Mortality Burden for Patients With Untreated Aortic Regurgitation

Abstract

Background: Aortic valve replacement (AVR) is indicated in patients with severe aortic regurgitation (AR); however, certain clinical factors may identify patients with less-than-severe AR at high mortality risk if untreated. Objectives: The authors sought to characterize key associations with mortality across the spectrum of AR in patients not treated with AVR from a large, contemporary database. Methods: We analyzed patients >18 years of age with documented AR assessment in a deidentified real-world data set from 27 U.S. institutions with appropriate permissions (egnite Database, egnite, Inc). Diagnosed AR severity was extracted from echocardiographic reports using a natural language processing–based algorithm. Cox multivariable analysis modeled the impact of key factors on untreated mortality according to AR severity. Results: In total, 81,378 patients were included for analysis. Hazard ratios for mortality were 1.26 (95% CI: 1.18-1.35) and 2.37 (95% CI: 1.96-2.87) for moderate and severe AR, respectively. Other significant associations included left ventricular (LV) ejection fraction ≤55% (1.09 [95% CI: 1.02-1.15]), LV dilation (1.34 [95% CI: 1.21-1.48]), left atrial dilation (1.09 [95% CI: 1.03-1.16]), atrial fibrillation (1.11 [1.04-1.17]), and elevated B-type natriuretic peptide/N-terminal pro-B-type natriuretic peptide (1.71 [95% CI: 1.60-1.84]). Modeled mortality risk increased with the presence of these key factors both alone and in combination. Conclusions: In patients with untreated AR, LV remodeling, left atrial remodeling, and other markers of cardiac damage are associated with substantial mortality risk, both for severe and moderate AR. Further study is needed to determine whether AVR is warranted in patients with less-than-severe AR with at-risk factors.

Keywords

• aortic insufficiency
• aortic regurgitation
• aortic valve
• aortic valve replacement
• database
• natural language processing