Association between Vancomycin Pharmacokinetic Parameters and Clinical and Microbiological Efficacy in a Cohort of Neonatal Patients

ABSTRACT

Vancomycin pharmacokinetic/pharmacodynamic (PK/PD) targets have not been validated in the neonatal population as no specifically designed studies are available. The main goal of this study was to analyze the therapeutic vancomycin regimen, the 24-h area under the curve (AUC₂₄), and the trough plasma concentration (C_t) obtained that achieved clinical and microbiological effectiveness in a cohort of neonates. This was an observational, prospective, single-center study covering a period of 2 years. Eligible patients were neonates and young infants who were undergoing treatment with intravenous vancomycin for ≥72 h with ≥1 C_t available. The primary outcome was the association of C_t and AUC₂₄ with clinical and microbiological efficacy at the beginning (early clinical evolution [ECE]) and the end (late clinical evolution [LCE]) of treatment with vancomycin. A total of 43 patients were included, 88.4% of whom were cured. In ECE, the cutoff points of the receiver operating characteristic (ROC) curve were 238 mg · h/L (sensitivity of 61% and specificity of 88%) for AUC₂₄ and 6.8 μg/mL (sensitivity of 61% and specificity of 92%) for C_t. In LCE, the C_t value was 11 μg/mL, with a sensitivity of 80% and a specificity of 92%. In this analysis, AUC₂₄ was not considered a good predictor. Logistic regression showed that a vancomycin C_t of ≤6.8 μg/mL was associated with an unfavorable ECE (P = 0.001), being 18 times more likely to progress poorly compared to those with higher levels. AUC₂₄ and C_t are good predictors of ECE in this population. Concentrations close to 7 μg/mL and an AUC₂₄ of around 240 mg · h/L 48 h after antibiotic initiation seem to be sufficient to achieve clinical cure in most cases.

INTRODUCTION

Optimized antimicrobial dosing is frequently required to obtain adequate antibiotic exposure for effective treatment. Antimicrobial dosing is critical in neonates since they exhibit great inter- and intraindividual pharmacokinetic (PK) variability related to maturation (1). Therapeutic drug monitoring (TDM) consists of dose individualization based on plasma drug concentrations and allows dosage optimization according to the individual characteristics of each patient, achieving a greater efficacy of the antimicrobial treatment (2) and reducing its risk of toxicity (3).

Vancomycin is a glycopeptide antibiotic commonly used for the treatment of infections caused by coagulase-negative staphylococci (CoNS), methicillin-resistant Staphylococcus aureus (MRSA), and Enterococcus sp. (4). The pharmacokinetics of vancomycin in neonates vary considerably compared to those in older children or adults. This is due to a higher percentage of body water, which affects the volume of drug distribution (5), in addition to the lower protein binding and reduced renal clearance in these patients (6).

Vancomycin is considered a time- and concentration-dependent antibiotic whose killing activity is predicted by the ratio of the 24-h area under the curve (AUC₂₄) to the MIC. This PK/pharmacodynamic (PD) parameter was established in a previous study conducted in adult patients with lower respiratory tract infections caused by Staphylococcus aureus where a positive correlation between a good clinical response and an AUC₂₄/MIC ratio of ≥400 was demonstrated (7). This ratio has been assumed to be the PK/PD parameter that best correlates with efficacy in the adult population. Thus, a meta-analysis carried out in 2016 showed that achieving an AUC₂₄/MIC ratio of vancomycin near 400 could significantly decrease mortality rates by 53% (8). However, trough plasma concentrations (C_ts) are routinely used as a surrogate parameter for AUC in the clinical setting. A C_t of between 10 and 20 μg/mL is recommended to obtain an AUC₂₄/MIC ratio of ≥400 when treating pathogens with an MIC of ≤1 mg/L (9). Nevertheless, it should be considered that these concentrations could vary depending on the type, location, and severity of the infection (10). A recent update to the Infectious Diseases Society of America (IDSA) guidelines advocates monitoring based on AUC₂₄ instead of C_ts (11). However, to calculate the AUC, it is necessary to have at least two samples, which makes its application difficult in the neonatal population.

These PK/PD parameter targets have not been validated in the neonatal population as there are no available studies specifically designed to assess the correlation between vancomycin PK/PD parameters and clinical and microbiological efficacy against invasive infections in this population. Therefore, the plasma values have been extrapolated from adults and for MRSA infections (12–14). However, there is great controversy regarding whether these levels are the most appropriate to achieve the greatest safety and efficacy in this group of subjects, which differs enormously from the adult population (15–17).

The main goal of this study was to analyze the therapeutic vancomycin regimens, AUC₂₄, and C_t obtained that achieved clinical and microbiological effectiveness in a cohort of neonatal patients from the neonatal intensive care unit (NICU).

RESULTS

Study participants and clinical evolution.

A total of 43 patients were included. Twenty-five were classified into the favorable group and 18 were classified into the unfavorable group according to early clinical evolution (ECE). The baseline demographic and clinical characteristics of the study population are summarized in Table 1. No statistically significant differences were found between the groups.

TABLE 1.

Baseline demographic and clinical characteristics of the study population^a

Characteristic	Value for group			P value
	Total (n = 43)	Early clinical evolution
		Favorable (n = 25)	Unfavorable (n = 18)
Median GA (wks) (IQR)	30 (28–35)	30 (28–35)	29.5 (26.75–36.5)	0.684
Median PMA (wks) (IQR)	34 (30–39)	35 (31–38.5)	31 (29.75–41.25)	0.342
Median PNA (days) (IQR)	14 (9–34)	14 (9.5–38.5)	14.5 (8.75–26.5)	0.666
Median wt (kg) (IQR)	1.94 (1.25–2.90)	2.03 (1.22–2.61)	1.30 (0.86–3.35)	0.445
Median size (cm) (IQR)	41.5 (36–45.8)	42.5 (35.5–43)	40.5 (34.5–52.3)	0.686
No. of patients of sex (%)
Female	19 (44.2)	10 (40)	9 (50)	0.515
Male	24 (55.8)	15 (60)	9 (50)
No. of patients with primary diagnosis (%)
Prematurity with complications	30 (69.8)	18 (72)	12 (66.7)	0.919
Esophageal/intestinal atresia or intestinal malformation	6 (14)	3 (12)	3 (16.7)
Pulmonary atresia	2 (4.7)	1 (4)	1 (5.6)
Gastroschisis	2 (4.7)	1 (4)	1 (5.6)
Bronchiolitis	1 (2.3)	1 (4)	0 (0)
Others	2 (4.7)	1 (4)	1 (5.6)
No. of patients with type of infection (%)
Nosocomial sepsis/bacteremia	33 (76.7)	21 (84.0)	12 (66.7)	0.185
Catheter-related infection	3 (7.0)	0 (0)	3 (16.7)
Meningitis	1 (2.3)	1 (4)	0 (0)
Superinfected pneumothorax	1 (2.3)	0 (0)	1 (5.6)
Vertical sepsisb	1 (2.3)	1 (4)	0 (0)
Others	4 (9.3)	2 (8.0)	2 (11.1)
No. of patients with pathogen(s) isolated (%)
Staphylococcus epidermidis	28 (71.8)	15 (68.2)	13 (76.5)	0.521
Staphylococcus haemolyticus	5 (12.8)	4 (18.2)	1 (5.9)
Streptococcus mitis	2 (5.1)	1 (4.5)	1 (5.9)
Staphylococcus hominis	2 (5.1)	1 (4.5)	1 (5.9)
Enterococcus faecium	1 (2.6)	1 (4.5)	0 (0)
S. epidermidis and S. hominis	1 (2.6)	0 (0)	1 (5.9)
No. of patients, median MIC (mg/L) (IQR)	37, 1 (1–1)	21, 1 (0.63–1)	16, 1 (1–1)	0.262

^aGA, gestational age; PMA, postmenstrual age; PNA, postnatal age. Quantitative variables are expressed as numbers (percentages) or medians (interquartile ranges).

^bLife-threatening organ dysfunction caused by dysregulation of the host with infection. This infection is caused by germs located in the maternal genital canal.

The 25 patients who evolved favorably at the start of antimicrobial treatment achieved clinical cure. Thus, the remaining 18 patients were included in the late clinical evolution (LCE) analysis. Thirteen of the 18 patients were cured (median treatment duration of 18 [interquartile range {IQR}, 10 to 20] days), and 5 patients had unfavorable evolution. No significant differences were found in any of the parameters evaluated (see Table S2 in the supplemental material). In 4 of these patients, it was necessary to modify the antibiotic therapy to daptomycin (n = 2) and linezolid (n = 2) due to the persistence of positive blood cultures. The remaining patient died of necrotizing enterocolitis.

The cure rate in the overall cohort was 88.4% (n = 38), with a median treatment duration of 11 (IQR, 9.25 to 14) days.

Characteristics of treatment with vancomycin.

Vancomycin was initially administered as an intermittent infusion without a loading dose in all patients, as specified by Neofax guidelines. Nevertheless, 9.3% (n = 4) of patients switched to a continuous intravenous infusion during treatment due to a lack of clinical improvement and/or low plasma concentrations. Two of these patients belonged to the favorable group and the remaining two belonged to the unfavorable group according to ECE. The latter presented favorable evolution in the LCE analysis.

The vancomycin dosage was determined based on weight, postmenstrual age (PMA), and postnatal age (PNA). Thus, the initial dose of vancomycin analyzed in the ECE was 30 (IQR, 20 to 33.45) mg/kg of body weight/day in the favorable group, versus 21 (IQR, 20 to 31.50) mg/kg/day in the unfavorable group (P = 0.34).

In the LCE analysis, the median dose administered after making the corresponding dose adjustments was 40 (IQR, 30 to 60) mg/kg/day in the favorable group, versus 39 (IQR, 30 to 49.5) mg/kg/day in the unfavorable group (P = 0.841).

Regarding the safety of vancomycin, no adverse effects were recorded. It should be noted that 34.89% of patients (n = 15) were being treated with a nephrotoxic drug (data not shown). However, no patient experienced nephrotoxicity.

Pharmacokinetic outcomes.

Regarding the pharmacokinetic parameters of the total cohort (n = 43), the following values were obtained: a median volume of distribution (V_d) of 0.844 (IQR, 0.819 to 0.877) L/kg, a median total clearance (CL) of 0.19 (0.09 to 0.29) L/h, and a median half-life (t_1/2) of 7 (IQR, 5.25 to 8.32) h.

The median C_t values obtained in ECE were 9.90 (IQR, 7.30 to 13.40) μg/mL and 6.15 (IQR, 4.5 to 8.85) μg/mL in the favorable and unfavorable groups, respectively (P = 0.002). The AUC₂₄ of vancomycin was higher in the favorable group (332.26 [IQR, 269.41 to 380.63] mg · h/L) than in the unfavorable group (228.40 [IQR, 171.34 to 395.79] mg · h/L) (P = 0.001) in ECE (Fig. 1).

FIG 1.

Comparison of vancomycin AUC₂₄ values, C_t values, and doses based on the early clinical evolution of patients.

In the LCE analysis, the median C_t obtained in cured patients was higher than that in those who ultimately failed treatment, at 14.20 (IQR, 11.75 to 17.15) μg/mL versus 8.60 (IQR, 7.15 to 12.60) μg/mL, respectively (P = 0.023). However, the median AUC₂₄ values were similar between patients, being 392.34 (IQR, 346.94 to 419.35) mg · h/L for cured patients and 363.92 (IQR, 307.69 to 395.45) mg · h/L for patients whose treatment failed (P = 0.654) (Fig. 2).

FIG 2.

Comparison of vancomycin AUC₂₄ values, C_t values, and dose according to the late clinical evolution of patients.

Association between clinical evaluation and pharmacokinetic parameters.

To assess the prognostic value of vancomycin C_t and AUC₂₄, receiver operating characteristic (ROC) analysis was performed at the times of ECE and LCE.

In the ECE analysis, the AUC₂₄ area under the ROC curve was 0.796 (95% confidence interval [CI], 0.661 to 0.930), and 238 mg · h/L was the optimal cutoff value to predict favorable evolution (Fig. 3a), with a sensitivity of 61% and a specificity of 88%. The positive predictive value (PPV) and negative predictive value (NPV) were 79% and 77%, respectively. For C_t, the area under the ROC curve was 0.774 (95% CI, 0.629 to 0.920), with an optimal cutoff value of 6.8 μg/mL. The sensitivity was 61%, and the specificity was 92%, with PPV and NPV values of 85% and 77%, respectively.

FIG 3.

(a) ROC curves of AUC₂₄ and trough concentrations in ECE. (b) ROC curves of AUC₂₄ and trough concentrations in LCE.

In the LCE analysis, the ROC curve showed that the area under the curve of the C_t was 0.862 (95% CI, 0.616 to 1.000), establishing a cutoff value of 11 μL/mL (Fig. 3b). The sensitivity was 80%, and the specificity was 92%, with a PPV of 80% and an NPV of 92%.

In this analysis, the AUC₂₄ area under the ROC curve was 0.600 (95% CI, 0.275 to 0.925), obtaining a cutoff point of 334 mg · h/L.

For the ECE analysis, a multivariate analysis was performed to evaluate the predictive value of the AUC₂₄ and C_t of vancomycin adjusted for sex, weight, and PMA (Table 2). Logistic regression showed that a vancomycin C_t of ≤6.8 μg/mL was associated with an unfavorable ECE (P = 0.001), being 18 times more likely to progress poorly compared to those with higher levels. No statistically significant results were obtained for the rest of the variables analyzed.

TABLE 2.

Multivariate relationship between ECE and vancomycin C_t values^a

Parameter	Mean B (SE)	Wald value	P value	OR	95% CI of OR
Vancomycin Ct of ≤6.8 μg/mL	2.894 (0.882)	10.778	0.001*	18.071*	3.210–101.72
Model constant	1.190 (0.432)	7.594	0.006	0.304

^aLogistic regression was done using the Wald forward method. Nagelkerke’s R² value was 0.390. B, β coefficients; *, statistically significant differences.

DISCUSSION

To the best of our knowledge, this is the first study analyzing the correlation between vancomycin PK/PD parameters and clinical and microbiological effectiveness in neonatal patients. Our results provide an approximation of the adequacy of currently recommended PK/PD targets and a proposal for new cutoff points specific to the neonatal population.

International guidelines (11) recommend AUC-guided therapeutic drug monitoring and dose individualization to optimize vancomycin therapy. The recommended AUC₂₄ target of 400 mg · h/L has been extrapolated from the adult population and for MRSA infections assuming an MIC of ≤1 mg/L (18). According to previous reports (12, 19–21), CoNS were the most common pathogens isolated from the neonatal population (83.7% in our cohort) and were the main cause of late-onset neonatal sepsis requiring vancomycin therapy, with limited cases of MRSA sepsis (none in our study). Despite this, information on vancomycin PK/PD against CoNS is limited. In our cohort, the CoNS MIC distribution was similar to the one previously described for MRSA in adults (median of 1 [IQR, 1 to 1] mg/L) (7), so the AUC target might be similar. However, an experimental study (22) suggested a higher AUC₂₄/MIC ratio target (520 to 665 mg · h/L) for CoNS infection in neonates than that currently proposed for MRSA infection in adults. This study includes in vitro and in vivo experiments, although these results have not been validated in clinical studies. Our results showed that both AUC₂₄ and C_t could be considered acceptable prognostic indicators of ECE in a neonatal population with a high proportion of CoNS infections. Nevertheless, C_t was the best prognostic indicator of LCE.

Vancomycin therapy optimization in daily practice using AUC₂₄ presents certain difficulties. AUC₂₄ monitoring requires collecting samples to properly describe the vancomycin distribution and elimination, preferably integrated within a Bayesian approach (11). This is not a common practice due to the difficulty in extracting blood samples from the neonatal population. On the contrary, C_t might be obtained by taking advantage of routine blood tests, avoiding additional venipunctures. This may be beneficial to prevent blood sampling errors and, in premature or critically ill infants, to minimize iatrogenic anemia caused by repeated blood sampling (19, 23). After analyzing ECE, AUC₂₄ values of >238 mg · h/L were found to have a PPV of 79%. The ROC curve showed that the AUC₂₄ value had an influence on ECE. However, these results did not hold for LCE, where AUC₂₄ values were similar in patients with favorable and unfavorable evolution.

To address this issue, vancomycin C_t has been proposed as a surrogate parameter of AUC₂₄. However, recent studies (24, 25) have shown that this assumption may not be optimal in neonates. The results of our study suggest that C_t values of >6.8 μg/mL are sufficient to achieve favorable ECE in most cases, with a PPV of 85%. In addition, the multivariate analysis showed that vancomycin C_t is the best predictor of ECE, with an odds ratio (OR) value of 18. Previous studies (12, 14, 16, 20, 26) proposed a significantly higher target C_t (10 to 20 μg/mL). Nevertheless, none of those studies correlated these C_t values with clinical results, and how these higher levels might affect long-term safety in this population is unknown. In accordance with our results, another observational study (19), in which patients were also dosed according to Neofax guidelines, obtained a median C_t of 8 μg/mL (range, 5 to 10.5 μg/mL). In this group of patients, 80% of cultures were negative at 48 h, and no patient had clinical failure. Thus, after evaluating ECE, it could be concluded that it is not necessary to reach a high C_t to achieve clinical cure since optimal clinical results would be achieved with vancomycin concentrations of around 7 to 8 μg/mL in neonates.

In patients with unfavorable ECE, it was necessary to achieve a C_t of at least 11 μg/mL, with a sensitivity of 80%, a specificity of 92%, a PPV of 80%, and an NPV of 92%. These results are of great clinical interest since they allow us to guide the target plasma concentrations that should be achieved in this group of patients with great certainty. However, in order to be able to extrapolate these results to clinical practice, its external validation would be convenient given that the analysis was carried out in a sample of 18 patients who presented high variability (the microorganisms involved and the sources of infection, etc.).

Finally, it is clear that the dosage regimen plays a very important role in the efficacy of the treatment. Most studies proposing a vancomycin dose regimen in this population evaluated only C_t achievement and did not correlate these concentrations with clinical results (12, 14, 16, 20). Therefore, the dose regimen that achieves the best clinical results has not yet been established (17). In our study, patients who evolved favorably at the beginning of the antimicrobial treatment received a median dose of 30 mg/kg/day, whereas those who evolved unfavorably received a median dose of 21 mg/kg/day. Although statistically significant differences were not obtained for the latter, clinical relevance should be emphasized. The dose guidelines at our center are determined according to PMA and PNA, and patients who received lower vancomycin doses (13.3 to 20 mg/kg/day) were neonates with a lower PMA (mainly ≤29 weeks). This suggests that this group of patients could be underdosed according to the recommendations of Neofax guidelines, which might translate to a higher risk of unfavorable evolution, in accordance with the results of this study. In this sense, Ramos-Martín et al. performed simulations in an experimental model of neonatal central-line-associated bloodstream CoNS infections predicting that neonates with a PMA of <29 weeks are also underdosed with standard regimens (15 mg/kg/day) (22). The authors of that study propose a regimen of 30 mg/kg/day to achieve reductions in C-reactive protein (CRP) concentrations similar to those achieved in older neonates and infants (22). Another study in neonates (19) demonstrated that a higher rate of vancomycin dose adjustment is needed to achieve clinical efficacy in patients with a gestational age of <29 weeks. Recently, the dosing regimens in the Pediamecum guidelines (27) have been updated, recommending dosing preterm infants and children <1 month of age with loading doses of 15 mg/kg, followed by a regimen of 10 mg/kg every 12 h during the first week and every 8 h until 1 month of age.

In terms of safety, a recent systematic review also supported the favorable safety profile of vancomycin in neonates (17). One study (28) proposed that nephrotoxicity was not associated with vancomycin alone but may occur in the presence of other recognized risk factors. In our population, despite most patients presenting at least one of these circumstances, no patient experienced nephrotoxicity, and creatinine values were not affected, demonstrating the safety of vancomycin in this population.

The major limitation of our study was calculating the AUC₂₄ theoretically using only one blood sample, which might have influenced the results. Other limitations are the technique used to measure the plasma concentrations of vancomycin (total concentrations) instead of the unbound fraction and the limited sample size of the study. Larger prospective studies and randomized clinical trials are necessary to consolidate the results obtained.

Conclusion.

In conclusion, this is the first study to offer vancomycin PK/PD evidence against CoNS in the neonatal population. Our study has shown that both AUC₂₄ and C_t are good predictors of ECE in this population. Concentrations close to 7 μg/mL and an AUC₂₄ of around 240 mg · h/L 48 h after antibiotic initiation seem to be sufficient to achieve clinical cure in most cases in the cohort studied, contrary to the evidence published so far. However, further studies to validate these results in the neonatal population are warranted.

MATERIALS AND METHODS

Setting and study design.

This was an observational, prospective, single-center study covering a period of 2 years (2019 to 2021), which was conducted in the NICU of an academic tertiary care hospital equipped with 14 beds.

The study was conducted in accordance with the Declaration of Helsinki of the World Medical Association. This study protocol was reviewed and approved by the Ethics Committee of the Virgen del Rocío-Macarena University Hospital (FIS-VAN-2022-01).

Participants.

Eligible patients were neonates and young infants (from birth to 3 months of age) who were admitted to the NICU and were undergoing treatment with intravenous vancomycin with or without another antibiotic (directed or empirical). All patients were required to receive vancomycin treatment for ≥72 h with ≥1 C_t available. Patients were excluded if they were older than 3 months of age (chronological in term neonates and postmenstrual age in preterm neonates) or if they received vancomycin treatment through means other than intravenous perfusion (orally, lock therapy for catheter infection, or intraperitoneally).

Follow-up, assessment, and endpoints.

Sociodemographic, clinical, and laboratory parameter variables for all patients were recorded. In addition, variables related to vancomycin treatment and concomitant pharmacological treatment were collected, as were all pharmacokinetic parameters (volume of distribution [V_d], total clearance [CL], and median half-life [t_1/2]).

(i) Procedure.

All patients were initially dosed according to Neofax guidelines (29). Vancomycin TDM was scheduled according to the center’s protocol (30), developed by a multidisciplinary team consisting of intensivists and pharmacists. In those patients whose treatment was expected to last more than 3 days, the C_t was obtained 48 to 72 h after the start of antibiotic therapy. Neonates were classified into two groups, favorable and unfavorable, according to early clinical evolution (ECE). Favorable ECE was defined as fever reduction, an improvement in infectious symptoms and signs, good general condition, or a clear (>30%) or continuous reduction in C-reactive protein (CRP), evaluated 48 to 72 h after the initiation of vancomycin treatment. If the patient had a control blood culture at 72 h, this had to be negative in order to consider favorable ECE. On the contrary, unfavorable ECE was defined as the persistence or worsening of the patient’s infectious symptoms, the persistence of positivity in control blood cultures, or an increase in/maintenance of high CRP values evaluated 48 to 72 after the initiation of vancomycin treatment (19, 31).

The center’s protocol recommended not modifying the vancomycin dose in patients with favorable ECE, except for safety concerns due to the risk of toxicity from high concentrations of vancomycin (≥25 μg/mL). In patients with unfavorable ECE, the dose was optimized using Abbottbase Pharmacokinetics Systems (PKS) software, taking into account the MIC of the isolated microorganism and the location of the infection. C_t testing was requested 48 to 72 h after the start of the new regimen, again assessing the patient’s clinical evolution. The last step was repeated as many times as necessary until a favorable outcome (clinical and microbiological cure) or an unfavorable outcome (change of treatment due to inefficacy or death) was obtained. This variable was designated late clinical evolution (LCE).

(ii) Vancomycin measurement and AUC₂₄ estimation.

Microtainer tubes (citric acid-dextrose) were used to collect the blood samples. The vancomycin C_t was measured using a homogeneous enzyme immunoassay technique on the Roche Cobas 8000 c702 system (Roche Diagnostics, Mannheim, Germany) in the hospital clinical laboratory. MIC values were determined by broth microdilution (BMD).

The pharmacokinetic analysis was performed using PKS software (version 1.10; Abbott Diagnostics Division, Irving, TX, USA), adjusting experimental data according to a monocompartmental linear model for intravenous delivery using Bayesian estimates. Using this approach, dose optimization and pharmacokinetic parameters were predicted (see Text S1 in the supplemental material), except for the AUC₂₄, which was calculated theoretically with the quotient of the total dose/day and the patient’s clearance (32).

(iii) Assessed outcomes.

The primary outcome was the association between vancomycin C_t and clinical and microbiological efficacy at the beginning (ECE) and the end (LCE) of treatment with vancomycin. The association between AUC₂₄ and clinical and microbiological efficacy was also analyzed.

Safety was assessed as the occurrence of nephrotoxicity, defined as an increase in serum creatinine (S_Cr) of ≥0.5 mg/dL or a 50% increase from the baseline S_Cr level, or the appearance of other adverse effects related to the treatment.

Statistical analyses.

The results were expressed as median values and interquartile ranges (IQRs) for continuous variables and as numbers and percentages of cases for categorical variables. Categorical variables were compared using the χ² test or Fisher’s exact test. Quantitative variables were analyzed using Student’s t test or the Mann-Whitney nonparametric test, according to their distribution. Receiver operating characteristic (ROC) curve analyses were performed to assess the accuracy of the AUC₂₄ and vancomycin C_ts in predicting a patient’s evolution, using the Youden index to choose the best cutoff point. Multivariate logistic regression analysis was used to determine the relationship between AUC₂₄ cutoff points and vancomycin C_ts and patient outcomes.

Statistical analyses were performed using SPSS software (version 28.0; IBM, Chicago, IL) and R software (R Foundation for Statistical Computing, Vienna, Austria). P values of <0.05 were considered significant.

Data availability.

All data generated or analyzed during this study are included in this article and its supplemental material. Further inquiries can be directed to the corresponding author.