Ilich J Z, Hangartner T N, Skugor M, Roche A F, Goel P K, Matkovic V
Bone and Mineral Metabolism Laboratory, Davis Medical Research Center, Ohio State University, Columbus 43210, USA.
Skeletal Radiol. 1996 Jul;25(5):431-9. doi: 10.1007/s002560050111.
To evaluate the association between chronological age, skeletal age, pubertal stage, and basic anthropometry with bone mass of the total body, forearm, and second metacarpal bone in 456 healthy Caucasian females, aged 8-13 years.
Total body and forearm bone measurements were performed by dual X-ray absorptiometry, while bone mass of the second metacarpal was assessed by radiogrammetry. Skeletal age (SA) was assessed by the FELS method and pubertal stage was self-determined by selecting corresponding illustrations of breast and public hair development. The Cp criterion was used to select the best multiple regression model containing the subset of independent variables with the least bias and best predictive ability for each of the measured bone mass variables.
Of all the independent variables, weight, stature, and SA emerged as the most significant predictors for almost all the bone mass variables. Multiple regression models were created based on the Cp criterion with the resulting R2 (adjusted) for bone mineral content of total body, proximal forearm, ultradistal forearm, length of second metacarpal, as well as of total, medullary, and cortical areas: 0.793, 0.523, 0.390, 0.602, 0.232, 0.073, and 0.264, respectively. The measured bone variables were also regressed on SA using either quadratic or linear equations, depending on the shape of the cubic splines used for the best curve fitting. Significant positive association (p < 0.0001) of SA and each of the bone variables was noted, the highest being with bone mineral density and content of total body (R2 = 0.176, 0.338) and proximal and ultradistal forearm (R2 = 0.216, 0.203, 0.106, 0.201), respectively, as well as with the length of the second metacarpal bone (R2 = 0.339). Chronological age and pubertal stage did not have statistically significant predictive abilities for bone mass variables in the multiple regression models.
We conclude that skeletal age is a powerful determinant of bone mass in children. It can be used as the criterion for the selection of a biologically homogeneous population with regard to bone mass. This may be important for the design of intervention studies targeting bone mass of children and adolescents.
