自建生化检测系统的性能评价

Estimation on the capability of a self-developed biochemical detection system

目的探讨不同检测系统间实验结果的可比性,以证实自建检测系统的试验结果的可靠性。方法按美国CLIA'88能力比对检验的分析要求,比对实验应建立在检测系统的分析性能的基础上。本文分析性能评价包括不精密度,不准确度、线性,分析灵敏度(检测线),方法学比较。结果自建系统项目丙氨酸氨基转移酶(ALT)、尿素(Urea)、肌酐(Crea)、总胆固醇(TC)、血清蛋白A1(ApoA1)、人血清脂蛋白B(ApoB)、高密度脂蛋白胆固醇(HDL-C)与Roche可溯源参考检测系统试验结果相关系数>0.975。低密度脂蛋白胆固醇(LDL-C)存在方法学的差异试验结果相关系数<0.975。检测不精密度CV值满足美国CLIA'88能力比对检验的质量分析要求。线性检测范围满足临床检测线性要求,特别线性高值要求。结论自建检测系统性能评价试验结果与Roche可溯源参考检测系统试验结果比对具有相关性,不确定度也相应增加。各种性能相关性满足统计学原理并不证明其具有溯源性。

Objective To explore the comparability of results among different detection systems so as to prove the result reliability of self-developed biochemical detection system. Methods Proficiency testing should be based on performance analysis of detection system meeting to the demand of American CLIA88 proficiency testing, The performance estimation consisted of imprecision, inaccuracy, linearity, sensitivity of measurement （detection limit） and methodical comparison. Results There were no ststistical differences of results of alanine aminotransferase （ALT）, creatine （Cr）, total cholesterol （TC）, apolipoprotein A1 （ApoA1）, apolipoprotein B （ApoB）, as well as high-density lipoprotein cholesterol （HDL-C） detected by self-developed biochemical detection system and by traceable Roche Modular reference assay system. Their respective relative coeffieients were all above 0. 975, while there was significant difference of methodical analysis for low-density lipoprotein cholesterol （LDL-C） between two systems, and the relative coeffieient was below 0. 975. The CV of imprecision of measurement met the demand of American CLIA88 proficiency testing quality control analysis. The range of linear detection met the demand of clinical detection, eapecially for linearity of high value. Conclusion There is relativity of detected results by self-developed biochemical detection system and by traceable Roche Modular reference assay system, and the uncertainty increases accordingly. Traceability can＇t be proved only by significant performance relativity.