四种非肽激素常规免疫分析方法正确度研究

Study on the Accuracy of Routine Immunoassay Methods for Four Non-peptide Hormones

目的比较四种非肽激素参考方法与免疫分析方法一致性,研究免疫方法正确度。方法基于本实验室复现运行的国际检验医学溯源联合委员会(JCTLM)公布的皮质醇、孕酮、睾酮和甲状腺素的参考方法,对每种非肽激素分别选择A、B两种主流免疫分析方法(A为国产试剂,B为进口试剂)与对应参考方法同时测定对应的20份新鲜单人份血清样品,以Passing-Bablok回归分析2种方法之间的相关性,绘制Bland-Altman图并比较四个非肽激素的参考方法与两个免疫分析方法间绝对与相对偏差,以一致性相关系数(CCC)评估方法间的一致性。结果皮质醇两种免疫分析方法与参考方法相关性良好(r>0.95),但它们与参考方法相比斜率的95%CI不包含1及截距的95%CI不包含0;甲状腺素两种免疫分析方法与参考方法相关性良好(r>0.95),它们与参考方法相比斜率的95%CI均包含1,但截距的95%CI只有方法A包含0;孕酮两种免疫分析方法与参考方法相关性良好(r>0.95),但它们与参考方法相比斜率的95%CI只有方法A包含1,截距的95%CI均为包含0;睾酮两种免疫分析方法与参考方法相关性良好(r>0.95),但它们与参考方法相比斜率的95%CI只有方法A包含1,截距的95%CI均包含0,四个项目的平均偏差除孕酮无标准外,其余均未超过最低偏移要求且它们与参考方法的一致性良好。结论四个项目所选的两个免疫分析方法与参考方法相比一致性良好,有偏移标准的皮质醇、甲状腺素和睾酮均能满足偏移标准,但孕酮项目的两种免疫分析方法的平均相对偏差都很大且该项目暂未查到偏移要求,参考睾酮项目标准,孕酮项目不满足偏移标准。

Objective To study the accuracy of immune methods by comparing the consistency between four non peptide hormone reference methods and immunoassay methods.Methods Based on the reference methods of cortisol,progesterone,testosterone and thyroxin published by the International Joint Committee on Traceability of Laboratory Medicine(JCTLM),which is reproduced in our laboratory,two mainstream im-munoassay methods,A and B,are selected for each non-peptide hormone(A is a domestic reagent,B is an im-ported reagent).Simultaneously measure the corresponding 20 fresh single human serum samples with the corresponding reference method,analyze the correlation between the two methods by Passing-Bablok regres-sion,draw the Bland-Altman diagram and compare the absolute and relative bias between the reference meth-od of the four non-peptide hormones and the two immunoassay methods were evaluated by the Concordance correlation coefficient(CCC).Results The two immunoassay methods of cortisol have a good correlation with the reference method(r>0.95),but compared with the reference method,the 95%CI of the slope does not contain 1 and the 95%CI of the intercept does not contain 0;The two immunoassay methods of thyroxine have a good correlation with the reference method(r>0.95).Compared with the reference method,the 95%CI of the slope both contains 1,but the 95%CI of the intercept only contains 0 in the method A;The two im-munoassay methods for progesterone have a good correlation with the reference method(r>0.95),but com-pared with the reference method,the 95%CI of the slope only method A contains 1,and the 95%CI of the in-tercept contains 0;The two immunoassay methods for testosterone correlate well with the reference method(r>0.95),but compared with the reference method,the 95%CI of the slope only method A contains 1,and the 95%CI of the intercept both methods contains 0.The mean deviations of the four items did not exceed the minimum deviation requirements except for progesterone without standard and they were in good agreement with the reference method.Conclusion The two immunoassay methods selected for the four projects are in good agreement with the reference method.Cortisol,thyroxine,and testosterone with offset standards can all meet the offset criteria,but the average relative biases of the two immunoassay methods of the progesterone project are very large,and the bias requirements have not been found for this project.Refer to the testosterone project standard,the progesterone project does not meet the bias standard.