Evert den Drijver

86 Chapter 5 Table 3. Specificity (with 95% confidence intervals) of different agars Specificity of agars* ESBL-positive isolates AmpC/ESBL-negative isolates Cefotaxime 0/40 0.00 (0.00-0.08) 38/39 0.97 (0.86-1.00) Cefotaxime + Cefoxitin 23/40 0.58 (0.42-0.72) 38/39 0.97 (0.86-1.00) Ceftazidime 0/40 0.00 (0.00-0.08) 39/39 1.00 (0.92-1.00 Ceftazidime + Cefoxitin 25/40 0.63 (0.47-0.76) 39/39 1.00 (0.92-1.00) *Specificity based on no growth on agar (true negative) divided through all isolates of group (true negative + false positive). The prevalence of ESBL in the isolates tested was 25.16%, due to the artificial nature of the isolates chosen. Figures 1 and 2 show the relation between specificity of the cefotaxime 1mg/L + cefoxitin 8mg/L and of the ceftazidime 1mg/L + cefoxitin 8mg/L agar plotted versus ESBL prevalence. An increase in ESBL prevalence in the isolate collection would decrease specificity in a nonlinear function. According to our model a low ESBL prevalence e.g. 5% would lead to a specificity of the cefotaxime 1mg/L + cefoxitin 8mg/L agar of 92% and of the ceftazidime 1mg/L + cefoxitin 8mg/L agar of 95%. A high prevalence of e.g. 50% would decrease specificity of the cefotaxime 1mg/L + cefoxitin 8mg/L agar to 67% and of the ceftazidime 1mg/L + cefoxitin 8mg/L agar to 72%.

RkJQdWJsaXNoZXIy MTk4NDMw