Multiparametric Strategy to Predict Early Disease Decompensation in Asymptomatic Severe Aortic Regurgitation

Use of the current echocardiography-based indications for aortic regurgitation (AR) surgery might result in late valve replacement at the stage of irreversible myocardial damage. Therefore, we aimed to identify simple models combining multiple echocardiography or magnetic resonance imaging (MRI)-derived indices and natriuretic peptides (BNP [brain natriuretic peptide] or NT-proBNP [N-terminnal pro-B type natriuretic peptide]) to predict early disease decompensation in asymptomatic severe AR.

This prospective and multicenter study included asymptomatic patients with severe AR, preserved left ventricular ejection fraction (>50%), and sinus rhythm. The echocardiography and MRI images were analyzed centrally in the CoreLab. The study end point was the onset of indication for aortic valve surgery as per current guidelines.

The derivative cohort consisted of 127 asymptomatic patients (age 45±14 years, 84% males) with 41 (32%) end points during a median follow-up of 1375 (interquartile range, 1041-1783) days. In multivariable Cox regression analysis, age, BNP, 3-dimensional vena contracta area, MRI left ventricular end-diastolic volume index, regurgitant volume, and a fraction were identified as independent predictors of end point (all P<0.05). However, a combined model including one parameter of AR assessment (MRI regurgitant volume or regurgitant fraction or 3-dimensional vena contracta area), 1 parameter of left ventricular remodeling (MRI left ventricular end-diastolic volume index or echocardiography 2-dimensional global longitudinal strain or E wave), and BNP showed significantly higher predictive accuracy (area under the curve, 0.74-0.81) than any parameter alone (area under the curve, 0.61-0.72). These findings were confirmed in the validation cohort (n=100 patients, 38 end points).

In asymptomatic severe AR, multimodality and multiparametric model combining 2 imaging indices with natriuretic peptides, showed high accuracy to identify early disease decompensation. Further prospective studies are warranted to explore the clinical benefit of implementing these models to guide patient management.

URL: https://www.

gov; Unique identifier: NCT02910349.