Discussion

The ASP efforts at our health system appear to be successful at reducing inpatient fluoroquinolone use and antimicrobial resistance among P. aeruginosa and E. coli isolates.

The results of this study add to the growing body of evidence that demonstrates the importance of persistent and well-resourced hospital-based ASPs. Several meta-analyses have summarized favorable outcomes for individual stewardship interventions as well as full programs on antibiotic use and resistance, clinical outcomes, adverse events, C. difficile infections, and costs. ^{Reference Baur, Primrose Gladstone and Burkert15–Reference Schuts, Boyd, Muller, Mouton and Prins19} For these reasons, the ASP at our health system is valued and has expanded to provide comprehensive services across the health system. The desirable trends in fluoroquinolone use and susceptibility reported here provide further justification for our program.

We detected a 74% decrease in fluoroquinolone use over 5 years and an absolute increase of 30% and 10% in levofloxacin-susceptible P. aeruginosa and E. coli isolates, respectively, over 10 years. ASP implementation is typically accompanied by an ∼20% decrease in total antimicrobial use, which is consistent with studies reporting on fluoroquinolone use specifically. ^{Reference Karanika, Paudel, Grigoras, Kalbasi and Mylonakis17,Reference Chin, Green and McKamey20,Reference Jones, Onwubiko and Kubes21} Resistance reduction rates vary widely between studies. ^{Reference Boel, Andreasen and Jarløv22–Reference Wu, Liu, Lin, Hsueh and Lee25} We detected a greater reduction in use and resistance than has typically been reported. Although it is difficult to compare results between studies with heterogeneous methods, one possible explanation may be the longer observation period in our study compared with shorter 1–2-year intervention periods in most reports. In fact, a greater magnitude of resistance reduction is observed when fluoroquinolone use and resistance is assessed over a longer period. ^{Reference Cook and Gooch23,Reference Lewis, Fang, Gooch and Cook26} Interestingly, we saw a numerically greater improvement in P. aeruginosa susceptibility than in E. coli. These results are consistent with other studies that describe either no impact on E. coli susceptibility in contrast to P. aeruginosa ^{Reference Hecker, Son and Murphy24,Reference Falagas, Bliziotis and Michalopoulos27} or only marginal improvement. ^{Reference Boel, Andreasen and Jarløv22,Reference Wu, Liu, Lin, Hsueh and Lee25,Reference Gottesman, Carmeli, Shitrit and Chowers28}

Our results provide support for non–restriction-focused ASP services. Most current evidence for ASP impact on gram-negative fluoroquinolone resistance includes formulary restriction as the main intervention. ^{Reference Pitiriga, Vrioni, Saroglou and Tsakris13} However, 2 studies have described interventions that included prospective audit and feedback without a restriction component. ^{Reference Hecker, Son and Murphy24,Reference Lafaurie, Porcher, Donay, Touratier and Molina29} The interventions were successful in both studies, demonstrating a significant decrease in fluoroquinolone use and absolute increase of 9%–16% in fluoroquinolone-susceptible P. aeruginosa isolates over an ∼5-year period, which is in line with our results. Our study is unique because it not only provides fluoroquinolone susceptibility trends over 10 years but also has a longer stewardship intervention period than the 2 studies.

Another strength of our study is the utilization of order-set optimization and prospective audit and feedback as the core stewardship intervention, which can be adopted in virtually any hospital setting without the need for around-the-clock staff support, as might be required by some formulary restriction models. We might have observed even greater benefits if our system employed formulary restriction in addition to prospective audit and feedback and order-set optimization. When considering longitudinal antimicrobial stewardship activities, ASPs should periodically re-evaluate their approaches and adapt their activities based on needs and resources.

Antimicrobial stewardship activities at our health systems are primarily pharmacist driven, which is similar to other stewardship programs. ^{Reference Barlam, Childs and Zieminski30} Pharmacists, particularly those with infectious diseases expertise, are uniquely qualified to provide robust antimicrobial stewardship and therapeutic optimization. ^{Reference Garau and Bassetti31,Reference Bessesen, Ma, Clegg, Fugit, Pepe and Goetz32} Leveraging pharmacists’ drug expertise appears to have been successful in reducing both antimicrobial use and antimicrobial resistance at our health system. In addition to empowering pharmacists to perform stewardship activities and promoting their expertise within our health system, investment in staffing likely contributes to the successes described. Various recommendations of ratios of antimicrobial stewardship staff to hospital beds exist, but our ratio of 1 full-time infectious diseases pharmacist to 400 beds may allow for more robust prospective audit and feedback and higher impact relative to ASPs with fewer staff members. ^{Reference Pulcini, Morel and Tacconelli33}

This study had several limitations. Fluoroquinolone use data were not available for inpatients before 2016. How use correlated with susceptibility before this time remains uncertain, but other than the stewardship activities described, there were no other clinical interventions, operational changes, or changes in infection prevention practices that could have significantly influenced susceptibility. Based on first-hand knowledge of stewardship activities and prescribing patters prior to 2016, we know the use of fluoroquinolones was substantially higher in 2011 and decreased over time. We could not assess the impact of our program in a quasi-experimental study design due to lack of data prior to the intervention period. We also did not report use and antimicrobial resistance trends for other antibiotic classes, which may have inadvertently increased. However, this is unlikely given that our program did not focus solely on fluoroquinolone use but addressed other broad-spectrum antibiotics in parallel. Susceptibility interpretations for fluoroquinolones were designated based on CLSI breakpoints before the change in 2019. ³⁴ Application of updated breakpoints likely would have resulted in decreased susceptibility rates; however, this would have been a major confounding factor, and for the purposes of this review, utilizing a consistent breakpoint likely more accurately reflects trends over time.

In our health system, fluoroquinolone use substantially decreased and P. aeruginosa and E. coli levofloxacin susceptibility increased after the implementation of an ASP. These results demonstrate the value of stewardship services and highlight the effectiveness of an infectious diseases pharmacist–led ASP.