4种ELISA方法检测牛口蹄疫非结构蛋白抗体的ROC评价

ROC evaluation of four ELISAs for detection of cattle serum antibodies to the non-structural proteins of foot and mouth disease virus

目的比较4种口蹄疫病毒非结构蛋白抗体(FMDV-NSP)ELISA的临床检测效果。方法以Ceditest?ELISA试剂盒检测结果为"金标准",应用4种ELISA方法检测164份牛血清口蹄疫病毒非结构蛋白抗体,通过平行试验比较各检测方法的特异性,敏感性及检出率,然后应用受试者工作特征(ROC)曲线分析,分析各检测方法的Youden指数,并优化最适临界值,再次比较几种检测方法的特异性,敏感性及检出率。结果口蹄疫病毒非结构蛋白3ABC抗体检测ELISA试剂盒,口蹄疫病毒非结构蛋白抗体单抗阻断ELISA试剂盒,3B合成肽ELISA试剂盒和本实验室建立的8BF-ELISA的敏感性分别为94.2%(80/85),91.9%(78/85),83.7%(71/85),91.9%(78/85);特异性分别为97.4%(77/79),91.0%(72/79),97.4%(77/79),94.9%(75/79);检出率为95.8%(157/164),91.5%(150/164),90.3%(148/164),93.3%(153/164)。。经X2检验,4种ELISA的检出率之间无显著性差异(P=0.461>0.05)。但3B合成肽ELISA试剂盒的敏感性明显低于其他3种检测方法,不适合大规模临床样本的初筛试验。结论在优化临界值后的ELISA A,ELISA B和ELISA D的敏感性和特异性都大于90%,同时有着很高的检出率,均可用于大规模临床的样本的初筛试验。

The clinical testing effectiveness of four ELISAs for detection of nonstructural protein antibody against the foot and-mouth disease virus was compared in China in this study and 164 serum samples of bovine were detected. Specificity, sensitivity and the detection rate of the four ELISAs were compared by parallel test and then were compared with Ceditest~ ELISA kit which was considered as the ＂gold standard＂. The results were calculated and the receiver-operator characteristic （ROC） curves and Youden index for each test method were subsequently analyzed. Four methods were evaluated again after op- timization of optimum threshold value. The results showed that the sensitivity with 3ABC-I- ELISA, Blocking ELISA, 3B- Elisa, and 8BF-ELISA which was established in the laboratory were 94.2% （80/85）, 91.90//00 （78/85）, 83.7% （71/85）, and 91.9% （78/85）, respectively. Specificity of these four were 97. 4% （77/79）, 91.0G （72/79）, 97.4~ （77/79）, and 94. 9% （75/79）, respectively. The detection rate were 95.8%/00 （157/164）, 91.5%o （150/164）, 90.8% （148/164）, and 93.3% （153/ 164）, respectively. For the X2 test, the four ELISAs detection rates had no significant difference （P=0. 461）0.05）. But 3B- Elisa was not suitable for large-scale clinical samples screening test as the sensitivity of it was obvious lower than that of three others. ELISAs optimization after optimal cut-off showed that sensitivity and specificity of the ELISA A, ELISA B, and EI.ISA D were more than 90%, and with a higher detection rate in experimental serum, which all could be applied to large- scale clinical sample screening test.