Objectives: i) To assess bone (BMC) by BIA in apparently healthy Indian children. ii) To generate percentile curves for BIA measured BMC by age/gender. iii) To investigate relationship between BMC measured by BIA and DXA in separate subset.
Methods: In a multicentre study, 4154 children (2298 boys), 518 years underwent BMC assessment by BIA. BMC for age percentiles were computed using LMS method. In 41 children (not from the multicentre study) total body BMC was measured by DXA and BIA. Z-scores for BMC for age by DXA and BIA (derived from the present study) were computed. Pearsons correlation and Bland-Altman limits of agreement were derived between BMC by BIA and DXA. Receiver operating curve (ROC) analysis was performed for cut-off point for BIA Z-score corresponding to Z-score of −2 by DXA to discriminate between those with or without risk of low bone mass.
Results: Mean age of children was 10.7±3.3; height, weight and BMI for age Z-scores were close to zero for all ages. Correlation coefficient between BMC by BIA and DXA was 0.83. Bland-Altman limits of agreement indicated that a BIA reading would be −1.52 below or 1.3 above the DXA, thus ROC analysis was performed. Cut-off yielding the maximal sensitivity and specificity for predicting low bone mass by BIA was Z-score of −1.3 (10th percentile) (sensitivity=94.1%, specificity=40%, AUC 0.85) (95% CI: 0.6940.941).
Conclusions: A Z-score of −1.3 by BIA corresponds to −2 of DXA which may be a useful tool for screening for the risk of low bone status. Further studies are required to confirm our results.
22 Jun 2013 - 25 Jun 2013