Predictors of bacterial resistance using in vitro dynamic models: area under the concentration-time curve related to either the minimum inhibitory or mutant prevention antibiotic concentration.

Abstract

To examine the predictive power of the ratios of the 24 h AUC (AUC24) to the mutant prevention concentration (MPC) and the MIC, the selection of ciprofloxacin-resistant mutants of Pseudomonas aeruginosa was studied in an in vitro dynamic model. Four clinical isolates of P. aeruginosa with MPC/MIC ratios from 5.6 to 32 were exposed to twice-daily ciprofloxacin for 3 days over a 100- to 200-fold range of the AUC24/MIC ratio. The emergence of P. aeruginosa resistance to ciprofloxacin was concentration dependent: mutants resistant to 2-16 × MIC were enriched at antibiotic concentrations between the MIC and MPC, but not at concentrations below the MIC or above the MPC. Both AUC24/MIC and AUC24/MPC relationships with the area under the bacterial mutant concentration-time curve (AUBCM) were bell-shaped. These relationships predict highly variable 'anti-mutant' AUC24/MIC and AUC24/MPC ratios: e.g. with mutants resistant to 2 × MIC the ratios ranged from 220 to 1100 and from 7 to 180 h, respectively. Using combined data for the four studied organisms, correlations between AUBCM and AUC24/MIC or AUC24/MPC were established (r(2) = 0.75 and 0.65, respectively). Much stronger correlation was observed between AUC24/MIC and the area between the cut-off level at 10(8) cfu/mL and the time-kill curve (ABBC) as an integral index of the antimicrobial effect of ciprofloxacin on the parental strains (r(2) = 0.93). Findings obtained with ciprofloxacin-exposed P. aeruginosa are consistent with the mutant selection window hypothesis. AUC24/MIC and AUC24/MPC relationships with resistance were more bacterial strain specific than AUC24/MIC relationships with fluoroquinolone-induced killing of susceptible cells.