Children lupus nephritis patient group pharmacokinetics model and construction method and application thereof

Abstract

The invention relates to the technical field of biological medicine, and particularly discloses a children lupus nephritis patient group pharmacokinetics model and a construction method and application thereof. The model can accurately guide individualized medication, fill up the research blank and promote development of pharmacogenomics, and the group pharmacokinetics model constructed by the invention comprehensively considers multiple factors influencing MMF pharmacokinetics, such as genetic factors and clinical characteristics of child LN patients, and can accurately predict MPA exposure and pharmacokinetics parameters of different patients. Clinical doctors can formulate a more accurate MMF administration scheme for child LN patients according to model prediction results and in combination with specific conditions of the patients, the drug treatment effect is improved, adverse reactions are reduced, and prognosis of the patients is improved. According to the method, the pharmacokinetics model suitable for the specific group is constructed for the first time, the research blank in the field is filled, and important reference and basis are provided for follow-up related research.

Description

Technical Field

[0001] The present application relates to the field of biomedicine technology, and in particular to a pharmacokinetic model for a pediatric lupus nephritis patient population, a construction method thereof, and an application thereof. Background Art

[0002] Lupus nephritis (LN) is a common and serious complication of systemic lupus erythematosus (SLE), severely impacting children's health and growth. Mycophenolate mofetil (MMF), an important drug for the treatment of LN, is increasingly used in children.

[0003] Characteristics of Pediatric LN Patients and Current Status of MMF Treatment: Pediatric LN patients have physiological differences from adults. Their liver and kidney function, metabolic enzyme activity, and transporter protein function are constantly developing and maturing, resulting in significant differences in drug absorption, distribution, metabolism, and excretion in children. Currently, MMF dosing regimens for pediatric LN patients are often based on adult experience, lacking precise guidance based on children's pharmacokinetic characteristics. This results in variable clinical treatment outcomes, with some patients experiencing poor results or adverse reactions.

[0004] Individual variability in MMF pharmacokinetics: After oral administration, MMF is rapidly hydrolyzed in vivo to mycophenolic acid (MPA). Its pharmacokinetics are influenced by multiple factors. Genetic polymorphisms of metabolic enzymes such as UDP-glucuronosyltransferase (UGT) 1A9 and UGT2B7, and transporters such as ABCC2 and SLCO1B3, as well as factors such as patient age, gender, weight, and concomitant medications, can lead to significant individual variability in MPA exposure and pharmacokinetic parameters. This variability makes it difficult to tailor a fixed dose of MMF to meet the needs of every patient, increasing the complexity and uncertainty of treatment.

[0005] While some studies have examined the pharmacokinetics of MMF in renal transplant recipients or adult LN patients, relatively few studies have focused on the specific population of pediatric LN patients. Existing studies have limited sample size, the range of genetic loci examined, and the comprehensiveness of model construction, hindering accurate characterization of the pharmacokinetic profile of pediatric LN patients and effectively guiding individualized clinical medication use. Furthermore, significant individual variability in the pharmacokinetics (PK) of MMF makes it difficult to determine the optimal dosage in clinical treatment, impacting therapeutic efficacy and potentially increasing the risk of adverse reactions.

[0006] Compared with traditional pharmacokinetic studies, population pharmacokinetic studies can fully account for inter- and intra-individual variability. By integrating clinical data and genetic information from a large number of patients, more accurate models can be constructed, providing strong support for personalized clinical medication. In the treatment of pediatric LN patients, MMF, as an important immunosuppressant, has its pharmacokinetics influenced by multiple factors, including genetic factors, physiological factors, and disease status. However, existing research has mostly focused on adults or other disease populations, and research on pediatric LN patients is extremely scarce, failing to meet the urgent clinical need for precision medication.

[0007] Therefore, there is currently a lack of a population pharmacokinetic model that comprehensively considers multiple factors and has high accuracy to meet the clinical needs for the rational use of MMF in pediatric LN patients. Summary of the Invention

[0008] The purpose of this application is to overcome the deficiencies of the above-mentioned prior art and to provide a pharmacokinetic model for a pediatric lupus nephritis patient population, a method for constructing the model, and an application thereof.

[0009] To achieve the above objectives, the technical solutions adopted in this application are:

[0010] The present application provides a method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population, comprising the following steps:

[0011] S1. Collect clinical information, medication information, and blood samples from patients with lupus nephritis; determine the concentration of the blood samples; extract whole blood genomic DNA from patients with lupus nephritis; perform genotyping on single nucleotide polymorphism sites of candidate genes to obtain concentration data and genotyping data;

[0012] S2. Analyze the data collected in step S1 using a nonlinear mixed-effects model program. Determine the optimal model structure by comparing the structural model, random-effects model, and error model. Perform a covariate analysis and then screen for covariates that have a significant impact on the model using stepwise forward inclusion and stepwise backward elimination. Finally, determine the covariates and incorporate them into the pharmacokinetic model for the pediatric lupus nephritis population.

[0013] S3. Evaluate and validate the model obtained in step S2 using goodness-of-fit diagnostic plots, visual predictive checks, and nonparametric bootstrapping, and perform external validation.

[0014] This application focuses on the pediatric lupus nephritis (LN) patient population and is dedicated to building an accurate population pharmacokinetic model to optimize the clinical use of mycophenolate mofetil (MMF).

[0015] In the current medical field, precision medicine has become a development trend, and population pharmacokinetic research is one of the key links in achieving precision medicine. The innovation of this application is to target the special group of children with LN, deeply analyze the pharmacokinetic characteristics of MMF in this group, and fill the gaps in existing research. This application breaks through the limitations of traditional research, innovatively and comprehensively uses advanced experimental techniques and data analysis methods to comprehensively and systematically study the pharmacokinetic process of MMF in children with LN. In terms of data collection, it not only covers detailed clinical information, but also deeply detects multiple single nucleotide polymorphism (SNP) sites of genes related to drug metabolism and transport, providing a rich data basis for comprehensive analysis of the source of pharmacokinetic variation. In the process of model construction, advanced nonlinear mixed effect models are adopted, combined with rigorous model evaluation and verification methods to ensure the accuracy and reliability of the model. This innovative research method and model building strategy enables this application to accurately quantify the identifiable sources of MMF pharmacokinetic variability and identify the risk genes that affect MMF pharmacokinetic parameters, thereby providing a scientific and accurate basis for clinicians to formulate personalized dosing plans, significantly improving the effectiveness and safety of MMF in the treatment of pediatric LN patients and promoting the development of precision medicine for pediatric LN.

[0016] This application also analyzed the correlation between multiple gene loci and MMF pharmacokinetic parameters and found that eight genetic variants, including UGT1A9 (rs6717546, rs13418420, rs7586110, rs2070959, rs6759892), UGT2B7 rs7438135, ABCC2 rs7910642, and CES1 rs12149373, significantly affected the pharmacokinetics of MMF. This not only provides important evidence for explaining inter-individual variability in drug response, but also contributes to further research on the application of pharmacogenomics in the treatment of pediatric LN and promotes the development of this field.

[0017] As a preferred embodiment of the method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population described in the present application, in step S1, the clinical data include a disease activity index or liver and kidney function indicators;

[0018] The medication information includes the dosage form, dosage or administration time of mycophenolate mofetil;

[0019] The concentration of mycophenolate mofetil (MMF) in blood samples was determined by liquid chromatography-tandem mass spectrometry.

[0020] As a preferred embodiment of the method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population described in the present application, in step S2, the structural model selects a two-compartment model combined with lag time and linear elimination, the random effect model uses an exponential model to describe inter-individual variation, and the error model selects a mixed error model to characterize intra-individual variation;

[0021] For the observed concentration Cobs,ij and the model predicted concentration Cpred,ij, the mixed error model is expressed as:

[0022] Cobs,ij=Cpred,ij×(1+εprop,ij)+εadd,ij;

[0023] i represents the i-th individual (i=1,2,…,n), j represents the j-th observation of the i-th individual (j=1,2,…,mi); εprop,ij is the proportional error term, which follows the normal distribution N(0,σ 2 prop), where σprop is the standard deviation of the proportional error, which characterizes the correlation between the error size and concentration (the higher the concentration, the larger the absolute value of the error);

[0024] εadd,ij is the additive error term, which obeys the normal distribution N(0,σ 2 add), where σadd is the standard deviation of the additive error, representing fixed errors that are independent of concentration (such as detection instrument noise).

[0025] As a preferred embodiment of the method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population described in the present application, in step S2, the covariate analysis includes at least one of gender, age, weight, white blood cell count, blood neutrophil count, platelet count, lymphocyte count, hematocrit, hemoglobin, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, total bilirubin, albumin, glucose, blood urea nitrogen, systemic lupus erythematosus disease activity index 25, cystatin C and creatinine clearance.

[0026] As a preferred embodiment of the method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population described in the present application, in step S2, a scatter plot and Pearson correlation coefficient are used to analyze the correlation between the covariates, and covariates with a correlation coefficient greater than |0.7| are excluded.

[0027] As a preferred embodiment of the method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population described in the present application, in step S2, the parameters of the stepwise forward inclusion method are: P < 0.05, OFV decrease is greater than 3.84; the parameters of the stepwise backward elimination method are: P < 0.01, OFV increase is greater than 10.83.

[0028] As a preferred embodiment of the method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population described in the present application, in step S2, the covariate is finally determined as follows: body weight is determined to be the only covariate that significantly affects the peripheral distribution volume V2 / F.

[0029] As a preferred embodiment of the method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population described in the present application, in step S3, the mean prediction error MDPE, mean absolute prediction error MAPE, F20, and F30 are calculated by external validation;

[0030] If MDPE≤±20%, MAPE≤±30%, F20≥35%, and F30≥50%, the accuracy and effectiveness of the pharmacokinetic model for the pediatric lupus nephritis population are demonstrated.

[0031] The present application provides a method for constructing the above-mentioned pharmacokinetic model of the childhood lupus nephritis patient population to construct a pharmacokinetic model of the childhood lupus nephritis patient population.

[0032] The present application provides the application of the above-mentioned pediatric lupus nephritis patient population pharmacokinetic model in guiding the individualized medication of mycophenolate mofetil in pediatric lupus nephritis patients.

[0033] The pharmacokinetic model for the pediatric lupus nephritis patient population of this application can accurately guide drug dosage, and this model can accurately predict the pharmacokinetic parameters of mycophenolate mofetil (MMF) in different pediatric lupus nephritis patients. Taking the actual clinical scenario as an example, for patients with different weights and genotypes, the initial dose of MMF that achieves the best therapeutic effect and the lowest risk of adverse reactions can be calculated based on the model. For example, for patients carrying specific UGT1A9, ABCC2 and CES1 gene mutations, the model can clearly prompt that the dosage should be adjusted appropriately to avoid treatment failure or adverse reactions due to inappropriate dosage. Compared with the traditional experience-based medication method, the use of this model to guide medication can make the drug dosage more in line with the patient's individual metabolic capacity and improve the treatment efficiency. Studies have shown that in clinical practice using this model to guide medication, the treatment efficiency of patients has been improved.

[0034] Through the population pharmacokinetic model of this application, clinicians can more accurately grasp the drug metabolism of each patient. For children with lupus nephritis at different stages of disease activity, the model can optimize the MMF dosing regimen based on their specific conditions. During the active phase of the disease, the model can guide an increase in drug dosage to quickly control the condition; during the remission phase of the disease, the dosage can be reasonably reduced to reduce the adverse reactions caused by long-term drug exposure. This not only helps to improve the treatment effect of the disease, but also significantly reduces the occurrence of adverse reactions such as infection and gastrointestinal discomfort, improving the patient's quality of life and treatment compliance.

[0035] In the field of drug development for pediatric kidney disease, this model provides an important reference for new drug development and clinical trial design. When developing new immunosuppressants for pediatric lupus nephritis, this model can be used to predict the drug's pharmacokinetic profile in the target patient population, optimize drug dosage and dosing regimen, improve R&D efficiency, and reduce R&D costs. During the clinical trial design phase, selecting appropriate subjects based on this model can more accurately assess drug efficacy and safety, reduce experimental errors, and make clinical trial results more reliable and convincing.

[0036] The application of this model helps optimize the allocation of medical resources. In clinical practice, it has avoided treatment failures and recurrences caused by irrational medication use, reduced the number and duration of hospitalizations, and lowered medical costs. Furthermore, precise medication guidance reduces unnecessary drug monitoring and testing, improves the efficiency of medical resource utilization, and enables limited medical resources to better serve more patients. Statistics show that using this model to guide treatment reduces the average number of hospitalizations, shortens hospitalization duration, and reduces medical costs.

[0037] Compared with the prior art, this application has the following beneficial effects:

[0038] The present application provides a pharmacokinetic model for a population of children with lupus nephritis, its construction method, and its application. The model of the present application can accurately guide individualized medication, fill research gaps, and promote the development of pharmacogenomics. The population pharmacokinetic model constructed by the present application comprehensively considers various factors affecting the pharmacokinetics of MMF in children with LN, such as genetic factors and clinical characteristics, and can accurately predict the MPA exposure and pharmacokinetic parameters of different patients. Clinicians can formulate more accurate MMF dosing regimens for children with LN based on the model prediction results and the specific conditions of the patients, thereby improving the efficacy of drug treatment, reducing the occurrence of adverse reactions, and improving the prognosis of patients. This application is the first to construct a pharmacokinetic model suitable for this specific population, filling the research gap in this field and providing an important reference and basis for subsequent related research. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flow chart;

[0040] Figure 2 is the model fit goodness of fit plot;

[0041] Figure 3 This is the visual predictive check (VPC) graph. DETAILED DESCRIPTION

[0042] In order to better illustrate the purpose, technical solutions and advantages of this application, this application will be further described below with reference to the accompanying drawings and specific embodiments.

[0043] In the following examples, the experimental methods used are conventional methods unless otherwise specified, and the materials, reagents, etc. used are all commercially available unless otherwise specified, and the components and raw materials used in each parallel experiment are all the same.

[0044] Study Selection: Fifty-one pediatric LN patients were consecutively recruited from the First Affiliated Hospital of Sun Yat-sen University, strictly following the 2019 American College of Rheumatology classification criteria for Systemic Lupus Erythematosus (SLE) and diagnostic criteria for Lupus nephritis (LN). Patients with other conditions that could affect the pharmacokinetics of MMF (e.g., severe liver disease, renal failure) and recent use of medications that could interact with MMF were excluded. Informed consent was obtained from the patients and their guardians before the study began, and approval was obtained from the hospital's ethics committee.

[0045] Example 1: A method for constructing a population pharmacokinetic model of mycophenolate mofetil

[0046] This embodiment provides a method for constructing a population pharmacokinetic model of mycophenolate mofetil, comprising the following steps:

[0047] 1. Sample collection and testing: Patients with lupus nephritis took MMF capsules continuously ( RochePharma, Basel, Switzerland) or MMF dispersible tablets ( Zhongmeihuadong Pharma, Hangzhou, China) for at least 7 days. After reaching steady state, venous blood samples were collected before dosing and 0.5, 1.5, 2.5, 4, 6, 9, and 12 hours after dosing. Immediately after sample collection, the samples were centrifuged at 5,000g for 10 minutes to separate the plasma and store it at -80°C until testing. The concentration of MPA in plasma was determined using a validated LC-MS / MS method to ensure the accuracy and reproducibility of the test results. This method has high sensitivity and specificity and can accurately detect low concentrations of MPA in plasma. During the determination process, a calibration curve was drawn using an MPA standard of known concentration to ensure the accuracy and reproducibility of the test results.

[0048] At the same time, whole blood samples were collected from patients, and genomic DNA was extracted using the TIANGEN Blood Genome Extraction Kit. The Agena MassARRAY System technology was used to genotype 29 SNP sites in 13 pre-selected candidate genes (ABCB1, ABCC2, ABCG2, CES1, CES2, CYP2C8, HNF1A, IMPDH1, SLCO1B1, SLCO1B3, UGT1A9, UGT2B10, and UGT2B7). Among them, the UGT1A9 gene contains rs6717546, rs13418420, rs7586110, rs2070959, and rs6759892; the UGT2B7 gene contains rs7438135; the ABCC2 gene contains rs7910642; the CES1 gene contains rs12 149373, rs2231142 in the ABCB1 gene, rs17146905 in the CES2 gene, rs10929303 in the CYP2C8 gene, rs1058932 in the SLCO1B1 gene, rs12762549 in the SLCO1B3 gene, and rs2804402 in the UGT2B10 gene. The remaining 13 SNPs (including rs273697, rs1169310, and rs8330 and rs3842, which were excluded due to low detection rates, insufficient genotype numbers, or deviations from Hardy-Weinberg equilibrium) are tentatively assigned to unknown genes because their gene assignments have not been reported in previous studies or lack clear evidence for association. Genotyping experiments were performed strictly according to the manufacturer's instructions for the kits and instruments. Quality control was performed at every step, from DNA extraction and PCR amplification to mass spectrometry, to ensure the reliability of the genotyping results.

[0049] 2. Data Preprocessing and Analysis: Collected patient clinical information (including disease activity index, liver and kidney function indicators, etc.), medication information (MMF dosage form, dosage, and administration time, etc.), genotyping data, and MPA concentration data were systematically organized and integrated. Data were statistically analyzed using IBM SPSS Statistics Version 22.0 software. Descriptive statistical indicators such as mean, standard deviation, and median were calculated for each variable to understand the basic characteristics of the data. The above gene polymorphisms (rs6759892, rs2804402, rs2231142, rs10929303, rs273697, rs1169310, rs17146905, rs2070959, rs2306283, rs6717546, rs8330, rs7910642, rs11075646, rs12149373, rs7311358, rs2228075, rs13418420, rs10 The correlations between SNPs (rs58932, rs12762549, rs3842, rs56097722, rs7438135, rs1885301, rs2278294, rs2464196, rs4149117, rs3740066, rs7586110, and rs28365062) and PK parameters (area under the concentration-time curve, half-life, clearance, time to peak concentration, and peak concentration) were analyzed in depth using appropriate statistical methods (e.g., chi-square test, Fisher's exact test). SNPs that did not conform to Hardy-Weinberg equilibrium (HWE), had a low detection rate (<95%), or had an abnormal number of genotypes (e.g., only two genotypes detected) were excluded. In this study, eight SNP loci (UGT1A9 (rs6717546, rs13418420, rs7586110, rs2070959, rs6759892), UGT2B7 rs7438135, ABCC2 rs7910642, and CES1 rs12149373) were finally included in the subsequent analysis to ensure the accuracy and reliability of the analysis results.

[0050] 3. Model Construction and Optimization: A population pharmacokinetic model for MMF was constructed using the nonlinear mixed-effects (NLME) model in Phoenix NLME software. During the structural model selection phase, both a one-compartment model and a two-compartment model were evaluated for fit. Considering different absorption and elimination mechanisms (e.g., first-order absorption and first-order elimination), as well as the presence of a lag time (Tlag), the optimal structural model was determined to be a two-compartment model combined with Tlag and linear elimination by comparing the Akaike Information Criterion (AIC), objective function value (OFV), goodness-of-fit diagnostic plots (including scatter plots of observed concentration versus population predicted concentration, observed concentration versus individual predicted concentration, conditional weighted residual versus time, and conditional weighted residual versus population predicted value), and relative standard errors (RSE) of parameter estimates.

[0051] In the random effects model, the exponential model was used to describe the inter-individual variation. By comparing the fitting effects of the proportional model, additive model and mixed model on the remaining unexplained variation, the mixed error model was selected to characterize the intra-individual variation.

[0052] Scale model: yij=f(xij;β)·exp(bi)+f(xij;β)·exp(bi)·σp·δij;

[0053] Additive model: yij = f(xij;β)·exp(bi)+σa·δij;

[0054] Mixed model: yij=f(xij;β)·exp(bi)+σa·δij+f(xij;β)·exp(bi)·σp·εij;

[0055] Among them, f(xij;β) is the fixed effect part, which is determined by the covariate xij and the parameter β;

[0056] exp(bi) is the exponential model of inter-individual variation, bi is the random effect of individual i, and it is usually assumed that bi~N(0,τ2), that is, the inter-individual variation is multiplied by the fixed effect through the exponential function;

[0057] eij is the intra-individual residual error term, corresponding to the characterization of three different models.

[0058] In the covariate analysis, the stepwise forward inclusion method (P < 0.05, OFV decrease greater than 3.84) and the stepwise backward elimination method (P < 0.01, OFV increase greater than 10.83) were used to screen for covariates that may affect the pharmacokinetics of MMF (such as gender, age, weight, disease activity index, and liver and kidney function indicators). Ultimately, weight was identified as a significant covariate affecting the peripheral volume of distribution (V2 / F) and was included in the model for optimization to obtain the final population pharmacokinetic model.

[0059] 4. Model evaluation and validation: The final model was comprehensively evaluated and validated using goodness-of-fit diagnostic plots, visual prediction checks (VPC), and non-parametric bootstrapping methods. The goodness-of-fit diagnostic plots demonstrate the consistency between the model predictions and the actual observations from multiple perspectives, such as observing the distribution of data points in relevant scatter plots to evaluate how well the model fits the data. VPC judges the predictive performance of the model by comparing the distribution characteristics of the simulated data and the actual observed data (including the distribution of concentration data at different time points, the distribution of the 5th, 50th, and 95th percentiles, etc.). The non-parametric bootstrapping method generates 1,000 data sets by repeatedly sampling with replacement from the original data set, fits the model to each data set, calculates the median and 95% confidence interval of the model parameter estimates, and compares them with the original model results to evaluate the stability and reliability of the model.

[0060] Here are the results:

[0061] 1. Typical values ​​of absorption rate constant (θ Ka , unit: h -1 )

[0062] Final model: Estimate 1.54, relative standard error (RSE) 10.25%;

[0063] Bootstrapping method: median 1.69, 95% confidence interval (95% CI) 1.42-1.06.

[0064] 2. Typical values ​​of central distribution volume (θ V / F , unit: L)

[0065] Final model: Estimate 7.99, RSE 22.76%

[0066] Bootstrap method: median 8.31, 95% CI 1.67-12.68.

[0067] 3. Typical values ​​of peripheral distribution volume (θ V2 / F , unit: L)

[0068] Final model: Estimate 1287.12, RSE 8.65%

[0069] Bootstrapping method: Median 1116.28, 95% CI 841.72-1682.86.

[0070] 4. Typical values ​​of central clearance (θ CL / F , unit: L·h -1 )

[0071] Final model: Estimate 15.23, RSE 1.85%

[0072] Bootstrap method: median 16.27, 95% CI 14.39-18.40.

[0073] 5. Typical values ​​of clearance between compartments (θ CL2 / F , unit: L·h -1 )

[0074] Final model: Estimate 37.65, RSE 3.32%

[0075] Bootstrap method: median 38.86, 95% CI 30.26-46.83.

[0076] 6. Lag time (T lag , unit: h)

[0077] Final model: Estimate 0.40, RSE 3.29%

[0078] Bootstrap method: median 0.42, 95% CI 0.17-0.49.

[0079] 7. Weight-related peripheral volume parameters (θ v2 ,weight)

[0080] Final model: Estimate 2.05, RSE 11.99%

[0081] Bootstrap method: median 2.05, 95% CI 0.39-2.60.

[0082] 8. Inter-individual variation (variance form, ω 2 )

[0083] Central distribution volume variation (ω 2 V / F ):

[0084] Final model: Estimate 1.02, RSE 14.89%;

[0085] Bootstrap method: median 1.10, 95% CI 0.98-1.22.

[0086] Peripheral distribution volume variation (ω 2 V2 / F ):

[0087] Final model: Estimate 1.24, RSE 13.80%;

[0088] Bootstrap method: median 1.25, 95% CI 1.01-1.48.

[0089] Compartmental clearance variation (ω2 CL2 / F ):

[0090] Final model: estimate 0.20, RSE 17.51%;

[0091] Bootstrap method: median 0.23, 95% CI 0.16-0.30.

[0092] At the same time, data from an additional nine patients with lupus nephritis were collected for external validation, and metrics such as the mean prediction error (MDPE), mean absolute prediction error (MAPE), and F20 and F30 were calculated. If MDPE ≤ ±20%, MAPE ≤ ±30%, F20 ≥ 35%, and F30 ≥ 50%, the model's predictive accuracy and clinical utility were considered good.

[0093] In this study, all evaluation indicators of the model met the above standards, proving the effectiveness of the model. Figure 1 shown.

[0094] The results are as follows Figure 2 The model goodness-of-fit plot (Figure 2) shows scatter plots of observed concentrations versus population-predicted concentrations, and observed concentrations versus individual-predicted concentrations. The plots intuitively show that the data points are concentrated near the diagonal, indicating good consistency between the model's predictions and observed values, and a good fit to the data, reflecting the model's ability to accurately describe the pharmacokinetic characteristics of MPA.

[0095] The results are as follows Figure 3 The Visual Predictive Check (VPC) plot compares the distribution of simulated and observed data. The observed concentrations mostly fall within the 95% confidence interval of the simulated data, with similar distribution trends. This demonstrates the model's excellent predictive performance and the ability to accurately predict MPA concentrations at different time points, providing a reliable basis for adjusting clinical dosages.

[0096] This model accurately predicts the pharmacokinetic parameters of mycophenolate mofetil (MMF) in different pediatric lupus nephritis patients. For example, in real-world clinical scenarios, the model can be used to calculate the initial MMF dose that achieves optimal therapeutic efficacy and minimizes the risk of adverse reactions for patients with varying weights and genotypes. For example, for patients with specific UGT1A9, ABCC2, and CES1 gene variants, the model can clearly indicate appropriate dosage adjustments to avoid treatment failure or adverse reactions due to inappropriate dosing. Compared with traditional empirical medication management, using this model to guide medication can better align drug dosages with individual patient metabolic capacity, thereby improving treatment efficacy. Studies have shown that clinical practice using this model to guide medication has improved patient efficacy. This model enables clinicians to more accurately assess drug metabolism in each patient. For pediatric lupus nephritis patients at different stages of disease activity, the model can optimize MMF dosing regimens based on their specific circumstances. During active disease, the model can guide dose increases to rapidly control the disease; during remission, the model can guide dose reductions to minimize adverse reactions associated with long-term drug exposure. This not only helps improve the treatment effect of the disease, but also significantly reduces the occurrence of adverse reactions such as infection and gastrointestinal discomfort, and improves patients' quality of life and treatment compliance.

[0097] Furthermore, this model provides an important reference for new drug development and clinical trial design. When developing new immunosuppressants for pediatric lupus nephritis, this model can be used to predict the drug's pharmacokinetic profile in the target patient population, optimize drug dosage and dosing regimen, improve R&D efficiency, and reduce R&D costs. During the clinical trial design phase, selecting appropriate subjects based on this model can more accurately assess drug efficacy and safety, reduce experimental errors, and make clinical trial results more reliable and convincing.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not intended to limit the scope of protection of the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application may be modified or replaced by equivalents without departing from the essence and scope of the technical solutions of the present application.

Claims

1. A method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population, characterized in that: The following steps are involved: S1. Collect clinical information, medication information, and blood samples from patients with lupus nephritis; Blood samples were tested for concentration, and whole blood genomic DNA was extracted from patients with lupus nephritis. Single nucleotide polymorphism sites of candidate genes were genotyped to obtain concentration data and genotyping data. S2. Analyze the data collected in step S1 using a nonlinear mixed-effects model program. Determine the optimal model structure by comparing the structural model, random-effects model, and error model. Perform a covariate analysis and then screen for covariates that have a significant impact on the model using stepwise forward inclusion and stepwise backward elimination. Finally, determine the covariates and incorporate them into the pharmacokinetic model for the pediatric lupus nephritis population. S3. Evaluate and validate the model obtained in step S2 using goodness-of-fit diagnostic plots, visual predictive checks, and nonparametric bootstrapping, and perform external validation.

2. The method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population according to claim 1, wherein: In step S1, the clinical data include disease activity index or liver and kidney function indicators; The medication information includes the dosage form, dosage or administration time of mycophenolate mofetil; The concentration of mycophenolate mofetil in blood samples was determined by liquid chromatography-tandem mass spectrometry.

3. The method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population according to claim 1, wherein: In step S2, the structural model selects a two-compartment model combined with lag time and linear elimination, the random effect model adopts an exponential model to describe inter-individual variation, and the error model selects a mixed error model to characterize intra-individual variation; For the observed concentration Cobs,ij and the model predicted concentration Cpred,ij, the mixed error model is expressed as: Cobs,ij=Cpred,ij×(1+εprop,ij)+εadd,ij; i represents the i-th individual, i = 1, 2, ..., n; j represents the j-th observation of the i-th individual, j = 1, 2, ..., mi; εprop,ij is the proportional error term, which obeys the normal distribution N(0,σ 2 prop), where σprop is the standard deviation of the proportional error, which characterizes the correlation between the error size and concentration; εadd,ij is the additive error term, which obeys the normal distribution N(0,σ 2 add), where σadd is the standard deviation of the additive error, representing the fixed error that is independent of concentration.

4. The method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population according to claim 1, wherein: In step S2, the covariate analysis includes at least one of gender, age, weight, white blood cell count, blood neutrophil count, platelet count, lymphocyte count, hematocrit, hemoglobin, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, total bilirubin, albumin, glucose, blood urea nitrogen, systemic lupus erythematosus disease activity index 25, cystatin C and creatinine clearance.

5. The method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population according to claim 1, wherein: In step S2, the correlation between the covariates is analyzed using a scatter plot and Pearson correlation coefficient, and covariates with a correlation coefficient greater than |0.7| are excluded.

6. The method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population according to claim 1, wherein: In step S2, the parameters of the stepwise forward inclusion method are: P < 0.05, OFV decrease greater than 3.84; the parameters of the stepwise backward elimination method are: P < 0.01, OFV increase greater than 10.

83.

7. The method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population according to claim 1, wherein: In step S2, the covariate is finally determined as follows: body weight is determined to be the only covariate that significantly affects the peripheral distribution volume V2 / F.

8. The method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population according to claim 1, wherein: In step S3, external verification is performed to calculate the mean prediction error MDPE, mean absolute prediction error MAPE, F20, and F30; If MDPE≤±20%, MAPE≤±30%, F20≥35%, and F30≥50%, the accuracy and effectiveness of the pharmacokinetic model for the pediatric lupus nephritis population are demonstrated.

9. The method for constructing a pharmacokinetic model for a pediatric lupus nephritis patient population as claimed in claim 1, wherein a pharmacokinetic model for a pediatric lupus nephritis patient population is constructed.

10. Application of the population pharmacokinetic model of pediatric lupus nephritis patients in guiding the individualized use of mycophenolate mofetil in pediatric lupus nephritis patients.

Images