Consequences of switching from a fixed 2:1 ratio of amoxicillin/clavulanate (CLSI) to a fixed concentration of clavulanate (EUCAST) for susceptibility testing of Escherichia coli

Objectives: The CLSI recommends a fixed 2: 1 ratio of co-amoxiclav for broth microdilution susceptibility testing of Enterobacteriaceae, while EUCAST recommends a fixed 2 mg/L clavulanate concentration. The aims of this study were: (i) to determine the influence of a switch from CLSI to EUCAST methodology on Escherichia coli susceptibility rates; (ii) to compare susceptibility results obtained using EUCAST-compliant microdilution with those from disc diffusion and the Etest; and (iii) to evaluate the clinical outcome of patients with E. coli sepsis treated with co-amoxiclav in relation to the susceptibility results obtained using either method. Methods: Resistance rates were determined in three laboratories that switched from CLSI to EUCAST cards with the Phoenix system (Becton Dickinson) as well as in 17 laboratories that continued to use CLSI cards with the VITEK 2 system (bioMerieux). In one laboratory, isolates were simultaneously tested by both the Phoenix system and either disc diffusion (n = 471) or the Etest (n = 113). Medical and laboratory records were reviewed for E. coli sepsis patients treated with co-amoxiclav monotherapy. Results: Only laboratories that switched methodology showed an increase in resistance rates - from 19% in 2010 to 31% in 2011 (P<0.0001). All isolates that tested susceptible by microdilution were also susceptible by disc diffusion or the Etest, but of 326 isolates that tested resistant by microdilution, 43% and 59% tested susceptible by disc diffusion and the Etest, respectively. Among the 89 patients included there was a better correlation between clinical response and measured MICs using the Phoenix system than the Etest. Conclusions: EUCAST methodology resulted in higher co-amoxiclav E. coli resistance rates than CLSI methodology, but correlated better with clinical outcome. EUCAST-compliant microdilution and disc diffusion provided discrepant results.