【摘要】 目的 评价我院在地中海贫血患儿造血干细胞移植（HSCT）术后群体中建立的伏立康唑群体药代动力学（PPK）模型的预测性能。方法 前瞻性收集接受HSCT并使用伏立康唑预防或治疗侵袭性真菌感染的地中海贫血儿童患者，基于前期建立的PPK模型，应用最大后验贝叶斯（MAPB）法估算个体预测值和一个给药间隔内游离药物浓度-时间曲线下面积（f AUCτ），采用基于模型预测和模拟的方法对模型进行性能评价。结果 共纳入28例患儿，获得118个伏立康唑血药浓度值，其中谷浓度49个，峰浓度30个，计算10例患儿伏立康唑的f AUCτ。伏立康唑浓度值的个体相对预测误差中位数（MDIPE）、个体绝对预测误差中位数（MDAIPE）、个体相对预测误差在±20%的比例（F20）和个体相对预测误差在±30%的比例（F30）分别为-3.2%、27.7%、35.6%、52.5%。模型对谷浓度和峰浓度预测误差在可接受范围。f AUCτ估算平均值和实测平均值分别为9.83 mg·L-1和13.82 mg·L-1，相对预测误差中位数（MDPE）和绝对预测误差中位数（MDAPE）分别为-22.7%和22.7%。拟合优度检验和可视化预测检验表明模型稳定、预测性能可靠。结论 在地中海贫血患儿HSCT术后群体中建立的伏立康唑PPK模型具有较好的预测性能，可用于辅助临床制订个体化给药方案。更多还原

【Abstract】 Objective To evaluate the prediction of voriconazole population pharmacokinetic(PPK)model in children with thalassemia after hematopoietic stem cell transplantation(HSCT) in our hospital. Methods Children with thalassemia undergoing HSCT who received voriconazole to prevent or treat invasive fungal infections were prospectively collected. Based on the PPK model,the maximum posteriori Bayesian method was used to estimate the individual predictive value and the area under the concentration-time curve of free drug in one administration interval(f AUCτ). The performance of the model was evaluated based on model prediction and simulation. Results Totally 118 voriconazole blood concentration values in 28 children were included, including 49 trough concentrations and 30 peak concentrations, and f AUCτ of 10 children was calculated. The median of individual relative prediction error, median of individual absolute prediction error, proportion of individual relative prediction error between±20%(F20) and proportion of individual relative prediction error between±30%(F30) of voriconazole concentration were-3.2%, 27.7%, 35.6%and 52.5%, respectively. The prediction errors of trough concentration and peak concentration were acceptable. The predicted and measured mean values of f AUCτ were 9.83 mg·L-1 and 13.82mg·L-1, respectively. The median of relative prediction error and median of absolute prediction error of f AUCτ were-22.7% and 22.7%, respectively. The goodness of fit plots and visual predictive check showed that the model was stable and the prediction was reliable. Conclusion The PPK model of voriconazole in children with thalassemia undergoing HSCT established in the early stage can predict and help individualize dosage regimen.更多还原