|
|
Background: When transitioning to new analytical platforms, laboratories must assess the analytical performances of their methods. This study aimed to verify the precision and trueness of the Maglumi X8 device for thyroid-stimulating hormone (TSH) and free thyroxine (FT4) tests and to compare its performance with that of the Advia Centaur XP system previously used in our laboratory.
Methods: Precision and trueness verifications for TSH and FT4 were performed using three levels of Bio-Rad Quality Control (QC) materials following the CLSI EP15-A3 guidelines. Each day consisted of one run with five replicates, resulting in 25 analyses performed using three levels of QC material over five days. Passing-Bablok regression and Bland-Altman analyses were performed for method comparison analysis, following the CLSI EP09c guidelines.
Results: The repeatability coefficients of variation (CVs) of TSH for levels 1, 2, and 3 were 2.170, 1.945, and 2.567%, respectively, whereas the within-laboratory (WL) CVs were 2.720, 2.786, and 2.609%, respectively. The repeatability CVs of FT4 for levels 1, 2, and 3 were 3.262, 1.326, and 0.696%, respectively, whereas the WL CVs were 4.848, 4.309, and 4.879%, respectively. The bias or overall mean values obtained in the study for TSH and FT4 levels were within the verification targets. TSH levels were found to be lower on Maglumi X8 [1.770 (1.190 to 2.790)] than Centaur XP [1.975 (1.185 to 3.315)]. FT4 levels were found to be higher on Maglumi X8 [1.305 (1.200 to 1.450)] than Centaur XP [1.210 (1.090 to 1.460)]. The bias between the two methods obtained from the Bland–Altman analysis for TSH and FT4 was -3.76% and 6.68%, respectively. Although the calculated bias for TSH fell within the desirable targets based on biological variation, FT4 did not meet these targets.
Conclusions: The precision and trueness verification results for TSH and FT4 demonstrated acceptable performance under the CLSI EP15-A3. TSH results between the two analyzers were consistent and can be transferred; FT4 results from Maglumi X8 require careful interpretation and may need harmonization and standardization due to the observed bias.
DOI: 10.7754/Clin.Lab.2024.241017
|
