Tjitske van Engelen

47 Biomarkers in sepsis population when developing or validating these host gene expression signatures. Along the same lines, direct comparison of different gene-signatures should be done with caution, especially when evaluated outside the context in which they were discovered and validated [45, 59]. Proteomics and metabolomics The developments of mass spectrometry have paved the way to study changes in proteomics and metabolomics during sepsis [60]. Initial enthusiasm about the application of proteomics to discover protein biomarkers was reduced due to technical difficulties and low reproducibility. Nonetheless, several studies have been reported using proteomics to discover biomarkers for infection diagnosis and sepsis prognosis. A recent proteomics study screened 600 proteins in 765 patients presenting to the ED with fever and identified a three-protein signature that could discriminate bacterial, viral and noninfectious disease accurately [61]. The plasma concentrations of tumor necrosis factor-related apoptosis-inducing ligand 1, chemokine (C-X-C motif) ligand 10 (CXCL10) and CRP distinguished bacterial from viral, and infectious (bacterial and viral) from noninfectious disease with an AUC of 0.94 (95% CI 0.92-0.96). The threeprotein signature was strongest in differentiating bacterial from viral etiologies in lower respiratory tract infections and fever without known source [61]. Notably, while this study suggests that proteomics can be useful for the discovery of biomarkers of infection, this population is different from patients with sepsis, and similar studies in sepsis are warranted to evaluate protein signatures as biomarkers in this context. A limitation of conventional proteomic techniques, such as 2D gel electrophoresis, is that the abundant plasma proteins, such as albumin, limit identification and measurement of changes in low-abundance plasma proteins (or peptides). A pilot study conducted in 20 patients with sepsis sought to circumvent this technical issue by making use of the fact that albumin is not glycosylated, and utilized a plasma glycoproteomic evaluation for analysis of low-abundance plasma proteins [62]. N-linked plasma glycopeptides were quantified by solid-phase extraction coupled with mass spectrometry, after which protein differences between sepsis survivors and non-survivors were validated by immunoblotting [62]. A total of 501 glycopeptides corresponding to 234 proteins were identified, of which 66 glycopeptides (54 proteins) were unique to survivors, 60 glycopeptides (43 proteins) were unique to non-survivors, and 375 glycopeptides (137 proteins) were common between groups. Non-survivors exhibited elevated levels of total kininogen and reduced levels of total cathepsin-L1, vascular cell adhesion molecule, periostin, neutrophil gelatinase-associated lipocalin, and glycosylated clusterin [62]. Recent studies have focused on plasma metabolomic profiles as potential predictive markers for ICU mortality in adult patients [63-68]. An investigation that sought to identify metabolic biomarkers for differentiation of sepsis and noninfectious acute disease analyzed 186 metabolites comprising six analyte classes (acylcarnitines, amino acids, biogenic amines, glycerophospholipids, sphingolipids and carbohydrates) to reveal two markers (acylcarnitine C10:1 and glycerophospholipid PCaaC32:0) that had higher concentrations in sepsis, suggesting that they may be helpful for differentiation of infectious from noninfectious systemic inflammation [63]. Carnitine, the product of acylcarnite, is used to transport fatty acids into the mitochondria whereas glycerophosphatidylcholines are major constituents of cell membranes and 3

RkJQdWJsaXNoZXIy MTk4NDMw