Several bisphosphonates are under investigation for the treatment and prevention of postmenopausal osteoporosis. Alendronate, one of these compounds, has been shown to inhibit bone turnover and induce substantial increases in bone mass, but little is known about the duration of its effects. This is considered important, keeping in mind the long half-life of bisphosphonate in bone. In this double-blind controlled study, two groups of 15 postmenopausal women with spinal bone mineral density (BMD) > 2 SD below adult mean peak without vertebral fractures were randomized to receive either alendronate, 20 mg/day, or placebo for 6 months. The treatment course with alendronate significantly suppressed all indices of bone turnover (hydrox-yproline, collagen crosslinks, and alkaline phosphatase activity) within 3 months, and a further slight decrease was observed in the subsequent 3 months. After treatment withdrawal, all indices of bone turnover slowly increased, and they attained the pretreatment values within 6–9 months. Lumbar spine BMD rose by 3.7% (± 1.7 SD) after 6 months of alendronate therapy but did not change 6 and 12 months after treatment withdrawal (4.6 ± 2.8 and 4.7 ± 2.6% versus baseline, respectively). In control patients a slow decrease in lumbar spine BMD was observed, but this was significant only at month 18 of the study. Femoral BMD did not significantly change in the alendronate group, but it slowly decreased in the control group at all sites of evaluation. The fractional loss became statistically significant versus both baseline and the active group by the end of the study only at the level of the femoral neck. These results may imply that the positive bone balance induced by alendronate is sustained long after stopping therapy, despite the reestablishment of pretreatment bone turnover.