In a prospective randomized open trial with 30-day follow-up, we compared a troponin-I-based protocol to 'standard management' for the diagnosis and risk stratification of patients with acute non-ST-elevation chest pain. Patients with acute chest pain (n=400) were randomized to standard diagnostic tests and management, or a protocol based on the admission ECG and the troponin-I result 6 h after onset of chest pain. Low-risk patients were discharged early from CCU; high-risk patients were treated with medical therapy or referred for in-patient angiography as appropriate. We measured length of CCU stay, and followed all patients for major adverse cardiac events (MACE) of death, non-fatal myocardial infarction (MI), or urgent revascularization during the admission and for 30 days post-discharge. The troponin protocol allowed earlier discharge in the low-risk group (10 vs. 30 h, p<0.001) with no excess of adverse events compared to standard management (3% vs. 5%, p=0.32). It identified a group of patients at moderate risk of cardiac events (15% MACE rate during admission and 30-day follow-up), and a high-risk group (75% MACE rate) more accurately than did standard management. The prognostic power of troponin testing in combination with the admission ECG was higher than with either test used alone. The protocol improved the efficiency of low-risk patient management, and improved patient risk stratification. This study adds to the evidence favouring troponin evaluation as part of the management of acute coronary syndromes.