardial oxygen delivery and increased myocardial oxygen consumption in regions of vulnerable myocardium.^214,215 Similar mechanisms for myocardial ischemia have been identified in studies of SDB patients; increased myocardial oxygen consumption associated with increased diastolic pressure and decreased oxygen supply have been observed to occur during peak changes in hemodynamics during the rebreathing phase of the OSA.²¹⁶ CAD patients with OSA have been also shown to have a higher frequency of noncalcified/mixed atherosclerotic plaques, along with more serious stenosis and a higher number of affected vessels than CAD patients without OSA.²¹⁷

The prevalence of OSA in patients with CAD is high (up to 87% in CAD patients referred for coronary artery bypass graft surgery) and is significantly increased compared with healthy controls.^{218,219,220,221,222,223} In a cohort of patients who had undergone revascularization for CAD, the prevalence of OSA was higher than that of obesity, hypertension, diabetes, and AF.²²⁴

Case-control studies have shown associations between CAD and SDB, with one study showing the persistence of this association even after taking into account potential confounding factors.^221,222,225 In a cross-sectional analysis of the Sleep Heart Health Study data, SDB was marginally associated with self-reported coronary heart disease (upper vs lower AHI quartile: OR = 1.27; 95% confidence interval [CI], 0.99-1.62).²²⁶ Prospective analyses showed a 70% increased 8-year incidence of CAD in men younger than 70 years of age in the Sleep Heart Health Study cohort.²²⁷ Over a 7-year study of the Gothenburg Sleep Cohort, CAD was observed in 16% of patients with SDB (defined as ≥ 30 oxygen desaturation events) compared with 5.4% of patients without SDB, with reduced rates of CAD among those with SDB who used CPAP.²²⁸ A Spanish observational study involving solely male patients assessed incidence of fatal and nonfatal cardiovascular events over a 10-year period and reported that patients with untreated severe SDB had a 2.9-fold increased risk of fatal and 3.2-fold increased risk of nonfatal cardiovascular events compared with healthy participants, even after taking into account potential confounders. The authors concluded that, in men, severe SDB significantly increases the risk of fatal and nonfatal cardiovascular events.¹³⁹

In a study in which cardiovascular mortality was prospectively investigated in consecutive patients with known CAD during a follow-up period of 5 years, cardiovascular death occurred in 38% of those with SDB compared with 9% of the non-SDB group (P = .02).¹⁴⁰ In another prospective cohort study involving 408 patients ≤ 70 years old with verified coronary disease who were followed for a median of 5.1 years, there was a significant 70% relative increase and a 10.7% absolute increase in the primary composite end point of death, cerebrovascular event, and MI in patients with SDB defined as an oxygen desaturation index of ≥ 5 (risk ratio = 1.70). Patients with an AHI ≥ 10 had a significant 62% relative increase and a 10% absolute increase in the composite end point (risk ratio = 1.62).²²⁹ These results suggest that individuals with CAD and SDB have a worse cardiovascular prognosis than individuals with CAD without SDB, and may benefit from SDB treatment in addition to more aggressive treatment of their vascular risk factors. The Sleep Apnea Cardiovascular Endpoints Trial is due to report in 2016²³⁰; this international, multicenter, open, blinded end point randomized trial recruited more than 2700 patients with CVD and OSA and is designed to determine the effect of CPAP therapy on cardiovascular end points (ClinicalTrials.gov identifier: NCT00738179). The average duration of follow-up by the time the trial reports results is likely to be more than 4 years.

Few reports have documented the association between SDB events and symptomatic ischemia.²³¹ One study reported a high prevalence of ST-segment depression in patients with SDB during sleep, even in the absence of documented CAD, with a reduction in periods of ST-segment depression occurring with initiation of CPAP therapy.²³² Another study noted that patients with SDB have a higher frequency of cardiac rhythm disturbances and ST-segment depression episodes than controls, and that the ST-segment changes were related to heightened sympathetic tone and sleep fragmentation.²³³ Cardiac ischemia may also disrupt sleep. In one study, more frequent and more severe arousals were observed during periods with myocardial ischemia than during control episodes, suggesting a vicious cycle where apneas may induce ischemia, which then further disrupts sleep.²³⁴

In patients with CAD, SDB treatment with CPAP has been reported to reduce rates of nocturnal ischemia in one nonrandomized study²¹⁶ and reduce rates of cardiovascular death, acute coronary syndrome, hospitalization for HF, and need for coronary revascularization in another observational study.²³⁵

Acute Coronary Syndrome

The prevalence of SDB in the setting of acute coronary syndrome has been reported to be high. Specifically, of 104 patients admitted with acute coronary syndrome to one center in the United States, 66% had mild-to-moderate SDB (AHI ≥ 10), and 26% had moderate-to-severe SDB (AHI ≥ 30), with the predominant apnea pattern being obstructive (72%).²³⁶ In the first few days after an acute MI, the prevalence of moderate-to-severe sleep apnea (AHI > 15) was as high as 55% in one German center.²³⁷ AHI has been shown to be independently associated with less myocardial salvage and a larger infarct size at 3 months,²³⁸ and the presence of OSA appears to inhibit the recovery of left ventricular function after MI.²²³ In the acute setting, the presence of OSA has been shown to be an independent predictor of cardiovascular events in patients with non-ST-segment elevation coronary syndromes in a study from a Brazilian center (OR = 3.4; 95% CI, 1.3-9.0; P = .0002).²³⁹ Cardiovascular event rates over a 5-year follow-up were 37.5% and 9.3% in CAD patients with and without OSA, respectively (P = .018).²³⁹ A randomized trial of CPAP in patients with acute coronary syndrome and nonsleep OSA is being conducted in 15 teaching hospitals in Spain, with plans to recruit more than 1250 patients (ClinicalTrials.gov identifier: NCT01355087).²⁴⁰ Until the results are available (likely in 2017), patients should be managed on an individual basis.

Heart Failure

The Sleep Heart Health Study identified OSA as an independent risk factor for the development of HF,²²⁶ with more impact in men than in women.²²⁷ Prospective data from the Wisconsin Sleep Cohort Study in a cohort of 1131 adults (men and women) aged 30 to 60 years and followed for 24 years show a 2.6-fold increase in the incidence of coronary heart disease and (self-reported) HF, after adjustment for age, sex, BMI, and smoking.²⁴¹

SDB is common in patients with HF, with prevalence rates of 50% to 75%.^242,243 SDB has been documented in patients with both HF with reduced ejection fraction^244,245 or HF with preserved ejection fraction (HFpEF),^246,247 with no difference in prevalence between the two groups,²⁴⁸ and in patients with acute decompensated HF, where the prevalence can be even higher (44%-97%).^249,250,251

The prevalence of CSA/CSR appears to increase as the symptomatic severity of the HF syndrome increases^242,246 (Fig. 77–7), and the severity of CSA/CSR seems to mirror underlying cardiac dysfunction.^252,253,254 Furthermore, CSA is independently associated with a worse prognosis in patients with HF, including increased mortality.^{210,255,256,257,258,259} Approximately 20% to 45% of patients with chronic HF have OSA,^242,243 and the predominant type of SDB in HFpEF appears to be OSA (70%-80%).^246,260 OSA is independently associated with a worse prognosis in HF patients,²⁵⁵ even in those who are receiving maximal and optimal HF therapy, including cardiac resynchronization.²⁶¹

One of the interesting features of SDB in patients with HF compared to general SDB patients is a relative lack of symptoms, especially of daytime somnolence,^262,263 which could contribute to the lack of recognition and detection of SDB in HF patients.²⁶⁴ One possible explanation for a lack of daytime sleepiness in HF patients with SDB is the increased sympathetic nervous system activity in HF patients compared with healthy subjects,^265,266 which is increased even further in the presence of OSA.^7,267 Increased sympathetic stimulation could stimulate alertness to counteract the effects of sleep fragmentation and sleep deprivation.²⁶⁵ A significant inverse correlation between the degree of subjective daytime sleepiness and daytime muscle sympathetic nervous system activity has been documented in patients with HF and OSA.²⁰ Furthermore, patients with HF are often taking a variety of medications that cross the blood-brain barrier, and these could also impact on sleep and SDB.²⁶⁸ One such group of agents is β-blockers, which have been shown to reduce daytime sleepiness and the prevalence of CSA.²⁶⁹

Although effective treatment of HF may improve CSA/CSR,^253,270 its negative prognostic impact persists even in patients who are receiving maximal and optimal HF therapy, including cardiac resynchronization.^210,261 In addition, when present, CSA in acute decompensated HF patients is usually severe (AHI > 30)²⁵⁰ and has been shown to be a predictor of hospital readmission and mortality.²⁷¹ It is important to note that even with optimal medical management, resolution of acute decompensation, and return to baseline cardiopulmonary status, the severity of CSA may not change.^{13,250,272,273} The effect of the specific treatment of SDB in HF is discussed later in this chapter.

Pulmonary Arterial Hypertension

Pulmonary hypertension, defined as a mean pulmonary arterial pressure of > 25 mm Hg at rest or > 30 mm Hg during exercise, is characterized by a progressive and sustained increase in pulmonary vascular resistance that may lead to right ventricular failure. Acute changes in pulmonary arterial pressures have been observed to occur during apneas.²⁷⁴ The primary mechanism likely to be involved with pulmonary hypertension development in SDB is hypoxemia, which reflexively increases pulmonary arterial pressures.²⁷⁵

Several studies have shown pulmonary hypertension in 20% to 40% of patients with SDB in the absence of other known cardiopulmonary disorders.^276,277,278 The pulmonary hypertension associated with SDB appears to be mild to moderate in severity and is thought to be a result of a combination of precapillary and postcapillary factors, including pulmonary arteriolar remodeling and hyperreactivity to hypoxia, left ventricular diastolic dysfunction, and left atrial enlargement.²⁷⁹ In a study involving 220 consecutive patients with moderate SDB, pulmonary arterial hypertension (mean arterial pressure > 20 mm Hg) was found in 17% of patients; however, the degree of pulmonary hypertension was relatively mild, and only 5% of patients had a pulmonary artery pressure > 35 mm Hg.²⁸⁰ Individuals with SDB and pulmonary hypertension tended to be more obese and to have hypoxemia and hypercapnia while awake, and some may have had underlying pulmonary disease.²⁸⁰ Severe SDB is independently associated with pulmonary hypertension in direct relationship with disease severity and presence of diastolic dysfunction.⁸¹ Although measurable changes in the structure and function of the right ventricle have been reported in association with SDB, the clinical significance of these changes is uncertain. Right ventricular failure in SDB appears to be uncommon and is more likely if there is coexisting left-sided heart disease or chronic hypoxic respiratory disease.²⁷⁹

In a study of patients with moderate SDB and mild pulmonary hypertension, CPAP treatment resulted in a significant decrease in pulmonary artery pressure, from an average of 16.8 mm Hg before CPAP to an average of 13.9 mm Hg after 4 months of CPAP, and a significant decrease in total pulmonary vascular resistance from 231.1 to 186.4 dyn/s/cm⁵.²⁸⁰ A randomized crossover trial demonstrated that effective CPAP induced a significant reduction in the values for pulmonary systolic pressure (from a mean of 28.9 to 24.0 mm Hg).²⁸¹ The reduction was greatest in patients with either pulmonary hypertension or left ventricular diastolic dysfunction at baseline. Another study also reported that CPAP was associated with a reduction in pulmonary systolic pressure levels,²⁸² including reduced hypoxic pulmonary vascular reactivity. The clinical relevance of these small changes in pulmonary artery pressure is not known.

Lifestyle Measures

Lifestyle modifications to improve activity and diet are important in the management of SDB. Weight loss significantly reduces AHI in obese patients with OSA; a meta-analysis of seven randomized controlled trials showed that weight loss programs resulted in a mean reduction in AHI of 6.04 events per hour (95% CI, -11.18 to -0.90).²⁸³ However, patients with HF and OSA are less likely to be obese, and the impact in this group is not known.

Patients in whom SDB occurs either exclusively or predominantly in a supine sleep position should be counseled regarding positional therapy. Using a wedge or cushion, or sewing a pocket filled with tennis balls on the back of a pajamas shirt, can discourage sleep in the supine position. Such an approach appears to work for patients with SDB caused by HF.²⁸⁴

Alcohol, sedatives, narcotics, and muscle relaxants should be avoided because they may reduce upper airway muscle tone. Drowsy driving precautions should be reviewed with the patient and advice documented.

General Medical Optimization

For patients with HF, optimal medical management is likely to improve the SDB. This should include the use of diuretics, disease-modifying therapy such as angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, sacubitril-valsartan, β-blockers, and aldosterone antagonists.^{269,270,272,285} Cardiac resynchronization therapy for patients with HF with reduced ejection fraction, a broad QRS complex, and CSA significantly reduces AHI²⁸⁶; meta-analysis of trials investigating the effect of cardiac resynchronization therapy on CSA reported a mean reduction in AHI of 13.05/h (95% CI, –16.74 to –9.36; P < .00001), whereas cardiac resynchronization therapy had no effect on AHI in patients with predominantly OSA²⁸⁷ (Fig. 77–8).

Oral Appliances

In patients with retrognathism, mandibular advancement devices may also significantly improve AHI in OSA.^288,289,290 Oral appliances, worn during sleep and usually fitted by a dentist, may be used to extend the dimensions of the airway and may be effective in select patients, such as those with mild SDB, with positional SDB, or without marked obesity.

Surgery

Bariatric surgery, increasingly used to treat multiple obesity-related comorbidities, also may benefit patients with OSA and a BMI > 35 kg/m².²⁹¹ Upper airway or craniofacial surgical interventions may be an option for SDB treatment; however, they require careful assessment and evaluation including a nasopharyngoscopic examination by an experienced otolaryngologist. Outcomes are often suboptimal, with < 50% of patients achieving a 50% reduction in the AHI and reduction to less than critical thresholds.²⁹² Procedures such as those involving injection of the soft palate resulting in stiffening, radiofrequency ablation, laser-assisted uvulopalatoplasty, and palatal implants can improve snoring, but minimal data exist to support their use for SDB.

Positive Airway Pressure

Obstructive Sleep Apnea

PAP therapy delivered through a nasal or nasal-oral mask stabilizes the airway and thus prevents collapse, and is the standard treatment for SDB.^293,294 There are a variety of different treatment modalities, as illustrated in Fig. 77–9.²⁹⁵

An overnight PAP titration study is required to determine the optimal pressure setting that reduces the number of apneas/hypopneas during sleep, improves hypoxia and sleep architecture, and reduces arousals. Additional beneficial cardiovascular effects of CPAP include increased intrathoracic pressure, reduced left ventricular preload and afterload, and reduced transmural cardiac pressure gradients, all of which can ameliorate impaired cardiac function.^296,297 CPAP improves daytime somnolence, some measures of quality of life, and physical vitality scores.²⁹⁸

Adherence with this therapy is highly variable, with average levels ranging from 50% to 80%,²⁹⁹ but with around 70% of patients still regularly using treatment after 5 years.³⁰⁰ Adherence is positively influenced by patient education, reviewing the health and quality-of-life benefits of SDB treatment, careful selection of PAP masks that best fit the patient, and supportive management of adverse effects such as nasal congestion or dryness and pressure intolerance. Patient perception of symptoms and improvement in sleepiness and daily functioning may be more important in determining patterns of use than physiologic aspects of disease severity.²⁹⁹

In a randomized controlled trial of 55 patients with HF and OSA, nocturnal CPAP for 3 months improved left ventricular ejection fraction (by 5.0% ± 1.0% vs 1.0% ± 1.4%, P = .04) and reduced urinary noradrenalin excretion.³⁰¹ Kaneko and colleagues³⁰² demonstrated that even one night of CPAP lowers systolic blood pressure (from 126 ± 6 to 116 ± 5 mm Hg, P = .02), reduces heart rate (from 68 ± 3 to 64 ± 3 bpm, P = .007), and improves left ventricular end-systolic diameter (from 54.5 ± 1.8 to 51.7 ± 1.2 mm, P = .009) in those with OSA and HF, compared to standard medical therapy. Another study, with echocardiographic and cardiac magnetic resonance follow-up, demonstrated that CPAP improves right ventricular function, left ventricular mass, and pulmonary hypertension after 3 months of treatment. These improvements persisted at 1 year.³⁰³ An observational study (88 patients) of CPAP versus medical therapy for those with HF and moderate-to-severe OSA demonstrated a significantly higher rate of hospitalization or death in the non-CPAP group (hazard ratio = 2.03; 95% CI, 1.07-3.68; P = .03) compared to those treated with CPAP.³⁰⁴ Patients who were not compliant with CPAP also had a higher risk of the composite end point. Two other large registry studies found similar results.^305,306

The 2010 Heart Failure Society of America Comprehensive HF guidelines recommend screening for SDB and CPAP therapy in those with confirmed OSA.³⁰⁷ The 2013 American College of Cardiology Foundation/AHA guidelines state that treating OSA with CPAP in patients with HF does have benefit.³⁰⁸

Central Sleep Apnea

A number of treatments for CSA/CSR have been studied, including oxygen, CO₂, CPAP, bilevel positive airway pressure (BiPAP), and ASV (see Fig. 77–9).

Although it does not trigger inspiration during CSA, CPAP improves CSA/CSR probably by increasing functional residual capacity (and, as a result, oxygen stores), decreasing blood volume in the lungs and upper airway when lying down,³⁰⁹ and reducing hyperventilation via a direct effect on the paravasal J-receptors of the lung.²⁵³ In addition, CPAP reduces preload and afterload and the cardiac transmural pressure and may benefit cardiac function in some patients.¹¹ It may have additional benefits in HF, as positive end-expiratory pressure prevents alveoli collapsing secondary to pulmonary edema and maintains alveoli at a greater diameter, thus reducing the work of breathing. It also increases alveolar recruitment, improves gas exchange, and reduces right-to-left intrapulmonary shunting of blood.

Early small trials of CPAP in CSA with HF demonstrated an improvement in AHI, reduced daytime plasma natriuretic peptide and catecholamine concentrations, and improved left ventricular ejection fraction, and in one small study of 29 patients, there was a trend toward improved survival with CPAP.³¹⁰ A larger randomized controlled trial (the CANPAP study) was designed to evaluate the effect of CPAP on transplant-free survival in patients with CSA and HF.³¹¹ This trial was stopped early after 258 patients had been randomized and followed up for over 2 years; there was no difference in transplant-free survival between CPAP and the optimal medical therapy alone arm. CPAP did result in an improved AHI (-21 ± 16 vs -2 ± 18, P < .001), left ventricular ejection fraction (2.2% ± 5.4% vs 0.4% ± 5.3%, P = .02), and 6-minute walk test distance and reduced plasma noradrenaline concentrations, but this did not translate into a survival advantage. Post hoc subgroup analysis suggested that there was a survival advantage in those in whom the AHI was suppressed by the CPAP therapy to < 15/h,³¹² suggesting a possible role for more efficacious ventilatory techniques, such as ASV.

Some patients with HF and OSA develop CSR during CPAP administration, a finding labeled complex sleep apnea.^313,314,315 BiPAP with backup rate has been shown to be more effective than CPAP at controlling CSA/CSR,³¹⁶ but it can also worsen CSA.³¹⁷ This is because BiPAP (with a backup rate) is designed to provide ventilation and reduce CO₂ levels, which does not ameliorate the hyperventilation and low CO₂ levels associated with CSA/CSR. For this reason, BiPAP is not considered an adequate treatment for CSA/CSR.

However, a high proportion of patients with CSA have residual apnea events despite CPAP therapy, suggesting that a more effective intervention is required.³¹⁸ ASV has been shown to be the most effective intervention for controlling (central) SDB in patients with HF.³¹⁹ ASV is a more sophisticated mode of noninvasive ventilation in which the ventilator increases inspiratory support during hypopnea, withdraws support during hyperventilation, provides mandatory breaths during apnea, and generates background positive airway pressure. Therefore, it is effective in both CSA and OSA and can suppress complex sleep apnea.³¹⁴ Teschler and colleagues³¹⁶ showed that ASV is more effective than CPAP or oxygen therapy in suppressing CSA events and is better tolerated by patients than CPAP (Fig. 77–10). ASV has also been shown by others to be more effective than BiPAP for treating CSA/CSR in HF³²⁰ and is better tolerated than CPAP,³²¹ resulting in improved compliance with therapy, which is an important part of successful treatment.³²²

Small studies of ASV have documented improvements in AHI, sleep quality, quality of life, left ventricular ejection fraction, New York Heart Association class, oxygen uptake, serum natriuretic peptide concentrations, inflammatory markers, and exercise capacity.^{323,324,325,326,327,328,329}

In randomized controlled clinical trials, beneficial effects of ASV treatment of CSA/CSR in HF patients include significant reductions in AHI,^{322,330,331,332,333,334} N-terminal pro-brain natriuretic peptide concentrations,^{322,333,334,335,336} urinary catecholamine release,³³³ and left ventricular end-systolic diameter³²²; increases in 6-minute walk distance³²² and left ventricular ejection fraction^330,331,332; and improved New York Heart Association class.³³¹

Given these beneficial effects, a large randomized controlled trial, SERVE-HF, was undertaken to assess the impact of ASV on hospitalization, life-saving cardiovascular intervention, or death in those with HF and CSA.^337,338 A total of 1325 patients with a left ventricular ejection fraction of 45% or less and moderate-to-severe (predominantly) CSA were enrolled. At 12 months, ASV was highly efficacious at reducing AHI (from a mean of 31.2 events per hour at baseline to 6.6 per hour). Despite the good control of the CSA, there was no difference in the primary end point (of all-cause mortality, unplanned HF hospitalization, or life-saving cardiovascular intervention) between the two groups, and the study demonstrated a higher overall mortality and cardiovascular mortality in those treated with ASV (hazard ratio for all-cause mortality, 1.28; 95% CI, 1.06-1.55; P = .01; and hazard ratio for cardiovascular mortality, 1.34; 95% CI, 1.09-1.65; P = .006) (Fig. 77–11). This trial did not find differences in plasma B-type natriuretic peptide concentration, 6-minute walk test, or health-related quality of life between the two randomized groups. Further analyses of the mode of death and data from the patients who had an implanted device with defibrillator capacity are awaited, but initial results suggest that the excess mortality is driven by an increase in sudden death. There was no difference in deaths from pump failure or admissions to hospital with HF decompensation. Various explanations have been proposed for these results, including chance, a direct toxic effect of PAP on patients with poor left ventricular function and a low pulmonary capillary wedge pressure,³³⁹ and the fact that CSA may be at least partially adaptive for patients with severe HF, as has been suggested previously.¹⁴ Further data will emerge from another randomized trial of an ASV device in patients with HF and reduced ejection fraction and either predominantly OSA or CSA (ADVENT-HF; ClinicalTrials.gov identifier: NCT01128816). In the meantime, the use of ASV for the treatment of predominantly CSA in patients with HF and reduced ejection fraction cannot be recommended. Patients already on ASV should be counseled about the potential risks of continuing with this therapy.

CSA is found in the majority of patients with acute decompensated (as opposed to chronic) HF, is usually severe, and is associated with an increased risk of readmission and death.³⁴⁰ A randomized trial of ASV in this patient group was initiated, but was terminated after the results of SERVE-HF became available (CAT-HF; ClinicalTrials.gov identifier: NCT01953874). The results will be presented in late 2016.

Another area of interest is the use of ASV in patients with HFpEF and CSA/CSR. Early results show that ASV can improve cardiac diastolic function, improve symptoms, and decrease B-type natriuretic peptide concentrations in this group of patients.^341,342 In addition, the proportion of HFpEF patients treated with ASV who were free of cardiac events was significantly higher than that in untreated patients.³⁴³

Oxygen Therapy for Central Sleep Apnea/Cheyne-Stokes Respiration

Oxygen therapy for CSA has been the subject of a few small-scale trials. Its use during sleep reduces the severity of CSA/CSR by approximately 50%, reduces nocturnal norepinephrine levels, and attenuates apnea-associated hypoxemia over time frames ranging from 1 night to 1 month,^344,345,346 but only one study has reported clinical improvements.³⁴⁶ The CHF-HOT trial randomized a small number of patients with CSA and HF to standard medical therapy with or without home oxygen therapy. Meta-analysis of the results from 97 patients demonstrated a decrease in AHI (–11.4 ± 11.0 vs –0.2 ± 7.6, P < .01) and an improvement in left ventricular ejection fraction (36.1% ± 11.8% vs 46.3% ± 16.2%, P = .014) in those with severe CSA treated with home oxygen at 3 L/min via an oxygen concentrator, at least out to 12 weeks.³⁴⁷ There was also an improvement in mean New York Heart Association class, but no overall improvement in ventricular ectopy or plasma catecholamine concentrations. The impact on prognosis is unknown. A meta-analysis of 14 studies concluded that oxygen therapy does reduce overnight desaturations, but prolongs apneas and hypopneas.³⁴⁸

Carbon Dioxide Therapy for Central Sleep Apnea/Cheyne-Stokes Respiration

Administration of CO₂ reduces AHI in CSA, but at the expense of hyperventilation and poor sleep quality and, thus, is not used clinically.³⁴⁹

Experimental Therapies for Sleep-Disordered Breathing

Phrenic Nerve Stimulation

Phrenic nerve stimulation is a new approach to the treatment of CSA/CSR, with initial results showing that it may improve central respiratory events by about 50%.^350,351,352 This device is similar to a pacemaker, with an electrode that stimulates the phrenic nerve via the left pericardiophrenic or right brachiocephalic vein (Fig. 77–12). It can be implanted percutaneously under sedation in the catheter laboratory. The device unilaterally stimulates the phrenic nerve when no impulse is sensed for a predetermined time period, inducing a breath. A nonrandomized study of 57 patients showed a mean reduction of 55% in AHI over 3 months (49.5 ± 14.6 to 22.4 ± 13.6, P < .0001), with the effect being largely confined to prevention of apnea and with little effect on hypopneas, as well as reductions in arousals and oxygen desaturation index and improved quality-of-life indices.³⁵⁰ Device- or procedure-related adverse events occurred in 26% of patients, predominantly as a result of lead displacement. A somewhat larger randomized study has completed recruitment to further evaluate the effect of this technology on the reduction in CSA events but is not powered to determine the effect on hospitalization or mortality (ClinicalTrials.gov identifier: NCT01816776).

Hypoglossal Nerve Stimulation

For those with OSA, a similar device exists that stimulates the hypoglossal nerve in response to apnea and hypopnea. An uncontrolled study has shown a significant mean reduction of 68% in AHI over 12 months in those treated with this stimulator.³⁵³ The impact on cardiovascular outcomes is not known.

Acetazolamide

Two small trials of medical therapy with acetazolamide have been reported, showing reductions in AHI and improved oxygen saturation in patients with HF and CSA, which may be a result of its respiratory-stimulating properties as well as diuretic action.^354,355 A slightly larger (n = 85) randomized study addressing the effect of acetazolamide on the severity of SDB in HF is currently being conducted (Predicting Successful Sleep Apnea Treatment With Acetazolamide in Heart Failure Patients; ClinicalTrials.gov identifier: NCT01377987). Whether furosemide achieves the same effect is unknown, although reduction in pulmonary congestion might be expected to lessen CSA by reducing pulmonary J-receptor stimulation.

SDB causes chronic episodic exposure to intermittent hypoxemia, sympathetic nervous system activation, intrathoracic pressure swings, and sleep fragmentation, which exert profound effects on the heart and vasculature. There is strong evidence for the association of SDB with systemic hypertension, glucose intolerance, obesity, CAD, AF, and HF, and SDB may be an independent risk factor for these conditions. Much of the evidence for the benefit of treatment of SDB comes from observational data sets or small randomized trials, but there is much circumstantial evidence to suggest that the diagnosis and treatment of OSA are worthwhile and may improve cardiovascular prognosis in addition to improving symptoms, including daytime sleepiness. A recent randomized trial of the treatment of CSA with noninvasive pressure support reported an unexpected increase in cardiovascular mortality, largely driven by an increase in sudden death. The explanation for this is unclear, but currently there is no therapeutic imperative to diagnose and treat CSA. Further research is ongoing to establish a stronger evidence base for the diagnosis and treatment of SDB in patients with, or at risk of, CVD.

The author would like to thank Dr. Reena Mehra and Dr. Susan Redline, who contributed to the previous version of this chapter in the 13th edition. The author's salary is supported by the National Institute for Health Research Cardiovascular Biomedical Research Unit at the Royal Brompton Hospital, London.