Bone turnover in elderly men: relationships to change in bone mineral density

BACKGROUND: It is not clear whether bone turnover markers can be used to make inference regarding changes in bone mineral density (BMD) in untreated healthy elderly men. The present study was designed to address three specific questions: (i) is there a relationship between bone turnover markers and femoral neck BMD within an individual; (ii) is there a relationship between baseline measurements of bone turnover markers and subsequent change in BMD; and (iii) is there a relationship between changes in bone turnover markers and changes in femoral neck BMD? METHODS: The present study was part of the on-going Dubbo Osteoporosis Epidemiology Study, which was designed as a prospective investigation. Men who had had at least 3 sequential visits with serum samples available during follow-up were selected from the study population. Serum C-terminal telopeptide of type I collagen (sICTP), N-terminal propeptide of type I collagen (sPINP) and femoral neck BMD were measured by competitive radioimmunoassays. Femoral neck bone mineral density (BMD) was measured by a densitometer (GE Lunar Corp, Madison, WI). Various mixed-effects models were used to assess the association between the markers and changes in BMD. RESULTS: One hundred and one men aged 70 ± 4.1 years (mean ± SD) met the criteria of selection for analysis. On average, sPINP decreased by 0.7% per year (p = 0.026), sICTP increased by 1.7% per year (p = 0.0002), and femoral neck BMD decreased by 0.4% per year (p < 0.01). Within-subject analysis indicated that changes in BMD were significantly associated with changes in sPINP (p = 0.022), but not with changes in sICTP (p = 0.84). However, neither baseline sPINP (p = 0.50) nor baseline sICTP (p = 0.63) was associated with subsequent changes in BMD. Moreover, changes in BMD were not significantly associated with previous changes in sPINP (p = 0.13) or sICTP (p = 0.95). CONCLUSION: These results suggest that in elderly men of Caucasian background, changes in sPINP were inversely related to changes in BMD within an individual. However, neither sPINP nor sICTP was sufficiently sensitive to predict the rate of change in BMD for a group of individuals or for an individual.