肌酐酶法与苦味酸法检测结果可比性分析

Comparability Analysis of Creatinine Enzymatic Method and Picric Acid Method

目的评价强生干化学检测系统与贝克曼湿化学检测系统两种不同检测方法对于肌酐项目检测结果的可比性。方法根据美国临床和实验室标准化协会(CLSI)发布的EP9-A3文件标准,以强生干化学肌酐酶法为参比方法,贝克曼肌酐苦味酸法为待评方法,选取40份在测量线性范围均匀分布临床患者新鲜样本,分别在两个系统上测定肌酐浓度,强生干化学酶法(Y)和贝克曼苦味酸法(X)之间的相对偏差,判断检验结果的可比性。结果绘制散点图,图中明显可见1个异常值点,经ESD法进一步检验发现第32号数据为离群值EP9-A3文件提供了OLR、WLS、Deming、Passing-Baklok 4种模型进行回归分析。拟合后回归方程分别Y=0.9715X+1.5367,Y=0.9712X+1.7536,Y=0.9725X+1.5271,Y=0.9716X+1.4775,斜率的95%CI分别为0.9346~0.9505、0.9451~0.9521、0.9367~0.9489、0.9502~0.9594,比例偏移分别为-5.16%^-4.75%、-5.21%^-4.60%、-5.96%^-4.94%、-5.57%^-4.87%,均小于可接受标准;肌酐的医学决定水平44μmol/L,133μmol/L,450μmol/L分别代入OLR、WLS、Deming、Passing-Baklok回归方程,其偏移也均小于可接受标准。结论方法比对实验能够及时发现方法学不同造成的系统误差,通过调整与校准可以确保相对偏差处于可接受范围,为不同检测系统检验结果的互认提供依据。

Objective To evaluate the comparability of two different detection methods of Johnson&Johnson dry chemical detection system and Beckman wet chemical detection system for creatinine project detection results.Methods According to the EP9-A3 document published by the American Society for Clinical and Laboratory Standards(CLSI),the Johnson&Johnson creatinine method was used as a reference method,and the Beckmann creatinine picric acid method was used as the method to be evaluated.A fresh sample of clinical patients was evenly distributed,and the relative deviation between creatinine concentration,Johnson&Johnson chemical enzymatic method(Y)and Beckman's picric acid method(X)was determined on two systems to judge the comparability of the test results.Results Plot a scatter plot.One abnormal value is clearly visible in the figure.Further examination by ESD method reveals that the 32nd data provides the OLR,WLS,Deming,Passing-Baklok models for the outlier EP9-A3 file.analysis.The fitted regression equations are Y=0.9715X+1.5367,Y=0.9712X+1.7536,Y=0.9725X+1.5271,Y=0.9716X+1.4775,and the 95% CI of the slope is 0.9346-0.9505,0.9451-0.9521,0.9367-0.9489,0.9502-0.9594,the proportional shifts were-5.16% to-4.75%,-5.21% to-4.60%,-5.96% to-4.94%,-5.57% to-4.87%,respectively,all less than the acceptable standard;creatinine medically determined levels of 44μmol/L,133μmol/L,and 450μmol/L were substituted into the OLR,WLS,Deming,and Passing-Baklok regression equations,respectively,and the offsets were also less than acceptable standards.Conclusion The method comparison method can timely discover the systematic errors caused by different methodologies.Through adjustment and calibration,the relative deviation can be ensured to be within the acceptable range,which provides a basis for mutual recognition of the test results of different detection systems.