The association of atrial fibrillation (AF) with coronary artery disease (CAD) remains controversial. In addition, the relation of AF to myocardial ischemia and outcomes in patients with known or suspected CAD referred for exercise stress testing has been poorly explored. In this study, 17,100 patients aged ≥50 years with known or suspected CAD who underwent exercise electrocardiography (n = 11,911) or exercise echocardiography (n = 5,189) were evaluated. End points were all-cause mortality, nonfatal myocardial infarction, and coronary revascularization. Overall, 619 patients presented with AF at the time of the tests. Patients with AF who had interpretable electrocardiograms had a lower likelihood of exercise-induced ischemic ST-segment abnormalities (adjusted odds ratio 0.51, 95% confidence interval 0.34 to 0.76, p = 0.001), and those with AF who underwent exercise echocardiography had a lower likelihood of new or worsening exercise-induced wall motion abnormalities (adjusted odds ratio 0.62, 95% confidence interval 0.44 to 0.87, p = 0.006). During a mean follow-up period of 6.5 ± 3.9 years, 2,364 patients died, 1,311 had nonfatal myocardial infarctions, 1,615 underwent percutaneous coronary intervention, and 922 underwent coronary artery bypass surgery. The 10-year mortality rate was 43% in patients with AF compared to 19% in those without AF (p <0.001). In multivariate analysis, AF remained an independent predictor of all-cause mortality (adjusted hazard ratio 1.45, 95% confidence interval 1.20 to 1.76, p <0.001), but not of nonfatal myocardial infarction or coronary revascularization. In conclusion, despite being associated with an apparently lower likelihood of myocardial ischemia, AF was an independent predictor of all-cause mortality in patients with known or suspected CAD referred for exercise stress testing.
Atrial fibrillation (AF) is the most common sustained cardiac rhythm disturbance, and its prevalence is increasing with the aging of the population. Patients with AF may present with symptoms suggestive of myocardial ischemia, such as chest discomfort due to a rapid and irregular heart rate, which do not necessarily translate into the presence of significant coronary artery disease (CAD). Although many of these patients are referred for stress testing, the prevalence of AF in patients with proved CAD seems to be low, and the association between AF and CAD remains controversial. In contrast, although there are data to support that the combination of AF and CAD confers a worse prognosis, the association of AF with outcomes has not been well characterized in patients with known or suspected CAD referred for exercise stress testing. Thus, our aim was to assess the relations of AF to myocardial ischemia, coronary angiographic results, all-cause mortality, and coronary ischemic events in patients with known or suspected CAD undergoing treadmill exercise testing.
Methods
This was a retrospective analysis of a prospectively collected database. From March 1995 to March 2008, we evaluated 17,100 consecutive patients with known or suspected CAD aged ≥50 years who were referred for first treadmill stress testing at our institution. Of them, 11,911 underwent exercise electrocardiography and 5,189 underwent exercise echocardiography. Demographic, clinical, and stress test data were entered in a dedicated database at the time of the tests. All patients gave informed consent before testing, and the study was approved by our local research ethics committee.
Blood pressure, heart rate, and a 12-lead electrocardiogram were obtained at baseline and at each stage of the exercise protocol. The electrocardiogram was considered interpretable in the absence of left bundle branch block, paced rhythm, preexcitation, left ventricular hypertrophy, repolarization abnormalities, or treatment with digoxin. Patients underwent a treadmill exercise test until they reached an end point, including physical exhaustion, severe angina, exercise-induced ST-segment deviation >2 mm, significant arrhythmias, severe hypertension (systolic blood pressure >240 mm Hg or diastolic blood pressure >110 mm Hg), or severe hypotensive response (decrease ≥20 mm Hg in systolic blood pressure from baseline value). Exercise protocols included the Bruce (95.6%), modified Bruce (2%), Naughton (1.1%), and other (1.4%) protocols. In patients with interpretable rest electrocardiograms, positive exercise electrocardiographic results were defined as the development of ST-segment deviation ≥1 mm that was horizontal or sloping away from the isoelectric line 80 ms after the J point. In the subset of patients who underwent treadmill exercise echocardiography, imaging was performed at rest, at peak exercise, and immediately after exercise, as previously described. Peak exercise images were acquired with the patient still exercising, when signs of exhaustion were present or an end point was achieved. The transducer was firmly placed on the apical and parasternal areas, and pressure was applied to the patient’s back with the left hand to decrease the relative movement between the transducer and the patient’s body. Peak and postexercise images were obtained using a continuous imaging acquisition system. The best quality images corresponding to each view were selected for comparison with those acquired at rest. Positive exercise echocardiographic results were defined as the appearance of new or worsening exercise-induced wall motion abnormalities, except worsening from akinesia to dyskinesia and isolated hypokinesia of the inferobasal segment.
Coronary angiography was performed at the discretion of the referring physician. A significant coronary stenosis was defined as a ≥50% luminal stenosis of the left main coronary artery or a ≥70% stenosis of any other major epicardial coronary artery. Coronary angiographic results <90 days after the tests were recorded.
Follow-up data were obtained by review of hospital databases, medical records, death certificates, and telephone interviews. End points were all-cause mortality, nonfatal myocardial infarction, and coronary revascularization procedures. We did not evaluate cardiac mortality, because the ascertainment of the cause of death may be liable to bias and misclassification.
Categorical variables were compared between groups using the chi-square test and are reported as percentages. Continuous variables were assessed using the unpaired Student’s t test or the Mann-Whitney U test as appropriate and are reported as mean ± SD. Survival free of the end point of interest was estimated using the Kaplan-Meier method, and survival curves were compared using the log-rank test. Multivariate logistic regression analyses were performed to determine the adjusted odds ratios for the association of AF with myocardial ischemia on exercise electrocardiography and exercise echocardiography, as well as with the presence of angiographically proved CAD. Adjusted hazard ratios for the association of AF with the end points were estimated using Cox proportional-hazard models. Whenever appropriate, multivariate analyses were adjusted by age, gender, diabetes mellitus, hypertension, hypercholesterolemia, smoking habit, family history of CAD, previous myocardial infarction, previous percutaneous coronary intervention, previous coronary artery bypass grafting, typical angina, left bundle branch block, β blockers, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, nitrates, calcium channel blockers, digoxin, diuretics, exercise-induced chest pain, exercise electrocardiographic results, METs, peak systolic blood pressure, and percentage of maximum age-predicted heart rate. Multivariate analyses performed in the subgroup of patients who underwent exercise echocardiography were also adjusted by rest left ventricular ejection fraction and exercise-induced wall motion abnormalities. A p value <0.05 was considered significant. Statistical analyses were carried out using SPSS version 15.0 (SPSS, Inc., Chicago, Illinois).
Results
Of the 17,100 patients, 619 (3.6%) presented with AF at the time of the exercise stress tests. Baseline characteristics of the 17,100 patients are listed in Table 1 , and exercise data are listed in Table 2 .
| Variable | All Patients (n = 17,100) | AF | p Value | |
|---|---|---|---|---|
| No | Yes | |||
| (n = 16,481) | (n = 619) | |||
| Men | 10,101 (59.1%) | 9,713 (58.9%) | 388 (62.7%) | 0.06 |
| Age (years) | 64.3 ± 8.2 | 64.1 ± 8.1 | 69.2 ± 7.6 | <0.001 |
| Diabetes mellitus | 2,675 (15.6%) | 2,585 (15.7) | 90 (14.5%) | 0.44 |
| Hypertension | 8,442 (49.4%) | 8,138 (49.4%) | 304 (49.1%) | 0.90 |
| Hypercholesterolemia | 7,619 (44.6%) | 7,430 (45.1%) | 189 (30.5%) | <0.001 |
| Smokers | 3,694 (21.6%) | 3,614 (21.9%) | 80 (12.9%) | <0.001 |
| Family history of coronary artery disease | 2,032 (11.9%) | 1,975 (12%) | 57 (9.2%) | 0.04 |
| Previous myocardial infarction | 2,963 (17.3%) | 2,871 (17.4%) | 92 (14.9%) | 0.10 |
| Previous percutaneous coronary intervention | 1,550 (9.1%) | 1,506 (9.1%) | 44 (7.1%) | 0.08 |
| Previous coronary bypass | 508 (3%) | 484 (2.9%) | 24 (3.9%) | 0.18 |
| Left bundle branch block | 600 (3.5) | 574 (3.5%) | 26 (4.2%) | 0.34 |
| Typical angina pectoris | 1,015 (5.9%) | 983 (6%) | 32 (5.2%) | 0.41 |
| β blockers | 2,098 (12.3%) | 1,998 (12.1%) | 100 (16.2%) | 0.003 |
| Calcium channel blockers | 2,555 (14.9%) | 2,447 (14.8%) | 108 (17.4%) | 0.07 |
| Digoxin | 387 (2.3%) | 141 (0.9%) | 246 (39.7%) | <0.001 |
| Nitrates | 3,983 (23.3%) | 3,792 (23%) | 191 (30.9%) | <0.001 |
| Angiotensin-converting enzyme inhibitors/angiotensin receptor blockers | 4,169 (24.4%) | 3,944 (23.9%) | 225 (36.3%) | <0.001 |
| Diuretics | 1,259 (7.4%) | 1,085 (6.6%) | 174 (28.1%) | <0.001 |
| Left ventricular ejection fraction at rest ⁎ | 56.3 ± 10.0 | 56.6 ± 9.7 | 52.8 ± 12.0 | <0.001 |
⁎ Available in the subset of 5,189 patients who underwent exercise echocardiography.
| Variable | All Patients (n = 17,100) | AF | p Value | |
|---|---|---|---|---|
| No | Yes | |||
| (n = 16,481) | (n = 619) | |||
| Peak systolic blood pressure (mm Hg) | 175.9 ± 31.1 | 176.6 ± 30.9 | 156.8 ± 30.6 | <0.001 |
| Peak heart rate (beats/min) | 143.3 ± 21.1 | 142.9 ± 20.5 | 156.9 ± 28.8 | <0.001 |
| % of age-predicted maximum heart rate | 92.0 ± 12.8 | 91.5 ± 12.3 | 103.9 ± 18.7 | <0.001 |
| Peak rate–pressure product (×10 3 mm Hg · beats/min) | 25.3 ± 6.2 | 25.4 ± 6.2 | 24.7 ± 6.9 | 0.03 |
| Exercise-induced chest pain | 2,133 (12.5%) | 2,086 (12.7%) | 47 (7.6%) | <0.001 |
| METs | 8.8 ± 2.8 | 8.9 ± 2.9 | 6.7 ± 2.8 | <0.001 |
A total of 15,324 patients had interpretable rest electrocardiograms. Of them, 2,516 patients (16.4%) developed significant exercise-induced ST-segment abnormalities, and the latter occurred less frequently in patients with AF (12.0% vs 16.5% in patients without AF, p = 0.028). In contrast, among the subgroup of 5,189 patients who underwent exercise echocardiography, 1,778 patients (34.3%) developed new or worsening wall motion abnormalities during exercise, which also were less likely in patients with AF (22.2% vs 35.1% in patients without AF, p <0.001). These associations remained significant after adjustment for other potential confounders ( Table 3 ).
