CLSI EP06-Ed2文件在定量测量程序线性验证中的应用

Application of CLSI EP06-Ed2 in the linear verification of quantitative measurement procedures

目的以三碘甲状腺原氨酸(T3)、游离三碘甲状腺原氨酸(FT3)项目为例,探讨美国临床实验室标准化协会(CLSI)EP06-Ed2文件在定量测量程序线性验证中的应用。方法收集接近线性区间上、下限的样本,按比例配制7份线性样本,各重复测定2次,以稀释比例和重复测定均值(x)进行加权线性回归分析,计算预期值,将线性偏差的95%可信区间(CI)与允许线性偏差(ADL)进行比较。当所有样本线性偏差的CI与ADL相交时,整个区间的线性是可接受的。结果T3项目第1~6号样本线性偏差的95%CI在ADL范围内,第7号样本线性偏差在ADL范围内,FT3项目第2~4号样本线性偏差的95%CI在ADL范围内,第1、5号样本线性偏差在ADL范围内,第6号样本线性偏差未在ADL范围内,但线性偏差的95%CI与ADL范围有重叠,第7号样本线性偏差的95%CI未在ADL范围内。结论CLSI EP06-Ed2文件相对于CLSI EP06-A等文件,对线性样本浓度选择的要求更明确和合理,数据分析处理更有效、更简便,是非常实用的定量测量程序线性验证指南。

Objective To investigate the application of the Clinical and Laboratory Standards Institute(CLSI)document EP06-Ed2 in the linear verification of quantitative measurement procedures,taking triiodothyronine(T3)and free triiodothyronine(FT3)as examples.Methods The samples close to the upper and lower limits of the linear interval were selected,7 samples were prepared proportionally,and each sample was repeated for determination twice.A weighted linear regression analysis was performed using dilution ratios and means(x).The predicted value was calculated,and the 95%confidence interval(CI)of linear deviation was compared with the allowable deviation from linearity(ADL).When the CI of all samples overlapped with ADL,the linearity of the whole interval was acceptable.Results The linear deviation 95%CI of samples No.1-6 of T3 was within the ADL,and the linear deviation of sample No.7 was within the ADL.The linear deviation 95%CI of samples No.2-4 of FT3 was within the ADL,the linear deviation of samples No.1 and No.5 was within the ADL,and the linear deviation of sample No.6 was not within the ADL,but the linear deviation 95%CI overlapped with the ADL,sample No.7 linear deviation 95%CI was not within the ADL.Conclusions Compared with EP06-A,CLSI EP06-Ed2 has clearer and more reasonable requirements for the selection of linear sample concentration,with more effective and simple datum analysis and processing.It is a practical guide for the linear verification of quantitative measurement procedures.