Azithromycin susceptibility testing for Salmonella enterica isolates: discordances in results using MIC gradient strips

Background Azithromycin resistance is emerging in typhoidal Salmonella. Confirmation of azithromycin MIC is the most frequent antibiotic susceptibility request made to the Gastrointestinal Bacteria Reference Unit (GBRU) laboratory in England by local diagnostic laboratories. Objectives (i) Determine concordance between local diagnostic and reference laboratory estimations of azithromycin MIC by gradient strip in Salmonella enterica serovars Typhi and Paratyphi. (ii) Consider causes of variation. Methods Isolates from patients with enteric fever attending a central London hospital between May 2011 and April 2019 were tested for azithromycin susceptibility using gradient strips, according to EUCAST methodology. Matched local diagnostic and reference laboratory estimations of azithromycin and ciprofloxacin (as a comparator) MICs were included; concordance in estimations was examined. Results Local diagnostic laboratory readings overestimated azithromycin MIC values compared with the reference laboratory, resulting in poor concordance in susceptibility/resistance attribution (concordant susceptibility interpretation in 8/19, κ = 0). In contrast, ciprofloxacin MIC estimation demonstrated superior concordance (concordant susceptibility interpretation in 16/17, κ = 0.85). None of the isolates was resistant to azithromycin at the reference laboratory and no known genes associated with azithromycin resistance were detected in any isolate using WGS. Conclusions Overestimation of azithromycin resistance is likely to be due to difficulty in interpreting the point of intersection of the ‘trailing edge’ with the gradient strip, used to determine MIC. We advise local diagnostic laboratories to review their experience and consider adopting a ‘second reader’ system to mitigate this.