Abstract

BACKGROUND
Many critically ill patients in the ED are hospitalized to the ICU, but most prognosis predicting systems have been developed based on the physiochemical variables of the critically ill in the ICU. The objective of this study is to identify prognostic predictors early in the ED when compared with well-known predictors in the ICU and estimate their predictive abilities.
METHODS
An observational prospective study was performed in an urban ED. Information of all the critically ill patients admitted to the ICU via the ED including vital signs, laboratory results, and physiochemical scoring systems were checked during 6 months and divided into the early stage for the ED and the late stage in the ICU. Poor outcome was defined as 28-days mortality. After checking for significant predictors among them through univariate analysis, we identified the most discriminating predictors in each stage using logistic regression and a decision tree analysis.
RESULTS
A total of 246 patients were enrolled. In univariate analysis, the significant predictors including central venous pressure, fraction of inspired oxygen (FiO2), pressure of arterial oxygen/fraction of inspired oxygen (PaO2/FiO2), albumin, mortality in emergency department sepsis, acute physiology and chronic health evaluation II, simplified acute physiology score II, and sequential organ failure assessment scores were identified in the early stage, while PaO2/FiO2, base excess, unmeasured anion, albumin, anion gap, albumin-corrected anion gap, APACHEII, SAPSII, SOFA, and rapid emergency medicine score were identified in the late stage. Through a decision tree analysis, PaO2/FiO2 and SAPSII were revealed as the most discriminating predictors in the ED and ICU, respectively.
CONCLUSIONS
The prognosis discriminating predictor in critical patients was different between the ED and ICU. Emergency physicians should pay more attention to the critical patients having low PaO2/FiO2.

• Keywords: critically ill;emergency department;intensive care unit;prognostic predictor;