Effective Exercise Stress Testing for CAD

from Emergency Medicine Magazine  May 2001  http://www.emedmag.com/stories/storyReader$193

Effective Use of the Exercise Stress Test to Detect and Predict Coronary Artery Disease

When interpreted in conjunction with the findings of the clinical examination and medical history of properly selected patients, the results of exercise stress testing can be an invaluable tool for determining the diagnosis or prognosis of CAD.

By Jennifer M.F. Kwok, MB, ChB, and Timothy F. Christian, MD

Ms. Kwok is a medical officer in the medical department at Princess Margaret Hospital in Lai Chi Kok, Kowloon, Hong Kong, and Dr. Christian is associate professor of medicine in the division of cardiovascular diseases and department of internal medicine at the Mayo Clinic and Mayo Foundation in Rochester, Minnesota.

Coronary artery disease (CAD) has become a global epidemic at this turn of the millennium. The disorder was implicated in 7.2 million deaths worldwide in 1996 and ranked first among leading causes of death.

In the acute care setting, the history and physical examination alone are often insufficient to detect severe CAD or determine the prognosis of known disease. Among the techniques that are used to make those determinations, treadmill exercise testing has become a popular and proven method supported by a number of advantages. It can be performed as an office procedure, it is convenient for patients, and it is relatively inexpensive in comparison with other stress-based methods, such as stress echocardiography or stress myocardial perfusion imaging.

The ultimate aim of subjecting patients to such testing is to determine the most effective therapy. The effectiveness of exercise testing can be improved through appropriate patient selection, proper execution of the test, and critical appraisal of the results to aid the therapeutic strategy.

Although it is useful in the evaluation of valvular heart disease, congenital heart disease, and arrhythmia, exercise testing is usually performed for the evaluation of CAD. The focus of this review, therefore, will be that purpose.

INITIAL EVALUATION OF PATIENTS WITH SUSPECTED CAD

Whether a patient with chest pain requires extensive examination for the detection of CAD depends on the likelihood suggested by the clinical information and baseline data gathered from the history, physical examination, resting electrocardiogram (ECG), chest radiograph, and baseline blood test results.

A significant body of evidence suggests that clinical parameters alone can predict the likelihood of CAD, and during the past 20 years, various methods have been devised to determine pretest risk of significant CAD among men and women. In 1997, the American College of Cardiology/American Heart Association (ACC/AHA) Task Force on Practice Guidelines included in its recommendations a table, derived from the analysis of angiographic and autopsy findings of 28,000 patients, for estimating the likelihood of CAD on the basis of age, sex, and chest pain character (Journal of the American College of Cardiology, vol. 30, p. 260) (see table below). Pryor and colleagues have also devised nomograms that base likelihood on sex, age, smoking history, lipid level, history of diabetes mellitus, history of myocardial infarction (MI), character of chest pain, and baseline ECG changes (American Journal of Medicine, vol. 75, p. 771, 1983).

The application of Bayes' theorem to the interpretation of noninvasive testing cannot be overemphasized. Bayes' theorem incorporates the pretest probability of a disease, as well as the sensitivity and specificity of a test, to calculate the posttest probability of the disease.

The baseline characteristics of patients who have chest pain determine the pretest probability of CAD. Patients with very low pretest probability of CAD will have a low posttest probability of CAD even if test results are positive. Consequently, a positive test cannot confirm the diagnosis of CAD in this particular population. On the other hand, in patients with high pretest probability, CAD cannot be confidently ruled out even if test results are negative. Those patients may benefit from a noninvasive test that determines cardiovascular risk relating to the prognosis but not the diagnosis of CAD.

For patients with intermediate pretest probability, a noninvasive test will provide the most information that is used to determine their risk and appropriate therapy. A test is useful only when it contributes incremental value to the baseline information that guides the plan of management. Thus, for these patients, exercise testing is the more cost-effective method for diagnosing CAD.

The list of absolute and relative contraindications to exercise testing is extensive and comprises mostly cardiovascular disorders (see table below). Patients who have any of these conditions should undergo pharmacologic perfusion imaging or cardiac catheterization as indicated by the specific condition.

In addition, since ST-segment changes during exercise testing are among the most important parameters to monitor, interpretation of the exercise ECG will not be possible if the baseline ECG reveals abnormalities produced by preexcitation syndrome, electronically paced ventricular rhythm, resting ST-depression greater than 1 mm, or complete left bundle branch block. Consequently, for patients with these findings exercise testing will provide no benefit in the detection of CAD or in determining its prognosis.

DETERMINING CARDIAC RISK IN PATIENTS WITH DIAGNOSED CAD

Treadmill exercise testing is an appropriate initial evaluation method for determining cardiovascular risk in patients with stable CAD. It is also warranted for those with known CAD whose clinical status changes significantly. The choice of therapeutic strategies, which include further noninvasive evaluations, cardiac catheterization, and medical therapy, is guided by the estimated prognosis. For these patients, exercise testing has been documented to be effective in predicting cardiac death, MI, and the necessity for revascularization. In addition, the procedure is useful for detecting severe CAD and estimating ventricular function and propensity of arrhythmias.

Treadmill exercise testing is also useful for determining cardiovascular risk after patients have had an acute MI. Prognostic assessment is usually based on a submaximal test conducted at about 4 to 7 days after MI or on a symptom-limited test at about 14 to 21 days.

EXERCISE TESTING AND PROTOCOLS

The treadmill and bicycle ergometer are the commonly used dynamic exercise devices. For untrained cyclists, the quadriceps muscles are less well developed and easily fatigued. As a result, exercise testing with bicycle ergometers may terminate prematurely before maximum oxygen uptake is reached. The treadmill is the more popular choice, since people are more used to walking than cycling and generally achieve a higher workload with that device.

A number of treadmill protocols have been devised, and the selection can be tailored to the patient's physical capacity and to the purpose of the test. Continuous progressive exercise lasting 6 to 12 minutes is optimal to achieve maximal myocardial oxygen demand for diagnostic and prognostic proposes. Exercise testing may be terminated prematurely if the protocol is too strenuous to the patient, or it may merely test his or her endurance if the protocol is too easy.

With extensive data supporting its value in the diagnosis and prognosis of CAD, the Bruce protocol is the most popular choice. A patient begins the test by walking at 1.7 mph on a 10% grade, a setting that corresponds to an oxygen consumption of about 5 metabolic equivalents (METs). In each three-minute stage of the protocol, the speed and grade are increased to generate a 3-MET increase in workload. Young, healthy patients are usually able to tolerate the Bruce protocol, but old or debilitated subjects may find it too strenuous.

Other protocols such as the modified Bruce, the Cornell, the Naughton, or the Ramp protocol are designed to produce a more gradual increment of METs between stages; they are suitable to patients who have limited exercise tolerance. Patients who have suffered an MI should undergo submaximal exercise testing or a symptom-limited exercise testing protocol soon after the event.

The exercise testing procedure should be adequately explained to patients before they start the test. Before the electrodes are applied, it is essential that the skin be properly prepared to lower the resistance at the skin-electrode interface and to improve the signal-to-noise ratio. Baseline blood pressure and ECG should be recorded while a patient is standing, so as not to miss orthostatic hypotension or ST-segment depression. In addition, blood pressure should be measured manually, as automated devices have not been proved to be accurate.

During the test, patients should walk upright and with a stable gait, without leaning on the support rails. To assess their fatigue, the Borg scale should be used. This linear scale grades a patient's perceived exertion in levels ranging from 7 (very, very light) to 19 (very, very hard).

Continuous ECG monitoring of heart rate, ST-segment changes, and arrhythmia is crucial. Blood pressure, heart rate, and 12-lead ECG should be obtained at the end of each stage, immediately after termination of exercise, and at every minute during recovery. Any symptoms should be observed closely during the test, and the origin of chest pain, whether anginal or noncardiac, must be identified.

Patients should reach maximum aerobic capacity or cardiac output by reaching their maximum exercise heart rate. Although the predicted maximum heart rate can be estimated [220 - the patient's age or 200 - (0.6 3 patient's age)], a wide spectrum of values straddle the regression line for that rate.

The Borg scale is a more practical tool for assessing a patient's maximum exercise capacity. A reading of 14 to 16 approximates a patient's anaerobic threshold, and 18 or higher approximates his or her maximum exercise capacity.

Signs that indicate immediate termination of an exercise test include severe or increasing chest pain or fatigue, neurologic symptoms, signs of poor perfusion, a hypertensive response, a drop in systolic blood pressure in the presence of other evidence of ischemia, ST-segment elevation abnormalities, and arrhythmia (see table below).

Adequate stress is achieved when patients reach 85% of their target heart rate or have a heart rate-systolic blood pressure product greater than 25,000.

INTERPRETING TEST RESULTS IN THE DIAGNOSIS OF CAD

The results of the exercise test should be interpreted on the basis of a patient's clinical information, risk profile, angina symptoms, baseline ECG, hemodynamic response, exercise capacity, and ST-segment response during exercise. The pretest probability of CAD should be estimated from the baseline clinical information using Bayes' theorem. Typical angina is the most important clinical predictor of CAD. The prevalence of disease in the population tested should be considered in the probability assessment of each patient who undergoes exercise testing.

Chest discomfort. Exercise-induced chest discomfort can be the only marker of CAD, although it usually occurs after the onset of ischemic ST-segment changes. The finding is particularly useful if it suggests classic angina.

Ischemic ST-segment changes. The ST-segment displacement is critical to the diagnostic interpretation of exercise testing. The PQ junction is chosen as the isoelectric point and the J point refers to the junction between the QRS complex and ST segment. The level of ST-segment displacement is measured from the isoelectric point 60 to 80 ms after the J point. An abnormal response is defined as a horizontal or down-sloping ST-segment depression of 1 mm or greater or as ST-segment elevation lasting at least 60 to 80 ms after the J point in three consecutive beats. If ST-segment depression is noted at baseline, an additional depression of 1 mm or greater is considered abnormal. For patients who have early repolarization and baseline ST-segment elevation, ST-segment depression should be measured from the PQ junction.

In the ACC/AHA Task Force on Practice Guidelines meta-analysis of exercise testing mentioned earlier, Gibbons and colleagues reported that as a diagnostic criterion, horizontal or down-sloping ST-segment depression of 1 mm was 67% sensitive and 72% specific in detecting CAD. Exercise-induced ST-segment depression is not specific in identifying diseased vessels, however.

It is important to note that among patients who have a normal ECG at rest, the ST-segment depression noted in the inferior leads is more likely to be false-positive than that detected by the precordial leads. Miranda and colleagues showed that the combined sensitivity and specificity of exercise-induced ST-segment depression in lead V5 (65% and 84%, respectively) was higher than in lead II (71% and 44%, respectively) (American Journal of Cardiology, vol. 69, p. 303, 1992). Although monitoring the inferior leads adds little to the diagnosis, marked, multiple-lead ST-segment depression that starts in the first stage and lasts at least six minutes into recovery is suggestive of left main CAD or three-vessel disease.

ST-segment elevation occurs in 0.1% of patients who undergo exercise treadmill testing. In patients with baseline ST-segment depression or elevation, exercise-induced ST-segment elevation is defined as additional ST-elevation from the baseline. ST-segment elevation noted in any lead except aVR or V1 suggests transmural ischemia, which can be caused by coronary spasm or a critical lesion. Unlike the leads that indicate ST-segment depression, those that reveal ST-segment elevation will identify the involved coronary arteries.

Patients with ST-T changes at rest. Among both the general population and those who present with chest pain, repolarization abnormalities accompanied by ST-segment depression or T-wave inversion or flattening are the most common anomalies recorded on the resting ECG. The prevalence of CAD, severe CAD, and left ventricular dysfunction is higher among patients with ST-T abnormalities, as are the rates of cardiac-related morbidity and death, than among patients whose resting ECG is normal.

Results obtained from the Framingham Study in the mid 1970s have shown that the overall prevalence of nonspecific electrocardiographic ST-segment or T-wave abnormalities was 8.5% among men and 7.7% among women in the study population. The age-adjusted CAD-related mortality and morbidity rates were about two times higher among men and women with ST-T abnormalities than among those without.

From a theoretical standpoint, ST-T abnormalities can cancel or obscure ST-T changes induced by ischemia and thereby produce false-negative test results. When an additional horizontal or down-sloping exercise-induced ST-segment depression of 1 mm was used as the criterion for a positive test, the sensitivity of exercise testing in patients with resting ST-T abnormalities ranged from 75% to 92% and the specificity ranged from 53% to 79%.

On the other hand, Miranda and colleagues have shown that among patients with nonspecific ST-T abnormalities on their resting ECGs but who have not had a previous MI, an additional ST-segment depression of 2 mm yielded a sensitivity of 67% and a specificity of 80% in the diagnosis of CAD; however, when an additional 1 mm of ST-segment depression was used as the criterion, the sensitivity increased to 83% but the specificity decreased to 20% (American Heart Journal, vol. 122, p. 1617, 1991).

Patients with right bundle branch block. Data are limited regarding the utility of exercise testing among patients with right bundle branch block. In a study from 23 years ago, Tanaka found that among 30 symptomatic patients with the disorder who had undergone treadmill exercise testing and coronary angiography, horizontal or down-sloping ST-segment depression of 1 mm from the baseline was 69% sensitive and 45% specific in the detection of CAD. The specificity appeared to be greater when ST-segment depression was recorded in leads V4 to V6. ST-segment depression limited to leads V1 to V3 was more likely to be false-positive. Although no large-scale studies of this type have been performed, the ACC/AHA guidelines recommend standard exercise treadmill testing as the initial tool for determining cardiac risk.

Patients with left ventricular hypertrophy secondary to hypertension. Exercise testing in patients with left ventricular hypertrophy is less specific, but its sensitivity is unaffected. A standard exercise test is recommended by the ACC/AHA guidelines as the initial step. Further testing is indicated for patients with abnormal test results.

Patients taking cardiac medications. Cardiac drugs can reduce the diagnostic accuracy of exercise testing. Beta-blockers can lower myocardial oxygen requirements by reducing blood pressure, heart rate, and ventricular contractility. In patients with CAD, the drugs attenuate or delay the onset of ST-segment depression that occurs in response to exercise. Beta-blockers can prevent patients from reaching their target heart rate and thereby render test results inconclusive.

To avoid obtaining an inadequate stress response, clinicians can withhold beta-blocker therapy 48 hours before diagnostic exercise testing is begun. However, in patients with angina, acute ischemia may develop in response to sudden withdrawal of beta-blocker therapy. The ACC/AHA guidelines for exercise testing do not recommend routine discontinuation of beta-blocker therapy before testing. The decision to withhold the medication should depend on each patient's circumstances.

Long-acting nitrate therapy can prevent, delay, or reduce the severity of ST-segment abnormalities that occur in response to exercise. For patients with CAD, calcium channel blockers have been known to increase exercise tolerance and reduce ECG ischemic response. When these drugs are not absolutely indicated, however, withholding such therapy is a reasonable method for improving the sensitivity of exercise testing.

When normal patients are given maintenance doses of digoxin, 25% to 50% of them have horizontal or down-sloping ST-segment depression greater than 1 mm in response to exercise. The depth of ST-segment depression is roughly correlated with the serum digoxin level. On the other hand, a normal ST-segment response to exercise in patients receiving digoxin suggests ischemia is unlikely. Ultimately, exercise testing remains sensitive but not specific for the detection of CAD in patients taking digoxin. Although two weeks are necessary for the effects of digoxin to wear off, discontinuing the regimen before diagnostic testing is begun is not practical.

Elderly patients. The data regarding exercise testing in elderly populations are limited. In one study of patients aged 65 years or older who underwent an ergometer stress test, the diagnosis of CAD was aided by exercise-induced ST depression, exercise heart rate, blood pressure response, and functional capacity. In another study, when ST-segment depression of 2 mm or greater was used as the diagnostic criterion, ergometer exercise testing was equally effective in detecting CAD in older (more than 65 years) and younger patients alike.

USE OF EXERCISE TESTING IN DETERMINING PROGNOSIS

The cardiac prognosis of patients with CAD depends on left ventricular function, the extent and severity of coronary stenosis, the presence of vulnerable plaque, and the electrical stability of the myocardium. Exercise testing provides prognostic information on left ventricular function and severity of CAD. Acute coronary syndromes are usually caused by rupture of vulnerable plaque; however, an effective means of identifying such plaque does not yet exist. Vulnerable plaque may not be flow-limiting or induce ischemia during noninvasive testing, and it may not be significant angiographically. Exercise testing can sometimes indicate a patient's risk for malignant ventricular arrhythmias, although other tests, including an electrophysiology study, may be necessary.

Maximum exercise capacity. Maximum exercise capacity depends on cardiac function, but it is influenced by environmental conditions and a patient's familiarity with the exercise test, activity status, age, and gender. Exercise capacity can be expressed in terms of duration under a particular protocol or of maximum heart rate or double product, but it is preferably expressed in multiples of basal resting requirements, or METs.

A MET is a unit of basal oxygen consumption equal to 3.5 mL/kg/min. Expressing exercise capacity in METs allows workloads obtained with different protocols to be compared in equal terms. For patients younger than 65 years, 5 METs or fewer is associated with a poor prognosis. Patients who have abnormal test results who can achieve 13 METS may have a good prognosis.

Maximum exercise capacity is inversely related to the severity and extent of CAD, and it predicts cardiac events. Multivariate models have shown that maximum exercise capacity is one of the independent prognostic indicators. Several treadmill scores incorporate maximum exercise capacity in determining a patient's prognosis.

The prognostic value of exercise capacity has been substantiated by Roger and colleagues in a population-based cohort study in which exercise capacity, exercise-induced angina, and ECG changes were associated with outcome in both sexes. However, after adjustment, exercise capacity was the only treadmill variable predictive of all-cause mortality and cardiac events. In addition, a strong linear relationship existed between exercise capacity and outcome. An increment of 1 MET in the workload was associated with a 20% to 25% reduction in all-cause mortality and cardiac events among patients of both sexes (Circulation, vol. 98, p. 2836, 1998).

Blood pressure. Systolic blood pressure increases progressively during exercise to a peak response ranging from 160 to 200 mm Hg. Abnormal responses that suggest left ventricular systolic dysfunction caused by ischemia point to inadequate cardiac output. Abnormal hemodynamic responses include a sustained decrease in systolic blood pressure of 10 mm Hg or more despite an increase in work load, a failure to increase the systolic blood pressure greater than 120 mm Hg, and a drop of systolic blood pressure below standing rest values.

Exercise-induced hypotension is associated with a high probability of severe CAD and with a poor prognosis. Less frequently, exercise-induced hypotension can occur in patients with chronic ventricular dysfunction, cardiomyopathy, valvular heart disease, papillary muscle dysfunction with mitral regurgitation, arrhythmias, or vasovagal response. It also has been known to occur after prolonged strenuous exercise or antihypertensive therapy.

Chest discomfort. Exercise-induced angina with or without ST-segment changes predicts a subsequent coronary event. Patients with exercise-induced angina associated with ischemic ST-segment changes are twice as likely as those who have only ST-segment changes to suffer cardiac events, such as cardiac death, MI, or progression of angina. In addition, chest pain that occurs early in the test during low workload is associated with an adverse prognosis.

ST-segment depression. Although the magnitude of exercise-induced ST-segment depression predicts the severity of CAD, it must be interpreted in conjunction with the time of onset and duration of ST-segment depression. In their study from 1978, McNeer and colleagues showed that over 97% of patients with positive test results at stage I or stage II of the Bruce protocol had significant CAD, and more than 60% of them had three-vessel disease. Multivessel or left main CAD is suggested by ST-segment depression that is down-sloping, occurring within the first stage of the Bruce protocol, and persisting more than eight minutes during recovery (Circulation, vol. 57, p. 64, 1978).

When both clinical and exercise variables are considered, the magnitude of ST-segment depression is an important component of the Long Beach-Palo Alto, Morise, and Detrano equations that are used to predict severe CAD. In addition, the magnitude of ST-segment depression is a powerful prognostic sign. Mark and colleagues have demonstrated that exercise-induced ST-segment deviation is the single most important variable in predicting cardiac outcomes from exercise testing (Annals of Internal Medicine, vol. 106, p. 793, 1983).

The time of onset of ST-segment depression is also prognostically significant. In McNeer's study, the survival rate among patients with positive test results at stage I or stage II of the Bruce protocol was 63% after 48 months of follow-up. The incidence of all coronary events is inversely proportional to the time of onset of ischemia. In 1975, Ellestad showed that the yearly incidence of coronary events was 15% among patients with ST-segment depression of 2 mm at a workload of 4 METs, as opposed to 4% among similar patients whose workload was 8 METs (Circulation, vol. 51, p. 363, 1975).

Treadmill score and prognostic assessment. Several treadmill scores are obtained to determine a patient's prognosis, and they rely to some degree upon the ST-segment response and duration of exercise. Such scores have had much stronger prognostic power than the ST-segment response alone and have been advocated as a replacement for the traditional method of reporting just a positive or negative test result. This approach makes sense, as a patient with 1 mm of ST-segment depression after 2 minutes of exercise is likely to have more severe disease than someone with the same response after 10 minutes of exercise. Ideally, a patient's risk should be classified as high, moderate, or low, according to one of these scores, but clinicians should understand that the score represents a spectrum of risk (see table below).

Patients with resting ST-T abnormalities. Coronary event rates are consistently higher among patients with resting ST-T abnormalities than among those with a normal ECG obtained at rest. In a recent study, we and our colleagues have found that the Duke treadmill score can be effective in determining the risk of coronary events in patients with resting ST-T changes (JAMA, vol. 282, p. 1047, 1999). With cardiac death as the primary outcome, the seven-year survival rates as determined by the Duke treadmill scores in the low-, intermediate-, and high-risk groups in our study population were 97%, 92%, and 76%, respectively. In addition, the score is capable of predicting not only cardiac death but also total cardiac events, including nonfatal MI and late revascularization.

The ACC/AHA guidelines recommend standard treadmill testing as the initial noninvasive method of evaluating patients who have either a normal ECG obtained at rest or ST-segment depression of less than 1 mm on a resting ECG. Data used to formulate those guidelines have shown that the standard exercise treadmill test combined with the Duke treadmill score is an effective initial tool for determining the prognosis of patients who have resting ST-T changes.

Elderly patients. In a study of 104 patients older than 65 years (average age, 68) who underwent treadmill testing, bicycle ergometer testing, or supervised walking, Glover and colleagues demonstrated that ST-segment depression of 1 mm or greater was associated with an increased risk of cardiac death. Other research had shown that the magnitude of exercise-induced ST depression is an independent predictor of three-vessel CAD.

Because the exercise capacity of elderly patients can be limited more by their general physical condition than by their cardiac function, the exercise variables that are prognostically useful for younger patients may not predict cardiac outcomes in this population.

Assessing prognosis after MI. According to the ACC/ AHA exercise testing guidelines, exercise testing is a class I method for assessing a patient's prognosis, devising an exercise regimen, and evaluating medical therapy and cardiac rehabilitation. Patients who have had an MI can undergo submaximal exercise testing without complications at about four to seven days after the event. Such testing should be stopped when a patient reaches a peak heart rate of 120 to 130 bpm, 70% of predicted maximum heart rate appropriate for his or her age, or a peak workload of 5 METs or when any of the following occurs: angina, dyspnea, ST-segment depression greater than 2 mm, exertional hypotension, or malignant ventricular arrhythmia.

At 14 to 21 days after an MI, patients can undergo symptom-limited exercise testing, without stopping for target heart rates or MET level. The popular treadmill protocols are the standard and modified Bruce and modified Naughton protocols. Predischarge exercise testing can detect ischemia that may be associated with coronary events early after discharge, but a more optimal functional assessment can be obtained with symptom-limited exercise testing performed three to six weeks after an MI.

The contraindications to exercise testing are the same as previously mentioned. The risk of death from performing exercise testing after MI is low; cardiac mortality is 0.03%.

As the treatment of MI improves through the use of thrombolytic therapy, revascularization, and beta-blocker and angiotensin-converting enzyme inhibitor therapy, the clinical presentation and prognosis change. We have known since the time before thrombolytic agents were available that mortality is highest among patients excluded from exercise testing. Poor exercise capacity and abnormal systolic blood pressure response are associated with poor prognosis. Exercise-induced ST-segment depression is an adverse prognosticator only for patients with inferior-posterior MI.

From the results of the GISSI-2 study, which included patients who underwent thrombolytic therapy, Villella and colleagues demonstrated that patients unable to perform exercise testing were at the highest predicted risk for death. The predictors of mortality included ST-segment depression of 1 mm or greater in the presence of exercise-induced angina, exercise tolerance of less than six minutes, and an increase of systolic blood pressure less than 28 mm Hg from rest. After six months of follow-up, the mortality among patients unable to exercise was 7.1%; among those with positive test results (typical angina or ST-segment depression 1 mm or greater) it was 1.7%; and among those with negative results, it was 0.9% (Lancet, vol. 346, p. 523, 1995).

The results of other studies have also shown a higher mortality rate among patients who are unable to exercise, achieve 5 METs during the treadmill exercise, or increase systolic blood pressure by 10 to 30 mm Hg. Notably, other studies have not consistently demonstrated that exercise-induced ST-segment depression increases a patient's risk for death.

Exercise testing after revascularization. For symptomatic patients who have undergone a coronary artery bypass procedure, exercise testing may help identify ischemia caused by incomplete revascularization or graft occlusion. However, the treatment plan depends on the extent and severity of ischemia, which exercise testing cannot define. Consequently, when compared with imaging studies, exercise testing has limited utility. Moreover, because resting ECG abnormalities are frequent among patients who have undergone coronary bypass surgery, the exercise ECG becomes less specific.

For patients who may have noncardiac chest pain after undergoing percutaneous transluminal coronary angioplasty, exercise testing may find objective evidence of ischemia that is linked to restenosis. However, in predicting restenosis, the exercise ECG is only 40% to 55% sensitive; imaging studies are much more sensitive. Ultimately, exercise testing has a limited role in the management of patients who have undergone revascularization.

Asymptomatic patients with undiagnosed CAD. The utility of exercise testing for detecting CAD in asymptomatic patients is controversial. The prevalence of CAD in this population is low when risk factors are absent. According to Bayes' theorem, for patients with a low pretest probability of CAD, a positive exercise test result is usually false-positive, and a negative result only confirms the low probability of the disease.

In 2000, Livschitz and colleagues demonstrated that the screening of young asymptomatic men for CAD by exercise testing alone is ineffective and does not have significant clinical impact (American Journal of Cardiology, vol. 86, p. 462, 2000). However, exercise test screening of such patients can be enhanced when their risk profile is assessed along with the test results. Bruce and colleagues from the Seattle Heart Watch study have shown that abnormal exercise test results do not predict cardiac events for normal asymptomatic patients, but such results do predict cardiac events for asymptomatic patients whose risk profile includes one or two CAD risk factors (Journal of the American College of Cardiology, vol. 2, p. 565, 1983).

Gibbons and colleagues have shown in a recent study that the sensitivity of abnormal exercise test results in predicting coronary death among asymptomatic patients was 61% (American Journal of Cardiology, vol. 86, p. 53, 2000). The age-adjusted relative risk for coronary death linked to an abnormal exercise test result was 21 among patients with no risk factors in their history, 27 among those with one risk factor, 54 among those with two risk factors, and 80 in those with three or more risk factors in their history.

The results of that study suggest that abnormal exercise test results can predict coronary death among asymptomatic patients, especially among those whose history includes conventional risk factors for CAD, such as cigarette smoking, diabetes mellitus, hypertension, hypercholesterolemia, and family history of premature cardiovascular disease.

SUGGESTED READING

Diamond GA and Forrester JS: Analysis of probability as an aid in the clinical diagnosis of coronary-artery disease. N Engl J Med 300:1350, 1979.

Gibbons LW, et al.: Maximal exercise test as a predictor of risk for mortality from coronary heart disease in asymptomatic men. Am J Cardiol 86:53, 2000.

Gibbons RJ, et al.: Guidelines for exercise testing: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Exercise Testing). J Am Coll Cardiol 30:260, 1997.

Kwok JMF, et al.: Prognostic value of a treadmill exercise score in symptomatic patients with nonspecific ST-T abnormalities on resting ECG. JAMA 282:1047, 1999.

Livschitz S, et al.: Limited clinical value of exercise stress test for the screening of coronary artery disease in young, asymptomatic adult men. Am J Cardiol 86:462, 2000.

Mahmarian JJ, et al.: Transdermal nitroglycerine patch therapy reduces the extent of exercise-induced myocardial ischemia: Results of a double-blind, placebo-controlled trial using quantitative thallium-201 tomography. J Am Coll Cardiol 24:25, 1994.

Mark DB, et al.: Exercise treadmill score for predicting prognosis in coronary artery disease. Ann Intern Med 106:793, 1987.

Mark DB, et al.: Prognostic value of a treadmill exercise score in outpatients with suspected coronary artery disease. N Engl J Med 325:849, 1991.

Morise A, et al.: Incremental evaluation of exercise capacity as an independent predictor of coronary artery disease presence and extent. Am Heart J 127:32, 1994.

Pryor DB, Harrell FE Jr, et al.: Estimating the likelihood of significant coronary artery disease. Am J Med 75:771, 1983.

Roger VL, et al.: Prognostic value of treadmill exercise testing. A population-based study in Olmsted County, Minnesota. Circulation 98:2836, 1998.

Villella A, et al.: Prognostic significance of maximal exercise testing after myocardial infarction treated with thrombolytic agents: the GISSI-2 database. Lancet 346:523, 1995.

Weiner DA, et al.: Identification of patients with left main and three-vessel coronary disease with clinical and exercise test variables. Am J Cardiol 46:21, 1980.