Validity of visceral adiposity estimates from DXA against MRI in Kuwaiti men and women

Abstract

Objectives - The prevalence of obesity and diabetes in the Middle East is amongst the highest in the world. Valid measures of abdominal adiposity are essential to understanding the metabolic consequences of obesity. Dual-energy X-ray absorptiometry (DXA) is increasingly being utilised to assess body composition in population studies, and has recently been used to estimate visceral adipose tissue (VAT). The aim of this study was to determine the accuracy of DXA-derived VAT in a Middle Eastern population using magnetic resonance imaging (MRI) as the criterion measure. ${\bf Method:}$ VAT was estimated from abdominal DXA measures in 237 adult men (n=130) and women (n=107), aged 18-65 years, participating in the Kuwait Wellbeing Study. These estimates were compared to MRI measures of the corresponding anatomical region. The agreement between methods was assessed using Bland-Altman as well as correlation analysis. ${\bf Results:}$ Median MRI-VAT was 1148.5 (95% CI: 594.2-1734.6) cm$^3$ in men and 711.3 (95% CI: 395.5-1042.8) cm$^3$ in women. DXA estimates of VAT showed high correlations with corresponding MRI measures (r= 0.94 (p<0.0001) in men; r= 0.93 (p<0.0001) in women). DXA overestimated VAT with a mean bias (95% limits of agreement) of 79.7 (-767; 963) cm$^3$ in men and 46.8 (-482; 866) cm$^3$ in women. The imprecision of DXA increased with increasing VAT adiposity in both men and women. ${\bf Conclusion:}$ DXA estimates of VAT are valid for use in Middle Eastern populations, although accuracy decreases with increasing level of obesity.