Urinary hydroxypyridinium crosslinks of collagen in population-based screening for overt vertebral osteoporosis: Results of a pilot study



The urinary pyridinium crosslinks pyridinoline (PYD) and deoxypyridinoline (DPD) have been shown to provide valid indices of bone resorption. At present, both crosslink components are determined by reversed-phase HPLC, a time-consuming method precluding the use of these markers for routine purposes. Therefore, efforts have been made to develop simple immunoassays for the rapid measurement of urinary crosslinks, and their application to large-scale osteoporosis screening has been proposed. To evaluate the applicability and diagnostic validity of pyridinium crosslink measurements for screening purposes, urinary concentrations of total and free PYD and DPD were determined by HPLC and immunoassay technique (ELISA) in a sample of 269 individuals (male to female ratio = 130:139; age 50-81 years) recruited at random within a population survey of vertebral osteoporosis. On a molar basis, ELISA measures of crosslink-related epitopes were highly correlated with both total and free PYD and DPD as determined by HPLC (r > 0.82, p < 0.001). Age-specific means for creatinine-corrected total and free pyridinium crosslinks were significantly higher in females than in males (p < 0.001). In both sexes, neither age nor anthropometric variables (weight, height, and body mass index) showed a linear effect on the urinary crosslink/creatinine ratio. On average, 50% of the total amount of urinary crosslinks were present in free form. For both PYD and DPD, this proportion was significantly higher in women than in men (p < 0.05), but no change was observed with age or anthropometric measures. The excretion of pyridinium crosslinks was higher in osteoporotic (n = 18) than in nonosteoporotic individuals (n = 208) from the same population. However, this difference was statistically significant (p < 0.05) only for total PYD, total DPD, and free DPD as determined by HPLC, whereas no significant difference was seen with the ELISA. However, ROC analysis showed none of the crosslink measurements to provide sufficient diagnostic validity for screening purposes with regard to vertebral osteoporosis. Our results demonstrate that (1) the immunoassay technique provides measurements of pyridinium crosslinks that are highly correlated with HPLC findings; (2) beyond age 50, urinary concentrations of total and free pyridinium crosslinks are not linearly influenced by age or anthropometric factors but change with sex independently of osteoporotic status; and (3) although pyridinium crosslink excretion provides a valid and clinically useful measure of bone resorption, a single determination lacks the sensitivity necessary to screen successfully for vertebral osteoporotic fractures.