Identifying coronary artery disease in women by heart rate adjustment of ST-segment depression and improved performance of linear regression over simple averaging method with comparison to standard criteria

Abstract

Performance of exercise electrocardiography for the detection of coronary artery disease (CAD) in women has been limited by relatively poor sensitivity and specificity of standard test criteria. Recent studies suggest that diagnostic methods incorporating heart rate (HR) adjustment of ST-segment depression during exercise may improve the accuracy of exercise testing in women, but the relative performance of different rate-adjusted methods for this purpose is unknown. To assess the effect of gender on relative test performance of the ST-segment/HR (ST/HR) slope, the simple ST/HR index, the HR-recovery loop, and standard ST-depression criteria for the identification of CAD, the exercise electrocardiograms of 254 patients with known or suspected CAD (67 women and 187 men) and of 150 clinically normal subjects (29 women and 121 men) were analyzed. Specificity of each method was comparable in men and women: ST/HR slope 98% (118 of 121) vs 97% (28 of 29), ST/HR index 97% (117 of 121) vs 97% (28 of 29), and HR-recovery loop 96% (116 of 121) vs 93% (27 of 29). In contrast, although there was no difference in sensitivity of the ST/HR slope (95% [177 of 187] vs 93% [62 of 67]; p = not significant [NS]) or HR-recovery loop (90% [168 of 187] vs 87% [58 of 67]; p = NS) between men and women, the ST/HR index was less sensitive for CAD in women than in men (82% [55 of 67] vs 93% [173 of 187]; p < 0.05). Thus, among women sensitivity of the simple ST/HR index was lower than that of the ST/HR slope for CAD (82 vs 93%; p < 0.05), but there was no significant difference in performance of these 2 methods in men (93 vs 95%; p = NS). Each of the HR-adjusted methods was significantly more sensitive than were standard test criteria at comparable specificity in men and women. These findings support use of the more complex ST/HR slope method for the detection of CAD in women.