The mutant selection window in rabbits infected with Staphylococcus aureus

Background. The mutant selection window hypothesis, originally based on agar plate assays, may lead to new antimicrobial dosing strategies that severely restrict the acquisition of resistance. However, it has not been directly tested in an animal model of infection. Methods. Local infection with Staphylococcus aureus was established in rabbits, and the infected animals were treated orally with various doses of levofloxacin. Changes in levofloxacin concentration, levofloxacin susceptibility, and counts of total and resistant viable bacteria were monitored at the site of infection. Results. S. aureus lost levofloxacin susceptibility when drug concentrations at the site of infection fluctuated between the lower and upper boundaries of the window, defined in vitro as the minimum inhibitory concentration (MIC) 99 and the mutant prevention concentration (MPC), respectively. The upper boundary of the selection window in vivo was estimated as an AUC(24)/MPC value of similar to 25 h, where AUC(24) is the area under the drug concentration time curve in a 24-h interval. The lower boundary was estimated as an AUC(24)/MIC value of similar to 20 h. Conclusions. The mutant selection window exists in vivo, and its boundaries fit well with those determined in vitro. Maintenance of antimicrobial concentrations above the window is expected to suppress the outgrowth of resistant mutant subpopulations.