Spuriously high value of serum free triiodothyronine (FT3: Amerlex free T3 kit, Amersham, UK.) was noted accidentally on routine laboratory examination of two clinically euthyroid patients (case 1: FT3; 18.5 pg/ml, FT4; 1.1 ng/dl, T3; 103 ng/dl, T4; 8.2 micrograms/dl, TSH; 1.74 microU/ml, case 2: FT3; 8.5 pg/ml, FT4; 1.1 ng/dl, T3; 137 ng/dl, T4; 8.9 micrograms/dl, TSH; 1.45 microU/ml), the former with poorly controlled diabetes (FBG 253 mg/dl, HbA1c 12.1%) and the latter with essential hypertension (184/108 mmHg). Although the hypertensive patient showed mild diffuse goiter, there was no evidence that the patients had autoimmune thyroid diseases because anti-thyroglobulin antibody tests measured by radioimmunoassay and MCHA, TGHA or TBII were all negative. Their serum levels of TBG were within the normal range. Further studies revealed that both patients' sera had unusual binding activity to labelled polyaminocarboxy T3 (125I-aT3) but not labelled T3 (125I-T3). Furthermore, this binding protein was precipitated by goat anti-human immunoglobulin G (IgG). The IgG purified from both patients' sera also showed strong binding activity to 125I-aT3, which was inhibited by unlabelled T3 in a dose dependent manner. In conclusion, we found anti-T3 antibody in two clinically euthyroid patients with no apparent evidence of complicating autoimmune thyroid diseases. The stronger binding activity to polyaminocarboxy T3 rather than T3 may lead to the spuriously high value of serum FT3. The mechanisms of the production of such autoantibodies in our cases should be further investigated.