Few studies have examined exercise capacity or cardiovascular responses to maximal exercise testing and recovery in patients with sleep-disordered breathing (SDB), and results from these studies are conflicting. The objective of this cross-sectional study conducted at a tertiary referral center was to examine the association between SDB and exercise testing outcomes independent of body mass index (BMI) and other cardiopulmonary risk factors. Between January 1, 2005 and January 1, 2010, 1,424 adults underwent exercise testing and within 6 months before first-time diagnostic polysomnography. Subjects were categorized by apnea-hypopnea index (AHI) into 4 groups: <5, 5 to 14, 15 to 29, and ≥30. A logistic regression model incorporated age, gender, BMI, smoking, hypertension, diabetes, beta-blocker use, and cardiac and pulmonary disease as covariates. The primary variable of interest was functional aerobic capacity (FAC). Mean age was 56.4 ± 12.4 years; 75% were men. Mean BMI was 32.4 ± 7.1 kg/m², and mean AHI 19.5 ± 22.1 per hour. On multivariate analysis, AHI as a continuous variable showed a negative correlation with FAC (R²adj = 0.30, p <0.001) and postexercise SBP (R²adj = 0.23, p = 0.03), and positively correlated with resting and peak DBP (R²adj = 0.09, p = 0.01 and R²adj = 0.09, p = 0.04 respectively). When comparing patients with severe SDB (AHI ≥30) with those without SDB (AHI <5), FAC and heart rate recovery were significantly lower, and resting, peak, and postexercise DBP were higher in those with severe apnea (all p <0.05), after accounting for confounders. In conclusion, SDB severity was associated with reduced FAC and increased resting and peak DBP. Even after accounting for confounders, severe SDB was associated with attenuated FAC, impaired heart rate recovery, and higher resting, peak, and postexercise DBP.
We hypothesized that decreased exercise capacity is a possible consequence of the cardiopulmonary and deconditioning effects of sleep apnea. Few studies have examined exercise capacity and responses to maximal exercise testing and recovery in patients with sleep-disordered breathing (SDB). The limited available data are derived from small studies, and results are conflicting. Furthermore, in most previous studies, it is difficult to discern the effects of SDB independent of BMI and cardiopulmonary disease. We aimed to conduct a cross-sectional study at our center to examine the association between SDB and exercise testing outcomes independent of body mass index (BMI) and other cardiopulmonary risk factors.
Methods
Patients who were referred to our center for comprehensive exercise testing between January 1, 2005 and January 1, 2010 were identified from the Cardiovascular Health Clinic (CVHC) database. Of these, subjects who underwent first-time diagnostic polysomnography (PSG) at the Center for Sleep Medicine at our facility within 6 months following the appropriate exercise test were identified using the relevant International Classification of Diseases (9th revision) procedure codes. Verification that this was the first diagnostic PSG and that the subject was treatment-naive was made by detailed review of the individual electronic medical records. Patients with amyloid or sarcoid heart disease; liver, kidney, or cardiac transplant; on dialysis; or with a history of previous lung resection were excluded.
Information on comorbidities and medications was collected by the clinician assessing the patient at the time of presentation to the CVHC for exercise testing. This was performed through patient interview and review of the electronic medical record, including results of other testing such as echocardiogram or pulmonary function tests where available. PSG measures collected by review of the individual sleep study reports were apnea-hypopnea index (AHI) and sleep efficiency. Patients were divided into 4 subgroups based on their AHI (<5, 5–14, 15–29, and ≥30). PSGs were manually scored by registered PSG technologists and reviewed by physicians board certified in sleep medicine, using standard American Academy of Sleep Medicine criteria.
Most patients underwent a treadmill exercise test using the Bruce, Naughton, or modified Naughton protocols. A minority had cycle ergometry. Exercise testing variables included predicted maximum exercise time, maximum exercise time achieved, predicted metabolic equivalents (METs), METs achieved, and functional aerobic capacity (FAC), the main outcome measure. FAC was calculated using a nomogram based on age, sex, baseline activity level and observed duration of exercise. Data on resting heart rate (HR), peak exercise HR, 1-minute post–peak exercise HR, resting systolic blood pressure (SBP) and diastolic blood pressure (DBP), peak exercise SBP and DBP, and 3-minute post–peak exercise SBP and DBP were collected. Heart rate recovery (HRR), SBP, and DBP recovery were calculated as the difference or ratio between peak exercise and postexercise values, respectively. Where applicable, reason for termination of the test, electrocardiogram abnormality, grade of dyspnea or ventricular arrhythmia, and chest pain index were recorded. Most patients were referred for an exercise test to the CVHC for symptoms of fatigue, chest pain, and palpitations or for testing before prescribing exercise in sedentary individuals. Forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC) were measured if required depending on test indication (e.g., complex cardiovascular disease, risk stratification in heart failure, dyspnea on exertion, or in athletes).
Covariates in the logistic regression analyses included age, gender, BMI (kg/m²), history of current smoking, hypertension, diabetes mellitus, beta-blocker use, cardiac disease (including coronary artery disease, cardiomyopathy, heart failure, known arrhythmia, valvular heart disease, congenital heart disease), and pulmonary disease (chronic obstructive pulmonary disease, asthma, restrictive lung disease, previous pulmonary embolism or presence of pulmonary artery stenosis, other pulmonary disease).
The primary variable of interest assessed was FAC. Secondary variables of interest were HR, peak exercise HR, 1-minute post-peak exercise HR, HRR, resting SBP and DBP, peak exercise SBP and DBP, 3-minute post-peak exercise SBP and DBP, and SBP and DBP recovery, electrocardiogram abnormality, and grade of ventricular arrhythmia. Univariate and multivariate analyses were conducted using AHI as a continuous variable to assess for a dose-response relationship between severity of SDB and the variable of interest. In subsequent analyses, subjects with severe SDB were compared with those without SDB using logistic regression analyses accounting for the complete list of confounders described earlier. Statistical analyses were performed using JMP software version 8 (SAS Institute, Cary, North Carolina).
The study was approved by the institutional review board, and all subjects provided research consent. No external funding sources were used.
Results
Between January 1, 2005 and January 1, 2010, 41,310 exercise tests were performed at the CVHC; 3,901 subjects underwent a first-time diagnostic polysomnogram during this time frame, and of these 2,716 had an exercise test. Reviewing these tests, we found 1,463 patients who underwent a first time diagnostic PSG within 6 months following the exercise test. After exclusions, the number of subjects analyzed was 1,424. The mean age of the subjects was 56.4 years (SD 12.4), and 75% were men. Subjects with moderate or severe SDB tended to be older with a higher BMI, and a greater proportion were male, compared with those with mild or no SDB. The most common comorbidity was hypertension ( Table 1 ).
| Characteristic Mean (SD) | AHI <5 (n = 339) | AHI 5–14 (n = 468) | AHI 15–29 (n = 311) | AHI ≥30 (n = 306) | All Subjects (n = 1,424) |
|---|---|---|---|---|---|
| Age (yrs) | 50.8 (12.4) | 56.4 (11.4) | 58.4 (12.0) | 60.4 (12.1) | 56.4 (12.4) |
| Men | 64% | 73% | 80% | 86% | 75% |
| BMI (kg/m²) | 30.7 (7.0) | 32.1 (6.8) | 32.6 (6.6) | 34.4 (7.2) | 32.4 (7.1) |
| Waist/hip ratio | 0.92 (0.09) | 0.96 (0.09) | 0.97 (0.13) | 0.99 (0.09) | 0.96 (0.10) n = 1,393 |
| Sleep efficiency (%) | 76.2 (14.4) | 74.1 (14.4) | 72.6 (15.8) | 69.3 (16.7) | 73.3 (15.4) |
| AHI (events/h) | 1.8 (1.2) | 8.4 (2.8) | 21.0 (4.3) | 54.4 (22.0) | 19.5 (22.1) |
| Beta blockers | 19.2% | 27.1% | 30.9% | 42.2% | 29.3% |
| Coronary artery disease | 11.8% | 15.2% | 18.3% | 25.8% | 17.3% |
| Valvular heart disease | 2.1% | 2.8% | 4.8% | 5.2% | 3.6% |
| Arrhythmia | 13.9% | 19.0% | 20.3% | 25.8% | 19.5% |
| Cardiomyopathy/heart failure | 7.4% | 11.8% | 13.2% | 20.3% | 12.8% |
| Pulmonary disease | 8.8% | 8.8% | 5.1% | 7.8% | 7.8% |
| Diabetes | 9.7% | 11.3% | 11.5% | 19.6% | 12.9% |
| Hypertension | 28.0% | 37.8% | 38.9% | 49.7% | 38.2% |
| Current smoker | 1.2% | 3.8% | 1.6% | 2.6% | 1.9% |
| Pulmonary embolism/Pulmonary artery stenosis | 0.9% | 2.1% | 1.0% | 0.7% | 0.8% |
| Congenital heart disease | 3.8% | 1.9% | 2.9% | 4.6% | 3.2% |
One hundred seven patients had a cycle ergometer test. All other patients underwent a treadmill test, using the Bruce (n = 790) or Naughton or modified Naughton (n = 78) protocols, and 499 patients had an oxygen consumption test. The most common reason for termination was the development of fatigue or other symptoms (n = 1,325). FEV1 and FVC were measured at the time of the exercise test in 198 subjects. All other exercise test measures were available for the entire cohort (n = 1,425). Maximum exercise time and METs achieved negatively correlated with AHI (all ps <0.001) and decreased across AHI categories from no SDB to severe SDB.
Overall, the mean FAC of the cohort was 76.3% (SD 22.1) and showed a progressive decline across AHI categories from no SDB to severe SDB. Univariate analysis results using AHI as a continuous variable showed that AHI negatively correlated with FAC (adj. R 2 = 0.05, p <0.001), peak HR (adj. R 2 = 0.04, p <0.001) and HRR (adj. R 2 = 0.01, p <0.001). AHI showed a positive correlation with resting SBP (adj. R 2 = 0.002, p = 0.04) and peak DBP (adj. R 2 = 0.003, p = 0.04; Table 2 ). One-minute postexercise HR showed a negative correlation (adj. R 2 = 0.03, p <0.001) and 3-minute postexercise DBP showed a positive correlation (adj. R 2 = 0.03, p = 0.03) with AHI, but resting HR, resting DBP, peak SBP, and SBP and DBP recovery were not significantly associated with AHI (all ps >0.05).
| Measure Mean (SD) | AHI <5 | AHI 5–14 | AHI 15–29 | AHI ≥30 | All Subjects | Estimate | p Value ∗ |
|---|---|---|---|---|---|---|---|
| Functional aerobic capacity (%) | 80.7 (21.0) | 78.8 (22.2) | 76.2 (22.3) | 68.0 (20.5) | 76.3 (22.1) | −0.231 | <0.001 |
| Resting heart rate (beats/min) | 80 (15.3) | 78.9 (39.6) | 76.8 (13.7) | 78.6 (14.1) | 78.6 (25.5) | −0.007 | 0.81 |
| Peak heart rate (beats/min) | 155.2 (24.9) | 144.6 (29.5) | 142.6 (27.5) | 136.1 (28.6) | 144.9 (28.6) | −0.267 | <0.001 |
| Heart rate recovery (beats/min) | 19.4 (17.0) | 18.4 (20.4) | 17.0 (18.7) | 13.0 (10.3) | 17.2 (17.6) | −0.098 | <0.001 |
| Resting SBP (mm Hg) | 120.2 (17.8) | 121.4 (17.6) | 122.2 (17.8) | 123.3 (18.6) | 121.7 (17.9) | 0.043 | 0.043 |
| Resting DBP (mm Hg) | 75.7 (11.4) | 76.2 (11.5) | 76.5 (10.9) | 76.6 (11.1) | 76.2 (11.2) | 0.015 | 0.26 |
| Peak SBP (mm Hg) | 167.4 (29.3) | 166.2 (30.1) | 171.1 (32.2) | 164.8 (34.1) | 167.3 (31.3) | −0.032 | 0.39 |
| Peak DBP (mm Hg) | 68.9 (17.6) | 70.4 (17.5) | 71.7 (17.6) | 72.8 (15.2) | 70.9 (17.1) | 0.042 | 0.038 |
| Abnormal electrocardiogram | 13.0% | 20.2% | 24.0% | 24.5% | 20.2% | 0.002 | |
| Any ventricular ectopy | 14.5% | 16.7% | 18.0% | 22.8% | 17.8% | 0.038 | |
| Frequent or repetitive ventricular ectopy | 3.2% | 3.2% | 2.9% | 2.0% | 2.9% | 0.74 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
