The aim of the work was to examine the relationship between thyroglobulin autoantibody (TgAb) levels and the Tg recovery rate (TgRR) using different concentrations of Tg (50, 10, 5, and 1 μg/l) in the recovery buffer. A total number of 225 serum samples from individual patients were analyzed. Samples were selected for their TgAb in 6 groups: TgAb<limit of detection (LoD; n=50), LoD<TgAb≤60 IU/ml (n=62), 60 IU/ml<TgAb≤100 IU/ml (n=30), 100 IU/ml<TgAb≤500 IU/ml (n=41), 500 IU/ml< TgAb≤1 000 IU/ml (n=13) and TgAb>1 000 IU/ml (n=28). TgAb were measured with 2 different assays (VARELISA and BRAHMS Anti-Tgn RIA). TgAb levels and the TgRR determined using the 50, 10, 5, and 1 µg/l buffers showed strong significant correlations with a Spearmans' rho of - 0.720, - 0.688, - 0.686, and - 0.356, respectively, for the VARELISA assay and - 0.670, -0.617, - 0.570, and - 0.274, respectively, for the Anti-Tgn assay (all p<0.001). TgRRs were a median of 94.8% (30.5-113.0%), 90.8% (40.6-127.6%), 90.0% (8.2-119.3%), and 89.4% (range - 43.6-121.6%) for the TgRR determined using recovery buffers with concentrations of 50, 10, 5, and 1 μg/l respectively. With decreasing Tg concentration in the recovery buffer the percentage of abnormal results increased, although the extreme increase we found in the 1 μg/l group is largely caused by a lack of analytical precision in the 73 sera with Tg levels exceeding 5 μg/l. Our results give cause for further investigation into reviving the concept of Tg-recovery measurement using 5 μg/l Tg in the recovery buffer instead of the traditional 50 μg/l.