During a consecutive 17 month period, 15 trauma patients were diagnosed as having a myocardial contusion on the basis of abnormal ECG in 14 patients, elevated creatinine phosphokinase (CPK) in 13 patients, and elevated CPK-Muscle Brain (CPK-MB) isoenzyme determination in 11 patients. Using these screening modalities, the incidence of myocardial contusion in patients with blunt chest trauma increased from 7% when viewed retrospectively to 15% when viewed prospectively. Five patients had determinations of cardiac index of which 4 were less than 2.9 L/min . M2. An additional 9 patients underwent a standard fluid challenge of 500 ml of 5% plasma protein faction infused over 30 min allowing construction of a Starling myocardial performance curve. Of these patients, 6 had biventricular dysfunction, 1 had isolated abnormal left ventricular function, and 2 had isolated abnormal right ventricular function in the absence of chronic obstructive pulmonary disease or preexisting heart disease. Multiple gated acquisition scans (MUGA) were abnormal in 6 patients and normal in 4 patients. Of the latter group, 3 had biventricular dysfunction and 1 had depressed cardiac index. Morbidity and mortality for myocardial contusion occurred in 40% (6 of 15) of patients in this series. Direct hemodynamic measurement with construction of a Starling curve was useful in monitoring the degree of impairment and subsequent recovery of myocardial function. This information was important in delaying semiurgent operations or in determining the best alternative of otherwise equally acceptable methods of patient management.