3种化学发光免疫分析法检测系统与同位素稀释液相色谱串联质谱法检测总甲状腺素的一致性评价  

Harmonization evaluation for results of total thyroxin determined by three chemiluminescence immunoassay methods compared with isotope diluted liquid chromatography-tandem mass spectrometry

作　　者：禹松林[1] 周伟燕[2] 程歆琦[1] 孟庆慧[2] 李洪雷[1] 侯立安[1] 程倩[1] 谢少伟[1] 张传宝[2] 邱玲[1] YU Songlin;ZHOU Weiyan;CHEN Xinqi;MENG Qinghui;LI Honglei;HOU Li'an;CHENG Qian;XIE Shaowei;ZHANG Chuanbao;QIU Ling(Department of Clinical Laboratory,Peking Union Medical College Hospital,Chinese Academy of Medical Sciences,Beijing 100730;National Center for Clinical Laboratories,Beijing Hospital,National center of Gerontology,Beijing100730,China)

机构地区：[1]中国医学科学院北京协和医院检验科,北京100730 [2]北京医院国家老年医学中心卫生部临床检验中心,北京100730

出　　处：《临床检验杂志》2018年第12期891-894,共4页Chinese Journal of Clinical Laboratory Science

基　　金：国家临床重点专科资助项目;国家自然科学基金(81702060;81201337);国家科技支撑计划(2015BAI32H00)

摘　　要：目的比较3种化学发光免疫分析法检测系统与同位素稀释液相色谱串联质谱法(ID-LC/MS/MS)检测总甲状腺素(tT4)结果的一致性。方法参考美国临床和实验室标准化协会(CLSI)EP-15A方案评价3种检测系统(Antolumo A2000、CL-1000i为国产,ADVIA Centaur~为进口)检测不同浓度水平tT4的不精密度。选择北京协和医院2015年10月至2016年1月144例剩余血清样品,浓度覆盖tT4的测量范围。以ID-LC/MS/MS候选参考方法为参比方法,分别评价Antolumo A2000、CL-1000i及ADVIA Centaur~检测系统tT4检测结果与参比方法的一致性。以Passing&Bablok回归分析2种方法之间的相关性,绘制Bland-Altaman图并比较医学决定水平处的偏倚,以一致性相关系数(CCC)评估方法间的一致性。结果 3种化学发光免疫分析法检测系统的实验室内不精密度均小于说明书标示值,但在低浓度水平的不精密度(以变异系数CV表示)均大于来源于生物学变异的允许不精密度(3.7%),各化学发光法检测系统tT4结果与ID-LC/MS/MS结果相关性良好(r均>0.975),但仅有Antolumo A2000与ID-LC-MS/MS相比的斜率的95%CI包含1,且3种方法的截距的95%CI均不包含0;与ID-LC-MS/MS比较,平均偏差超过最低偏倚要求(4.5%),且在医学决定水平处各法均未能全部满足最低偏倚要求。一致性分析显示,Antolumo A2000和CL-1000i与ID-LC/MS/MS均具有较好的一致性(CCC:0.95～0.99),ADVIA Centaur~与ID-LC/MS/MS具有中度一致性(CCC:0.90～0.94)。结论 2种国产和1种进口tT4检测系统与ID-LC-MS/MS方法一致性良好,但在精密度、偏倚等方面均仍需进一步提高。Objective To compare the harmonization for the results of total thyroxin(tT4)determined by three chemiluminescence immunoassays with those of isotope dilution liquid chromatography-tandem mass spectrometry(ID-LC-MS/MS)method.Methods The imprecision evaluation was carried out according to CLSI EP-15A program.The aliquots of samples from 144 patients with low and high level of tT4 were collected during October 2015 to January 2016 in Peking Union Medical College Hospital.The levels of tT4 were measured by automated chemiluminescence immunoassays with Siemens(ADVIA Centaur?),Autobio(Autolumo A2000)and Mindray(CL-1000i),and ID-LC-MS/MS method was used as the candidate reference measurement procedure.Passing&Bablok regression and Concordance Correlation coefficient(CCC)were performed to analyze the correlations between the methods,and Bland&Altaman plot diagram were plotted to compare the bias at the level of medical decision.Results The imprecisions within laboratory of all three chemiluminescence systems were less than those of the indicated value in the the instruction of manufactures,however,at the lower level of t T4,the imprecisions of all the three methods were higher than the minimum performance goal(≤3.7%)derived from biological variation.The correlation coefficients of the results between the chemiluminescence assays and ID-LC/MS/MS were higher than 0.975.However,only the 95%CI of the slope from Antolumo A2000 contained 1 compared with that of ID-LC/MS/MS,and the 95%CI of the intercept from all the three assays did not contain zero which means that bias existed between chemiluminescence assays and ID-LC/MS/MS.Comparing with the ID-LC/MS/MS method,the mean deviation of all the three chemiluminescence assays methods exceeded the minimum bias requirement(≤4.5%).Not all the bias from the anticipants met the minimum requirement of 4.5%at the medicine decision level.Theconcordance correlation coefficient(CCC)of the performance of both Antolumo A2000 and CL-1000i showed substantial agreement(CCC:0.95-0.99)and A

关 键 词：甲状腺素 化学发光法 同位素稀释质谱法 

分 类 号：R446[医药卫生—诊断学]