The CT-derived fat-free muscle fraction is associated with inflammation and predicts outcome in patients with aortic stenosis undergoing TAVR

Abstract Background The ESC guidelines underscore the importance of objective risk stratification in making patient-centered decision for valve interventions. However, they primarily rely on surgical risk scores, not considering essential factors such as frailty. Computed tomography (CT) is a routine part of the pre-interventional workup in patients undergoing TAVR and allows for body composition analysis, including the assessment of fat-free muscle fraction (FFMF) as an indicator of muscle quality. Purpose This study aimed to assess the correlation between FFMF and both inflammation and endocrine disorders, and to investigate the predictive value of FFMF in patients with aortic stenosis undergoing TAVR. Methods The study included 789 patients undergoing TAVR between 2017 and 2019. Evaluations involved CT scans to assess body composition. Using densitometric thresholds, the skeletal muscle area at L3/L4 level was separated into fatty and lean muscle to calculate the FFMF. The cohort was divided into tertiles of high, medium, and low FFMF, Figure 1. Results The study population was 45.5% female with a mean age of 81.2±6.0 years. Overall, 261 (33.1%) patients had a high FFMF, 293 (37.1%) had a medium FFMF, and 235 (29.8%) had a low FFMF. Patients with a low FFMF were older (82.3±5.9 vs 81.6±5.6 vs 79.8±6.2 years, p<0.01) and had higher rates of atrial fibrillation (52 vs 43 vs 37%, p<0.01), compared to those with a medium and high FFMF. Regarding body composition, low FFMF patients had a higher percentage of body fat (28.7±11.2 vs 26.8±9.1 vs 24.2±9.6%, p=0.01) and a lower percentage of body muscle (30.9±5.3 vs 31.3±4.2 vs 32.7±4.4%, p=0.03). Moreover, a low FFMF was associated with inflammation, as indicated by increased levels of C-reactive protein (6.3 vs 4.2 vs 3.1 mg/l, p<0.01) and procalcitonin (0.06 vs 0.05 vs 0.05 µg/l, p=0.01). Endocrinological assessment revealed lower levels of estradiol (6.8 vs 14.7 vs 15.0 pg/ml, p=0.02) and testosterone (0.18 vs 0.48 vs 2.35 ng/ml, p<0.01) as well as higher levels of cortisol (8.9 vs 7.7 vs 7.6μg/dl, p=0.05) in low FFMF patients. Regarding clinical outcomes, low FFMF was associated with increased one-year mortality (23.4%) compared to medium (11.6%) or low FFMF (3.8%, p<0.01), Figure 2A. This difference persisted over five years (p<0.01), Figure 2B. ROC curve analysis showed that FFMF had the strongest association with one-year mortality (AUC 0.75, p<0.01) and was superior to the STS-PROM (AUC 0.65) and EuroSCORE II (AUC 0.64). Multivariate analysis revealed that low FFMF was independently associated with mortality following TAVR (OR: 3.19 [CI: 1.64 – 6.22], p<0.01). Conclusion The CT-derived FFMF is associated with inflammation and endocrinological disorder in patients undergoing TAVR. Importantly, a low FFMF is a strong and independent predictor of dismal outcomes. FFMF assessment can seamlessly integrate into routine pre-interventional CTs, offering a promising objective parameter for outcome prediction.