How does susceptibility prevalence impact on the performance of disk diffusion susceptibility testing?

Abstract

Antimicrobial disk diffusion susceptibility testing, devoted in a clinical context to predicting whether an antibiotic regimen will be effective, should be evaluated through predictive values. This approach implies that the susceptibility prevalence (frequency of susceptible, intermediate, and resistant isolates) affects the predictive value of a result. We quantified the influence of the susceptibility prevalence variation on the disk diffusion method performance through a modeling approach. Simulations based on a resampling procedure from two distinct minimum inhibitory concentration/diameter data sets were performed. Experimental variability on minimum inhibitory concentration and diameters was taken into account in the simulations. Results show that the susceptibility prevalence impact depends on the antibiotic and may be significant when prevalence variation is high enough. Consequences of these results on zone diameter breakpoint determination policy are discussed. This implies that the following should be done: (i) consider more rigorously the susceptibility prevalence in studies dealing with zone diameter breakpoint determination and performance evaluation, (ii) re-evaluate disk diffusion breakpoint consistency when the weight of prevalence variation is noteworthy, (iii) estimate consequences of a breakpoint international consensus on prediction quality and appropriate patient management.