6个检测系统4种分析方法测定人血清葡萄糖的精密度和正确度调查  

Evaluation of the precision and accuracy of the results from the 6 systems using 4 serum glucose analysis methods

作　　者：刘欧[1] 徐国宾[1] 王兰珍 夏昌宇[1] 刘岩[1] 

机构地区：[1]北京大学第一医院检验科,北京100034 [2]北京利德曼生化技术有限公司,北京100176

出　　处：《临床检验杂志》2010年第3期227-229,共3页Chinese Journal of Clinical Laboratory Science

基　　金：国家十一五科技支撑计划(2007BA105B09);国家高技术研究发展计划(863计划)(2006AA020909)

摘　　要：目的调查6个检测系统4种分析方法测定人血清葡萄糖(Glu)的精密度和正确度,为选择准确的Glu测定比对目标系统提供依据。方法2个候选参考实验室使用美国CDC推荐的血清Glu己糖激酶参考方法,为5个浓度水平的冰冻混合人血清定值;以NIST有证参考物质SRM965a评价参考方法的正确度,以是否通过IFCC组织的参考实验室间Glu比对(RELA)进一步验证参考实验室的测定能力;每个常规分析系统(n=3)采用配套的校准品按厂家的要求进行校准;以测定定值冰冻人血清的均值与靶值的偏倚评价其正确度,以变异系数(CV)评价其不精密度。结果2个候选Glu参考实验室测定4个水平的有证参考物,获8组数据,其中4组平均偏差介于2.4%~2.8%,另外4组小于2%;RELA比对结果显示参加本实验的2个参考实验室与其他参考实验室测定结果等效一致。6个(A^F)检测系统间5水平标本测定的CV范围1.81%~2.89%,各测定系统检测结果与靶值的相对偏倚分别为:-2.47%^-0.32%、1.69%~4.5%、-2.27%~0.27%、3.55%~4.97%、1.69%~3.5%、-0.97%~0.32%。各检测系统5个水平的相对总误差均小于CLIA′88规定的Glu检测的可接受范围(±10%)的1/2,同时也小于Glu生物学变异允许总误差(±6.9%)。结论2个Glu测定候选参考实验室定值准确,参加本次调查的6个检测系统4种Glu分析方法具有理想的精密度和正确度。Objective To evaluate the precision and accuracy of the results from six systems using four serum glucose analysis methods.Methods The glucose value of five frozen human serum pools was determined by the glucose hexokinase reference method which was recommended by the CDC in the USA.The accuracy of the reference laboratories was evaluated by the SRM965a and further validated through attending inter-laboratory comparison organized by IFCC（RELA）.Each routine analysis system（n=3） used in the matched calibrators was calibrated by the procedure according to the manufacturer recommended.Bias between the values from routine systems and the target value was used to evaluate the accuracy of routine system.The imprecision of six routine systems was expressed by coefficient variation.Results Two candidate reference laboratories measured four levels NIST′s SRM965a.Eight groups′ results were obtained.Average deviations of four groups were from 2.4%-2.8%,while four other groups′ average deviations were less than 2%.RELA comparison showed that results of the two laboratories were similar to the results from other reference laboratories.Coefficient variation of routine systems was from 1.81% to 2.89%.Relative bias was as follows：-2.47%--0.32%,1.69%-4.5%,-2.27%-0.27%,3.55%-4.97%and-0.97%-0.32%,respectively.The relative total error of five levels in each routine analysis system was less than 1/2 of CLIA′88 glucose which defined maximum acceptable range（±10%）,and also less than of the glucose biological variation total error（±6.9%）.Conclusions Two candidate reference laboratories′ results are accuracy.In this study,six systems using four different serum glucose analysis methods have good precision and accuracy.

关 键 词：血清葡萄糖 常规方法 参考测量程序 正确度 

分 类 号：R446.1[医药卫生—诊断学]