Background: The new generation of dual energy X-ray absorptiometry (DXA) scanners provide visceral adipose tissue (VAT) estimates by applying different algorithms to the conventional DXA-derived fat parameters such as total fat, trunk fat and android fat for the same image data.
Objective: This cross-sectional study aimed to investigate whether VAT estimates from Hologic scanners are better predictors of VAT than conventional DXA parameters in pre-pubertal children, and to explore the discrimination ability of these VAT methods for cardiometabolic risks.
Methods: Healthy pre-pubertal children aged 7-10 years were recruited for basic anthropometric, DXA and magnetic resonance imaging (MRI) measurements. Laboratory tests included lipid profile, glycaemic tests and blood pressure.
Results: All VAT methods had acceptable to excellent performance for the diagnosis of dyslipidaemia (area under the curve [AUC] = 0.753-0.837) and hypertensive risk (AUC = 0.710-0.821) in boys, but suboptimal performance for these risks in girls, except for VAT by MRI in the diagnosis of dyslipidaemia. In both sexes, all VAT methods had no or poor discrimination ability for diabetes risk.
Conclusions: DXA-derived VAT estimates are very highly correlated with standard methods but has equivalent discrimination abilities compared to the existing DXA-derived fat estimates.