Exercise testing in women with chest pain: applications and limitations of computer analysis

The frequent occurrence of false-positive results of standard electrocardiographic (ECG) treadmill exercise testing in women limits its diagnostic value. In men, a commercially available computer-derived treadmill exercise score (TES), which quantifies the exercise ECG, has been reported to have high diagnostic accuracy. The purpose of this study was to test TES accuracy in women.

We prospectively evaluated 87 middle-aged women with chest pain and no previous history of cardiac events who underwent exercise testing and coronary arteriography (Group 1). Also, in 15 young symptomatic women with mitral valve prolapse and 15 age- and sex-matched controls (Group 2) we compared the TES score with standard exercise testing. In 25 Group 1 patients, TES could not be evaluated because of technical reasons.

By standard ECG criteria, 18 out of the 28 women in Group 1 with coronary artery disease (55%) had a false-positive test, whereas the TES score reduced the false positives to 5 (15%, P < 0.001). In Group 2, the false-positive rate was improved from 12 (40%) to zero with TES (P < 0.001). Overall specificity was thus improved from 52% to 92% by TES (P < 0.0001). There was a trend towards improving the positive predictive value (52% versus 74%) and diagnostic accuracy (57% versus 69%) (P = 0.1), but sensitivity and negative predictive value were not improved with TES.

The treadmill exercise score improves the specificity of exercise ECG especially in women with mitral valve prolapse and atypical chest pain, but its overall utility and accuracy in women is lower than that reported in men.