Abstract
Early differential diagnosis of bloodstream infections (BSIs) caused by different sources and species of bacteria in hospitalized patients is crucial for the timely targeted interventions including appropriate use of antibiotics. The aim of this study was to identify 9 biomarkers for the early differentiation of gram-negative-bloodstream infection (GN-BSI), gram-positive (GP)-BSI, and fungal-BSI.A prospective study was conducted for a total of 390 inpatients who underwent blood culture in the Chinese PLA General Hospital from September 2015 to March 2018. Patients with positive culture of a single pathogen were divided into GN-BSI, GP-BSI, and Fungal-BSI groups, and a culture-negative disease control group was also established. The serum levels of macrophage inflammatory protein 1β (MIP-1β), tumor necrosis factor α (TNF-α), interleukin (IL)-3, interferon (IFN)-γ, IL-17A, IL-4, IL-12p70, and P-selectin were detected and the NLR was calculated from routine blood test. Receiver-operating characteristic analysis was used to determine the efficacy of various indicators in the differential diagnosis of BSIs. Prediction and validation experiments on clinical patient samples (263 cases) were also performed.The level of IL-3 in the GP-BSI group was significantly higher than those in the other 3 groups. The level of IFN-γ in the fungal-BSI group was significantly higher than those in the other 3 groups. NLR, MIP-1β, TNF-α, IL-17A, and IL3 exhibited some efficacy when distinguishing between GN-BSI and GP-BSI and NLR had the largest area under curve (AUC) (0.728), followed by MIP-1β with an AUC of 0.679. IFN-γ and IL-3 exhibited some value in differential diagnosis between GN-BSI and Fungal-BSI. IL-3, MIP-1β, TNF-α, IFN-γ, NLR, IL-17A, and IL-4 exhibited some value in distinguishing fungal-BSI and GP-BSI, with IL-3 had the largest AUC (0.722), followed by MIP-1β with an AUC of 0.703.NLR and MIP-1β may be valuable in differentiating GN-BSI from GP-BSI in hospitalized patients. IFN-γ and IL-3 may be helpful in differential diagnosis GN-BSI and fungal-BSI. IL-3 and MIP-1β exhibited some diagnostic efficacy in distinguishing fungal-BSI and GP-BSI. Additionally, IL-3 with high serum level may be a marker for GP-BSI and IFN-γ with high serum level may be a valuable marker for the prediction of Fungal-BSI. The utility of these biomarkers to predict BSIs owing to different pathogens in hospitalized patients needs to be assessed in further studies.