Metabolomics Study of ARDS Diagnosis, Heterogeneity, and Mortality
Date
2020-08-06
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Background: Acute Respiratory Distress Syndrome (ARDS) represents 10.4% of all Intensive Care Unit (ICU) admissions with an overall mortality rate of 35.3%. Early diagnosis of ARDS is an important requirement for the timely institution of proper supportive care but it is hampered by the unreliable tissue diagnosis, lack of an early diagnostic biomarker and lack of specific therapy. Metabolomics is a systems biology approach to examine changes in metabolites in response to physiological and pathological stimuli. It is more closely related to observed phenotypes and more responsive to real-time pathophysiological changes, thus constitutes an attractive platform for studying ARDS.Hypothesis: We hypothesize that the diverse inflammatory, pathobiological and adaptive processes involved in ARDS would manifest as unique metabolic fingerprint that sets ARDS apart from other ICU conditions. Biomarker identification may help early diagnosis of ARDS, explore its mechanism, differentiate ARDS subclasses and predict ARDS mortality. Methodology: Serum samples from 108 ARDS patients and 27 ICU ventilated age- and sex-matched controls were included in this study. Samples were analyzed by proton nuclear magnetic resonance spectroscopy (1H-NMR) and gas chromatography mass spectrometry (GC-MS). The proteins/cytokines IL-6, IL-8, IL-10, TNF-?, Ang-2, RAGE, vWF, TNF-RI, ICAM-1, PAI-1, SP-D and Protein-C were also measured. Multivariate statistical models that distinguish ARDS from ICU controls (diagnose ARDS), subclassify ARDS and predict ARDS non-survival were built. Results were externally validated in an independent cohort of 188 patients comprising 118 ARDS patients and 70 ICU ventilated controls.Results: Our data indicates that NMR and GC-MS metabolomics are more accurate than proteins/cytokines in differentiating ARDS from ICU controls and ARDS subgroups. Pathway analysis of ARDS versus controls identified a dominant involvement of serine-glycine metabolism. In external validation, ARDS patients were correctly identified in 93% using NMR and 96% using GC-MS. Metabolomics was not able to predict mortality in the ARDS population studied.Conclusion: ARDS subgroups are metabolically distinct. Further research is needed to explore the role of dysfunctional folate metabolism in the development of ARDS. Metabolomics provides a novel technology to not only understand but intervene on the pathophysiology of ARDS.
Description
Keywords
ARDS, metabolomics, biomarkers, statistics, cytokines, endotyping, folate hypothesis, waning of the initial insult effect, machine learning, visualization, Probabilistic universal model approximator, PUMA
Citation
Mohamed Metwaly, S. (2020). Metabolomics Study of ARDS Diagnosis, Heterogeneity, and Mortality (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.