真实世界中苦碟子注射液不同疗程对肝肾功能指标变化影响的分析

Data analysis of real world clinical changes in indexes of liver and kidney function to use of parenterally administered Kudiezi at different doses and time periods

该研究是基于全国18家大型三甲医院信息管理系统(hospital information system,HIS)中所有使用苦碟子注射液(KDZ)住院患者临床用药信息数据仓库的回顾性分析。根据使用KDZ不同疗程分为2组,1 707例病人使用疗程≤14 d,275例病人使用疗程>14 d。运用倾向性评分方法对比2组人群在用药后肝功能指标谷丙转氨酶(ALT)和谷草转氨酶(AST)、肾功能指标肌酐(Cr)和尿素氮(BUN)的变化,为KDZ临床安全用药提供有用信息。运用generalized boosted models(GBM)倾向评分加权法,消除71个已知混杂因素在组间的差异,如年龄、性别、住院病情等,使结果更加贴近临床真实情况。同时使用另外2种Logistic回归分析确证GBM的分析结果,发现2组人群肝肾功能指标除了AST估计出来的平均处理效应有统计学差异外(P<0.05),其余3个指标均无统计学差异。这不足以说明疗程>14 d使用KDZ会导致肝肾功能发生异常变化。倾向性评分方法为一种有效控制混杂偏倚的统计分析方法,但真实世界还存在诸多未知潜在混杂因素,因此有待将来前瞻性研究或更高级的混杂因素分析方法进行验证。

This is a retrospective study based on a 18 hospital information system data warehouse. Records of 1 982 patients who used Kudiezi intravenous infusion （KDZ） were extracted from the data warehouse. All the patients were divided into two groups, one group of 1 707 patients used KDZ, ~〈 14 days under the instruction, the other group of 275 patients used KDZ, 〉 14 days, off label use. Generalized boosted models （GBM） with propensity score were applied to compare the two groups on four indexes of liver and kid- ney functions, such as alanine aminotransferase（ALT）, aspartate aminotransferase （AST）, creatinine （Cr） and blood urea nitrogen （BUN）. Seventy one confounder~ were identified and balanced by GBM. And another two logistic analysis methods were used to con- firm the results from GBM. The results indicated that there were no significant difference on the four indexes except the AST （P 〈 O. 05 ） by all the three analyzing models. It is hard to conclude that ultra long, off label using Kudiezi intravenous infusion could influ- ence the four indexed of liver and kidney from this data analysis. More conclusive evidence should be collected by further prospective study.