静脉血栓栓塞症风险评估工具的系统评价

Venous thromboembolism risk assessment tools:a systematic review

目的系统评价静脉血栓栓塞症(venous thromboembolism,VTE)风险评估工具。方法计算机检索Embase、PubMed、CNKI、CBM、WanFang Data、VIP数据库和22个相关机构与学会网站,搜集VTE风险评估工具,检索时限均为建库至2022年12月31日。由2位研究者独立筛选文献、提取资料并交叉核对。定性分析发布国家、机构、年份、适用疾病类型、适用人群、工具形成方法等基本特征;重点比较VTE风险评估工具的评估维度、工具性能、风险分层能力、是否验证等关键要素。结果共纳入42个VTE风险评估工具,其中美国最多(n=17),中国4个,集中发布于1996—2021年,适用疾病类型和人群不完全相同;其中19个工具构建基于病例-对照或回顾性队列研究,16个基于前瞻性队列研究,5个基于横断面研究和随机对照试验;另外基于Logistic回归模型构建的工具达20个;各工具评估维度不同,出现频次较高的为VTE既往史、年龄、BMI值和确诊为肿瘤,分别占比64.29%、54.76%、54.76%和47.62%。33个工具进行了风险分层,30个工具以风险评分的方式呈现;部分工具缺失临床验证数据,同时进行了特异性、敏感性、阴性预测值、阳性预测值和曲线下面积分析的工具仅有12个。结论现有VTE风险评估工具的评估维度和证据来源不完全一致、工具实施方法和结果呈现形式不完全相同,适用范围有所差别;部分工具构建方法和过程欠清晰,缺乏相关外部效度的验证研究,在我国临床实践推广存在一定局限性。

Objective To systematically review venous thromboembolism(VTE)risk assessment tools.Methods The Embase,PubMed,CNKI,CBM,WanFang Data,VIP databases and 22 relevant institutions and associations were searched to identify all VTE assessment tools from inception to December 31,2022.Two researchers independently screened the literature,extracted data,and cross-checked the data.A qualitative analysis was used to describe the country's essential characteristics,publishing organization,year,applicable disease type,applicable population,tool formation method,etc.Key elements and techniques were compared in terms of evaluation dimension,methods,and procedures to form the tool,risk stratification ability,and whether to verify.Results A total of 42 VTE risk assessment tools were included,of which 16 were in the United States,and only 4 were in China.They were released between 1996 and 2021,and the applicable disease types and populations differ.Nineteen tools were constructed based on case-control or retrospective cohort studies,16 were conducted using prospective cohort studies,and 5 were based on cross-sectional and RCT studies;Additionally,20 tools were built based on logistic regression models;The evaluation dimensions of each tool differed,and the most common frequency of occurrences were VTE history,age,BMI value,and confirmed tumor,accounting for 64.29%,54.76%,54.76%,and 47.62%,respectively.Thirty-three tools were stratified for risk,and 30 tools were presented in the form of risk scores;Some tools lacked clinical validation data,and only 12 tools were analyzed for specificity,sensitivity,NPV,PPV,and AUC.Conclusion The evaluation dimensions and evidence sources of existing VTE risk assessment tools are not completely consistent,the implementation methods and results presentation forms of the tools are not completely the same,and the scope of application is different;Some tool construction methods and processes are not clear enough,and there is a lack of validation research on external validity,which has certain limitations in promoting clinical practice in China.