Objectives: We sought to define regional variations in fluoroquinolone non-susceptibility (FQ-NS) among bloodstream isolates of Escherichia coli across the Veterans Health Administration (VHA) in the United States.
Methods: We analyzed a retrospective cohort of patients managed at 136 VHA hospitals who had a blood culture positive for E.coli between 2003 and 2013. Hospitals were classified based on US Census Divisions, and regional variations in FQ-NS were analyzed.
Results: Twenty-four thousand five hundred twenty-three unique E.coli bloodstream infections (BSIs) were identified between 2003 and 2013. 53.9 % of these were community-acquired, 30.7 % were healthcare-associated, and 15.4 % were hospital-onset BSIs. The proportion of E.coli BSIs with FQ-NS significantly varied across US Census Divisions (p < 0.001). During 2003-2013, the proportion of E.coli BSIs with FQ-NS was highest in the West South-Central Division (32.7 %) and lowest in the Mountain Division (20.0 %). Multivariable analysis showed that there were universal secular trends towards higher FQ-NS rates (p < 0.001) with significant variability of slopes across US Census Divisions (p < 0.001).
Conclusion: There has been a universal increase in FQ-NS among E.coli BSIs within VHA, but the rate of increase has significantly varied across Census Divisions. The reasons for this variability are unclear. These findings reinforce the importance of using local data to develop and update local antibiograms and antibiotic-prescribing guidelines.
Keywords: Antimicrobial resistance; Epidemiology; Escherichia coli; Fluoroquinolones.