Searchable abstracts of presentations at key conferences on calcified tissues
Bone Abstracts (2013) 2 OC14 | DOI: 10.1530/boneabs.2.OC14

ICCBH2013 Oral Communications Diagnostics (6 abstracts)

What DXA measurement sites are best for bone health assessment in children? Effect of inter-machine differences on bone outcomes from the Bone Mineral Density in Childhood Study

Babette Zemel 1, , Heidi Kalkwarf 3 , Mary Leonard 1, , Vicente Gilsanz 4 , Joan Lappe 5 , Justine Shults 1, , John Shepherd 7 , Sharon Oberfield 6 & Karen Winer 8

1Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA; 2University of Pennsylvania, Philadelphia, Pennsylvania, USA; 3Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio, USA; 4Children’s Hospital, Los Angeles, California, USA; 5Creighton University, Omaha, Nebraska, USA; 6Columbia University Medical Center, New York, New York, USA; 7University of California, San Francisco, California, USA; 8National Institute of Child Health and Human Development, Bethesda, Maryland, USA.

Objective: Total body less head (TBLH) and spine are the recommended DXA sites for bone health assessment in children and adolescents. However, inter-machine differences will affect use and interpretation of results in clinical care and research applications. We examined BMC and areal-BMD (aBMD) at 4 skeletal sites among healthy children to identify the magnitude of inter-machine differences in Z-scores.

Methods: BMDCS evaluated healthy participants, ages 5–20 years, as previously described. All five centers used Hologic devices and scans were analyzed centrally. Data from the first study visit for subjects with complete data for the distal 1/3 radius, hip, spine and and TBLH, and covariates (demographic information, height and BMI Z-score, race, puberty stage, calcium intake and physical activity) were included in the analysis. BMDCS Z-scores for BMC and areal-BMD were calculated, and differences among study centers assessed by ANCOVA adjusting for covariates. Logistic regression assessed the probability of having a BMC or aBMD Z-score ≤−1.5 among centers after adjusting for covariates.

Results: 1889 subjects (48% females, 24% African American, 17% Hispanic, 47% non-Hispanic white, 12% other) were evaluated. There were significant differences in Z-scores adjusted for covariates among study centers for all measures except total hip aBMD Z-score. Differences from the group mean in adjusted Z-scores were lowest for spine and hip measures (−0.14 to 0.14 SDS), and highest for TBLH aBMD (−0.19 to 0.53 SDS). After adjusting for covariates, the probability of having a Z-score ≤−1.5 (expected probability .067) was not significantly different between centers for most skeletal sites, except for TBLH aBMD (probabilities ranged from 0.02 (95% CI: 0.1–0.04) to 0.09 (95% CI: 0.07–0.15)).

Conclusions: Spine aBMD and hip measurements showed relatively good agreement between centers. There was wide variation in DXA whole body scan results obtained on healthy children measured on comparable Hologic DXA devices. This can potentially result in misdiagnosis of children with low bone status in clinical care and research. The recommendation of whole body scans as an optimal measurement site in children should be reconsidered.

Support: R01 HD058886, U54 RR024134, N01-HD-1-3331.

Volume 2

6th International Conference on Children's Bone Health

Rotterdam, The Netherlands
22 Jun 2013 - 25 Jun 2013


Browse other volumes

Article tools

My recent searches

No recent searches.