We evaluated cardiac troponin T (cTnT) and creatine kinase-MB (CK-MB) for risk stratification of chest pain unit (CPU) patients. We studied 383 consecutive patients with chest pain assigned to our CPU by emergency department physicians. At baseline all had normal or nondiagnostic electrocardiograms, no high-risk clinical features, and negative CK/CK-MB. CK-MB and electrocardiograms were taken at 0, 4, 8, and 12 hours and cTnT at 0, 4, and 8 hours. Eight patients (2.1%) were CK-MB positive and 39 (10.2%) were cTnT positive, including all but 1 CK-MB-positive patient. All marker-positive patients were detected by 8 hours. Seven cTnT-positive patients and 1 cTnT-negative patient had myocardial infarction (p <0.0001). cTnT-positive patients were older, less likely to be women or smokers, and more often had diabetes mellitus or known coronary disease (CAD). Seventy-one percent of patients underwent diagnostic testing. cTnT-positive patients more often underwent angiography (46% vs 20%) and underwent stress testing less often (28% vs 57%) than cTnT-negative patients. When performed, their stress tests were more often positive (46% vs 14%) and they more often had angiographically significant lesions (89% vs 49%) and multivessel disease (67% vs 29%). There were no short-term deaths. Long-term mortality was higher in cTnT-positive patients (27% vs 7%, p <0.0001). Thus, cTnT identified more CPU patients with myocardial necrosis and multivessel CAD than CK-MB and a population with high long-term mortality risk. Routine use of cTnT in CPUs could facilitate risk stratification and management.