Evaluation of Systemic Inflammatory Response Syndrome (SIRS) Score as a Predictor of Mortality in Trauma Patients

Abstract Background: Systemic Inflammatory Response Syndrome (SIRS) score has been widely used for mortality risk assessment, out come and cost containment, it provides an excellent indication of the risk of mortality for trauma patients, but its validity is not assessed well in literature. In this study, we validate this score by measurement of its discrimination and calibration in a large population of trauma patients. Methods: Two large trauma centers in central part of Iran participated in the study. From December 2001 to May 2004; centers collected the information required for SIRS on all trauma patients at the time of admission. The performance of SIRS model was evaluated by assessing the discrimination (area under the receiver operator characteristic plot) and calibration (by calculating the expected and observed number of survivors and deaths in deciles of risk). Results: 4,300 patients were selected for analysis. Male patients comprised 90.4% of study population SIRS was present in 33% of patients at admission. Mortality rate was 9% (n = 387). It was 2.8% in negative SIRS patients and 9.7% in positive SIRS patient. The area under the Roc plot was 0.90 (0.89–0.91). The fit in groups of patients with different risks of death was 7.32. Conclusion: In summary the SIRS model is simple enough for it to be wildly used in trauma, it requires the collection of only four variables at the time of admission and it has good predictive power.