This study was designed to compare cefepime exposures with microbiological outcomes in ESBL and non-ESBL infections and determine the pharmacodynamic profiles associated with successful outcome.Cefepime pharmacodynamic exposures of unbound drug [time above MIC (fT>MIC), minimal concentration over MIC (fC(min)/MIC), and area under the curve over MIC (fAUC/MIC)] for 18 patients with ESBL and non-ESBL infections were determined by using a published population pharmacokinetic model. Classification and regression tree analysis was used to identify pharmacodynamic breakpoints that predicted eradication. A 5000-patient Monte Carlo Simulation was conducted to estimate the probability of target attainment for the goal pharmacodynamic exposures.Eradication was 80% when fT>MIC was 50% compared with 0% when T>MIC was less than 50% (p7.6 and only 33.3% were eradicated when fC(min)/MICMIC and fAUC/MIC>1654 were also predictive of eradication. While conventional dosage regimens of 2g q 12h and q 8h failed to achieve adequate target attainment, 4 g continuous infusion and 2g q 6-8h prolonged infusion could attain more than 90% of target attainment at the MIC of 2 microg/ml for the breakpoint of fCmin/MIC=7.6.Microbiological eradication in patients receiving cefepime was best predicted by fCmin/MIC ratio greater than 7.6 regardless of the presence of an ESBL. Continuous or prolonged infusion regimens provided the greatest probability of attaining this exposure.