Exercise testing in suspected coronary artery disease

Abstract

1. 1. The interpretation and selection of exercise tests depends on the pretest probability of CAD. 2. 2. Imperfect tests (like exercise tests) provide probability estimates, not definite statements (such as “the patient has CAD” or “the patient does not have CAD”). 3. 3. In patients with a low pretest probability of CAD (asymptomatic persons or men and women with nonanginal chest pain), abnormal exercise test results provide probability estimates that are much too low to conclude that the patient has CAD. 4. 4. In patients with anginal pain and normal exercise tests, the probability of CAD is too high to conclude that the patient has a normal coronary circulation. Exercise tests are not useful for trying to rule out CAD in patients with anginal pain. 5. 5. In patients with an intermediate pretest probability of CAD (men and women with atypical angina and women with typical angina), abnormal exercise tests (particularly the myocardial scintiscan) provide probability estimates that are high enough to justify starting treatment for CAD. Exercise tests are most useful in this group, a conclusion that has been reached by other methods of analysis. 67 6. 6. The myocardial scintiscan is much more useful than the exercise ECG in women. 7. 7. When CAD is strongly suspected, exercise tests have relatively little diagnostic value but may be useful for prognosis. However, clinical evidence of poor ventricular function may alone suffice to select patients with angina pectoris for coronary arteriography. Conversely, when clinical indicators of congestive heart failure are absent, the prognosis in chronic stable angina is so favorable that any further testing may be unnecessary. 8. 8. Screening asymptomatic persons for CAD is a very low yield practice. Patients who have no cardiac risk factors (hypercholesterolemia, family history of CAD, cigarette smoking, and hypertension) are at especially low risk of a primary cardiac event. 9. 9. Older men with stable typical angina are particularly likely to have left main coronary artery stenosis or three-vessel disease with poor ventricular function. The exercise ECG can identify groups of older men with a relatively high risk of having left main coronary artery stenosis. 10. 10. Physicians should be cautious when applying these recommendations to a primary care practice. The foregoing analysis is based on data obtained from patients who had been selected for coronary arteriography. There are two principal effects of biased selection of study patients: • • The pretest probability of CAD in clinical subgroups is probably lower than as shown here. The effect of this selection bias is to reduce the probability of CAD when exercise tests are abnormal and to overestimate the yield of an aggressive diagnostic approach. • • If the test performance of exercise tests could be measured in primary care practice, the sensitivity would probably be lower and the specificity higher than as shown here. Therefore, in primary care practice, the probability of disease for both normal and abnormal test results is likely to be higher than as shown here. When interpreting an abnormal test result, these two sources of bias have opposite effects and may partially offset one another.