Validation of longitudinal DXA changes in body composition from pre- to mid-adolescence using MRI as reference

Abstract

Dual-energy X-ray absorptiometry (DXA) has been used extensively for bone mineral density and body composition assessments. Surprisingly, the role of DXA in monitoring changes in children's body composition, using direct imaging methods such as magnetic resonance imaging (MRI) as reference, is still yet to be validated. We aimed at validating the use of DXA in monitoring change in the thigh lean soft tissue mass (LSTM) and fat mass (FM) when compared with thigh skeletal muscle mass (SM) and FM, measured using MRI as the reference standard, from childhood to midadolescence. At baseline, 22 healthy children (16 boys and 6 girls) aged 8-11yr were included, and then recalled at pubertal stage Tanner2-Tanner4. LSTM-DXA and FM-DXA of the mid-third femur and SM-MRI and FM-MRI of the same region were measured on the same day. The same protocol was repeated 26-48mo later. At baseline, DXA overestimated LSTM-DXA on average by 222g (95% confidence interval [CI]: 33-410g) with a concordance C-LSTM=0.576. FM-MRI and FM-DXA were not significantly different (95% CI=213 to 199g, the C-FM=0.907). At follow-up, change in LSTM-DXA and FM-DXA were not significantly different to change in SM-MRI and FM-MRI, respectively (95% CI of the difference was -278 to 208g for LSTM, and -148 to 236g for FM). The coefficient of concordance between the 2 techniques was 0.88 for both LSTM and FM. This study validates the use of DXA in monitoring changes in LSTM and FM in children, confirming its significant potential in clinical and research roles in pediatric body composition.