OBJECTIVE The association between established hypothyroidism and high cholesterol levels is well known. The aim of the present study was to investigate the effect of thyroxine (T4) administration on cholesterol levels in hypercholesterolemic subjects with TSH levels within the normal range ('high-normal' TSH compared with 'low-normal' TSH). METHODS We determined TSH levels in 110 consecutive patients referred for hypercholesterolemia (serum cholesterol >7.5 mmol/l). Those with 'high-normal' TSH (2.0-4.0 microU/ml) as well as those with 'low-normal' TSH (0.40-1.99 microU/ml) were randomly assigned to receive either 25 or 50 microg T4 daily for two months. Thus, groups A and B (low-normal TSH) received 25 and 50 microg T4 respectively and groups C and D (high-normal TSH) received 25 and 50 microg T4 respectively. Serum T4, tri-iodothyronine (T3), TSH, free thyroxine index, resin T3 uptake and thyroid autoantibodies (ThAab) as well as total cholesterol, high and low density lipoprotein cholesterol (HDL, LDL), and triglycerides were determined before and at the end of the two-month treatment period. RESULTS TSH levels were reduced in all groups. The most striking effect was observed in group D (TSH levels before: 2.77+/-0.55, after: 1.41+/-0.85 microU/ml, P < 0.01). Subjects in groups C and D had a higher probability of having positive ThAabs. A significant reduction in total cholesterol (P < 0.01) and LDL (P < 0.01) was observed after treatment only in group D. In those subjects in group D who were ThAab negative, there was no significant effect of thyroxine on cholesterol levels. CONCLUSIONS Subjects with high-normal TSH levels combined with ThAabs may, in fact, have subclinical hypothyroidism presenting with elevated cholesterol levels. It is possible that these patients might benefit from thyroxine administration.