외상 중환자에서 Systemic Inflammatory Response Syndrome(SIRS) 점수와 다른 예후예측 지표들과의 비교

Background: Since the systemic inflammatory response syndrome (SIRS) was first introduced in 1992, it has been well-known as an significant indicator to predict the level of severity and outcomes in medical, surgical and trauma patients for the last ten years. There are several other indicators such as triage Revised Trauma Score (t-RTS), Triage Score (TS), Trauma and Injury Severity Score (TRISS), Acute Physiology and Chronic Health Evaluation (APACHE), and Simplified Acute Physiology Score (SAPS) to predict trauma patients` severity and outcomes, and the authors focused on comparing the SIRS with those five other indicators to predict severity and outcome of traumatic intensive care unit (ICU) patients. Material and Method: The retrospective medical records of four hundred forty one consecutive ICU trauma patients from the emergency center, Masan Samsung Hospital from March 2002 and February 2003 have been carefully examined, and among them, three hundred fifty six were included as the target of this research. The t-RTS, TS and TRISS were calculated based on the record from the emergency center and operating rooms while SIRS, APACHE Ⅱ and SAPS Ⅱ were measured by the data for the first twenty four hours after their ICU hospitalization. The predictability of SIRS and other indicators was evaluated by using two by two decision matrices and receiver operating characteristic (ROC) curve analysis. Result: Using injury severity score 15 as the gold standard, the accuracy of SIRS was higher than t-RTS, yet it was not higher than the one of TS (p<0.01). And the areas under the ROC curves of the SIRS, TS and t-RTS were 0.645±0.073, 0.803±0.054 and 0.766±0.054, respectively (p<0.01 vs. SIRS). Using survival as the gold standard, the accuracy of SIRS was remarkably lower than the ones of TRISS, APACHE Ⅱ and SAPS Ⅱ(p<0.01 vs. SIRS). And the areas under the ROC curves of the SIRS, TRISS, APACHE Ⅱ and SAPS Ⅱ were 0.665±0.052, 0.925±0.023, 0.946±0.025, and 0.965±0.017, respectively(p<0.001 vs. SIRS). Conclusion: The authors concluded that even though SIRS was closely related to death rate and injury severity score, yet its predictability was lower than the ones of other indicators mentioned above.