83 Detection of AmpC beta-lactamases in E. coli using different screening agars the recovery of the isolates is described in the additional file 1. Additional file 2 shows an overview of all strains. Evaluation of AmpC agars MacConkey agar no3 medium was obtained from Oxoid Limited (ThermoFisher, Basingstoke, Hampshire, United Kingdom). Both cefotaxime sodium salt and ceftazidime sodium salt were obtained from VWR International (Radnor, Pennsylvania, USA). Cefoxitin sodium salt was obtained from Sigma-Aldrich (Poole, Dorset, UK). MacConkey agar no3 was used as the basal medium to which different dilutions of antibiotics were added. Four different concentration combination were evaluated; 1. cefotaxime 1mg/L agar, 2. ceftazidime 1 mg/L agar, 3. cefotaxime 1mg/L with cefoxitin 8mg/L agar, and 4. ceftazidime 1 mg/L with cefoxitin 8mg/L agar. All strains with AmpC or ESBL were cultured in a selective enrichment broth consisting of a TSB containing 0, 25 mg/L cefotaxime and 8mg/L vancomycin (TSBVC, Mediaproducts, Groningen, The Netherlands). For strains without AmpC or ESBL a brain heart infusion broth was used (BHI, Mediaproducts, Groningen, The Netherlands), as using selective enrichment broth might inhibit growth. After overnight aerobic incubation for 18-24 hours (35°-37°), 10 µl of the broth was subcultured on the four different AmpC agars. Agars were incubated aerobically for 18-24 hours (35°-37°). Results were interpreted as growth or no growth. Statistical analysis Data were analyzed with Statistical Package for Social Science software (SPSS; IBM Corp., Armonk, New York, US; version 22). For sensitivity all strains positive for an acquired AmpC gene or alterations in promoter/attenuator AmpC related to hyperproduction were considered true positives. For specificity were all strains negative for an acquired AmpC gene and alterations in promoter/attenuator AmpC related to hyperproduction were considered true negatives. Confidence intervals (95% CI) were calculated according to the adjusted Wald method. The prevalence of ESBL in the panel of isolates was 25.16% (40 out of 159 isolates), due to the artificial nature of the collection set up. To analyse the influence of variations in ESBL prevalence on the specificity of both the cefotaxime 1mg/L + cefoxitin 8mg/L and the ceftazidime 1mg/L + cefoxitin 8mg/L agar, we remodelled our data for an ESBL prevalence range from 0 to 99%. Specificity of the cefotaxime 1mg/L + cefoxitin 8mg/L agar was calculated using following formula: 38 + (23/40 x number of ESBL isolates) / (38 + (23/40 x number of ESBL isolates) + (1 + 17/40 x number of ESBL isolates)). 5
RkJQdWJsaXNoZXIy MTk4NDMw