Correlation between 61 risk factors and vancomycin blood concentration reached based on real world data

Background : Vancomycin, an antibiotic produced by microbial fermentation, is used to treat gram-positive infections. With extensive broad-spectrum antibiotics use, bacterial resistance to vancomycin continues to increase. We explored the factors influencing vancomycin blood concentration compliance rate, established a predictive equation for such concentrations, and provided a reference for individual applications. Methods: We used a single-center, retrospective, case-control study design based on real-world data from the Hospital Information Management System of the First Affiliated Hospital of the Teaching Hospital of Bengbu Medical School from January 1, 2017 to December 31, 2018. Inpatients whose vancomycin blood concentration was monitored were selected. Single factor and multivariate logistic regression analyses were performed using SPSS 21.0 software to screen factors affecting, vancomycin blood concentration compliance rate. Vancomycin blood concentration was then determined. A receiver operator characteristic (ROC) curve of influencing factors was then used to establish a prediction model. Results: In total, 168 patients (122 males and 46 females) were enrolled. Eighty-one had their vancomycin blood concentration monitored, and 87 had concentrations that did not reach the standard. Multivariate logistic model analysis showed that patient drug allergy history, alanine aminotransferase (ALT), aspartate aminotransferase (AST), patient infusion volume, and urine volume influenced vancomycin blood concentration compliance rate. According to logistic model analysis, a history of drug allergy (95% CI: 1.225-24.850, P<0.05), ALT (95% CI: 0.979-0.999, P<0.05), AST (95% CI: 1.003-1.027, P<0.05), patient infusion volume (95% CI: 0.996-0.998, P<0.05), patient urine volume (95% CI: 1.001-1.003, P<0.05), and combined predictors were used to construct ROC curves. For combined predictive factors, area under the ROC curve (0.829, 95% CI: 0.755-0.902, P<0.005) was greater than that of the five individual indicators and was predicted to be the preferred value. Conclusions: In clinical practice, the patient’s daily average infusion and urine volumes can be substituted into the joint predictor calculation formula: P=1/[1+e-(0.940-0.004*X infusion +0.002*X urine )]. By calculating the combined predictive factor to predict vancomycin blood concentration compliance rate, the dosing regimen can be adjusted.


Background
Vancomycin (VAN), a natural antibiotic produced by microbial fermentation, is a tricyclic glycopeptide antibiotic and was the first glycopeptide antibiotic [1]. Clinically, it is used to treat gram-positive bacterial infections, such as those caused by hemolytic streptococcus, or pneumococcus,among others. VAN is one of the first-line drugs for the treatment of methicillin-resistant Staphylococcus aureus (MRSA) and related infections, as directed by many authoritative guides [2]. Differences in the blood concentration of VAN have been observed in different individuals. If its blood concentration is too low, the drug might not be effective or might not reach its effective trough concentration. However a concentration that is too high can lead to adverse drug reactions, such as ototoxicity and nephrotoxicity [3,4]. If administered according to the recommended dosage found in the instructions, a significant portion of the patient's blood concentration does not appear to be within the therapeutic window [5][6][7]. Thus, it is clinically necessary to monitor VAN blood concentrations. Methods for this mainly include high performance liquid chromatography, radioimmunoassay, and enzyme expanded immunoassay [8]. Since VAN is a time-dependent antibacterial drug, the key to ensuring its efficacy is extending the antibacterial drug infusion time such that the drug concentration is always higher than the target organism's minimum inhibitory concentration [9]. However, routine monitoring of peak concentration [10] does not correlate with monitoring efficacy. Furthermore, as the distribution of VAN into tissues is slow, its peak concentration is difficult to monitor [1].
Some studies [11] suggest that the peak concentration of VAN does not correlate well with its efficacy and its concentration can be used as a surrogate for its area under the curve, 4 which currently is widely accepted. According to the Guide to the Monitoring of Chinese Vancomycin Therapeutic Drugs [12], patients with normal renal function have a VAN halflife of 6-12 h, with stable blood concentration achieved after 4 to 5 half-lives. On the third day (48 h after the first dose), monitoring of patient VAN can be initiated. However, for patients with renal insufficiency, upon first administration, monitoring is initiated 72 h after drug administration. Accurately determining VAN trough concentrations can not only avoid poor therapeutic effects due to drug resistance or adverse reactions caused by concentrations that are too low or too high, but also help clinicians develop individualized drug delivery plans for patients, as revealed by Lijun and other researchers [13][14][15][16][17][18][19].
Xingru et al. [5,20,21] reported that VAN blood concentration is related to age, serum albumin, dose, dose, urine volume, creatinine clearance, serum creatinine, blood urea nitrogen, and plasma albumin in critically ill patients. Moreover, VAN blood concentrations in children are related to serum creatinine, alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatinine clearance, glutamyl transaminase, and gestational age.
Similar to other reports, Qingrong and other researchers [22] pointed out that the VAN plasma concentration in patients ≥18 years of age is related to age, albumin, liver and kidney function, usage, and dose. Although the aforementioned factors have been widely discussed as influencing factors related to VAN blood concentration, only a few reports have discussed them with regard to VAN plasma concentrations. In addition, the research populations used for these analyses were relatively small (e.g., studies on children only), the entire population was not analyzed, and a prediction model has not been established.
In this paper, a real-world single-center, retrospective, case-control study was conducted to investigate the effects of age, sex, liver and kidney function, and solvent on the plasma concentration of VAN. The factors involved were found to be more extensive and more comprehensive. By combining single factor analysis and multi-factor logistic analysis, a receiver operating characteristic (ROC) curve was generated and the prediction equation for VAN blood concentration was fitted to provide a basis for clinical pharmacists to guide clinicians in the use of this drug.

Research design
We performed a single-center, retrospective, case-control, research design study[The First Affiliated Hospital of Bengbu Medical College, the ethics committee of clinical medical research (scientific research project) approval number: 2018KY008, the approval date was August 28, 2018; and on June 23, 2019, the registration number of the China Clinical Trial Registration Center was ChiCTR1900024062.]based on real-world data present in the Hospital Information Management System (HIS). HIS is a database that stores the information recorded during an inpatient's stay at the hospital. It includes their basic information, daily medication during hospitalization, care, admission, and expenses incurred. According to the Guide to the Monitoring of Chinese Vancomycin Therapeutic Drugs [12], when patient blood concentration was monitored, VAN had to be within the effective concentration range (10-20 μg/mL). The population was divided into the standard group, including patients whose blood concentrations, as measured after the intake of VAN, were within the range of 10-20 μg/mL, and the substandard group, including patients with VAN trough concentrations outside the range of 10-20 μg/mL.

Blood drug concentration monitoring method
The VAN blood concentration was monitored by enzyme amplification immunoassay, using enzyme reagent 1 (vancomycin labeled with bacterial glucose-6-phosphate dehydrogenase

Research object
This study was performed at the First Affiliated Hospital of Bengbu Medical College from January 1, 2017 to December 31, 2018. Inpatients whose VAN blood concentrations were being monitored were selected as subjects. The study protocol was reviewed and approved by the hospital ethics committee. Inclusion criteria were as follows: (1) patients treated with VAN for clinical infections who had their VAN blood concentration monitored; (2) trough concentration of VAN blood concentration and not peak concentration or concentration found immediately after administration; (3) influencing factors of the control group had to be consistent with the influencing factors in the observation group; (4) complete patient data. Exclusion criteria were as follows: (1) vancomycin was not used as the clinical treatment for infection; however, cases treated with norvancomycin were included; (2) patients whose VAN  Data processing and assignment Univariate and multivariate logistic regression analyses were used to screen independent risk factors for the compliance of patients receiving VAN. According to the information retrieved for each patient, the assignment of various influencing factors is shown in Table   1.
[Insert Table 1] Statistical analysis Data were processed using SPSS 21.0 statistical software (IBM, Armonk, NY, USA). A Pearson's chi-square test was used to analyze classification data, whereas a t-test was used to analyze continuous data. Univariate risk factors affecting VAN plasma concentration were determined by single factor and multivariate regression analyses.
According to the multivariate logistic model analysis results, joint predictors were set and included in the logistic model for analysis. Model equation and ROC curves were then derived. Curve, screen cut points, area under the curve, and the predictive efficacy of various indicators were then derived. All tests were two-sided, and P < 0.05 was considered statistically significant. Unless otherwise stated, results are expressed as the mean ± standard deviation, and lost data is processed using the mean substitution method.  Table 2. Univariate analysis of compliance and non-standardized data 8 VAN blood concentration was found to adhere to the standard. In addition, patients's drug allergy history, rheumatoid arthritis, combined use of human immunoglobulin, hospital days, AST, G, daily average infusion of patients, and urine volume showed statistical differences (to avoid missing important factors in the single factor analysis process, The range of P values was set to< 0.1, to include more potential influencing factors, see Table   3).

Results
[Insert Table 3] Multi-factor analysis of compliance and non-standardized data Multivariate logistic regression analysis showed that allergic history, ALT, AST, daily average infusion volume, and daily mean urine volume were significantly associated with patient VAN blood concentration (P < 0.05, Table 4). The general trend of VAN blood concentration as a function of ALT, AST, daily average infusion volume, and daily average urine volume is presented in Figures 1-3.

VAN blood concentration standard prediction index
With the extensive use of broad-spectrum antibiotics, the severity of bacterial resistance to VAN continues to increase [23]. Because creatinine clearance was found to increase by 26.5 μmol/L or 50% within 14 days after VAN use, and/or because urine volume was <0.5 ml/(kg·h) in more than 6 h [24], kidney damage was determined. Therefore, urine volume was identified as an important factor. In this study, daily average infusion volume of patients and daily average urine volume were independent factors, as reported by Jiyao [18]. Typically VAN is not considered to be metabolized in the body, and at 24 h, its level is maintained at 90% for elimination via the kidney by the prototype [25]. However, because of different treatment options and doses required by different patients, a previous study could not report a stable clearance rate for clinical reference [26,27].
Infusion volume and urine volume serve as factors that can affect the distribution volume and clearance rate of VAN in patients, thereby affecting its concentration. By fitting the prediction equation, patient's daily average infusion volume and daily average urine volume were substituted into the formula to predict their VAN blood concentration compliance rate and facilitate the timely adjustment of their medication regimen during treatment. When the ratio of the daily average infusion volume to the daily average urine volume was in the range of 0.35-1, a high VAN blood concentration was found; when the ratio was in the range of 0.48-1, VAN blood concentration was at its highest, and the rest of the concentrations were either higher or lower than the normal value; Further, when the ratio was greater than 1 (only three cases in this study), infusion volume was found to be greater than urine volume. Subsequently, a higher utilization rate of pigment led to a corresponding reduction in free VAN and an increase in the trough concentration.
However, it cannot be excluded that urine volume still contained a large amount of free VAN when patient input (including the amount of daily intake such as infusion volume and drinking water) is greater than the infusion volume.

Conclusions
In summary, although the dosing regimen and monitoring procedures for VAN have been largely established via pharmacodynamic and pharmacokinetic studies, personalized drug regimens are still needed for specific patients. In the present study, the combined 11 predictive factor formula for patient's daily average infusion volume and urine volume could predict the compliance rate of VAN blood concentration. This will allow clinicians to identify problems and adjust the medication plan to promote the safe and effective use of VAN [39]. ChiCTR1900024062. At the same time, the informed consent is not signed because this study is a low-risk study and meets the IRB exemption informed consent requirements:

Abbreviations
Research using anonymous or no-risk tests, surveys, interviews, or observations.

Consent to publish
Not Applicable Availability of data and materials The dataset analyzed during the current study is available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.    Relationship between Alanine aminotransferase (ALT) and blood serum.

Figure 3
Relationship between ratio of infusion to urine and blood serum.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.