Prospective validation of an acute respiratory distress syndrome predictive score

Abstract

We derived an Acute Respiratory Distress Syndrome Score (ARDS Score) from previously described training set data. To validate its diagnostic accuracy for identifying a complicated course (early death or prolonged intubation) in acute lung injury, 50 patients were prospectively scored using an ARDS Score decision threshold of > or = 2.5 to discriminate between an uncomplicated (successful extubation after < or = 14 d) and complicated course. Predictor factors incorporated in the ARDS Score were collected on Day 4 and Day 7 of ARDS and included PaO2/PAO2 ratio, required positive end-expiratory pressure (PEEP), and chest radiograph progression. The diagnostic accuracy of the ARDS Score for determining a complicated course as well as overall survival was compared with three other available indices. Using receiver operating characteristic (ROC) analysis, the ARDS Score and Lung Injury Score (LIS) had the greatest diagnostic accuracy for determining a complicated course, but the Simplified Acute Physiology Score (SAPS Score) (score > or = 14) more accurately identified survival. The LIS components of static respiratory system compliance (Crs) and chest radiograph description did not differ between patient groups. The interobserver concordance of the dynamic chest radiograph interpretation included in the ARDS Score was significant (p < 0.05). We conclude that the previously derived ARDS Score has valid diagnostic accuracy for identifying patients with ARDS who will follow a complicated course.