摘要
目的 分析γ-干扰素释放试验（IGRA,本研究采用IGRA-ELISA法）与齐尔-尼尔森（Ziehl-Neelsen）染色法（简称＆quot;Z-N法＆quot;）、结核抗体（TB-Ab）胶体金快速检测方法（简称＆quot;TB-Ab法＆quot;）和荧光探针PCR法（简称＆quot;荧光PCR法＆quot;）等检测技术在北京同仁医院应用于TB辅助诊断中的作用,对IGRA-ELISA检测技术作出客观评价.方法 选取本院2012年8月至2013年7月进行TB诊断的1228例患者作研究对象,其中包括IGRA-ELISA 827例、TB-Ab法306例、Z-N法324例和荧光PCR法83例,共1540例次.采用SPSS 16.0软件统计分析和绘制受试者工作特征（ROC）曲线,计数资料采用x2检验,P＜0.05为差异有统计学意义,分别计算各技术ROC曲线下的面积（AUC）,比较各技术辅助诊断价值的大小.结果 IGRA-ELISA的阳性检出率为36.0％ （298/827）,均高于TB-Ab法（6.2％,19/306）、Z-N法（3.1％,10/324）和荧光PCR法（3.6％,3/83）,差异有统计学意义（x2值分别为98.6、128.9、35.8,P值均＜0.001）.IGRA-ELISA的阳性诊断率10.3％ （85/827）,均高于TB-Ab法（1.3％,4/306）、Z-N法（2.5％,8/324）和荧光PCR法（3.6％,3/83）,与TB-Ab法和Z-N法相比较差异有统计学意义（x2值分别为24.8、19.1,P值均＜0.001）;而与荧光PCR法相比较差异无统计学意义（x2=3.8,P＞0.05）.IGRA-ELISA的敏感度为97.7％ （85/87）,均高于TB-Ab法（17.4％,4/23）、Z-N法（28.6％,8/28）和荧光PCR法（25.0％,3/12）,差异有统计学意义（x2值分别为76.0、65.4、56.4,P值均＜0.001）.IGRA-ELISA的特异度为71.2％ （527/740）,均低于TB-Ab法（94.7％,268/283）、Z-N法（99.3％,294/296）和荧光PCR法（100.0％,71/71）,差异有统计学意义（x2值分别为65.1、101.6、27.7,P值均＜0.001）.IGRA-ELISA的阳性预测值为28.5％ （85/298）,均低于Z-N法（80.0％,8/10）和荧光PCR法（100.0％,3/3）,差异有统计学意义（x2 =12.1,7.3,P值均＜0.01）;但高于TB-Ab法（21.1％,4/19）且差异无统计学意义（x2=0.5,P＞0.05）.IGRA-ELISA的阴性预测值为99.6％ （527/529）,均高于TB-Ab法（93.4％,268/287）、Z-N法（93.6％,294/314）和荧光PCR法（88.8％,71/80）,差异有统计学意义（x2值分别为28.9、27.8、46.3,P值均＜0.001）.IGRA-ELISA的ROC曲线的AUC为0.846,P＜0.001,95％CI为0.815～0.877;AUC均大于TB-Ab法（0.521）、Z-N法（0.622）和荧光PCR法（0.625）.结论 在本院的TB辅助诊断中,IGRA与其他3种检测技术比较有较高的敏感度,阳性预测值不高却有较高的阳性检出率和阳性诊断率,特异度较低因阴性预测值较高有较好的阴性排查能力,AUC最大有相对较好的辅助诊断价值.
Objective To compare the auxiliary diagnostic value of interferon gamma release assays （IGRA, IGRA-ELISA） for TB with TB antibody Colloidal gold rapid detection method （TB-Ah）, Ziehl-Neelsen Staining （Z-N） and fluorescent probe PCR （fluorescent PCR） in Beijing Tongren hospital so as to make an objective evaluation of IGRA-ELISA detection technology. Methods A study was performed on 1228 TB cases （IGRA-ELISA： 827; TB-Ab.. 306; Z-N.. 324; fluorescent PCR： 83） from Beijing Tongren hospital from August 2012 to July 2013. Chi square tests were performed on the enumeration data using SPSS 16.0. P〈0.05 was considered statistically significant. ROC （receiver operating characteristic） curves were drawn with SPSS 16.0 to calculate the area under the curve （AUC） by which auxiliary diagnostic values were compared. Results The positive rate of IGRA-ELISA （36.0%, 298/827）, was significantly higher than that for TB-Ab （6.2%, 19/306）, Z-N （3.1%, 10/324） and fluorescent PCR （3.6%, 3/83） （χ^2= 98. 6, 128.9 and 35.8, respectively, P〈0. 001）. The positive diagnostic rate of IGRA-ELISA （10.3%, 85/827） was higher than that for TB-Ab （1.3%, 4/306）, Z-N （2.5%, 8/324） and fluo rescent PCR （3.6%, 3/83）. IGRA-ELISA was significantly different in the positive diagnostic rate from TB-Ab and Z-N （χ^2 =24.8 and 19.1, respectively, P〈0. 001） but not from fluorescent PCR （χ^2 =3.8, P〉0.05）. The sensitivity of IGRA ELISA （97.7%, 85/87） was significantly higher than TB-Ab （17.4%, 4/23）, Z-N （28.6%, 8/28） and fluorescent PCR （25.0%, 3/12） （Z2 =76.0, 65.4 and 56.4, respectively, P〈0. 001）. The specificity of IGRA-ELISA （71.2%, 527/740） was significantly lower than TB-Ab （94. 7%, 268/283）, Z-N （99.3%, 294/296） and fluorescent PCR （100. 0%, 71/71） （χ^2 =65.1, 101.6 and 27.7, respectively, P〈0. 001）. The positive predictive value of IGRA-ELISA （28.5 %, 85/298） was significantly lower than that for Z-N （80. 0%, 8/10） and fluores cent PCR （100.0%, 3/3） （χ^2= 12.1, P〈0. 001; χ^2 =7.3, P〈0. 01）, but higher than that for TB-Ab （21.1 %, 4/19）, although this difference was not significant （χ^2 =0.5, P〉0. 05）. The negative predictive value of IGRA- ELISA （99.6%, 527/529） was significantly higher than that for TB-Ab （93.4%, 268/287）, Z-N （93.6%, 294/314） and fluorescent PCR （88.8%, 71/80） （χ^2 =28. 9, 27.8 and 46.3, respectively, P〈0. 001）. The AUC of IGRA-ELISA was 0. 846 （P〈0. 001） and its 95% confidence interval was 0. 815-0. 877, which was larger than that of TB-Ab （0. 521）, Z-N （0. 622） and fluorescent PCR （0. 625）. Conclusion IGRA-EI.ISA has greater auxiliary diagnostic value for TB in our hospital than the other three detection techniques evaluated because it has higher sensitivity and a higher positive rate and positive diagnostic rate in spite of its low positive predictive value, and it has a good negative exclusion for TB and a higher negative predictive value in spite of its low specificity and the largest AUC.
出处
《中国防痨杂志》
CAS
2015年第12期1211-1216,共6页
Chinese Journal of Antituberculosis
关键词
干扰素γ释放试验
酶联免疫吸附测定
聚合酶链反应
结核/诊断
对比研究
Interferon-gamma release assay
Enzyme-linked immunosorbent assay
Polymerase chain reaction
Tuberculosis/diagnosis
Comparative study
