Page 69 - 《中国药房》2024年24期
P. 69
3 款个体化给药工具用于肾移植患者术后他克莫司的预测准确
性及影响因素分析
Δ
王国徽 ,李兴德,潘 娅,毛盼盼,张函舒,马雪娇,宋沧桑(昆明市第一人民医院药学部/云南省临床疾病个体
#
*
化用药研究中心,昆明 650224)
中图分类号 R979.5 文献标志码 A 文章编号 1001-0408(2024)24-3023-06
DOI 10.6039/j.issn.1001-0408.2024.24.10
摘 要 目的 评估JPKD、SmartDose、NextDose 3款个体化给药工具用于预测肾移植患者术后他克莫司给药剂量和血药浓度的
准确性,并分析影响预测准确性的因素。方法 回顾性纳入2021年1月-2023年6月在昆明市某三级甲等医院住院的肾移植术后
使用他克莫司治疗的成年患者的临床资料。使用JPKD、SmartDose、NextDose 3款软件分别预测他克莫司给药剂量和血药浓度,
计算实测值与预测值的绝对权重偏差(APE)、相对预测误差(PE)和预测良好率(APE<30%为预测效果良好);采用Pearson检验
或Spearman检验分析3款软件预测给药剂量与实际给药剂量、血药浓度预测值与实测值的相关性,采用单因素方差分析考察影响
3 款软件预测准确性的因素。结果 本研究共纳入 110 例住院患者,收集他克莫司给药剂量和血药浓度监测数据各 193 例次。
JPKD、SmartDose、NextDose 软件预测的他克莫司给药剂量分别为(2.0±0.7)、(2.7±1.9)、(1.8±0.8)mg,与实际给药剂量[(1.9±
0.6)mg]的相关系数分别为 0.841、0.450、0.247(P 均小于 0.001);APE 中位数分别为 6.00%、52.07%、30.40%,PE 中位数分别为
5.00%、18.50%、-3.50%,预测良好率分别为 98.45%、30.05%、49.22%。上述 3 款软件预测的他克莫司血药浓度分别为(6.74±
3.36)、(6.93±5.02)、9.00(5.80,12.60)ng/mL,与实测值[8.64(7.11,9.77)ng/mL]的相关系数分别为 0.997、-0.066、0.920(P 分别
为<0.001、0.360、<0.001);APE中位数分别为5.54%、45.91%、35.56%,PE分别为-4.94%(中位数)、-17.05%(中位数)和36.93%
(平均值),预测良好率分别为97.93%、32.64%、37.31%。单因素方差分析显示,给药剂量、血药浓度、体重、移植时间等与各软件的
预测准确性有关(P<0.05)。结论 3款个体化给药工具用于肾移植患者术后他克莫司给药剂量和血药浓度的预测良好率由高到
低均为JPKD、NextDose、SmartDose软件,临床可优先考虑JPKD软件。
关键词 他克莫司;给药剂量;血药浓度;JPKD软件;SmartDose软件;NextDose软件;肾移植术后
Analysis of predictive accuracy and its influential factors of three individualized administration tools for
tacrolimus after kidney transplantation
WANG Guohui,LI Xingde,PAN Ya,MAO Panpan,ZHANG Hanshu,MA Xuejiao,SONG Cangsang(Dept. of
Pharmacy, Kunming First People’s Hospital/Yunnan Provincial Research Center of Individualized Medication for
Clinical Diseases, Kunming 650224, China)
ABSTRACT OBJECTIVE To evaluate the accuracy of three individualized drug delivery tools, i. e. JPKD, SmartDose and
NextDose, in predicting tacrolimus dose and blood concentration after kidney transplantation, and analyze the influential factors of
prediction accuracy. METHODS The clinical data of adult hospitalized patients treated with tacrolimus after kidney transplantation
from January 2021 to June 2023 were retrospectively collected. Three individualized dosing tools, i. e. JPKD, SmartDose and
NextDose, were used to predict the dose and plasma concentration of tacrolimus. The absolute prediction error (APE) and
prediction error (PE) between the measured value and the predicted value, and prediction success rate were calculated (APE<30%
indicating a good forecast). Pearson assay or Spearman assay was used to analyze the correlation between the predicted dosage and
actual dosage, as well as the predicted and measured blood concentration values using three software; univariate analysis was used
to investigate the influential factors for prediction accuracy of JPKD, SmartDose and NextDose. RESULTS A total of 110
hospitalized patients were included in this study, and 193
Δ 基金项目 云 南 省 科 技 厅 重 大 科 技 专 项 计 划 项 目(No.
tacrolimus doses and plasma concentrations were monitored.
202302AA310018);云南省卫生健康委医学领军人才培养计划项目
(No.L-2018012);云南省卫生健康委临床药学中心建设项目;云南省 The predicted doses of JPKD, SmartDose and NextDose were
临床药学中心开放课题(No.2023CPC001-A-2);“春城计划”高层次人 (2.0±0.7), (2.7±1.9), (1.8±0.8) mg, their measured value
才创新创业团队专项(No.2022SCP002) was (1.9±0.6) mg, and the correlation coefficients between
* 第一作者 药 师 ,硕 士 。 研 究 方 向 :临 床 药 理 学 。 E-mail:
the predicted values and the measured value were 0.841,
1219816254@qq.com
# 通信作者 主任药师,硕士生导师。研究方向:临床药理学。E- 0.450, 0.247 (P<0.001); the median APEs were 6.00%,
mail:songcs163@163.com 52.07% and 30.40%, and the median PEs were 5.00%,
中国药房 2024年第35卷第24期 China Pharmacy 2024 Vol. 35 No. 24 · 3023 ·