摘要
目的探讨血清载脂蛋白AⅠ(Apo AⅠ)吸光度-浓度校准曲线的最适拟合模型,以建立并优化定量检测系统,并通过对不同检测系统进行方法比对分析,探讨各系统检测载脂蛋白Apo AⅠ和B100结果是否具有可比性。方法Apo AⅠ校准物吸光度检测采用免疫透射比浊法,对Apo AⅠ校准物吸光度与浓度进行线性模型、二次多项式模型、三次多项式模型、s模型、Logistic模型等模型的曲线拟合,并根据各回归模型决定系数(r2)的大小来确定最适拟合模型。用经过自身校准后的3个检测系统分别检测不同浓度的血清和质控物指标:Apo AⅠ和Apo B100,然后用最可靠的A系统校准另外2个系统后再重复测定血清和质控物,比较经过A系统校准前后另外2个系统测定的相对偏差大小,并按照1988年美国临床实验室修正案(CLIA'88)标准进行临床可接受性能判断。结果 Apo AⅠ校准物吸光度与浓度的散点图呈非线性趋势;线性模型、二次多项式模型、三次多项式模型、s模型Logistic模型的回归方程均有意义(P<0.05),其中二次多项式模型、三次多项式模型、s模型的拟合效果较好,以三次多项式模型为最佳。所测的Apo AⅠ和B100与标准系统的相关性好,结果可接受。结论三次多项式模型是载脂蛋白吸光度与浓度校准曲线的最适拟合模型。当实验室内存在2套和2套以上检测系统时,应对其进行方法比对和偏倚评估,才能保证检测结果的准确性和一致性。
Objective To study the serum apolipoprotein AⅠ（ Apo AⅠ） best fit model of 5 kinds of the absorbance- concentration calibration curve models and to establish and optimize the quantitative detection system. To discuss the comparison of apolipoprotein Apo AⅠand Apo B100 results among different detection systems. Methods Apo AⅠ calibrator absorbance were detected by immune nephelometry. The Apo AⅠabsorbance- concentration calibration curve models were analyzed including linear model,quadratic model,cubic polynomial model,s model and Logistic model to observe the curve fitting the regression model based on the coefficient of determination（ r2）,to determine the optimal size of the fitting model. Three detection systems after corresponding calibration were used to detect different concentrations of serum and quality control material including detecting items such as Apo AⅠand Apo B100. The most reliable A system was used to calibrate the two other systems. The serum and quality control material were measured again with the two other systems. The differences of bias between two other detection systems were compared before and after calibration. That was judged by the clinical acception according to clinical laboratory improvement amendment（ CLIA＇88）. Results Apo AⅠabsorbance- concentration calibration curve was non- linear trend. Linear model,quadratic model,cubic polynomial model,s model and Logistic regression models were significant（ P〈0. 05）. Which quadratic model,the cubic polynomial model and s model fitting accuracy was better,the best was cubic polynomial model. The measured apolipoprotein AⅠand B100 associated with the standard system was good and the results were acceptable. Conclusion Cubic polynomial model is the best fitting model of the apolipoprotein absorbance- concentration calibration curve. If there are two or more detective systems in lab,the method comparison and bias evaluation should be performed to ensure the accuracy and the consistency of the report.
出处
《湖北中医药大学学报》
2015年第5期113-115,共3页
Journal of Hubei University of Chinese Medicine
关键词
载脂蛋白
非线性校准
线性比对
相对偏差
apolipoprotein
nonlinear calibration
linear comparison
relative deviation
