Objective: To evaluate the impact of long-term, non-suppressive levothyroxine (L-T 4 ) treatment on quantitative ultrasonometry in women.Design: This was a case-control study. Subjects and methods: Altogether 667 women (mean age Ϯ S.D., 49.5 Ϯ 13.1 years) were studied. Of these, 156 (23%) had non-toxic goitre or hypothyroidism and had been taking L-T 4 (75-100 mg/day) for at least 5 years (mean Ϯ S.D., 12.5 Ϯ 7.5 years); the remaining 511 (77%) women were not receiving L-T 4 . All women had completed a questionnaire on risk factors for thyroid dysfunction and osteoporosis, and those with diseases or treatments known to effect bone metabolism -other than thyroxine or hormone replacement therapy (HRT) -were excluded. Women underwent quantitative ultrasonometry (QUS) at the heel. Speed of sound (SOS), broadband ultrasound attenuation (BUA) and the stiffness index (SI) were compared, first, in all women taking L-T 4 and controls and, secondly, in women taking L-T 4 and controls pair-matched for age, weight, body mass index (BMI), menopausal status and HRT use. Results: Even after matching for age, weight, BMI, menopausal and HRT status, women taking L-T 4 had significantly lower values for SOS and SI (P < 0.05), but not for BUA. However, absolute T-and Z-scores for SI were not low in either the study or control groups. Lower values were associated, but not significantly so, with years since the menopause and duration of L-T 4 treatment. Conclusions: Long-term, non-suppressive L-T 4 treatment in women with goitre or hypothyroidism was associated with a slight reduction in QUS values, which was more pronounced in postmenopausal women. This group could be at higher risk for osteoporotic fracture.