Comparison of DXA- and CT-based measurements of abdominal visceral adipose tissue in premenopausal women.

OBJECTIVE

To compare dual-energy X-ray absorptiometry (DXA) and computed tomography (CT) for estimating visceral adipose tissue (VAT) area in premenopausal women, and to evaluate whether DXA can track VAT changes over time.

METHODS

VAT area (cm²) was measured using DXA and CT on the same day in 66 premenopausal women before (baseline) and after a 20-week prospective intervention trial; the current study is a secondary analysis of these data. Correlation, Bland-Altman, and regression analyses were performed. A regression-based adjustment equation was developed to align DXA with CT values and applied to baseline and follow-up data, with robustness assessed using bootstrap resampling. Analyses were stratified by body mass index (BMI: normal, overweight, obese).

RESULTS

At baseline, DXA significantly overestimated VAT compared with CT (DXA: 98.7 ± 49.5 cm² vs CT: 67.7 ± 33.1 cm²; mean difference +31.0 ± 7.3 cm², p ≤ 0.001), though measures were strongly correlated (r² = 0.87, p ≤ 0.001). Overestimation increased with higher BMI. However, DXA and CT measures of VAT change over 20 weeks were not correlated (r² = 0.00, p = 0.834). A regression-based adjustment equation reduced baseline bias (+31.0 cm² to +4.0 ± 5.8 cm² difference) but did not improve agreement for longitudinal changes.

CONCLUSIONS

DXA provides a correlated but biased estimate of VAT area at a single time point compared with CT, with greater overestimation at higher BMI. A regression-based adjustment equation improves baseline agreement but does not enhance the detection of VAT change over time. DXA should be used cautiously for monitoring longitudinal VAT changes in women.