Ceftolozane/tazobactam (C/T) has emerged as a potential agent for the treatment of extensively drug-resistant (XDR) Pseudomonas aeruginosa infections. As it is a time-dependent antimicrobial, prolonged infusion may help achieve pharmacokinetic/pharmacodynamic (PK/PD) targets. To compare alternative steady-state concentrations (Css) of C/T in continuous infusion (CI) against three XDR P. aeruginosa ST175 isolates with C/T minimum inhibitory concentration (MIC) values of 2 to 16 mg/L in a hollow-fiber infection model (HFIM). Duplicate 10-day HFIM assays were performed to evaluate Css of C/T in CI: one compared 20 and 45 mg/L against the C/T-susceptible isolate while the other compared 45 and 80 mg/L against the two C/T-non-susceptible isolates. C/T resistance emerged when C/T-susceptible isolate was treated with C/T in CI at a Css of 20 mg/L; which showed a deletion in the gene encoding AmpC β-lactamase. The higher dosing regimen (80 mg/L) showed a slight advantage in effectiveness. The higher dosing regimen has the greatest bactericidal effect, regardless of C/T MIC. Exposure to the suboptimal Css of 20 mg/L led to the emergence of C/T resistance in the susceptible isolate. Antimicrobial regimens should be optimized through C/T levels monitoring and dose adjustments to improve clinical management.

Combining antibiotics is one of the few strategies available to overcome infections caused by multidrug-resistant bacteria. Time-kill curves are usually performed to evaluate antibiotic combinations, but obtaining results is too laborious to be routinely performed in a clinical laboratory.

This study correlates in vivo findings in a patient with an extensively drug-resistant (XDR) P. aeruginosa infection who developed resistance to ceftazidime-avibactam (CAZ-AVI) with in vitro results of a 7-day hollow-fiber infection model (HFIM) testing the same bacterial strain. The patient was critically ill with ventilator-associated pneumonia caused by XDR P. aeruginosa ST175 with CAZ-AVI MIC of 6 mg/L and was treated with CAZ-AVI in continuous infusion at doses adjusted for renal function. Plasma concentrations of CAZ-AVI were analyzed on days 3, 7, and 10. In the HIFM, the efficacy of different steady-state concentrations (Css) of CAZ-AVI (12, 18, 30 and 48 mg/L) was evaluated. In both models, a correlation was observed between the decreasing plasma levels of CAZ-AVI and the emergence of resistance. In the HIFM, a Css of 30 and 48 mg/L (corresponding to 5× and 8× MIC) had a bactericidal effect without selecting resistant mutants, whereas a Css of 12 and 18 mg/L (corresponding to 2× and 3× MIC) failed to prevent the emergence of resistance. CAZ/AVI resistance development was caused by the selection of a single ampC mutation in both patient and HFIM. Until further data are available, strategies to achieve plasma CAZ-AVI levels at least 4× MIC could be of interest, particularly in severe and high-inoculum infections caused by XDR P. aeruginosa with high CAZ-AVI MICs.