Morphology and body composition

2018 The Journal of Physical Fitness and Sports Medicine  
Aims] The purpose of this study was to validate the newly developed visceral fat area (VF) measurement from dual-energy X-ray absorptiometry (DXA) on Japanese adults by comparing to a reference method, magnetic resonance imaging (MRI). [Methods] The study population included 391 adults (266 males and 125 females) with a BMI of 18.6 to 36.7 kg/m 2 , aged 40 years and over, who participated in WASEDA'S Health Study. VF (cm 2 ) was measured with both DXA (Horizon A, Hologic) and MRI (SIGNA 1.5T,
more » ... MRI (SIGNA 1.5T, GE). [Results] DXA_VF was strongly correlated to MRI_VF (all participants: r = 0.896, standard error of the estimate (SEE) = 17.9 cm 2 , males: r = 0.867, SEE = 17.3 cm 2 , females: r = 0.909, SEE = 12.5 cm 2 ). The Bland-Altman plot showed a VF bias (MRI as standard method) of 4.1 cm 2 (r = 0.356, p < 0.001) for males and -3.8 cm 2 (r = 0.124, p = 0.170) for females. Combined, the bias was 1.6 cm 2 (r = 0.124, p = 0.014) with 95% limits of agreement of -35.7 to 38.8 cm 2 . Regression of the differences on the average values for males yielded a statistically significant slope of -0.201 and intercept of 22.6 cm 2 (p < 0.001, respectively). [Conclusion] For Japanese adults, VF measured with DXA is highly correlated to VF measured with MRI. The current DXA method moderately overestimates the VF for males and has systematic error compared with MRI method.
doi:10.7600/jpfsm.7.390 fatcat:5adfs6whq5b75fmrr5jbt272dm