To analyze the prognostics value of 5 biomarkers IFIH1, IRF1, STAT1, GBP1 and IFIT3 in acute respiratory distress syndrome (ARDS).

All of 26 ARDS patients admitted to the Department of Critical Care Medicine, Zhongda Hospital Affiliated to Southeast University from January 2017 to December 2017 were included as the study objects. Peripheral blood was collected for microarray detection, and the expression levels of IFIH1, IRF1, STAT1, GBP1, IFIT3 were extracted. The relationship of 5 biomarkers and ARDS severity with 28-day cumulative mortality was examined using univariate Cox regression and receiver operating characteristic curve (ROC), with a cutoff value of P<0.05. In order to create a combined model of the five chemicals and the severity of ARDS, multivariate COX regression was utilized.

In ARDS patients, expression of five molecules was linked with 28-day cumulative mortality (P<0.05) according to a univariate COX regression analysis, but not with ARDS severity (P>0.05): IFIH1[HR 2.309 95%CI(1.162, 4.590)]; IRF1[HR 3.152 95%CI(1.043, 9.528)]; IFIT3[HR 1.793 95%CI(1.103, 2.917)]; GBP1[HR 2.342 95%CI(1.087, 5.047)]; STAT1[HR 2.492 95%CI(1.388, 4.474)]; ARDS severity[HR 1.665 95%CI(0.598, 4.628)]. According to ROC analysis, the prediction accuracy (area under the ROC curve) of five molecules and the severity of ARDS to 28-day mortality were, respectively, 0.883, 0.833, 0.900, 0.925, 0.908, and 0.625. With an accuracy of 0.950, the combined model predicted the 28-day mortality.

Five molecules were more accurate at predicting the prognosis of ARDS patients than the Berlin criteria for ARDS severity. The integrated model based on five molecules can accurately categorize patients with severe ARDS who have a bad prognosis and aid in ARDS therapy.