B-009 Verification of Result Interchangeability Between Two Analyzers in the Determination of Uric Acid and Urea

Abstract

Background Uric acid and urea are two parameters widely used in the diagnosis and management of renal pathology. In our laboratory, they are determined using Atellica Solution® analyzers (Siemens Healthineers). After validating a new Siemens equipment, Atellica CI®, it is essential to ensure result interchangeability in our laboratory accredited by ISO 15189:2012. Objectives evaluate the interchangeability of results for uric acid and urea measured by Atellica Solution® and CI® analyzers. Methods A total of 40 serum samples from patients with uric acid values between 1.3 and 15.2 mg/dL, and another 40 serum samples from patients with urea values between 7 and 142 mg/dL were processed in parallel. Statistical analysis of the results was performed using Passing-Bablok regression and Bland Altman mean difference analysis with Method Validator Version 1.19. Results are expressed with their 95% confidence interval. Results Uric Acid: After analyzing Bland-Altman differences, a statistically significant slight systematic error was found, as the 95% CI does not include 0, with slightly higher results in CI 1900®. Regarding Passing-Bablok regression analysis, there were no systematic errors of constant type (since the 95% CI of the intercept includes 0) but a discreet significant proportional systematic error was observed (the confidence interval of the slope does not include 1).Urea: In Bland-Altman analysis, a discreet systematic error was observed as the 95% CI of the mean difference does not include the null value. Regarding Passing-Bablok regression analysis, there were no systematic errors of constant type (since the 95% CI of the intercept includes 0) or proportional type (since the 95% CI of the slope includes 1), indicating interchangeability between both analyzers. Conclusions The results obtained by both Atellica Solution® and CI® analyzers for uric acid and urea assays are interchangeable since the statistically significant differences observed can be considered analytically and clinically insignificant.