ABSTRACT
PURPOSE: Mortality in emergency department sepsis (MEDS), sepsis-related organ failure assessment (SOFA), multiple organ dysfunction score (MODS), and serum lactate levels have shown their efficacy in the early detection of patients with a bad prognosis. However, those studies did not consider differences in treatment protocols and could not rule out the interference of these differences in treatment modalities. Hence, we aimed to assess the performance of MEDS, MODS, SOFA, and serum lactate levels for predicting a bad prognosis in patients scheduled for identical, standardized treatment protocols, EGDT. METHODS: Medical records of patients who visited a tertiary level teaching hospital and were enrolled in an EGDT program between October 2009 and May 2010, were retrospectively reviewed. MEDS, SOFA, and MODS scores were calculated and recorded along with serum lactate levels. Receiver operating characteristics (ROC) curves of those predictors of mortality were plotted, Bivariate correlation analyses with overall lengths of admission and ICU lengths of stay were done for surviving patients. RESULTS: None of the diagnostic methods (serum lactate level, MEDS, SOFA, MODS) showed a significant difference on ROC analysis (p=0.819, 0.506, 0.811, 0.873, respectively). Bivariate correlation analyses of MEDS, SOFA, MODS and overall lengths of admission showed significant results (p=0.048, 0.018, and 0.003, respectively. Pearson correlation coefficients were, 0.263, 0.312, and 0.381). Only MEDS showed a significant correlation with intensive care unit (ICU) length of stay (p=0.032, Pearson correlation coefficient = 0.332). CONCLUSION: Neither MEDS, SOFA, MODS, nor serum lactate level can predict mortality in EGDT-enrolled patients. MEDS may be correlated with ICU length of stay.