Subclinical systolic dysfunction in aortic stenosis

Abstract

Aims: Aortic stenosis (AS) imposes chronic pressure overload on the left ventricle (LV), leading to myocardial remodeling and eventual systolic dysfunction. This study aims to evaluate longitudinal patterns of LV systolic function, identify early predictors, and assess prognostic implications. Methods and Results: We retrospectively analyzed 602 patients (mean age 77 years, 54% female) with severe AS who had previous serial echocardiograms available (median 3 per patient over 4 years). Patients were categorized as low fraction (LF, Left Ventricular Ejection Fraction (LVEF) <50%, n=124, 21%) or normal fraction (NF, LVEF ≥50%, n=478, 79%) at severe AS diagnosis. Demographics, comorbidities, LV morphology, and functional measurements derived from deep learning-based strain (DLS) analysis (using EchoNet-Dynamic) were compared between groups. LVEF deterioration preceded severe AS diagnosis, with abrupt decline in the preceding 2 years in LF patients, while remaining >58% in NF patients. LVEF and DLS measurements emerged as the most significant predictors of subsequent LVEF deterioration 4 years before AS became severe. A decision tree model identified three distinct echocardiographic risk profiles with 89.5% internal validation accuracy: high risk (LVEF <56%), intermediate risk (LVEF ≥56% and |DLS| <16%), and low risk (LVEF ≥56% and |DLS| ≥16%). These profiles independently predicted distinct risks of developing depressed LVEF before AS becoming severe and mortality, regardless of aortic valve replacement, after AS severe diagnosis. Conclusion: Subclinical myocardial dysfunction precedes measurable LVEF decline in AS patients. DLS integration with conventional assessment enables earlier identification of at-risk individuals, potentially optimizing follow-up scheduling and intervention timing.