应用EP6-A2方法验证迈瑞血糖试剂盒的线性范围

Validation of the linear range of Mindray blood glucose kit by EP6-A2 method

目的通过验证本实验室迈瑞血糖试剂盒的线性范围,初步评价该血糖试剂盒的性能。方法依据美国临床实验室标准化协会(CLSI)2003年4月发布的EP06-A2文件,用高值和低值血糖浓度的样本按比例精确配成等间距的6个实验样本,每个浓度水平测量2次,所得实验数据做一次、二次和三次多项式回归分析,二次多项式的非线性系数b2、三次多项式的非线性系数b2、b3进行t检验,判断非线性系数是否有统计学意义,即与0之间有无差异。计算标准回归误差(Sy.x),确定最适多项式模型。结果未发现离群的实验结果。实验样本重复测定的恒定误差(SDr)为0.40%,SDr小于本室的批内不精密度要求1/4CLIA'88允许总误差(2.5%)。确定最适多项式模型为二次回归多项式。6个浓度水平的线性偏差(DLi)最大值为1.73%,小于美国临床病理学家协会(CAP)要求的非线性差异指标(2.5%)。结论该迈瑞葡萄糖试剂盒在0.5～25.0 mmol/L范围内呈线性,本验证实验结果与该试剂盒标示的线性范围一致。

Objective To preliminarily evaluate the performance of the blood glucose kit by verifying the linear range of the Mindray blood glucose kit in our laboratory. Methods According to the EP06-A2 document issued by the American Clinical Laboratory Standardization Institute(CLSI) in April 2003, six equal-interval experimental samples with high and low blood glucose concentrations were accurately proportionally prepared. Each concentration level was measured twice. The obtained experimental data were analyzed by simple, quadratic and cubic polynomial regression analyses. The t-tests were carried out for non-linear coefficients b2, b2 and b3 obtained from the above quadratic and cubic polynomial regression analyses respectively to judge whether the nonlinear coefficients had statistical significances or not, that was whether there were differences between them and 0. The optimal polynomial model was determined by calculating the standard regression error(Sy.x). Results No experimental results were outliers. The constant error(SDr) of repeated determinations of experimental samples was 0.40%. The SDr was less than the total allowable error(2.5%) of 1/4 CLIA’88 for the in-batch imprecision in our laboratory. The quadratic regression polynomial type was determined as the optimal polynomial model. The maximum linear deviation(DLi) of the six concentration levels was 1.73%, which was less than the non-linear difference index(2.5%) required by the American Association of Clinical Pathologists(CAP). Conclusion The Mindray glucose kit was linear in the range of 0.5-25.0 mmol/L, demonstrating that these experimental results were consistent with the linear range indicated by the kit.