论文部分内容阅读
目的建立万古霉素在新生儿和小婴儿患者的群体药动学(PPK)模型,为临床个体化用药提供参考。方法收集85例新生儿科患者静脉注射使用万古霉素后的血药浓度数据和临床资料。将患者分为两组,模型组(n=71)采用Phoenix~NLME~(TM)1.3软件进行PPK分析,建立一房室药动学模型(个体间变异采用指数模型,个体内变异采用混合误差模型),考察各协变量对参数V和CL的影响。用拟合优度、自举法对最终模型的性能进行内部验证。采用验证组(n=14)患者的血药浓度,计算平均预测误差(MPE)、平均绝对预测误差(MAE)、平均预测误差均方(MSPE)对最终模型进行外部验证。结果 PPK最终模型为V(L)=3.167,CL(L·h~(-1))=0.413×(WT/3.32)~(0.747)×(PNA/25)~(0.402)×e~(ηCL),体重(WT)和产后日龄(PNA)对CL有影响。拟合优度、自举验证的结果表明最终模型稳定、预测结果可靠。外部验证最终模型计算MPE、MAE和MSPE值分别为(-0.843±1.347)、(1.462±1.175)和(2.432±4.293)mg·L~(-1)。血药浓度实测值和最终模型的个体预测值的决定系数R=0.955,外部验证说明最终模型预测准确度较好。结论本研究建立的万古霉素新生儿和小婴儿患者的PPK模型预测能力和稳定性良好,可为其个体化给药方案的制订提供参考。
Objective To establish a population pharmacokinetic (PPK) model of vancomycin in neonates and small infants, and provide a reference for clinical individualized medication. Methods Blood samples were collected from 85 neonate pediatric patients for intravenous injection of vancomycin and clinical data. The patients were divided into two groups. The model group (n = 71) was analyzed by PPK using Phoenix ~ NLME ~ (TM) 1.3 software to establish a one-compartment pharmacokinetic model (the exponential model was used for individual variation and the intramural variation was mixed Error model) to examine the influence of each covariate on parameters V and CL. Using the goodness of fit and bootstrap method to verify the performance of the final model. The final model was externally validated using the plasma concentration of patients in the validation group (n = 14), the mean predictive error (MPE), the mean absolute predictive error (MAE), and the mean forecast error square (MSPE). Results The final model of PPK was V (L) = 3.167, CL (L · h -1) = 0.413 × (WT / 3.32) ~ (0.747) × (PNA / 25) ~ (0.402) × e ~ (ηCL ), Body weight (WT) and postnatal day (PNA) had effects on CL. Goodness of fit, bootstrap verification results show that the final model is stable and the prediction result is reliable. The final MPE, MAE and MSPE values calculated by external validation were (-0.843 ± 1.347), (1.462 ± 1.175) and (2.432 ± 4.293) mg · L -1, respectively. The coefficient of determination (R = 0.955) between the measured value of the plasma concentration and the individual prediction value of the final model indicates that the final model has better prediction accuracy. Conclusion The PPK model of vancomycin in infants and young infants established in this study has good predictive ability and stability, which can provide a reference for the formulation of individualized dosing regimens.