Obstructive sleep apnoea (OSA) affects 26% of adults and is associated with multiple cardiovascular comorbidities including hypertension, diabetes, congestive heart failure (HF), AF, and coronary artery disease.1 In this condition, airway collapse disrupts breathing, which, in turn, can fragment sleep, leading to excessive daytime sleepiness. OSA has been found to have a higher prevalence in patients with chronic HF compared to the general population, estimated to be between 12% and 43% of patients.2
In contrast, central sleep apnoea (CSA) is driven by hyperventilation or hypoventilation as the main physiological mechanisms.3 One of the more prevalent forms of CSA associated with HF is Cheyne-Stokes respiration (CSR), a cyclical crescendo-decrescendo breathing pattern of hyper- and hypoventilation. CSR is estimated to occur in 30–40% of HF patients.3 Intermittent hypoxia, commonly seen in sleep apnoea of both types, can increase sympathetic tone over time and worsen HF symptoms, leading to increased hospitalisations, worse quality of life, and increased rates of cerebrovascular and major adverse cardiac events.4
In HF patients, untreated sleep disordered breathing (SDB) can worsen HF and is associated with increased mortality.5,6 The pathophysiology of how SDB and HF influence one another is incompletely understood. In OSA, changes in intrathoracic pressure and increased sympathetic drive are thought to be implicated, while in CSA, processes altering CO2-driven respiratory mechanisms and sympathetic nervous system activity are likely to be involved.7
When properly managed, both forms of SDB can improve HF patient quality of life and comorbidities; however, accurate diagnosis in a laboratory setting can be challenging. The current gold standard for SDB testing is polysomnography (PSG), a complex procedure that actively monitors at least seven physiological channels using EEG, electrooculogram (EOG), chin electromyogram (EMG) and ECG, while measuring airflow, oxygen saturation, respiratory effort and sometimes heart rate.8–11 Though comprehensive, in-lab evaluations are burdened with long wait times, patient discomfort, difficulty attending test appointments due to limited mobility or need for caregivers, and highly specialised personnel.
Given these diagnostic bottlenecks, home sleep apnoea testing (HSAT) is an accessible alternative to traditional in-lab PSG that may provide timely diagnosis and management of these patients. These devices vary in testing parameters and design, enabling selection tailored to a patient’s specific limitations. Compared to PSG, HSAT is less sensitive, but for high-probability individuals with low comorbidity burden, it is considered an acceptable alternative for diagnostic purposes.12 This review explores the current and emerging role of HSAT in HF patients with suspected sleep apnoea.
History and Limitations of Home Sleep Apnoea Testing
Emergence of Home Sleep Apnoea Testing and Types
The technical and logistical challenges of PSG testing have been recognised since the early 1990s, and the search for an ideal portal sleep monitoring system has been an evolutionary process.13 The ApneaLink device (ResMed) is one of the earliest Food and Drug Administration-approved HSAT devices for the diagnosis of OSA.14 This single-channel screening tool was designed to measure airflow through a nasal cannula connected to a pressure transducer. Early trials with the ApneaLink also attempted to characterise CSR, with relatively high diagnostic accuracy (sensitivity 87.1%, specificity 94.9%).15
Devices have increasingly evolved to measure a variety of physiologic variables to distinguish sleep patterns. Sleep monitoring methodologies are classified into four levels (Table 1), with the majority of HSAT devices categorised as level III or level IV.12,16 There are several key technologies commonly employed in HSAT monitoring.17 Respiratory inductance plethysmography (RIP) can be used to record chest and abdominal expansion and contraction to evaluate respiratory movement during sleep. Nasal pressure transducers are used to assess airflow. Microphones may be used to record snoring volume. Accelerometers and gyroscopes are used to record body movements and positions during sleep and may involve devices worn on the wrist or ankle. Photoplethysmographic (PPG) signals are often used to determine oxygen saturation and pulse rate. Radar and sonar have started to be explored as novel non-wearable methods of assessing body movement while asleep. Table 2 includes FDA-approved devices for HSAT testing and the primary measurements recorded for each, adapted from Park et al.18
Peripheral arterial tonometry (PAT) is another popular technique for assessing sympathetic activation during sleep.17 This method uses beat-to-beat PPG signals to noninvasively record peripheral arterial vascular tone, or fluctuations in fingertip vasoconstriction as a result of sympathetic activity.19,20 In SDB, apnoeic episodes can lead to increased sympathetic tone and peripheral vasoconstriction. PAT affords several conveniences to patients, including minimal equipment and even disposable devices; however, the pressure on the digits applied by the device may be uncomfortable for patients.21,22 Importantly, these devices alone do not measure airflow or brain activity during sleep. PAT diagnostic accuracy has been shown to be comparable to other HSAT types but remains insufficient to definitively rule out sleep apnoea compared to PSG.
There are several significant advantages to using HSAT devices for the diagnosis of OSA, including improved access to testing, cost-effectiveness, and home delivery options.23 Kim, et al. performed an economic analysis of HSAT versus in-lab diagnosis of OSA and found that for Medicare, use of HSAT decreased patient costs by US$264 per patient, even if additional in-lab testing was required for inconclusive home studies.24 This calculation likely underestimates the true value of HSAT, as it does not capture the indirect costs of delayed diagnosis, such as poor patient quality of life and increased healthcare usage (often related to cardiovascular comorbidities).25,26 This finding was not consistent for provider costs, which demonstrated increased cost of US$40 per patient using HSAT.24 Early assessment of suspected severe OSA and initiation of positive airway pressure (PAP) therapy has been shown to improve treatment adherence.27
Limitations of Home Sleep Apnoea Testing
Notable limitations of HSAT should be considered when selecting candidates for these methods. HSAT testing is often unsupervised and, therefore, is subject to errors in patient execution, device dislodgement, and poor signal quality.12 Additionally, patient physicality may limit HSAT use, including patients with a BMI >35 kg/m2, which is a known risk factor for OSA.28
A major limitation of type III and type IV HSAT use is the inability to confidently estimate total sleep time, which requires simultaneous EEG, EOG, and EMG monitoring for sleep staging.12 In PSG, the apnoea-hypopnoea index (AHI) reflects the patient burden of hypopnoeas (airflow reduction) and apnoeas (airflow cessation) during the study, with a higher index indicating worse sleep disturbances.29 The formula for AHI is given below, where 5–15 events per hour are classed as mild, 15–30 events per hour moderate and >30 events per hour severe.
Because of decreased sensitivity of HSAT, the respiratory event index (REI) has been derived to characterise sleep quality in these devices. The REI uses estimated sleep time as opposed to total sleep time used in PSG-derived AHI; estimated total sleep time tends to be higher than actual sleep time, thus the REI generally underestimates PSG-derived AHI.12
When compared to PSG, HSAT is 61% accurate with a 0.82 correlation in estimating AHI in a meta-analysis by Massie et al., and so should be used in uncomplicated patients with moderate to high clinical suspicion for OSA based on comprehensive assessment of the clinical history, physical exam and sleep questionnaires.12,30 To account for the lower sensitivity of HSAT compared to PSG, the American Academy of Sleep Medicine recommends comprehensive patient assessment in tandem with HSAT to formally diagnose OSA, with escalation to PSG if HSAT is inconclusive or negative for sleep apnoea with a high index of suspicion for the condition.12 PSG is also preferred over HSAT to diagnose OSA in patients with significant cardiorespiratory disease, neuromuscular respiratory muscle weakness, chronic opioid use, history of stroke, severe insomnia and awake hypoventilation. These conditions make PSG preferable due to the higher risk of CSA or, in the case of insomnia, that the estimated total sleep time would be unreliable.
Furthermore, if the primary diagnostic concern is CSA, HSAT is not recommended, because these devices have not been cleared for CSA detection. However, one study using PPG to detect CSA demonstrated similar AHI estimation compared to PSG for this population (sensitivity 81%, specificity 99% at an AHI cutoff of 10 events per hour).31
Another limitation of HSAT is that it is approved for diagnostic purposes only and cannot be used to titrate therapy. PSG is advantageous in that one study can be both diagnostic and therapeutic by allowing for PAP titration to occur in the second portion of the night (i.e. split-night study).
Finally, HSAT is currently approved for SDB diagnosis alone, and is just starting to be validated for detection and characterisation of other comorbidities, which will be discussed later in this review. More stringent screening strategies are required to determine ideal candidates for HSAT versus PSG, which should include the likelihood of CSA and presence of comorbidities.
Home Sleep Apnoea Testing Usage in Heart Failure Patients
HF patients have a higher prevalence of CSA compared to the general population, occurring in 20–50% of patients with either HF with reduced ejection fraction (HFrEF) or HF with preserved ejection fraction (HFpEF).32 It is a class IIA recommendation by the American College of Cardiology, American Heart Association, and Heart Failure Society of America that patients with New York Heart Association class II–IV and suspicion of SDB undergo a formal sleep assessment.33 In the clinical setting, HF nurses can be instrumental in eliciting sleep patterns and reviewing health information accrued from digital wearable devices to refer patients for HSAT evaluation.34
The pathophysiology of CSA differs in HF patients compared to patients with intact cardiovascular health. In HF patients with CSA, impaired cardiac output increases sympathetic response and hyperventilation by delaying CO2 chemoreceptor signalling. The resulting hypocapnia leads to compensatory apnoea. This is particularly evident in HF patients with CSR, which can occur while in wakeful states as well.32 Fluid derangements further disrupt sympathetic regulation and can worsen sleep quality.
The mixed presentation of OSA and CSA in HF patients may make assessment using HSAT devices challenging. Li et al. examined this diagnostic complexity in 84 patients with chronic, stable HF and no prior SDB diagnosis, comparing sleep metrics using a type III HSAT to in-lab PSG.35 The Nox T3 (Nox Medical) device was used for the study and measured nasal pressure, respiratory movement using RIP, snoring, body position, activity, heart rate and oxygen saturation. Technologists were able to classify apnoeic episodes captured via nasal pressure signal as obstructive (apnoea associated with respiratory effort), central (apnoea during which respiratory effort was absent), or mixed (apnoea during which respiratory effort was initially absent but appeared in the latter part of the event). Additionally, they classified CSR as at least three central apnoeas separated by a crescendo-decrescendo change in breathing amplitude over at least 40 seconds. Compared to PSG, the HSAT device had similar event detection for obstructive (MD −0.3 events/h, p=0.6) and central (MD −1.1 events/h, p=0.1) apnoea indices as well as detection of CSR.
PAT technology has also been tested in differentiating OSA and CSA in HF patients using the WatchPAT system (Itamar Medical).36 Central versus obstructive apnoeic events can be identified with PAT by comparing accelerometer data on chest wall respiratory movements and dynamic systolic upstrokes, representing increased peripheral tone, on PAT signal waveform. In centrally derived apnoeic episodes, systolic upstrokes occurred regardless of respiratory effort, while for obstructive apnoeic episodes, intrathoracic pressure changes directly impacted systolic upstrokes measured by PAT.
Future studies may focus on dynamic fluid changes in HF patients and the effects on SDB using HSAT. Peripheral oedema in HF can lead to fluid shifts rostrally into the chest while sleeping, leading to upper airway and pulmonary volume overload. These fluid shifts can initiate CSA and CSR by increasing hyperpnoea, and can worsen OSA by distending jugular veins and parapharyngeal tissues.37,38 Devices that employ PAT have the potential to detect changes in vascular tone from increased volume. Integration of these data with other existing volume monitoring devices, such as the CardioMEMS HF system (Abbott), which measures pulmonary arterial pressure, might characterise how these fluid shifts impact SDB and HF severity.
Home Sleep Apnoea Testing Detection of AF
HSAT is also being validated to evaluate for the presence of AF, one of the most frequently reported comorbidities in both HF and SDB due to gradual atrial electrical and structural remodelling from the progression of chronic insults (i.e. high-frequency oxygen desaturation and reoxygenation, enhanced sympathetic tone related to atrial stretching from volume overload, large changes in intrathoracic pressure during sleep).39 Studies have shown that identification and PAP management of SDB ameliorate the recurrence of AF after catheter ablation.40,41 PAT devices have been predominantly explored for HSAT testing of cardiac arrhythmias. Pillar et al. compared WatchPAT signalling in 84 patients with HF to concurrent cardiac monitoring via ECG during PSG.42 The WatchPAT was able to assess cardiac electrical activity by timing QRS complexes detected in peripheral pulsation and irregularly spaced QRS complexes suggesting underlying AF. Compared to ECG, this method of detecting AF was highly sensitive (91.7%) and specific (98.6%) for episodes greater than 6 minutes but was less sensitive for AF duration <60 seconds (sensitivity 77.3%, specificity 98.7%).
Advanced Heart Failure Population
The use of HSAT devices has been explored in patients with advanced HF, defined as patients with class III or IV New York Heart Association symptoms that limit daily activity despite optimal medical therapy.43 Carey et al. used the WatchPAT device to assess SDB in patients with advanced HF or who had received heart transplant with a positive STOP-Bang survey, though participants were not evaluated with PSG for comparison.44 The study team attempted to collect PAT data for left ventricular assist device patients, but results were inconclusive, likely due in part to lack of native pulsatility in this population. An alternative HSAT technology that does not rely on peripheral pulsatility may be more suitable for evaluation of SDB in left ventricular assist device patients.45
Home Sleep Apnoea Testing for Monitoring Therapy Response
While HSAT has not been approved for PAP initiation and titration, there are several non-PAP therapies that have been evaluated using HSAT. The efficacy of mandibular advancement devices that serve to reposition the jaw for improved respiratory flow can be evaluated using HSAT.46,47
Another PAP alternative is hypoglossal nerve stimulation (HGNS), an implantable therapy that detects inspiratory effort during sleep and stimulates the hypoglossal nerve to recruit muscles in an effort to decrease airway collapse.48 Huyett et al. used PAT and a pressure sensor sleep mat compared to PSG to estimate AHI in patients using HGNS.49 Compared to PSG, PAT had a sensitivity of 57% with a specificity of 77%, while the mat pressure sensor had a sensitivity of 61% and specificity of 82% for AHI <15. Use of portal monitoring systems may allow for patient self-titration of HGNS using objective measures, as opposed to the subjective self-titration currently employed.
For CSA, transvenous phrenic nerve stimulation is a PAP alternative that induces breathing when the respiratory drive is inadequate. When successful, phrenic nerve stimulation in HF patients reduces AHI and improves quality of life and heart rate variability during sleep; the efficacy of this device can be monitored by HSAT.50,51
Future Directions
Pre-screening tools should be developed to identify ideal candidates for diagnosis of sleep apnoea using HSAT versus PSG.52 The integration of widely available technologies, such as smartwatches, has the potential to screen patients for formal diagnostic testing.53 Head-to-head studies comparing multiple HSAT systems to PSG should be performed to identify the ideal combination of measured parameters to effectively estimate AHI. Contactless methods incorporating radar technology that monitor movements during sleep could be combined with FDA-approved HSAT modalities to better identify patients who should be referred for PAP therapy.54
There is a need for further validation of HSAT devices in HF populations and approval for CSA detection. Phenotyping of HF patients with mixed CSA and OSA presentations should be identified to guide optimal treatment. HF cohort analyses stratified by HF aetiology, HF subtype and with the concurrent use of guideline-directed medical therapy should be examined. Additionally, pre-screening tools for HSAT should be developed that incorporate specific HF biomarkers and cardiovascular parameters to best account for the high prevalence of CSA in this population. There is also a need for investigation into the consequences of delayed sleep apnoea diagnosis on long-term outcomes in HF patients and if early HSAT screening can meaningfully improve negative consequences.
Conclusion
The use of HSAT has considerably improved accessibility to diagnostic testing for sleep apnoea but continues to undergo investigation in cases of CSA and in HF populations. Careful patient selection for HSAT testing is required but has the potential to identify SDB phenotypes and important concurrent comorbidities of SDB, such as arrhythmias. Continued characterisation of SDB in HF patients using HSAT may improve clinical outcomes and quality of life but requires dedicated investigation in this often-excluded population.