In this 2-year, randomized study, we compared the efficacy and tolerability of tibolone 2.5 mg (n = 75), tibolone 1.25 mg (n = 76) and estradiol 2 mg plus norethindrone acetate 1 mg (E2/NETA: n = 74) for preventing bone loss in postmenopausal women. Bone mineral density (BMD), measured by dual-energy X-ray absorptiometry, and bone remodeling markers were assessed every 6 months. Side-effects were assessed quarterly. After 24 months, the mean increase (+/- SD) in lumbar spine BMD from baseline was 3.6% +/- 2.9%, 1.9% +/- 3.5% and 6.8% +/- 4.5% in the tibolone 2.5 mg, tibolone 1.25 mg and E2/NETA groups, respectively. All pairwise differences were significant. The proportion of responders (women with a change from baseline in lumbar spine BMD of > or = -2% after 2 years) was 95.7%, 89.0% and 98.5% with tibolone 2.5 mg, tibolone 1.25 mg and E2/NETA, respectively. Similar results were obtained for femoral BMD, although the difference between tibolone 2.5 mg and E2/NETA was not significant at 24 months. Decreases in bone remodeling markers were similar in the three groups. Vaginal bleeding was more common in the E2/ NETA group (33.8%) than with tibolone 2.5 mg (12.0%) or tibolone 1.25 mg (9.2%), as was breast pain (23.0%, 2.7% and 2.6%, respectively). Each treatment effectively prevented bone loss. Overall, tolerability of tibolone was better than with E2/NETA, because of less frequent vaginal bleeding and breast pain. This may promote long-term adherence.