ARDS Metabolic Fingerprints: Characterization, Benchmarking and Potential Mechanistic Interpretation

Aims: In this study we aimed to identify ARDS metabolic fingerprints in selected patient cohorts, compare the metabolic profiles of direct vs indirect ARDS and hypoinflammatory vs hyperinflammatory ARDS. Hypothesis: We hypothesize that the biological and inflammatory processes in ARDS would manifest as unique metabolomic fingerprints which set ARDS apart from other ICU conditions, help examine ARDS subphenotypes and clinical subgroups.Subjects: 108 ARDS patients and 27 ICU ventilated controls were analyzed. Samples were randomly divided into 2/3 training and 1/3 test sets. Methods: Samples were analyzed using proton nuclear magnetic resonance spectroscopy (1H-NMR) and gas chromatography mass spectrometry (GC-MS). 12 proteins/cytokines were also measured. Orthogonal partial least squares discriminant analysis (OPLS-DA) was utilized to select the most differentiating ARDS metabolites and protein/cytokines. Predictive performance of OPLS-DA models was measured in the test set. Temporal changes of metabolites were examined as patients progressed through ARDS until clinical recovery. Metabolic profiles of direct vs indirect ARDS subgroups and hypoinflammatory vs hyperinflammatory ARDS subgroups were compared.Results: Serum metabolomics and proteins/cytokines have similar AUROC when distinguishing ARDS from ICU controls. Pathway analysis of ARDS differentiating metabolites identified a dominant involvement of serine-glycine metabolism. In longitudinal tracking, the identified pathway metabolites generally exhibit correction by 7-14 days, coinciding with clinical improvement. ARDS subphenotypes and clinical subgroups are metabolically distinct.Limitations: Our identified metabolic fingerprints are not ARDS diagnostic biomarkers. Further research is required to ascertain generalizability.Conclusions: ARDS patients are metabolically different from ICU controls. ARDS subphenotypes and clinical subgroups are metabolically distinct.