THE GENOMIC POPULATION STUDY OF BLOODSTREAM ASSOCIATED ESCHERICHIA COLI IN 2020 IN SOUTHWEST, UK

Abstract

Intro: Bloodstream infections (BSIs), predominantly caused by Escherichia coli, are an increasing clinical challenge. Rapid empiric antimicrobial therapy is vital for management for BSIs, therefore understanding the genotypic basis of resistance can contribute to improving empiric choice. Methods: This study investigates the molecular epidemiology of 669 deduplicated, sequential E. coli isolated from BSI by a regional diagnostic laboratory serving a population of 1.5 million people between January and December 2020. E. coli were sequenced on the Illumina-HiSeq-2500. Sequencing data were interrogated for AMR determinants and phylogenetic relationships. Using a combination of antimicrobial susceptibility and sequencing data, the concordance of antimicrobial resistance prediction based on detection of resistant determinants and resistance was evaluated. Findings: ST131 was the predominantly identified sequence type (20%) harbouring resistance genes associated with ß-lactams, aminoglycosides and sulphonamides. Genotypic-phenotypic concordance across the antimicrobial panel was lowest and highest for amoxicillin-clavulanate and cefotaxime respectively. Phylogenetic relationships showed resistance was dispersed and intermixed with isolates from different phylogroups. Phylogroups represented were: A (5%), B1 (6%), B2 (66%) and D (16%). χ2 analysis of phylogroup B2 versus A and B1 revealed there was a strong association (p<0.0001) between carriage of “farm-animal specific resistance genes” – those encoding resistance to streptomycin, neomycin and florfenicol, used in the UK only to treat farmanimals. Discussion: Disconcordance was observed with amoxicillin-clavulanate and ciprofloxacin. Amoxicillin-clavulanate resistance often involves changes in gene expression and ciprofloxicin resistance involves point mutations, both of which are not factored with bioinformatics tools that only consider presence of mobile resistance genes. χ2 analysis suggests that BSI isolates from phylogroups A and B1 have a farm-animal origin. Conclusion: Overall, this study reveals that commonly human associated phylogroups B2 and D were most predominant groups in E. coli BSI. WGS-based prediction of AMR still requires further optimisation before implementing this exclusively for surveillance studies.