A-061 Analytical Performance and R55G Mutant Detection Ability of an Improved Thyroid-Stimulating Hormone Assay, TSH3-Ultra II, on the ADVIA Centaur Immunoassay System and the Atellica IM Analyzer

Abstract Background Measurement of thyroid stimulating hormone (TSH) is a critical tool for the investigation of thyroid disorders. Previous studies have demonstrated that some immunoassays are unable to detect TSH from patients with the R55G mutation in the TSHβ subunit. This study evaluated the analytical performance of the ADVIA Centaur® and Atellica® IM TSH3-Ultra II (TSH3ULII) assay and its ability to detect the R55G TSH mutant. Methods Limit of quantitation (LoQ) and linearity were determined as described in CLSI EP17-A2 and EP06, respectively. Precision and method comparison (MC) studies were performed with three reagent lots according to CLSI EP05-A3 and CLSI EP09c using native and contrived human serum and plasma samples. For precision, one aliquot of each sample was tested in duplicate in two runs per day >2 hours apart on each analyzer for ≥20 days. MC studies were performed by testing native human serum samples using ADVIA Centaur TSH3-UL assay (current TSH assay) and the newer ADVIA Centaur TSH3ULII assay. An additional MC was performed by testing native human serum samples with the TSH3ULII assay on both the ADVIA Centaur XP and the Atellica IM Analyzer. Samples from 61 patients with TSH results below the LoQ when using TSH3-UL were sequenced and confirmed to have the R55G mutant. Each of these samples was tested in duplicate using the ADVIA Centaur TSH3ULII assay. Results LoQ on both the ADVIA Centaur XP and the Atellica IM Analyzer was estimated to be 0.010 μIU/mL (mIU/L). Linearity was verified from 0.010 to 150.000 μIU/mL (mIU/L). ADVIA Centaur TSH3ULII repeatability %CVs ranged from 1.4 to 3.6% at 0.099 and 91.993 μIU/mL (mIU/L), and within-lab %CVs ranged from 2.3 to 6.2% at 4.908 and 91.993 μIU/mL, respectively. Atellica IM TSH3ULII repeatability %CVs ranged from 1.1 to 3.4% at 4.800 and 93.097 μIU/mL, and within-lab %CVs ranged from 2.2 to 4.6% at 0.103 and 93.097 μIU/mL, respectively. For the MC studies, slopes determined by the Passing-Bablock regression model were approximately 1 (0.96-0.97 range) across the sample interval tested. All TSH3ULII results from the 61 mutant patient samples were greater than LoQ, ranging from 0.864 to 23.443 μIU/mL (mIU/L). Conclusions Evaluation of the TSH3ULII assay using the ADVIA Centaur system demonstrated performance comparable to the commercially available TSH3-UL assay for LoQ, linearity, and precision, when using normal patient samples. TSH3ULII results on the Atellica IM Analyzer were compara