INTRODUCTION — Treatment-emergent central sleep apnea (CSA), previously referred to as complex sleep apnea, is detected in approximately 5 to 15 percent of patients who undergo positive airway pressure (PAP) titration for obstructive sleep apnea (OSA).
In more than half of cases, treatment-emergent CSA is a transient phenomenon that resolves within the first few months of PAP. Less commonly, the abnormality is persistent and requires a change in mode of ventilation.
The epidemiology, pathogenesis, clinical presentation, clinical findings, diagnostic criteria, and management of treatment-emergent CSA are reviewed here. The diagnosis and management of OSA and central sleep apnea are described separately. (See "Clinical presentation and diagnosis of obstructive sleep apnea in adults" and "Management of obstructive sleep apnea in adults" and "Central sleep apnea: Risk factors, clinical presentation, and diagnosis" and "Central sleep apnea: Treatment".)
DEFINITION — Treatment-emergent CSA is defined as the persistence or emergence of central apneas and hypopneas during the initiation of PAP therapy for obstructive OSA, with a central apnea index >5 events/hour of sleep and the number of central events is ≥50 percent of the total events, despite significant resolution of obstructive respiratory events [1].
Alternative definitions of treatment-emergent CSA have been used [2-6], but they all share the presence of new or increased central apneas and central hypopneas that prevent the apnea-hypopnea index (AHI) from normalizing despite the disappearance of obstructive apneas and hypopneas.
PREVALENCE AND RISK FACTORS — Risk factors for treatment-emergent central sleep apnea (CSA) are not well understood, and there is wide variability in the reported incidence depending upon the study and the population. In a large prospective study, the prevalence during the first night with stable continuous positive airway pressure (CPAP) was 12 percent [7]. Smaller studies have reported an incidence ranging from 5 to 20 percent [8]. The incidence tends to be higher in studies using split-night versus full-night PAP titration.
Factors associated with an increased likelihood of treatment-emergent CSA in small studies include (table 1):
●Male sex [3,9,10]
●Older age [7,11]
●Comorbid heart failure or coronary artery disease [8,9,12]
●Severe obstructive sleep apnea (OSA) [13]
●Presence of a mixture of obstructive, mixed, and central apneas during initial polysomnography [7,9,13-15]
●Higher arousal index on baseline polysomnography [7,9]
●Use of higher levels of CPAP (ie, over-titration) [10,16]
●Use of bilevel PAP or a high-pressure support level [10,17]
●High altitude [18]
●Oral breathing [4]
●Opioid use [5,10]
●Supine sleep position [19,20]
●Elevated nasal resistance with CPAP intolerance [21]
PATHOGENESIS — Whether or not treatment-emergent central sleep apnea (CSA) and obstructive sleep apnea (OSA) are separate disorders is controversial [2-4]. Treatment-emergent CSA shares similar pathophysiology with OSA and CSA but is a dynamic process, which often resolves spontaneously with ongoing positive pressure therapy [11,22,23].
There are numerous potential mechanisms by which treatment of OSA may induce central apneas, including the following [4,22]:
●Loop gain – Loop gain refers to the ratio of a corrective ventilatory response to a disturbance. If the response exceeds the disturbance (ie, loop gain >1), then self-sustaining periodic breathing with central apneas may result. Patients with OSA frequently exhibit "high loop gain" or "ventilatory instability" at baseline, which is augmented by PAP therapy. As a result, minor disturbances (eg, arousals from sleep) result in an exuberant corrective ventilatory response (ie, hyperpnea) and then central apnea if the arterial carbon dioxide tension (PaCO2) falls below the apneic threshold. The central apnea, in turn, prompts a reciprocal corrective response, establishing a pattern of periodic breathing. High loop gain is more prevalent in patients with severe OSA than those with mild OSA [13], as well as in patients who receive BPAP rather than CPAP [17].
●Positive airway pressure (PAP) – PAP therapy relieves repetitive upper airway obstruction, which increases the elimination of CO2, lowers the PaCO2, and induces treatment-associated central apnea if the PaCO2 falls below the apneic threshold [13,24-26].
●Intermittent hypoxia – Acute intermittent hypoxia is associated with increased chemoreflex sensitivity and increased propensity to develop central apnea [27]. This may explain the higher likelihood of developing treatment-emergent CSA during split-night versus full-night PAP titration studies. Likewise, chronic intermittent hypoxia, as in OSA, is associated with increased peripheral chemoreflex activity. Accordingly, patients with OSA are more susceptible to the development of hypocapnic central apnea secondary to increased CO2 chemoreflex sensitivity relative to healthy control subjects. Treatment with nasal CPAP for three months was associated with decreased chemoreflex sensitivity and decreased propensity to develop central apnea [28]. This may also explain the finding that treatment-emergent CSA resolved after a few months of PAP therapy in a majority of the patients [5,11,23].
●Air leak – Air leak during PAP titration may be associated with the development of acute central apnea, owing to CO2 washout from the mask and the anatomic dead space [29].
●Stretch receptor activation – PAP therapy activates stretch receptors, which inhibits central respiratory output and causes central apnea [25].
While these mechanisms directly link PAP therapy to the development of central apneas, it is important to realize that other factors may contribute in some situations. These include loss of the wakefulness stimulus to breath during sleep [30], baseline hypocapnia due to heart failure [31], reduced drive to breathe due to opioids or high altitude [18], and concomitant hypoxia [13]. Moreover, OSA itself increases the propensity to central apneas in association with a higher loop gain [28].
CLINICAL FEATURES
Presentation — Treatment-emergent central sleep apnea (CSA) typically presents as an incidental polysomnographic finding during the initial in-laboratory titration of continuous positive airway pressure (CPAP) therapy (or bilevel positive airway pressure without a backup respiratory rate) for obstructive sleep apnea (OSA). Specifically, central apneas and hypopneas emerge or increase during the titration, which prevent the apnea-hypopnea index (AHI) from normalizing despite the disappearance of obstructive apneas and hypopneas. The treatment-emergent apneas almost invariably occur during non-rapid eye movement (NREM) sleep [1,32,33] and may be increased during supine sleep [20].
Less commonly, treatment-emergent CSA begins after several months of positive airway pressure (PAP) therapy [7].
History — Patients with an increased number of central apneas while being treated with PAP may remain asymptomatic or manifest symptoms and signs of disrupted sleep due to the central apneas and accompanying arousals. The symptoms and signs may include excessive daytime sleepiness, poor subjective sleep quality, repetitive awakenings, insomnia, fatigue, inattention, poor concentration, and moodiness.
Patients may also report symptoms related to recurrent oxyhemoglobin desaturation, including morning headaches and nocturnal angina. Some patients describe paroxysmal nocturnal dyspnea due to the compensatory hyperpnea that follows a central apnea.
Snoring is not a typical finding because patients with treatment-emergent CSA have had their recurrent upper airway obstruction alleviated by the PAP therapy.
Examination — There are no physical findings that are specific for treatment-emergent CSA. However, most patients have physical features that are typical of patients with OSA, including obesity, a large neck circumference, and/or a crowded upper airway [3,32]. (See "Clinical presentation and diagnosis of obstructive sleep apnea in adults", section on 'Physical examination'.)
DIAGNOSIS — Treatment-emergent central sleep apnea (CSA) is diagnosed when central apneas develop in patients with obstructive sleep apnea (OSA) after the application of continuous positive airway pressure (CPAP). No additional diagnostic testing is indicated or required beyond polysomnography.
With the increased adoption of home sleep apnea testing (HSAT) and auto-titrating PAP devices for OSA, emergent CSA may be detected in the download from an auto-titrating PAP device as an elevated residual apnea-hypopnea index (AHI). This finding usually warrants further confirmatory evaluation by an in-laboratory polysomnography with PAP titration. (See "Downloading data from positive airway pressure devices in adults", section on 'Continuous positive airway pressure downloads'.)
Diagnostic criteria for treatment-emergent CSA according to the third edition of the International Classification of Sleep Disorders (ICSD-3) include all of the following [1]:
●Diagnostic polysomnography shows ≥5 predominantly obstructive respiratory events (obstructive or mixed apneas, hypopneas, or respiratory effort related arousals) per hour of sleep
●Polysomnography during the use of positive airway pressure without a backup rate shows significant resolution of obstructive events and emergence or persistence of central apnea or central hypopnea with all of the following:
•Central apnea and central hypopnea index ≥5 per hour
•Number of central apneas and central hypopneas is ≥50 percent of the total number of apneas and hypopneas
●The CSA is not better explained by another CSA disorder, such as CSA with Cheyne-Stokes breathing or CSA due to a medication or substance
Criteria used by the Centers for Medicare and Medicaid Services for reimbursement of treatment-emergent CSA also include a requirement for concomitant daytime sleepiness or disrupted sleep [34]:
●Patients must have confirmed OSA, defined as predominantly obstructive or mixed apneas occurring at a frequency of ≥5 events per hour (see "Clinical presentation and diagnosis of obstructive sleep apnea in adults", section on 'Diagnosis')
●Upon exposure to CPAP or bilevel positive airway pressure (BPAP) without backup rate, all of the following must be present:
•AHI ≥5 events per hour
•Central apneas and hypopneas make up more than 50 percent of the AHI
•Central AHI ≥5 events per hour
•Symptoms of excessive daytime sleepiness or disrupted sleep
Obstructive apneas, central apneas, mixed apneas, hypopneas, and the AHI are defined separately. (See "Polysomnography in the evaluation of sleep-disordered breathing in adults".)
NATURAL HISTORY — The natural history of treatment-emergent central sleep apnea (CSA) is not well defined prospectively. In observational studies, central events resolve spontaneously in 50 to 85 percent of patients who return for repeat polysomnography after several months of continuous positive airway pressure (CPAP) therapy [2,4,5,7,8,11,23,35]. Similar numbers were reported in the CPAP arm of a small, randomized trial, in which approximately two-thirds of patients with treatment-emergent CSA achieved an apnea-hypopnea index (AHI) <10 per hour after 90 days of CPAP therapy [36]. This proportion improved to 73 percent when limited to patients who were adherent with therapy.
Patients with treatment-emergent CSA appear to be at increased risk for adverse outcomes. In one large database study, patients with treatment-emergent CSA at week 1 or 13 of PAP therapy were at higher risk of early PAP termination [11]. In a separate study, treatment-emergent CSA was associated with higher rates of postoperative complications, including unplanned intensive care admissions, 30-day readmission rate, and mortality rate [37].
TREATMENT — Management of treatment-emergent central sleep apnea (CSA) is controversial, and treatment recommendations are derived from clinical experience, small observational studies, and a limited number of clinical trials. Since many patients improve spontaneously after several months of continuous positive airway pressure (CPAP) therapy, expectant management with CPAP is appropriate in most cases.
In patients who do not improve, options include changing the mode of positive airway pressure to either adaptive servo-ventilation (ASV) or bilevel positive airway pressure (BPAP) with a backup rate. If these modes are unavailable or too costly, however, continuing CPAP is preferable to discontinuing the treatment of obstructive sleep apnea (OSA) altogether.
Expectant management with CPAP — For patients with treatment-emergent CSA, we suggest expectant management with CPAP therapy, rather than making an immediate change in therapy. We perform a repeat clinical evaluation and diagnostic sleep test in two to three months after the initial study. The test should be performed with patients wearing their device.
This approach is supported by retrospective and limited prospective data showing that central apneas improve or resolve spontaneously in approximately two-thirds of patients who continue CPAP for 90 days [36]. (See 'Natural history' above.)
As part of expectant management, patients should be evaluated for comorbidities and medications that may increase the likelihood of central apneas. This includes optimization of the heart failure regimen in patients with symptomatic heart failure and rigorous assessment of the need for continuing opioid therapy in patients taking opioids.
Predictors of persistent central apneas on continued CPAP are not well defined, and clinical symptoms may or may not be present. Repeat polysomnography or home sleep apnea testing (HSAT) is therefore required after two to three months to determine whether further changes are necessary, considering that limited data suggest that ASV or bilevel PAP with a backup rate may be more effective in controlling CPAP-emergent central apneas [36,38]. Clinical symptoms and adherence should be reassessed at this time as well.
The decision to continue CPAP therapy or change to an alternative mode of ventilation should be made based upon the persistence of CPAP-emergent central apneas, the stability of the patient’s medical condition (such as coexistent heart failure), and the severity of sleep-related symptoms. Data in heart failure patients with CSA suggest that improvements in left ventricular ejection fraction and transplant-free survival are dependent upon adequate suppression of central apneas [38].
Alternative modes of ventilation — ASV and BPAP with a backup rate are alternative modes of ventilation in patients with persistent central apneas on CPAP. The data reviewed below suggest that both ASV and BPAP with a backup respiratory rate decrease the frequency of obstructive and central apnea and hypopneas, but ASV may have a slightly larger magnitude of effect. In addition, limited longer-term data suggest that re-emergence of central events may be of concern with BPAP but not ASV [39].
Adaptive servo-ventilation — Small clinical trials have consistently shown that the apnea-hypopnea index (AHI) is also lowered by ASV in patients with treatment-emergent CSA, with the AHI typically improving to <10 events per hour both acutely and after several months of therapy in a large majority of patients [12,39-42]. Our confidence in this evidence is limited, however, because it consists of small trials that measured primarily physiological outcomes rather than patient-important outcomes. Notably, most of these trials were funded by the makers of the PAP device being studied. Moreover, ASV devices are considerably more expensive than CPAP devices.
While ASV may be more effective at reducing the AHI compared with CPAP, the clinical relevance of this advantage is not yet clear [36,43]. In a multicenter trial, 66 patients with treatment-emergent CSA and a residual central apnea index (CAI) >10 events per hour during the CPAP titration study were randomly assigned to ASV or CPAP therapy for 90 days [36]. There were no significant differences in baseline clinical characteristics between the two groups, although more patients in the CPAP arm were male (91 versus 79 percent) and had heart failure (15 versus 3 percent). Six patients failed to complete the study, and the adherence rate (defined as use of the study device for ≥4 hours per night and on ≥70 percent of the nights during the study) in the remaining patients was 60 percent. Results included the following:
●After 90 days of therapy, the AHI was lower in patients treated with ASV versus CPAP (4.4 versus 9.9 events per hour, p = 0.002), but the absolute difference of 5.5 events per hour was lower than the prespecified clinically relevant difference of 10 events per hour.
●The proportion of patients who achieved an AHI <10 events per hour was greater for ASV (90 versus 65 percent). An analysis in adherent patients yielded similar results (94 versus 74 percent).
●Patients in both groups had improved Epworth Sleepiness Scale (ESS) scores and minimal change in measures of quality of life.
●No baseline clinical or polysomnographic characteristics could be identified that predicted lack of improvement on CPAP.
ASV provides a varying amount of inspiratory pressure superimposed on a low level of CPAP, with a backup respiratory rate [41,44-46]. The magnitude of the inspiratory pressure provided by the device is reciprocal to changes in peak flow, determined over a three- to four-minute moving window (figure 1). Thus, peak flows that are lower than the average peak flow induce an increase in the amount of inspiratory pressure; conversely, peak flows that are higher than the average peak flow induce a decrease in the amount of inspiratory pressure. The backup respiratory rate can be set automatically by the device or manually and is an important aspect of ASV [47,48].
Of note, ASV should be avoided in patients with CSA associated with heart failure with reduced ejection fraction. This is based on results of SERVE-HF, a randomized trial of ASV versus standard medical therapy in patients with predominantly CSA due to symptomatic heart failure and a low ejection fraction (EF ≤45 percent) in which ASV was associated with a 6 percent absolute increase in all-cause mortality and cardiovascular mortality compared with standard medical therapy [49] (see "Central sleep apnea: Treatment", section on 'Patients with ejection fraction ≤45 percent'). Moreover, in an industry-sponsored randomized controlled cross-over trial of patients with treatment-emergent CSA with preserved left ventricular ejection fraction, there were significant differences in minute ventilation and sleep architecture while receiving ASV therapy from various available devices [50]. In this study, higher minute ventilation was associated with small but statistically nonsignificant prolongation of QT interval, suggesting that there may be device-specific rather than ASV class-specific effects that may have been responsible for the observed increased cardiovascular mortality in the SERVE-HF trial.
BPAP with a backup respiratory rate — BPAP with a backup rate can also lower the AHI compared with levels during CPAP initiation in patients with treatment-emergent CSA, at least initially [33,41,44].
Small trials have demonstrated that BPAP with a backup rate and ASV perform similarly, although re-emergence of central apneas might be of concern with BPAP with a backup rate:
●In a randomized trial that included nine patients with treatment-emergent CSA, ASV reduced the AHI from 52 to 1 event per hour, while BPAP with a backup rate reduced the AHI from 52 to 6 events per hour [41].
●In a larger randomized trial that included 30 patients with treatment-emergent CSA, both ASV and BPAP, with a backup rate just below the average sleep-related respiratory rate, reduced the first-night AHI (28 to 9 events per hour and 29 to 9 events per hour, respectively); however, there was re-emergence of central events in the BPAP group after six weeks [39]. This observation requires further study and independent confirmation, given that the pathophysiologic mechanisms of recurrence with BPAP, but not with ASV, with backup rate remains unexplained if patients were adherent to therapy.
BPAP delivers positive airway pressure at different, but fixed, levels during inspiration and expiration. The level during inspiration is called the inspiratory positive airway pressure (IPAP) and the level during expiration is called the expiratory positive airway pressure (EPAP). BPAP has two major effects. First, it splints the upper airway open. Second, it increases alveolar ventilation by augmenting the tidal volume. The tidal volume is directly related to the difference between the IPAP and EPAP. As an example, the tidal volume is greater when the IPAP is set at 15 cm H2O and the EPAP at 5 cm H2O (difference of 10 cm H2O), than when the IPAP is set at 10 cm H2O and the EPAP at 5 cm H2O (difference of 5 cm H2O). (See "Titration of positive airway pressure therapy for adults with obstructive sleep apnea", section on 'Bilevel positive airway pressure, spontaneous mode (BPAP-S)'.)
Modes to avoid — In contrast to BPAP with a backup rate, BPAP without a backup rate does not decrease the AHI during BPAP compared with during the initiation of CPAP in patients with treatment-emergent CSA, and it may even worsen the AHI.
This was suggested by a retrospective cohort study of 63 patients with treatment-emergent CSA that found that BPAP without a backup rate increased the AHI from 30 to 52 events per hour [33] and supported by indirect evidence from a retrospective cohort study of 95 patients with various types of central apnea that found that BPAP without a backup rate was more likely to be associated with increased rather than decreased central events [17]. Therefore, BPAP without a backup rate is not recommended in patients with treatment-emergent CSA [51].
Other potential therapies — Supplemental nocturnal oxygen has been tried with some success in patients with CSA due to heart failure [52,53], providing potential rationale for its use in treatment-emergent CSA. The proposed mechanism of action is via a reduction in loop gain with resultant amelioration of CSA [54,55]. However, there are no controlled trials of oxygen alone or in conjunction with PAP in patients with treatment-emergent CSA.
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Sleep-related breathing disorders in adults".)
SUMMARY AND RECOMMENDATIONS
●Definition – Treatment-emergent central sleep apnea (CSA) is the persistence or emergence of central apneas and hypopneas during the titration of positive airway pressure therapy without a backup respiratory rate for obstructive sleep apnea (OSA). (See 'Definition' above.)
●Clinical features – Treatment-emergent CSA presents as an incidental polysomnographic finding during the initial in-laboratory titration of continuous positive airway pressure (CPAP) therapy (or bilevel positive airway pressure [BPAP] without a backup respiratory rate) for OSA. (See 'Presentation' above.)
Patients are usually asymptomatic but sometimes manifest symptoms and signs of disrupted sleep due to the central apneas and accompanying arousals. They may also report symptoms related to recurrent oxyhemoglobin desaturation. (See 'Clinical features' above.)
●Diagnosis – Diagnostic criteria for treatment-emergent CSA include the following (see 'Diagnosis' above):
•Patients must have confirmed OSA, defined as predominantly obstructive respiratory events occurring at a frequency of ≥5 events per hour of sleep.
•Upon exposure to positive airway pressure without a backup respiratory rate, all of the following must be present: significant resolution of obstructive events; central hypopnea index ≥5 events per hour; central apneas and hypopneas make up ≥50 percent of the total number of apneas and hypopneas; and symptoms of excessive daytime sleepiness or disrupted sleep.
●Natural history – Treatment-emergent CSA resolves spontaneously in approximately two-thirds of patients on continued CPAP therapy. (See 'Natural history' above.)
●Expectant management – For patients with treatment-emergent CSA, we suggest continuing CPAP as initial therapy rather than switching to an alternative mode of ventilation (Grade 2C). Patients should undergo repeat polysomnography or home sleep apnea testing (HSAT) while wearing their device in two to three months to determine whether central apneas have resolved. (See 'Expectant management with CPAP' above.)
●Persistent central apneas – For most patients with persistent central apneas on CPAP, we suggest switching to adaptive servo-ventilation (ASV) (Grade 2C). In situations in which ASV is not available or too expensive, BPAP with a backup rate is a reasonable alternative. If neither mode is available, continuing CPAP is preferable to discontinuing therapy altogether. (See 'Alternative modes of ventilation' above.)
●Modes to avoid – ASV should be avoided in patients with heart failure and a reduced ejection fraction, based on concerns that it may increase cardiovascular mortality when used to treat CSA in this patient population. (See "Central sleep apnea: Treatment", section on 'Patients with ejection fraction ≤45 percent'.)
BPAP without a backup rate may be ineffective or harmful in patients with treatment-emergent CSA. (See 'Modes to avoid' above.)
24 : Susceptibility to periodic breathing with assisted ventilation during sleep in normal subjects.
27 : Effect of episodic hypoxia on the susceptibility to hypocapnic central apnea during NREM sleep.
