We have shown previously in animal model and in vitro systems that antimicrobial therapy intensity has a profound influence on subpopulations of resistant organisms. Little attention has been paid to the effect of therapy duration on resistant subpopulations. We examined the influence of therapy intensity (area under the concentration/time curve for 24 h:minimum inhibitory concentration [AUC24:MIC] ratio) and therapy duration on resistance emergence using an in vitro model of Staphylococcus aureus infection. AUC24:MIC ratios of>or=100 were necessary to kill a substantial portion of the total population. Importantly, we demonstrated that therapy duration is a critical parameter. As the duration increased beyond 5 days, the intensity needed to suppress the antibiotic-resistant subpopulations increased, even when the initial bacterial kill was>4 log10 (cfu/mL). These findings were prospectively validated in an independent experiment in which exposures were calculated from the results of fitting a large mathematical model to all data simultaneously. All of the prospectively determined predictions were fulfilled in this validation experiment.