Recognition of an acute myocardial infarction in the patient with chest pain is a frequent challenge to the clinician. Previous studies suggest that cardiac enzymes are of limited value in identifying patients with acute MI in the emergency department. Such studies have not evaluated the use of cardiac enzyme tests to complement decision making in the population of patients clinically designated for ED release. We studied 773 ED visits by patients age greater than or equal to 30 years presenting with chest pain unexplained by thoracic trauma or radiographic abnormalities. Cardiac enzyme levels were not available to the clinicians at the time of the initial visit and disposition of these patients was determined solely by clinical and ECG evaluation. Of the 291 admitted patients, 46 had an MI; 22 of the MI patients had a normal creatine kinase (CK) level. Of the 482 patients released from the ED, 181 patients had an elevated CK level. Among the released patients were five patients with MI. Four released MI patients had a CK level greater than or equal to 200 IU/L and three had an elevated CK-MB fraction (greater than or equal to 12 IU/L). In the population of patients scheduled for release, an elevated CK-MB had sensitivity, specificity, and positive predictive value for MI of 60%, 100%, and 60%, respectively. Although cardiac enzymes cannot be used in isolation to make admission decisions, selective use of CK-MB for final screening of patients otherwise scheduled for ED release may enhance the initial admission of patients with MI at risk for unintentional release.