OBJECTIVE Hypothyroidism is a common disorder and while the association of overt hypothyroidism with hypercholesterolaemia is clear, the effect upon lipids of the minor abnormalities of thyroid function often found in those receiving T4 replacement therapy is unclear. The aim of the present studies was to define in those with hypothyroidism the effect upon circulating lipids of subtle changes in thyroid status, indicated by serum TSH values below or above the normal range. METHODS (i) Prospective study of short-term T4 treatment of those with subclinical hypothyroidism. (ii) Cross-sectional study of long-term T4 treatment, comparing non-T4 treated controls and T4 treated patients with serum TSH values either below or within the normal range. METHODS Short-term T4 therapy was examined in 11 post-menopausal females with subclinical hypothyroidism (high TSH, normal free T4) studied before therapy and 6 weeks after T4 at incremental doses (50, 100, 150 micrograms/day). Long-term T4 treatment for hypothyroidism was investigated in 105 females on therapy for at least 1 year compared with 105 controls matched for age and menopausal status. METHODS Fasting levels of total cholesterol, low density lipoprotein (LDL) cholesterol, high density lipoprotein (HDL) cholesterol and HDL subfractions were determined in patients and controls. RESULTS (i) T4 treatment of subjects with subclinical hypothyroidism resulted in reductions in total and LDL cholesterol with increasing T4 dose (P < 0.001). Comparison of lipid results pretreatment with those when TSH was restored to normal revealed no significant difference in lipids; in contrast, comparison of lipids in those with normal TSH compared with the same subjects when TSH was suppressed revealed reductions in total and LDL cholesterol. (ii) Long-term T4 treatment was associated with a reduction in total and LDL cholesterol measurements in those over 55 years receiving suppressive doses of T4 but no significant difference in lipids in those with normal serum TSH compared with non-T4 treated controls. CONCLUSIONS Effects of T4 upon lipid measurements suggest that patients with subclinical hypothyroidism should receive replacement therapy. Doses of T4 which suppress TSH to below normal may have a more significant influence upon lipids than doses of T4 which restore TSH to the normal range.