Author + information
- Received August 8, 2000
- Revision received March 13, 2001
- Accepted March 28, 2001
- Published online July 1, 2001.
- Ronald J Krone, MD, FACC∗,*,
- Regina M Hardison, MS†,
- Bernard R Chaitman, MD, FACC‡,
- Raymond J Gibbons, MD, FACC§,
- George Sopko, MD, MPH, FACC∥,
- Richard Bach, MD, FACC∗ and
- Katherine M Detre, MD, DrPH†
- ↵*Reprint requests and correspondence:
Dr. Katherine Detre, c/o BARI Coordinating Center, University of Pittsburgh, 127 Parran Hall, 130 DeSoto Street, Pittsburgh, Pennsylvania 15261
The objective of this study was to evaluate the American College of Cardiology/American Heart Association (ACC/AHA) guidelines for exercise testing (EXT) after successful coronary revascularization (CR) using the Bypass Angioplasty Revascularization Investigation experience.
The ACC/AHA guidelines state that EXT within three years of successful CR is not useful.
The 1,678 patients randomized to CR by either angioplasty or bypass surgery were required to take symptom-limited treadmill tests one, three and five years after revascularization.
Patients who took the test at each specified time had a much lower subsequent two-year mortality than those who did not (1.9% vs. 9.4%, 3.5% vs. 12.6% and 3.3% vs. 11.0% at one, three and five years, respectively, after CR [p < 0.0001 for each]). Exercise parameters at the one- and three-year test did not improve a multivariable model of survival after including clinical parameters. Exercising to Bruce stage 3 or generating a Duke score >−6 were independently predictive of two-year survival after the five-year test. ST depression on the one-year test was associated with more revascularizations (relative risk = 1.6; p < 0.001).
Patients with stable multivessel coronary disease who took a protocol-mandated exercise test at one, three and five years after revascularization were at low risk for mortality in the two years subsequent to each test. Exercise parameters did not improve prediction of mortality in the two years after the one- and three-year tests. The ACC/AHA guidelines on exercise testing after CR (no value for routine testing in stable patients for three years after revascularization) are supported by these results.
Routine stress testing in asymptomatic patients for three years after successful coronary revascularization (CR) is considered a class III indication (of no value) by the current American College of Cardiology/American Heart Association (ACC/AHA) exercise guidelines unless special anatomical situations are present (1,2). Noninvasive testing has been used in an attempt to discover asymptomatic occurrence of restenosis or conduit failure (1,3–7). This strategy assumes that high-grade stenoses detected by noninvasive tests require revascularization even if asymptomatic.
The Bypass Angioplasty Revascularization Investigation (BARI) is an ongoing randomized trial comparing an initial strategy of surgical to percutaneous revascularization on mortality and recurrent infarction in patients with multivessel coronary artery disease (CAD) (8). The protocol required treadmill tests (EXTs) periodically after revascularization. The goal of this study was to assess the ability of the protocol-mandated EXTs to predict new cardiac events in this population of stable patients with revascularized multivessel CAD.
The details of the BARI protocol and seven-year survival results have been published (8). To be eligible for the BARI, multivessel CAD with clinically severe angina or ischemia requiring revascularization and suitable for both percutaneous transluminal coronary angioplasty (PTCA) and coronary bypass graft revascularization (CABG) was required. Between August 1988 and August 1991, 1,829 patients from 20 centers in North America were enrolled in the study. The average age of patients at baseline was 61.5 years; 27% were women; 64% had unstable angina; 41% had triple-vessel disease; 9% had congestive heart failure, and 19% had diabetes treated with either oral hypoglycemics or insulin.
A symptom-limited treadmill test was mandated one, three and five years after the index revascularization procedure. The Bruce protocol was utilized for most patients. Exercise electrocardiograms (ECGs) were interpreted at the Core ECG Laboratory (CEL) in addition to the local site. ST segment horizontal depression of 0.1 mV or more or upsloping with 1.5 mm depression 0.08 s after the J point was required for a positive result (1). The Duke prognostic score, which incorporates exercise time, development of angina and ST displacement was calculated using the CEL ST analysis (9).
Patients were followed for a minimum of five years after the one-year exercise test. Patients were contacted annually. Mortality from all causes was the primary end point. New cardiac events or mortality were timed from the date of the EXT.
An independent committee assessed causes of death. A Q-wave myocardial infarction (MI) was defined by new persistent Q-waves or left bundle branch block with abnormal enzymes.
Data from patients treated with either PTCA or CABG were combined. However, treatment allocation was included in all multivariate analyses (8). Kaplan-Meier survival curves were constructed and compared using the log-rank test. Multivariable Cox models adding exercise variables to other clinical variables previously determined to predict events were constructed for the end points of survival, survival free of Q-wave MI and subsequent revascularization.
Of the 1,698 patients eligible to take the EXT one-year after CR, 310 (18%) patients did not take it. At three years, 1,632 patients were eligible, and 424 (26%) did not take the exercise test; at five years, 1,521 were eligible, and 474 (31%) did not take the exercise test. Demographics of the patients at the time of the one-year test are seen in Table 1. As expected, there was more comorbidity in the nonexercisers. Treatment allocations were similar in the two groups. Only 9% developed angina during the test, consistent with the “protocol-not symptom driven” indication for the test. No relationship was seen between reason for not exercising (patient refusal, orthopedic limitation, peripheral vascular disease and other) and survival, but there was a suggestion that orthopedic limitation decreased five-year survival (66.7% vs. 74.1%).
Nonexercisers had a 3.9% mortality rate in the year after the scheduled test and 25.9% mortality after five years. This compared with the 0.4% one-year mortality (relative risk [RR] = 9.2, p < 0.0001) and the 8.1% five-year mortality in the group who did take the test. Within all subgroups considered, the exercising patients had a better five-year survival than nonexercising patients (Fig. 1). Nonexercisers also had a higher combined rate of death or Q-wave infarction than did the exercising patients (2.9% vs. 1.0% at one year and 18.0% vs. 6.8% at five years). The nonexercisers underwent more revascularizations in the first year after the test (8% vs. 4%, p < 0.001). However, by the fifth year, the cumulative rate was similar (18.6% vs. 16.5%, respectively).
We evaluated survival for the two years after each of the scheduled tests. The risk of two-year mortality in those who did not take the EXT ranged from 9.4% after the one-year test to 12.6% and 11.0% after the three and five year tests, respectively. This compares with a two-year mortality of 1.9%, 3.5% and 3.3%, respectively, for the EXT takers (Tables 2 and 3). ⇓⇓The results for combined mortality or MI show a similar difference (11.9% vs. 3.6%, multivariate RR = 2.86, p < 0.0001) for the two years after the one-year test, for example.
The BARI trial has shown a difference in seven-year mortality between treated diabetics who received PTCA as compared with those receiving CABG (8). It was found that 73% of the CABG patients with diabetes and 71% of the PTCA patients with diabetes performed the one-year EXT. The five-year Kaplan Meier survival estimates for patients with diabetes were 70.0% for the nonexercising CABG patients compared with 47.6% in the nonexercising PTCA patients with diabetes (p = 0.0098) and 88.9% vs. 79.2% for the patients who could exercise (p = 0.0273), respectively (Fig. 2). The corresponding estimates for patients without treated diabetes were 80.2% vs. 80.3% for the nonexercising patients and 93.1% compared with 93.6% for the patients who exercised. All multivariate analyses were corrected for the diabetes.
Relationship between EXT parameters and mortality after each exercise test
The independent contribution of exercise parameters to multivariable models for survival and survival free of MI was evaluated for the two years after each test (Tables 4 and 5). ⇓⇓The addition of exercise parameters did not improve the models for survival or survival free of infarction after the one and three year exercise tests. Exercise duration was independently associated with decreased survival for two years after the five-year test (RR = 0.42, p = 0.0069) and was independently associated with decreased survival free of MI for the two years after the three-year test (RR = 0.51, p = 0.0372).
The CEL was able to evaluate the ST segment in 1,322 of the 1,388 exercising patients. ST depression, interpreted at either site, was not related to subsequent infarction-free survival. However, ST segment depression diagnosed at the clinical site was strongly associated with revascularization procedures five years after the one-year test (RR = 1.6, p < 0.001 compared with RR = 1.22, p = 0.16, for the CEL measurements).
Twenty-eight patients (2%) underwent revascularization within three months of the EXT. Of the 255 patients with symptoms of angina before the test, 17 (6.7%) were revascularized within three months compared with 11 of the 1,132 patients (1%) who did not report symptoms of angina. Of the 11, seven had a negative EXT (no ST depression and no angina during the test). Thus, of the 28 patients who were revascularized within three months of the EXT, only four (0.3% of the total patients) might be construed to be revascularized only as result of ischemia documented on an EXT (Table 6).
Long-term (five-year) mortality was also evaluated after the one-year test. Exercising to Bruce stage 3, heart rate >85% and Duke Score >−6 were associated with an improved five-years survival, but only Bruce Stage 3 was an independent predictor of survival on the multivariate analysis (RR = 0.69, p = 0.04) (Fig. 3).
Coronary atherosclerosis is a progressive disease, and, therefore, even after a successful revascularization, recurrent cardiac events are to be expected. In the first 12 to 18 months after revascularization, these events are more common in patients after PTCA than after CABG (8,10,11). After 18 months, the number of events (primarily recurrence of angina) stabilizes and is similar for patients undergoing either procedure (8).
Additional prognostic information from the EXT parameters
The prognostic value of the results of the EXT for survival or survival free of new Q-wave infarction in this largely stabilized and asymptomatic population of patients after revascularization was limited. Specific exercise parameters were disappointingly not predictive of major cardiac events, with the exception of the Duke score and the exercise capacity, which only predicted events five years after the test. The results of ST segment analysis were not correlated with infarction-free survival. ST depression in patients not reporting angina before the EXT was not predictive of early revascularization. The predictive value of the EXT is limited by the sensitivity and specificity of the individual parameters and especially by the low incidence of events (Table 7). Approximately nine in 10 patients with low exercise tolerance and five in six patients with a high-risk Duke score at one year will survive five years (6,12,13). In this patient population, a more specific test, such as radionuclide imaging or exercise echocardiography, would still be expected to have a high incidence of false positive results; only a fraction of those with a significant perfusion defect will have a subsequent event.
These results in the context of plaque biology
Cardiac events are caused by the transformation of stable plaques to unstable plaques (14–16). White et al. (17)showed that, in over 50% of patients with an angiogram two years before an MI, the ultimate culprit lesion was of only borderline significance on the first study. An EXT will only screen for coronary stenosis and cannot be expected to predict which or whether plaques will become unstable. The most common symptom of progressive disease is angina (18,19). Multiple studies have shown a low incidence of cardiac events in stable patients after CR without angina (18).
Routine ET one year after successful coronary revascularization to identify patients at risk for subsequent cardiac events presumes that the identification of stenosis in patients with stable revascularized coronary disease will be able to isolate patients at risk for events better than a strategy of symptom-driven evaluation. Since the important events, sudden death or infarctions, are infrequent in asymptomatic individuals, the best early detection strategy is not defined. The paradigm of aggressively seeking out potential restenosis to revascularize the stenotic vessel must be proven before being accepted as a routine strategy (6,20).
Exercise testing in patients with diabetes
We did not specifically evaluate individual exercise parameters in patients with diabetes since the number of events and the number of patients with diabetes who exercised was quite small. However, the five-year mortality in treated patients with diabetes who did not exercise was quite striking: 30% for patients undergoing CABG and 52.4% for those undergoing PTCA. An examination of Figure 2shows a very limited mortality in the patients with diabetes who exercised for the first year and a half after the exercise test, so it seems also likely that exercise parameters would be of limited value in these patients.
No additional imaging, echocardiography or myocardial scintigraphy was used with the EXT (21,22). Secondly, 7.8% of patients who experienced angina on the EXT underwent revascularization within three months of the study compared with 1.4% of those who did not demonstrate angina. It is possible that the subsequent procedures selected for patients with EXT abnormalities attenuated a difference in cardiac events between the two EXT groups. However, most of the patients undergoing revascularization already reported angina at the time of the test and were, therefore, not truly asymptomatic. Lastly, these revascularizations were performed in 1988 through 1991, and considerable progress has occurred both in revascularization techniques and management of coronary disease itself, further lowering recurrence rates and making prediction of recurrent events even more difficult and less precise.
How does this relate to the ACC/AHA guidelines for EXT in asymptomatic patients after coronary revascularization?
The ACC/AHA guidelines for EXT consider EXT for prognosis in asymptomatic individuals for at least three years after revascularization to be a class III indication, i.e., the procedure is not useful and, in some cases, may be harmful (1). Our data clearly support that position. The ACC/AHA task force suggested that stress testing could have a role for patients with particularly compelling anatomy, where the angioplastied lesion controls a large amount of myocardium or where the patient clearly has a defective warning system, i.e., when the patient suffers from silent ischemia (1,4). We did not specifically evaluate these situations, and it is not clear whether the event rate is higher in these patients.
The ability or willingness to take an EXT is a simple, but powerful, discriminator of risk. In asymptomatic patients, the results of a routine (protocol driven) EXT one year after revascularization does not reliably identify those patients who will go on to have cardiac events in the next year or two. However, it does delineate a small group at high risk within five years after the test. We found that the best stratification occurred with recurrent evaluation for the ability or willingness to exercise one, three and five years after revascularization. Our findings are consistent with the ACC/AHA guidelines on exercise testing suggesting no role for routine exercise testing for at least three years after successful coronary revascularization (1). These data suggest that efforts to improve survival should be concentrated in the patients who are not able to take the EXT.
☆ Supported by grants from the National Heart, Lung and Blood Institute, the National Institutes of Health, Bethesda, Maryland.
- American College of Cardiology/American Heart Association
- Bypass Angioplasty Revascularization Investigation
- coronary artery bypass grafting
- coronary artery disease
- Core Electrocardiographic Laboratory
- coronary revascularization
- exercise test
- myocardial infarction
- percutaneous transluminal coronary angioplasty
- relative risk
- Received August 8, 2000.
- Revision received March 13, 2001.
- Accepted March 28, 2001.
- American College of Cardiology
- Gibbons R.J,
- Balady G.J,
- Beasley J.W,
- et al.
- Miller D.D,
- Verani M.S
- Multicenter Myocardial Ischemia Research Group,
- Krone R.J,
- Gregory J.J,
- Freedland K.E,
- et al.
- Miller T.D,
- Christian T.F,
- Hodge D.O,
- Mullan B.P,
- Gibbons R.J
- BARI Investigators
- Hernández R.A,
- Macaya C,
- Iñiguez A,
- et al.
- Weintraub W.S,
- Ghazzai Z.M.B,
- Douglas J.S Jr.,
- et al.
- Farb A,
- Burke A.P,
- Tang A.L,
- et al.
- Little W.C,
- Constantinescu M,
- Applegate R.J,
- et al.
- Pfisterer M,
- Rickenbacher P,
- Kiowski W,
- Muller-Brand J,
- Burkart F
- Wijns W,
- Serruys P.W,
- Simoons M.L,
- et al.
- Multicenter Postinfarction Research Group,
- Krone R.J,
- Gillespie J.A,
- Weld F.M,
- Miller J.P,
- Moss A.J
- Iskandrian A.S,
- Chae S.C,
- Heo J,
- Stanberry C.D,
- Wasserleben V,
- Cave V
- Olmos L.I,
- Dakik H,
- Gordon R,
- et al.