A multimodal risk model integrating global longitudinal strain, BNP, and diastolic function for predicting remodeling and outcomes in severe asymptomatic aortic stenosis.

BACKGROUND AND AIMS The timing of aortic valve replacement (AVR) in severe asymptomatic aortic stenosis (AS) remains debated. Preserved ejection fraction (EF) may mask subclinical dysfunction, while global longitudinal strain (GLS), brain natriuretic peptide (BNP), and diastolic indices (E/E') provide complementary prognostic information. A predictive model for adverse outcomes after AVR integrating GLS, BNP, and E/E' has not been previously investigated. METHODS Ninety-six patients with severe asymptomatic AS and preserved EF (>50%) undergoing AVR were assessed at baseline and 1, 3, and 6 months. Echocardiography (GLS, EF, LVMI, IVSd, LVIDd, E/E'), BNP, and clinical outcomes were analyzed. Primary endpoint was LV remodeling; secondary endpoint was major adverse cardiovascular events (MACE). RESULTS Despite preserved EF, 76% had impaired GLS (<15%), and 64% remained in negative remodeling at 6 months. Baseline GLS ≤15% was the only independent predictor of adverse remodeling in multivariable logistic regression (OR 4.7 at 3 months; OR 3.5 at 6 months). For MACE, baseline E/E' >13 was the strongest independent predictor (OR 3.15, 95% CI 1.58-7.57, p = 0.004). The integrated GLS-BNP-E/E' model demonstrated superior predictive strength compared with individual parameters, with Nagelkerke R2 values of 0.41 for remodeling and 0.31 for MACE. CONCLUSION A multimodal risk model integrating GLS, BNP, and E/E' predicts adverse remodeling and MACE in severe asymptomatic AS. These findings highlight the complementary role of imaging and biomarkers in risk stratification before AVR-a concept that warrants confirmation in future multicenter studies.