The purpose of this study was to determine whether the parameters commonly used to evaluate a training effect can be compared when serial tests are performed using different protocols. Thirty-two patients with stable coronary artery disease performed both the standard and modified Bruce protocols in a random order, 1 h apart. Physiologic variables at matched exercise stage I including heart rate (99 +/- 12 vs 101 +/- 14 bpm), rate-pressure product (15 +/- 3 vs 15 +/- 3 bpm x mmHg x 10(3)), and VO2 (13.5 +/- 2.1 vs 13.0 +/- 3.0 ml.kg-1 x min-1) were not significantly different for exercise tests performed using the standard vs modified Bruce protocols respectively. Similarly, these parameters were also nearly identical at matched exercise stage II. However, peak VO2 was significantly higher using the standard vs modified Bruce protocol, although the difference was small. Therefore, these data indicate that a difference in the heart rate response at matched submaximal workrates on tests using these two protocols before and after a training program is most likely due to a training effect. Conversely, improvements in peak VO2 using the standard vs modified Bruce are at least in part due to inherent differences in responses between these two protocols.