两种方法测定血清人附睾蛋白4的方法学比对及偏倚评估

摘要

Objective To compare the performance of 2 kinds of quantitative detection method the chemiluminescence immu-moassay and the enzyme-linked immunosorbent assay(ELISA)for detecting serum human epididymis geneproduct 4(HE4),and to explore the accuracy and practical value of the two methods.Methods In accordance to the document EP9-A2 of NCCLS,the Roche automatic electrochemical immunoassay system as the comparative method(X)and the ELISA as the testing method(Y )were per-formed according to the requirements of the detection kits.Results The serum HE4 level could be accurately reflected by the two methods.The correlation coefficient(r)in comparing these two kinds of method was ≥0.975,which showed that there was a high correlation between them.The detection results of the two methods were compared by paired sample t-test,and no statistically sig-nificant difference was found between them(P >0.05).Conclusion Because of CLIA′88 giving no specific medical decision level of HE4 and clinical acceptable level,therefore the expected bias of credibility interval can not be calculated or whether being clinically acceptable is unable to be judged.Based on the comparison of paired sample t test,these two kinds of detection methods can be used interchangeably to a certain extent.%目的：对电化学发光法与 ELISA 法两种定量检测血清人附睾蛋白4（HE4）的方法进行性能比较，探讨其检测结果的准确性及实用价值。方法根据美国临床实验室标准化委员会（NCCLS）EP9-A2文件，用罗氏全自动电化学免疫分析系统为比较方法（X），以 ELISA 为实验方法（Y ），两种方法均严格按照试剂盒要求进行操作检测。结果ELISA 法与电化学发光法均能较准确地反映血清中 HE4水平，两种方法比较相关系数 r≥0．975，显示高度相关。二者经配对样本 t 检验比较，差异无统计学意义（P ＞0．05）。结论由于 HE4因 CLIA′88未给出具体医学决定水平以及临床可接受水平，因此不能计算出预期偏倚的可信区间，未能判断临床是否可接受。根据配对样本 t 检验的比较，只能说明这两种检测方法在一定程度上可以交替使用。