个体化给药辅助决策系统pharmVAN、SmartDose对肾功能亢进患者万古霉素血药浓度预测能力的回顾性评估

目的：评估个体化给药辅助决策系统pharmVAN与SmartDose对肾功能亢进（ARC）患者万古霉素稳态谷浓度的预测能力。方法：以2018年4月 ~ 2022年3月在东部战区总医院住院且静脉使用万古霉素的ARC患者为研究对象进行回顾性研究，利用上述两款辅助决策系统预测万古霉素初始给药方案的稳态血药浓度，分析比较实测值与预测值之间的线性相关系数、绝对权重偏差和相对预测误差，以绝对权重偏差< 30%代表预测良好。结果：两款辅助决策系统的预测值与实测值之间均不存在相关性（pharmVAN，r = 0.143，P = 0.245；SmartDose，r = 0.122，P = 0.320）。SmartDose预测良好率高于pharmVAN（35.2% vs 17.6%，P = 0.020）；pharmVAN和SmartDose的绝对权重偏差分别为（57.58 ± 26.23）%、（74.45 ± 95.88）%，两组无统计学差异（P = 0.164）。pharmVAN的相对预测误差为（-48.01 ± 41.39）%，显著低于SmartDose的（40.11 ± 114.8）% （P < 0.001）。结论：个体化给药辅助决策系统pharmVAN和SmartDose对ARC患者万古霉素初始给药方案稳态谷浓度的预测能力不佳。

Predictive Performance of Individualization Dosage Auxiliary Systems PharmVAN and SmartDose on Plasma Vancomycin Concentrations in Patients with Augmented Renal Clearance: A Retrospective Study

Objective: To evaluate the ability of individualization dosage auxiliary systems ''pharmVAN'' and ''SmartDose'' to predict steady-state trough concentrations of vancomycin in patients with augmented renal clearance (ARC). Methods: A retrospective study was conducted on patients with ARC who were hospitalized in the General Hospital of Eastern Theater Command from April 2018 to March 2022 with intravenous use of vancomycin. The two systems were used to predict steady-state trough concentrations of vancomycin based on initial dosage regimens. The linear correlation, the absolute weight deviation and relative prediction error between the measured concentrations and the predicted concentrations were calculated. Absolute weight deviation < 30% represented a good forecast. Results: There was no linear correlation between the predicted values and the measured values for either of the system (pharmVAN, r = 0.143, P = 0.245; SmartDose, r = 0.122, P = 0.320). The good prediction rate of SmartDose was higher than that of pharmVAN (35.2% vs 17.6%, P = 0.020). There was no statistical difference (P = 0.164) between the absolute weight deviations of pharmVAN (57.58 ± 26.23)% and SmartDose (74.45 ± 95.88)%. The relative prediction error of pharmVAN (-48.01 ± 41.39)% was significantly lower than that of SmartDose (40.11 ± 114.8)% (P < 0.001). Conclusion: Neither of pharmVAN and SmartDose had good ability to predict steady-state trough concentrations based on initial vancomycin regimens in patients with ARC.