Prediction of outcome in patients with acute respiratory distress syndrome by bronchoalveolar lavage inflammatory mediators.

Lin, Wei-Chieh; Lin, Chiou-Feng; Chen, Chia-Ling; Chen, Chang-Wen; Lin, Yee-Shin

Acute respiratory distress syndrome (ARDS) is characterized by overwhelming lung inflammation. This study explored the inflammatory mediators in bronchoalveolar lavage fluid (BALF) for prognostic relevance in patients with infection-induced ARDS. Thirty-nine patients with infection-induced ARDS (28 pneumonia and 11 extrapulmonary sepsis) and two patients with cardiogenic lung edema as the control were included. The expression profiles of inflammatory mediators in BALF were compared between ARDS and cardiogenic lung edema. A group of inflammatory mediators that showed higher expression in ARDS was analyzed for their relationships with clinical features and outcome. We found that 17 patients who died had higher levels of interleukin (IL)-6 (P = 0.012), IL-8 (P = 0.001) and monocyte chemoattractant protein-1 (P = 0.036) in BALF compared with those who survived. Furthermore, there was an inverse relationship between the BALF levels of IL-6 (P = 0.026), IL-8 (P = 0.008) and macrophage inflammatory protein (MIP)-1 alpha (P = 0.048) and the changes of lung compliance between days 1 and 4, whereas the BALF levels of IL-8 (P = 0.033) and MIP-1 alpha (P = 0.029) were positively correlated with the changes of sequential organ failure assessment scores between days 1 and 4. In multivariate logistic regression analysis, only IL-8 (P = 0.013) and lung injury score (LIS) (P = 0.017) independently predicted the mortality, and IL-8 (P = 0.002) was most likely predictive of mortality in analysis of area under the receiver operating characteristic curve. In conclusion, we show the expression profiles of inflammatory mediators in BALF of infection-induced ARDS. Among the mediators, IL-8 is the most significant predictor for mortality, and several mediators are correlated with clinical severity. However, potential selection bias due to limited control subjects and lack of serum inflammatory mediator data suggest a necessity of further studies to confirm our findings.