Background: Antimicrobial resistance is now a leading cause of death due to infections worldwide. Antimicrobial stewardship can help mitigate antimicrobial resistance by reducing and optimizing broad-spectrum antibiotic use. Urinary tract infections (UTIs) are one of the common indications for use of antibiotics in the hospital setting. Carbapenem use for extended-spectrum β-lactamase producing Enterobacterales (ESBL-PE) UTIs remains controversial with some studies showing benefit while others demonstrating similar efficacy with non-carbapenem beta-lactams or non-beta-lactam agents. National guidelines recommend the utilization of carbapenem sparing strategies for UTIs if there is demonstrated susceptibility to other first line agents such as nitrofurantoin or trimethoprim-sulfamethoxazole (TMP/SMX) for uncomplicated UTIs, and fluoroquinolones or TMP/SMX for complicated UTIs/Pyelonephritis. The purpose of this study is to develop and implement a pharmacist-driven carbapenem de-escalation algorithm for hospitalized patients with UTIs in an effort to reduce further development of antimicrobial resistance.

Methods: This is a biphasic, retrospective chart review of patients who received carbapenems with a diagnosis of UTIs in a 900+ bed tertiary community hospital. Patients were included if they were ≥18 years old and receiving carbapenems with a diagnosis of UTI without bacteremia. Patients were excluded if they were receiving carbapenems for other indications, had a urinary stent at baseline, had abnormal genitourinary anatomy, or had a UTI involving renal calculi. Phase I was the pre-intervention period consisting of a chart review from August 2022 to November 2022 prior to the utilization of the carbapenem de-escalation algorithm. Phase II was the post-intervention period, consisting of a chart review of the intervention period from December 2022 to March 2023. During the intervention period, a pharmacist proactively screened for carbapenem de-escalation eligibility using the algorithm and intervened as appropriate. The primary outcome was the percent of eligible patients de-escalated to non-carbapenem agents before and after implementation of the algorithm. Secondary outcomes included intervention acceptance rate post intervention, median duration of carbapenem therapy pre and post intervention and hospital length of stay pre and post intervention. A 95% confidence interval was calculated for the primary outcome and the Mann-Whitney U test was utilized to assess significance of the secondary outcomes.

Results: In phase I of the study, 60 patients met inclusion criteria and 40 were eligible for de-escalation to non-carbapenem agents. In phase II, 60 patients were included and 37 were eligible for de-escalation via algorithm. Baseline characteristics were similar in both groups, except the post-intervention period had less females 48% versus 35%. Percent de-escalation before and after implementation of the algorithm was 68% versus 90% (95%Cl [0.03 - 0.38]) respectively. In phase II of the study, pharmacists were responsible for 41% of de-escalations and the intervention acceptance rate was 79%. Median duration of carbapenem therapy pre and post intervention was 3 versus 2.4 days (p=0.4237) respectively. Median hospital length of stay pre and post intervention was 5 versus 7 days (p=0.1615) respectively.

Conclusion: A pharmacist-driven carbapenem de-escalation algorithm for patients with ESBL-PE UTIs helped increase carbapenem de-escalation by 22% in the post-intervention period with a 79% intervention acceptance rate.

Publication Date


Presented At:

Florida Residency Conference

Content Type



Available to BHSF community via local IP address or BHSF login.