EARLY PREDICTIVE FACTORS OF SURVIVAL IN THE ACUTE RESPIRATORY-DISTRESS SYNDROME - A MULTIVARIATE-ANALYSIS

To identify the potential impact of novel therapeutic approaches, we s tudied the early predictive factors of survival at the onset of acute respiratory distress syndrome (ARDS) in a 24-bed medical ICU of an aca demic tertiary care hospital. Over a 48-mo period, a total of 3,511 ad ult patients were admitted and 259 mechanically ventilated patients me t ARDS criteria, as defined by American-European consensus conference, i.e., bilateral pulmonary infiltrates and Pa-O2/FIO2 lower than 200 w ithout left atrial hypertension. These patients were randomly included in a developmental sample (177 patients) and a validation sample (82 patients). Demographic variables, hemodynamic and respiratory paramete rs, underlying diseases, as well as several severity scores (SAPS, SAP S-II, OSF) and Lung Injury Score (LIS) were collected. These variables were compared between survivors and nonsurvivors and entered into a s tepwise logistic regression model to evaluate their independent progno stic roles. The overall mortality rate was 65%. SAPS-II, the severity of the underlying medical conditions, the oxygenation index (mean airw ay pressure x FIO2 x 100/Pa-O2), the length of mechanical ventilation prior to ARDS, the mechanism of lung injury, cirrhosis, and occurrence of right ventricular dysfunction were independently associated with a n elevated risk of death. Model calibration was very good in the devel opmental and validation samples (p = 0.84 and p = 0.72, respectively), as was model discrimination (area under the ROC curves of 0.95 and 0. 95 respectively). Thus, the prognosis of ARDS seems to be related to t he triggering risk factor, the severity of the respiratory illness, an d the occurrence of a right ventricle dysfunction, after adjustment fo r a general severity score.