Comparison between handheld ultrasound and regional and whole-body dual energy x-ray absorptiometry (DXA) for body fat assessment.

Abstract

To determine the extent of agreement between a handheld ultrasound (US) attached to an android tablet and the reference method dual energy x-ray absorptiometry (DXA) for the measurement of adiposity. A whole-body DXA scan and abdominal adipose tissue thickness measurements using a handheld US were obtained from 104 adults (63 females, 41 males). Body fat percent (BF%), total fat mass (kg), and trunk fat mass (kg) were obtained from DXA. Subcutaneous adipose tissue (SAT), superficial subcutaneous adipose tissue (SSAT), and deep subcutaneous adipose tissue (DSAT) thickness were obtained from US. Sex-specific total fat mass, trunk fat mass, and BF% estimates by US were compared with DXA. Spearman's correlations and Bland-Altman plots were used to assess agreement between the methods. US SAT correlated strongly with total fat mass for both females (rs=0.74) and males (rs=0.87) as did trunk fat mass (females, rs=0.81; males, rs=0.83); as did SSAT and DSAT (females: rs=0.65 and rs=0.66; males: rs=0.63 and rs=0.85, respectively, all p-values < 0.0001). Bland-Altman plots demonstrated strong agreement for total and trunk fat mass for both males and females. For BF%, acceptable limits of agreement were observed for males but not for females, substantial proportional bias as indicated by a negative slope was noted for BF% using SAT (r=-0.298, p=0.0177). The handheld US and technique to analyze abdominal adipose tissue thickness showed strong agreement with DXA results and generated highly comparable estimates for total and trunk fat mass for both sexes.