Collagen markers deoxypyridinoline and hydroxylysine glycosides: pediatric reference data and use for growth prediction in growth hormone-deficient children.

BACKGROUND In children and adolescents, markers of bone and collagen metabolism reflect the dynamics of skeletal growth and development. The aim of this study was to assess the relationship of the urinary collagen markers deoxypyridinoline (DPD) and hydroxylysine (Hyl) and its glycosides [galactosyl-Hyl (Gal-Hyl) and glucosyl-Gal-Hyl] with growth. METHODS Urine samples from 240 apparently healthy children and adolescents (6-19 years; 124 girls) and from 51 prepubertal children with growth hormone (GH) deficiency (3-14 years; 14 girls) were analyzed. Urinary Hyl and its glycosides were quantified by HPLC, and DPD was assessed by chemiluminescence assay. Urinary concentrations of all markers were related to urinary creatinine. RESULTS Multiple regression analysis revealed that only age and height velocity were independently associated with these markers in healthy children. In GH-deficient patients, the urinary excretion of both analytes after 4 weeks of GH therapy correlated significantly with the height increase during the first treatment year (r = 0.79 for Gal-Hyl; r = 0.70 for DPD; P <0.001 each). In a multivariate linear regression model using Gal-Hyl concentrations at 4 weeks, baseline concentrations of insulin-like growth factor 1 and height velocity after 3 months accounted for 80% of the variability in height gain during the first treatment year. A model using DPD concentrations at 4 weeks, in place of Gal-Hyl concentrations, as well as baseline concentrations of insulin-like growth factor 1 and height velocity after 3 months accounted for 83% of the variability. CONCLUSIONS These urinary bone and collagen markers give some early indication of growth response, but the prediction of an individual marker is too imprecise to serve as a basis for clinical decisions. Markers of bone and collagen metabolism might be more useful as components of multivariate growth prediction models.