In the present study, the risk of developing nephrotoxicity was assessed using logistic regression curves in the High-risk group with risk factors for nephrotoxicity when using VCM and the Low-risk group without risk factors, and measures to set the appropriate AUCss for each case were investigated. Based on comparisons of clinical parameters, the peak value, trough value, and initial blood sampling measured AUC were excluded as risk factors, and the AUCss was adopted. These parameters all depend on PK factors, such as the drug dosage, dosing interval, absorption, distribution, metabolism, and excretion, leading to correlations and multicollinearity. Correlation coefficients were 0.403 for the peak value, 0.498 for the trough value, and 0.432 for initial blood sampling measured AUC, all indicating correlations. Therefore, to avoid using these parameters simultaneously as risk factors, the AUCss was selected as the representative indicator. The relationship between the duration of VCM administration and nephrotoxicity was investigated. A comparison between the nephrotoxicity group and no nephrotoxicity group showed that the duration of administration was significantly longer in the former, while there was no significant difference in the period before the onset of nephrotoxicity between the groups. These results suggest that the administration of VCM continued even after the onset of nephrotoxicity in the nephrotoxicity group. Additionally, the lack of a significant difference in the no nephrotoxicity period suggests that the duration of VCM administration did not directly affect nephrotoxicity itself. On the other hand, the difficulties associated with controlling VCM blood concentrations with long-term use may contribute to the occurrence of nephrotoxicity. The AUCss was 573 mg·h/L [230–1984] (median [range]) in the High-risk group and 572 mg·h/L [306–1217] in the Low-risk group. Despite both groups being controlled within the effective concentration range, the High-risk group had a higher incidence of nephrotoxicity than the Low-risk group. Furthermore, the mean AUCss in the no nephrotoxicity group was 551 mg·h/L [230–1036], whereas it was significantly higher in the nephrotoxicity group at 656 mg·h/L [287–1984]. This result indicates that fluctuations in the AUCss associated with long-term administration increase the risk of nephrotoxicity. Therefore, considering fluctuations in and the difficulties associated with controlling blood concentrations with long-term administration, a duration of more than 14 days was selected as a potential risk factor. Based on these results, we identified the following factors as being closely related to VCM-induced nephrotoxicity: AST, ALT, a high dose of VCM (4 g/day), AUCss, dosing for more than 14 days, TAZ/PIPC, AGs, diuretics, NSAIDs, and chronic hepatic disease. After optimizing the model and performing a multivariate analysis, AUCss [7], TAZ/PIPC [19], diuretics [20], and chronic liver disease [16] were identified as independent risk factors for VCM-induced nephrotoxicity. These results are consistent with previous findings.

Diuretics [20] and hepatic disease [16] contribute to ischemia and cause nephrotoxicity by decreasing renal blood flow. The administration of VCM in the setting of a reduced circulating blood volume and renal blood flow due to infection may trigger acute kidney injury. TAZ/PIPC has been reported to increase the risk of VCM nephrotoxicity [19]. PIPC has the side effect of acute interstitial nephritis; however, the underlying mechanisms remain unclear. The combination of TAZ/PIPC and VCM is considered to cause tubular necrosis. The nephrotoxicity of AGs has been attributed to glomerular-filtered AGs being taken up by proximal tubular epithelial cells through endocytosis from the tubular lumen, resulting in tubular necrosis. Rybak et al. demonstrated that AGs caused nephrotoxicity at lower AUCss when administered with VCM than when given alone [8]. In the present study, 3 of 4 patients (Gentamicin: 2, Amikacin: 1) who received concomitant AGs developed nephrotoxicity; however, it was not identified as a risk factor. This result was due to the strong nephrotoxicity of AGs, resulting in a low number of cases combined with VCM, which, in turn, led to a low detection power.

Patients were divided into a High-risk group, with one or more independent risk factors for nephrotoxicity, and a Low-risk group, without any risk factors. The incidence of nephrotoxicity was compared between the two groups. The AUCss was 573 mg·h/L [230–1984] (median [range]) in the High-risk group and 572 mg·h/L [306–1217] in the Low-risk group, with blood concentrations controlled within the therapeutic range in both groups. However, the incidence of nephrotoxicity was significantly higher in the High-risk group (31.7%) than in the Low-risk group (13.0%). The incidence of nephrotoxicity in effective concentration range AUCss (400, 500, 600) were 2.1, 4.7, and 10.2% in the Low-risk group and 16.8, 23.3, and 31.4% in the High-risk group, respectively, which were higher in the latter. This result suggests the need for separate target concentration ranges for the two groups. Therefore, new target concentration ranges were defined for the High- and Low-risk groups using the threshold values obtained from the CART analysis and ROC curves as well as the probability of developing nephrotoxicity from logistic regression curves. In the High-risk group, the AUCss threshold was set at 575, and the probability of developing nephrotoxicity was compared in three groups: 400 ≤ AUCss < 500, 500 ≤ AUCss < 575, and 575 ≤ AUCss. The probability of developing nephrotoxicity in the 500 ≤ AUCss < 575 group did not significantly differ from that in the 400 ≤ AUCss < 500 group. However, the probability of developing nephrotoxicity in the logistic regression analysis was high at 23.3% at AUCss 500 and 29.3% at AUCss 575. Although the AUCss threshold for the High-risk group was 575.4, the discriminatory ability of the ROC curve was low at 0.665. The sigmoid curve obtained from the logistic regression analysis showed a smooth increase without abrupt changes at specific points, indicating a high rate of nephrotoxicity in the whole area. Therefore, the target concentration range appears to be 400 ≤ AUCss < 500 for safety reasons. A previous study reported that a VCM AUC24/MIC ratio of ≥ 505 was required for the treatment of severe infections [22]. The newly established target concentration range (400 ≤ AUCss < 500) is lower than the previously recommended range, necessitating caution to ensure its efficacy. Therefore, if the treatment effect is inadequate, a dose increase up to AUCss 505 is recommended and renal function needs to be carefully monitored with frequent measurements of blood levels. Additionally, if the treatment effect remains inadequate, switching to other anti-MRSA agents is suggested.

On the other hand, the AUCss threshold was set to 650 in the Low-risk group, and the probability of developing nephrotoxicity was compared among three groups: 400 ≤ AUCss < 575, 575 ≤ AUCss < 650, and 650 ≤ AUCss. The results obtained showed no significant difference in the probability of developing nephrotoxicity between 575 ≤ AUCss < 650 and 400 ≤ AUCss < 575.

The probability of developing nephrotoxicity by the logistic regression analysis was 8.4% for AUCss575 and 14.6% for AUCss650. In previous studies, the probability of developing nephrotoxicity due to VCM ranged between 12 and 48% [23, 24]. The AUCss threshold for the Low-risk group was 638.7, and the AUC of the ROC curve was 0.832, indicating a high discriminatory ability. Additionally, the sigmoid curve obtained from the logistic regression analysis showed an initial gradual increase, with a sharp elevation in the probability of developing nephrotoxicity near the threshold. Since the Low-risk group had a high safety profile, the target concentration range was newly set at 400 ≤ AUCss < 650, suggesting the safe administration of the drug up to AUCss650 while aiming for AUCss600 from the initial dose design. The upper limit of the target concentration range was considered to be 650 from the viewpoint of safety because the probability of developing nephrotoxicity was significantly higher at 650 ≤ AUC.

The increased probability of developing nephrotoxicity due to VCM was previously shown to be confounded by the concomitant use of nephrotoxic drugs, hemodynamic changes, and the effects of underlying diseases [25]. Although VCM is not classified as a nephrotoxic drug in the Kidney Disease Improving Global Outcomes acute kidney injury guidelines, the present study indicates that VCM itself is associated with a risk of nephrotoxicity [26].

The present results may contribute to the individualization of VCM dosing plans for the treatment of infectious diseases. As mentioned in the Introduction, the recommended effective concentration range for VCM is AUCss 400–600 mg·h/L for efficacy and safety [2]. However, the present study demonstrated that the probability of developing nephrotoxicity was high in the nephrotoxicity group, even within the effective concentration range, and, thus, a target AUCss needs to be set for each patient in consideration of the balance between treatment efficacy and the prevention of side effects. Holford et al. recommended a Target Concentration Intervention (TCI) as an alternative strategy to TDM [27]. TCI focuses on achieving specific target drug concentrations tailored to each patient’s condition. This approach involves establishing target drug concentrations that maximize therapeutic effects while minimizing side effects, and then adjusting the dosage to reach these concentrations. TCI utilizes PK models that take into account individual factors, such as the patient’s weight, age, renal function, and hepatic function. This enables the provision of optimal therapeutic effects for each patient while minimizing the risk of adverse effects. On the other hand, the concept of a therapeutic range for TDM has been reported to reduce the expected clinical benefit to patients because measured values below the lower end of the range, within the range, and above the upper end are classified as ‘sub-therapeutic’, ‘therapeutic’, and ‘toxic’, respectively, leading to uniform dosing recommendations based on measured blood levels [27, 28]. In the present study, the target concentration range was divided into High- and Low-risk groups, and a logistic regression curve was used to quantify the risk of nephrotoxicity for each AUCss. This allowed for proposals of strategic individual target concentrations based on the balance between risk and benefit. This approach aligns with the principles of TCI, which utilizes PK models that consider individual patient factors.

The novelty of this study lies in the identification of risk factors for nephrotoxicity in the design of VCM dosing and the establishment of optimal AUCss for each risk group. VCM is an antimicrobial agent used in many healthcare facilities and, thus, setting appropriate guidelines to reduce the risk of nephrotoxicity is extremely important. Previous studies reported the risk of nephrotoxicity with the administration of VCM, and the present study proposes an individualized dosing strategy based on specific risk factors. Furthermore, by examining the incidence of nephrotoxicity in detail in patients divided into High- and Low-risk groups, we provide useful insights for personalized medicine. This study is unique in that it classifies VCM into High- and Low-risk groups based on the presence or absence of risk factors and proposes a target concentration range appropriate for each group.

The present study has several limitations that need to be addressed. The number of cases in this study was small at only 212. In addition, the number of patients who received the combination of AGs and VCM, which is highly nephrotoxic, was small (only 4). Future multicenter studies on a more diverse patient population are needed. Since this study focused on reducing the risk of nephrotoxicity, the evaluation of the efficacy of VCM was insufficient. Therefore, therapeutic effects within the proposed target concentration range need to be confirmed. The 7-day monitoring of renal function after the cessation of VCM administration was useful for evaluating short-term changes in renal function, but was considered to be insufficient for assessing long-term effects. In the future, follow-up periods of 3 months or longer will be necessary to evaluate the risk of acute kidney injury progressing to chronic kidney disease. Although the group with nephrotoxicity risk factors was classified as the High-risk group, the probability of developing nephrotoxicity may differ between patients with a single risk factor and those with multiple risk factors. The regression curve obtained from the logistic regression analysis suggested a gradual increase in the incidence of nephrotoxicity from the Low-risk line to the High-risk line, according to the level of risk. Conducting large-scale multicenter collaborative studies and securing a sample size that maximizes detection power will allow for a detailed examination of the incidence of nephrotoxicity according to each risk factor and combinations of multiple risk factors. However, the development of nephrotoxicity largely depends on the patient’s condition. Therefore, it is crucial to understand the condition of patients with risk factors and the probability of developing nephrotoxicity at each AUCss, while frequently measuring blood concentrations and carefully monitoring renal function.

In recent years, higher doses have been required to prevent hyposensitization to VCM [29]. The development of prerenal acute renal failure due to septic shock or endotoxin shock is common in the ICU. Difficulties are associated with establishing whether renal failure in severe cases is due to prerenal renal failure caused by renal ischemia or drug-induced acute kidney injury. Therefore, a prospective study with data that refutes renal ischemia is needed to clarify the risk factors for nephrotoxicity. The validity of this method will be confirmed in the future by increasing the number of patients and tracking treatment outcomes in the Low- and High-risk groups.