OUP user menu

Obstructive sleep apnoea–hypoapnoea syndrome reversibly depresses cardiac response to exercise

Alberto Alonso-Fernández , Francisco García-Río , Miguel A. Arias , Olga Mediano , José M. Pino , Isabel Martínez , José Villamor
DOI: http://dx.doi.org/10.1093/eurheartj/ehi621 207-215 First published online: 2 November 2005


Aims To evaluate cardiac response to exercise in middle-aged normotensive obstructive sleep apnoea–hypoapnoea syndrome (OSAHS) adults with normal resting left ventricular systolic function and to test the hypothesis that nasal continuous positive airway pressure (CPAP) therapy might improve cardiac performance during exercise.

Methods and results We performed a prospective, randomized, double-blind, placebo-controlled, cross-over clinical trial including 31 consecutive newly diagnosed OSAHS patients and 15 healthy subjects. Cardiopulmonary exercise testing with cardiac output measurement, blood pressure (BP) recordings, and urinary excretion of catecholamine levels were obtained at baseline and after 3 months on both effective and sham CPAP. OSAHS subjects had higher systolic and mean nocturnal BP and higher nocturnal levels of catecholamines. In contrast, they had lower increments in cardiac output (Qt) and in stroke volume (SV) in response to exercise than control subjects. CPAP therapy was associated with highly significant improvements in all the indices of left ventricular systolic performance response during exercise, whereas with sham CPAP, all of them remained unchanged.

Conclusion OSAHS patients with normal resting left ventricular systolic function and no hypertension had a worse cardiac response to exercise than healthy subjects. In these patients, 3 months of CPAP improved both Qt and SV responses to exercise.

  • Obstructive sleep aponea
  • Cardiac output
  • Left ventricular function
  • Exercise
  • Continuous positive airway pressure


Obstructive sleep apnoea–hypoapnoea syndrome (OSAHS) is characterized by repetitive partial or complete closure of the upper airway during sleep, despite increased respiratory effort. It is a common disorder, and prevalence surveys estimate that 2% of women and 4% of men of middle age are affected by this syndrome.1 Cardiovascular disturbances are the most important complications producing severe morbidity and mortality.2 Many risk factors for OSAHS, including male gender, increasing age, and obesity, are the same for cardiovascular disease, making it difficult to recognize the role of OSAHS as an independent risk factor. At the present time, only the association with systemic hypertension has been well established.3

Congestive heart failure (CHF) remains common in developed countries and contributes appreciably to the burden of morbidity and mortality.4 The relationship between OSAHS and left ventricular dysfunction and CHF is less known, although OSAHS is frequent in systolic heart failure patients5 and cross-sectional results from a large study have shown a significant association between OSAHS and CHF.6 Proposed mechanisms affecting left ventricular performance in patients with OSAHS include several mechanical,7 neurohumoral,8 metabolic,9 inflammatory,10 oxidative stress,11 and vascular endothelial effects.12

Normally, cardiac output (Qt) increases with exercise to support the increasing metabolic demands of the tissues. In healthy subjects, Qt is a linear function of oxygen uptake (V′O2) and does not vary as a function of either sex or state of fitness. Increases in Qt are first accomplished by increases in stroke volume (SV) and heart rate (HR), and then at moderate-to-high-intensity exercise almost exclusively by increases in HR. The increase in Qt is largely driven by vagal withdrawal and by increases in either circulating or neurally produced catecholamines that elicit a greater ventricular contractile response and increments in venous return.13 It has been shown that there are lower increments in Qt and SV during exercise in CHF even at early stages of the disease when there is still no impairment.14

Some authors have suggested that REM sleep in OSAHS can produce cardiac stress as great as that produced by exercise.15 There is little information about the cardiovascular performance during exercise in OSAHS patients without other risk factors affecting the left ventricular systolic function and there are few data about the possible role of OSAHS as an independent cause for left ventricular systolic dysfunction in otherwise healthy patients. The aims of this study were (i) to evaluate cardiac response to exercise in middle-aged OSAHS adults with normal resting left ventricular systolic function and (ii) to test the hypothesis that nasal continuous positive airway pressure (CPAP) therapy, the standard treatment for OSAHS, might improve cardiac performance during exercise.


Patients and protocol

Forty-one consecutive newly diagnosed OSAHS patients [apnoea–hypoapnoea index (AHI) >10 h−1, Epworth sleepiness scale >10, and no current drug or mechanical treatment] and 15 healthy subjects (AHI <5 h−1 and Epworth sleepiness scale <10) were selected to be studied. Control subjects were recruited from a list of healthy subjects from our sanitary area that had had a routine health test in the previous 3 months. We randomly selected a control subject similar in gender, age (±2 years), weight (±2 kg), and height (±5 cm) with regard to the two preceding patients included in the study. Exclusion criteria from the study for both patients and control subjects were (i) unwillingness or inability to perform the testing procedure, (ii) problems with sleepiness when driving, and (iii) comorbid disorders or situations that could affect cardiac response to exercise (see Supplementary material online). Withdrawal criteria were (i) clinical exacerbation which leads to change in medication, (ii) hospital admission for 11 or more days, (iii) angina or ST-segment depression episodes during cardiopulmonary exercise testing, and (iv) CPAP average nightly usage <3.5 h.

Seven subjects refused to take part in the study. Three patients deemed ineligible for inclusion after initial assessments (one had acute coronary syndrome during exercise testing and two because they had unknown mitral stenosis), therefore 31 OSAHS patients were randomized (Figure 1).

Figure 1 Subjects' flow chart.

The study was approved by the Institutional Ethics Committee at the hospital. All subjects gave their written informed consent prior to enrolment. We performed a single-centre, prospective, randomized, double-blind, placebo-controlled, and cross-over clinical trial, in which patients received effective CPAP, Aria LX CPAP device (Respironics, PA, USA), and sham CPAP therapy16 for two 12-week periods. The sham CPAP device consisted of a conventional CPAP device, in which the area of the exhalation port was amplified, thereby nearly cancelling nasal pressure; an orifice resistor was connected between the tubing and the CPAP unit that loads the blower with the same airflow resistance as in effective CPAP.16 All patients underwent full-night CPAP titration studies using an automated pressure-setting device (Auto Set; ResMed, Sydney, Australia). Patients were given detailed instructions on using the CPAP equipment, but they were not informed of the type of therapy they were receiving. Compliance data were measured with a run-time counter. All subjects included in the study underwent baseline sleep study and cardiopulmonary exercise testing with cardiac output measurement. Ambulatory blood pressure monitoring (ABPM) and catecholamines in urine were determined as previously described17 (see Supplementary material online). Patients were randomized to receive either therapeutic CPAP or sham CPAP during 12 weeks. Then, they were readmitted to the hospital, and the CPAP device was switched to the alternate mode of therapy for another 12 weeks. ABPM recordings, urine specimens, demographic data, and cardiopulmonary exercise testing with cardiac output measurement were obtained just after completing the period with either effective CPAP or sham CPAP.


All subjects underwent overnight sleep studies in the sleep laboratory using a portable respiratory recording device (Sibel Home-300; Sibel S.A., Barcelona, Spain) that had been previously validated.18 More details and descriptions of each of the respiratory events are available in a data supplement.

Spirometry was performed by means of a pneumotachograph (MasterLab 6.0, Erich Jaeger GmbH, Würzburg, Germany), according to the American Thoracic Society standardization.19

Systolic function was assessed by left ventricular shortening fraction (LVSF) and left ventricular ejection fraction (LVEF)20 using high-quality echocardiograph with 2.0–4.0 MHz probes (Hewlett Packard Sonos 5500, Andover, MA, USA). LVSF ≥28% and LVEF ≥50% were considered normal.

Symptoms-limited progressive cardiopulmonary exercise testing was performed with a cycle ergometer Ergobex (Bexen, Madrid, Spain), according to the American Thoracic Society/American College of Chest Physicians standardization21 (see Supplementary material online).

During exercise, cardiac output (Qt) was measured according to the Fick equation by the CO2 rebreathing equilibrium method (Oxycon Alpha, Jaeger).22 SV was calculated with the simultaneous HR registration. The rebreathing manoeuvre was carried out at rest and at 20, 40, and 60% of the theoretical maximum work intensity (W). Increments in Qt and SV during exercise were related to increments in work intensity (W) and oxygen uptake (V′O2). More descriptions are available in a data supplement.

Statistical analysis

Values are expressed as mean±SD or percentage. All statistical tests were two-sided. The comparisons between the groups were performed by means of the Student's t-test, whereas the χ2 test was used for evaluating frequencies. Relationships between variables were determined by Pearson's correlation analysis. Comparisons of the effects of the treatments over time were made with repeated measures ANOVA, with treatment as a within-subject factor and order as a between-subject factor. When ANOVA results showed significant differences between treatment conditions, posthoc multiple comparisons were performed with the Bonferroni test. Statistical analysis was performed using the SPSS statistical package, version 10.0 for Windows (SPSS, Chicago, IL, USA). P-values of less than 0.05 were considered to be statistically significant.


Characteristics of the subjects

All subjects, except one, were men. Anthropometric characteristics, smoking habit, spirometric parameters, sleep study data, ABPM, and urinary catecholamines in OSAHS patients and control subjects are shown in Table 1. OSAHS patients had higher systolic and mean nocturnal ABPM as well as higher nocturnal levels of norepinephrine and epinephrine compared with control subjects. Furthermore, OSAHS patients had lower epinephrine nocturnal dip than healthy subjects (Table 1). No significant difference in either daytime urinary catecholamines or diurnal ABPM was found between the two groups.

View this table:
Table 1

General characteristics

Exercise response and left ventricular performance systolic function

Table 2 shows the respiratory, cardiovascular, and metabolic variables at rest and at peak exercise. No difference in any of them was found between OSAHS and healthy subjects. However, we found significant differences in left ventricular systolic function response to exercise between OSAHS patients and healthy subjects (Figure 2). In fact, OSAHS patients showed a lower Qt response to the increments in W (0.11±0.04 vs. 0.23±0.07 L/min/W, P=0.001) and in V′O2 (0.01±0 vs. 0.02 ±0.01 L/mL, P=0.001) during exercise than control subjects. We did not find significant differences in HR response to exercise between both study groups, and the poorer Qt response in OSAHS subjects was due to lower increments in SV during exercise compared with healthy subjects (0.61±0.42 vs. 1.34±0.62 mL/W and 0.05±0.04 vs. 0.10±0.05 min, P=0.001).

Figure 2 Cardiac output (Qt) and SV responses to exercise, related to work intensity (W) (A and B) and oxygen uptake (V′O2) (C and D), in OSAHS patients and control subjects. Data are presented as mean±SD. At each exercise level, comparisons between patient groups were made by analysis of variance adjusted for W or V′O2 as covariates. The Bonferroni test was employed for post hoc comparisons.

View this table:
Table 2

Exercise response

However, Qt and SV responses to exercise in OSAHS patients were positively correlated with heart rate reserve (HRR). No correlations were found among left ventricular systolic function during exercise testing and sleep parameters, spirometric data, urinary excretion of catecholamines, and blood pressure (BP).

Effects of CPAP

Five patients failed to complete the trial, leaving 26 patients for complete analysis. One patient did not tolerate CPAP, and two patients withdrew because of change of residence to another city. The two remaining patients were excluded due to average nightly usage of CPAP <3.5 h (Figure 1). There were no significant differences in either baseline anthropometric characteristics, smoking habit, lung function data, sleep study indices, sympathetic tone, or ABPM between patients who withdrew and those who completed the trial. Moreover, cardiopulmonary exercise response was not significantly different among these subgroups.

Complete measurements were available in 26 patients who went home on CPAP therapy for an average of 104±31 days and on sham CPAP for 98±22 days. The mean CPAP pressure was 8.8±1.4 mmHg, and the average nightly usage was similar on CPAP and sham CPAP (6±1 vs. 6±2 h, respectively).

Body weight remained virtually unchanged during the trial (88.6±15.2 kg on CPAP and 88.4±15.8 kg on sham CPAP). Findings for treatment effects on BP and urinary catecholamines and exercise response are detailed in Tables 3 and 4. CPAP treatment did not change either BP or urinary excretion of catecholamines. There was a significant reduction in the HR peak (and consistent upturn in HRR) after CPAP therapy (P=0.034 and P=0.014, respectively) (Table 3). There were no significant differences in the remaining variables in cardiopulmonary exercise testing with both treatment modalities.

View this table:
Table 3

Effects of CPAP on exercise response in OSAHS patients

View this table:
Table 4

Effects of CPAP on left ventricular systolic function response to exercise in OSAHS patients

In contrast, 3 months of CPAP therapy resulted in highly significant improvements in all the indices of left ventricular systolic response during exercise, whereas all of them remained unchanged during the sham CPAP period (Figure 3). In fact, CPAP therapy doubled all the increases in Qt and SV responses to exercise expressed in relation to both work intensity and oxygen uptake (Table 4).

Figure 3 Individual values of cardiac output (Qt) (A and B) and SV (C and D) responses to exercise in OSAHS patients at baseline and after sham CPAP and therapeutic CPAP. Horizontal lines represent mean values.


The present study provides evidence suggesting that OSAHS results in worse cardiac response to exercise than healthy subjects. Our major findings are (i) OSAHS patients with normal resting left ventricular systolic function and without hypertension had lower Qt and SV responses to exercise than control subjects, (ii) 3 months of CPAP led to significant improvement in cardiac response to exercise compared with placebo.

The CO2 rebreathe manoeuvre has been commonly used to assess Qt during exercise. It is a simple and non-invasive method that provides accurate determinations of Qt during exercise conditions in healthy subjects and in patients with hypertension or CHF.23,24 It has also been shown that the reliability of the CO2 rebreathe is usually similar to that of other non-invasive procedures and improves with increasing exercise intensities.23,25 Besides, it has been argued that the reliability of this method is analogous for both men and women.26 The rebreathing equilibrium method is uncomfortable to perform during heavier stages of submaximal exercise due to high concentration of CO2; therefore, its use during maximal exercise is limited.23,24 However, we measured Qt only at 20, 40, and 60% of the theoretical maximum W, and none of the patients stopped before the last measurement was taken. In OSAHS patients, we have shown a significant relationship between Qt and SV at rest (using the CO2 rebreathe manoeuvre) with LVSF (r=0.648, P<0.001 and r=0.567, P<0.01, respectively) and with LVEF (r=0.690, P<0.001 and r=0.582, P<0.01, respectively) determined by echocardiography.27

A variety of exercise responses can be seen in obesity, depending mostly on its severity. V′O2 peak may be reduced when expressed per kilogram of actual body weight or normal when expressed per kilogram of ideal body weight. There is an excessive metabolic requirement but with a normal slope in obesity.28 Furthermore, a lower cardiac index (also determined by the CO2 rebreathing method) has been shown during exercise in obese patients compared with normal subjects.29 Nevertheless, they were more obese [mean body mass index (BMI) 39.9 kg/m2] than the patients of the present study, the cardiac index was measured below anaerobic threshold (AT) and the authors did not rule out OSAHS. We think that obesity did not seem to play a major role in less-efficient cardiac performance in the present article, because both study groups had similar BMI, which did not change throughout the trial.

Physical deconditioning might also be a confounder, but peak V′O2 and peak oxygen pulse were analogous in OSAHS patients and healthy subjects, and there were no differences in AT, HR slope, and HRR among both study groups (Table 2). Therefore, the authors believe that this factor did not contribute to their results. In addition, no significant difference in SaO2, minute ventilation, respiratory frequency, breathing reserve, ventilatory equivalent for carbon dioxide, or ratio of physiological dead space to tidal volume between both groups was found (Table 2). Thus hypoxaemia, inefficient ventilation or increased dead space ventilation, did not bias the exercise response pattern of our patients.

There is a substantial evidence that sleep-related breathing disorders are common in CHF;5 however, it remains unclear whether OSAHS per se could cause CHF. Previous studies have reported some data about the relation between OSAHS and left ventricular systolic function.3032 The effects of cardioactive drugs, systemic hypertension, diabetes mellitus, or other possible diseases affecting left ventricular systolic function could influence the results in some studies. Indeed, the absence of a control group of subjects and lack of placebo-controlled design of the studies constitute their limitations.

Two previous studies have reported cardiac response to exercise in OSAHS patients, assessed by radionuclide ventriculography.15,33 It is difficult to generalize the results from these studies to other settings, because they had small samples and were non-controlled studies.15,33 Moreover, all of the subjects in one had hypertension as well as elevated left ventricular end-systolic and left atrial end-diastolic diameters.33 However, as we found in our study, OSAHS was shown to be associated with impaired cardiac performance during exercise.

OSAHS has several detrimental effects that may impair left ventricular systolic function. Inspiration against an occluded upper airway generates exaggerated negative intrathoracic pressure, which increases systolic transmural pressure, leading to both an increase in left ventricular afterload and a decrease in left ventricular pre-load. The net effect is a reduction in SV and Qt.15,33 Hypoxia during apnoea can also elicit myocardial ischaemia and arrhythmias,34 and may even reduce myocardial contractility. Apnoea-induced hypoxia, hypercapnia, and arousals from sleep increase sympathetic system activity, which results in systemic vasoconstriction and further increases in left ventricular afterload.33 Accordingly, inflammatory cytokines10 and endothelial dysfunction12 may also play a role in the development of left ventricular dysfunction.

A low SV response to exercise could be due to a decreased pre-load, contractility, or an increased afterload. Some studies have shown that patients with OSAHS with or without mild pulmonary hypertension at rest could develop more severe pulmonary hypertension with exercise.35,36 This could result in reduced left heart return, resulting in a relatively lower pre-load, and consequently, a decrease in SV. However, peak Qt was measured and normal values found, in spite of the increments in mean pulmonary pressure.36 We do have no data about pulmonary pressure during exercise, but a pulmonary haemodynamic limitation in our group of patients is unlikely because neither the relationship between ventilation and carbon dioxide production (V′E/V′CO2) nor the AT was different from the healthy subjects.21,22 In addition, left ventricular impaired relaxation is frequently observed in OSAHS patients37 and probably contributes to decrease pre-load.

In contrast, Tryfon et al.38 demonstrated that normotensive OSAHS patients develop diastolic blood hypertension during exercise, but resting left ventricular function was not measured and other causes of hypertension such as diabetes mellitus were not excluded. In addition, an excessive rise in diastolic BP with exercise may be indicative of abnormal BP control, or may also be a sign of cardiac dysfunction if left heart function does not keep up with the increases in cardiac output.21 Unfortunately, we cannot rule out these mechanisms because we did not measure blood or pulmonary pressure during exercise. It is difficult to define how each of those mechanisms affects SV response to exercise in OSAHS, as they would act in synergy. Therefore, data from new studies are warranted to verify this hypothesis.

It is well known that CPAP application in heart failure patients abolishes intermittent apnoea-related hypoxia,39,40 lowers systolic BP,39 and reduces sympathetic tone40 and left ventricular afterload, by elevations in intrathoracic pressure,41 leading to improvement in LVEF.

In OSAHS subjects without heart failure, it has also been shown that CPAP decreases muscle sympathetic traffic41 and increases LVEF, even after long-term treatment,30,31 and right ventricular ejection fraction.42 Furthermore, it improves diastolic left ventricular function37 and endothelial function.12 Nevertheless, caution should be taken when generalizing these results, because most of them are uncontrolled studies. The mechanisms underlying the improvement in Qt and SV during exercise may include the above data. However, we were unable to demonstrate a significant improvement in both BP and urinary catecholamine levels with CPAP. This may be explained because patients in the present study were normotensive with normal resting LVEF, which would have limited the magnitude of any further improvement. In contrast, it could also be because urinary excretion of norepinephrine and epinephrine and ABPM lack sensitivity to detect the elimination of surges in BP and nocturnal sympathetic nerve traffic drive in OSAHS patients receiving CPAP. In our study, however, there was a fall in HR peak and an increase in HRR after 3 months on CPAP (with similar V′O2 peak) that could show better cardiovascular adaptation to exercise. It could be hypothesized that the demonstrated objective benefits of CPAP on intrathoracic pressure41 and a ‘direct’ effect on systolic ventricular function could play a main role, but this must be further evaluated.

Our results could represent an early pathological cardiac manifestation, prior to the development of CHF, related to cyclic obstructive events. It is tempting to speculate that repeated nocturnal hypoxaemia, abrupt decreases and increases in cardiac output, and sympathetic activation in association with changes in intrapleural pressure over many years could progress to structural ventricular changes that may lead to symptomatic CHF if OSAHS is left untreated or if additional risk factors develop over time. In fact, the improvement in cardiac response to exercise after 3 months on CPAP in a randomized controlled trial supports the pathogenic relationship between OSAHS and left ventricular performance impairment during exercise. However, further studies are needed to evaluate whether CPAP prevents cardiovascular morbidity and heart failure in these patients.

Supplementary material

Supplementary material is available at European Heart Journal online.


This work was supported by F.I.S. (exp. 01/0278) and NEUMOMADRID (2000) grants. We thank A. Alvarez. P. Librán, J. Lacacci, M.J. Martín, A. Pérez, and C. Suárez for technical assistance.

Conflict of interest: none declared.


View Abstract