Objectives: Rapid diagnostic tests for bacteremia are important for early treatment to improve clinical outcome. We sought to identify plasma biomarkers that can identify patients with bacteremia using an untargeted global metabolomic analysis.
Methods: Plasma metabolomic profiles were analyzed for 145 adult patients with (cases) and without (controls) bacteremia using ultra-high-performance liquid chromatography/quadrupole-time-of-flight mass spectrometry (UHPLC-Q-TOF-MS). All metabolites were compared between cases and controls using a 2-tier filtering approach, and each metabolite underwent receiver operating characteristic (ROC) curve analysis. Individual metabolites that distinguish between cases and controls were characterized. Subgroup analysis was performed to identify metabolites with prognostic significance.
Results: After 2-tier filtering, 128 molecular features were identified to be potential biomarkers that could distinguish cases from controls. Five metabolites had an area under the ROC curve (AUC) of >0.8 in ROC curve analysis, including a sphingolipid, an acylcarnitine, a fatty acid ester, and 2 glycerophosphocholines. These metabolites could distinguish cases from controls in the unsupervised hierarchical clustering analysis. Subgroup analysis of bacteremic patients showed that the level of trans-2,3,4-trimethoxycinnamate was lower in fatal than non-fatal cases.
Conclusions: Plasma lipid mediators of inflammation can distinguish bacteremia cases from non-bacteremia controls. These biomarkers may be used as targets for rapid test in clinical practice.
Keywords: Bacteremia; Biomarker; Lipid; Metabolomic; Organic acid; Plasma.
Copyright © 2015 The British Infection Association. Published by Elsevier Ltd. All rights reserved.