Summary The aim of this study was to determine the correlation between changes in collagen metabolites (ICTP, mature cross-linked carboxyterminal telopeptide of type collagen; PINP, the amino-terminal propeptide of type procollagen) and bone mineral density (BMD) in 206 pre-and post-menopausal breast cancer patients with non-metastatic disease. All patients received adjuvant cancer treatmentpremenopausal patients chemotherapy and post-menopausal patients anti-oestrogens. In addition, the patients were also randomized to receive oral clodronate 1600 mg daily for 3 years. BMD was measured at baseline and at 1 and 2 years, the collagen metabolites at baseline and at 1 year. There was a highly significant negative correlation between the changes in PINP and BMD in lumbar spine and femoral neck from baseline to 12 months in all patients (r, = -0.68, P < 0.0001, and -0.45, P < 0.0001, respectively), and in pre-and post-menopausal patients separately.The changes in PINP levels at 12 months predict further changes in BMD at 24 months (rs = -0.70, P < 0.0001, and -0.51, P < 0.0001, respectively). ICTP and BMD changes correlated significantly only in lumbar spine of premenopausal patients who developed rapid bone loss due to chemotherapy-induced amenorrhoea (rs = -0.34, P = 0.0003). The PINP levels at 12 months were significantly lower in the clodronate group than in the control group (P < 0.0001). Our results indicate that PINP is a sensitive marker of bone turnover rate. Changes in PINP levels significantly predicted changes in BMD and correlated with the antiresorptive efficacy of clodronate treatment.Keywords: adjuvant chemotherapy; breast neoplasm; collagen metabolites; post-menopausal osteoporosis Post-menopausal osteoporosis is a common disorder. After menopause, bone turnover rate increases rapidly as a result of oestrogen deficiency. There is an imbalance between resorption and formation, the resorption exceeding the formation, with accelerated bone loss as a result (Parfitt, 1979). Early menopause, low bone mass at menopause and fast rate of bone loss after menopause are the risk factors of osteoporosis. BMD is the most accurate way of measuring bone mass and diagnosing osteoporosis. However, the rate of bone loss after menopause varies significantly from one woman to another and can not be predicted by a single BMD measurement Hansen et al, 1991). Serial BMD measurements are needed, but because of the relatively small changes in bone mass per year in comparison with the precision of the measurement methods, it may take a long time to predict the rate of bone loss with BMD measurements (Riggs et al, 1986;Hansen et al, 1990;Pouilles et al, 1993Pouilles et al, , 1995.An intriguing possibility would therefore be to use biochemical markers of bone turnover as indicators of the rate of bone loss. Both resorption (urinary excretions of hydroxyproline, pyridinoline cross-links of collagen, and cross-linked telopeptides of type I