There is mounting evidence that exercise tolerance is an important predictor of heart disease. Our objective was to determine if decreased exercise tolerance, as estimated by physicians, may be useful in stratifying risk in Emergency Department (ED) patients with potential acute coronary syndromes. We conducted a prospective cohort study on a convenience sample of ED patients at an urban teaching hospital. Patients with chest pain, dyspnea, syncope, or epigastric pain who were evaluated for acute coronary syndromes were included. Clinical and laboratory data were recorded. In addition, the Emergency Physicians were asked to estimate the exercise tolerance of the patient as excellent, good, bad, or very poor. The primary outcome of the study was myocardial infarction (MI) or death in patients stratified by physician-perceived exercise tolerance (excellent or good vs. bad or very poor). There were 166 patients enrolled in the study. Nine patients (5%) had an MI; there were no deaths. Physicians reported exercise tolerance as excellent in 33 patients, good in 63, bad in 50, and very poor in 20. The unadjusted risk of MI was significantly elevated in patients with physician-perceived decreased exercise tolerance (relative risk = 4.8, 95% confidence interval 1.03-22). After adjustment for age, sex, and major cardiovascular risk factors, decreased exercise tolerance remained a significant predictor of MI (adjusted odds ratio = 7.3, 95% confidence interval 1.2-46). Exercise tolerance, as estimated by clinical impression, may be an important predictor of complications in ED patients presenting with potential acute coronary syndromes.