Whither the ST segment during exercise.

THE WIDESPREAD POPULARITY of using the ST-segment response to exercise to detect individuals with ische-mic heart disease is based largely on epidemiologic data and arteriographic correlative studies. Epidemiologic studies of asymptomatic individuals have shown that the presence of a positive ST-segment response to exercise distinguishes a group of individuals at relatively high risk of developing coronary events over the next several years.'`I Conversely, the presence of a negative test response distinguishes those individuals with a very low probability of such events. Arterio-graphic studies, in which the results of the exercise test response have been correlated with evidence of coronary luminal narrowing, also have shown that in symptomatic subjects the presence of an abnormal ST-segment response to stress in most instances correctly denotes the presence of important anatomic abnormalities of the coronary arteries .6-10 From this evidence, it has been concluded that the ST-segment response to exercise is highly accurate in predicting the presence or absence of coronary disease. However, this conclusion has been questioned in the light of recent coronary angiographic studies of asymptomatic subjects with positive exercise tests. In such an asymptomatic population the predictive accuracy of a positive ST segment response (that is, the percentage of positive test results which correctly predict important angiographically-demonstrable coronary arterial disease [table 1]) is relatively low. Thus, in a population of asymptomatic subjects, Froelicher found that only 44% of all positive tests are true positives." We also found that only 37% of positive tests in asymptomatic subjects accurately reflect important coronary disease.'2 It would seem that the results of these epidemiologic and arteriographic studies contradict one another. However, closer scrutiny of published data indicates that the disparity is only apparent. Thus, the low predictive accuracy of a positive exercise test in the angiographic studies of asymp-tomatic subjects is entirely compatible with the conclusions of epidemiologic studies, once the difference between risk ratio and predictive accuracy is appreciated (table 1). The results of epidemiologic studies2-0 (table 2) show that a group of asymptomatic subjects with positive test results is indeed at higher risk (10-15 times) of eventually developing cor-703 onary events than a group of asymptomatic subjects with negative tests (i.e., the risk ratio is high). This, however, should not be taken to imply that the large majority of subjects with positive tests have disease. In fact, the risk of subsequent coronary events in an asymptomatic individual with a positive test response (i.e. …