Vancomycin is a glycopeptide antibiotic that inhibits bacterial cell wall synthesis by binding to the D-alanyl-D-alanine terminus of cell-wall precursors [1]. Due to its excellent antibiotic effect against gram-positive bacteria, vancomycin is used as the first-line therapy for methicillin-resistant Staphylococcus aureus (MRSA) infection, and other various indications including neutropenic fever, coagulase negative staphylococcus (CoNS), and other gram-positive bacterial infections [1]. However, since vancomycin has a narrow therapeutic index, therapeutic drug monitoring (TDM) is necessary for improving efficacy while preventing possible side effects, especially nephrotoxicity [2].

Since the release of the first consensus guideline on vancomycin monitoring, the 24-h AUC (AUC)/minimum inhibitory concentration (MIC) has been the primary pharmacokinetic (PK)/pharmacodynamic (PD) index. Nevertheless, serum vancomycin trough concentration (C_trough) has been suggested to be a surrogate monitoring index because it is easily available [2, 3]. However, the increase in vancomycin PK data raised concerns regarding C_trough use since C_trough greater than 15 mg/L frequently caused high AUC, which was associated with frequent vancomycin toxicity. Vancomycin TDM guidelines have been changed to reflect these new findings. The American Society of Health-System Pharmacists (ASHP), the Infectious Diseases Society of America (IDSA), the Pediatric Infectious Disease Society (PIDS), and the Society of Infectious Disease Pharmacists (SIDP) published a new consensus guideline in 2020 suggesting the use of AUC/MIC instead of C_trough for vancomycin monitoring in patients with serious MRSA infection [4]. Moreover, the revised guideline of the Chinese Pharmacological Society included the recommendation for monitoring AUC along with C_trough for vancomycin TDM [5].

However, due to the many drawbacks of AUC-based TDM, particularly the difficulty in obtaining AUC and unfamiliarity, many clinicians are still using C_trough for vancomycin monitoring. Surveys on the implementation and perceptions of vancomycin monitoring revealed that significant number of pharmacists, physicians, and institutions were unaware of the transition to AUC-based TDM or were not planning to adopt AUC-based TDM [6-9]. Therefore, intuitive investigation of the relationships between AUC and C_trough and the AUC attainment rate of C_trough-based TDM would help physicians use the adjusted TDM target in clinical practice.

This retrospective in silico study analyzed previous TDM data to assess the correlation between C_trough and AUC, the attainment rate of AUC target range, and the change in vancomycin dose based on a recent American guideline and the commonly used C_trough-based TDM in a Korean population.

We evaluated the TDM data of patients who were referred for TDM consultation from April 1st, 2019, to April 30th, 2020, at the Department of Laboratory Medicine and Genetics, Samsung Medical Center Seoul, Korea. Patients with at least one vancomycin trough measurement were enrolled. Patients who were younger than 18 years, admitted to an intensive care unit, or on renal replacement therapy were excluded. If patients were consulted multiple times during the given period, data from first TDM consultation were used for analysis. For impact analysis when transitioning to AUC-based TDM from trough-based TDM, C_trough data within 15-20 mg/L with the given regimen were included.

The following demographic characteristics, clinical information, and laboratory data were obtained for analysis and TDM: sex, age, height, body weight, diagnosis, vancomycin indication, vancomycin regimen, sampling time, serum vancomycin concentration, and serum creatinine. Creatinine clearance (CL_Cr) was calculated from serum creatinine concentration using Cockcroft-Gault equation. Vancomycin concentration was measured by the kinetic interaction of microparticles in solution immunoassay on Roche Cobas c702 analyzer (Roche, Basel, Switzerland). Based on our institution protocol, initial vancomycin measurements were performed prior to the third or fourth vancomycin injection. Follow-up vancomycin measurements, mostly prior to the third injection, were performed at the physician’s discretion.

We used target AUC and C_trough ranges recommended by the American guidelines. The primary C_trough target range was established based on vancomycin indication: 15-20 mg/L for bacteremia, sepsis, pneumonia, catheter-related blood stream infection (CLABSI), central nervous system infection, and endocarditis, and 10-15 mg/L for neutropenic fever, skin infection, post-operative wound infection, otitis, and other non-complicated infections. The therapeutic range for AUC was 400-600 mg*h/L for serious MRSA infection and was extrapolated to other indications to explore the PK dynamics of vancomycin. The AUC attainment rate was defined as the proportion of patients with AUC within 400-600 mg*h/L divided by the total patients in the given C_trough target range. Similarly, the rate of failure to obtain AUC therapeutic range (failure rate) was defined as the proportion of patients who did not have AUC within the therapeutic range divided by the total patients in the given C_trough target range.

Estimated C_trough and AUC were calculated with commercial Bay-esian PK software, MWPharm++ (MW) (Mediware, Praha, Czech Republic), using two-compartment vancomycin KKGT (Dutch Association for Quality Assessment in TDM and Clinical Toxicology) population model [10]. The mean and standard deviation of PK fitting parameter were as follows: C01.V (L/KgLBMc)=0.21±0.04, C02.kxy (1/h)=1.12±0.28, C02.kyx (1/h)=0.48±0.12, and RE.k (1/h/(mL/min/1.73 m²))=0.00327±0.00109. Patient covariates were age, sex, height, weight, and CL_Cr. For each patient, the recommended regimen for previous C_trough-based TDM consultations was simulated. The correlation between C_trough and AUC was evaluated, and the rate of failure to achieve AUC therapeutic range was calculated for C_trough target range with lower limit of 0-20 mg/L and range interval of 3.0 mg/L, 3.5 mg/L, 4.0 mg/L, 4.5 mg/L, and 5.0 mg/L. Furthermore, to investigate the expected changes in dose and trough when changing monitoring index from C_trough to AUC, additional analysis was performed on patients with C_trough of 15-20 mg/L. The doses needed to obtain AUC 400 mg*h/L (AUC400), 500 mg*h/L (AUC500), and 600 mg*h/L (AUC600) and the corresponding C_trough values were calculated.

MedCalc^® version 77.5.1.0 (MedCalc Software Ltd., Ostend, Belgium) and IBM SPSS^® Statistics version 25 (IBM Corp., Armonk, NY, USA) were used for statistical analyses. Pearson correlation analysis for estimated C_trough and AUC data from all enrolled patients was performed. Paired t-test and Wilcoxon signed rank test were performed to demonstrate statistically significant differences between C_trough and vancomycin dose of AUC-based TDM. A P-value less than 0.05 was considered statistically significant.

This study was approved by the Institutional Review Board (IRB) of Samsung Medical Center, Seoul, Korea (IRB number. SMC 2021-02-125-001), and the need for informed consent was waived.

The detailed characteristics of patients are described in Table 1. Vancomycin TDM data from 104 patients were available for correlation analysis between C_trough-based and AUC-based TDM. Under the C_trough-based recommended regimen, estimated C_trough was categorized into three groups: less than 10 mg/L (n=7), 10-15 mg/L (n=66), and 15-20 mg/L (n=31), as was estimated AUC: less than 400 mg*h/L (n=7), 400-600 mg*h/L (n=79), and 600 mg*h/L or more (n=18).

The AUC was within the therapeutic range in 18 of 31 patients with C_trough of 15-20 mg/L (18/31; attainment rate, 58.1%). On the other hand, 57 patients with C_trough of 10-15 mg/L had AUC within the therapeutic range (57/66; attainment rate, 86.4%). These results suggest that the AUC therapeutic range is difficult to obtain with the previously suggested target range of C_trough. Instead, a lower C_trough is needed to achieve the AUC therapeutic range. Pearson correlation analysis revealed moderate correlation between estimated C_trough and AUC (R=0.707, P<0.001) (Fig. 1).

In estimating failure rate when using C_trough range, lower failure rates were observed at target ranges of 11.9 mg/L-14.9 mg/L, 11.0 mg/L-14.5 mg/, 10.7 mg/L-14.7 mg/L, 10.2 mg/L-14.7 mg/L, and 9.5 mg/L-14.5 mg/L (Table 2). The upper limits of all suggested C_trough target ranges were near 15 mg/L, and the lower limits varied by range interval. Each candidate C_trough target range corresponded with AUC therapeutic range in 90% of patients.

Of the 104 patients, 31 met the 15-20 mg/L target based on previous TDM C_trough target. Among them, data from 28 patients were included for the assessment of change in vancomycin dose and C_trough when using AUC-based TDM. There were statistically significant decreases in both vancomycin dose and C_trough when targeting AUC 400 mg*h/L and 500 mg*h/L compared to the values when targeting C_trough of 15-20 mg/L (Fig. 2).

To the best of our knowledge, this is the first study comparing AUC-based and trough-based TDM using data from real clinical practices in Korean patients. In this retrospective study comparing C_trough-based TDM and AUC-based TDM, the achievement of target range and the correlation between C_trough and AUC were assessed in a Korean population. A moderate correlation was observed between AUC and C_trough. However, a considerable number of patients with C_trough of 15-20 mg/L were over-treated based on the recent AUC therapeutic range guidelines. Moreover, most patients with lower than recommended C_trough had an AUC therapeutic range of 400-600 mg*h/L. This unsatisfactory compatibility between C_trough target range and AUC target range was consistent with previous findings [4, 11-17]. Compared to trough-based TDM, AUC-based TDM was expected to achieve lower vancomycin dose and C_trough.

For AUC-based TDM, the revised American guideline has shown that lower vancomycin dosage could be adopted to prevent possible adverse effects of vancomycin, such as nephrotoxicity. The current American guideline based on several studies suggests an upper limit of AUC of 600 mg*h/L to prevent nephrotoxicity [4, 11, 18-21], and AUC-guided dosing based on this target range reduced nephrotoxicity without a significant increase in treatment failure compared to C_trough-based dosing [15, 20, 22-26]. Based on currently published data, adopting AUC-based TDM would help prevent unnecessary risk of nephrotoxicity.

However, not all institutions have access to Bayesian PK software, and some clinicians still use C_trough for vancomycin monitoring. Indeed, AUC is not intuitive as C_trough because it requires additional PK parameters and is derived from the calculation of such parameters. In the most recent survey conducted in 2020, 70% of responding clinicians answered that they had not implemented AUC-based TDM and 43.0% were not planning to adopt AUC-based TDM [6]. Therefore, although AUC-based TDM appears to be superior to trough-based TDM, alternative methods are needed to avoid the hesitancy associated with adopting AUC-based TDM. One option could be establishing an individual target C_trough based on the initial AUC calculation using the first-order PK equation and monitoring according to an established target C_trough range until significant conditional change occurs [27]. Further practical assessment and adjustment of vancomycin monitoring would be helpful.

Although this article mainly focused on the American guideline, the Chinese guideline does not discourage the use of C_trough for monitoring vancomycin. This might be because the Chinese guideline recommends a lower vancomycin C_trough lower limit (10 mg/L) and higher AUC upper limit (650 mg*h/L) for serious MRSA infections compared to that recommended by the American guideline [3-5]. In this study, to assess whether simply lowering target C_trough is acceptable and explore the optimal target for C_trough range, we evaluated the rate of AUC therapeutic range attainment with lower vancomycin C_trough target range. Regardless of range interval, the upper limits of the optimal candidate ranges were near 15 mg/L, which is lower than the C_trough suggested by the 2009 American guideline. Considering this low failure rate, using lower target C_trough would be acceptable when AUC-based TDM is not conducted. However, further study is needed to determine if a low C_trough target yields the same clinical outcome as AUC-based TDM.

There are some limitations in this study. First, as this study was an in silico prediction, the comparison of clinical outcomes by TDM strategy was not performed. Moreover, when performing TDM simulation, we included all previous TDM consultation data including those with only one vancomycin measurement or non-steady state serum vancomycin concentration and real steady state C_trough could not be obtained. However, as stated in the current guideline and other studies, vancomycin TDM with the Bayesian approach can yield reliable prediction even with single non-steady state concentration [4, 28]. In addition, by simulating TDM as it is applied currently in practice, the data reflect real-world practice. Finally, this study was based on the assumption of AUC 400-600 mg*h/L as the accepted target range and assessed only PK factors affecting vancomycin administration. The actual applicability of AUC for indications other than serious MRSA infection, such as vancomycin-susceptible Enterococcus infection, was not addressed in this study. A further large-scale prospective study on the extrapolation of AUC target range is needed.

In conclusion, targeting C_trough of 15-20 mg/L is not appropriate for attaining the target AUC range and is associated with high AUC values. Using AUC-based TDM could result in a lower vancomycin dose. Applying a low C_trough might be acceptable, but further large study on the actual clinical outcomes is needed.