VAT was associated with higher cardiovascular risk and improved risk reclassification when added to an established prediction model. In a population-based analysis of UK Biobank participants published in Progress in Cardiovascular Diseases, MRI-derived VAT enhanced reclassification for ASCVD, HF, and total CVD outcomes despite no significant improvement in model discrimination.

The analysis included 38373 adults without known CVD at baseline who underwent abdominal MRI. Median VAT volume was 3.58 L. Using this threshold, participants with higher VAT showed increased risk of ASCVD (hazard ratio [HR] 1.32; 95 % confidence interval [CI] 1.15-1.51), HF (HR 1.55; 95 % CI 1.27-1.89), and total CVD (HR 1.38; 95 % CI 1.23-1.55) after adjustment for age and sex.

Adding VAT to the PREVENT model did not significantly change discrimination for ASCVD (C-statistic 0.731 vs 0.729; P = 0.85), nor for HF or total CVD. However, VAT significantly improved reclassification, with an NRI of 0.37 (95 % CI 0.30-0.33) for ASCVD, 0.48 (95 % CI 0.35-0.61) for HF, and 0.37 (95 % CI 0.28-0.46) for total CVD.

These findings indicate that VAT captures cardiovascular risk not fully reflected in traditional prediction models. Incorporating VAT improved individualized risk classification, particularly for HF, without altering overall discrimination performance.