We examined the effect of endogenous free fatty acids (FFA) on the measurement of free thyroxin (FT4) by five different methodologies represented in 16 different assays in a large number of patients with nonthyroidal illness (NTI). Some, but not all, one-step (analog) FT4 RIAs negatively correlated with FFA concentration. All two-step FT4 RIAs, equilibrium dialysis FT4, and the dialyzable (free) fraction of T4 positively correlated. In contrast, a binding-rate-based FT4 RIA, FT4 indices based on T3 macroaggregated albumin uptake, and T4/TBG ratios did not correlate. We also analyzed the FT4-FFA relationship with a second, more sensitive approach by correlating test results with FFA/albumin molar ratio as an estimate of the "excess" (nonalbumin bound) FFA. We found that all FT4 RIAs, equilibrium dialysis FT4, FT4 indices based on T3 uptake, the dialyzable fraction of labeled T4 in equilibrium dialysis, the fraction of labeled T4 bound to solid phase antibody in the binding-rate-based RIA, and T3 uptake correlated with the FFA/albumin molar ratio. This FFA dependency was comparable among all the various techniques and was relatively small. Thus, increases or decreases in FT4 results due to varying FFA (and albumin) concentrations are highly likely with most currently available methods (only the T4/TBG ratio did not reveal FFA-dependency), but the magnitude of changes varies with the "excess" FFA.