Background: Modified Sequential Organ Failure Assessment (MSOFA) scoring system has been developed as critical care triage in centers with limited resources. Previous studies evaluated the performance of MSOFA only in critically ill surgical patients and showed low precision in predicting mortality. Addition of another variable to improve MSOFA’s performance merits further investigation. Hyperglycemia in critically ill patients without previous history of diabetes (stress hyperglycemia) has been shown as an independent risk factor of mortality. Objective: To evaluate the performance of MSOFA and to assess the addition of admission blood glucose to predict mortality in critically ill patient without previous history of diabetes.
Methods: A validation study with prospective cohort design recruited critically ill medical and surgical patients admitted to Cipto Mangunkusumo Hospital, Jakarta during August to December 2013. History taking, physical examination, peripheral oxygen saturation, Glasgow coma scale, creatinine, blood glucose and HbA1C were obtained within 24 hour of admission. Assessed outcome was mortality within 28 days. Performance of MSOFA was evaluated with Hosmer-Lemeshow goodness of fit test and measurement of area under the curve (AUC). Results: A total of 150 patients completed the study protocols. Mortality was observed in 52 (34.67%) subjects with sepsis as the most prevalent diagnosis. Calibration of MSOFA is shown by Hosmer-Lemeshow test x2=13.748 (p=0.05). Receiver Operating Curve (ROC) of MSOFA shows an AUC of 0.83 (CI95% 0.76-0.89). Stress hyperglycemia was evident in 79 (52.67%) subjects. Addition of blood glucose to MSOFA scoring system does not improve its performance. Conclusion: MSOFA shows good calibration and discrimination in both critically ill medical and surgical patients. Addition of blood glucose to MSOFA scoring system does not improve its performance.
Key words: MSOFA, blood glucose, mortality, critically ill.