三个不同血细胞分析检测系统间检验结果的偏差评估与比对性研究

Bias Estimation and Comparative Analysis and Acceptability Evaluation of the Complete Blood Count with Three Different Detection Systems

目的通过对同一检验科不同检测系统进行方法比对和偏差评估，探讨不同检测系统间WBC．RBC，HGB，HCT，MCV，MCH，MCHC和PLT8项血细胞分析测定结果是否具有可比性，为血细胞分析的标准化和临床实验室认可提供实验数据。方法参考CLSI的EP9-A2文件，以SysmexSF3000血细胞分析检测系统为比较（参考）方法（x），其它不同检测系统（Y1：BC-5300和Y2：xT-1800i）为试验方法，检测病人新鲜全血中8项血细胞分析结果，用单因素方差分析法和回归统计法分析试验方法（Y）和比较方法（x）测定结果在医学决定水平（xc）处的绝对偏差和相对偏差，以美国临床实验室修正法规（CLIA’88）规定的室间质量评价允许误差范围的1／2为标准。判断不同检测系统的临床可接受性。结果BC-5300和xT-1800i试验检测系统（Y1，Y2）与SF3000比较检测系统（x）检测80份血样8项血细胞分析结果的相关性研究表明：Y1与X间有75％的项目r〉0．975，分别是WBC，RBC，HGB，HCT和PLT，25％的项目r在0．810～0．975之间，分别是MCH和MCHC。Y2与X间有75％的项目r〉0．975，分别是wBC，RBC，HGB，HCT和PLT，25％的项目r在0．880～0．975之间，分别是MCH和MCHC。临床接受性能研究表明：Y1在22．2％的Xc处血细胞分析项目结果基本接受，在77．78％的Xc处血细胞分析项目结果完全接受，所有Xc均接受的项目有RBC，HGB和MCV。Y2在18．5％的Xc处血细胞分析项目的结果基本接受，在81．5％的xc处血细胞分析项目的结果完全接受，所有Xc均接受的项目有HGB，MCV，MCH和MCHC。两检测系统均无不接受的项目。通过同型新鲜混合全血对日常校准品重新赋值，从而达到试验检测系统（Y1，Y2）和比较检测系统（X）间检测结果的一致性。结论当同一实验室同一检验项目存在2个以上的检测系统时，应进行方法比对和偏差评估，判断其临床可接受性，以保证其检验结果的可比性。

Objective The comparability of the results of eight items in whole blood analysis,including WBC, RBC, HGB, HCT, MCV, MCH, MCHC and PLT, measured by different detection system in a same laboratory department,were investi- gated to provide experimental data for standardization of whole blood analysis and medical laboratory accreditation by bias estimation and method comparison. Methods Referring to the protocol of CLSI EP9-A2,compared the results of eight items of blood analysis in detection system of Sysmex SF3000 analyzer, which was applied as the comparative methods to detect the fresh whole blood samples of the patients, with those in other different detection systems（Y1 ： BC-5300 and Y2： XT-1800i）, which were used as test methods to detect the same samples. One-way within ANOVA and linear regression statistic proce- dure were applied to analyze the absolute and relative bias of the test mean results respectively with the comparative method （X） and test methods（Y） at every medical decision level（Xc）,which were compared with the criteria of the half allowable errors from PT regulated by CLIA＇ 88 in USA, to evaluate the clinical acceptability between the comparative method（X） and the test methods（Y） in different systems. Results The correlation study between BC-5300/XT-1800i test detection system （Y1, Y2） and SF3000 comparative detection system （X） shows that： r〉0. 975 in the 75 % items（WBC, RBC, HGB, HCT and PLT）,r 0. 810～0. 975 in 25% items（MCH and MCHC） between Y1 and X. r〉0. 975 in 75% items（WBC,RBC,HGB, HCT and PLT） ,r 0. 880～0. 975 in 25% items（MCH and MCHC） between Y2 and X. The Y1 and Y2 clinical acceptability studies had shown that：22.2% results at Xc levels of the items in blood cells analysis of Y1 were accepted basically and77.8% results of the Xc levels of those were fully accepted. Items accepted fully at all Xc levels of Y1 were RBC, HGB and MCV. Similarly, 18.5% results of the items at Xe levels in blood cells analysis of Y2 were accepted basically and 81.5 % re- sults at Xc levels of those were fully accepted. Items accepted fully at all Xc levels of Y2 were HGB, MCV, MCH and MCHC. There were no unaccepted items in two test systems. The test results consistency were achieved between the test de- tection systems （Y1, Y2） and the comparative detection system （X） by reassigning values from fresh whole blood to routine verifiers. Conclusion If there are more than two systems for the same test used in a same laboratory, method comparison and bias estimation is necessary to be carried on to assess their clinical acceptability and insure the comparability of their test results.