Is Early Monitoring Better? Impact of Early Vancomycin Exposure on Treatment Outcomes and Nephrotoxicity in Patients with Methicillin-Resistant Staphylococcus aureus Infections

Optimal early vancomycin target exposure remains controversial. To clarify the therapeutic exposure range, we investigated the association between vancomycin exposure and treatment outcomes or nephrotoxicity in patients with methicillin-resistant Staphylococcus aureus (MRSA) infection. This retrospective study reviewed clinical data obtained from 131 patients with MRSA infections between January 2017 and September 2019. Clinical outcomes included treatment failure, 30-day mortality, microbiological failure, and acute kidney injury. We measured serum vancomycin levels after the first dose to 48 h and estimated vancomycin exposure using the Bayesian theorem. The minimum inhibitory concentration (MIC) of antimicrobial agents was determined using the broth microdilution method. Classification and Regression Tree analyses identified day 1 and 2 exposure thresholds associated with an increased risk of failure and nephrotoxicity. Treatment failure (27.9% vs. 33.3%) and 30-day mortality (26.6% vs. 31.74%) were numerically but not significantly reduced in patients with the area under the curve (AUC)24–48h/MICBMD ≥ 698. Patients with AUCss/MICBMD ≥ 679 exhibited a significantly increased risk of acute kidney injury (27.9% vs. 10.9%, p = 0.041). These findings indicate that AUCss/MICBMD ratios > 600 may cause nephrotoxicity. AUC/MICBMD at days 1 and 2 do not appear to be significantly associated with particular clinical outcomes, but further studies are needed.


Introduction
Methicillin-resistant Staphylococcus aureus (MRSA) is the most common cause of nosocomial infections. The Asian Network for Surveillance of Resistant Pathogens (ANSORP) study reported that MRSA accounts for 57% of hospital-acquired infections in Thailand [1].

Results
From January 2017 to September 2019, we obtained results from 131 MRSA infection patients with 315 measured vancomycin concentrations. Of the 131 patients with MRSA infections, 88 (67.2%) were men, the mean age was 70.1 years, and there were 82 (62.6%) critically ill patients. All patients received intravenous vancomycin for more than 48 h. Most patients were diagnosed with pneumonia and bacteremia. The MIC range, MIC50, and MIC90 for vancomycin against the studied MRSA isolates were 0.5-2.0 µg/mL, 1 µg/mL, and 1 µg/mL, respectively (Table 1). Table 3 shows the vancomycin exposures (AUC/MIC BMD ) and clinical outcomes.

Nephrotoxicity
Eighty nine patients had a baseline serum creatinine level <2.0 mg/dL. According to CART analysis-derived vancomycin exposure values, an AUC ss /MIC BMD of 679 was the breakpoint for AKI. The number of patients with AKI was 17, and was significantly higher in the high vancomycin exposure groups (AUC ss /MIC BMD 679 (27.9% vs. 10.9%, p = 0.041)) ( Figure 1). Kaplan-Meier survival analysis of the time to AKI demonstrated a significantly increased risk of AKI in the high AUC group (>670) (log-rank test p = 0.042) ( Figure 2).

Nephrotoxicity
Eighty nine patients had a baseline serum creatinine level <2.0 mg/dL. According to CART analysis-derived vancomycin exposure values, an AUCss/MICBMD of 679 was the breakpoint for AKI. The number of patients with AKI was 17, and was significantly higher in the high vancomycin exposure groups (

Discussion
This study evaluated vancomycin exposure during the first dose of its administration, and the resultant clinical outcomes, in patients with MRSA infections. These results were obtained using broth microdilution methods to measure MICs (the gold standard method for vancomycin susceptibility determination) [9,10,13]. We estimated the AUC values using individual vancomycin levels. We used vancomycin concentrations after the first dose to calculate the AUC values that reflected vancomycin PK parameters in patients during the first 24 h. In contrast, previous studies often determined vancomycin concentrations at steady-state, which may be in the transition to the recovery phase of sepsis.
The current guidelines, provided by the American Society for Health System Pharmacists (ASHP)/Infectious Diseases Society of America (IDSA)/the Pediatric Infectious Diseases Society (PIDS/the Society of Infectious Diseases Pharmacists (SIDP), no longer recommend trough concentrations of 15-20 mg/L. AUC-based dosing is preferable because AUC24/MICss is the most accurate PK/PD parameter. In patients with suspected or definitive serious MRSA infections, an individualized target of an AUC/MICBMD ratio between 400 and 600 (assuming a vancomycin MICBMD of 1 µ g/mL) is recommended to achieve clinical efficacy and improve patient safety [6]. This target value came from studies on patients with MRSA bacteremia; most studies estimated PK/PD targets using formulas [8,[14][15][16]. This study used Bayesian software to estimate AUC values, which provides an advantage in critically ill patients since vancomycin concentrations could be collected within the first 24 to 48 h, rather than at steady-state conditions (after the third or fourth dose) [6].
Moreover, the current guidelines recommend Bayesian-derived AUC monitoring based on Bayes's theorem. Bayesian dose optimization software uses a well-developed vancomycin population PK model as the Bayesian prior, together with the individual patient's measured drug concentrations in the data file, to calculate a Bayesian posterior parameter value distribution for that patient. Pea et al. assessed therapeutic drug monitoring (TDM) using a Bayesian approach to determine vancomycin dosages compare with a nomogram in critically ill patients. The study suggested that a TDM-guided Bayesian-based approach should be considered an invaluable tool for clinicians to appropriately monitor real-time vancomycin therapy in critically ill patients [17]. In this study, peak and trough vancomycin concentrations were measured in over half the patients for whom the previous study showed peak-trough based vancomycin TDM improved the therapeutic cure rate [18].

Discussion
This study evaluated vancomycin exposure during the first dose of its administration, and the resultant clinical outcomes, in patients with MRSA infections. These results were obtained using broth microdilution methods to measure MICs (the gold standard method for vancomycin susceptibility determination) [9,10,13]. We estimated the AUC values using individual vancomycin levels. We used vancomycin concentrations after the first dose to calculate the AUC values that reflected vancomycin PK parameters in patients during the first 24 h. In contrast, previous studies often determined vancomycin concentrations at steady-state, which may be in the transition to the recovery phase of sepsis.
The current guidelines, provided by the American Society for Health System Pharmacists (ASHP)/Infectious Diseases Society of America (IDSA)/the Pediatric Infectious Diseases Society (PIDS/the Society of Infectious Diseases Pharmacists (SIDP), no longer recommend trough concentrations of 15-20 mg/L. AUC-based dosing is preferable because AUC 24 /MICss is the most accurate PK/PD parameter. In patients with suspected or definitive serious MRSA infections, an individualized target of an AUC/MIC BMD ratio between 400 and 600 (assuming a vancomycin MIC BMD of 1 µg/mL) is recommended to achieve clinical efficacy and improve patient safety [6]. This target value came from studies on patients with MRSA bacteremia; most studies estimated PK/PD targets using formulas [8,[14][15][16]. This study used Bayesian software to estimate AUC values, which provides an advantage in critically ill patients since vancomycin concentrations could be collected within the first 24 to 48 h, rather than at steady-state conditions (after the third or fourth dose) [6].
Moreover, the current guidelines recommend Bayesian-derived AUC monitoring based on Bayes's theorem. Bayesian dose optimization software uses a well-developed vancomycin population PK model as the Bayesian prior, together with the individual patient's measured drug concentrations in the data file, to calculate a Bayesian posterior parameter value distribution for that patient. Pea et al. assessed therapeutic drug monitoring (TDM) using a Bayesian approach to determine vancomycin dosages compare with a nomogram in critically ill patients. The study suggested that a TDM-guided Bayesian-based approach should be considered an invaluable tool for clinicians to appropriately monitor real-time vancomycin therapy in critically ill patients [17]. In this study, peak and trough vancomycin concentrations were measured in over half the patients for whom the previous study showed peak-trough based vancomycin TDM improved the therapeutic cure rate [18].
Using this validated Bayesian method to estimate the daily AUC in a single-center, retrospective study of patients with MRSA bloodstream infections, Lodise et al. found that failure outcomes were maximized when day 1 and day 2 AUC/MIC BMD ratios exceeded 521 and 650, respectively [9]. A multicenter, observational prospective study was performed to evaluate the relationship between prespecified day 2 AUC/MIC ratios (AUC/MIC BMD of ≥ 650) and outcomes in adult patients with MRSA bacteremia. The study, performed by Ho et al. in MRSA bacteremia patients, showed that treatment outcomes were improved by adjusting the dose to achieve an AUC 24 /MIC BMD of > 400 based on individual vancomycin clearances and the vancomycin MIC of the infection-causing strain [19].
In our study, higher vancomycin exposures were not associated with treatment failure, 30-day mortality, or microbiological failure. In contrast, AKI was associated with higher AUC ss /MIC BMD values ≥679. In the previous multivariate analyses, treatment failure rates were not significantly different between the pre-specified day 2 AUC/MIC groups [10], as our study found that treatment failure was not significantly associated with higher vancomycin exposures on day 1 (AUC/MIC BMD of ≥626) and day 2 (AUC/MIC BMD of ≥698). Approximately 32% of patients in our study presented with hemodialysis and renal failure in critical condition, including continuous renal replacement therapy (CRRT). Even though the AUC/MIC values were higher than in previous studies, patient condition trends in the higher 30-mortality rate from a study in end-stage renal disease (ESRD) patients in an ICU setting were about 20-40%, compared with those without renal failure [20,21]. Based on our findings (data not shown in the results part), the patients with ESRD had a significantly higher mortality rate than the non-ESRD group (42.86% vs. 22.47%).
Not surprisingly, the CART-derived exposure AUC ss /MIC BMD values showed that the rates of AKI were significantly higher in patients in the high vancomycin exposure groups (AUC ss /MIC BMD 679 (27.9% vs. 10.9%, p = 0.041)]). The AUC breakpoint associated with an increased risk of AKI in this study was notably consistent with previous reports [10,11,[22][23][24][25]. However, in our analysis, there was no correlation between AUC 0-24h /MIC BMD (626) and AUC 24-48h /MIC BMD (698) with nephrotoxicity, unlike in a previous study [22]. The data from a previous meta-analysis [26,27] were similar to the updated guidelines using AUC-guided dosing for reduced nephrotoxicity at values lower than 650. Notably, the higher AUC value cited (above 600 mg h/L) provides a new index that incorporates both efficacy and AKI, and is within the recommended AUC range of 400 to 600 mg h/L (assuming an MIC of 1 mg/L) [6,28].
This study has several limitations. First, our study used Bayesian software that uses a one-compartment model as a prior PK parameter, which may not reflect that our study included about 60% critically ill patients. However, previous population PK studies in critically ill or sepsis/septic shock patients used both a one-and two-compartment model. Second, the sample size in our study was small. Thus, a larger MRSA population may show significant differences in treatment outcomes. Third, vancomycin measurement assays in this study had been reported to impair reaction kinetics leading to the incorrect results, especially below the measuring range (<4.0 µg/mL). However, in this study, patients in the higher group of vancomycin exposure tended to show better treatment outcomes than those in the lower exposure groups.

Study Design
This retrospective study was conducted between January 2017 and September 2019 at Phramongkutklao Hospital, a 1200-bed university hospital in Bangkok, Thailand. Patients who were admitted with an MRSA infection, treated with intravenous vancomycin infusion, and had serum vancomycin concentration data were included.
The ethics review committee of the Royal Thai Army Medical Department, Bangkok, Thailand (approval no. Q0007h/63_Exp) approved this study. It was performed following the Declaration of Helsinki. Data were collected after obtaining ethical approval and permission from the Director of the Phramongkutklao Hospital.
Patient data were collected and analyzed anonymously and confidentially. Identifying information of participants was not collected, and only the researcher could access the data. The ethics review committee did not require patient informed consent for using serum vancomycin levels, retrospective chart review studies, and confidential and anonymized data.

Microbiological Data
We determined the MIC of the non-repeated clinical isolates. The MIC of the antimicrobial agents was determined using automated susceptibility testing (Thermo Scientific™ Sensititre™ ARIS™ 2X Instrument) based on the broth microdilution method. The MIC value of the antimicrobials in each strain was denoted as susceptible, intermediate, or resistant using the Clinical and Laboratory Standards Institute (CLSI) breakpoint [29].

Vancomycin Assay
Serum vancomycin concentrations were determined using a fluorescence immunoassay (VANC3, Cobas, Roche Diagnostics, Indianapolis, USA). The limit of detection of this assay for vancomycin was 4 µg/mL. The coefficient of variation for this assay was <10%.

Treatment Data and Outcomes
Patient data were obtained from electronic medical records. Information collected included demographic data, infection dates, specimens for microbial culture, types of infections, the severity of illness, antimicrobial susceptibility, treatment duration, and treatment outcomes.
Serum vancomycin concentrations were collected at 1-2 h after the end of the infusion of the loading dose, and 30 min before the second dose or 15 min before the fourth vancomycin dose. The 24 h AUCs for vancomycin were estimated using Bayesian dose optimizing software by Precise PK (San Diego, CA, USA).
The primary outcome measure was treatment failure, defined as any of the following: (1) death within 30 days after treatment (30-day mortality), (2) microbiological failure (specimen culture growing MRSA obtained seven days after the initiation of therapy). Secondary outcome measures included the occurrence of AKI among patients with a baseline serum creatinine <2 mg/dL. AKI was based on post-baseline serum creatinine levels ≥1.5×baseline serum creatinine or an increase >0.5 mg/dL, and was assessed from the initiation of vancomycin to 48 h after initiation [30]. Clinical outcomes were indicated with an AUC 0-24h /MIC BMD of 521 and AUC 24-48h /MIC BMD of 650, from previous studies [9,10].

Statistical Analysis
We reported the descriptive statistics for the clinical characteristics and the MICs at the 50th (MIC50) and 90th percentile (MIC90). CART analysis was used to determine the AUC/MIC cut-off value for vancomycin treatment outcomes (included treatment failure, 30-day mortality, microbiological failure, and nephrotoxicity). The Chi-square test was used to compare categorical variables between groups. The survival analysis of AKI used Kaplan-Meier estimates. Data were analyzed in Statistical Package for the Social Sciences (SPSS) version 27.0 (IBM Corp., Armonk, NY, USA), and a p-value <0.05 was considered statistically significant.

Conclusions
Our findings suggest that AUC/MIC BMD values on days 1 and 2 were not associated with treatment outcomes for patients with MRSA infections. However, this study confirmed that nephrotoxicity increased in patients with a high AUC ss /MIC BMD of 679. AUC/MIC BMD values of 400-600 remain the best target window to achieve favorable outcomes and minimize the likelihood of AKI, whereas the early threshold of vancomycin exposures remains unclear.
Funding: This research was funded by Payap University and Faculty of Pharmacy, Silpakorn University. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.

Conflicts of Interest:
The authors declare no conflict of interest.