4种脂蛋白相关磷脂酶A2活性测定试剂的性能评价

The performance evaluation of four Lp-PLA2 activity assays

目的评价4种脂蛋白相关磷脂酶A2 （Lp－PLA2）活性试剂在Beckman AU5800自动生化分析仪上的分析性能。方法试剂性能验证。收集2017年3至7月北京协和医院214例患者和140名表观健康人剩余血清，分别用于Lp－PLA2方法学比对及参考区间验证。参考美国临床和实验室标准化协会（CLSI） EP15－A、EP6－A、EP17－A和EP7－P文件，评价精密度、线性范围、灵敏度及常见干扰（胆红素、血红蛋白、乳糜）。参考EP9－A2方案，比较4种试剂 （A博源、B德赛、C恒晓、D中元）检测患者样本的一致性，并计算其在医学决定水平（328、391、485 U/L）处的预期偏差。结果A～D各水平重复性（CV%）分别为：0.5%～1.7%、0.7%～3.0%、0.9%～2.0%和0.5%～3.3%，实验室内不精密度（CV%）分别为：0.7%～2.9%、1.4%～3.2%、1.3%～1.9%和0.8%～4.1%，均小于实验室5%的质量目标；A～D线性范围分别为：44 ～1 992 U/L、39 ～1 798 U/L、13～540 U/L、75 ～1 717 U/L，回归决定系数R2在0.997～1.000之间，相关系数（r）在0.998～1.000之间；4种试剂抗乳糜干扰能力均达到制造商标注或满足临床需求，胆红素在低Lp－PLA2水平对C试剂干扰明显；B、C在血红蛋白4.5 g/L时受明显负干扰，而D在血红蛋白2.45 g/L时呈现正干扰。A、C、D与B测定结果的回归决定系数R2在0.978～0.995之间，相关系数（r）在0.989～0.998之间，在医学决定水平处的预期偏差在－240～113 U/L间。A～D分别有131（93.6%）、140（100%）、82（58.6%）、128例（91.4%）Lp－PLA2活性检测结果落在制造商标注参考区间内。结论4种应用于自动生化分析仪的Lp－PLA2活性检测试剂精密度、线性均良好，但抗干扰能力有待提高。（中华检验医学杂志，2018, 41：208－213）

ObjectiveTo evaluate the analytical performance of four lipoprotein associated phospholipase A2（Lp-PLA2） activity reagents on Beckman AU5800 automatic biochemical analyzer.MethodsThe remaining serum samples of 214 patients and 140 apparently healthy individuals were collected from March to July 2017 in Peking Union Medical College Hospital.These samples were used for method comparison and reference interval evaluation.According to the guidelines of EP15-A, EP6-A, EP-17 and EP7-P from Clinical and Laboratory Standards Institute （CLSI）standards, the precision, linearity, sensitivity and common interferences （e.g free bilirubin, conjunct bilirubin, hemoglobin and chyle） were assessed. According to EP9-A2, method comparisons of differents regents（Evermed, DiaSys, Hengxiao and Zhongyuan were labeled as A, B, C and D, respectively） were conducted and the differences were estimated at medical decision levels （328U/L, 391U/L and 485U/L）.ResultsThe precision of four reagents were acceptable. The repeatability （CV%） of A to D were 0.5%-1.7%, 0.7%-3.0%, 0.9%-2.0% and 0.5%-3.3%, respectively. The reproducibility（CV%） were 0.7%-2.9%, 1.4%-3.2%, 1.3%-1.9% and 0.8%-4.1%, respectively.Both of those achievedlaboratory defined quality objective （〈5%）. The linearity of A to D were 44 -1 992 U/L, 39 -1 798 U/L, 13 -540 U/L and 75 -1 717U/L, respectively.The regression coefficient R2 was between 0.997 and 1.000, and the correlation coefficient （r） was between 0.998 and 1.000. The interference of chyle were acceptable among these four reagents andmet the manufacturer′s requirementsor clinical needs.In a low level of Lp-PLA2, bilirubin had an obvious interferenceonreagent C; B and C were negatively affected when the hemoglobin was 4.5 g/L; and D was positively affected when the hemoglobin was 2.45 g/L. The regression coefficients R2 of A, C, D compared with B were between 0.978 and 0.995, and the correlation coefficients （r） were between 0.989 and 0.998. The expected differences at medical decision levels ranged from -240 U/L to 113 U/L. For A to D, the Lp-PLA2 activity results of 131 （93.6%）, 140 （100%）, 82 （58.6%）, and 128 （91.4%） cases were analysed within the manufacturer′s claimed reference intervals.ConclusionThe precision and linearity of the four Lp-PLA2 activity detection reagents used in automatic biochemical analyzer are good, but the anti-interference ability needs to be improved.（Chin J Lab Med, 2018, 41： 208-213）