Predictors of death following ICU discharge.

In order to study possible predictors of early after-discharge mortality (EADM), 700 consecutive patients discharged alive from three intensive care units (ICUs) were followed up 2 months after their admission. The observed mortality was 7.3%. Univariate analysis indicated a strong statistically significant association of the following variables with mortality: initial simplified acute physiological score (SAPS) greater than 10, therapeutic intervention score in the first 24 h of admission greater than 20, age greater than or equal to 65, length of stay in the ICU greater than 10 days, and low educational level. The multiple logistic regression analysis included as predictive independent variables the SAPS, organs or systems failure, age, and length of stay. The model built upon these four variables was able to identify a group of patients at high risk (21-46%) of EADM. We conclude that some simple variables can be used as useful markers of patients groups at high risk of EADM.