The aim of this study is to investigate the long-term (9 months) effects of variable doses (200/100 microg/day) of L-selenomethionine on autoimmune thyroiditis (AIT) and the parameters affecting the success rate of this therapy. The present study was designed in three steps: (1) 88 female patients with AIT (mean age = 40.1 +/- 13.3 years) were randomized into two groups according to their initial serum TSH, thyroid peroxidase antibody (TPOAb) concentrations, and age. All the patients were receiving L-thyroxine to keep serum TSH <or=2 mIU/l. Group S2 (n = 48, mean TPOAb = 803.9 +/- 483.8 IU/ml) received 200 microg L-selenomethionine per day, orally for 3 months, and group C (n = 40, mean TPOAb = 770.3 +/- 406.2 IU/ml) received placebo. (2) 40 volunteers of group S2 were randomized into two age- and TPOAb-matched groups. Group S22 (n = 20) went on taking L-selenomethionine 200 microg/day, while others (group S21) lowered the dose to 100 microg/day. (3) 12 patients of group S22 (group S222) went on taking L-selenomethionine 200 microg/day, while 12 patients of group S21 (S212) increased the dose to 200 microg/day. Serum titers of TPOAb decreased significantly in group S2 (26.2%, P < 0.001), group S22 (23.7%, P < 0.01) and group S212 (30.3%, P < 0.01). There were no significant changes in group C and group S222 (P > 0.05). TPOAb titers increased significantly in group S21 (38.1%, P < 0.01). A significant decrease in thyroglobulin antibody titers was only noted in group S2 (5.2%, P < 0.01). L-selenomethionine substitution suppresses serum concentrations of TPOAb in patients with AIT, but suppression requires doses higher than 100 microg/day which is sufficient to maximize glutathione peroxidase activities. The suppression rate decreases with time.