The interference of familial dysalbuminemic hyperthyroxinemia (FDH), antithyroxine (anti-T<sub>4</sub>) or antitriiodothyronine (anti-T<sub>3</sub>) antibodies in serum free T<sub>4</sub> (FT<sub>4</sub>) or free T<sub>3</sub> (FT<sub>3</sub>) assays depends on the assay method: two-step immunoextraction, one-step labelled ligand (analogue or derivative tracer) or labelled antibody (solid-phase antigen-linked technique: SPALT) method. For FDH subjects, only FT<sub>4</sub> measurements by one two-step method gave reliable results in all cases. However, the overestimation was less marked with new SPALT assays than with previous analogue-based assays. In sera containing anti-T<sub>4</sub> antibodies, two-step assays and SPALT assays with a T<sub>3</sub>-coated solid phase can be considered as reliable for FT<sub>4</sub> determination. All other methods including SPALT assays with a T<sub>4</sub>-coated solid phase may give falsely high results. In sera containing anti-T<sub>3</sub> antibodies, SPALT FT<sub>4</sub> and FT<sub>3</sub> assays with a T<sub>3</sub>-coated solid phase may give spuriously high values. Anti-T<sub>3</sub> antibodies do not interfere in two-step FT<sub>3</sub> assays and SPALT FT<sub>3</sub> assays with a T<sub>2</sub>-coated solid phase, but may give high FT<sub>3</sub> values as measured by analogue or derivative methods. Moreover, SPALT free thyroid hormone assays may also be subject to interference from antibodies directed towards the assay solid phase.