Exercise treadmill testing (ETT) is considered an unreliable method for detection of restenosis after percutaneous transluminal coronary angioplasty (PTCA). The studies on which this belief is based have used quantitative coronary angiography (QCA) as a reference. The inherent limitations of angiography have been demonstrated by intravascular ultrasound (IVUS). To determine the value of ETT for detecting restenosis when IVUS criteria are used to define restenosis, we studied 29 patients with angiographically documented 1-vessel coronary disease (<35% stenosis in all nondilated segments) who underwent angioplasty. ETT was performed < or =2 weeks before follow-up angiography and IVUS imaging. Only patients without any abnormalities precluding an accurate reading of the ST segment during ETT were included. Restenosis was defined as a > or =50% diameter stenosis by QCA criteria and as a cross-sectional area narrowing of > or =75% by IVUS. The ETT was positive in 12 patients and restenosis was documented by QCA and IVUS in 38% and 48% of cases, respectively. Sensitivity, specificity, and positive and negative predictive values of ETT when QCA was used as a reference were 55%, 67%, 50%, and 71%, respectively. This corresponded to an accuracy of 62% when compared with QCA. Sensitivity, specificity, and positive and negative predictive values were 79%, 93%, 92%, and 82% when ETT was compared with IVUS, with an accuracy of 86% (p = 0.002). Thus, ETT is a reliable noninvasive method for detecting the presence of restenosis after PTCA in patients with 1-vessel coronary artery disease. The presence of > or =75% cross-sectional narrowing shown by IVUS is well correlated with > or =1-mm ST-segment depression at follow-up ETT after PTCA.