We directly compared the competitor potency for serum T4 binding of 11 nonsteroidal antiinflammatory drugs; the diuretics furosemide, ethacrynic acid, and bumetanide; diphenylhydantoin; the cholecystographic contrast agents iopanoate and ipodate; and six long-chain nonesterified fatty acids (NEFA) using equilibrium dialysis. To avoid artefacts that occur in competitor studies with diluted serum or isolated binding proteins, we used undiluted normal serum, with drugs added at concentrations that achieved high therapeutic total and free serum levels at equilibrium. Drug addition was based on the measured free fraction of each drug in serum. The free T4 fraction in normal serum (Tris buffer, pH 7.4; 37 C) was between 1.40 X 10(-4) and 1.53 X 10(-4). Drug-induced increases in T4 free fraction were: fenclofenac, 90%; aspirin, 62%; meclofenamic acid, 39%; diflunisal, 37%; mefenamic acid, 31%; and furosemide, 31%. Significant increases of 7-15% occurred with diclofenac, flufenamic acid, phenylbutazone, and diphenylhydantoin. Indomethacin, ketoprofen, tolmetin, ethacrynic acid, bumetanide, iopanoate, and ipodate were inactive at the concentrations studied. Addition of 2.0 mmol/L oleic acid had a negligible effect, but 3.5 mmol/L oleic acid inhibited T3 and T4 binding significantly. Other long chain NEFA (addition of 1.5 mmol/L) gave increases in free T4 fraction as follows: arachidonic acid, 26%; linolenic acid, 23%; and linoleic acid, 11%. Stearic and palmitic acids were inactive. The effect of 5 mmol/L oleic acid in serum could be reproduced by addition of 0.5 mmol/L to serum diluted 1:10, indicating that protein binding of NEFA is the major determinant that limits their competitor potency. These findings provide a basis for anticipating which potential inhibitors may cause important changes in serum thyroid hormone binding. The time course of such effects will be influenced by the pharmacokinetics of the inhibitor itself as well as the equilibrium findings described here.