EUROISMAR 2019

Acute Respiratory Distress Syndrome (ARDS), as characterized by the onset of clinically significant hypoxemia and diffuse pulmonary infiltrates, has been a challenge to the critical care physicians due to high death toll rate. Categorization of the severity of ARDS is based on degree of hypoxemia enumerated by partial pressure of oxygen to the fraction of inspired oxygen (PaO2/FIO2) ratio and chest X-ray. ARDS diagnostic criteria is based on Berlin definition and classified as mild ARDS (P/F between 200-300), moderate ARDS (P/F between 100-200), severe ARDS (P/F between <100). Due to complex etiology of ARDS, efforts are required to apply system biology tools to understand disease progression and to improve survival prediction. In this direction, we have applied nuclear magnetic resonance (NMR) based metabolomics to understand heterogeneous biology of ARDS. The NMR spectroscopy of mini – bronchoalveolar lavage fluid (mBALF) was optimised and several small molecular weight metabolites were identified which are indicator of lung pathology. NMR spectroscopy of human serum samples helps in identifying the metabolites associated with ARDS severity. Further sub classifying the progression, outcome and the metabolites contributing to pulmonary and non-pulmonary causes of ARDS, mBALF and serum samples were being used in larger sample size for which initial model was tested with respect to control showing good separation and accuracy. The sensitivity and specificity of individual serum metabolites and mBALF metabolites as resultant serum and mBALF endotypes were used further to determine their clinical predictability when combined with clinical APACHE and SOFA score. The accuracy increased to AUROC 1 indicating the clinical relevance of the above determined metabolic endotypes. Pathway analysis of serum endotype and mBALF endotype predictive of mortality gave important metabolic pathway symbolic of ARDS correlated changes in metabolism.