常用肌酐检测系统的准确度评估

Accuracy investigation of commonly used creatinine assay systems

目的评估常用cr检测系统的准确度，观察系统间Cr测量值的相对变异及所带来的eGFR差异。方法参照CLSIEP14-A2实验方案，在10套酶法和1套苦味酸法Cr检测系统上验证CAPLN24的互通性。LN24共6个浓度，分别为68．1、126．9、185．7、244．5、303．2和361．9μmol/L，Cr定值基于NIST—IDMS方法。以LN24定值为靶值，评估各检测系统的准确度，分析各系统问变异。eGFR采用基于苦味酸法cr的MDRD公式和基于酶法cr的CKD—EH公式进行计算。结果LN24在11套cr检测系统间具有互通性。4套系统在各浓度的偏倚均〈4．4μmoL／L，2套系统在各浓度的偏倚均〉4．4μmol/L，1套系统存在较固定偏倚[（-4．2±0．7）μmoL／L]，其余系统在各浓度均有不同程度的偏倚。Cr在68．1μmoL／L时，eGFR偏差最高可达14．9ml·min^-1·（1．73m^2）^-1。系统间sD随cr浓度升高而升高（2．6～6．1μmoL／L）；系统问CV随Cr浓度升高而下降（4．0％～1．7％）。去除负偏倚显著的2套系统，系统间变异及eGFR偏差均明显下降。测量新鲜血发现，各酶法系统间Cr差异大部分〈10μmoL／L；苦味酸法与酶法的差异离散度大，差异范围-15～20μmoL／L。Cr〈100μmoL／L时，MDRD公式和CKD—EH公式计算出的eGFR可相差～18～40ml·min^-1·（1．73m2）～。结论部分酶法Cr检测系统具有良好的准确度；各酶法系统间的差异较为固定，可通过数学校正达到统一。苦味酸法和酶法肌酐测量结果间存在较大差异，用数学公式不能校正。低浓度时，肌酐测量偏倚和采用不同公式计算eGFR均可导致较大的eGFR偏差。建议临床实验室重视低浓度下Cr测量的准确度和一致性，并采用统一的公式计算eGFR。

Objective To evaluate the accuracy of Cr measurement value from commonly used homogenous detection systems, to investigate the variation among different systems and the corresponding bias of eGFR. Methods According to the CLSI EP14-A2 protocol, commutability of LN24 was validated among 10 enzymatic assays and 1 picrate assay. LN24 included 6 vials of solution with Cr values assigned by IDMS at NIST, and concentrations of Cr for each vial were 68. 1, 126.9, 185.7,244. 5, 303.2 and 361.9 μmoL/L. LN24 was used to evaluate the accuracy of the included systems and the variation among them, and the assigned values were taken as the target values, eGFR were calculated by MDRD equation using IDMS- traced picrate Cr and CKD-EPI equation using enzymatic Cr. Results Commutability was exist among the 11 systems fnr LN24 detection. Four systems showed bias 〈 4. 4 μmoL/L at each level of LN24, two system showed bias 〉 4. 4 μmoL/L at each level of LN24, one system showed a fixed negative bias （ - 4. 2 ± 0. 7 ）μmoL/L, the other 4 systems showed diverse bias at difterent levels. Cr-bias-caused eGFR bias could reach 14. 9 ml ·min^-1·（1.73m^2 ）^-1 at Cr level of 68. 1 μmoL/L. SD among systems ascended with Cr level （2. 6 -6. μmoL/L） ;CV among systems descended with Cr level（4. 0% - 1.7% ） ;After the 2 systems with obvious negative bias were removed, SD, CV among systems and eGFR bias decreased obviously. By measuring fresh serum, it was found that Cr bias among enzymatic systems was mostly 〈 10 μmoL/L;that between enzymatic assays and picrate assay was much diffused（from - 15 to 20 μmoL/L）. When Cr 〈 100 μmoL/L, the eGFR difference between result of MDRD equation and that of CKD-EPI equation ranged from -18 to 40 ml · min-1 （1.73 m2） 1. Conclusions Some enzymatic systems show good accuracy. Difference of Cr value is relatively fixed among enzymatic systems, and comparability can be reached through mathematic, way. Un-aeeeptable difference between pierate assay and enzymatic assays still exists, thus comparability cannot be reached through mathematic way. At low Cr level, bias of Cr and using different equations may lead to significant bias of eGFR. We recommend that clinical laboratory should pay much attention to the accuracy and comparability at low level of Cr, and use uniform equation to calculate eGFR.