Introduction
Sleep-disordered breathing (SDB), obstructive sleep apnea (OSA) in particular, is highly prevalent in patients with heart failure (HF) [
1,
2]. Epidemiological studies suggest that OSA is an independent risk factor for HF development and has a negative effect on prognosis in patients with HF [
3‐
5]. OSA is characterized by repeated partial or complete collapse of the upper airway during sleep [
6], accompanied by complete absence of airflow and paradoxical respiration with opposing respiratory movements of the thorax and abdomen [
7]. Obstructive respiratory events result in negative intrathoracic pressure swings, which influence venous return and preload for the right ventricle, and trigger cardiac arrhythmias [
8]. Another mechanism by which OSA might contribute to the pathophysiology of HF is by increasing sympathetic activation [
9]. Endothelial dysfunction, inflammation, hypercoagulability and apoptosis have also been associated with OSA [
10] and may contribute to the development or worsening of HF. Furthermore, nocturnal hypoxemia was a robust and independent predictor of all-cause mortality in patients with stable HF with reduced ejection fraction (HFrEF) in a cohort study [
11].
Continuous positive airway pressure (CPAP) maintains upper airway patency by applying positive pressure during sleep [
7,
12] and is the gold standard treatment for moderate-to-severe OSA in symptomatic patients with or without HF [
7,
13]. Recently, CPAP has been shown to improve the clinical course of cardiac recompensation and to have a beneficial impact on pulmonary hypertension in patients with acute HF and OSA [
14].
Automatically titrating CPAP (APAP) devices include an algorithm that adjusts delivered pressures to a patient’s individual demands. Patients with HF are expected to benefit from application of lower positive airway pressures delivered by APAP in terms of cardiac filling pressures and, especially, right ventricular function [
15], making APAP an attractive option for the treatment of OSA in patients with HF. However, the randomized controlled SAVE (Sleep Apnea Cardiovascular Endpoints) trial of 2717 OSA patients with coronary or cerebrovascular disease did not show any significant beneficial effects of APAP on the rate of cardiovascular events and mortality over a mean follow-up of 3.7 years [
16]. Hereby, device use during the trial was low, averaging only 3.3 h per night. Current literature suggests that positive airway pressure therapy for OSA needs to be used for a minimum of 4 h per night to achieve measurable benefits [
7], including reductions in blood pressure [
17] and prevention of recurrent atrial fibrillation [
18].
Cardiopulmonary exercise capacity (peak VO
2) is a well-known predictor of mortality in HFrEF and remains one of the major parameters defining qualification for heart transplantation [
19]. Peak VO
2, reflecting maximal oxygen consumption, is closely related to cardiac output, representing the best validated non-invasive parameter to classify HFrEF stage and severity [
19]. However, there is a current lack of data on the impact of APAP therapy on exercise capacity in patients with HFrEF.
This randomized controlled pilot study investigated the impact of APAP therapy on cardiopulmonary exercise capacity, echocardiography measures of cardiac function, quality of life, and nocturnal OSA parameters in patients with HFrEF.
Discussion
This is the first randomized controlled trial to show improvement in percent-predicted peak VO
2 as a primary study endpoint in APAP therapy in HFrEF patients with OSA determined in cardiopulmonary exercise capacity testing (Fig.
1). Moreover, this study shows enhanced global cardiac function after 6 months of APAP therapy and our study demonstrates statistically significant and clinically meaningful improvements of APAP therapy for polysomnography parameters, functional variables, along with parallel improvements in echocardiographic parameters of both right and left ventricular function (Tables
2,
3 and
4).
A previous randomized trial evaluated the effects of 6 weeks of APAP in 26 patients with chronic stable HFrEF but failed to document any significant improvement in cardiopulmonary exercise capacity and quality of life, and did not report respiratory or sleep parameters [
21]. However, the treatment duration in that study (6 weeks) was much shorter than in our study and may have been insufficient to show any benefit of APAP on cardiopulmonary exercise. Furthermore, device use in the previous study was low, averaging 3.5 h per night and this might also have been not enough to obtain potential benefit from APAP therapy, while patients in our study revealed good device use averaging 6.0 ± 1.6 h per night. The importance of device use in realizing the beneficial effects of APAP was indicated in our study by the fact that the per-protocol analysis showed a significant improvement in percent-predicted peak VO
2 in the APAP group, but a secondary analysis including patients with exercise capacity data at only 3 months failed to reach statistical significance. This suggests that only patients with ongoing use of APAP obtained greater benefit from therapy.
Our results were similar to those from another randomized study investigating use of CPAP for 3 months in patients with HF and LVEF < 55% [
22]. As in our trial of APAP, CPAP treatment was associated with significant improvements in LVEF and quality of life compared with controls, and the magnitude of changes in LVEF was similar to that seen in our study (≈ 5%) [
22]. Improvements of > 5% in the LVEF during drug therapy for HF have been shown to predict improved survival and decreased HF-related hospitalization [
23]. Therefore, the effects of APAP on left ventricular function documented in our study have the potential to be clinically meaningful.
Another important finding from this study is that patients treated with APAP had a significant improvement in average oxygen saturation compared with baseline and the control group. This is also clinically relevant, because hypoxemic burden has been shown to be a robust and independent predictor of mortality in HFrEF [
11].
Multiple comorbidities are common in patients with HFrEF and these show a complex and bidirectional relationship with HF [
24]. Management of these comorbidities, including SDB, is a key component of modern holistic care for HFrEF [
25]. OSA is often undiagnosed in patients with HFrEF but has been shown to be associated with increased mortality [
26]. It can be difficult to identify typical OSA symptoms in HFrEF patients because the two conditions share a number of common symptoms, including fatigue, sleepiness and exercise intolerance. This means that traditional tools, such as the ESS, often fail to identify OSA in patients with cardiovascular disease [
27]. In our study, APAP had no effect on the ESS score, but patients did not have excessive daytime sleepiness at baseline (mean ESS score of just over 7 whereas scores > 10 indicate sleepiness).
The results of cardiopulmonary exercise testing have been shown to predict prognosis in patients with HFrEF [
28]. In particular, peak VO
2 is an important predictor of survival in HF [
29]. In the multicenter Heart Failure and a Controlled Trial Investigating Outcomes of Exercise Training (HF-ACTION) trial [
30], exercise training was associated with a 6% increase in peak VO
2 over 3 months and a significantly lower risk of cardiovascular mortality or HF hospitalization after adjustment for highly prognostic baseline characteristics (hazard ratio versus control of 0.85, 95% confidence interval 0.74–0.99;
p = 0.03) [
30]. Our study patients had an initial peak VO
2 of 15.01 mL/kg/min and through APAP intervention peak VO
2 improved to 16.17 mL/kg/min, which represents a total improvement of 1.16 ± 2.18 mL/kg/min (7.7%). Although we missed the aspired 10% difference in peak VO
2 values between APAP and control group, we can demonstrate a clinically important 7.7% improvement in our trial, because each 6% increase in peak VO
2 over only 3 months was shown to be accompanied with significant reduction in the risk of all-cause mortality and all-cause hospitalization [
31]. Another study showed that peak VO
2 was a strong predictor of mortality [
31]. After adjustment for other significant predictors, each 6% increase in peak VO
2 was associated with a 5% reduction in the risk of all-cause mortality and all-cause hospitalization (
p < 0.001) [
31].
Although our study failed to demonstrate statistical significance for all three pre-defined primary endpoints [
13], our study results picture a very important finding with significant improvement specifically for percent-predicted peak VO
2. Percent-predicted peak VO
2 has been shown to be superior to other endpoints. Ross Arena et al clearly demonstrate percent-predicted peak VO
2 to be the most interesting variable [
32] because in their analysis, percent-predicted peak VO
2 value derived from the Wasserman/ Hansen equations had outperformed all other expressions of cardiopulmonary exercise testing variables [
32]. This is of particular interest because our study population reveals small differences in age and weight. Although these baseline characteristic differences are not statistically significantly different, one may argue that they may influence our study results. Percent-predicted peak VO
2 controls for such variables, making percent-predicted peak VO
2 the most important variable, supporting the meaningful impact of our study result as this endpoint has effect on mortality [
32].
The largest trial on mortality in this field is the SERVE HF [
33] study which investigated predominant central sleep apnea in HFrEF patients through application of adaptive servo-ventilation. Our trial studied predominant obstructive sleep apnea in HFrEF through application of APAP therapy, which makes comparison of both studies extremely difficult as two different disease patterns have been explored. However, the SERVE HF study [
33] revealed adverse events for the use of adaptive servo-ventilation, while our study results are encouraging APAP treatment in HFrEF patients with predominant obstructive sleep apnea and our results yield hope for further trials to improve outcome in HFrEF patients by treating obstructive sleep apnea.
Our randomized, controlled clinical pilot study showed that use of APAP therapy for 6 months in patients with HFrEF and OSA had beneficial effects, not only on an important primary endpoint, but also on a number of surrogate markers for cardiovascular outcome. Nevertheless, there is a need for more randomized controlled trials to confirm the effects of positive airway pressure therapy on cardiopulmonary exercise capacity, quality of life, global cardiac function, and hypoxemic burden, and to determine whether these improvements translate into better outcomes for HFrEF patients with OSA.