Analytical performance evaluation of thyroid-stimulating hormone receptor antibody (TRAb) immunoassays

Background: Thyroid-stimulating hormone receptor (TSHR)-activating autoantibodies stimulate thyroid growth and hormone synthesis/secretion, causing hyperthyroidism of Graves' disease (GD). TRAb measurement helps diagnose GD and is an important first test in evaluating hyperthyroidism according to the recent American Thyroid Association guidelines. We compared the performance of the BRAHMS TRAK Kryptor (Thermo Scientific) and Roche cobas TRAb immunoassays for use in GD. Method: Method comparison (n = 40) and clinical agreement were assessed between the Kryptor, cobas e411, and cobas e601. The analytical performance of Kryptor and cobas e411 were assessed for within- and between-day imprecision across 20 days, linearity, functional assay sensitivity (FAS), dilution recovery, and cut-off verification. Results: The Kryptor, e411, and e601 TRAb immunoassays correlated well (r > 0.95, overall percent agreement = 0.95, Cohen's kappa = 0.90). With a total allowable error of 20%, percent bias was within 13%, which was minimally negative at< 20 IU/L, but highly positive (33%-34%) > 20 IU/L. The Kryptor, but not e411, was linear across the claimed analytical measuring range (AMR). The claimed functional assay sensitivity (FAS), which was close to the clinical GD cut-off 1.8 IU/L, was verified for Kryptor and e411. Conclusion: Overall, our evaluation demonstrates acceptable comparability between TRAb immunoassays with in-house imprecision up to 13% and 10% on Kryptor and e411, respectively. While Roche has preferable calibration frequency and on-board reagent stability, both platforms demonstrate acceptable imprecision using patient samples at their claimed FAS, which is important for GD diagnosis. Diluted results (using a negative patient pool as diluent) exhibits proportional positive bias on the Kryptor relative to the Roche methods.