4种促甲状腺素免疫法检测系统结果的一致性评价
|
摘要
目的比较4种促甲状腺素(TSH)免疫法检测系统结果的一致性。方法收集2018年3至4月北京协和医院临床检测剩余新鲜血清标本共102份。参考CLSI EP15-A2及EP9-A2方案进行精密度评价和方法学比对。分别采用雅培i2000、贝克曼DXI 800、西门子ADVIA Centaur XP及罗氏e601全自动化学发光免疫分析仪及其配套试剂(标为A～D)检测102份血清的TSH水平,应用passing-bablok及Bland Altman进行比对分析。以0.27μIU/m L、5.33μIU/m L为医学决定水平,比较各检测系统的预期偏倚。结果 A～D 4种检测系统的精密度分别为1.7%～3.3%、2.3%～3.9%、0.7%～2.3%、0.6%～1.5%,中位数(25分位数,75分位数)分别为1.898(0.518,4.809)μIU/m L、2.819(0.719,7.020)μIU/m L、2.502(0.692,6.888)μIU/m L、3.105(0.886,7.905)μIU/m L。4种检测系统回归方程的决定系数(R2)均>0.975,相关系数(r)在0.993 5～0.997 1之间,截距和斜率分别为0～0.06、0.59～1.15,其中B与C两个检测系统直线回归图斜率为1.02,截距为0,相关性最好;偏差图显示不同检测系统的百分偏倚在-48.1%～17.3%,A与D检测系统偏倚最大,B与C检测系统偏倚最小。4种检测系统在医学决定水平处的预期偏倚为-40.7%～37.2%。结论贝克曼与西门子TSH检测试剂一致性较好,罗氏、雅培TSH检测试剂与另外两家一致性存在差异。
Objective To compare the consistency of thyroid stimulating hormone( TSH) results from four chemiluminescence assays.Methods A total of 102 fresh serum samples from Peking Union Medical College Hospital during March 2018 and April 2018 were collected for precision evaluation and methodological comparison referring to CLSI EP15-A2 and EP9-A2 protocols. The levels of serum TSH were detected by Abbott i2000( system A),Beckman DXI800( system B),Siemens ADVIA Centaur XP( system C) and Roche e601( system D) automatic chemiluminescence analyzers and their matching reagents,respectively. The obtained results were compared with the passing-bablok and Bland Altman methods. Taking 0.27 μIU/m L and 5.33 μIU/m L as the medical decision level,the expected bias of each detection system was compared. Results The precisions of systems A,B,C and D were 1.7%-3.3%,2.3%-3.9%,0.7%-2.3% and 0.6%-1.5%,respectively. The median( P25,P75) of TSH concentrations detected by systems A,B,C and D were1.898( 0.518,4.809) μIU/m L,2.819( 0.719,7.020) μIU/m L,2.502( 0.692,6.888) μIU/m L and 3.105( 0.886,7.905) μIU/m L,respectively. The coefficients of determination( R2) of regression equation were above 0.975 for 4 detection systems. The correlation coefficients( r),intercepts and slopes of 4 detection systems were 0.993 5-0.997 1,0-0.06 and 0.59-1.15,respectively,and systems B and C had the best correlations with 1.02 of slope and 0 of intercept. The deviation plot showed that the bias% of 4 detection systems was between-48.1% and 17.3%. Among them,systems A and D had the largest bias,while systems B and C had the lowest bias. The expected bias of 4 detection systems at the medical decision level was-40.7%-37.2%. Conclusion The consistency between Beckman and Siemens TSH detection systems is good,while those of Roche and Abbott TSH detection systems are different from the other two.
Objective To compare the consistency of thyroid stimulating hormone( TSH) results from four chemiluminescence assays.Methods A total of 102 fresh serum samples from Peking Union Medical College Hospital during March 2018 and April 2018 were collected for precision evaluation and methodological comparison referring to CLSI EP15-A2 and EP9-A2 protocols. The levels of serum TSH were detected by Abbott i2000( system A),Beckman DXI800( system B),Siemens ADVIA Centaur XP( system C) and Roche e601( system D) automatic chemiluminescence analyzers and their matching reagents,respectively. The obtained results were compared with the passing-bablok and Bland Altman methods. Taking 0.27 μIU/m L and 5.33 μIU/m L as the medical decision level,the expected bias of each detection system was compared. Results The precisions of systems A,B,C and D were 1.7%-3.3%,2.3%-3.9%,0.7%-2.3% and 0.6%-1.5%,respectively. The median( P25,P75) of TSH concentrations detected by systems A,B,C and D were1.898( 0.518,4.809) μIU/m L,2.819( 0.719,7.020) μIU/m L,2.502( 0.692,6.888) μIU/m L and 3.105( 0.886,7.905) μIU/m L,respectively. The coefficients of determination( R2) of regression equation were above 0.975 for 4 detection systems. The correlation coefficients( r),intercepts and slopes of 4 detection systems were 0.993 5-0.997 1,0-0.06 and 0.59-1.15,respectively,and systems B and C had the best correlations with 1.02 of slope and 0 of intercept. The deviation plot showed that the bias% of 4 detection systems was between-48.1% and 17.3%. Among them,systems A and D had the largest bias,while systems B and C had the lowest bias. The expected bias of 4 detection systems at the medical decision level was-40.7%-37.2%. Conclusion The consistency between Beckman and Siemens TSH detection systems is good,while those of Roche and Abbott TSH detection systems are different from the other two.
引文
[1]Thyroid Disease Manager. Guidelines for diagnosis and management of thyroid disease[OL]. https://www. thyroidmanager. org/guidelines/.(Accessed February 2017).
[2]Cooper DS,Biondi B. Subclinical thyroid disease[J]. The Lancet,2012,379(9821):1142-1154.
[3]Sheehan MT. Biochemical testing of the thyroid:TSH is the best and,oftentimes,only test needed-a review for primary care[J]. Clin Med Res,2016,14(2):83-92.
[4]Taylor R. Interpretation of the correlation coefficient:a basic review[J]. J Diagn Med Sonogr,1990,6(1):35-39.
[5]Thienpont LM,Van Uytfanghe K,De Grande LAC,et al. Harmonization of serum thyroid-stimulating hormone measurements paves the way for the adoption of a more uniform reference interval[J]. Clin Chem,2017,63(7):1248-1260.
[6]Clerico A,Ripoli A,Fortunato A,et al. Harmonization protocols for TSH immunoassays:a multicenter study in Italy[J]. Clin Chem Lab Med,2017,55(11):1722-1733.
[2]Cooper DS,Biondi B. Subclinical thyroid disease[J]. The Lancet,2012,379(9821):1142-1154.
[3]Sheehan MT. Biochemical testing of the thyroid:TSH is the best and,oftentimes,only test needed-a review for primary care[J]. Clin Med Res,2016,14(2):83-92.
[4]Taylor R. Interpretation of the correlation coefficient:a basic review[J]. J Diagn Med Sonogr,1990,6(1):35-39.
[5]Thienpont LM,Van Uytfanghe K,De Grande LAC,et al. Harmonization of serum thyroid-stimulating hormone measurements paves the way for the adoption of a more uniform reference interval[J]. Clin Chem,2017,63(7):1248-1260.
[6]Clerico A,Ripoli A,Fortunato A,et al. Harmonization protocols for TSH immunoassays:a multicenter study in Italy[J]. Clin Chem Lab Med,2017,55(11):1722-1733.