Imaging and biomarker-based risk stratification in TAVI: the role of epicardial fat, visceral fat, and adiponectin

Aims: Transcatheter aortic valve implantation (TAVI) has become a standard treatment for severe aortic stenosis, yet clinical outcomes vary significantly. Epicardial adipose tissue (EAT), visceral adipose tissue (VAT), and circulating adiponectin may contribute to this heterogeneity. This study aimed to determine how adipose tissue characteristics, in combination with adiponectin levels, affect clinical outcomes following TAVI. Methods and results: This single-centre retrospective study included 243 patients with severe aortic stenosis who underwent TAVI. Pre-procedural computed tomography scans were processed using artificial intelligence to quantify EAT density (EATHU) and VAT area at the level of the L3 vertebra (VATL3A). Patients were stratified into groups based on median adiponectin (6.54 µg/mL), VATL3A (135 mm), EATHU (−82.0 HU), and EATVOL (31 mL). The primary endpoint was 1-year major adverse cardiac and cerebrovascular events (MACCE), defined as all-cause mortality, heart failure hospitalization, or stroke. Over a mean follow-up of 344 ± 77 days, there were 25 deaths (10%), 12 heart failure hospitalizations (5%), and 7 strokes (3%). Circulating adiponectin (adjusted HR 1.34 per 5 μg/mL, 95% CI 1.05–1.70; P = 0.017) and VATL3A (adjusted HR 0.63 per 100 mm2, 95% CI 0.44–0.91; P = 0.014) independently associated with MACCE, EATHU, and EATVOL did not. All three combined adipose–adiponectin phenotypes demonstrated significant graded risk (P-for-trend 0.003–0.048), with the high EATVOL/high adiponectin phenotype showing the strongest association (adjusted HR 4.25, 95% CI 1.80–10.07). Conclusion: Circulating adiponectin and VATL3A independently predict 1-year MACCE after TAVI. EAT parameters become prognostically informative only when combined with adiponectin.