945-118 Validation of a Simple Exercise-Test Based Method for Early Identification of Multivessel Coronary Disease After Acute Myocardial Infarction

Abstract

The presence of multivessel coronary disease (MVD) is a powerful determinant of prognosis after acute myocardial infarction (AMI). The use of early exercise testing (ExT) in the detection of MVD has produced conflicting results in part due to the use of submaximal protocols and the selective use of angiography. We routinely perform coronary angiography after AMI irrespective of the results of ExT. In this setting, we studied the use of multivariate analysis of clinical and exercise parameters derived from maximal ExT, to detect the presence of MVD in consecutive patients after AMI. Of 240 patients with AMI (October 1992-May 1994), 117 had both early (8 ± 2 days) maximal ExT and coronary angiography (13 ± 8 days) after uncomplicated AMI. No complications occurred related to ExT. Using a discriminant analysis, a statistical model was established in the first 72 patients (learning group) and validated in the next 45 patients (testing group). Baseline variables were comparable in terms of age (56 yrs vs 59 yrs), Q-wave AMI (83% vs 76%), inferior AMI (65% vs 69%), use of thrombolytic therapy (64% vs 60%) and MVD (44% vs 36%). Of 9 clinical and 15 exercise variables which were incorporated in the discriminant analysis, the number of METs achieved (p < 0.0005), maximal ST segment depression in V5 (p < 0.0051 and maximal exercise heart rate were the only variables significantly associated with the presence of MVD in the learning group. The discriminant function, using these 3 variables, correctly classified 75% of patients with MVD in the learning group and 79% of patients in the testing group, whereas the use of maximal ST depression alone, the most conventional index, correctly classified only 60% of patients in the testing group (p < 0.005). Thus, maximal ExT can be safely performed early after uncomplicated AMI. The accuracy of early ExT in the detection of MVD can be improved using a discriminant analysis that incorporates 3 easily derived exercise variables which were prospectively validated.