Combination of Leukocyte Cell Surface Biomarkers Measured by Cytometry, to Differentiate Bacterial From Viral Infections in Emergency Department: a Multicentric Cohort for the Validation of Diagnostic Performances

The characterization of the bacterial or viral etiology of an infectious event is required for both isolation decisions and rationale use of antibiotics. In emergency room (ER), the direct identification of the causal pathogen is rarely available in real-time. Alternative is the identification of the host-response to either a bacterial or viral infection. One of this host-response is the expression of peripheral leukocytes cell surface markers, measured by flow cytometry. Investigators and others have reported the high diagnostic performances of combination of cell surface biomarkers to differentiate bacterial from viral infection. The CYTOBACT study aims to confirm on a 500 patients multicentric cohort (200 having already been collected during another study: SEPTIMET), the best combinations for this diagnostic issue. The study will be conducted in 3 emergency departments of APHP hospitals network in Paris, France. Patients with a suspicion of infection will be proposed to participate. No intervention will be introduced during the routine care in the (ER) which will be let at the discretion of the treating emergency physician. During the routine blood sampling in the ER, an additional 30 ml volume of whole blood will be collected, centrifugated, aliquoted and stored at -80°C for further measurement of the expression of a panel of cell surface markers. The participants will be followed up during their hospitalization (if any) and no longer than 28 days. Clinical data at admission, usual blood tests and all microbiological investigations performed during the hospital stay will be recorded into an electronic case record form (eCRF). Based on all those recorded data (excepted the results of flow cytometry for cell surface biomarkers) 2 independent adjudicators will qualify the infectious episode into bacterial,viral or no infection, and (if any) into infection, sepsis or septic shock (according to Sepsis 3.0 definitions). Using different "machine learning" statistical