Discussion

This study evaluated treatment strategies and outcomes for CRAB and S. maltophilia before and after implementation of institutional treatment guidance. This study demonstrated a significant change in the way infections caused by CRAB and S. maltophilia were treated following the publication of a treatment algorithm on our internal antimicrobial stewardship website and after education to the ID Division. A previous study that evaluated guideline implementation for the treatment of skin and soft tissue infections suggested that implementing an evidence-based guideline may improve prescribing of empiric antimicrobials. ^{Reference Ryan, Caverzagie, Ross and Chandorkar9}

A single-center, retrospective study conducted by Pant et al. assessed the adherence to institutional treatment guidelines for several infection types including pneumonia, urinary tract infections, and skin and soft tissue infections. ^{Reference Pant, Corwin and Adhikari10} Their study showed an adherence rate of 37.9% after publication of treatment guidelines, which highlights the importance for institution-level treatment guidelines. ^{Reference Pant, Corwin and Adhikari10} The rate of adherence in the present study after institutional treatment guidelines was 5/8 (62.5%) for CRAB and 38/75 (50.7%) for S. maltophilia. While adherence increased after guideline implementation, there remains significant room for improvement.

As previously mentioned, our treatment guidance document was published on our institution’s antimicrobial stewardship program (ASP) website. In addition to this intervention, repeated educational sessions on local treatment guidelines have also been shown to change empiric antibiotic use. ^{Reference Smiley, Rizzuto and White11} A study conducted by Smiley et al. improved empiric antimicrobial prescribing for intraamniotic infections via electronic clinical decision support tools and repeated education sessions. ^{Reference Smiley, Rizzuto and White11} Another study by Zahlanie et al. integrated a computerized order set for outpatient treatment of pediatric bacterial respiratory infections. ^{Reference Zahlanie, Mang, Lin, Hynan and Prokesch12} In this study, a higher percentage of first-line antibiotic prescribing and shorter durations of treatment were observed after the order set was coupled with educational sessions to providers. ^{Reference Zahlanie, Mang, Lin, Hynan and Prokesch12} None of the previously mentioned strategies were used in the present study. If electronic clinical decision support tools including order sets and repeated educational sessions are combined with local published treatment guidance, this could possibly further improve targeted therapy for these difficult-to-treat infections.

For notable changes in antimicrobial use, in CRAB there was a decrease in utilization of both polymyxins and aminoglycosides in the post-intervention group. This finding is in concordance with the IDSA document for the treatment of MDR gram-negative infections with sulbactam being a preferred agent for CRAB. ^{Reference Tamma, Aitken, Bonomo, Mathers, van Duin and Clancy4} It is also worth mentioning that even though sulbactam-durlobactam was approved during the post-intervention period no patients were treated with this agent. This is likely due to our institution not adding this agent to formulary until February 2024, and providers being unfamiliar with this agent. We also saw more patients treated with SXT and less with levofloxacin for S. maltophilia. The decrease in levofloxacin use for S. maltophilia is noteworthy, considering that PK/PD data has shown variable activity in isolates with an MIC ≥ 2 and the possibility of isolates developing resistance after levofloxacin exposure. ^{Reference Nys, Cherabuddi, Venugopalan and Klinker13,Reference Fratoni, Nicolau and Kuti14}

Another significant change that was observed post-intervention was the decrease of ceftazidime monotherapy for patients with S. maltophilia. S. maltophilia ceftazidime breakpoints were not removed by CLSI until February 2024 due to intrinsic L1 and L2 β-lactamases expressed by these isolates, which have been shown to hydrolyze ceftazidime.^15,16 Clinical success reported for ceftazidime monotherapy is as low as 66%. ^{Reference Falagas, Valkimadi, Huang, Matthaiou and Hsueh17} No differences in clinical outcomes were found in our study, but we anticipate that less ceftazidime and levofloxacin monotherapy may lead to improved outcomes over time. We also acknowledge that even though cefiderocol was included in this treatment algorithm, no patients were treated with this agent for S. maltophilia. Cefiderocol has an additional restriction at our institution for approval by antimicrobial stewardship program approval only. It also likely came down to provider or pharmacist preference of using ceftazidime/avibactam + aztreonam over cefiderocol as a metallo-β-lactamase active β-lactam regimen at our institution, and not every patient with an infection caused by S. maltophilia being included in our study.

Secondary outcomes for the present study were exploratory and none of them were found to be statistically significant despite an increased number of patients treated with guideline directed antimicrobial therapy for CRAB and S. maltophilia. All-cause 30-d mortality and clinical efficacy were similar pre- and post-intervention for CRAB, but slightly higher in the post-intervention group for S. maltophilia. A possible explanation of this difference is likely due to the random sampling of patients for the S. maltophilia cohorts. Microbiologic recurrence was relatively similar pre- and post-intervention for both organisms which is as expected since both organisms are colonizers. It is noteworthy that no patients in the CRAB post-intervention group experienced an antimicrobial related adverse effect while there were three in the pre-intervention group which were all attributed to either amikacin or colistimethate.

There were several limitations to this study. First, given the retrospective nature of this study, there could have been a selection bias toward certain regimens based on perceived severity of infection or belief that an isolate was more or less likely to represent a pathogen. In addition, we only included a randomly selected sample of patients in the S. maltophilia cohort and may have missed treatment or outcome trends only evident with a greater sample size. For secondary outcomes, limitations in interpreting these include small sample size and retrospective methodology. Other limitations of this study include possible confounding factors that we did not include in our analyses or that were not documented in the medical record. Specifically, our microbiology lab stopped reporting ceftazidime susceptibilities for S. maltophilia isolates in March 2024 after CLSI removed this breakpoint. Another possible confounding factor is the update of minocycline breakpoints which were lowered in 2023 for S. maltophilia from ≤4 µg/mL to ≤1 µg/mL. ¹⁵ This update was based on data showing doses of minocycline 200 mg q12h have a >90% probability of achieving PK/PD targets in neutropenic models for organisms with MICs of 1 of µg/mL. ^{Reference Fratoni, Nicolau and Kuti18} Both interventions could have resulted in reduced ceftazidime use and higher doses of minocycline for S. maltophilia.

In summary, after implementing a local treatment guideline based on IDSA guidance for CRAB and S. maltophilia on our institution’s ASP website, we saw an increase in guideline directed antimicrobial therapy post-intervention for both of these organisms. These findings highlight an opportunity for ASP interventions to improve antibiotic selection according to published guidelines and local susceptibility patterns. Additional studies are needed to define the impact of this treatment guidance on clinical outcomes. Future work should also include the intervention of repeated educational sessions combined with electronic order sets in addition to institutional treatment algorithms to further improve treatment guidance adherence.