Prognostic value of left atrial reverse remodelling in heart failure with preserved ejection fraction.

Abstract

Left atrial reverse remodelling (LARR) reflects an improvement in the function or structure. However, it is unclear whether the presence of LARR is associated with better outcomes in patients with heart failure with preserved ejection fraction (HFpEF). The study was a monocentric retrospective cohort one. Consecutive HFpEF patients admitted to the hospital between 1 January 2018 and 30 June 2020 were included. This cohort was divided into LARR and non-LARR groups based on the recovery of the left atrium. The primary endpoints were all-cause mortality, rehospitalization for heart failure, and the composite of death or readmission. Significant predictors of LARR were examined. A total of 409 patients were enrolled, including 90 cases in the LARR group and 319 in the non-LARR group. Kaplan-Meier analysis showed that compared with the non-LARR group, the LARR group had a lower incidence of rehospitalization for heart failure and the composite of death or readmission but not all-cause mortality. Similar results were observed in a subgroup analysis of patients with and without atrial fibrillation. Cox regression analysis demonstrated that the non-LARR group experienced higher risks of heart failure-related readmission [hazard ratio: 1.785, 95% confidence interval (CI) 1.236-3.215, P = 0.037] and the composite outcome (hazard ratio: 1.684, 95% CI 1.254-2.865, P = 0.044), but not all-cause mortality (hazard ratio: 1.475, 95% CI: 0.481-3.527, P = 0.577) compared with the LARR group after adjusting for significant confounders. Logistic regression analysis showed that mild mitral regurgitation and the use of loop diuretics were two positive predictors of LARR in patients with HFpEF. LARR is an effective echocardiographic index that can be used to predict heart failure-related readmission in HFpEF. Therefore, regular assessment of left atrial size can provide a useful marker for risk stratification of heart failure.