Author + information
- Received March 13, 2007
- Revision received May 29, 2007
- Accepted June 18, 2007
- Published online October 2, 2007.
- Andrew Cassar, MD, MRCP⁎,
- Timothy I. Morgenthaler, MD, FCCP†,
- Ryan J. Lennon, MS‡,
- Charanjit S. Rihal, MD, FACC§ and
- Amir Lerman, MD, FACC§,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Amir Lerman, Division of Cardiovascular Disease and Department of Internal Medicine, Mayo College of Medicine, 200 First Street SW, Rochester, Minnesota 55902.
Objectives Our purpose was to compare outcomes of patients treated for obstructive sleep apnea (OSA) versus patients with untreated OSA, all of whom had undergone percutaneous coronary intervention (PCI).
Background Obstructive sleep apnea has been associated with increases in fatal and nonfatal cardiovascular events. It is not known whether treatment of OSA in patients who have had PCI results in a better outcome.
Methods In a retrospective cohort study, a group of patients with OSA diagnosed with polysomnography between 1992 and 2004 (apnea-hypopnea index ≥15) who subsequently underwent a PCI (n = 371) were stratified according to whether they were treated for OSA (n = 175) or not (n = 196). Main outcome measures were cardiac death, general mortality, major adverse cardiac events (MACE) (severe angina, myocardial infarction, PCI, coronary artery bypass grafting, or death), and major adverse cardiac or cerebrovascular events (MACCE).
Results Patients treated for OSA had a statistically significant decreased number of cardiac deaths on follow-up when compared with untreated OSA patients (3% [95% confidence interval (CI) 0% to 6%] vs. 10% [95% CI 5% to 14%] after 5 years, p = 0.027), as well as a trend toward decreased all-cause mortality (p = 0.058). There was no difference in the number of MACE or MACCE between the 2 groups (p = 0.91 and 0.96, respectively).
Conclusions Treatment of OSA is associated with a reduction in the number of cardiac deaths, but not in MACE or MACCE, after PCI. Screening for and treating OSA in patients with coronary artery disease who may undergo PCI may result in decreased cardiac death.
Both sleep-disordered breathing and coronary artery disease (CAD) are common health problems. Sleep-disordered breathing affects approximately 24% of men and 9% of women between the ages of 30 to 60 years (1). Coronary artery disease is a leading cause of death (2) in this and older age groups. Obstructive sleep apnea (OSA) has been associated with a 70% relative increased risk of cardiovascular morbidity and mortality (3).
To clarify whether treatment of OSA is associated with a better prognosis in patients who have undergone percutaneous coronary intervention (PCI), we performed a retrospective analysis of the Mayo Sleep and PCI registries. The current study was designed to test the hypotheses that in patients who had undergone PCI, those treated for OSA would have improved outcomes compared with those not treated for OSA.
Study design, study population, and data collection
After approval by the Mayo Clinic Institutional Review Board, all patients who had undergone a polysomnogram at the Mayo Clinic, Rochester, Minnesota, from 1992 to 2004 were identified (Fig. 1).This database of 23,602 patients was crossed with the PCI database of the Mayo Clinic to identify those patients (n = 519) who had a PCI up to the end of 2005. Only those patients with an apnea-hypopnea index (AHI) of ≥15 were selected (n = 376). Patients with central sleep apnea (n = 5) were excluded. The patients remaining (n = 371) by definition had moderate (AHI 15 to 30) OSA (n = 113) or severe (AHI >30) OSA (n = 258) (4). These patients’ medical records were reviewed for compliance to treatment for OSA, and the patients were classified as treated (n = 175) or untreated (n = 196). The treated group was composed of 3 subgroups: 1) patients who at follow-up visits showed compliance to treatment >3 months after initiation of continuous positive airway pressure (CPAP) (n = 141); 2) patients who showed initial compliance to treatment to CPAP over the first 3 months (n = 31) but further information about compliance beyond 3 months was not available; and 3) patients treated surgically with uvulopalatopharyngoplasty (n = 3). Those patients who declined treatment, did not show compliance to treatment, or in whom no information about treatment was available were assigned to the untreated group (n = 196). Follow-up of patients was done with a telephone call or questionnaire by mail at 6 months, 1 year, and yearly thereafter following their PCI procedure (5). The medical records were reviewed by one investigator, and the patients’ characteristics, risk factors, and procedural details were extracted. Six months of follow-up is available in 95% of the patients and at least 12 months of follow-up in 89%. We compared the outcomes between untreated and treated OSA patients who underwent PCI.
The main outcome measures of interest were cardiac death, general mortality, major adverse cardiac events (MACE) (severe angina, myocardial infarction [MI], PCI, coronary artery bypass grafting [CABG], or death), and major adverse cardiac or cerebrovascular events (MACCE) (MACE or stroke).
Continuous variables are presented as mean ± standard deviation unless otherwise noted. Discrete variables are summarized as frequency (group percentage). Kaplan-Meier estimates are used to summarize time-to-event variables, with day of PCI as the start date. Comparisons between groups are made using 1-way analysis of variance, Pearson’s chi-square statistic, and the log-rank test, respectively. All tests are 2-tailed with a 0.05 significance level. The overall type I error rate may be >0.05 due to multiple outcomes being tested. Analyses were completed with SAS 9.1 software (SAS Institute, Cary, North Carolina).
The baseline patient characteristics for the treated and untreated groups of patients with OSA are shown in Table 1.The treated group had more severe OSA than the untreated group at baseline as indicated by a higher average AHI (50 vs. 39, p = 0.013) and the longer percentage time with an oxygen saturation below 90% (8.9% vs. 5.5%, p = 0.06). Body mass index was higher in the treated group (35.4 ± 7.1 kg/m2vs. 33.0 ± 5.6 kg/m2, p < 0.001). Otherwise the 2 groups were similar for age, gender, family history of CAD, smoking, diabetes mellitus, hypertension, hyperlipidemia, previous MI, PCI or CABG, peripheral arterial disease, cerebrovascular disease, chronic obstructive pulmonary disease, and renal failure. The percentage of patients presenting to PCI with acute coronary syndrome or acute MI was similar in both groups. There was no statistical difference in use of medications between the 2 groups.
Angiographic and procedural characteristics and outcome
There were no differences between the 2 groups in the percentage of urgent/emergent PCI, number of vessels diseased, total number of vessels treated, number of stents and type used, glycoprotein IIb/IIIa use, and Thrombolysis In Myocardial Infarction score before and after PCI.
As regards complications during or just after PCI, the treated sleep apnea group had less episodes of bradycardia during the PCI than the untreated sleep apnea group (1% vs. 7%, p = 0.007). All other arrhythmias, in-hospital death, and in-hospital revascularization were similar for the 2 groups.
Morbidity and mortality
Over a median follow-up of at least 3 years, the treated OSA group had a statistically significant reduced number of cardiac deaths when compared with untreated OSA (3% [95% confidence interval (CI) 0% to 6%] vs. 10% [95% CI 5% to 14%] after 5 years, p = 0.027) (Table 2,Figs. 2 and 3).⇓⇓Comparison of all-cause mortality between the treated and untreated sleep apnea groups was just short of significance (p = 0.058). At 5 years post-PCI, 3% in the treated OSA group suffered cardiac death compared with 10% in the untreated group. A similar difference was observed for all-cause mortality with 11% and 17%, respectively. There was no difference in the number of MACE or MACCE between groups (p = 0.91 and 0.96, respectively).
After subgroup analysis, the severe (AHI >30) OSA patients (131 treated vs. 127 untreated) also had a significant decrease in cardiac death (3% [95% CI 0% to 6%] vs. 9% [95% CI 3% to 14%] at 5 years, p = 0.045) but not mortality, MACE, or MACCE. There were no significant differences between the moderate (AHI 15 to 30) OSA-treated (n = 44) and untreated (n = 69) groups (for cardiac death 3% [95% CI 0% to 6%] vs. 11% [95% CI 1% to 20%] at 5 years, p = 0.24). This was likely due to the smaller number of moderate OSA patients under study.
Our study shows that patients with treated OSA undergoing PCI have a significantly lower cardiac death rate than patients with untreated OSA. The current study further supports an important role for the diagnosis and treatment of OSA in patients with CAD.
Obstructive sleep apnea is thought to be mechanistically linked to atherosclerosis through the initiation of endothelial injury by repetitive bursts of sympathetic activity, surges of blood pressure, and oxidative stress brought on by apneas and episodic hypoxemia (6). The endothelial injury promotes expression of adhesion molecules (7) with resultant leukocyte accumulation and adhesion. Elevations of the inflammatory mediators such as C-reactive protein (8), elevated plasma fibrinogen levels, exaggerated platelet activity, and reduced fibrinolytic capacity (9), increased plasma viscosity (10), and endothelial dysfunction (11) are reported in patients with OSA, even without CAD (12). Treatment of OSA has been associated with reductions in circulating inflammatory and thrombogenic factors (13) and improvement in endothelial dysfunction (14), suggesting that treatment of OSA should retard progression of atherosclerosis. Consistent with this, some studies suggest that treatment of OSA reduces cardiovascular mortality and event rates (15). Hayashi et al. (16) determined that the oxygen desaturation index over the night after PCI was the most significant independent determinant of loss of lumen after 3 to 6 months.
Our findings extend those of previous studies that demonstrated an increased risk of both fatal and nonfatal cardiovascular events in men with untreated severe sleep apnea when compared with patients treated with CPAP (15). However, patients in the study by Marin et al. (15) were not selected on the basis of known CAD, whereas ours were limited to patients already showing advanced CAD. Our study adds further information to the study of Milleron et al. (17), which showed a significantly decreased incidence of new cardiovascular events in patients with OSA and CAD with CPAP treatment, but our study includes more patients. In this context, it is noteworthy that we did not find a significant overall difference in MACE or MACCE between treated and untreated patients. This may be secondary to the demographic differences between the studies. In our study, all patients included had undergone PCI for CAD (70% for acute coronary syndrome) and our population included 10% to 15% women. Our patients were put on a more aggressive medical plan with aspirin, clopidogrel, statins, and beta-blockers than those in the studies referenced in the preceding text. This treatment, together with a shorter follow-up period (at least 3 years) compared with other studies, may have resulted in decreased MACE events in our patients resulting in no significant difference in events between our treated and untreated OSA patients.
Apart from increased mortality from acute cardiovascular events (3), patients with sleep apnea are also at increased risk of sudden cardiac death due to arrhythmias. Gami et al. (18) described a 2.57 relative risk increase in cardiac death between midnight and 6 amin patients with OSA. The presumed mechanism (19) is that repetitive apnea-related hypoxia and arousals stimulate the sympathetic nervous system resulting in surges in blood pressure, increased afterload, and myocardial oxygen demand. This may, in turn, provoke myocardial ischemia and arrhythmias with increased risk of sudden cardiac death. Studies have shown that CPAP treatment might abolish this trend (20). Obstructive sleep apnea is also associated with an increased prevalence of atrial fibrillation (21), which is, in turn, associated with an increased risk of death (22).
Studies have demonstrated a higher prevalence of OSA and nocturnal oxygen desaturation in patients with CAD when compared with control subjects (23,24). Mooe et al. (25) showed a statistically significant higher occurrence of sleep-disordered breathing in patients with CAD compared with control patients (37% vs. 20%). With over one-third of patients with CAD having sleep-disordered breathing, screening for and subsequently treating OSA in patients who may undergo PCI for CAD might be beneficial in terms of decreasing mortality. A recent study by Wang et al. (26) recommends such screening and treatment of OSA in patients with heart failure.
Limitations to our study include its retrospective nature. Treatment was not randomly allocated. All patients were offered CPAP, but the patients who received treatment were those who accepted and could tolerate CPAP. Patients were considered “treated” if they were compliant to treatment during the first 3 months after starting CPAP and “untreated” if they were not. It is possible that some patients may have been mistakenly classified as treated due to difficulty in assessing long-term CPAP compliance especially before the new CPAP machines with in-built compliance meters were introduced. Also, patients who were classified as untreated may have started CPAP treatment at a later date. Our retrospective method could have compared patients who were more health conscious and compliant with therapies in general with less adherent patients. Our study also does not demonstrate that treatment for OSA after PCI may decrease mortality, as OSA diagnosis and treatment occurred before the PCI. A randomized prospective trial to test such a hypothesis would be ideal but may not be possible, since withholding CPAP therapy from patients with OSA is not justifiable given the data available.
Treatment of OSA in patients who then undergo PCI resulted in a statistically significant reduction in cardiac death but not in MACE or MACCE when compared with untreated OSA patients. Screening for, and treating, OSA in patients with CAD undergoing PCI may result in decreased cardiac death and overall mortality.
The authors would like to thank Ms. LaVon N. Hammes and Mr. Jon L. Kosanke for their help in extracting data from the Mayo Clinic PCI database.
- Abbreviations and Acronyms
- apnea-hypopnea index
- coronary artery bypass graft
- coronary artery disease
- confidence interval
- continuous positive airway pressure
- major adverse cardiac or cerebrovascular events
- major adverse cardiac events
- myocardial infarct
- obstructive sleep apnea
- percutaneous coronary intervention
- Received March 13, 2007.
- Revision received May 29, 2007.
- Accepted June 18, 2007.
- American College of Cardiology Foundation
- Rosamond W.,
- Flegal K.,
- Friday G.,
- et al.
- Jaber W.A.,
- Lennon R.J.,
- Mathew V.,
- Holmes D.R. Jr..,
- Lerman A.,
- Rihal C.S.
- Shamsuzzaman A.S.,
- Winnicki M.,
- Lanfranchi P.,
- et al.
- Yokoe T.,
- Minoguchi K.,
- Matsuo H.,
- et al.
- Milleron O.,
- Pilliere R.,
- Foucher A.,
- et al.
- Gami A.S.,
- Pressman G.,
- Caples S.M.,
- et al.
- Benjamin E.J.,
- Wolf P.A.,
- D’Agostino R.B.,
- Silbershatz H.,
- Kannel W.B.,
- Levy D.
- Wang H.,
- Parker J.D.,
- Newton G.E.,
- et al.