Metabolomic Analysis in Sepsis and ARDS

Journal: American Journal of Respiratory and Critical Care Medicine

Publication Date: Jan. 1, 2014

Authors: Martin GS, Martin RT, Uppal K, Jones DP, Esper AM

Cite As: Martin RT, Uppal K, Martin GS, Jones DP, Esper AM. Metabolomic Analysis in Sepsis and ARDS. Am J Respir Crit Care Med 2014;189:A4964.

Studies:

Abstract

Rationale: Metabolomic analysis can be useful in critically ill patients to enhance our ability to diagnose, guide clinical care, and prognosticate. The combination of liquid chromatography with mass spectrometry (LCMS) provides rapid analysis of metabolites from various biological source materials, such as bronchoalveolar lavage (BAL) fluid, with minimal sample preparation. The objective of this study was to evaluate the potential of metabolomics to accurately discriminate sepsis patients according to development of ARDS and mortality. Methods: Prospective cohort of adult ICU patients at Grady Memorial Hospital meeting the ACCP/SCCM definition of severe sepsis or septic shock. Each subject underwent research bronchoscopy with BAL and data were collected on infection, co-morbidities, illness severity, organ dysfunction, development of ARDS (by the American-European Consensus Conference definition) and clinical outcomes. Statistical analysis employed chi-squared analysis; P-values < 0.05 were considered significant. BAL fluid was analyzed on a LCMS metabolomics platform using C18 chromatography and high-resolution mass spectrometry (Thermo Qexactive) and analyzed using LIMMA to identify differentially expressed metabolites (DEMs) after adjusting for false discovery rate (FDR). Classification accuracies were evaluated using 10-fold cross-validation (CV) using support vector machine and network analysis using Spearman correlations to identify metabolites correlated with DEMs. Results: Metabolomic analysis was performed on 79 BAL fluid samples from patients with severe sepsis or septic shock, 42 of whom had ARDS. The mean age of the cohort was 53.9 years (±16), 59 % male, 68% black. Sources of sepsis were: respiratory (68%), blood (12%), GU (11,3%) , other (3.7%), CSF (2,5%), and GI (1%). The overall mortality for the cohort was 48%. There was no mortality difference between ARDS and non-ARDS patients (42.9% vs. 54.1%, p=0.32). FDR analysis between survivors and non-survivors of sepsis revealed 37 DEMs with 85.9% 10-fold CV classification accuracy. Top 500 metabolites from the network analysis were found to be significantly enriched in the glycerophospholipid metabolism, cytochrome P450/ drug metabolism, valine, leucine and isoleucine biosynthesis pathways. Analysis between ARDS and non-ARDS patients showed 35 DEMs with 92.1% CV accuracy, with pathways for phenylalanine, tyrosine, D-arginine and D-ornithine metabolism, and cytochrome P450, significantly enriched. Conclusions: BAL fluid metabolites discriminate between sepsis non-survivors and survivors, and those developing ARDS or not. High resolution metabolomics to identify metabolic markers may be useful for predicting development of ARDS and other clinical outcomes in critically ill patients.