We compared the predictive value of the Rush score with the Thrombolysis In Myocardial Infarction (TIMI) risk score in unselected patients with an acute coronary syndrome and evaluated the effect of compliance with established guidelines on the accuracy of these models. The Registry of Acute Coronary Syndromes is a retrospective registry of 3,754 consecutive patients (38% women; mean age 67 years) who presented with acute coronary syndrome to the emergency department between April 1, 1999, and December 31, 2000, at 9 hospitals. The primary end point was death, myocardial infarction, or urgent revascularization during hospitalization. Rush classification was based on quartiles of predicted risk of cardiac complication (<2% for class I vs >15% for class IV). The TIMI score was implemented as published. Compliance with guidelines for acute coronary syndrome was assessed with a 4-point scale based on the aggregate use of aspirin, beta blockers, heparin, and glycoprotein IIb/IIIa inhibitors. Fifteen percent of patients met the primary end point. The primary end point rates for TIMI scores 0/1, 2, 3, 4, 5, and 6/7 were 11%, 14%, 13%, 11%, 14%, and 12%, respectively (p = NS). The primary end point rates for Rush classes I, II, III, and IV were 6%, 8%, 9%, and 17%, respectively (p <0.001). After controlling for compliance with established guidelines, the odds ratio of an event increased by 46% for each unit increase in Rush score (p <0.001). After adjusting for the Rush score, the odds ratio decreased by 54% for each unit increase in compliance (p <0.001). Thus, compliance with current American College of Cardiology/American Heart Association guidelines significantly improves prognosis, regardless of the risk score. The use of established risk scores may overestimate event rates in unselected populations.