Effect of precision of ST-segment measurement on identification and quantification of coronary artery disease by the ST/HR index.

To assess the effect of varying precision of ST-segment depression measurement on test performance of the ST-segment/heart rate (ST/HR) index for the identification and quantification of coronary artery disease, the exercise electrocardiograms (ECGs) of 100 clinically normal subjects and 154 patients with angiographically proved coronary disease were reviewed. The ST/HR index was calculated by dividing the maximal additional ST-segment depression at end exercise by the exercise-induced change in heart rate. ST-segment depression was measured to the nearest 10 microV (ST10) at a point 60 ms after the J point on a computerized exercise ECG system, and was subsequently rounded down to the nearest 50 microV (ST50) and the nearest 100 microV (ST100) to simulate measurements to these precisions. An ST10/HR index partition of 1.60 microV/bpm with a specificity of 95% (95/100) in normal subjects identified the presence of coronary disease with a sensitivity of 94% (144/154). Precision of ST-segment measurement significantly affected sensitivity for coronary disease. At matched specificity of 95%, an ST50/HR index partition of 1.55 microV/bpm had a sensitivity of 88% (135/154, p less than 0.01) and an ST100/HR index partition of 1.22 microV/bpm had a sensitivity of 84% (130/154, p less than 0.001) for the detection of coronary obstructions. Comparison of receiver-operating characteristic curves (ROC) confirmed the superior overall performance of the ST/HR index using ST10 measurements for the identification of coronary disease. By contrast, test performance for the identification of three-vessel coronary disease was not affected by the precision of ST-segment measurement with no significant difference in test sensitivity or areas under respective ROCs.(ABSTRACT TRUNCATED AT 250 WORDS)