"MB Leak" patients who develop an elevated MB relative index with a normal total creatine kinase (CK) level are not as well characterized as those who have diagnostic enzyme elevations in the setting of ST elevation (elevation) or non-ST elevation acute myocardial infarction (AMI). During a 1-year period, we studied all patients hospitalized in an urban academic hospital with suspected AMI who developed an elevated MB relative index within 24 hours of presentation. Of 595 patients, 44% had MB Leak, 34% had non-ST elevation AMI and 22% had ST elevation AMI. Patients with MB Leak and non-ST elevation AMI were significantly older than those with ST elevation AMI (mean ages 69, 71, and 63 years, respectively; p <0.001), and were more likely to have previous AMI (55%, 46%, 12%; p <0.001) or past coronary revascularization (40%, 19%, 12%; p <0.001). The in-hospital death rate of patients with MB Leak was half that of patients with non-ST elevation AMI or ST elevation AMI (6%, 12%, 12%; p = 0.03). By 1 year after presentation, the death rate of patients with MB Leak (17%) was intermediate between that of non-ST elevation AMI (24%) and ST elevation AMI (14%). Within the MB Leak group, those with elevated absolute CK-MB levels were at highest risk. In a multivariable model using MB Leak as the referent, the relative risks for 1 year death were 1.4 (95% confidence interval, 0.9 to 2.2) for patients with non-ST elevation AMI and 1.7 (0.8 to 3.4) for patients with ST elevation AMI. Patients with MB Leak are at high risk for cardiovascular events in the hospital and for death by 1 year. Therefore, they may benefit from early aggressive therapy and risk stratification. These results suggest that CK-MB should be measured in all patients with suspected AMI, regardless of their total CK level.