Abstract
OBJECTIVES: 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.
METHOD: 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: 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%).
CONCLUSIONS: 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.
Originalsprog | Engelsk |
---|---|
Tidsskrift | Clinical Biochemistry |
Vol/bind | 108 |
Sider (fra-til) | 5-9 |
Antal sider | 5 |
ISSN | 0009-9120 |
DOI | |
Status | Udgivet - okt. 2022 |