To evaluate bone biochemical markers as predictors of the efficacy of a hormone replacement therapy (HRT), we studied the bone changes induced by the cessation and return of ovarian function in 28 patients treated for 6 months with a GnRH agonist. This model reproduced the effects observed in postmenopausal women with high bone turnover treated with HRT. At the end of the treatment, Z scores were 1.8 +/- 0.3 for Crosslaps (CTx) and deoxypyridinoline (D-Pyr), and 1.1 +/- 0.2 for bone alkaline phosphatase (B-ALP) and osteocalcin (OC). This indicated an imbalance in bone remodeling with a high bone resorption. Bone mineral density (BMD) fell by 4.2 +/- 2.5%. The changes in BMD between the 6th and 12th months were 0. 34 +/- 2.24 and -1.73 +/- 3.25% at the lumbar spine and the femoral neck, respectively. Biochemical markers except urinary calcium and hydroxyproline measured at 6 months were positively correlated with the BMD changes at the lumbar spine. After the resumption of menstruation, 13 of 28 women displayed positive spine BMD changes between the 6th and 12th months; in this group, bone biochemical markers measured at 6 months were significantly higher (P = 0.02). Stepwise regression analysis showed that the association of B-ALP and D-Pyr measured at 6 months explained 40% of BMD variance and the association of B-ALP, PTH, and estradiol 56%. We conclude that measuring individual biochemical bone markers can help to predict the bone effect of an increase in the circulating estradiol in women with ovarian deficiency.