Does serum cholesterol contribute to vertebral bone loss in postmenopausal women?

Recent in vitro and animal studies suggest that cholesterol and its metabolites inhibit the functional activity of osteoblasts and thereby induce reduced bone mineralization. However, scant information is available on the clinical implication of these findings with special regard to postmenopausal bone loss. Therefore, the aim of the present study was to investigate cross-sectional and longitudinal associations between serum cholesterol, bone mineral density (BMD), and bone turnover in 340 postmenopausal women aged 50-75 years (mean 59 years), who were followed for 8.3 +/- 1.1 years. BMD in the lumbar spine, distal forearm, and total hip was measured by dual energy X-ray absorptiometry. Other study variables were physical measures, serum cholesterol, serum markers of bone turnover, and self-reported information on various risk factors for osteoporosis. At baseline, serum cholesterol showed significant negative correlation with BMD at the lumbar spine (r = -0.21, P < 0.0001) and distal forearm (r = -0.14, P = 0.013), but not at the hip. No associations of serum cholesterol with serum osteocalcin (r = 0.054, P = 0.317) and CTX (r = -0.027, P = 0.623) were, however, noted. After adjustment for age and BMI, the negative correlation remained significant at the lumbar spine (r = -0.16, P = 0.004), but not at the distal forearm (r = -0.018, P = 0.738). At the end of the 8-year follow-up, the correlation between serum cholesterol and spine BMD was not observed. Those with the largest increases in serum cholesterol, however, showed the greatest decreases in spine BMD independently of the changes in BMI (r = -0.16, P = 0.004). The correlation between the changes in serum cholesterol and the simultaneous changes in osteocalcin (r = 0.081, P = 0.140) and CTX (r = 0.042, P = 0.441) were statistically insignificant. Thus, our results suggest that the weak associations between spine BMD and serum cholesterol can be explained by the fact that both variables are simultaneously affected by estrogen deficiency rather than by a direct influence of serum cholesterol on osteoblast function.