Improving risk stratification in TAVI: the role of epicardial fat, visceral fat and adiponectin

Abstract Background Transcatheter aortic valve implantation (TAVI) has become a standard treatment for patients with severe aortic stenosis, particularly those unsuitable for surgical valve replacement. However, the clinical outcomes following TAVI can vary significantly. Prior studies indicate that epicardial adipose tissue (EAT) and visceral adipose tissue (VAT) may affect cardiovascular outcomes. Moreover, adiponectin, an anti-inflammatory hormone produced by adipocytes, may also play a critical role in predicting post-TAVI outcomes. Purpose We aimed to determine how adipose tissue characteristics in combination with adiponectin levels influence clinical outcomes following TAVI. Methods This retrospective, single-centre, observational study included 226 patients with severe aortic stenosis who underwent TAVI. Using a custom artificial intelligence pipeline, pre-procedural CT scans were examined to quantify epicardial and visceral adipose tissue compartments. EAT was characterized by its average Hounsfield Units (EATHU), while VAT was measured by the average VAT area at the level of the L3 vertebra (VATL3A). Patients were then stratified into four groups based on the median EATHU and adiponectin levels: EATHU -82 HU and adiponectin ≥ 6540 ng/ml (n=56); EATHU -82 HU and adiponectin 6540 ng/ml (n=51); EATHU -82 HU and adiponectin 6540 ng/ml (n=60); EATHU -82 HU and adiponectin ≥ 6540 ng/ml (n=59). The primary outcomes were cardiovascular mortality, heart failure hospitalization, stroke and acute myocardial infarction within 12 months after TAVI. Results The study population’s average age was 82 ± 6.2 years, and 54% were female. Notably, the group with reduced EATHU and elevated adiponectin levels had the highest proportion of females (77%) compared to other groups (p0.001). Regardless of EATHU, patients with adiponectin levels below 6540 ng/ml had significantly higher VATL3A compared to those with ≥ 6540 ng/ml (p0.0001 for all). Over the 12-month follow-up, 31 patients experienced a primary endpoint event, with the highest event rate of 25% observed in the group with elevated EATHU and adiponectin levels (p=0.018). A logistic regression model incorporating EATHU, VATL3A and adiponectin improved mortality prediction, yielding an AUC of 0.78 (95% CI: 0.59–0.97, p0.0001), with a sensitivity of 83%, a specificity of 75%. For stroke and arrhythmia prediction, the model demonstrated an increased AUC of 0.73 (95% CI: 0.62–0.83, p0.0001) and 0.72 (95% CI: 0.57–0.86, p0.0001), respectively, with balanced sensitivity and specificity. Conclusions The combined assessment of EATHU, VATL3A, and adiponectin levels enhances the prediction of mortality, stroke, and arrhythmias after TAVI. Clinically, this suggests that evaluating these adipose tissue characteristics and adiponectin levels can aid in identifying patients at high risk for adverse events post-TAVI, enabling more tailored and proactive patient management.