国内4厂家15套血清葡萄糖检测系统精度的比对研究

Investigation of precision and trueness of the glucose results from 15 measure systems in 4 domestic factories

目的:了解国产葡萄糖检测试剂在不同自动化分析仪上测定结果的精度。方法:向参加调查的国内4个葡萄糖试剂生产厂家发放经美国CDC推荐的血清葡萄糖己糖激酶参考方法定值的5水平候选葡萄糖标准物质,各厂家自行选取自动生化分析仪,分别采用试剂配套校准品和第2水平候选标准物质校准后测定。结果:15个检测系统采用试剂配套校准品校准后5水平候选标准物质测定结果的变异均明显高于以C2为校准品校准后测定结果的变异;经原厂校准品校准后,8个GOD法检测系统的变异明显高于7个HK法检测系统的变异,但以C2为校准品校准后GOD法与HK法检测系统的变异无此明显差异。15个检测系统以原厂校准品校准后C1～C5候选标物的偏倚分别是-2.92%～11.68%、-3.37%～9.79%、-3.75%～8.75%、-3.11%～9.55%、-4.38%～8.28%,其中3个检测系统的偏倚超出了葡萄糖常规测定方法中欧洲生物学目标和计算出的生物学允许的总误差5.5%,2个检测系统的偏倚超出了美国标准的总分析误差6.0%,1个检测系统的偏倚超出了CLIA′88能力比对检验的可接受偏差10%的要求;以C2为校准品校准后C1～C5候选标物的偏倚分别是-1.95%～3.57%、-1.52%～1.85%、-2.72%～2.27%、-1.82%～2.9%、-2.83%～2.19%。15个检测系统的偏倚均小于欧洲生物学目标和计算出的生物学允许的总误差5.5%的要求,同时小于CLIA′88能力比对检验的可接受偏差10%的1/2。结论:部分常规血清葡萄糖检测系统的偏倚较大,采用具有互通性的候选标物为校准品可以提高常规检测系统测定结果的精密度和正确度。

Objective To investigate the precision and trueness of the glucose results from the domestic kits on different analyzers. Methods To provide 5 levels serum glucose standard reference materials which the target values are assigned by reference methods for 4 domestic glucose kits factories, factories choose analyzers by themselves, they calibrated with factory calibrator and C2 respectively, then measure the 5 levels serum. Results The results of all 15 measure systems variation by factory calibrator calibrated are higher than the results by C2 calibrated; when calibrated by factory calibrator, the variation of 8 measure systems adopt GOD method are higher than the variation of 7 measure systems adopt HK method, there are no difference between the two methods measure systems when calibrated by C2. When calibrated by factory calibrator, Bias of C1 - C5 include 15 measure systems are respectively： -2.92% - 11.68%,-3.37% -9.79%,-3.75% - 8.75%,-3.11% - 9.55%,-4.38% - 8.28%, there are 3 measure systems bias are beyond the limit of the Europe biological variation total error （5.5%）, there are 2measure systems bias are beyond the limit of the U.S.A. criterion analyze total error（6% ）, there is lmeasure system bias is beyond the limit of the CLIA＇ 88 defined acceptable error（10%） ;When calibrated by C2, Bias of C 1 - C5 include 15 measure systems are respectively： -1.95% - 3.57%,-1.52% - 1.85%,-2.72% - 2.27%,-1.82% - 2.9%,-2.83% - 2.19%, 15 measure systems bias are within the limit of the Europe biological variation total error （5.5%） and the half limit of CLIA＇88 defined acceptable error （10%）. Conclusions The bias of some routine measure systems are larger. Precision and trueness of routine measure systems can be improved by calibrated with serum standard reference materials.