Tacrolimus blood concentration prediction method based on multi-factor regression analysis

Abstract

The invention discloses a tacrolimus blood concentration prediction method based on multi-factor regression analysis, which comprises the following steps: step 1, clinical data acquisition, blood concentration monitoring and patient grouping comparison, step 2, obtaining independent risk factors by adopting multi-factor Logistic regression analysis, step 3, establishing a column diagram risk prediction model, and step 4, establishing a tacrolimus risk prediction model. 4, evaluating the column graph risk prediction model; the risk that the postoperative tacrolimus blood concentration of the kidney transplantation patient does not reach the standard can be effectively predicted through the constructed column diagram risk prediction model, the model has good accuracy and distinguishing ability, meanwhile, the model has good calibration ability, the situation that the tacrolimus blood concentration of the kidney transplantation patient does not reach the standard can be accurately predicted, and the risk of clinical application is effectively predicted. The clinical application value is relatively high.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical technology, and in particular to a method for predicting tacrolimus blood concentration based on multi-factor regression analysis. Background Art

[0002] Tacrolimus (FK506) is a drug with a powerful immunosuppressive effect. It forms a complex by binding to FK506 binding protein-12 in T lymphocytes, inhibiting the activation of calcineurin, thereby interfering with the synthesis and expression of related cytokines, reducing cytotoxicity, lowering the risk of acute rejection reactions, and improving the survival rate of transplanted organs.

[0003] However, FK506 has a narrow therapeutic index and is greatly affected by pharmacokinetics and pharmacodynamics. It has significant individual differences in biological metabolism. Too low a blood concentration may lead to treatment failure or an increased risk of drug resistance, while too high a blood concentration may increase nephrotoxicity and neurotoxicity and reduce the survival time of organ transplant patients. Previous studies have shown that the blood concentration of FK506 is mainly related to demographic factors, medication dosage, and genetic polymorphism. Among them, CYP3A5 genetic polymorphism is an important factor affecting the blood concentration of FK506. CYP3A5-expressing individuals usually require a higher dose to achieve the same blood concentration as non-expressing individuals. However, there are large differences in research conclusions between different reports, and there is a lack of research on early prediction of substandard FK506 blood concentration in renal transplant patients after surgery. Therefore, the present invention proposes a tacrolimus blood concentration prediction method based on multivariate regression analysis to solve the problems existing in the prior art. Summary of the invention

[0004] In view of the above problems, the purpose of the present invention is to propose a method for predicting tacrolimus blood concentration based on multivariate regression analysis. The method for predicting tacrolimus blood concentration based on multivariate regression analysis can provide clinicians with a practical tool to help them better manage the use of tacrolimus (FK506) and improve the treatment effect and safety of patients.

[0005] To achieve the purpose of the present invention, the present invention is implemented by the following technical scheme: a method for predicting tacrolimus blood concentration based on multivariate regression analysis, comprising the following steps: Step 1: Obtain clinical data of patients taking tacrolimus after renal transplantation and monitor tacrolimus blood concentrations. According to the blood concentration data, divide the patients into a target group and a non-target group, and use a chi-square test to compare the groups to obtain comparison data. Step 2: Based on the comparison data, the blood drug concentration that does not meet the standard is set as the dependent variable, and the factors with statistical differences in the clinical data analysis are selected as independent variables. Multivariate Logistic regression analysis is used to obtain the independent risk factors for blood drug concentration that does not meet the standard; Step 3: Based on the identified independent risk factors, a nomogram risk prediction model was established to predict the risk of tacrolimus blood concentration not being within the therapeutic range in renal transplant patients, and the stability and predictive performance of the model were verified and evaluated using cross-validation and bootstrap methods; Step 4: Use HL goodness-of-fit test, ROC curve and consistency index to evaluate the nomogram risk prediction model.

[0006] A further improvement is that the clinical data in step 1 include age, gender, body weight index, fasting blood glucose, triglycerides, total bilirubin, white blood cell count, red blood cell count, albumin, hemoglobin, platelet count, aspartate aminotransferase, alanine aminotransferase, blood creatinine and blood urea nitrogen.

[0007] Further improvements are as follows: in the step 1, the blood drug concentration monitoring is specifically performed by using an enzyme amplification immunoassay method for blood pretreatment, and then a fully automatic biochemical analyzer is used for detection to determine the tacrolimus blood drug concentration; the patient grouping is specifically performed by dividing the patients with a blood drug concentration of 5-15 ng / ml into a target group, and vice versa into a non-target group.

[0008] A further improvement is that the independent variable in step 2 includes fasting blood sugar exceeding Total bilirubin exceeds , white blood cell count less than , hemoglobin exceeds and serum creatinine exceeds .

[0009] A further improvement is that the nomogram model in step three specifically uses the scikit-learn database to build a regression model, uses the statsmodels database to extract coefficients, and finally uses the matplotlib database to retrieve the nomogram to build a nomogram risk prediction model.

[0010] The beneficial effects of the present invention are as follows: the nomogram risk prediction model constructed by the present invention can effectively predict the risk of substandard tacrolimus blood concentration in renal transplant patients after surgery. The model has good accuracy and discrimination ability, and at the same time has good calibration ability, and can accurately predict the situation where the tacrolimus blood concentration in renal transplant patients does not meet the standard. It has high clinical application value, provides a practical decision support tool for clinicians, and helps to improve the treatment effect and safety of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 This is a flow chart of the method of Example 1 of the present invention.

[0012] Figure 2 This is a nomogram model diagram for predicting that the tacrolimus blood concentration in renal transplant patients will not meet the standard in Example 2 of the present invention.

[0013] Figure 3 This is a schematic diagram of the confidence interval of the ROC curve of the nomogram of the tacrolimus blood concentration substandard model in Example 2 of the present invention.

[0014] Figure 4 This is a ROC curve validation diagram of the nomogram of the tacrolimus blood concentration substandard model in Example 2 of the present invention.

[0015] Figure 5 This is a calibration curve verification diagram of the nomogram for predicting substandard tacrolimus blood concentration in renal transplant patients according to Example 2 of the present invention.

[0016] Figure 6 This is a clinical decision curve diagram of the nomogram for predicting substandard tacrolimus blood concentration in renal transplant patients according to Example 2 of the present invention. DETAILED DESCRIPTION

[0017] In order to deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with examples. The examples are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0018] Example 1 according to Figure 1 As shown, this embodiment provides a method for predicting tacrolimus blood concentration based on multivariate regression analysis, comprising the following steps: Step 1: Obtain clinical data of patients taking tacrolimus after renal transplantation and monitor tacrolimus blood concentrations. According to the blood concentration data, divide the patients into a target group and a non-target group, and use a chi-square test to compare the groups to obtain comparison data. The clinical data included age, sex, body mass index, fasting blood glucose, triglycerides, total bilirubin, white blood cell count, red blood cell count, albumin, hemoglobin, platelet count, aspartate aminotransferase, alanine aminotransferase, serum creatinine, and blood urea nitrogen; The blood drug concentration monitoring specifically uses enzyme amplification immunoassay to pre-treat the blood, and then uses a fully automatic biochemical analyzer to measure the tacrolimus blood drug concentration; Specifically, patients with blood drug concentrations of 5-15 ng / ml were divided into the target group, and those with blood drug concentrations of 5-15 ng / ml were divided into the non-target group.

[0019] Step 2: Based on the comparison data, the blood drug concentration that does not meet the standard is set as the dependent variable, and the factors with statistical differences in the clinical data analysis are selected as independent variables. Multivariate Logistic regression analysis is used to obtain the independent risk factors for blood drug concentration that does not meet the standard; The independent variables include fasting blood glucose over Total bilirubin exceeds , white blood cell count less than , hemoglobin exceeds and serum creatinine exceeds .

[0020] Step 3: Based on the identified independent risk factors, a nomogram risk prediction model was established to predict the risk of tacrolimus blood concentration not being within the therapeutic range in renal transplant patients, and the stability and predictive performance of the model were verified and evaluated using cross-validation and bootstrap methods; The nomogram model specifically uses the scikit-learn database to build a regression model, uses the statsmodels database to extract coefficients, and finally uses the matplotlib database to return the nomogram to build a nomogram risk prediction model.

[0021] Step 4: Use HL goodness-of-fit test, ROC curve and consistency index to evaluate the nomogram risk prediction model.

[0022] Example 2 according to Figure 2-Figure 6 As shown, this embodiment provides an example of a method for predicting tacrolimus blood concentration based on multivariate regression analysis, which is as follows: 1. Data Acquisition Subject screening: Patients who were hospitalized and outpatients in a hospital from January 2023 to December 2023, took tacrolimus for treatment and monitored blood drug concentration after renal transplantation were selected as research subjects. Inclusion criteria: renal transplant patients, taking tacrolimus after surgery; age>18 years old;>1 year after surgery, taking drugs regularly and monitoring drug concentration regularly as required; complete clinical data and related examination results. Exclusion criteria: obvious adverse drug reactions after surgery; tacrolimus was suspended to enhance immunity; other immunosuppressants were replaced within one year; drugs that are currently known to affect tacrolimus blood concentration were taken; other chronic diseases were combined and affected the drug metabolism and distribution of tacrolimus in the body; clinical data were incomplete.

[0023] Medication regimen: Oral sustained-release tacrolimus was used as the basic immunosuppressant after surgery, and it was combined with mycophenolate mofetil and prednisone to form a triple immunosuppressive regimen. Tacrolimus dosage: oral, 0.15-0.3 mg / kg per day, twice a day; mycophenolate mofetil dosage: oral, 1 g each time, twice a day (daily dose 2 g). Prednisone dosage: oral, 5-10 mg / d. The study patients who met the inclusion criteria had basically the same medication schedule.

[0024] Blood drug concentration monitoring: All monitored patients were fasting in the morning, and 2 ml of venous blood was drawn from the patients and placed in a standard blood collection tube with EDTA anticoagulant for processing. The blood was pre-treated according to the tacrolimus detection reagent (enzyme amplification immunoassay), and the tacrolimus whole blood blood drug concentration was determined using the Viva-ProE fully automatic biochemical analyzer and enzyme amplification immunoassay. Take tacrolimus according to the doctor's instructions and monitor the tacrolimus drug concentration regularly. Patients with a target concentration range of 5-15 ng / ml using tacrolimus triple therapy were included in the target group, and those with a target concentration range of 5-15 ng / ml were included in the non-target group.

[0025] Data collection: Clinical data of 340 outpatients and inpatients who took tacrolimus according to doctor's orders in a certain hospital in 2023 were collected. Drugs that are currently known to affect the blood concentration of tacrolimus were excluded. Including age, gender, body mass index (BMI), fasting blood glucose (FPG), triglycerides (TG), total bilirubin (TB), white blood cell count (WBC), red blood cell count (RBC), albumin (ALB), hemoglobin (HB), platelet count (PLT), aspartate aminotransferase (AST), alanine aminotransferase (ALT), blood creatinine (Cr) and blood urea nitrogen (BUN).

[0026] 2. Comparison of clinical data Of the 340 tacrolimus drug concentration monitorings performed on patients, 224 concentrations reached the target range of 5-15 ng / ml, with a compliance rate of approximately 66%. , total bilirubin (TB) exceeds , white blood cell count (WBC) less than , hemoglobin (HB) exceeds and serum creatinine (Cr) exceeds The proportion of patients in the control group was significantly higher than that in the control group (P values ​​were all less than 0.05). However, there was no significant difference in the proportion of patients in the two groups in terms of body mass index (BMI), triglycerides (TG), red blood cell count (RBC), albumin (ALB), hemoglobin (HB), platelet count (PLT), aspartate aminotransferase (AST), alanine aminotransferase (ALT), and urea nitrogen (BUN) (P values ​​were all greater than 0.05). The specific data are shown in Table 1.

[0027] Table 1 Comparison of clinical data between the target group and the non-target group

[0028]

[0029] 3. Logistic regression analysis of influencing factors of tacrolimus blood concentration not reaching the target In the study of whether the tacrolimus blood concentration in renal transplant patients reached the target, the tacrolimus blood concentration that did not reach the target was set as the dependent variable (no = 0, yes = 1), and the factors with statistical differences in the clinical data analysis were selected as independent variables: fasting blood glucose (FPG) exceeding , total bilirubin (TB) exceeds , white blood cell count (WBC) less than , hemoglobin (HB) exceeds and serum creatinine (Cr) exceeds Multivariate logistic regression analysis showed that white blood cell count (WBC) ≤ ,TB> ,Cr> 、BUN≤ The above data are independent risk factors for substandard tacrolimus blood concentration in renal transplant patients (P < 0.05), indicating that they have a significant impact on substandard conditions. The specific data are shown in Table 2.

[0030] Table 2 Logistic regression analysis of factors affecting drug concentration not meeting the standard

[0031] 4. Establishment of a risk prediction model for tacrolimus blood concentration not meeting the target White blood cell count (WBC) based on identification ≤ ,TB> ,Cr> 、BUN≤ It is an independent risk factor for the substandard tacrolimus blood concentration in renal transplant patients. Based on the above risk factors, a nomogram model was established to predict the risk of tacrolimus blood concentration in renal transplant patients not being within the therapeutic range, as shown in the attached manual. Figure 2 shown.

[0032] 5. Evaluation of the efficacy of the nomogram risk prediction model for substandard tacrolimus blood concentration in renal transplant patients Discrimination evaluation: The discrimination ROC of this model is 0.84, 95%CI (0.73, 0.93,) as shown in the instruction manual. Figure 3 The ROC curve showed that the nomogram prediction model had a good prediction effect on the substandard tacrolimus blood concentration in renal transplant patients, with an area under the curve of 0.84 and a 95% confidence interval of (0.73, 0.93). Compared with the prediction AUC values ​​of other risk factors, the AUC value of this model was higher, indicating that the nomogram prediction model had good accuracy and discrimination ability, as shown in the attached manual. Figure 4 .

[0033] Consistency evaluation: The verification results of the model calibration curve showed that the chi-square value of the HL deviation test was 5.134 and the P value was 0.386, indicating that the model has good calibration ability, the predicted value and the actual value are highly consistent, and it can accurately predict the situation where the tacrolimus blood concentration in renal transplant patients does not meet the standard, as shown in the attached manual. Figure 5 .

[0034] Evaluation of clinical application: A decision curve analysis was conducted on various risk factors and nomogram prediction models that affect the substandard tacrolimus blood concentration in renal transplant patients. The results showed that the nomogram prediction model performed better than various risk factors within the threshold probability range (0.05-0.95), and the area between the red curve and the gray curve and the horizontal axis of the nomogram prediction model was the largest, indicating that the model has a high clinical application value, as shown in the attached manual. Figure 6 .

[0035] This application constructs a nomogram model to more intuitively display the risk of tacrolimus blood concentration not reaching the target in renal transplant patients, thereby strengthening the monitoring of tacrolimus drug therapeutic concentration. ,TB> ,Cr> 、BUN≤ Multivariate Logistic regression analysis was performed and a nomogram model was constructed. Figure 2As you can see, each variable (TB, Cr, WBC, BUN) has a score line, which represents the impact of the variable on the risk of substandard concentration. For example, the score line for TB (total bilirubin) means that the higher the TB value, the higher the corresponding risk score. According to the patient's specific value, find the corresponding score on the corresponding variable score line. For example, if TB is 25, find the risk score corresponding to 25 on the TB score line, add up the scores of all variables, and get a total score. Based on the total score, assess the risk of the patient's substandard concentration. The higher the total score, the greater the risk of substandard concentration. To take a simple example, suppose a patient's TB is 25. 、Cr is 140 、WBC 、BUN is 8 , TB is 25, the corresponding risk score is about 25*0.0205=0.5125, Cr is 140, the corresponding risk score is about 140*-0.0120=-1.68, WBC is 6, the corresponding risk score is about 6*-0.1352=-0.8112, BUN is 8, and the corresponding risk score is about 8*0.1536=1.2288. Then, add these scores to get a total score of 0.5125-1.68-0.8112+1.2288≈-0.75. The total score is positive: If the total score is positive, it means that the combination of these variables increases the risk of substandard concentration. If the total score is negative (such as -0.75 in the example), it means that the combination of these variables reduces the risk of substandard concentration. In this model, because the coefficients of some variables are negative (such as Cr and WBC), their higher values ​​actually reduce the overall risk score, and vice versa. The larger the absolute value of the total score, the more confident the model is in predicting the risk of substandard concentration.

[0036] In summary, the higher the total score, the greater the risk of substandard concentration. Doctors can judge the patient's risk level based on this total score and take corresponding intervention measures. The substandard risk model showed excellent predictive ability. The calibration curve verification showed that the predicted probability of the nomogram was highly consistent with the actual probability, and it had excellent accuracy and discrimination.

[0037] The prediction model of this application helps to identify high-risk patients, thereby reducing unnecessary tacrolimus monitoring and lowering medical costs. Based on current research results, it is estimated that the cost of each tacrolimus drug concentration monitoring will be reduced by approximately RMB 350. For patients after kidney transplantation who need to monitor tacrolimus blood concentrations for a long time, reducing the monitoring frequency will significantly reduce total medical expenses. Clinical decision curve analysis shows that the model has high practical value in clinical applications.

[0038] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for predicting tacrolimus blood concentration based on multivariate regression analysis, characterized in that: The following steps are involved: Step 1: Obtain clinical data of patients taking tacrolimus after renal transplantation and monitor tacrolimus blood concentrations. According to the blood concentration data, divide the patients into a target group and a non-target group, and use a chi-square test to compare the groups to obtain comparison data. Step 2: Based on the comparison data, the blood drug concentration that does not meet the standard is set as the dependent variable, and the factors with statistical differences in the clinical data analysis are selected as independent variables. Multivariate Logistic regression analysis is used to obtain the independent risk factors for blood drug concentration that does not meet the standard; Step 3: Based on the identified independent risk factors, a nomogram risk prediction model was established to predict the tacrolimus blood concentration in renal transplant patients that was not within the therapeutic range, and the stability and predictive performance of the model were verified and evaluated using cross-validation and bootstrap methods; Step 4: Use HL goodness-of-fit test, ROC curve and consistency index to evaluate the nomogram risk prediction model.

2. The method for predicting tacrolimus blood concentration based on multivariate regression analysis according to claim 1, characterized in that: The clinical data in step 1 include age, gender, body mass index, fasting blood glucose, triglycerides, total bilirubin, white blood cell count, red blood cell count, albumin, hemoglobin, platelet count, aspartate aminotransferase, alanine aminotransferase, blood creatinine and blood urea nitrogen.

3. The method for predicting tacrolimus blood concentration based on multivariate regression analysis according to claim 1, characterized in that: The blood drug concentration monitoring in the step 1 specifically adopts enzyme amplification immunoassay to perform blood pretreatment, and then uses a fully automatic biochemical analyzer to detect and determine the tacrolimus blood drug concentration; the patient grouping specifically divides the patients with a blood drug concentration of 5-15ng / ml into a target group, and vice versa into a non-target group.

4. The method for predicting tacrolimus blood concentration based on multivariate regression analysis according to claim 1, characterized in that: The independent variables in step 2 include fasting blood glucose exceeding Total bilirubin exceeds , white blood cell count less than , hemoglobin exceeds and serum creatinine exceeds .

5. The method for predicting tacrolimus blood concentration based on multivariate regression analysis according to claim 1, characterized in that: The nomogram risk prediction model in step three specifically uses the scikit-learn database to build a regression model, uses the statsmodels database to extract coefficients, and finally uses the matplotlib database to return the nomogram to build the nomogram risk prediction model.

Images