Quantitative grading of exercise stress test for patients with coronary artery disease using multivariate discriminant analysis

A group of 111 male patients who had undergone maximal or symptom-limited maximal exercise stress testing for evaluation of coronary artery disease (CAD) were subjected to coronary angiography. Group I comprised 33 patients with normal or single-vessel disease (SVD), while 78 patients with double-vessel disease (DVD) or triple-vessel disease (TVD) formed Group II. On univariate analysis of the exercise test, the following variables were found to be of significance in discriminating between the two groups: age, exercise duration, double product (heart rate X systolic blood pressure) at peak exercise, duration of ST-segment depression, number of leads showing ST depression, ST depression of 1 mm or more, configuration of the depressed ST segment, and diastolic blood pressure response to exercise. Multivariate analysis however revealed that only the following five variables had significant discriminant power: number of leads showing ST depression, exercise duration, double product, diastolic pressure response, and ST-segment configuration. On the basis of their relative importance, a regression equation was developed to give a quantitative score to individual patients. A score of less than zero detected multivessel disease (MVD) with high specificity (94%) and sensitivity (70%), while a score of 15 or more almost excluded MVD (sensitivity 87%). The scoring system as reported here improved the exercise stress test interpretation when compared with the conventional reporting system.