Changing Immunochemistry Platforms: Thyroid Function Test Comparison and Reference Intervals Based on Clinical Needs

Background: Diagnosis of thyroid dysfunction relies on thyroid stimulating hormone (TSH), free thyroxine (FT4), and free tri-iodothyronine (FT3) tests against valid reference intervals (Rls). We changed the immunoassay platform from Abbott Architect to Siemens Atellica and aimed to establish Atellica Rls based on laboratory information system (LIS) patient data. Methods: Atellica thyroid hormone immunoassays were verified against those of Architect. Real-life patient results were retrieved from LIS. A single result per patient dataset was used to establish the Rls by the indirect method. Results: Atellica and Architect assays correlated well but Atellica showed a positive bias between 13% and 53%, the largest for FT4. Variations of the Atellica assays were <= 4%. The 95% Atellica Rls were 0.4-3.8 mU/L for TSH, 0.9-1.6 ng/dL for FT4, and 227-416 pg/dL for FT3. Considering the accumulating clinical experience with Atellica, the Rls for clinical use were adjusted as 0.5-4.0 mU/L, 0.9-1.8 ng/dL, and 169-409 pg/dL, respectively. Conclusions: We verified thyroid hormone Rls for Atellica by the indirect method for the first time. Our model proved reliable for selecting results of presumably healthy individuals from LIS data. Critical review of the Rls with local endocrinologists is essential.