Purpose: To verify whether estimation of bone mineral density (BMD) using digital X-ray radiogrammetry (DXR) is possible on children and to determine normative values of both such a DXR-BMD estimate and a corresponding metacarpal index (DXR-MCI) on.
Patients and methods: In retrospect, X-rays were selected of the hands of 200 healthy Caucasian children (120 boys and 80 girls, aged 4-18 yr). The involved children were selected among a larger group of children submitted to the surgical department of our institute for evaluation of a suspected fracture after an occurred trauma. All children with a verified fracture or a chronic bone-related disease, including bone age retardation or acceleration, were excluded from the study. Furthermore, only conventional X-rays with the same film and capture parameters were included. The images were scanned and analyzed using the Pronosco X-posure system V.2 (Sectra Pronosco, Denmark). DXR-BMD, DXR-MCI, and a porosity index (DXR-PI) were automatically calculated using the midshafts of the metacarpals II-IV. Mean values of DXR-BMD and DXR-MCI were calculated separately for girls and boys in 2-yr intervals.
Results: In the present study the system has been demonstrated to be capable of calculating DXR-BMD from conventional X-rays of the hand from children down to approx 6 yr of age. This ability depended somewhat on the diameter and the length of the involved metacarpals. The success rate was higher for large bones than for small bones. The system succeeded in analyzing the images of 110 boys and 68 girls. Values of DXR-BMD were observed to increase with age from 0.40 g/cm2 to 0.62 g/cm2 in the male group and from 0.39 g/cm2 to 0.54 g/cm2 in the female group. Girls aged 11-12 yr had a higher DXR-BMD than did boys, corresponding to the earlier entry to puberty of girls. Standard deviations (SDs) reached values of up to 0.05 g/cm2. DXR-MCI increased with age from 0.36 to 0.47 for boys and from 0.34 to 0.49 for girls with a maximum SD of 0.06. The correlation between DXR-BMD and age was r=0.83 and r=0.84 for boys and girls, respectively. The corresponding correlations for DXR-MCI was lower, with observed correlations of r=0.63 (boys) and r=0.68 (girls), respectively, with p<0.01 in all cases. The DXR-PI did not reveal a significant correlation to age (r=-0.31 and r=0.04. respectively) and showed SDs marginally higher than the calculated mean values.
Conclusion: The newly available DXR-methodology seems to offer the ability to determine DXR-BMD and DXR-MCI in children starting with a bone age of 6. This possibility may be of special relevance for children suffering from chronic bone diseases that require repeated X-rays of the hand (e.g. to determine bone age). The acquired normative data suggest that the measurements are of clinical value owing to low age-dependent variability (SDs) relative to an observed high increase with age. The clinical value of the porosity index (DXR-PI) remains uncertain and is limited owing to a high inter-individual variability.