| Objective:In recent years,great progress has been made in basic research in order to provide better,safe and effective medicines for children,but the development and clinical use of medicines for children is still far behind the level of medicines for adults,and medicines for children are still an "orphan field".This situation is mainly due to the lack of clinical pharmacokinetic data,and the theoretical mechanism research results have not been fully verified by clinical data.Although mathematical tools have reduced the demand for clinical data to a certain extent,they still cannot solve the shortcomings of many subjects required for children’s clinical trials,and the long development cycle,especially the lack of individual exposure curve data,makes it difficult to meet the requirements of application materials and standard.There is therefore an urgent need to develop a test method that ethically characterizes the pharmacokinetic behavior of each individual.Non-invasive surrogate matrix sampling is the preferred strategy,and urine is the preferred study object.In this paper,drugs with a wide range of clinical applications were selected as the test drugs,and the selection criteria were:(1)special drugs for children or shared preparations for adults.(2)there were enough clinical patients.(3)the modes of administration covered common types.This paper establishes a mathematical model of the drug exposure relationship between urine and blood,and then conducts a comprehensive experimental validation of this new method,which is ultimately expected to be applied in the clinic.The principles of modeling and validating experimental designs are from simple to complex,from one-compartment models to multi-compartment models,from intravenous to oral administration,from single to multiple administrations,from animals to clinical,In order to promote clinical research on children’s(especially young children)drug use,promote the development of pediatric drugs,and provide theoretical and data support for the approval of new drugs.The research contents of this topic include:1.Use the classical compartment model to establish the mathematical relationship between the cumulative urinary excretion curve and the plasma exposure curve after intravenous injection,oral administration(extravascular),infusion,etc.2.The intravenous injection model used desloratadine as test drug.By measuring the urinary excretion and plasma exposure of rats after iv injection of desloratadine,we comprehensively evaluate the effect of different compartment models and different single-point plasma data on the plasma drug.The fitting results of concentration and pharmacokinetic parameters illustrate the optimization process of the model.3.The oral model uses desloratadine as the test drug.By measuring the urinary excretion curve and plasma exposure curve of rats after oral administration of desloratadine,the effect of different compartment models and different single-point plasma data on the plasma concentration and drug concentration is comprehensively evaluated.The fitting results of the kinetic parameters,and the optimization process of the model is explained.4.The intravenous infusion model uses busulfan as the test drug,and the pharmacokinetics and urinary and fecal excretion in rats(single-dose and multi-dose)after intravenous infusion are measured.best modeling method.5.To study the effects of renal injury on the pharmacokinetics and urinary excretion after intravenous injection of busulfan and aristolochic acid I in rats,and to explain the pharmacokinetics and urinary and fecal excretion changes from the injury mechanism.To study the effect of kidney injury on CLr and to explore the feasibility of fitting blood exposure.6.To recruit adolescent leukemia patients aged 3-18 years and complete clinical studies,verifying the feasibility of the infusion model in clinically fitting the blood exposure of pediatric patients.Main research results:1.The conversion of the cumulative urinary excretion equation to the residual urine dose equation has the same confounding parameters as the plasma exposure equation of the corresponding compartment model,so it is theoretically feasible to inversely infer blood exposure from urinary excretion.However,in the process of model fitting,compartments need to be determined,and single-point plasma drug concentration data is required to calculate the urinary excretion rate to anchor the plasma exposure curve.This paper mainly focuses on the above two subjective judgment factors,and establishes and optimizes the inversion process by intravenous injection,oral administration,single-dose infusion and multi-dose infusion,and validates the model efficacy.2.After single intravenous injection of DL 0.5 mg/kg in rats,the t1/2 was 2.14 h,widely distributed in tissues and organs,and the plasma concentration was low.The cumulative excretion rate within 24 h was 5.19%.By establishing the equations of the one-,two-and three-compartment models,it is found that the two-or three-compartment models have better fitting effects than the one-compartment model.After the comprehensive comparison of MAPE,AFE,AAFE and t1/2,Vd,AUC fitting errors of the model,the three-compartment model combined with T=1h blood drug data has the most ideal fitting effect,among which MAPE<100%,AFE<1,AAFE<3,AUC and other major pharmacokinetic parameters fitting errors were all within 0.5～2 times.However,there are also individual rats showing poor fitting effect.For example,the residual urine drug curve of Rat06 is a two-compartment model,while the plasma C-t curve is three-compartment,which leads to the poor fitting effect of the C-t curve obtained by anchoring the T=1 h single-point data.The reason is that the duration of theα phase is short,the urine excretion is limited by the sampling time period,and the αphase cannot be fully displayed;or the blood drug concentration of the alpha phase is high,the renal excretion is saturated,and the urine excretion curve fails to show the αphase.3.After oral administration of 2 mg/kg of DL to rats,the drug is rapidly absorbed in the body,Tmax is about 2 h,the t1/2 is about 4 h,and the absolute bioavailability of oral administration is 34.15%,but the individual differences are large.The cumulative excretion rate of the prototype in urine within 60 hours after oral administration in rats was 1.76%.After the optimization of the two subjective judgment factors,the two-compartment model and the CLr calculated from the T=4 h blood drug data were used to reverse the plasma exposure curve from the urinary excretion curve of rats after oral administration of DL.The drug concentration fitting error MAPE<100%,AFE=0.63,AAFE=2.02,the fitting range of the main pharmacokinetic parameters such as AUC,Cmax,Tmax,t1/2 are all within 0.5～2.0 times,and the actual measurement valueswere not significantly different.In the oral model in this chapter,there is a large error in the calculated value of tlag,which does not meet the expected value of 5 min,so the tlag correction is not considered in the equation modeling.At 8～12 hours after oral administration of DL to rats,the plasma C-t curve of individual animals appeared double peaks,which may be due to the existence of enterohepatic circulation.Although the animal’s urine excretion curve and the plasma C-t curve show a certain correlation in the presence of enterohepatic circulation,this more complex urine-blood exposure relationship cannot be fitted by this model(based on the compartment model),while the PBPK may be more advantageous.4.After intravenous infusion of Bu in rats,the in vivo exposure conformed to the one-compartment model,and the t1/2 was about 2.3 h.The cumulative excretion rate of the prototype through urine accounted for 4%of the total dose,while the excretion through feces was lower,about 0.4%.In this paper,an atrioventricular infusion model was established.By comparing the fitting results of different single-point plasma data,it was found that the early plasma drug data had a small fitting error after the end of administration.The fitting results show that the model has the best fitting effect when using T=1 h or T=1.033 h single-point plasma data to calculate CLr.The fitting error of plasma drug concentration MAPE=15.07%,AFE=1.04,AAFE=1.37,the fitting error of the main PK parameters such as t1/2,C0,AUC,Vss are within 0.9～1.1 times,and there is no significant difference with the measured values.The one-compartment single-infusion model in this chapter is suitable for the single-infusion administration process that does not reach the steady-state plasma drug concentration,only the measured data of the single-point plasma drug concentration of T=1 h or T=1.033 h and after the end of administration is required.The urinary excretion data can be reversibly extrapolated to the plasma exposure curve to obtain ideal pharmacokinetic parameter fitting results.5.Referring to the clinical dosing schedule of busulfan and its half-life in rats,in this chapter,τ is set to 4 h(2 times of the half-life),every infusion is 1 h,and total 7 times.After the first administration to rats,the excretion of the prototype drug in urine accounted for about 0.91%of the total dose,and about 0.05%in the feces.After the last administration,the cumulative urinary excretion rate within 24 hours was 3.61%,and the fecal cumulative excretion rate was 0.10%.In this chapter,a model equation for a total of "n" times of infusion was established under a constant dose,interval τ(h),each infusion T(h),and the single-point blood drug data after the first dose and the urinary excretion after the last dose were used.The data completes the inversion of the plasma exposure curve.Using the blood drug data immediately after the first infusion(T=1 h)has the best fitting effect,where MAPE=19.40%,AFE=0.97,AAFE=1.17.The error range of pharmacokinetic parameters such as t1/2,Cmaxs s,Cmins s,AUC(0-t),and CL were between 0.8～1.2 times.6.In this chapter,the Adriamycin kidney injury model(ARN)and the aristolochic acid I kidney injury model(AAN)were established.Taking into account the body weight,urine volume,urine creatinine,urine microalbumin,urine microalbumin and pathology of the model groups,the rats in both ARN and AAN group reached the degree of slight renal injury.ARN and AAN models lead to distinct pharmacokinetic and excretion trends for Bu and AAI.ARN led to a significant increase in the blood exposure of Bu in ARN rats,a significant decrease in the clearance rate,and a slight decrease in the cumulative excretion rate of the drug prototype in urine and feces;it led to an increase in the urinary excretion of AAI,a significant decrease in blood exposure,and a significant increase in the clearance rate.AAN led to a significant decrease in the cumulative excretion rate of AAI in rats,but did not cause significant changes in blood exposure and clearance;the PK characteristics and urinary excretion of Bu in AAN rats were not significantly affected.The above pharmacokinetic parameters and excretion changes may be related to the induction of the expression of drug metabolizing enzymes and transporters.The C-t curve and residual urine drug curve of rats in the control group and model groups after intravenous injection of Bu were in line with the one-compartment model,and the two curves had a good parallel relationship.The plasma C-t curve was inversely derived using the one-compartment intravenous injection model and T=2 min single-point plasma drug concentration data.The fitting errors of plasma drug concentration in the three groups were AFE<2,AAFE<3.The fitting errors of major pharmacokinetic parameters such as Vss,CL,and MRT were within 0.8～1.2 times.The results of this experiment showed that the parameters α and keA obtained by fitting the urinary excretion data were lower than the measured values,which showed a time-dependent trend of large variation of CLr,t calculated at each moment and a time-dependent trend of low before and after high.This greatly affects the representativeness and application prospects of urinary excretion data to PK studiesAfter intravenous injection,AAI conformed to the three-compartment model in rats.However,the residual urine drug curve conformed to the two-compartment model.Because the urinary excretion data is not dense enough,the slope of the α-phase of the C-t curve and the residual urine drug curve is quite different,and the mathematical model cannot achieve the ideal fitting effect.Preliminary experiments show that for toxins or drugs with short t1/2 and low urinary excretion rate,bladder intubation or urethral intubation can improve the fitting effect.7.This chapter presents the clinical validation of the Bu multiple infusion model.After the first dose of Bu 0.8 mg/kg,t1/2 was 2.51 h,Cmax was 833.51 ng/mL,AUC(0-6h)was 2904.00 h·ng/mL,and CL was 234.61 mL/h/kg.The cumulative urinary drug excretion rate was 1.41%from 0 to 6 hours after the first administration,and was 1.43%after the last administration.Using the urine excretion data after the first administration to reverse the C-t curve,MAPE<32.21%,RSD<2.55,AFE<1.07,AAFE<1.31.The fitting errors of PK parameters such as t1/2,Cminss,Cmaxss,AUC(0-6h),AUC(0-t)were between 0.99～1.25 times,which met the error requirements.When using urinary excretion data after last-dose,MAPE<38.32%,RSD<4.4,the fitting error of PK parameters is between 0.5～2.0 times,which can achieve accurate fitting of PK parameters.In another girl patient,the MAPE of the reverse C-t curve using the urinary excretion data after the first or last administration was 58.94%and 215.51%,respectively.The reason for the large error of the latter may be the variation of the CL during the treatment.In conclusion,in order to realize the purpose of using urine as a non-invasive biological matrix and inversely infer the C-t curve,this subject constructed the mathematical relationship between the plasma C-t curve and the urinary cumulative excretion curve based on the compartment model.For intravenous,oral,single-dose infusion,multi-dose infusion,and renal injury,the model was optimized through sensitivity analysis of 2 subjective factors(compartmental judgment and single-point plasma data selection).The efficacy of the model was verified by the measured blood concentration,and the PK research strategy that mainly relied on the urinary excretion data was initially realized.It cannot be denied that the limitations of the premise(drugs have a constant CLr)and principles(compartment model)of this model make it not applicable to all drugs,such as drugs with no(or low)urinary excretion,short half-life,complex disposes.Therefore,it is still necessary to further expand the technology,for example,continuous sampling of urine,using PBPK model to fully explore the kinetic data of drug metabolites in urine,and promote the PK research strategy of urine as a non-invasive biological matrix. |
PDF Full Text Request