The Methodology Of Data Management And Analysis For Population Pharmacokinetics/pharmacodynamics Data

It has been several decades since the population principle was firstpublished. With the more and more complicated clinical trials areperformed, these certainly induce to enlarge the database to collect theincreasing values. All the researchs in the field may be focused on datamanagement and data analysis of the population database. In this paper,we provided our solutions and tried to solve these problems.OBJECTIVE: To make up the insufficiencies of the traditionalpaper data capture and data management, a completely new system onelectronical data capture (EDC) and database management (DM) wasdeveloped and applied. By comparing several kinds of methods andsoftwares, we try to provid a feasible solution on data analysis andmodeling of population data. We also tried to improve the efficacy andfacility of the population data modeling. METHODS:â‘ Based on the functionalities of databasemanagement and maintain provided by Microsoft SQL Server, wedeveloped and applied a web-based data capture and management systemnamed CT-DMS. All the following functionalities were integrated to thesystem, such as project architecture, electronical data capture, role andpermission management, double data entry, double check, data correctionplatform and reporting etc.â‘¡We compared results of the most twopopular methods, standard two steps (STS) and nonlinear mixed effectmodel (NONMEM), to deal with analysis and modeling for the data ofKAV-RPZ. It showed that STS could only provid the pharmacokineticsparameters of each subject enrolled, inter-individual error, and covariable.In the KAV-RPZ project, STS can also give us the linear relationshipbetween the partial parameters and doses. By using NONMEM method,we tried to model the serial models of KAV-RPZ population data, such aspopulation pharmacokinetics model (PPK), populationpharmacodynamics model (PPD) and PPK-PPD link model. According tothese models, we tried further to find out the population structuralparameters, fix effects and the distribution of errors.â‘¢To realize thefollowing expected functionalities, such as management of populationprojects, automatic generation of contol stream files, enhancedmanagement of the output files etc., we used C#.NET with S-PLUS todevelop two softwares: Data Format Transfer (DFT) and Win-NONMEM.RESULTS:â‘ Developed and applied severals softwares to solvethe problems occurred in data management and analysis of the populationclinical trial phaseâ… of KAV-RPZ project.â‘¡Used STS mothod to fitthe data of KAV-RPZ, got the pharmacokinetics parameters of eachsubject and inter-individual error. Successfully found a significantlycovariable, the genotype of CYP2C 19. Also found the linear relationshipbetween doses and the pharmacokinetics parameters AUC, AUC_e and C_max.â‘¢Successfully found serial models for KAV-RPZ (High dose) withNONMEM mothod, included PPK model, PPD model and PPK-PPD linkmodel. The PPK model of KAV-RPZ is a one-order absorption,one-compartment with two mixed effects genotype of CYP2C19 andbody weight. The PPK-PPD link model is full Sigmoid model witheffect-compartment. These models were verified by the other group (Lowdose).CONCLUSION:â‘ CT-DMS provides necessary technicalsupport for the quality control of the database, and makes it possible toachieve the clinical trial phaseâ… of KAV-RPZ ahead of schedule. Thesystem is powerful enough to be extended to other large scale trials.â‘ Web-based EDC and CDM technic with the mid-term data analysis canhelp the sponsor by providing early visibility to clinical data for fasterdecision-making, optimized resources and lower execution risk.â‘¢By using STS and NONMEM methods, we built a serial of models ofKAV-RPZ, these models are representative and valuable. It also provideda feasible solution for population data modeling.â‘£Using NONMEMcombined with DFT and Win-NONMEM software, it makes the analysisand modeling more convenient and efficient.