Abstract: Objective　To systematically evaluate the diagnostic value and efficiency of the water swallowing test （WST） and standardized swallowing assessment （SSA） for respiratory aspiration in patients with stroke.Methods　Studies published from 1980 to 2017 regarding the application of WST and SSA for respiratory aspiration in patients with stroke were retrieved through searching of CNKI，VIP，the Wanfang database，Cochrane Library，PubMed and Web of Science.Literature screening was performed by two researchers strictly following the literature inclusion criteria recommended by the Cochrane Collaboration Network.The key information in the included studies，including true positive value，false positive value，true negative value，false negative value，sensitivity and specificity，were extracted.The bias risk of the included studies was assessed.Meta-analysis was conducted in Meta-DiSc version 1.4，Stata 12.0 and RevMan 5.3 software.Results　Fourteen studies were included；the pooled sensitivity of WST （<50 ml） in diagnosing respiratory aspiration in patients with stroke was 0.54 〔95%CI（0.48，0.61）〕，and the pooled specificity was 0.76 〔95%CI（0.69，0.82）〕.The pooled sensitivity of WST （≥50 ml） in diagnosing respiratory aspiration in patients with stroke was 0.80 〔95%CI（0.72，0.86）〕，and the pooled specificity was 0.67 〔95%CI（0.60，0.73）〕.For SSA，the pooled sensitivity and specificity were 0.86 〔95%CI（0.82，0.90）〕 and 0.69 〔95%CI（0.64，0.73）〕.The pooled sensitivity of WST （<50 ml） and SSA were significantly different （Z=-2.449，P<0.05），whereas WST （≥50 ml） and SSA were not significantly different in pooled sensitivity （Z=-0.980，P>0.05）.WST （<50 ml） and WST（≥50 ml） were significantly different in pooled sensitivity （Z=-2.193，P<0.05）.For pooled specificity，WST （<50 ml）
and SSA were not significantly different （Z=-0.735，P>0.05），and WST （≥50 ml），SSA （Z=-0.490，P>0.05），WST （<50 ml） and WST （≥50 ml） all showed no significant differences in pooled specificity （Z=-0.104，P>0.05）.The AUCs of WST （<50 ml），WST （≥50 ml） and SSA in the SROC curves were 0.700，0.818 and 0.858，respectively.Conclusion　The diagnostic efficiency of WST in detecting respiratory aspiration in patients with stroke depends on the drinking amount.The accuracy of WST （≥50 ml） and SSA is better than that of WST （<50 ml）.

Key words: Stroke, Respiratory aspiration, Multiphasic screening, Sensitivity, Specificity, Meta-analysis

摘要： 目的　系统评价洼田饮水试验（WST）和标准吞咽功能评估（SSA）对脑卒中患者误吸筛查的诊断价值与效能。方法　计算机检索中国知网、维普网、万方数据知识服务平台、Cochrane Library、PubMed、Web of Science数据库1980—2017年公开发表的WST和SSA对脑卒中患者误吸筛查诊断准确性的文献，由两名研究者严格按照Cochrane协作网关于诊断性试验推荐的文献纳入标准独立地进行文献筛选，提取各纳入研究中包含的诊断信息：真阳性值、假阳性值、真阴性值、假阴性值、灵敏度、特异度，并评价纳入研究的偏倚风险，采用Meta-DiSc version 1.4软件、Stata 12.0软件、RevMan 5.3软件进行Meta分析。结果　共纳入文献14篇，WST（<50 ml）对脑卒中患者误吸筛查的合并灵敏度为0.54〔95%CI（0.48，0.61）〕，合并特异度为0.76〔95%CI（0.69，0.82）〕；WST（≥50 ml）对脑卒中患者误吸筛查的合并灵敏度为0.80〔95%CI（0.72，0.86）〕，合并特异度为0.67〔95%CI（0.60，0.73）〕；SSA对脑卒中患者误吸筛查的合并灵敏度为0.86〔95%CI（0.82，0.90）〕，合并特异度为0.69〔95%CI（0.64，0.73）〕。WST（<50 ml）与SSA合并灵敏度比较，差异有统计学意义（Z=-2.449，P<0.05）；WST（≥50 ml）与SSA合并灵敏度比较，差异无统计学意义（Z=-0.980，P>0.05）；WST（<50 ml）与WST（≥50 ml）合并灵敏度比较，差异有统计学意义（Z=-2.193，P<0.05）。WST（<50 ml）与SSA合并特异度比较，差异无统计学意义（Z=-0.735，P>0.05）；WST（≥50 ml）与SSA合并特异度比较，差异无统计学意义（Z=-0.490，P>0.05）；WST（<50 ml）与WST（≥50 ml）合并特异度比较，差异无统计学意义（Z=-0.104，P>0.05）。集成受试者工作特征（SROC）曲线显示WST（<50 ml）、WST（≥50 ml）、SSA对脑卒中患者误吸筛查的曲线下面积分别为0.700、0.818、0.858。结论　WST对脑卒中患者误吸筛查的诊断效能与其饮水量有关，WST（≥50 ml）和SSA对脑卒中患者误吸筛查的准确性优于WST（<50 ml）。

关键词: 卒中, 误吸, 多相筛查, 灵敏度, 特异度, Meta分析