经导管主动脉瓣置换术后死亡风险预测模型系统评价

Mortality Risk Prediction Models After Transcatheter Aortic Valve Replacement:A Systematic Review

目的系统评价经导管主动脉瓣置换术后死亡风险预测模型。方法在Cochrane Library、EMBASE、PubMed、Web of Science、中国生物医学文献数据库、中国知网、维普和万方数据库中检索经导管主动脉瓣置换术后死亡风险预测模型,检索时限为2000年1月1日—2021年2月1日。由两名研究者独立筛选文献和提取数据,再采用PROBAST工具对纳入文献中的模型进行质量评价。结果共纳入14项研究,其中11项为前瞻性队列研究,3项为回顾性研究;12个模型进行了内部或外部验证,3个模型未验证;纳入模型最常见的预测因子依次为年龄、非股动脉入路、肾小球滤过率、肺动脉高压、体质量指数、NYHA心功能Ⅳ级、性别、左室射血分数、血液透析、血清白蛋白以及肺部疾病;住院、术后30 d和1年的死亡风险预测模型的曲线下面积分别为0.66~0.74、0.59~0.76和0.60~0.83。偏倚结果方面,有3项研究的偏倚风险较低,3项研究的偏倚风险不清楚,8项研究的偏倚风险较高。偏倚主要来源于数据分析部分,包括缺失数据的处理和预测变量的选择等。结论经导管主动脉瓣置换术后死亡风险预测模型的研究还处于发展阶段,未来应开发性能优良和偏倚风险低的本土预测模型,并进行内部验证和外部验证。

Objective To systematically review the mortality risk prediction models after transcatheter aortic valve replacement.Methods We searched Cochrane Library、EMBASE、PubMed、Web of Science,CBM,CNKI,VIP and Wanfang Data from January 1,2000 to February 1,2021.Two researchers independently screened the literature,extracted information,and then used prediction model risk of bias assessment tool(PROBAST)to evaluate the quality of the models included in the literature.Results A total of 14 studies were included,of which 11 were prospective cohort studies and 3 were retrospective studies.12 models were internally or externally verified,while 3 models were not verified.The most common predictors in the included model followed by age,non-femoral approach,glomerular rate filtration,pulmonary hypertension,body mass index,NYHA function classⅣ,gender,left ventricular ejection fraction,hemodialysis,serum albumin,and lung disease.The area under the curve of the mortality risk prediction models for hospitalization,30 days,and one year after surgery were 0.66 to 0.74,0.59 to 0.76,and 0.60 to 0.83 respectively.In terms of results of bias,three studies had a low risk of bias,three studies had an unclear risk of bias,and 8 studies had a higher risk of bias.Bias mainly comes from the data analysis,including the processing of missing data and the selection of predictive variables.Conclusion The research on the prediction model of death risk after transcatheter aortic valve replacement is still in the development stage.In the future,a local prediction model with excellent performance and low risk of bias should be developed and internally or externally verified.