Congestive heart failure (CHF) is a highly prevalent chronic disease, affecting around 2 % of the general population (5–6 million Americans) and up to 10 % of those above age 65 [26]. CHF is the most commonly recognized cause of CSA. Although the existing literature is limited by referral and participatory bias, it is estimated that in heart failure patients with reduced left ventricular ejection fraction (<45 %), as much as 31 % have CSA [27–29]. In those patients with preserved ejection fraction (diastolic dysfunction), the prevalence of CSA may reach up to 23 % [30].

Classic signs and symptoms such as paroxysmal nocturnal dyspnea (PND) and orthopnea in those with heart failure might be explained in many instances by CSA-CSB. Even though patients with heart failure and sleep-disordered breathing do not commonly report sleep-related complaints such as excessive daytime sleepiness, presenting clinical characteristics of patients with CSA-CSB may be otherwise undistinguishable from those with OSA [31]. Patients in CHF with CSB are more likely to have atrial fibrillation and poorer functional status (NYHA class), supporting the notion that CSB is a consequence of progressive heart failure and possible indicator of higher morbidity and mortality [29]. Although found more frequently in CHF patients, CSB is not pathognomonic as it can also be observed in stroke and in chronic renal failure patients [32].

Non-hypercapnic central apnea in heart failure patients with ventricular dysfunction are the result of the interaction of the following factors: (1) low awake steady state of PaCO₂, (2) lack of increase of PaCO₂ at sleep onset, with subsequent reduced difference between the apnea threshold and the sleeping eucapnic level (reduce PCO₂ reserve), (3) increased response to PaCO₂ changes (high loop gain system), and (4) impaired cerebrovascular reactivity to CO₂.

The classification of sleep disorders as obstructive sleep apnea (OSA) or CSA is important when determining treatment options in CHF. In CSA with CSB, the polysomnography during diagnostic or positive airway pressure titration should show at least 5 or more central apneas and/or central hypopneas per hour of sleep, representing more than 50 % of the total number of apneas and hypopneas. The characteristic periodic breathing pattern of heart failure, known as Cheyne–Stokes breathing, is defined polysomnographically by cycles of ≥3 consecutive central apneas and/or central hypopneas separated by a crescendo–decrescendo shape tidal volume [40]. It is commonly found during transition from wakefulness to NREM sleep (N1 and N2), dissipating in N3 and REM. The cycle length, measured from the beginning of a central apnea to the end of the next crescendo–decrescendo respiratory phase, averages ≥40 s, and it likely reflects a prolong circulatory time (high mixing gain) [8]. Central apneas that occur within a run of CSB are reported to be associated with less marked oxygen desaturation than similar obstructive and mixed apneas [41]. Arousals, usually located during the ventilatory period, have not shown to have any critical role in termination of apneas or protective mechanism as in obstructive sleep apnea [42] (Fig. 33.4).

Optimization of cardiovascular pharmacotherapy should be the first intervention in the management of CSA due to CHF. The use of beta-blockers, ACE-inhibitors, angiotensin II receptor blockers (ARB), aldosterone antagonists, and diuretics will have a salutary effect in the neurohormonal physiopathology of heart failure, with potential benefits in Cheyne–Stokes breathing. For example, β-blockers are known to influence hypoxemic chemosensitivity (controller gain) [43]. Also, ACE-inhibitors and diuretic therapy may influence the occurrence of CSB by lowering the intracardiac filling pressures [44]. Despite current efforts to optimize cardiovascular therapy in severe CHF, no significant change of Cheyne–Stokes prevalence could be appreciated in small observational studies [29, 45, 46].

Even though not commonly used, other pharmacological therapies are theophylline and acetazolamide. Acetazolamide, by respiratory stimulation and diuretic effect, may decrease the plant gain and widen the difference between the sleep eucapnic levels and the sleep apneic threshold [47]. Clinically, this translates into a reduction in AHI and improvement in patient’s perception of sleep quality [48, 49]. However, side effects associated with acetazolamide, such as paresthesias, tinnitus, metabolic acidosis, electrolyte imbalance, and dizziness, should be considered at time of prescription. Theophylline, by competing with adenosine at a central level, stimulates respiratory centers resulting in a 51 % reduction in the apnea–hypopnea index, mostly because of a reduction in the number of episodes of central apnea. It also shows a decrease in the duration of arterial oxyhemoglobin desaturation during sleep [49–51]. However, the benefits of theophylline should be weighed against the narrow therapeutical range and stimulant effect.

Because the pathophysiology of Cheyne–Stokes breathing is thought to be related to increase chemosensitivity to PaCO₂ at the central controllers (high controller gain), avoiding relative hypocapnia may stabilize the respiratory system. This can be reached by CO₂ supplementation (gas modulator) or by increasing anatomic dead space (DS). Low CO₂ supplementation can be bled from a positive airway pressure gas modulator into a positive airway pressure device. However, limitations relate to cost of the gas modulator, supply of CO₂, and side effects of treatment have affected the practical application of this therapy in the clinical field [52]. Increasing physiologic dead space during positive pressure therapy has been reported to control CSAs in those with CHF [53, 54]. Low levels of CO₂ inhalation mildly raise the sleep eucapnic level, increasing the difference with the PaCO₂ apneic threshold. This is achieved by the sequential connection of a non-vented oronasal mask to 4–6 in. of tubing (rebreathing space), an exhalation valve, and a conventional positive pressure circuit. This expiratory rebreathing reservoir represents an increase of PaCO₂ in the order of 2–3 mm Hg. Unfortunately, pressure leaks and sleep fragmentation are common [55]. When compared to ASV therapy (discussed later) in patients with CHF, DS and ASV caused a similar reduction in the apnea–hypopnea index (AHI), but total sleep time was significantly decreased by DS, attributed to a high arousal index and disruption of sleep architecture [56].

Uncontrolled studies of oxygen supplementation via nasal cannula report a decrease in the AHI, as well as an increase the left ventricular ejection fraction (LVEF) at 3–12-month follow-up without clinically significant adverse effects [57, 58]. These effects may be achieved by different pathways: (1) by reducing the controller gain, (2) by increasing the difference between the eucapnic PaCO₂ level and the PaCO₂ at apneic threshold, and 3) by increasing cerebral CO₂ levels [59, 60]. Even though oxygen supplementation cannot replace CPAP treatment benefits, oxygen supplementation could be considered in those patients with poor compliance with noninvasive ventilation [25].

Several assistive devices modalities have been studied in patients with CHF. Continuous positive airway pressure (CPAP), bilevel positive airway pressure without backup rate (BPAP-S) and with respiratory backup rate (BPAP–ST), and adaptive servo-ventilation (ASV) have been investigated.

There are a number of beneficial effects on the cardiovascular system when continuous positive airway pressure is applied to the patient with CHF. CPAP can affect the response of the respiratory control system by several pathways. It may decrease the circulatory time (mixing gain) and increase the difference between the eucapnic sleep levels and the apneic threshold. Positive airway pressure may also increase the functional residual capacity (decrease in plant gain) and overcome resistance to airflow in the congested upper airway.

Despite these effects, a large multicenter controlled trial named the Canadian Continuous Positive Airway Pressure for Patients with CSA and Heart Failure Trial (CANPAP), failed to show survival benefit with CPAP treatment in advanced HF and CSA at interim analysis, leading to early termination of the study. Worth noting in this study, CPAP implementation showed attenuation of CSA, improvement in nocturnal oxygenation, and an increase in the ejection fraction [61]. This negative result in survival benefit might be attributed to a lack of PAP titration to achieve a therapeutical reduction in AHI or limited compliance with the device. The authors suggested the lower than expected prevalence of CSA led to under-powering, possibly as a result of better medical management in the era of beta-blockers. However, in a subsequent subgroup analysis, CPAP showed positive effect on LVEF and transplant-free survival when CPAP therapy was able to normalize the apnea–hypopnea index to less than 15 events per hour [62, 63]. Subsequent studies have also shown CPAP to increase the LVEF by 6 % and decrease the AHI between 21/h [95 % CI 17–25] and 30/h [95 % CI 23–37] [25]. In conclusion, CPAP may be appropriate therapy for selected patients in whom respirator events are controlled [25].

There is paucity of publications regarding the impact of bilevel positive pressure ventilation in patient with CHF. The data available up to now discourage the use of BPAP-S mode in this population, as it may aggravate central apneas/periodic breathing by hyperventilation. It has been shown that CPAP and BPAP-ST are equally effective in lowering the AHI and NYHA class [64]. BPAP-ST could be an effective alternative to those patients with high residual AHI while on CPAP, benefiting from an increase in LVEF of up to 12.7 ± 10 % [65].

The adaptive servo-ventilation is a feedback control system targeting minute ventilation or peak respiratory flow, adding a component of ventilation that is anticyclic to the patient’s own respiratory drive periodicity. Short-term studies have shown ASV to effectively suppress the Cheyne–Stokes breathing pattern seen in CHF patients with CSA, with improvements in some sleep-related outcomes, as outlined below. When compared to CPAP, compliance may be better with ASV [66]. However, as final results of ongoing large clinical trials are awaited, and in the context of an unexpected increase in mortality associated with ASV as detailed below, the role of the device in the management of CSA in patients with CHF remains to be defined.

Patients with systolic CHF-CSB on treatment with ASV show improvement in polysomnographic parameters (normalization of number of respiratory events), cardiac function (left ventricular ejection fraction, LVEF, and N terminal pro B-type natriuretic peptide NT-proBNP), NYHA functional class, and cardiopulmonary exercise tolerance parameters (VO₂-ATOxygen consumption at anaerobic threshold or peak exercise, VO₂ peak, and 6-min walking distance) [67]. One observational study reported a decrease in the number of cardiac events (cardiac death and rehospitalization) at 6-month follow-up [68, 69]. Even in cases of coexistence of obstructive sleep apnea (OSA), CSA, and Cheyne–Stokes breathing (CSB) in patients with and without heart failure, ASV reduced the central apnea hypopnea index and BNP levels significantly more effectively as compared with CPAP over an eight-month and twelve-month follow-up period [70, 71]. A similar profile of improvements has also been shown in patients with diastolic CHF-CSB on ASV, as well as CHF patients with obstructive and CSA occurring concurrently within the same night, independent of the severity of the sleep-disordered breathing [72–74]. Limited published data are available to infer the impact of ASV on mortality [69, 74]. Recently, preliminary results of the SERVE-HF trial, which assessed the effects of treatment of CSA with ASV on mortality and morbidity in patients with symptomatic chronic heart failure (NYHA 2–4) with reduced ejection fraction (LVEF ≤ 45 %), showed an increased risk of cardiovascular mortality for those treated with ASV in comparison with those with best medical care alone. The increased risk appears to be greatest in those with pure CSA with more severe ventricular dysfunction. However, additional analysis of the phase IV SERVE-HF data and further longitudinal studies are needed to accurately identify the long-term impact of ASV in CSA-CHF patients with varies left ventricular ejection fractions, as well as those with preserved ejection fraction.

There are few published studies with direct comparison of ASV with CPAP, BPAP-ST, and oxygen supplementation. At one night study, ASV suppresses CSA and/or CSB (CSA-CSB) in heart failure and improves sleep quality better than CPAP or 2 L/min of oxygen supplementation. In comparison with BPAP-ST, ASV performs better in CSB-CHF but equivalent to CPAP [75, 76].

Some smaller case series show improvement in the AHI when atrial overdrive pacing (AOP) is used to improve cardiac function in patients with CSA-CHF in comparison with those without AOP [77]. This may be attributed to an increase in cardiac output with subsequent decrease in pulmonary wedge pressures and circulatory time. Similar to AOP, cardiac resynchronization therapy (CRT) reduces the AHI without altering sleep stages [78]. When both AOP and CRT are combined, a minor additional improvement in the AHI is obtained by decrease in the central AHI [79]. Of note, preliminary data from unilateral transvenous phrenic nerve stimulation in patients with CSB–CHF showed a trend toward stabilization of breathing and improvement in oxygen saturation [80].

In those patients with CSA secondary to impaired cardiac function from valvular disease, surgical treatment has been shown to improve sleep-disordered breathing. Even though improved, CSB may persist even after post-transplant normalization of cardiac function [81].

A projected 201.9 million opioid prescriptions were dispensed in the USA in 2009 [82]. As the quantity of opioid prescription in the USA has grown rapidly in recent years, the number of patients on opioids presenting to the sleep clinic for evaluation of sleep-related breathing disorders has also increased. It is estimated that the prevalence of CSA among opioid users, on at least six months therapy, is about 24 %. In methadone maintenance therapy, CSA is identified in up to 30 % of patients [83]. A direct relationship between the total daily dose of opioid and the AHI is found only among methadone users and not among other opioid users including oxycodone, hydrocodone, morphine, hydromorphone, and tramadol [84]. Sleep-disordered breathing has been reported to be reversed with discontinuation of methadone [85].

Opioid-related breathing disorders in sleep can present in the form of obstructive sleep apnea, CSA, alveolar hypoventilation, or a combination of them. Although symptoms of excessive daytime sleepiness, sleep fragmentation, and insomnia are common complaints, it is difficult to disentangle independent effects of the sleep disorder from the influences of the underlying pain syndrome and non-sleep-related influences of narcotics.

Opioids are naturally occurring or synthetic agents which bind to a class of four G protein-coupled receptors in the central and peripheral nervous systems and respiratory tract, leading to decreased neuronal excitability [86]. These receptors fall into four classes: δ, μ, nociception/orphanin, and the κ receptor. Opioid medications, mimicking endogenous ligands (endorphins, enkephalins, dynorphins, etc.), act through these receptors at different levels of the peripheral and central nervous system. Ligands that stimulate μ and κ receptors particularly suppress central respiratory pattern generation, resulting in decreases in respiratory rate and tidal volume. Both naturally occurring and synthetic opioids may exhibit preferential receptor affinity and may act as agonists, mixed agonists/antagonists, or antagonists.

Animal research, supported by limited work in humans, describes opioids acting on medullary respiratory neurons with the suppression of respiration rate and respiratory drive (pre-Botzinger complex), central chemoreceptors’ response to hypercapnia, peripheral response to hypoxemia (glomus cell of carotid body), and depression of the arousal system. Generally, in humans, opioids tend to decrease hypoxic and hypercapnic ventilatory responsiveness. Opioid agonists, particularly fentanyl, tend to decrease upper airway muscle tone and the compensatory response to resistive loading. In addition, some agents have been associated with increased rigidity of accessory respiratory muscles, further limiting ventilation.

These influences suggest that opioids may increase susceptibility to sleep-related breathing disorders. The effect of opioids on human ventilatory control is influenced by dose and ligand specificity. At lower doses, respiratory depression occurs secondary to a decrease in tidal volume, and at higher doses, secondary to a decrease in respiratory rate [86]. A decreased ventilatory response to hypercapnia/hypoxia also occurs [83].

Polysomnographic effects on sleep during acute opioid administration include decreases in slow-wave and REM sleep with decreased sleep efficiency and increases in arousals and N1 and N2 sleep. Other studies have shown a decrease in slow-wave sleep with increased N2 sleep and no change in sleep efficiency or total sleep time in patients receiving either morphine or methadone during polysomnography [87]. During chronic opioid administration, the decreases in slow-wave and REM sleep may normalize with improvement in sleep efficiency [88].

As mentioned in the pathophysiology section, a decreased ventilatory response to hypercapnia/hypoxia will develop CSAs in two characteristic patterns of breathing: the cluster period breathing pattern and the ataxic/Biot’s breathing pattern. The cluster breathing is characterized by cycles of hyperventilation with tidal volumes of stable amplitude, separated by central apneas of variable duration. The ataxic/Biot’s breathing pattern is characterized by variable amplitude and rate of tidal volume as well as variable central apnea duration (Fig. 33.5)

The literature assessing treatment of CSAS due to drugs or substance is limited. As the effect of opioids on central apneas is probably dose-dependent, discontinuation or a decrease in opioid dose to the lowest tolerated by the patient is recommended [89]. Only limited data support use of PAP devices. CPAP may reduce the AHI, generally by controlling obstructive respiratory events, but not CSAs. In fact, the use of CPAP may aggravate CSA [90]. Regarding intervention with ASV, there are conflicting data from small studies about its role in this group of patients. Further study is needed before routine use of PAP devices can be recommended in the setting of chronic opioid use [90, 91].

After a high-altitude ascend to approximately 3500 m above sea level or greater, most healthy individuals will develop a periodic breathing pattern driven by acute alveolar hypoxia [92–94]. This typically develops during ascension itself, immediately after or during the acclimatization period. Development of CSA due to high altitude will depend not only on altitude itself, but also on speed of ascension as well as on individual variability to hypoxic responsiveness.

Arousal events increase with altitude. This increase in total arousals will translate into sleep fragmentation which may result in subsequent complaints of fatigue and excessive daytime sleepiness [95, 96]. Despite the fact that the number of arousals is higher in periodic breathers at different altitudes, periodic breathing pattern itself is not associated with a poorer subjective sleep quality when compared to non-periodic breathers [92, 97].

The major effects of high altitude on humans relate to the changes in barometric pressure (PB) and its consequential changes in the ambient pressure of oxygen (O₂). Since the fraction of inspired oxygen remains constant at approximately 20.93 % during ascension, the most important determinant of inspired PO₂ and subsequent arterial PO₂ (PaO₂) at any altitude is the barometric pressure. For example, at an altitude of 3000 m, the barometric pressure and inspired PO₂ are only about 70 % of that at sea level. During ascension, as arterial PO₂ falls due to a low atmospheric pressure (hypobaric hypoxia), ventilation will be stimulated, with resultant hypocapnia and respiratory alkalosis. Despite high individual variability in acclimatization to the increasing hypoxia at high altitude, sustained hypoxic stimulus during ascension will trigger a periodic breathing pattern in clusters during the night [97]. It is the hyperventilation triggered by hypoxia, with subsequent fall in PaCO₂ below the apneic threshold, that results in central apneas [98]. With breathing cessation, PaCO₂ increase will trigger a subsequent hyperventilation period based on high tidal volumes, in which the PaO₂ rises and PaCO₂ falls to near wakefulness level. In high altitude, periodic breathing is the result of the hypoxia-mediated increase in controller gain, with subsequent narrowing of the difference between eupneic PaCO₂ and apneic threshold, as well as the oscillation in cerebral blood flow, which elevates the cerebrovascular responsiveness to hypercapnia and hypocapnia [99–101]. Even though acclimatization to high-altitude results in an overall increase in saturation of oxygen, periodic breathing may increase in duration and persist in time, consistent with progressive increase in loop gain of the respiratory control system [96, 102, 103].

In healthy individuals, sleep architecture changes at high altitude with concomitant development of periodic breathing. When altitude starts increasing from 3500 m above sea level, duration in N1 stage increases, at the same time slow-wave sleep decreases. However, the time spent in REM sleep is well preserved along ascension up to 5000 m above sea level. Total arousal events increase as altitude increases, as a result of an increase in spontaneous arousals, as well as an increase in arousal related to periodic breathing and/or obstructive sleep apnea [95, 97, 104]. A periodic breathing pattern can develop at any altitude above 3500 m. In comparison with those individuals without periodic breathing, periodic breathing subjects experience an increase in the total arousal index, without a disruption of sleep architecture or sleep oxyhemoglobin saturation [92, 97].

At high altitude, periodic breathing is characterized by recurrent central apneic events in cycles of 12–34 s. It is usually associated with mild oxygen desaturation. Worse in NREM sleep, it stabilizes during REM sleep. As noted in different studies, there is high individual variability in intensity of periodic breathing to hypoxia at any certain altitude [94]. Because it is considered an adaptation to high altitude, there are no established polysomnographic criteria regarding the frequency of apneic events above which it is considered abnormal.

Slow ascension and acclimatization are crucial interventions to ameliorate the development of sleep and breathing disorders. By physiologic adaptation of oxygen content and carrying capacity to high altitude, slow ascent can decrease respiratory drive triggered by hypoxia. Unfortunately, at elevations above 3500–4000 m, acclimatization does not restore normal sleep, even for healthy individuals born at high altitude [96, 105].

At this time, limited data support pharmacological interventions, such as theophylline, temazepam, and acetazolamide for the prevention and treatment of high-altitude periodic breathing. If pharmacological intervention is contemplated, side effects and limitations of the intervention should be cautiously considered. Theophylline and acetazolamide have being shown to normalize the sleep-disordered breathing at high altitude. Different from theophylline, acetazolamide significantly improves basal oxyhemoglobin saturation during sleep, with no major side effects. The effect of temazepam on periodic breathing, overnight SaO₂, and next-day cognitive performance has also been studied. Even though temazepam has shown to reduce periodic breathing, it has also been associated with a small but significant decrease in overnight SaO₂. If pharmacological intervention is considered, theophylline and acetazolamide are recommended to be started 3 days before ascension [106]. On the other hand, temazepam is recommended to be taken on two consecutive nights soon after arrival at 5000 m [107].

Since all of the deleterious effects of high altitude are caused by the low inspired PO₂, supplementary oxygen can improve the inspired PO₂, thereby decreasing respiratory drive and ensuing central apneas. Unfortunately, implementation of this intervention is challenging at high altitude, and data in oxygen supplementation are limited to intervention in acute high-altitude illness such as acute mountain sickness [108, 109].