两台化学发光仪检测人生长激素的比对分析及偏倚评估

目的对Immulite 1000化学发光仪和DXI 800化学发光仪两种不同检测系统检测人生长激素(hGH)进行方法比对及偏倚评估,探讨不同检测系统间测定结果是否具有可比性。方法依据美国临床和实验室标准化协会(CLSI)EP9-A2文件要求,每天至少收集8份不同浓度的新鲜血清,分别用两种检测系统测定hGH,连续测定至少5d,每个标本均按正反序重复测定,记录测定结果,检查离群点,目测线性及数据分布,计算相关系数和直线回归方程,根据回归方程计算hGH在临界值处的预期偏倚和预期偏倚的95%可信区间并判断偏倚是否可以接受。结果两台仪器检测hGH的相关系数r=0.999,均值差异呈负偏倚,但这种偏差均在允许的误差范围内,临床均可接受。结论同一项目在不同检测系统检测时应进行方法比对和偏倚评估,以确保检测结果的可比性和连续性。

Comparative analysis and bias assessment of detecting results of human growth hormone using two chemiluminescence ap-paratus

Objective To explore the comparability of human growth hormone（hGH） detected by chemilumi- nescence apparatus of SIEMENS Immulite 1000 and Beckman Coulter DXI 800 detection systems,through methodology comparing and bias assessing. Methods According to Clinic and Laboratory Standrd Instritate（CLSI） EP9-A2 document,at least 8 copies of different concentrations of fresh serum were collected everyday, then we tested hGH with two detection systems for at least 5 days continuously. Every sample was detected repeatedly with obverse and reverse sequence,then we recorded the results, checked outlier,observed linearity and data distribute,and calculated correlation coefficient and linear regression equation. According to the regression equation,we calculated hGH in the critical value of the expected bias and the expected bias of the 95 % confidence interval,and determined the bias ac- cepted or not. Results The correlation coefficient（r） of two detection systems in hGH was 0. 999 ,and the difference of mean results was negative bias. The bias was in the permissible error range, and it was acceptable. Conclusion Methodology comparison and bias estimation of the same test items in different detection systems should be carried out regularly in order to ensure comparability and continuity of test results.