Raloxifene, a nonsteroidal selective estrogen receptor modulator (SERM), increases bone mineral density (BMD), decreases biochemical markers of bone turnover, and prevents incident vertebral fractures in postmenopausal women, while sparing the breast and endometrium from the undesirable stimulation caused by estrogen. How the long-term beneficial effects of raloxifene on bone turnover, as assessed by bone histomorphometry, compare with hormone replacement therapy (HRT) and placebo are not known. We studied 66 healthy postmenopausal women (age 55 to 75 years, mean 63 years) who were randomized to either raloxifene 150 mg/day, HRT (Premarin 0.625 mg/day, and Provera 2.5 mg/day), or placebo for 1 year. All women received 1-1.5 g of calcium/day. Following double tetracycline labeling, transiliac bone biopsies were obtained at baseline and 1 year and analyzed for changes in histologic indexes of bone remodeling on the cancellous surface as well as at the endocortical subdivision of the endosteal envelope, the location of the greatest fraction of postmenopausal bone loss. BMD and biochemical markers of bone turnover were also determined at baseline and 1 year. Four paired biopsies were obtained in the HRT group, six in the raloxifene group, and five in the placebo group. The frequency of remodeling events on cancellous bone and rate of bone formation in both cancellous and endocortical bone increased in the placebo group, while these measurements decreased in both drug treatment groups. Using analysis of mean percentage changes, when compared with the placebo group, these changes were significantly different for both raloxifene and HRT treatment groups (p<0.02). In all subjects, the bone was lamellar with discrete tetracycline labels and there was no evidence of marrow fibrosis or abnormal bone cells. BMD increased from baseline at the lumbar spine (p<0.05 in the HRT group) and in the total body (p<0.05 for both raloxifene and HRT). Compared with that of the raloxifene group, the increase in BMD was greater in the HRT group at the lumbar spine but not in the total body. Serum bone alkaline phosphatase, serum osteocalcin, and urine C-terminal cross-linking telopeptide of type I collagen significantly decreased (p<0.05) in both active treatment groups, changes significantly different from those seen with placebo. Overall, these results support the hypothesis that raloxifene preserves bone mass by reducing the elevated bone turnover found in postmenopausal women receiving placebo, by mechanisms similar to those operative in postmenopausal women receiving HRT.