中文摘要：

【摘要】目的评价肺炎链球菌尿抗原检测（SpUAT）对成人社区获得性肺炎链球菌肺炎（SPP）的诊断价值。方法通过计算机检索和手工检索方法，收集中国知网、中国生物医学文献数据库、中文科技期刊全文数据库、万方数据库等中文文献数据库，以及美国国立医学图书馆PubMed、美国EBSCO学术信息数据库、荷兰ElsevierScience学术期刊数据库、美国OVID医学电子期刊数据库、德国Springer电子期刊数据库等英文数据库发表的采用SpUAT诊断成人社区获得性SPP的中文和英文临床研究。按纳入与排除标准选择文献，采用Meta—disc1．4和Stata12．0软件对数据进行统计分析，计算合并敏感度、合并特异度、阳性似然比、阴性似然比、诊断优势比（DOR），并绘制合并受试者工作特征曲线（SROC），计算SROC曲线下面积（AUC），分析SpUAT对社区获得性SPP诊断的准确性。结果共纳入15篇文献6866例患者，文献诊断性研究质量评价（QUADAS评分）均≥10分，提示文献质量较高。异质性检验结果显示，各研究间存在非阈值效应引起的异质性，采用随机效应模型进行Meta分析。结果显示，与常规病原学诊断方法（血培养、痰涂片、痰培养及气管支气管吸出物培养等）相比，SpUAT诊断社区获得性SPP的合并敏感度为73％[95％可信区间（95％CI）=71％．76％]、合并特异度为91％（95％CI=90％～92％）、合并阳性似然比为6．97（95％CI=4．13～11．77）、合并阴性似然比为0．30（95％CI=0．26—0．34）、合并DOR为24．34（95％CI=13．14～45．11）；SROC的AUC为0．8051±0．0362。异质性分析显示，实验设计（前瞻性或回顾性）、病例数（〈200例或≥200例）、参照检测方法（〈3种或≥3种）、文献来源（西班牙或非西班牙）4个因素对文献异质性均无影响，提示存在其他未知因素与纳入文献异质性有关。Deck漏斗

英文摘要：

Objective To evaluate the diagnostic accuracy of Streptocuccus pneumoniae urinary antigen test （SpUAT） in patients with community acquired Streptocuccus pneumoniae pneumoniae （SPP）. Methods The clinical studies relating SpUAT diagnostic accuracy for community acquired Streptocuccus pneumoniae infection were searched via computer and manual screening of Chinese databases including China National Knowledge lnternet （CNKI）, China Biology Medicine disc, China Technical Journal Full-text Database, and Wanfang Database as well as English databases such as PubMed, EBSCO, Elsevier Science, Ovid Technologies, and Springer. Data were extracted according to appropriate inclusion and exclusion criteria and analyzed with Meta-disc 1.4 and Stata 12.0 software. Pooled sensitivity, specificity, positive likelihood ratio, negative likelihood ratio as well as diagnostic odds ratio （DOR） were calculated, and summary receiver operating characteristic curve （SROC） was plotted and area under SROC （AUC） was calculated to analyze the diagnostic accuracy of SpUAT for adult community acquired SPP. Results A total of 15 studies enrolling 6 866 patients were included, all of which with a quality assessment of diagnostic accuracy studies （QUADAS） score not lower than 10, indicating a high study quality. Random effect model was applied owing to non-threshold effect associated heterogeneity. It was shown by Meta analysis that compared with routine pathogen diagnosis （hemocuhure, sputum smears, sputum culture and trachea and bronchus attractive culture）, the pooled sensitivity, pooled specificity, pooled positive likelihood ratio, and pooled negative likelihood ratio was 73% [95% confidence interval （95%CI） = 71%-76%,] 91% （95%CI = 90%-92%）, 6.97 （95%CI = 4.13-11.77）, and 0.30 （95%CI = 0.26-0.34）, respectively, pooled DOR was 24.34 （95%CI = 13.14-45.11）, and AUC of SROC was 0.8051 ± 0.0362. It was indicated by heterogeneity test that experiment design （prospective or retrospective）, sampl