Visceral fat mass prediction from spine DXA scans

Body mass index (BMI) does not account for weight distribution. Visceral fat mass obtained from total body dual-energy x-ray absorptiometry (DXA) is a stronger predictor of cardiovascular risk than BMI. Sagittal abdominal diameter, the distance from the small of the back to the anterior surface of the abdomen, can be accurately measured from lumbar spine DXA scans. We hypothesized that sagittal abdominal diameter from spine DXA, alone and in conjunction with other measures, can be used to predict visceral fat mass from total body DXA.

The study utilized data from the Manitoba Bone Mineral Density Program registry, which includes all DXA scans performed in Manitoba, Canada. We identified 1,981 adults aged 18 years and older individuals undergoing both total body DXA and lumbar spine DXA at the same visit between April 1999 and April 2025. The population was randomly divided into a derivation subgroup (n = 1321) and a validation subgroup (n = 660). Regression models were developed to predict visceral fat mass (grams) and visceral fat mass index (kg/m²). Model 1 used sagittal abdominal diameter as the sole covariate. Model 2 additionally included age, sex and spine fat fraction Model 3 additionally included squared terms for sagittal abdominal diameter and spine fat fraction.

Sagittal abdominal diameter demonstrated high Pearson correlations with both visceral fat mass (r = 0.864) and visceral fat mass index (r = 0.858). Model 3 showed the high predictions for visceral fat mass (derivation subgroup r² = 0.837, validation subgroup r² 0.852) and visceral fat mass index (derivation subgroup r² = 0.854, validation subgroup r² 0.862).

Sagittal abdominal diameter derived from DXA, alone or in conjunction with other measures, can accurately predict visceral fat mass. This may allow for objective assessment of abdominal obesity in the many individuals who undergo routine DXA scans.