Circulating iodothyronine-binding autoantibodies interfere with total T4 and T3 RIAs, giving falsely high or falsely low values depending on the assay separation method used. Direct serum free T4 (FT4) measurement should compensate for such abnormal binding if the method is free of artefact. We assessed 5 different FT4 methods in 12 patients with immunoglobulin binding of iodothyronines, with limited assessment of a sixth method in 2 subjects. Thyroid status was assessed clinically and by measurement of TSH and its response to TRH. T4 methods which use analog tracers, i.e. Amerlex, and Clinical Assays One-step FT4, gave spuriously high values in almost all hypothyroid and some euthyroid patients, due to immunoglobulin binding of the tracers used in these techniques. The kinetic indirect method (Corning Immunophase FT4) gave inappropriately high values in 4 of 6 hypothyroid patients. FT4 by equilibrium dialysis or by adsorption chromatography and RIA, accurately assessed thyroid status. These findings suggest that FT4 methods are valid in patients with circulating iodothyronine-binding immunoglobulins only if the free hormone fraction is physically separated from serum binding proteins before the assay procedure.