This study evaluated whether an exercise treadmill test could predict restenosis in 289 patients 6 months after a successful emergency angioplasty of the infarct-related artery for acute myocardial infarction. After excluding those with interim interventions (64), medical events (36), or medical contraindications to follow-up testing (25), both a treadmill test and a cardiac catheterization were completed in 144 patients, 88% of those eligible for this assessment. Four patients with left bundle branch block or pacemaker rhythm at the time of treadmill testing were also excluded from analysis. Of six follow-up clinical and treadmill variables examined by multivariable logistic regression analysis, only exercise ST deviation was independently correlated with restenosis at follow-up (chi 2 = 5, p = 0.02). The clinical diagnosis of angina at follow-up, although marginally related to restenosis when considered by itself (p = 0.04), did not add significant information once ST deviation was known. The sensitivity of ST deviation of 0.10 mV or greater for detecting restenosis was only 24% (13 of 55 patients), and the specificity was 88% (75 of 85 patients). The sensitivity of exercise-induced ST deviation for detection of restenosis was not affected by extent or severity of wall motion abnormalities at follow-up, by the timing of thrombolytic therapy or of angioplasty, or by the presence of collateral blood flow at the time of acute angiography. A second multivariable analysis evaluating the association of the same variables with number of vessels with significant coronary disease at the 6-month catheterization found an association with both exercise ST deviation (p = 0.003) and exercise duration (p = 0.04). Angina symptoms and exercise treadmill test results in this population had limited value for predicting anatomic restenosis 6 months after emergency angioplasty for acute myocardial infarction.