| Objective1.To establish the population pharmacokinetic（PPK）modelling of Oxcarbazepine（OXC）active metabolite 10-monohydroxy derivative（MHD）in Chinese adult epilepsy patients,investigate the influence of the genetics factors on MHD Pharmacokinetics（PK）of adult patients,and propose a dosage guideline for adult population.2.To establish the PPK modelling of OXC active metabolite MHD in Chinese paediatrics epilepsy patients,investigate the influence of the genetics factors on MHD PK of paediatrics patients,and propose a dosage guideline for paediatrics population.3.To develop software for individualising dosage regimens of OXC based on our established PPK models.Methods1.MHD steady-state trough concentrations were prospectively collected from 149adult epilepsy patients who received OXC at least 1 week as model building group.Patient sex,age,body weight（BW）,hepatorenal function,and concomitant medications were recorded.Trough concentrations from 38 additional adult patients were collected as the external validation group.2.MHD steady-state trough concentrations were prospectively collected from 141paediatrics epilepsy patients who received OXC at least 1 week as model building group.Patient sex,age,BW,hepatorenal function,and concomitant medications were recorded.Trough concentrations from 20 additional adult patients were collected as the external validation group.3.Enzyme multiplied immunoassay technique（EMIT）was used to determine MHD plasma concentration of patients,and High performance liquid chromatography（HPLC）was used for comparison to test the accuracy and reliability of EMIT.4.The four single nucleotide polymorphisms（UGT2B7 802T>C,UGT1A9 I399C>T,ABCB1 3435C>T,and ABCB2 1249G>A）were detected by Dideoxy chain termination method（Sanger method）.Genotype distribution in adult and paediatric model building group was assessed by Hardy-Weinberg equilibrium（HWE）.5.The adult and paediatric PPK models were developed to investigate sex,age,BW,hepato-renal function,concomitant medications,and genetics factors on MHD pharmacokinetics（PK）by Nonlinear Mixed Effects modelling（NONMEM）software.6.The graphical method was used to detect the goodness-fit of the models,the Bootstrap method was used to detect the validity and stability of the model,Normalized predictive distribution error（NPDE）was utilized to evaluate the internal predictive performance of the models,and Mean prediciton error%（MPE%）and Mean absolute prediction error（MAE%）for the samples in the validation group were calculated to evaluate the external prective performance of the basic and final models.7.The dosage guidelines for the dosage regimens of the adult and paediatric patients were established based on the final models by Monte Carlo simulation.8.OXC Dosing software was developed using MyEclipse,SQL Server,and JRE,which could run NONMEM,predicte the trough concentrations under the various dosing regimens automatically.Results1.Comparing with HPLC,EMIT has the characteristics such as good accuracy,high sensitivity,non-essential preprocessing,easy operation,and rapid determination,which is fit for clinical routine monitoring of MHD.2.Among 149 epileptic adult patients,the frequency of genotype distribution for UGT2B7 802T>C is the following:the number of TT genotype is 17 cases（11.4%）,the number of TC genotype is 63 cases（42.3%）,and the number of CC genotype is 69cases（46.3%）;the frequency of genotype distribution for UGT1A9 I399C>T is the following:the number of CC genotype is 20 cases（13.4%）,the number of CT genotype is 82 cases（55.0%）,and the number of TT genotype is 47 cases（31.5%）;the frequency of genotype distribution for ABCC2 1249G>A is the following:the number of GG genotype is 98 cases（65.8%）,the number of GA genotype is 46 cases（30.9%）,and the number of AA genotype is 5 cases（3.4%）;the frequency of genotype distribution for ABCB1 3435 C>T is the following:the number of CC genotype is 69 cases（46.3%）,the number of CT genotype is 59 cases（39.6%）,the number of TT genotype is 21 cases（14.1%）.All genotype frequencies in adult model building group exhibited HWE.3.Among 141 epileptic paediatric patients,the frequency of genotype distribution for UGT2B7 802T>C is the following:the number of TT genotype is 15 cases（10.6%）,the number of TC genotype is 60 cases（42.6%）,and the number of CC genotype is 66cases（46.8%）;the frequency of genotype distribution for is UGT1A9 I399C>T the following:the number of CC genotype is 24 cases（17.0%）,the number of CT genotype is 61 cases（43.3%）,and the number of TT genotype is 56 cases（39.7%）;the frequency of genotype distribution for ABCC2 1249G>A is the following:the number of GG genotype is 81 cases（57.4%）,the number of GA genotype is 56 cases（39.7%）,and the number of AA genotype is 4 cases（2.8%）;the frequency of genotype distribution for ABCB1 3435 C>T is the following:the number of CC genotype is 50 cases（35.5%）,the number of CT genotype is 66 cases（46.8%）,and the number of TT genotype is 25 cases（17.7%）.All genotype frequencies in paediatric model building group exhibited HWE.4.The model development dataset included 278 serum samples from 149 adult patients with epilepsy,the final model of adult patients was:CL/F（L/h）=2×（GFR/80）0.7 54×（BW/70）0.482V/F（L）=96.2The final model revealed that Gomerular filtration rate and BW were significant factors which influenced adult MHD apparent clearance（CL/F）,and the other covariates such as polymorphisms of UGT2B7 802T>C,UGT1A9 I399C>T,ABCB1 3435C>T,ABCB21249G>A and co-administration with levetiracetem,lamotrigine,and topiramate did not significantly influence adult MHD CL/F.5.The model development dataset included 301 serum samples from 141 paediatric patients with epilepsy,the final model of pediatric patients was:CL/F（L/h）=（?）V/F（L）=14.7The final model revealed that BW was a significant factor which influenced paediatric MHD CL/F,BW-dependent exponent model was utilized to describe the effect of BW on CL/F,and the other covariates such as polymorphisms of UGT2B7 802T>C,UGT1A9 I399C>T,ABCB1 3435C>T,ABCB2 1249G>A and co-administration with levetiracetem,lamotrigine,and topiramate did not significantly influence paediatric MHD CL/F.6.Graphic method presented that the final model exhibited better fit than the basic model for data of adult population;the success fitting rate of Bootstrap method was97.50%,the relative bias between the median PK parameters calculated by the bootstrap method and the corresponding NONMEM results of the final model were all less than10%;the NPDE method revealed that the final model had good internal predictive performance;the external validation showed that the final model had better predictive capability than the basic model in terms of accuracy and precision（（MPE%=-1.50%,MAE%=18.12%）vs.（MPE%=2.34%,MAE%=23.58%））.7.Graphic method presented that the final model exhibited better fit than the basic model for data of paediatric population;the success fitting rate of Bootstrap method was92.10%,the relative bias between the median PK parameters calculated by the bootstrap method and the corresponding NONMEM results of the final model were all less than10%;the NPDE method revealed that the final model had good predictive performance;the external validation showed that the final model had better predictive capability than the basic model in terms of accuracy and precision（（MPE%=1.90%,MAE%=16.88%）vs.（MPE%=9.09%,MAE%=25.15%））.8.The new dosage guidelines were proposed based on the final models for individualising initial OXC regimens of adult and paediatric patients.9.The developed software realized the functions such as input and management of patient information,prediction of trough concentrations under various dosing regimens which could help initial dosage design,and prediction of trough concentrations more accurately based on Therapeutics drug monitoring results and Bayesian method which could help dosage adjustment.Conclusions1.The established PPK models for adult and paediatric epileptic patients were effective,stable and had good predictive performance,and could estimate MHD CL/F and individualise OXC dosage regimens accurately.2.The new developed software based on our established PPK models could provide a useful tool in the clinical setting to facilitate the individualised therapy for epilepsy patients. |
PDF Full Text Request