Introduction
Vancomycin (VCM) is commonly used for treating gram-positive bacterial infections, including methicillin-resistant Staphylococcus aureus. High doses of VCM have been associated with acute kidney injury (AKI), primarily through mechanisms involving oxidative stress and mitochondrial damage. Reference Dieterich, Puey, Lyn, Swezey, Furimsky and Fairchild1 Dosing strategies based on the area under the concentration–time curve (AUC) and trough serum concentrations (Cmin) have been developed to help balance efficacy and toxicity. Reference Filippone, Kraft and Farber2,Reference Hirano, Sakamoto, Kitazawa, Yamamoto and Tachibana3
Previous therapeutic drug monitoring guidelines used Cmin as a surrogate marker for AUC; Reference Matsumoto, Takesue and Ohmagari4 however, recent studies have highlighted discrepancies between these 2 parameters. Reference Patel, Pai, Rodvold, Lomaestro, Drusano and Lodise5–Reference Pai, Neely, Rodvold and Lodise7 As a result, AUC-based dosing has been recommended in the updated guidelines released in the United States in May 2020 and in Japan in February 2022. Reference Rybak, Le and Lodise8–Reference Moise-Broder, Forrest, Birmingham and Schentag10
Evidence suggests that appropriate loading doses do not increase the risk of renal impairment and can improve treatment success, particularly in patients with an estimated glomerular filtration rate of 30–80 mL/min/1.73m2. Reference Asakawa, Shimizu, Saito, Kamada and Soehara11 Nevertheless, the administration rate of the loading dose remains suboptimal in clinical practice, limiting the achievement of target AUC levels in the early phase of therapy. Reference Reuter, Stocker and Alffenaar12 For instance, a loading dose was implemented in less than one-third of cases in an Australian public hospital. Reference Carland, Stocker and Baysari13 Furthermore, underdosing was reported in approximately 70% of cases at a Belgian tertiary care center. Reference Van Der Heggen, Buyle and Claus14
Pharmacist-led interventions have recently gained attention as a strategy for supporting appropriate antimicrobial use by assisting physicians in designing a VCM dosing support system. Reference Thaler and Sunstein15,Reference Tai, Motoki and Yamaguchi16 However, there is limited evidence on the impact of such support on clinical outcomes, including loading dose administration, AUC0–24, AKI incidence, and mortality. This study aimed to evaluate the effects of a pharmacist-led support system for initial VCM dosing introduced at Kagawa University Hospital, focusing on its impact on loading dose administration, AUC0–24, AKI incidence, and all-cause mortality.
Methods
Facility overview and description of the support system
At our institution, VCM serum concentrations are measured 24/7 using the ARCHITECT i2000 SR immunoassay system. Pharmacists estimate Cmin and AUC values using EasyTDM Ver. 3-6-1-1 M (Kagawa Society of Hospital Pharmacists, Japan). Reference Rodvold, Pryka, Garrison and Rotschafer17 Ward pharmacists performed blood concentration evaluations on weekdays; on-duty pharmacists during nights and holidays.
In line with the 2022 guideline revision, the institution transitioned from Cmin-based to AUC-based dosing. These changes, implemented in February 2022, included pharmacist intervention when a loading dose was not prescribed or provided, with direct feedback to the prescribing physician. The Antimicrobial Stewardship Team pharmacists also reviewed prescriptions from the day of ordering to the next business day. Pharmacists, including but not limited to Antimicrobial Stewardship Team members, provided AUC-based dosing recommendations—covering both loading and maintenance doses—based on pharmacokinetic calculations or the renal function–adjusted dosing nomogram described in the Japanese Society of Therapeutic Drug Monitoring guidelines. Reference Matsumoto, Oda and Shoji9 Pharmacists involved in dosing receive regular training; to support consistent administration, system updates are shared annually with the entire hospital staff.
Study population and study period
This retrospective study included patients who received VCM over 179 weeks from January 1, 2021, to May 31, 2024. A total of 587 cases of intravenous VCM prescription were identified during the study period. Of these, 117, 58, and 18 cases were excluded due to discontinuing therapy was within 2 days of initiation, ongoing renal replacement therapy (hemodialysis or peritoneal dialysis), and presence of AKI prior to VCM initiation, respectively. In addition, 32, 4, and 32 cases were excluded because they were younger than 15 years, VCM was initiated at another facility, and administration was not via the intravenous route, respectively. After these exclusions, 326 patients met the inclusion criteria and were evaluated. The study period was divided into a preintervention phase (57 wk: January 1, 2021–February 3, 2022) and a postintervention phase (122 wk: February 4, 2022–May 31, 2024), considering the date of introducing the system. The exclusion criteria included patients who completed VCM within 2 days after initiation; younger than 15 years; on continuous renal replacement therapy, hemodialysis, or continuous ambulatory peritoneal dialysis; who did not receive intravenous administration; with AKI before VCM administration; and whose VCM was initiated at another facility.
Comparison of patient background characteristics before and after introducing the support system
We compared the background characteristics of patients receiving VCM before and after the introduction of the support system. The following parameters were evaluated on the day of initiating VCM therapy: age, sex, body weight, serum creatinine, body mass index, blood urea nitrogen/creatinine ratio, disease severity assessed by the Pitt Bacteremia Score, distribution of causative microorganisms (including methicillin-resistant Staphylococcus aureus), and the concomitant use of nephrotoxic agents. The concomitant use of nephrotoxic agents—defined as the use of any of the following: diuretics, non-steroidal anti-inflammatory drugs, aminoglycoside antibiotics, amphotericin B, or tazobactam/piperacillin—was evaluated as a single composite variable.
Impact of the support system on loading dose administration rate
To assess the system’s impact, the biweekly administration rate of the loading dose was evaluated using a quasi-experimental interrupted time series analysis. The loading dose was defined as an initial dose of 25–30 mg/kg.
Impact of the support system on AUC optimization
We compared the impact of the new system on the optimization of VCM AUC before and after its introduction. AUCs were evaluated using AUC0–24, AUC24–48 (24–48-h AUC), and AUC_ss (steady-state AUC), all of which were calculated based on measured serum concentrations. Individual VCM AUCs were calculated by simulated concentration–time profiles generated by Bayesian estimation with EasyTDM. AUCs were determined using the trapezoidal rule, considering a 2-compartment model and each pharmacokinetic phase, including drug administration, distribution, and elimination. Reference DeRyke and Alexander18
Impact of the support system on patient outcomes
The impact of the system on patient outcomes was assessed by comparing data before and after its implementation. Outcomes included the incidence of adverse events such as AKI and 90-day all-cause mortality. AKI was defined as an increase in serum creatinine of ≥0.3 mg/dL or ≥50% from baseline in 2 consecutive measurements. Reference DeRyke and Alexander18 Other adverse events, including liver dysfunction, leukopenia, and thrombocytopenia, were evaluated based on the criteria of the Japanese Society of Chemotherapy Antimicrobial Agents Safety Evaluation Committee. Reference Khwaja19
Statistical analysis
The effects of the system on AUC and patient outcomes were evaluated using Fisher’s exact test for categorical variables and the Mann–Whitney U test for continuous variables. A P value of <.05 was considered statistically significant. A quasi-experimental interrupted time series analysis was conducted using a seasonal autoregressive integrated moving average model to adjust for baseline trends, autocorrelation, seasonality, and changes in level and slope before and after introducing the pharmacist-led support system. Reference Watanabe, Tokue and Aoki20,Reference Wagner, Soumerai, Zhang and Ross-Degnan21 Statistical analyses were performed using R software (version 4.4.3; R Foundation for Statistical Computing, Vienna, Austria), with the seasonal autoregressive integrated moving average model implemented via the Forecast package.
Ethics statements
Informed consent was obtained via an opt-out on the website in accordance with the Ethical Guidelines for Medical and Biological Research Involving Human Subjects in Japan. This study was conducted in compliance with the “Clinical Guidelines for Medical Research Involving Human Subjects” and approved by the Ethics Committee (Clinical Research Review Committee) established at our hospital (approval number 2024-175).
Results
Comparison of patient background characteristics before and after introducing the support system
Among the 587 patients who received VCM, 326 met the inclusion criteria and were evaluated. There were no significant differences in patient background characteristics or causative organisms between the pre- and postintervention periods (Table 1). Methicillin-resistant Staphylococcus aureus was identified in 26.6% and 21.2% of the patients in the preimplementation and postimplementation groups, respectively.
Table 1. Comparison of patient background characteristics before and after introducing the support system
For the percentage of each group for each item, a) median (25% IQR–75% IQR). Scr, serum creatinine; BMI, body mass index; BUN/Scr ratio, blood urea nitrogen/creatinine ratio; PBS, Pitt Bacteremia Score; MRSA, methicillin-resistant Staphylococcus aureus; MSSA, methicillin-sensitive Staphylococcus aureus.
Impact of the support system on loading dose administration rate
Results from the quasi-experimental interrupted time series analysis are presented in Figure 1 and Table 2. During the preintervention period, the loading dose administration rate remained stable (0.32% per 2 wk; P = .342). After introducing the support system, an immediate and significant increase of 43.15% was observed (95% confidence interval: 8.40–77.90; P = .0156). No significant upward trend was observed during the postintervention period ( −.21% per 2 wk, 95% confidence interval: −0.9 to 0.5; P = .546).
Figure 1. Time trend in loading dose administration rate following implementation of the support system.
Table 2. Impact of the support system on loading dose administration rate
CI, confidence interval.
Influence of the support system on AUC optimization
Following the administration of the support system, both AUC0–24 and AUC24–48 values were significantly higher compared with the preintervention group (241.0 vs 307.0, P < .001; 424.0 vs 485.0, P = .018). However, there was no significant difference in AUC_ss between the groups (459.0 vs 486.0, P = .644) (Table 3).
Table 3. Impact of the support system on patient outcomes
For the percentage of each group for each item, a) median (25% IQR–75% IQR), b) Mann–Whitney U test, c) Fisher’s exact test. AUC, daily area under the concentration–time curve; AUC0–24, 0–24-hour AUC; AUC24–48, 24–48-hour AUC; AUCss, steady-state AUC; AKI, acute kidney injury; IQR, interquartile range. Some patients experienced multiple adverse events.
Impact of the support system on patient outcomes
The 90-day mortality rate was 10.1% before and 7.8% after introducing the support system (P = .532), while the incidence of AKI was 12.8% and 12.4%, respectively (P = 1.000). No statistically significant differences were observed (Table 3).
Discussion
This study uniquely assessed the impact of a newly implemented dosing system using quasi-experimental interrupted time series analysis. A significant and immediate increase in the loading dose rate was observed following administration, with the effect sustained over 122 weeks. The loading dose was associated with a significant improvement in AUC0–24, without increasing the incidence of AKI or mortality. These findings suggest that the system may safely and sustainably enhance the quality of initial VCM dosing through AUC optimization.
Previous studies have shown that a uniform loading dose, such as 1.75 g for Staphylococcus aureus bacteremia, can improve composite outcomes, including mortality, without increasing nephrotoxicity. Reference Gillings, Makuc and Siegel22 Another report from an intensive care unit setting demonstrated that implementing a uniform 2 g loading dose increased the administration rate from 0% to 52.4% at 5 months and 63.9% at 16 months. Reference Mei, Wang, Che, Wang and Cai23 However, uniform dosing may lead to overdosing and an increased risk of nephrotoxicity, particularly when Cmin exceeds 20 mg/L if renal function and body weight are not considered. Reference Truong, Levkovich and Padiglione24 In this study, we introduced a pharmacist-led support system that recommends an individualized loading dose instead of a fixed-dose approach. The implementation of the pharmacist-led support system led to an appropriate increase in loading dose administration, which remained stable at approximately 50%–60%. This rate is comparable to that reported in facilities using uniform dosing, Reference Mei, Wang, Che, Wang and Cai23 suggesting that the support system has been established as standard practice and promotes sustained administration. However, because the system considers the physician’s final decision, further increases in the administration rate may have been limited.
These findings show that the appropriate increase in loading dose administration after introducing the support system significantly raised AUC0–24. Both groups achieved median AUC(24–48) values ≥400 µg·h/mL, the efficacy threshold, suggesting that the postintervention group might have reached effective exposure earlier. From a safety perspective, renal toxicity has been associated with AUC(24–48) ≥ 515 µg·h/mL and AUC_ss ≥ 600 µg·h/mL. Therefore, the recommended target AUC range is 400–500 µg·h/mL. Reference Lodise, Patel, Lomaestro, Rodvold and Drusano25 In this study, although AUC(24–48) increased significantly after administration, it remained within the target range. Moreover, AUC_ss did not change significantly. These findings align with previous reports indicating that loading doses do not affect AUC_ss, Reference Lodise, Rosenkranz and Finnemeyer26 suggesting that the implemented support system improved efficacy while maintaining safety.
In this study, although a trend toward lower 90-day mortality was observed in the postintervention group, the difference was not statistically significant. Similarly, no significant difference in AKI incidence was found between the two groups. These findings suggest that an increase in AUC0–24 may not be directly associated with 90-day mortality or AKI risk. However, based on PK/PD principles, achieving an appropriate AUC0–24 is expected to enhance efficacy. Previous studies have linked adequate AUC0–24 to improved clinical outcomes of Staphylococcus aureus infections Reference Moise-Broder, Forrest, Birmingham and Schentag10 and reduced AKI incidence. Reference Gillings, Makuc and Siegel22 In this study, AUC increased significantly without a rise in AKI incidence, indicating that the introduced support system may enhance efficacy while maintaining safety. Further large-scale studies with sufficient statistical power and subgroup analyses are warranted to clarify the clinical relevance of the observed AUC increase.
A limitation of this study is its single-center design. Future research should incorporate data from multiple institutions as this approach would enhance the generalizability and reliability of the findings.
In conclusion, the pharmacist-led support system increased the loading dose rate without negatively impacting AKI incidence or mortality, maintaining a balance between efficacy and safety. The improvement was immediate and sustained over the 122-week postintervention period. This system might serve as a practical model for broader administration in other healthcare settings.
Data availability statement
Data are available upon request and subject to privacy, ethics, and other restrictions.
Acknowledgments
We would like to express our deepest gratitude to the ward pharmacists of the Pharmacy Department and of the Antimicrobial Stewardship Team, who are working on the appropriate use of antimicrobial agents in our hospital.
Author contributions
All authors meet the ICMJE authorship criteria. Tatsuya Tai participated in the design of the study, collected the study data, and prepared the manuscript. Yuichi Muraki and Masahiro Watanabe participated in the study design and editing of the manuscript. Takahiro Motoki, Kazunori Yamaguchi, Naohiro Kurokawa, Sayaka Yamashita, Hiroaki Tanaka, and Shinji Kosaka assisted in the preparation of the manuscript.
Financial support
None reported.
Competing interests
YM received grants from Pfizer Japan Inc., Kowa Company Ltd., the Japan Pharmaceutical Association, the Japan Society for the Promotion of Science, and the Ministry of Health, Labour, and Welfare. YM is also a board member of the Japanese Society of Pharmaceutical Health Care and Sciences and the Japanese Society for Infection Prevention and Control, as well as a committee member of the AMR Clinical Reference Center.
The other authors report no conflicts of interest relevant to this article.
Open access
