- The ability to provide an accurate prognosis in an intensive care unit is of major importance. Numerous risk scores have been developed to predict hospital mortality based on demographic, physiologic, and clinical data. These scores were universally developed in general medical or surgical intensive care units. Patients admitted to a cardiac care unit differ in many aspects from those admitted to general medical intensive care units. Few patients require mechanical ventilation and prolonged intensive care. Performance of risk scores developed for patients with acute myocardial infarction (AMI) in this subgroup is unknown. We prospectively studied 51 consecutive patients who were admitted to a cardiac care unit from September 2006 to March 2008 for AMI and received mechanical ventilation for >24 hours. Acute Physiology and Chronic Health Evaluation II (APACHE II), Thrombolysis In Myocardial Infarction, and Global Registry of Acute Coronary Events risk scores were calculated for each patient. Mortality rates were extrapolated based on these 3 risk scores. Twenty-two of 51 patients (43%) died in hospital. Age, mean arterial pressure, urea, albumin, hemoglobin, need for vasopressors, and estimated glomerular filtration rate were predictive of mortality. APACHE II and Global Registry of Acute Coronary Events scores were higher in nonsurvivors but Thrombolysis In Myocardial Infarction risk score was not predictive of mortality. APACHE II score had the highest value for area under receiver operator characteristics curve for mortality prediction. In conclusion, patients with AMI requiring mechanical ventilation have a high mortality rate. This risk is predicted by co-morbidities better than by direct cardiac parameters. Consequently, conventional AMI risk scores do not perform well in this very sick population and the APACHE II score better predicts their short-term outcome.