Department of Laboratory Medicine, Green Cross Laboratories, Yongin, Korea.
Department of Laboratory Medicine, National Cancer Center, Goyang, Korea.
Korean J Intern Med. 2018 Nov;33(6):1119-1128. doi: 10.3904/kjim.2017.034. Epub 2017 Oct 25.
BACKGROUND/AIMS: In multicenter clinical trials, laboratory tests are performed in the laboratory of each center, mostly using different measuring methodologies. The purpose of this study was to evaluate coefficients of variation (CVs) of laboratory results produced by various measuring methods and to determine whether mathematical data adjustment could achieve harmonization between the methods.
We chose 10 clinical laboratories, including Green Cross Laboratories (GC Labs), the central laboratory, for the measurement of total cholesterol, high density lipoprotein cholesterol (HDL-C), low density lipoprotein cholesterol (LDL-C), serum triglycerides, creatinine, and glucose. The serum panels made with patient samples referred to GC Labs were sent to the other laboratories. Twenty serum samples for each analyte were prepared, sent frozen, and analyzed by each participating laboratory.
All methods used by participating laboratories for the six analytes had traceability by reference materials and methods. When the results from the nine laboratories were compared with those from GC Labs, the mean CVs for total cholesterol, HDL-C, LDL-C, and glucose analyzed using the same method were 1.7%, 3.7%, 4.3%, and 1.7%, respectively; and those for triglycerides and creatinine analyzed using two different methods were 4.5% and 4.48%, respectively. After adjusting data using Deming regression, the mean CV were 0.7%, 1.4%, 1.8%, 1.4%, 1.6%, and 0.8% for total cholesterol, HDL-C, LDL-C, triglyceride, creatinine, and glucose, respectively.
We found that more comparable results can be produced by laboratory data harmonization using commutable samples. Therefore, harmonization efforts should be undertaken in multicenter trials for accurate data analysis (CRIS number; KCT0001235).
背景/目的：在多中心临床试验中，实验室测试在每个中心的实验室进行，主要使用不同的测量方法。本研究的目的是评估不同测量方法产生的实验室结果的变异系数（CV），并确定数学数据调整是否可以实现方法之间的协调。
我们选择了 10 个临床实验室，包括格林十字实验室（GC 实验室）作为中心实验室，用于测量总胆固醇、高密度脂蛋白胆固醇（HDL-C）、低密度脂蛋白胆固醇（LDL-C）、血清甘油三酯、肌酐和葡萄糖。由患者样本制成的血清板被送到其他实验室。每个分析物准备 20 个血清样本，冷冻后发送并由每个参与实验室进行分析。
参与实验室用于六种分析物的所有方法均通过参考物质和方法具有可追溯性。当将九个实验室的结果与 GC 实验室的结果进行比较时，使用相同方法分析的总胆固醇、HDL-C、LDL-C 和葡萄糖的平均 CV 分别为 1.7%、3.7%、4.3%和 1.7%；使用两种不同方法分析的甘油三酯和肌酐的平均 CV 分别为 4.5%和 4.48%。使用 Deming 回归调整数据后，总胆固醇、HDL-C、LDL-C、甘油三酯、肌酐和葡萄糖的平均 CV 分别为 0.7%、1.4%、1.8%、1.4%、1.6%和 0.8%。
我们发现，使用可互换样本进行实验室数据协调可以产生更可比的结果。因此，在多中心试验中应进行协调努力，以进行准确的数据分析（CRIS 编号；KCT0001235）。
