Enhanced evaluation of treadmill tests by means of scoring based on multivariate analysis and its clinical...

Six hundred eight patients being evaluated for chest pain who did not have valvular disease, cardiomyopathy, left ventricular hypertrophy or bundle branch block, and were not receiving digitalis, had treadmill tests and coronary angiograms. In 351, various exercise variables were correlated by multivariate analysis to coronary artery disease (CAD). In men, significant variables were: (1) maximal heart rate achieved less than 80% of maximal predicted heart rate (Mx PHR), (2) ST-T change greater than or equal to 1 mm, (3) age greater than or equal to 55 years and (4) treadmill time (TT) less than 8 minutes. These variables rated diagnostic scores of 9, 6, 5, and 3, respectively. A score of greater than or equal to 7 was considered diagnostic of CAD. In a test group of 192 men in which ST-T change was compared with treadmill score, sensitivity was 65 versus 85%, specificity 79 versus 74% and accuracy 69 versus 83%. In women, maximal heart rate less than 90% of Mx PHR and TT of less than 6 minutes were significant, with an accuracy of 75%. Moreover, 89% of incomplete tests and 70% of tests in patients with previous myocardial infarction were also correctly diagnosed. This method allows convenient use of significant exercise variables for clinical purposes with improved results.