Background: The aim of the present study was to compare 16 routine clinical and laboratory parameters, acute physiologic and chronic health evaluation (APACHE) and sequential organ failure assessment (SOFA) score for their value in predicting mortality during hospital stay in patients admitted to a general intensive care unit (ICU).
Methods: A retrospective observational clinical study was carried out in a 15-bed ICU in a university hospital. Nine hundred and thirty-three consecutive patients with ICU stay > 24 h (36.2% surgical, 29.1% medical and 34.7% trauma) were observed. Blood sampling, patient surveillance and data collection were performed. The primary outcome was mortality in the hospital. We used receiver operating characteristic (ROC) analyses and logistic regression to compare the 16 relevant parameters, APACHE II and SOFA scores.
Results: Two hundred and thirty-three out of the 933 patients died (mortality 25.0%). One laboratory parameter, serum osmolality [area under the curve (AUC) 0.732] had a predictive value for mortality which lay between that of APACHE II (AUC 0.784) and SOFA (AUC 0.720) scores. When outcome prediction was restricted to long-term patients (ICU stay > 5 days), serum osmolality (AUC 0.711) performed better than either of the standard scores (APACHE AUC 0.655, SOFA AUC 0.636). Using logistic regression analysis, the association of clinical parameters, age and diagnosis group with mortality was determined.
Conclusion: Elevated serum osmolality at ICU admission is associated with an increased mortality risk in critically ill patients. Serum osmolality is cheaper and more rapid to determine than the scoring systems. However, further studies are needed to evaluate the predictive value of serum osmolality in different patient populations.