BackgroundThe failure of therapies aimed at modulating systemic inflammatory response syndrome and decreasing multiple organ failure (MOF) has been attributed in part to the inability to identify early the population at risk. Our objective, therefore, was to develop predictive models for MOF at admission and at 12, 24, and 48 hours after injury.MethodsLogistic regression models were derived in a data set with 411 adult trauma patients using indicators of tissue injury, shock, host factors, and the Acute Physiology Score-Acute Physiology and Chronic Health Evaluation III (APS-APACHE III).ResultsMOF was diagnosed in 78 patients (19%). Injury Severity Score, platelet count, and age emerged as predictors in all models. Transfused blood, inotropes, and lactate were significant predictors at 12, 24, and 48 hours, but not at admission. The APS-APACHE III emerged only in the 0- to 48-hour model and offered minimal improvement in predictive power. Good predictive power was achieved at 12 hours after injury.ConclusionPostinjury MOF can be predicted as early as 12 hours after injury. The APS-APACHE III added little to the predictive power of tissue injury, shock, and host factors.