Department of Clinical Biochemistry, Steno Diabetes Center Copenhagen, Herlev, Denmark; Department of Pathology, Rigshospitalet, Copenhagen, Denmark.
Department of Clinical Biochemistry, Steno Diabetes Center Copenhagen, Herlev, Denmark; Department of Clinical Microbiology, Hvidovre Hospital, Hvidovre, Denmark.
Clin Biochem. 2022 Oct;108:5-9. doi: 10.1016/j.clinbiochem.2022.06.014. Epub 2022 Jul 3.
Assessment of albuminuria through albumin to creatinine ratio (ACR) is a widely used method to identify and monitor kidney damage. Significant interassay differences in urinary albumin quantification have been documented, which may affect ACR. This study was conducted to assess quantification of urinary albumin and urinary creatinine by two different analytical platforms, Cobas 6000 and Atellica CH 930, to examine the concordance of ACR stratification.
60 urinary albumin and 60 urinary creatinine concentrations were analyzed by Cobas 6000 and Atellica CH 930 using immunoturbidimetric assays for urinary albumin quantification and enzymatic assays for urinary creatinine quantification. Analytical performance was evaluated using Passing Bablok regression, Bland Altman plots, desirable specifications for inaccuracy and imprecision, and Wilcoxon test. Clinical significance was assessed through total error allowance (TEa) and concordance of ACR categories through Cohen's Kappa coefficient. Imprecision was assessed using control material of two levels.
Results were within desirable specifications for inaccuracy. Statistical differences were found (p < 0.05) for both analytes. TEa results were exchangeable for urinary creatinine, whereas no exchangeability was found for urinary albumin. Cohen's Kappa confirmed an almost perfect agreement of ACRs between the two methods (K = 0.87), testing 42 samples. Five of the 42 samples were stratified into different categories of ACR. Results from control material were within limits of acceptable imprecision (CV < 5%).
The findings of the study suggest that while differences in urinary creatinine results are not clinically significant, differences in urinary albumin results are. Despite an almost perfect agreement between the ACR results from Cobas 6000 and Atellica CH 930, there is a risk of incorrectly understanding a patient's kidney disease progression.
通过白蛋白与肌酐比值(ACR)评估白蛋白尿是识别和监测肾脏损害的常用方法。已经记录到尿液白蛋白定量存在显著的分析间差异,这可能会影响 ACR。本研究旨在评估两种不同分析平台 Cobas 6000 和 Atellica CH 930 对尿液白蛋白和肌酐的定量检测,并检查 ACR 分层的一致性。
使用免疫比浊法对 60 例尿液白蛋白和 60 例尿液肌酐进行 Cobas 6000 和 Atellica CH 930 分析,对尿液白蛋白进行定量分析,对尿液肌酐进行酶法分析。采用 Passing-Bablok 回归、Bland-Altman 图、不精密度和不准确度的理想规格以及 Wilcoxon 检验评估分析性能。通过总允许误差(TEa)和 ACR 分类的一致性(Cohen Kappa 系数)评估临床意义。使用两个水平的控制材料评估不精密度。
结果在不准确度的理想规格内。两种分析物均发现具有统计学差异(p<0.05)。对于尿液肌酐,TEa 结果是可互换的,而对于尿液白蛋白,TEa 结果则不可互换。Cohen Kappa 证实了两种方法之间 ACR 的一致性几乎为完美(K=0.87),检测了 42 个样本。42 个样本中有 5 个被分为不同的 ACR 类别。控制材料的结果在可接受的不精密度范围内(CV<5%)。
研究结果表明,虽然尿液肌酐结果的差异没有临床意义,但尿液白蛋白结果的差异是有临床意义的。尽管 Cobas 6000 和 Atellica CH 930 的 ACR 结果几乎完全一致,但仍存在错误理解患者肾脏疾病进展的风险。