不同检测系统钙、磷、镁测定结果的偏倚评估

Bias evaluation on results of serum calcium,phospholipids and magnesium with different biochemical detecting system

目的探讨不同检测系统钙、磷、镁测定结果的可比性。方法根据EP-9A文件,取朗道水平2和水平3质控物,以及41例不同浓度的患者新鲜血清,在3个不同的生化检测系统(检测系统:日立7170A生化分析仪、Roche校准物、试剂及质控品;检测系统:强生Vitros250干式生化分析仪及其配套的校准物、试剂,德灵质控品;检测系统:日立7170A生化分析仪、Randox校准物、试剂及质控品)上进行钙、磷、镁检测,并对数据进行相关的统计学分析。以美国临床实验室修正法规(CLIA'88)或澳大利亚室间质量评价标准规定的室间质量评价允许误差范围的1/2为临床接受范围,判断不同检测系统的可比性。结果朗道质控物和新鲜血清标本钙、磷、镁测定结果经随机区组设计资料的方差分析,除检测系统Ⅲ的磷、镁外,其余各检测系统测定结果差异均有显著性(P<0.001)。各检测系统新鲜血清标本钙、磷、镁测定可靠性系数α分别为0.9928、0.9908、0.9277,各系统间的相关系数均>0.975。各检测系统测定钙、磷的批间精密度变异系数均<5%,镁的批间精密度变异系数均<8.3%,以可溯源的检测系统Ⅰ为目标检测系统,钙、磷、镁均有超过CLIA88允许范围1/2的待评检测系统。结论检测系统2与目标检测系统间有明显偏倚,主要与方法学差异有关。实验室应经常进行同一项目不同检测系统间的偏倚评估,判断临床接受性能,采取整改措施,保证结果的可比性。

[Objective] To explore comparability of serum calcium （Ca^2＋）, phospholipids （P） and magnesium （Mg^2＋） with different biochemical detecting system. [Methods] Three different kinds of biochemistry detection systems were used to detect Ca^2＋, IP and Mg^2＋ concentrations at 2 levels of Randox quality controls and 41 clinical specimen according to EP-9A file. The collected data were deal with statistics analysis. [Results] Analysis of variance showed Ca^2＋, IP and Mg^2＋ results from different controls and patients sera have significant difference in various detection system （P 〈0.001） except P and Mg^2＋ results in detection system Ⅲ. The correlation coefficient of each detection system was above 0.975 and the reliability coefficients of them were 0.9928, 0.9908 and 0.9277 respectively. The CV value of Ca^2＋, P at Ⅲ detection systems were less than 5.0%. The CV value of Mg^2＋ at Ⅲ detection systems was less than 8.3%. System Ⅰ was regarded as Target detection system and evaluated the acceptability of other detection systems. As a result, the items of Ca^2＋, IP, Mg^2＋ at Ⅲ detection systems partly excelled permitted error of clinic assessment. [Conclusion] There is a significant bias between the detection system Ⅱ and the target detection system, which mainly caused of the differences in their test methodology.