Performance of Thyroid-Stimulating Immunoglobulin Bioassay and Thyrotropin-Binding Inhibitory Immunoglobulin Assay for the Diagnosis of Graves' Disease in Patients With Active Thyrotoxicosis

Graves' disease (GD) is caused by the stimulation of thyrotropin receptors by autoantibodies. We compared the diagnostic accuracy of the thyroid-stimulating immunoglobulin (TSI) bioassay and thyrotropin-binding inhibitory immunoglobulin (TBII) assay in differentiating GD from other causes of thyrotoxicosis.We retrospectively evaluated 493 patients with thyrotoxicosis who were tested with the third-generation TSI and TBII assays simultaneously. Patients were classified according to the clinical, histopathologic, and imaging criteria into the following groups: positive reference group (PRG) (patients with GD), negative reference group (NRG) (patients without GD), and inconclusive group (patients without a definitive diagnosis).TSI and TBII assays were concordant in 88% of the cases and showed a strong positive correlation (rs = 0.844, P < .01). When analyzed collectively, TSI and TBII assays confirmed the diagnosis of GD in 79% of the PRG cases and excluded GD in 92.5% of patients in NRG. Combined TSI and TBII assays or TBII assay alone showed similar accuracy to the diagnosis of GD (81.4% and 77.5%, respectively). Tests in 40 of 191 patients in PRG were negative for both TSI and TBII assays, whereas 3 of 40 cases in NRG had at least 1 positive thyrotropin receptor antibody test. False-negative cases were associated with subclinical hyperthyroidism, normal radionuclide uptake, longer duration of thyrotoxicosis, and absence of goiter or Graves' ophthalmopathy.TSI and TBII assays showed similar performance in differentiating GD from other causes of thyrotoxicosis in a real-world sample of patients with active thyrotoxicosis. In combination, both tests showed little benefit compared with the TBII assay alone. Thyrotropin receptor antibody assay results should be carefully interpreted in patients with mild GD or longstanding disease.