Antimicrobial stewardship (AMS) encourages appropriate antimicrobial use to improve outcomes. Initiatives include preventing infection and optimizing antimicrobial use (i.e., right drug, dose, and duration). Infections are associated with significant morbidity and mortality among patients receiving hemodialysis. Although this population would benefit from AMS, initiatives are limited. In response, three studies were conducted.

The first was a study of hemodialysis bloodstream isolates (n = 1024). Pathogen distributions and antimicrobial susceptibilities were used to evaluate empiric antimicrobial combinations. Most isolates (>70%) were gram-positive. Using cefazolin over vancomycin reduced gram-positive coverage by >30%. Gram-negative coverage varied by <1%, with no advantage for broader agents, e.g., meropenem. These findings assist in optimizing empiric antimicrobial selection for bloodstream infections (BSIs) in the hemodialysis setting.

The second was a study of hemodialysis catheter-related exit-site infections (ESIs, n = 113) and BSIs (n = 64). The impact of new infection prevention measures on infection rates and risk factors for infection were evaluated. ESIs dropped by 60% with new measures, then increased slightly over time. BSIs declined by 85% over time, which was associated with new measures and reduced temporary catheter use. ESIs preceded one-third of BSIs. These findings highlight the necessity for ongoing surveillance, reinforcement of prevention measures, and strategies to address ESIs to reduce BSI risk.

The third was a pharmacokinetic study of cefazolin (n = 20) and ceftazidime (n = 18) in patients receiving hemodialysis. The pharmacokinetics, pharmacodynamics, and toxicodynamics were characterized and used to evaluate current dosing practices. For both drugs, the protein binding and volume of distributions were similar to other populations, while the elimination half-lives were >15-fold longer compared to patients with normal kidney function. In the cefazolin cohort, there was greater interindividual variation in half-life attributed to residual kidney function. The analysis identified limitations in using a one-dose-fits-all approach. It highlighted the importance of considering residual kidney function, body weight, the interdialytic period, and pathogen susceptibility to optimize dosing. These findings can be used to develop adaptive dosing approaches that improve the treatment of infections in this high-risk population.

Given the lack of data, this work represents significant progress in facilitating AMS in the hemodialysis setting.