10种市售总25-羟基维生素D免疫检测系统测定偏差评价

Detection bias evaluation of ten commercial total 25-hydroxyvitamin D immunoassays

目的评价10种25-羟基维生素D[25（OH）D]免疫检测系统的测定偏差。方法本研究属于方法学比对。先选用含有不同25（OH）D类型的4浓度水平的人血清标准物质SRM972a，应用本室建立的液相色谱-同位素稀释质谱法（ID-LC/MS/MS）和迈瑞、新产业、迈克、科美、雅培、西门子、罗氏、梅里埃8个厂家的10种标记免疫检测系统共同测定，利用Excel软件计算测量值和标准物质标示值之间的相对偏差。再选用2支含不同浓度水平25（OH）D3的临床血清样本进行共同测定，利用Excel软件计算免疫检测系统测量值和ID-LC/MS/MS测量值之间的相对偏差。对各免疫检测系统测定偏差结果进行评价，要求免疫检测系统相对偏差不大于10%，同时简要分析各检测系统分析特异性。结果通过对血清25（OH）D标准物质测试，自建ID-LC/MS/MS测量结果和标准物质标示值最接近，测量偏差在1.6%～2.8%，各免疫检测系统对不同类型25（OH）D成分的检测能力有差别，有8个检测系统对高含量25（OH）D3标准物质（水平1）测量偏差较小（1.5%～3.5%），但只有2个免疫检测系统对高含量25（OH）D2标准物质（水平3）测量偏差较小（3.6%～3.7%），5个免疫检测系统对高含量3-epi-25（OH）D3标准物质（水平4）测量偏差较小（－0.3%～9.0%）。总体而言检测系统7表现出较小的测定偏差，对25（OH）D2、25（OH）D3和无活性异构体3-epi-25（OH）D3有较好的分析特异性。对2支25（OH）D3血清临床样本的测试显示，与同位素稀释质谱法测量结果相比，免疫检测系统高值样本检测时测定偏差较小。结论不同免疫检测系统对25（OH）D检测结果与ID-LC/MS/MS相比，测量偏差的差异较大。各25（OH）D免疫检测的特异性检测能力，尤其是对25（OH）D2和25（OH）D3的等效识别上还有提升的空间。

ObejectiveTo evaluate the accuracies of ten commercial total 25-hydroxyvitamin D[25（OH）D]immnoassays.MethodsNIST 25（OH）D reference material SRM 972a, which consisted of four vials of frozen serum with different concentration levels of different 25（OH）D types, and two human serumsamples provided by our lab（BIMT）, which had different concentration levels of 25（OH）D3,were analyzed by ten total 25-hydroxyvitamin D immnoassays from Biomerieux, Mindray, Maccura, Chemclin, Abbott（2）, Siemens, SNIBE（2）, Roche, and by isotope-dilution liquid chromatography-tandem mass spectrometry（ID-LC/MS/MS） founded by BIMT. For the measurements of SRM 972a, the biases between tested values and certified values were calculated as a evaluating indicator, meanwhile the test biases between immnoassays and ID-LC/MS/MS were used as a evaluating indicator for the measurements of BIMT 25（OH）D3 serum samples. Test bias lower than 10% was deemed acceptible.ResultsThe ID-LC/MS/MS exhibited low biases at （1.6%-2.8%） in the measurements of all levels of SRM 972a. 8 immnoassays showed low biases at（1.5%-3.5%）in the measurements of level 1 of SRM 972a, which had a high 25（OH）D3 concentration level, but only 2 immnoassays gave low biases at（3.6%-3.7%）in the measurements of high 25（OH）D2 concentration sample（level 3）. While, 5 immnoassays gave low biases at（-0.3%-9.0%）in the measurements of high 3-epi-25（OH）D3 concentrationsample （level 4）. It seems that, when SRM 972a were analyzed, only one of the ten commercial total 25（OH）D immnoassays were in good accuracy and analytical specificity agrements with ID-LC/MS/MS. When two human serum25（OH）D3 samples from BIMT were tested, most immunoassays were overall in relative good agreement with ID-LC/MS/MS at high 25（OH）D3concentration level.ConclusionThe test biases in the total 25（OH）D measurements are differences between differentimmnoassays and ID-LC/MS/MS, and the specificities of current commercial total 25（OH）D immnoassays should be improved, especially the performance on the equivalent recognition of 25（OH）D2 and 25（OH）D3.