Myocardial contusion remains an elusive clinical entity, which consumes a disproportionate amount of scarce and expensive critical care resources for the purpose of cardiac monitoring. This study attempts to define a group of patients at high risk who can be identified from the available data present at the time of admission. All patients admitted with the suspicion of a myocardial contusion over a 3-year period were retrospectively studied. The records were examined for history, physical findings, electrocardiographic (ECG) results, creatine kinase levels, Injury Severity Score (ISS), and echocardiographic findings. A diagnosis of a myocardial contusion was made if patients had an ECG consistent with acute injury, increased creatine kinase-MB, or an abnormal echocardiogram consistent with acute injury. Patients were stratified into two groups: Group 1 patients satisfied the criteria for a myocardial contusion and Group 2 patients lacked sufficient evidence to substantiate this diagnosis. The records were then examined for the presence of factors available in the emergency room that might be predictive of a myocardial contusion or its complications. A total of 88 patients were evaluated; 27 of these were found to have a myocardial contusion (Group 1) with 61 patients placed in Group 2 (no myocardial contusion). Group 1 patients had an abnormal admission ECG (p less than 0.05), and an ISS greater than or equal to 10 (p less than 0.05). Multivariate analysis identified two factors predictive of a myocardial contusion: an abnormal ECG and an ISS greater than 10. When these two predictors were absent, the probability of a myocardial contusion was 1%. No predictors of a complication of a myocardial contusion were identified. These data suggest that a combination of easily obtained variables in the emergency department can be used to select a patient population at high risk for myocardial contusion. Prospective evaluation of these variables is necessary.