Our aims were: 1) to analyze the effect of the methodology used to derive clinically feasible cut-off values for thyroid peroxidase antibodies (TPOAb) and thyroglobulin antibodies (TgAb), which exhibit highly skewed distributions; and 2) to describe the influence of thyroid antibodies on thyroid stimulating hormone (TSH) and free thyroxine (FT4) reference intervals among thyroid disease-free aged subjects. The reference population consisted of 1086 individuals with a mean age of 73 years. The impacts of TPOAb and/or TgAb positivity on the reference intervals of TSH and FT4 were evaluated by both including and excluding subjects with elevated thyroid antibodies. The exclusion of subjects with elevated thyroid antibodies had no effect on the FT4 reference interval in either gender or on the TSH reference interval in men. Among women, the exclusion of 196 (34%) thyroid antibody-positive subjects resulted in lowering of the upper reference limit of TSH from 7.2 to 5.8 mIU/L. When the more stringent "mode-method" by summing mode+(mode-2.5th percentile) was used, 334 women (58%) were excluded and the upper reference limit of TSH remained essentially identical. Regardless of the statistical methodology used to derive cut-off values, the effect of antibody positivity was found to be less than expected.