三种方法检测糖化血红蛋白结果的一致性评价  

Comparison of different methods for HbA 1c measurement

作　　者：宋旖川 徐安平[2] 王默 石洁 赵瑞[1] 纪玲[2] 张瑞[1] Song Yichuan;Xu Anping;Wang Mo;Shi Jie;Zhao Rui;Ji Ling;Zhang Rui(Department of Clinical Laboratory,Beijing Chao-yang Hospital,Capital Medical University,Beijing 100020,China;Department of Laboratory Medicine,Peking University Shenzhen Hospital,Shenzhen 518036,China)

机构地区：[1]首都医科大学附属北京朝阳医院检验科,北京100020 [2]北京大学深圳医院检验科,广东深圳518036

出　　处：《中华检验医学杂志》2024年第10期1197-1205,共9页Chinese Journal of Laboratory Medicine

基　　金：北京市医院管理局临床医学发展专项经费资助“扬帆”计划重点医学专业(ZYLX20213);北京朝阳医院科技创新基金重点项目(22kcjjzd-7)。

摘　　要：目的比较毛细管电泳法(CE)与离子交换层析-高效液相色谱法(CE-HPLC)和基质辅助激光解吸电离飞行时间质谱法(MALDI-TOF MS)间糖化血红蛋白(HbA1c)测定结果。方法方法学评价。收集2020年临床检测后的902份未含变异血红蛋白的正常全血样本及83份含变异血红蛋白的全血样本,其中变异血红蛋白样本主要分为α链变异、β链变异、γ链变异和δ链变异。902份正常样本分别用CE、CE-HPLC和MALDI-TOF MS 3种方法进行检测。以液相色谱串联质谱法(LC-MS/MS)作为参比方法,3种方法分别检测α链,β链,γ链3种变异血红蛋白与其进行比较。根据EP9-A3指南,分别用Passing-Bakolk回归分析和一致性相关系数(CCC)评估方法间的相关性及一致性。同时,评价3种方法间的不精密度,并以生物学变异最低偏倚2.3%为标准,评价方法性能。结果3种方法的不精密度均<2%。3种方法检测902份正常HbA1c结果的R2值和CCC均大于0.95。其中,MALDI-TOF MS与CE-HPLC的平均偏移绝对值高于生物学变异最低偏倚2.3%。以LC-MS/MS为参比方法,CE与LC-MS/MS检测α链,β链,γ链3种变异血红蛋白结果间的R2均大于0.95,γ链变异平均偏移绝对值高于生物学变异最低偏倚2.3%;CE-HPLC与LC-MS/MS检测β链变异结果间的R2为0.66,3种变异血红蛋白的平均偏倚绝对值均高于2.3%;MALDI-TOF MS与LC-MS/MS 3种变异血红蛋白检测结果间的R2均大于0.95,γ链变异平均偏倚绝对值高于2.3%。结论正常HbA1c样本中MALDI-TOF MS和CE-HPLC与CE检测结果间一致性较好。变异血红蛋白不同程度干扰3种方法的检测结果,CE法受到变异血红蛋白干扰较小。ObjectivesTo compare the HbA 1c results on capillary electrophoresis(CE),cation exchange-high performance liquid chromatography(CE-HPLC),and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry(MALDI-TOF MS).Methods902 normal samples without hemoglobin variants and 83 samples with hemoglobin variants were collected in 2020.The variants were divided intoα-chain,β-chain,γ-chain,andδ-chain variants.The 902 samples without hemoglobin variants were measured by CE,CE-HPLC,and MALDI-TOF MS.Three methods were used for measuringα-chain,β-chain,andγ-chain compared with the liquid chromatography-tandem mass spectrometry(LC-MS/MS)reference method.According to the EP9-A3 guideline,Passing-Bablok regression and concordance correlation coefficient(CCC)were performed for correlation and consistency,respectively.The imprecision was assessed by coefficient of variance(CV).The mean relative bias was calculated and compared with the lowest biological variation bias of 2.3%.ResultsAll three methods met the acceptance of imprecision requirements(<2%).The R 2 and CCC for the normal samples were all above 0.95 between the pairwise comparison of CE,CE-HPLC,and MALDI-TOF MS.The mean relative bias between MALDI-TOF MS and CE-HPLC was higher than the lowest biological variation bias of 2.3%.The R 2 results between CE and LC-MS/MS results ofα-chain,β-chain,γ-chain variants were>0.95,and the mean relative biases forγ-chain variants exceeded the lowest biological variation bias of 2.3%.A bad correlation(R 2=0.66)forβ-chain variants was shown between CE-HPLC and LC-MS/MS and the mean relative biases of all variant samples exceeded 2.3%.MALDI-TOF MS showed good correlations with LC-MS/MS for three groups,and the mean relative biases forγ-chain variants were higher than 2.3%.ConclusionMALDI-TOF MS and CE-HPLC with CE had good comparability in the measurement of HbA 1c in normal samples,butthe three methods showed interferences from different types of Hb variants and the CE method was affected with less interfe

关 键 词：血红蛋白A 糖基化 毛细管电泳法 飞行时间质谱法 离子交换层析-高效液相色谱法 方法学比对 

分 类 号：R446.11[医药卫生—诊断学]