Department of Pathology, Virginia Commonwealth University, Richmond, VA, USA.
Clin Chem. 2011 Jan;57(1):76-83. doi: 10.1373/clinchem.2010.148106. Epub 2010 Nov 19.
Good laboratory practice includes verifying that each new lot of reagents is suitable for use before it is put into service. Noncommutability of quality control (QC) samples with clinical patient samples may preclude their use to verify consistency of results for patient samples between different reagent lots.
Patient sample results and QC data were obtained from reagent lot change verification records for 18 QC materials, 661 reagent lot changes, 1483 reagent lot change-QC events, 82 analytes, and 7 instrument platforms. The significance of between-lot differences in the results for QC samples compared with those for patient samples was assessed by a modified 2-sample t test adjusted for heterogeneity of QC and patient sample measurement variances.
Overall, 40.9% of reagent lot change-QC events had a significant difference (P < 0.05) between results for QC samples compared with results for patient samples between 2 reagent lots. For QC results with differences <1.0 SD interval (83.1% of total), 37.7% were significantly different from the changes observed for patient samples. For QC results with differences ≥1.0 SD interval (16.9% of total), 57.0% were significantly different from those for patient samples.
Occurrence of noncommutable results for QC materials was frequent enough that the QC results could not be used to verify consistency of results for patient samples when changing lots of reagents.
良好的实验室规范包括在试剂投入使用前验证每批新试剂是否适用。质控（QC）样本与临床患者样本的不可互换性可能会阻止它们用于验证不同试剂批次之间患者样本结果的一致性。
从 18 种 QC 材料、661 个试剂批次变化、1483 个试剂批次变化-QC 事件、82 个分析物和 7 个仪器平台的试剂批次变化验证记录中获得患者样本结果和 QC 数据。通过调整 QC 和患者样本测量方差异质性的改良两样本 t 检验评估 QC 样本结果与患者样本结果之间批间差异的显著性。
总体而言，40.9%的试剂批次变化-QC 事件在 2 个试剂批次之间的 QC 样本结果与患者样本结果之间存在显著差异（P<0.05）。对于差异<1.0 SD 间隔（总差异的 83.1%）的 QC 结果，有 37.7%与患者样本观察到的变化显著不同。对于差异≥1.0 SD 间隔（总差异的 16.9%）的 QC 结果，有 57.0%与患者样本的结果显著不同。
QC 材料不可互换的结果发生频率足够高，以至于在更换试剂批次时，QC 结果不能用于验证患者样本结果的一致性。
