首页|个体化给药辅助决策系统软件对卡马西平稳态谷浓度的预测能力分析

个体化给药辅助决策系统软件对卡马西平稳态谷浓度的预测能力分析

扫码查看
目的 验证群体药代动力学软件JPKD对卡马西平稳态谷浓度的预测能力,并分析其影响因素.方法 收集 2017 年 1 月~2022年 12 月使用卡马西平治疗并同时进行血药浓度监测(TDM)的患者信息.依据JPKD软件确定卡马西平初始给药方案并预测或调整稳态谷浓度.采用卡马西平预测稳态谷浓度(C预)与实测稳态谷浓度(C实)的权重偏差(WRES)评估JPKD软件对卡马西平稳态谷浓度的预测能力.根据WRES将初始方案TDM数据分为预测准确组(WRES<20%)与预测不准确组(WRES≥20%),记录两组患者的年龄、性别、体重等指标以及卡马西平用药及合并用药情况和TDM结果.上述两组变量采用差异性分析方法比较其差异性,而后运用多因素Logistic回归分析筛选影响JPKD软件预测卡马西平稳态谷浓度的因素变量.结果 共纳入 170 例患者,收集卡马西平的TDM数据 209 例次,其中初始方案 170 例次,调整方案 39 例次.与初始方案相比,根据初始C实调整方案后的WRES显著减小[10.48%(9.01%,15.85%)vs.15.55%(12.15%,20.56%),P<0.001],WRES<20%的比例显著升高[84.62%(33/39)vs.57.65%(98/170),P<0.01],提示JPKD软件预测调整方案卡马西平稳态谷浓度的准确性优于对初始方案稳态谷浓度的预测.预测准确组 98 例次,预测不准确组 72 例次;差异性分析比较结果显示:使用和未使用氯硝西泮组预测准确率分别为 12.50%、59.88%,两组预测准确率差异有统计学意义(差值为 0.47,差值95%CI 0.23~0.71,P<0.05);多因素Logistic回归分析显示:合并服用氯硝西泮(OR=9.17,95%CI为 1.04~80.78,P=0.046)是JPKD软件预测不准确的独立危险因素.结论 JPKD软件对卡马西平调整给药方案后的稳态谷浓度的预测准确能力较初始方案更高;JPKD软件对使用氯硝西泮患者的预测能力较差.
Predictive performance of JPKD software on carbamazepine steady-state trough concentration
Objective To estimate the predictive performance of the population pharmacokinetics software JPKD on predicting the vancomycin steady-state trough concentration,and and to analyze its influencing factors.Methods The information of inpatients who were treated with CBZ from January 2017 to December 2022 and admitted therapeutic drug monitoring(TDM)at the same time were collected.Based on the JPKD software,the initial dosage regimen of CBZ is determined and the steady-state trough concentration is predicted or adjusted.The weight residual(WRES)between the predicted steady-state trough concentration(Cpre)and the measured steady-state trough concentration(Creal)was used to evaluate the ability of the JPKD software for predicting the CBZ steady-state trough concentration.The TDM results of initial regimen were divided into accurate prediction group(WRES<20%)and the inaccurate prediction group(WRES≥20%)according to the WRES value.Patient characteristics including age,gender,weight,carbamazepine therapy,other concomitant medications therapy and TDM results were collected from electronic medical records.The difference analysis method was adopted to compare the differences between the two groups of variables and multivariate Logistic regression analysis was used to screen the related factors that influence the predictive performance of JPKD software.Results A total of 170 patients were enrolled,and 209 steady-state trough concentrations of CBZ were collected,including 170 concentrations of initial regimen and 39 concentrations of adjustment regimen.Compared with the initial regimen,the WRES of adjusted regimen was significantly reduced[10.48%(9.01%,15.85%)vs.15.55%(12.15%,20.56%),P<0.01],and the proportion of WRES<20%increased significantly[84.62%(33/39)vs.57.65%(98/170),P<0.01].These results indicated that JPKD software had a better accuracy prediction for steady-state trough concentration of the adjusted regimen than the initial regimen.There were 98 concentrations in the accurate prediction group and 72 in the inaccurate prediction group.Correlation analysis showed that the prediction accuracy of the group with and without clonazepam was 1.02%and 98.98%,respectively.There was a statistical difference in the prediction accuracy between the two groups(the difference was 0.98,and 95%CI was 0.78~1.17,P<0.05).Multivariate Logistic regression analysis indicated that clonazepam(OR=9.17,95%CI 1.04~80.78,P=0.046)were independent risk factors for inaccurate prediction of JPKD software.Conclusion JPKD software had a better predictive performance for the CBZ steady-state trough concentrations of adjustment regimen than initial regimen.JPKD software had a poor predictive performance while the patient was given clonazepam.

CarbamazepineEpilepsyPopulation pharmacokineticsJPKDPredictive abilityHigh risk factor

张芹、霍继浩、韩海燕、原君丽、杨东斌

展开 >

鹤壁市人民医院临床药学室,河南 鹤壁 458030

卡马西平 癫痫 群体药代动力学 JPKD 预测能力 高危因素

2024

中国处方药
南方医药经济研究所

中国处方药

影响因子:0.649
ISSN:1671-945X
年,卷(期):2024.22(12)
  • 1