PURPOSE: Previously we have reported a significant increase in bone mineral density (BMD) of the spine and the hip and reductions in biochemical indices of bone turnover in postmenopausal women with osteoporosis treated with alendronate at various doses over 1 to 2 years. We have followed BMD and biochemical parameters in these patients for 1 or 2 years after discontinuation of alendronate to determine resolution of alendronate effects. PATIENTS AND METHODS: Participants received daily oral doses of placebo, 5 or 10 mg of alendronate for 2 years, or 20 or 40 mg of alendronate for I year followed by I year of placebo. No treatment was given in the third year of study. RESULTS: Lumbar spine BMD changes in the 5-arid 10-mg groups (-1.4 and -0.4%) were similar to those in the placebo group (-1.2%) 1 year after discontinuation of drug and lumbar spine BMD changes in the 20- and 40- mg groups (-1.2% and 0.8%) were similar to those in the placebo group (- 0.9%) 2 years after discontinuation of drug. BMD of the total hip followed the same pattern of resolution. The difference in BMD between alendronate and placebo groups at the end of alendronate treatment was maintained up to 2 years. Residual reductions in the bone resorption markers urinary deoxypyridinoline (D-Pyr) and collagen type 1 cross-linked N telopeptides and the bone formation markers serum bone-specific alkaline phosphatase and osteocalcin remained for I year after discontinuation of 5 and 10 mg of alendronate and for 2 years after discontinuation of 20 and 40 mg of alendronate, other than return of D-Pyr to baseline 1 year after cessation of treatment with the 5- and 10-mg doses. CONCLUSIONS: A residual decrease in bone turnover may be found up to 2 years after discontinuation of alendronate. Accelerated bone loss is not observed when treatment is discontinued. However, continuous therapy with alendronate is required to achieve a continuous gain in BMD.