The ST segment shift relative to exercise-induced increments in heart rate, the ST/heart rate slope (ST/HR slope), has been proposed as a more accurate ECG criterion for diagnosing significant coronary artery disease (CAD). Its clinical utility, with the use of a standard treadmill protocol, was compared with quantitative stress thallium (TI) and standard treadmill criteria in 64 unselected patients who underwent coronary angiography. The overall diagnostic accuracy of the ST/HR slope was an improvement over TI and conventional ST criteria (81%, 67%, and 69%). For patients failing to reach 85% of their age-predicted maximal heart rate, its diagnostic accuracy was comparable with TI (77% and 74%). Its sensitivity in patients without prior myocardial infarctions was equivalent to that of thallium (91% and 95%). The ST/HR slope was directly related to the angiographic severity (Gensini score) of CAD in patients without a prior infarction (r = 0.61, p less than 0.001). The ST/HR slope was an improved ECG criterion for diagnosing CAD and compared favorably with TI imaging.