Combined Bacterial Identification and Antimicrobial Susceptibility Testing Directly from Whole Blood

Abstract

BackgroundTimely identification of a causative pathogen and its antimicrobial susceptibility profile is important for effective therapy. This is especially true in the case of bloodstream infections caused by the ESKAPE (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter cloacae) pathogens where inappropriate antibiotic prescription often leads to higher mortality and increased selection for multi-drug resistant strains. However, current standard protocols for pathogen identification (ID) and antimicrobial susceptibility testing (AST) take days to complete and despite the advancement of molecular diagnostics, none can concurrently provide reliable ID and AST information.MethodsWe developed a method of direct ID and AST of ESKAPE pathogens using real-time PCR-HRM (high resolution melt) as the end-point analysis coupled with whole blood sample preparation. Our assay utilizes blood cell lysis, removal of background human DNA and protein, pathogen enrichment, antibiotic exposure, and broad-range PCR-HRM analysis targeting bacterial internal transcribed spacer region to determine ID and AST in less than 10 hours. We then assessed antimicrobial susceptibility/resistance by calculating the difference in threshold cycle compared with untreated sample (ΔCt), and determined species identity by its melt curve using an automated computer algorithm against a reference curve database of 89 bacterial species.ResultsOur assay was able to reach a limit of detection of 10 CFU/mL for all tested ESKAPE organisms except for E. faecium (100 CFU/mL). Using a final concentration of 100 CFU/mL and a calculated ΔCt cutoff value of 1.5, we observed significant ΔCt changes with antibiotic treatment and established minimum inhibitory concentrations (MICs) for each strain. Comparison of the observed values to reference MICs showed overall similar results. Furthermore, our automated machine-learning based computer algorithm was able to correctly identify each organism based on its melt curve.ConclusionOur results suggest that reliable rapid bacterial ID and AST information can be simultaneously obtained directly from whole blood using our combined ID-AST assay.Disclosures All authors: No reported disclosures.