OBJECTIVES: To compare the bacterial killing and emergence of resistance of intermittent versus prolonged (extended and continuous infusions) infusion dosing regimens of piperacillin/tazobactam against two Escherichia coli clinical isolates in a dynamic hollow-fibre infection model (HFIM). METHODS: Three piperacillin/tazobactam dosing regimens (4/0.5 g 8 hourly as 0.5 and 4 h infusions and 12/1.5 g/24 h continuous infusion) against a ceftriaxone-susceptible, non-ESBL-producing E. coli 44 (Ec44, MIC 2 mg/L) and six piperacillin/tazobactam dosing regimens (4/0.5 g 8 hourly as 0.5 and 4 h infusions and 12/1.5 g/24 h continuous infusion; 4/0.5 g 6 hourly as 0.5 and 3 h infusions and 16/2 g/24 h continuous infusion) were simulated against a ceftriaxone-resistant, AmpC- and ESBL-producing E. coli 50 (Ec50, MIC 8 mg/L) in a HFIM over 7 days (initial inoculum ∼107 cfu/mL). Total and less-susceptible subpopulations and MICs were determined. RESULTS: All simulated dosing regimens against Ec44 exhibited 4 log10 of bacterial killing over 8 h without regrowth and resistance emergence throughout the experiment. For Ec50, there was the initial bacterial killing of 4 log10 followed by regrowth to 1011 cfu/mL within 24 h against all simulated dosing regimens, and the MICs for resistant subpopulations exceeded 256 mg/L at 72 h. CONCLUSIONS: Our study suggests that, for critically ill patients, conventional intermittent infusion, or prolonged infusions of piperacillin/tazobactam may suppress resistant subpopulations of non-ESBL-producing E. coli clinical isolates. However, intermittent, or prolonged infusions may not suppress the resistant subpopulations of AmpC- and ESBL-producing E. coli clinical isolates. More studies are required to confirm these findings.