使用模拟目标试验框架对脑出血患者是否应停用他汀类药物相关研究的方法学质量评价

Methodological Quality Evaluation on Studies Related to Whether Statins Should be Discontinued in Patients with Intracerebral Hemorrhage by Using Target Trial Emulation Framework

目的:使用模拟目标试验(TTE)框架,对脑出血(ICH)患者是否停用他汀类药物的观察性研究进行方法学质量评价,总结可以避免的常见方法学缺陷,为后续研究设计提供方法学参考。方法:通过文献研究法筛选符合纳入与排除标准的文献,使用TTE框架评价所纳入文献的方法学质量,包括纳入与排除标准、治疗策略、随机分配方法、随访的开始及结束、结局指标、因果对比、统计分析方案等7个方面。结果:共纳入8篇队列研究,将患者分为两组,其中暴露组患者停用他汀类药物,非暴露组患者继续使用他汀类药物。方法学评价存在以下问题,2项研究将基线后的信息作为排除标准;4项研究未具体描述治疗策略;4项研究无模拟随机分配的方法;3项研究的协变量选择没有明确的文献依据;1项研究未描述随访的时间节点;所有研究均未明确指出采用意向性治疗(ITT)分析或符合方案(PP)分析;仅2项研究描述了处理缺失数据的方法,但选择了直接删除有缺失数据的患者。结论:未来ICH后中止他汀类药物对临床预后影响的类似研究应避免使用基线后的信息来定义纳入与排除标准,明确因果推断(理想情况下同时包括ITT和PP分析),并考虑使用文献综述或临床经验进行协变量的筛选(最好使用有向无环图)。

OBJECTIVE:By using target trial emulation(TTE)framework,to perform methodological quality evaluation on observational studies related to whether statins should be discontinued in patients with intracerebral hemorrhage(ICH),so as to summarize common methodological deficiency that can be avoided and provide methodological references for subsequent study design.METHODS:Literature research method was used to screen the studies that meet the inclusion and exclusion criteria,the TTE framework was used tot evaluate the methodological quality of the enrolled studies,including seven aspects of inclusion and exclusion criteria,treatment strategy,random allocation method,start and end of follow-up,outcome indicator,causal comparison and statistical analysis protocols.RESULTS:A total of 8 cohort studies were enrolled,and the patients were divided into two groups,with patients in the exposed group discontinuing statins and patients in the non-exposed group continuing to use statins.The methodological evaluation presented the following problems,2 studies used post-baseline information as an exclusion criterion;4 studies did not specifically describe the treatment strategy;4 studies had no simulate random allocation method;3 studies had no clear literature basis for the selection of covariates;1 study did not describe the time points of follow-up;and none of the studies explicitly stated that an intention-to-treat(ITT)analysis or a protocol-conformant(PP)analysis was used.Only 2 studies described the method of handling missing data,by choosing to directly delete patients with missing data.CONCLUSIONS:Future similar studies on the clinical prognostic impact of discontinuing statins after ICH should avoid using post-baseline information to define inclusion and exclusion criteria,define causal inferences(ideally including both ITT and PP analyses),and consider to use literature reviews or clinical experience for covariate screening(preferably using directed acyclic graphs).