To determine the utility of graded treadmill exercise test (ETT) in estimating severity of obstruction in postoperative patients with aortic stenosis, we compared results to a nonoperated aortic stenosis group. Cardiac catheterization was performed in all patients. The interval between surgery and ETT in the postoperative group was 7.5 years (range 1.1-12.9). Mean and range values for postoperative versus nonoperative aortic stenosis groups respectively were: age 14.5 (7.9-20.8) versus 12.5 years (5.9-19.7); peak left ventricular pressure 150 (101-190) versus 144 mmHg (100-183); peak systolic ejection gradient (PSEG) 41 (15-75) versus 38 mmHg (14-80); left ventricular end-diastolic pressure (LVEDP) 13.4 (7-20) versus 13.5 mmHg (8-20); and cardiac index 3.7 (2.4-5.6) versus 3.8 liters/min/m2 (2.6-4.9) (p for each = NS). ETT in the nonoperated patients correctly identified all those with PSEG greater than or equal to 50 mmHg (sensitivity 100%), but specificity was 56%, (i.e., the proportion correctly identified with PSEG less than 50 mmHg). By contrast, sensitivity and specificity for postoperative aortic stenosis patients were 67% and 63%, respectively. Furthermore, among nonoperated aortic stenosis patients ST depression on ETT identified those with increased left ventricular end-diastolic pressure. For LVEDP greater than or equal to 12 mmHg, sensitivity was 75% and specificity 100%, corresponding values for postoperative aortic stenosis patients were 45% and 33%, respectively. In the nonoperated aortic stenosis group with a PSEG greater than or equal to 50 mmHg, sensitivity and specificity for ETT-induced change in systolic blood pressure of less than or equal to 30 mmHg were 100% and 56%, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)