Non-positive Airway Pressure Therapies for Obstructive Sleep Apnoea and Central Sleep Apnoea

Abstract

Obstructive sleep apnoea and central sleep apnoea are common disorders, particularly in people with congestive heart failure. Continuous positive airway pressure provides transformative benefits for some patients, but is not always well tolerated. Thus, considerable focus is now being placed on alternative therapies for sleep-disordered breathing, including non-positive airway pressure therapy. Advances in pharmacology have led to a variety of options to treat both obstructive and central sleep apnoea. In addition, these approaches may have major benefits from the perspective of preventing the complications of sleep-disordered breathing. In the case of tirzepatide, the Food and Drug Administration recently approved this medication for the treatment of moderate-to-severe sleep apnoea in people with obesity. However, further data are required regarding how best to improve both subjective and objective hard outcomes in people with congestive heart failure and sleep-disordered breathing.

Obstructive sleep apnoea (OSA) is a common condition, estimated to affect up to 1 billion people worldwide.1 People with congestive heart failure (CHF) may be at particular risk, as studies have suggested up to 70–80% of these patients will have OSA and/or central sleep apnoea (CSA).2,3 Nasal positive airway pressure (PAP) therapy is the first-line treatment for OSA because it has been shown in randomised trials to improve symptoms, lower blood pressure and improve quality of life.4–8 However, definitive data regarding cardiovascular benefits of PAP therapy are lacking.9,10 Patients with OSA can experience transformative benefits from PAP treatment, and in experienced hands, the majority of patients will show good adherence to treatment.11,12 However, there is general acknowledgement that alternative therapies may be helpful for patients who are PAP intolerant. In addition, the optimal treatment of OSA for people with CHF is less clear, with definitive data currently lacking.13–16 This review focuses on non-PAP therapies for sleep apnoea, with special emphasis on issues regarding CHF when data are available.

Pathophysiology

The importance of OSA in CHF has been the subject of much discussion. OSA is defined by repetitive collapse of the pharyngeal airway during sleep, leading to intermittent desaturations with re-oxygenation, which can lead to oxidative stress.17,18 In heart failure, overnight rostral fluid shifts from the legs to the neck, and pharyngeal oedema may contribute to upper airway narrowing, at least in theory.19,20 In addition, surges in catecholamines can occur with each respiratory event, which can contribute to the neurohormonal activation thought to be deleterious in CHF.10,21–23

Patients with CHF can also develop CSA in the form of Cheyne–Stokes breathing (CSB).24 CSB is a crescendo–decrescendo ventilatory pattern in which cessations of airflow can occur without respiratory effort. In contrast, in OSA, ongoing respiratory effort occurs during cessations of airflow (Figure 1).25 Both OSA and CSA can lead to recurrent arousals from sleep, which can contribute to paroxysmal nocturnal dyspnoea, particularly in patients with CHF and CSB.

Figure 1: Obstructive versus Central Apnoea.

Complete cessation of respiratory effort and flow is observed in central sleep apnoea; whereas in obstructive sleep apnoea, only cessation of flow occurs, and effort is preserved.

The treatment of OSA and CSA in CHF can reduce the burden of hypoxia and help to suppress catecholamines. However, definitive data regarding the effects of PAP on CHF outcomes are limited. Additionally, in patients with CHF and predominantly central sleep apnoea, the CANPAP and SERVE-HF studies showed possible deleterious effects on outcomes.15,26 More recently, ADVENT-HF provided some reassurance that adaptive servo-ventilation could be used safely in CHF patients, although the results were negative overall for the primary outcomes and most secondary objective outcomes.13,14

In terms of alternative therapies to PAP, the choice depends on many factors and should be an individualised decision. Oral appliances are an option, but can be expensive and are not always covered by insurance.27–29 Surgical and pharmacological therapies are also available for select patients, although each has risks and benefits. In general, we recommend conservative measures for all patients, including diet and exercise, avoiding alcohol at night, maintenance of nasal patency, and avoiding sleep deprivation.

Oral Appliances

Patients can obtain customised oral appliances via qualified dentists that can be used to treat OSA. Most oral appliances are mandibular advancement devices that serve to pull the jaw forward and prevent collapse at the back of the throat during sleep.30 These devices have been shown in randomised trials to improve daytime symptoms and may improve blood pressure.29,31 These devices likely work best in lean and younger individuals with mild-to-moderate OSA disease. A network meta-analysis suggested comparable improvement in blood pressure with oral appliances compared with PAP therapy.32

Continuous PAP is associated with greater reductions in the apnoea– hypopnoea index (AHI) than oral appliances, but adherence to oral appliance therapy is better overall than for PAP therapy, making overall effectiveness of the oral appliance equivalent to PAP therapy.28,33 There is little research on oral appliances in patients with CHF and OSA, but they may be a suitable alternative for patients who do not tolerate PAP therapy.34

Upper Airway Surgery Including Hypoglossal Nerve Stimulation

For select patients, upper airway surgery may have a role in treating OSA.35–37 A wide array of surgical procedures targeting different anatomical sites exist, and careful selection of both the procedure and patient is essential. Adenotonsillectomy can be helpful for patients with hypertrophy; however, this procedure is generally reserved for paediatric patients because the number of adults who have major benefits from this procedure is relatively low.38–41 The uvulopalatopharyngoplasty was commonly performed, although outcome data for this procedure are relatively weak.42,43

Approximately 40% of patients will achieve a good response to therapy (defined by an AHI <20/h with at least a 50% improvement in AHI). However, an argument can be made that adherence is no longer required once the surgery has been performed and thus even incomplete responses may have favourable outcomes compared with medical therapies if they are being used infrequently.37 However, uvulopalatopharyngoplasty has surgical-related morbidity (including nasal regurgitation) and can affect voice and speech, making it undesirable for some patients.

More recently, hypoglossal nerve stimulation (HGNS) received Food and Drug Administration (FDA) approval in 2014.44,45 The technology relies on a surgically implanted device that stimulates branches of the hypoglossal nerve, which are important for innervating the upper airway dilator muscles (Figure 2). HGNS has been shown to improve OSA in select cases.

Figure 2: Hypoglossal Nerve Stimulator Placement.

An implantable pulse generator or neurostimulator is surgically implanted in a subcutaneous pocket in the right upper chest (usually infraclavicular); this device generates electrical pulses synchronised with respiration. A stimulation lead is routed from the chest to the hypoglossal nerve in the neck (usually right side); this lead selectively stimulates branches of the hypoglossal nerve that innervate the tongue protrusor muscles (primarily genioglossus), avoiding detrusor fibres. A respiratory sensing lead is implanted in the intercostal muscles of the right chest wall. This lead detects respiratory effort by sensing thoracic movement, signalling inspiratory phase timing to the pulse generator.

The STAR study in 2014 showed approximately 70% of patients with moderate-to-severe obstructive sleep apnoea who had difficulty accepting or adhering to continuous PAP achieved clinically significant improvement in AHI with this device.44 However, the study was not a randomised trial and did not examine hard outcomes, making additional data imperative. Of note, the study had careful selection criteria, including a BMI <32 kg/m² and absence of complete concentric airway collapse on drug-induced sleep endoscopy, making the results less generalisable to many OSA patients. It also excluded patients with New York Heart Association class III or IV heart failure. Newer devices are being studied in randomised trials, which may offer advantages over the initially approved device.46–48 Positive results from a Phase III randomised trial were recently reported via press release, but a peer-reviewed manuscript is not yet available.49 However, we are not aware of data in CHF patients, making this approach more speculative in these patients.

Medical and Surgical Weight Loss

Bariatric surgery has been extensively studied to facilitate weight loss.50 Considerable data show that surgically achieved weight loss will lead to marked improvements in patients with OSA and obesity.50–52 However, in some studies, the improvement in AHI is relatively modest. Moreover, some patients regain the weight over time, and in others fat redistribution can occur with reappearance of OSA in longer-term follow-up studies.53,54 One randomised study showed favourable outcomes with PAP therapy compared with bariatric surgery, although further data are needed.55 However, due to the major advances in pharmacotherapy for obesity, surgical therapy for weight loss is becoming less frequent.

The SURMOUNT-OSA study randomised patients with moderate-to-severe OSA and obesity to tirzepatide, glucagon-like peptide receptor 1 agonist, with gastric inhibitor peptide (or glucose-dependent insulinotropic polypeptide) or placebo. The study involved two arms, one using PAP therapy and the other in patients who were not using PAP therapy. In both arms of the study, tirzepatide was superior to the placebo for the primary outcome of AHI change.56 In addition, secondary outcomes were prespecified and controlled for multiple comparisons, including systolic blood pressure, sleep apnoea-specific hypoxic burden, high-sensitivity C-reactive burden, patient-reported outcomes and body weight.56,57 The improvements in systolic blood pressure were up to 9.5 mmHg, making tirzepatide quite a powerful intervention. The observed improvement in bodyweight was 18–20% at 1 year (Table 1).

Table 1: Tirzepatide for Obstructive Sleep Apnoea.

Outcome	Trial 1 (Treatment Difference)	Trial 2 (Treatment Difference)
Per cent change in AHI (events/h)	-50.7% (-47.7)	-58.7% (-56.2)
Per cent change in weight	-17.7% (-16.1)	-19.6% (-17.3)
CRP concentration	-1.4 (-0.7)	-1.4 (-1.0)
Hypoxic burden change (%min/h)	-95.2 (-70.1)	-103.0 (-61.3)
Systolic blood pressure	-9.5 (-7.6)	-7.6 (-3.7)

The SURMOUNT-OSA study involved two arms, one using positive airway pressure therapy and the other not. Both parallel arms involved randomising patients to receive either tirzepatide – glucagon-like peptide receptor 1 agonist with gastric inhibitor peptide (or glucose-dependent insulinotropic polypeptide) – or a placebo. Tirzepatide was superior to the placebo for the primary outcome of the apnoea–hypopnoea index.56 Secondary outcomes included improvement in systolic blood pressure up to 9.5 mmHg; the observed improvement in body weight was 18–20% at 1 year. AHI = apnoea–hypopnoea index; CRP = C-reactive protein.

Although CHF patients were not included in SURMOUNT-OSA, a subsequent study showed a lower risk of a composite of death from cardiovascular causes or worsening heart failure, and improved health status with the use of tirzepatide in patients with HF with an ejection fraction of at least 50% and obesity.58 Studies are ongoing regarding tirzepatide effects on hard cardiovascular outcomes. Based on the results of SURMOUNT-OSA, tirzepatide received FDA approval for treatment of moderate-to-severe OSA in people with obesity in December 2024. Tirzepatide does have considerable cost and gastrointestinal side-effects, leading to some challenges with its widespread use.59

Pharmacotherapy

Apart from tirzepatide, no pharmacotherapy is currently approved for the management of OSA, although many have been studied.

Acetazolamide is a carbonic anhydrase inhibitor that reduces proximal tubular sodium and bicarbonate reabsorption, and has been used as a diuretic in CHF patients.60–64 The addition of acetazolamide to standardised loop diuretic therapy has been shown to lead to greater diuresis in acute decompensated heart failure patients.65 Acetazolamide also induces bicarbonaturia, thereby causing a hyperchloraemic metabolic acidosis that increases ventilation, and it has also been used in central sleep apnoea.66

A meta-analysis of short-term acetazolamide for OSA and CSA patients showed some modest associated benefits, but more rigorous studies with long-term follow-up are needed to better understand the role of acetazolamide in patients with CHF and sleep apnoea.62 Studies are ongoing regarding the definitive role of acetazolamide with and without combinations of other therapies in CHF patients with regard to hard cardiovascular outcomes.62–64,67–69

Atomoxetine/oxybutynin drug combination has been reported to have some efficacy for patients with OSA.70 Atomoxetine is a norepinephrine reuptake inhibitor, while oxybutynin is an anticholinergic; both agents are thought to increase genioglossus and pharyngeal muscle activity. The MARIPOSA study randomised 211 patients with mild-to-severe OSA to atomoxetine/aroxybutynin at varying doses or a placebo, and showed somewhat modest improvements in the AHI with drug therapy over 1 month compared with the placebo.71 More definitive Phase III randomised trials are now ongoing.

Atomoxetine/oxybutynin can have important side-effects, including tachycardia and hypertension, and while we are not aware of data in CHF patients, we caution use in this context until more safety data are available. A recently explored combination therapy with acetazolamide and atomoxetine/oxybutynin in 19 patients with moderate-to-severe OSA found no added benefit of the combination in reducing the AHI compared with atomoxetine/oxybutynin alone.72 Top-line results from a positive Phase III study were recently reported via press release.73 However, a peer-reviewed manuscript is not yet available.

Sodium-glucose cotransporter 2 (SGLT2) inhibitors have been shown to have cardioprotective and renoprotective effects in patients with heart failure. Both the DAPA-HF and EMPEROR-Reduced trials showed a reduction in the combined risk of cardiovascular death or hospitalisation in patients with heart failure with reduced ejection fraction.74,75 The EMPEROR-Preserved and DELIVER trials showed reductions in composite death or HF events in patients with heart failure with preserved ejection fraction.76,77

SGLT2 inhibition may also reduce cardiovascular risk in the context of OSA. The VERTIS CV trial showed ertugliflozin, a SLGT2 inhibitor, nearly halved the incidence of OSA in patients with type 2 diabetes and cardiovascular disease.78 In a post-hoc analysis of the EMPA-REG OUTCOME trial, empagliflozin, another SLGT2 inhibitor, showed more pronounced weight loss in individuals with comorbid OSA; discontinuation of therapy resulted in a higher incidence of newly diagnosed OSA.79 A small randomised controlled trial showed a significant decrease in AHI, improvement in SpO₂ nadir and a reduction in Epworth Sleepiness Scale scores in patients with type 2 diabetes and OSA treated with dapagliflozin, also a SGLT2 inhibitor, combined with metformin compared with those receiving glimepiride and metformin.79 SGLT2 inhibitors may reduce AHI and cardiovascular mortality by preferentially reducing central obesity and epicardial fat, and ameliorate nocturnal fluid shifts via natriuresis.80,81

Positional Therapy

Approximately 50% of OSA patients have improved breathing pattern in the lateral posture as compared with supine.82 Commercial products have been developed to facilitate avoiding supine sleep in patients with supine predominant OSA.83 However, there is a lack of clinical trials evaluating clinical outcomes with the various commercial devices available, although a few studies suggest non-inferiority of this approach in improving sleepiness compared with PAP therapy. CSB may also improve with avoidance of supine sleep, although outcome data are lacking.84 Patient preferences must be considered since adherence to this approach in longer-term follow-up studies is relatively poor.

Phrenic Nerve Stimulation

Phrenic nerve stimulation has been FDA-approved for treatment of central apnoea. Phrenic nerve stimulation can serve to normalise breathing patterns in people with CSA; however, it does not have established utility in OSA. A pivotal study showed improvement in AHI using this.85–87 Subsequent studies have examined cardiac function in people with CHF and shown some improvements in left ventricular ejection fraction via normalising breathing patterns.88 Phrenic nerve stimulation has been used in various forms of central apnoea, including heart failure with reduced ejection fraction, heart failure with preserved ejection fraction and opioid-induced central apnoea, as well as some neurological conditions.

Oxygen

Supplemental oxygen improves oxygen saturation in OSA, but can prolong the duration of an apnoea–hypopnoea episode.89,90 In a randomised control trial, continuous PAP, but not supplemental oxygen, resulted in significant reductions in blood pressure in patients with cardiovascular disease or multiple cardiovascular risk factors and moderate-to-severe OSA.91

The LOFT-HF study was designed to study the effect of nocturnal oxygen in patients with heart failure and central sleep apnoea, but was terminated during the COVID-19 pandemic. The study was under-powered due to poor enrolment during the pandemic. However, a harmful safety signal was noted with oxygen therapy, suggesting deleterious outcomes with this approach. These data are currently in peer review, but have been presented at national meetings. Thus, oxygen alone is not currently recommended to treat sleep-disordered breathing in CHF.

Glossal Neuromuscular Electrical Stimulation

Non-invasive electrical stimulation devices have been developed with the aim to increase the tongue muscle tone for patients with OSA and primary snoring.92–94 Unlike other therapies for sleep-disordered breathing, it is used during daytime wakefulness. One device (eXciteOSA, Signifier Medical Technologies) delivers subtle electrical stimulation through a mouthpiece worn for 20 minutes daily over an initial 6-week period, followed by maintenance therapy. In a study of n people with mild OSA, a total of 50% of those with mild OSA had a ≥50% reduction in AHI, but the efficacy of neuromuscular electrical stimulation in moderate-to-severe OSA is not yet established, and long-term outcomes are unknown.95 Adherence is generally high with this device, probably due to the convenience of daytime use and lack of interference with sleep. Transient tongue tingling or excess salivation were reported in this study.

Conclusion

Given suboptimal adherence to PAP therapy, and the potential deleterious effects of PAP therapy in patients with heart failure with reduced ejection fraction and CSA, non-PAP therapies in patients with heart failure may be desirable. There is growing availability of non-PAP therapies for OSA and CSA, but their role in patients with heart failure remains poorly studied. Further research exploring non-PAP therapies in patients with heart failure is needed, but is challenged by the heterogeneity of SDB in these patients. Phenotypic characterisation of SDB in patients with heart failure is essential. Additionally, combination modalities of therapies may be necessary to treat these patients optimally.

Funding Statement

ZOLL Respicardia funded the article processing charges for the articles in the special collection ‘Sleep-disordered Breathing in Heart Failure’.