摘要目的 探讨部分临床实验室FCM淋巴细胞亚群分析参考范围应用的合理性及不同生产厂家的流式细胞仪和试剂组合对淋巴细胞亚群分析结果的影响.方法 根据国内临床实验室常用的3个流式细胞仪型号(Beckman Coulter Epics XL、Beckman Coulter Cytomics FC500、BD FACS Calibur),分别选取3家北京地区临床流式细胞室(A、B、c室),按照各室的实际检测方案分别测定50份健康人静脉血标本,以验证各室淋巴细胞亚群分析参考范围是否合理.调查3家实验室室内全血质控品使用情况,并将商品化全血质控品分发各室,按照各自的实际实验方案在20个工作日内与常规标本平行处理、检测和分析.针对不同生产厂家的试剂,在同一流式细胞仪(型号为Beckman Coulter Epics XL)上用a、b、c、d 4种不同的试剂组合对20份患者标本进行检测,其中试剂组合a为美国Beckman Coulter公司同厂配套试剂和仪器,试剂组合b、c、d的检测结果分别与试剂组合a比较,计算b、c、d试剂组合偏倚>10%的概率.采用相同试剂和溶血素(美国Beckman Coulter公司)对24份患者标本进行前处理,分别在2台不同厂家和型号的流式细胞仪(型号为Beckman Coulter Epics XL和BD FACS Calibur)上检测,比较相同试剂处理标本后不同仪器对淋巴细胞亚群分析结果的影响.采用同厂配套试剂和仪器,比较Beckman Coulter Epics XL和BD FACS Calibur两个流式细胞检测系统对20份患者标本检测结果的影响.结果 A室的自然杀伤(NK)细胞及CD+4 T淋巴细胞/CD+8 T淋巴细胞(T4/T8),B、c两室的T4均有大于10%的结果落在相应的参考范围之外,超出相应参考范围的概率分别为16%(9/50)、24%(12/50)、22%(11/50)、12%(6/50).3家实验室20个工作日内的室内质控均在参考范围内.与试剂组合a比较,试剂组合b、c的所有项目偏倚均较大,其中偏倚>10%的概率最低为试剂组合b的T8,为70%(14/20);最高为试剂组合b、c的T淋巴细胞(T3)、T4,均达到100%(20/20).试剂组合d的T3、T8和B淋巴细胞(B)偏倚较大,偏倚>10%的概率分别为35%(7/20)、85%(17/20)、75%(15/20).不同生产厂家的试剂、仪器处理和分析标本的结果,与采用同一生产厂家的试剂、仪器处理和分析的结果相比,T3、T4、T8、B、NK均存在较大偏倚,偏倚>10%的概率分别为71%(17/24)、80%(19/24)、38%(9/24)、33%(8/24)、92%(22/24).Beckman Coulter Epics XL和BD FACS Calibur两个流式细胞检测系统相比较,T8、NK和B的偏倚均较大,偏倚>10%的概率分别为55%(11/20)、70%(14/20)、55%(11/20).结论 流式细胞实验室需要建立自己的参考范围并定期验证,以便合理进行调整.建议定期采用全血质控品,并累计质控数据.各实验室应选择同厂配套试剂处理标本.

abstractsObjective To investigate the appropriate setting up of normal reference ranges of lymphocyte subsets in some flow cytometry laboratories and to study the effects of different flow cytometers and various reagents by different manufacturers on the analysis of peripheral blood lymphocyte subsets. Methods Three FCM labs (named A, B and C) in Beijing region were selected representing 3 commonly used flow cytometers (Beckman Coulter Epics XL, Beckman Coulter Cytomics FC500, BD FACS Calibur). 50 samples from healthy donors were distributed to 3 labs and tested according to individual lab's standard operating procedure to verify whether the normal reference ranges of peripheral blood lymphocyte subsets established were appropriate. The application of internal quality control was also investigated. Commercial blood quality control reagents were given to the 3 FCM labs and tested within 20 working days paralleled with routine samples. In addition, 20 patients' samples were prepared using 4 different combinations of reagents ( a , b , c and d). The results from combination a, which used the Beckman Coulter reagents and instrument, were compared to the results from combination b, c and d, which used reagents from different manufacturers. Then the prepared samples were tested on Beckman Coulter Epics XL to evaluate the effects of different combinations of reagents on the results of peripheral blood lymphocyte subsets analyzed by the same instrument. Furthermore, 24 patients' samples prepared by same reagents from Beckman Coulter company were tested on both Beckman Coulter Epics XL and BD FACS Calibur respectively to assess the effects of different instruments on peripheral blood lymphocyte subsets. 20 patients' samples prepared by same reagents and instruments were analyzed by Beckman Coulter Epics XL analytic system and BD FACS Calibur analytic system respectively to assess the effects of the two analytic systems on the lymphocyte subsets. Results Over 10% of the results for NK and T4/T8 in lab A as well as T4 in labs B and C fell outside of their normal reference ranges. The probabilities exceeding corresponding normal reference ranges were 16% ( 9/50 ), 24% ( 12/50 ), 22% (11/50) and 12% ( 6/50 ), respectively. The results using internal blood quality control in 3 FCM labs within 20 working days were all within the reference ranges of the quality control provided by the kit. The biases from b and c reagent combinations were substantial compared with that of reagent a combination. Among the biases from b and c reagent combinations, the lowest probability of bias exceeding 10% was T8 of combination b, which had probability of 70% (14/20). The highest probabilities of hias exceeding 10% were T3 and T4 of b and c reagent combinations, which reached 100% (20/20) . Furthermore, the biases of T3, T8 and B of d reagent combination compared with that of reagent a combination were also substantial. The probabilities of bias exceeding 10% were 35% (7/20) ,85% (17/20) and 75% (15/20), respectively. Comparing the results of samples prepared and analyzed by reagents and instruments from different manufacturers to that of samples prepared and analyzed by the same company's reagents and instruments showed that there were great discrepancies in T3, T4 , T8 , B and NK. The probabilities of bias exceeding 10% were 71% ( 17/24), 80% (19/24) ,38% (9/24), 33% (8/24) and 92% (22/24), respectively. The biases of T8, NK and B were substantial when compared the results from Beckman Coulter Epics XL analytic systems and BD FACS Calibur analytic systems. The probabilities of bias exceeding 10% were 55% (11/20 ), 70% ( 14/20 ) and 55% (11/20), respectively. Conclusions FCM labs should set up their own normal reference range for peripheral blood lymphocyte subsets. The normal reference range should be verified periodically. It is important to apply internal blood quality control regularly and accumulate the quality control results. The reagents and instrument for preparing peripheral blood samples should be from the same manufacturers.