Three hundred and one healthy women between 45 and 65 yr of age and at least 1 yr postmenopausal were randomly assigned to 12-month double-blind therapy with levormeloxifene [1.25 (n = 51), 5, 10, or 20 mg/day], low dose continuous combined hormone replacement therapy [HRT; 1 mg 17 beta-estradiol and 0.5 mg norethisterone acetate/day], or placebo (all n = 50). All of the women were also given a daily supplement of calcium (500 mg). Serum CrossLaps decreased by about 50% in the levormeloxifene groups, with no dose-response effect. The group receiving HRT decreased more (>60%), and the placebo group (500 mg calcium alone) decreased by about 10%. The pattern was similar for bone alkaline phosphatase, except that the decreases were smaller, about 30% for the levormeloxifene groups and 50% for the HRT group. Serum osteocalcin also showed highly significant decreases, of the same magnitude in the levormeloxifene and HRT groups. Spinal bone mineral density (BMD) decreased by less than 1% in the placebo group and increased by about 2% in the levormeloxifene groups and by almost 5% in the HRT group (P < 0.001 for the difference between levormeloxifene and HRT vs. placebo). BMD of the total hip and total body changed in the same direction, although differences between groups were not as pronounced as those for BMD spine. Total cholesterol decreased by about 13--20% during levormeloxifene therapy, whereas daily doses of 1 mg estradiol and 0.5 mg norethisterone acetate produced a decrease of only about 8%. Levormeloxifene decreased low density lipoprotein cholesterol by about 22-30% compared with about 12% in the low dose HRT group. High density lipoprotein cholesterol was unchanged in all groups. Endometrial thickness increased both clinically and statistically significantly in the levormeloxifene groups independently of the dose; the difference from the placebo and HRT groups was significant (P < 0.001). There was no significant difference between the HRT and placebo groups. Other adverse events of interest include hot flushes, which did not occur more frequently in the levormeloxifene than the placebo groups, but occurred significantly less frequently in the HRT group (P < 0.05). Breast tenderness was much more common in the HRT group (<0.001) than in all other groups. In conclusion, the study shows that levormeloxifene, a new selective estrogen receptor modulator, has positive effects on BMD and bone turnover and apparently strong estrogenic effects on the serum concentrations of different cholesterol subfractions. Levormeloxifene at the doses tested had an estrogen-like effect on endometrium and no effect on hot flushes. The study was unable to differentiate between the effects of the different doses of levormeloxifene.