重症感染患者多黏菌素B群体药代动力学模型的外部验证

External validation of the population pharmacokinetic models for polymyxin B in patients with severe infection

目的:对既往发表的多黏菌素B群体药代动力学(population pharmacokinetics, PPK)模型进行外部验证,评估各模型对重症感染患者血药浓度的预测性能。方法:在PubMed数据库、Web of Science核心合集、中国知网、万方数据库检索建库至2022年3月公开发表的多黏菌素B PPK模型的相关文献,提取模型结构和参数信息。同时收集南京大学医学院附属鼓楼医院接受多黏菌素B治疗重症感染患者的临床数据作为外部数据集。根据患者是否采用连续肾脏替代疗法(continuous renal replacement therapy, CRRT)分为非CRRT组和CRRT组。利用2组数据对已发表模型进行基于模型预测和模型模拟的评价。结果:检索到14个多黏菌素B PPK模型。CRRT组,预测误差检验结果显示,所有已发表模型均不满足选定的标准。非CRRT组,有一个模型具有较好的血药浓度预测性能。结论:可以尝试利用预测结果良好的模型在非CRRT患者进行前瞻性个体化给药。

OBJECTIVE To evaluate the predictive performance of the previously published population pharmacokinetic(PPK) models for polymyxin B plasma concentration in patients with severe infection.METHODS The literature on PPK model of polymyxin B was retrieved from PubMed database, Web of Science Core Collection, CNKI and Wanfang database from inception to March 2022. Model structure and parameter information were extracted. Meanwhile, the patients with severe infection treated with polymyxin B in Nanjing Drum Tower Hospital were included, and the clinical and demographic information were collected as the external data set. According to the adoption of continuous renal replacement therapy(CRRT), patients were divided into non-CRRT group and CRRT group. The approaches of model prediction and model simulation were conducted to evaluate the predictive performance.RESULTS A total of 14 PPK models of polymyxin B were retrieved. In the CRRT group, the prediction errors showed that none of the published model met the predefined criteria.In the non-CRRT group, one model achieved satisfied prediction performance for polymyxin B plasma concentration.CONCLUSION The model with good predictive results can be used for prospectively individualized dosing of polymyxin B in the non-CRRT patients.