Area of technical uncertainty for susceptibility testing of amoxicillin/clavulanate against Escherichia coli: analysis of automated system, Etest and disk diffusion methods compared to the broth microdilution reference

Abstract

ObjectivesThe European Committee on Antimicrobial Susceptibility Testing (EUCAST) recently warned about an area of technical uncertainty (ATU) of amoxicillin/clavulanate (AMX/C) disk susceptibility testing against members of the Enterobacterales. Thus, we aimed to compare the reliability of three routine methods and to evaluate the impact of the ATU. Methods286 Escherichia coli strains (including 159 AMX-resistant strains) were categorized for the two EUCAST AMX/C breakpoints by disk diffusion (Bio-Rad), the Phoenix automated system (Becton Dickinson) and the Etest (AES) compared to the broth microdilution reference method. ResultsBy microdilution, 84.2% of strains were AMX/C-susceptible using the urinary breakpoint (MIC ≤32 mg/L) and 62.2% using the systemic breakpoint (MIC ≤8 mg/L), with 63.6% of MICs between 4 and 16 mg/L. For the systemic breakpoint, category agreement (CA) and very major error (VME) were unacceptable for the Etest (71.7% and 27.3%), disk (73.1% and 23.4% at 19-mm cut-off) and to a lesser extent for the Phoenix system (83.6% and 10.5%). For disks, an unacceptable VME rate was observed for diameters up to 22 mm, probably due to overcharged disks. For the Etest, VMEs were high at 6 mg/L (46/63) and 8 mg/L (22/29). For the urinary breakpoint, CA was more acceptable for disk (88.9%) and Etest (84.3%) but was unevaluable for Phoenix. ConclusionAMX/C susceptibility testing of E. coli for systemic breakpoint was unreliable with the three routine methods, explained mainly by the high prevalence (~60%) of strains with microdilution MICs around the breakpoint (8 mg/L). Our data confirmed the EUCAST 19–20-mm ATU for disk and suggest introducing ATU for Etest MIC values of 6 and 8 mg/L.