Whether the ST segment shift used to evaluate the presence and severity of myocardial ischemia should include the additional deviation due to decreasing amounts of baseline ST segment elevation was examined in 100 clinically normal subjects and in 124 patients with coronary disease. Exercise ST segment depression was calculated in two ways: as the difference between exercise and resting ST segment depression, but excluding any resting ST elevation (STdep), and as the total ST segment difference or excursion, including any baseline resting ST elevation (STdiff). These values were also used for separate calculation of the maximal ST/heart rate slope and delta ST/heart rate index in each case. Given partition values with matched specificity of 95% in clinically normal subjects, 150 μV of STdep was significantly more sensitive for coronary disease than 220 μV of STdiff (61% [76 of 124] versus 50% [62 of 124], p<0.005). Comparison of receiver operating characteristic curves confirmed the superior test performance of ST dep for the identification of coronary disease in this population (area under the curve 0.920 versus 0.869, p=0.0019). In contrast, detection of three-vessel coronary obstruction by standard ST segment criteria was not affected by definition of ST segment excursion. Substitution of STdiff for STdep did not change the performance of the ST/heart rate slope or the delta ST/heart rate index for either the detection of coronary disease or for the identification of three-vessel coronary obstruction. We conclude that incorporation of resting ST segment elevation into the measurement of exercise-induced ST segment depression decreases the sensitivity of standard electrocardiographic criteria for the detection of coronary disease and does not improve the performance of heart rate-adjusted ST segment criteria for the assessment of either the presence or severity of coronary obstruction. © 1990.