OBJECTIVE To compare trabecular and compact bone response and relationship to oestrogen status using continuous oestradiol valerate 2 mg and norethisterone 0.7 mg daily as hormone replacement and to determine the therapeutic range of 17 beta-oestradiol. METHODS Open label trial. METHODS Independent endocrine clinic METHODS One hundred and thirty-one patients were compared at point of entry and at 36 months. METHODS Postmenopausal women were assessed using a Lunar dual photon and single photon bone scanner, and bone mineral density of the lumbar spine, right hip and left forearm were annually correlated with 17 beta-oestradiol and oestrone levels over three years. Total alkaline phosphatase was compared between improvers and decliners of bone mineral density. RESULTS Significant improvement in bone mineral density (P < 0.0001) occurred at all sites except the left forearm, where bone loss was prevented. There was no correlation between oestrogen levels and bone mineral density improvements at hip sites. However, in the lumbar spine larger improvements in bone mineral density occurred in women with 17 beta-oestradiol levels > 185 pmol/L compared with those below, which were statistically significant for those with 17 beta-oestradiol levels > 248 pmol/L. Bone turnover, as quanitifed by total alkaline phosphatase measurements, was suppressed in most patients, but there were no differences in the mean alkaline phosphatase levels between the best improvers and worst decliners for lumbar spine bone mineral density. Improvers had an age mean of 5.21 years greater than decliners (P = 0.01) and a mean duration difference since the menopause of 5 1 years compared with decliners (P = 0.007). CONCLUSIONS This combined continuous preparation of hormone replacement therapy improves not only trabecular bone but prevents compact bone loss, and the data suggest that the therapeutic range of 17 beta-oestradiol is between 200 pmol/L and 350 pmol/L.