Abstract
Elderly patients in the USA account for 26-50% of all intensive care unit (ICU) admissions. The applicability of validated ICU scoring systems to predict outcomes in the "Oldest Old" is poorly documented. We evaluated the utility of three commonly used ICU scoring systems (SAPS II, SAPS III, and APACHE II) to predict clinical outcomes in patients > 90 years. 1,189 surgical procedures performed upon 951 patients > 90 years (between 2000 and 2010) were analyzed. SAPS II, SAPS III, and Acute APACHE II were calculated for all patients admitted to the SICU. Differences between survivors and nonsurvivors were analyzed using the Student's t-test and binary logistic regression analysis. A receiver operating characteristic (ROC) curve was constructed for each scoring system studied. The area under the ROC curve (aROC) for the SAPS III was 0.81 at a cut-off value of 57, whereas the aROC for SAPS II was 0.75 at a cut-off score of 44 and the aROC for APACHE II was 0.74 at a cut-off score of 13. The SAPS III ROC curve for prediction of hospital mortality exhibited the greatest sensitivity (84%) and specificity (66%) with a score of 57 for the "Oldest Old" population.