Context: Thyroglobulin autoantibodies (TgAb) have been proposed as a surrogate marker of thyroglobulin in the follow-up of differentiated thyroid carcinoma. Commercially available TgAb assays are often discordant. We investigated the causes of discrepancy.
Design: TgAb were measured by three noncompetitive immunometric assays and three competitive RIA in 72 patients with papillary thyroid carcinoma and associated lymphocytic thyroiditis (PTC-T), 105 with papillary thyroid carcinoma and no lymphocytic thyroiditis (PTC), 160 with Hashimoto's thyroiditis, and in 150 normal subjects. The results of the six assays were correlated. TgAb epitope pattern, evaluated by inhibition of serum TgAb binding to thyroglobulin by TgAb-Fab regions A, B, C, and D, were compared in sera which were positive in all six assays (concordant sera) and positive in only one to five assays (discordant sera) were compared. TgAb International Reference Preparation (IRP) was measured in 2007 and 2009.
Results: The correlations of the six assays ranged from -0.01 to 0.93 and were higher in PTC-T and Hashimoto's thyroiditis than in PTC and normal subjects. Two uncorrelated components, one including the three immunometric assays, the other the three RIA, explained 40 and 37% of the total variance of the results of the six assays. The levels of inhibition were higher in concordant sera than in discordant sera by TgAb-Fab region B (27.0%, 21.2-34.0 vs. 6.0%, and 2.7-12.7%) and region C (30.5%, 21.3-37.7 vs. 4.0%, and 1.0-6.5%); thus, the epitope pattern was more homogeneous in concordant sera than in discordant sera. TgAb IRP ranged from 157 to 1088 (expected 1000) IU/ml in 2009; results in 2007 were similar in all but two assays.
Conclusions: TgAb assays are highly discordant. Discrepancy is lower when comparing assays with similar methodology. Results of TgAb from PTC-T are more concordant than those from PTC because their epitope pattern is more restricted. The internal standardization of TgAb is generally, but not completely, satisfactory.