Challenges to accurate susceptibility testing and interpretation of quinolone resistance in Enterobacteriaceae: results of a Spanish multicentre study

Abstract

Abstract Objectives The objective of this study was to evaluate the proficiency of Spanish laboratories with respect to accurate susceptibility testing and the detection and interpretation of quinolone resistance phenotypes in Enterobacteriaceae. Methods Thirteen strains of Enterobacteriaceae were sent to 62 participating centres throughout Spain; strains harboured GyrA/ParC modifications, reduced permeability and/or plasmid-mediated quinolone resistance genes. The centres were requested to evaluate nalidixic acid and five quinolones, provide raw/interpreted clinical categories and to detect/infer resistance mechanisms. Consensus results from reference centres were used to assign minor, major and very major errors (mEs, MEs and VMEs, respectively). Results Susceptibility testing in the participating centres was frequently performed using the MicroScan WalkAway, Vitek 2 and Wider systems (48%, 30% and 8%, respectively). CLSI/EUCAST breakpoints were used in 71%/29% of the determinations. The percentage of VMEs for all quinolones was well below 2%. Only ofloxacin and moxifloxacin showed higher values for raw VMEs (6.6%), which decreased to 0% and 2.9%, respectively, in the interpreted VMEs. These errors were particularly associated with the CC-03 strain [qnrS2 + aac(6′)-Ib-cr]. For MEs, percentages were always <10%, except in the case of ofloxacin and nalidixic acid. There was a significantly higher percentage of all types of errors for strains whose MICs were at the border of clinical breakpoints. Conclusions The use of different breakpoints and methods, the complexity of mutation-driven and transferable resistance mechanisms and the absence of specific tests for detecting low-level resistance lead to high variability and represent a challenge to accuracy in susceptibility testing, particularly in strains with MICs on the border of clinical breakpoints.