绞窄性肠梗阻CT诊断的Meta分析

Meta-analysis of application of computed tomography in strangulated obstruction

目的采用循证医学Meta分析(Meta-analysis)的方法研究绞窄性肠梗阻CT诊断的准确性。方法收集1990～2011年国内外公开发表的绞窄性肠梗阻CT诊断的文献,并根据QUADAS质量评价标准评价文献的质量,分析偏倚和变异产生的来源,同时对文献间异质性来源进行分析,对产生阈值效应的异质性和其他异质性原因进行Meta-回归分析,并用定量方法SROC对有阈值效应的数值进行合并,对无阈值效应的数据进行随机效应分析(有异质性)和固定效应(无异质性)分析合并。数据分析采用Metadisc1.4,Ex-cell2007和Stata10.0软件,得到合并的SROC曲线,合并的诊断敏感性和特异性,以及合并的阳性和阴性似然比。结果共检索到文献261篇,符合纳入标准的文献共15篇,均为绞窄性肠梗阻的CT诊断准确且质量较高的研究,合并敏感性为0.698(95%CI0.655～0.739),合并特异性0.904(95%CI0.883～0.922),阳性似然比7.418(95%CI4.468～12.316),阴性似然比0.228(95%CI0.129～0.404)。结论肠梗阻的CT检查有较高的特异性,但研究有较多的异质性和偏倚,需要进一步研究。

Objective To summarize the accuracy of strangulated obstruction （SO）with meta-analysis for computed tomography examination. Methods Published studies were identified by Pub/Medline and Ovid databases and Chinese databases （1990-2011）to search SO for computed tomography. A standard extractive data were obtained. Inclusion and Extraction criteria were used to extract relevant data, QUADAS was analysed. Using software metadisc 1.4,Excell 2007 and StatalO. 0, the summary sensitivity and specificity and summary positive likelihood ratio, summary negative likelihood ratio were caculated, and heterogeneity, biase were also evaluated respectively. Results Of 261 identified studies, 15 involving adult patients were considered appropriate for pooled analysis, the pooled sensitivity was 0. 698（95% CIO. 655-0. 739） ,and the pooled specificity was 0. 904（95% CIO. 883-0. 922）, The pooled positive likelihood ratio was 7.418 （95 % CI 4. 468-12. 316）. The pooled negative likelihood ratio was 0. 228 （95% CI,0. 129-0. 404）. Conclusions There are high specificity for SO of CT, but relative low sensitivity; but there were still much biase in these studies,appropriate studies should be evaluated.