Biochemical markers of bone turnover may correlate with rates of bone loss in a group of postmenopausal women, but it is uncertain how useful they are in predicting rates of bone loss in the individual. The aim of this study was to determine the value of measurements of biochemical markers for the prediction of rates of bone loss in the individual. We studied 60 postmenopausal women (ages, 49-62 years), 43 of whom had gone through a natural menopause 1-20 years previously and 17 of whom had undergone hysterectomy 3-22 years ago. Lumbar spine bone mineral density (BMD) was measured using dual-energy X-ray absorptiometry (DXA) over 2-4 years. Bone formation markers (bone-specific alkaline phosphatase [ibAP] and amino terminal of type I collagen [PINP] and osteocalcin [OC]) were measured in serum. Bone resorption markers (N-telopeptide of type 1 collagen [NTx] and immunoreactive free deoxypyridinoline [iFDpd]) were measured in urine and corrected for creatinine (Cr). Rates of bone loss were calculated as percent change per year. We found significant negative correlations (Spearman rank) between all measured biochemical markers and rate of change in bone density with r values ranging from -0.35 to -0.52. When markers and rates of bone loss were divided into tertiles, prediction of bone loss in an individual was poor (kappa < 0.2). There was an exponential relationship between rate of bone loss and years since menopause (YSM) in the 43 women having a natural menopause (r2 = 0.44; p = 0.008) indicating higher rates of loss in the early postmenopausal period. Levels of NTx, iFDpd, and PINP also showed a significant negative correlation with YSM. We conclude that there is a strong relationship between rates of spinal bone loss and levels of bone turnover markers. Although this is a small study, the results also suggest that using DXA measurements of the lumbar spine as the "gold standard," it is not possible to use biochemical markers to predict rate of bone loss in the individual.