Several methods for evaluation of exercise stress testing (EST) have been described in order to analyze the relationship between ST segment changes and heart rate. The ST/HR slope has demonstrated to be worthwhile in identifying severe coronary artery disease (CAD). We applied this method in patients catalogued as borderline in the traditional exercise test to find out if they could be considered to have a severe CAD. The patients were divided into two groups: the A, which included 41 patients with borderline EST, and the group B with 41 patients with normal EST. Age, risk factors, double product and ST/HR slope were evaluated. The testing was done on a treadmill with the Bruce protocol. Four patients in group A had ST/HR slope greater than 6.0 mu Volt/beat/min (two of them with borderline EST). Whereas all patients in group B had ST/HR slope values less than 6.0. We concluded this is a sensitive method for discrimination between normal and borderline EST. We found no correlation among age, sex, risk factors, double product and ST/HR slope. Approximately 10 percent of borderline EST would be underestimated with the traditional method. The calculation of the slope obtained its maximum applicability in patients with almost maximum EST.