Congestion phenotypes in elderly patients with acute heart failure: distinct patterns in preserved vs. reduced ejection fraction

Background

Congestion is the main driver of worsening acute heart failure (AHF), yet whether congestion phenotypes differ by left ventricular ejection fraction (LVEF) in the elderly remains uncertain.

A multiparametric approach to congestion is increasingly recommended in AHF to allow for better identification of congestion and more tailored management to optimize patient outcomes.

Purpose

The aim of this study is to characterize clinical, echocardiographic, and biomarker congestion profiles by LVEF phenotype (HFpEF ≥50% vs. HFrEF <50%) in elderly patients hospitalized with AHF, and secondarily to assess associations with rehospitalization and mortality.

Methods

We conducted a retrospective cohort study of 830 consecutive patients admitted with AHF to a tertiary hospital. Congestion was assessed clinically, by echocardiography, and with biomarkers (BNP, CA125) and point-of-care ultrasound (POCUS). Outcomes were heart failure hospitalization and all-cause mortality over a median 310 days (IQR 62–543). Cox models adjusted for age, sex, systolic blood pressure, creatinine, haemoglobin, pleural effusion, edema, LVEF, BNP, and CA125; LVEF was also modelled using restricted cubic splines.

Results

Median age was 87 years, 65.6% were women and 81.7% of patients had HFpEF. Traditional congestion signs (peripheral edema, rales, jugular venous distension) and NYHA class were similar across phenotypes, and systolic blood pressure was slightly lower in HFrEF. HFrEF showed clear structural/functional differences—lower LVEF, larger LV dimensions, reduced TAPSE, and modestly larger indexed LA volume—whereas diastolic indices (E/e′) and POCUS metrics (IVC dilation/collapsibility, B-lines, pleural effusion) were similar. HFrEF had substantially higher BNP and CA125 concentrations. During follow-up, 301 patients (36.3%) were rehospitalized and 418 (50.4%) died. LVEF phenotype was not associated with rehospitalization or mortality (figure 1). LVEF as a continuous variable was not associated with mortality (HR 1.004 per 1%; p=0.470) but showed a modest positive association with rehospitalization (HR 1.013 per 1%; p=0.019). Spline analyses provided no evidence of a non-linear association with mortality and no clear non-linearity for rehospitalization (figure 2).

Conclusion

In this large cohort of elderly patients with acute heart failure, HFpEF was the predominant phenotype. Despite similar clinical and POCUS congestion signs, HFrEF displays distinct echocardiographic remodelling and higher biomarker levels (CA125, BNP). Prognosis was not differentiated by LVEF phenotype, underscoring the value of a multiparametric, biomarker-informed assessment beyond LVEF categories.Figure 1:Kaplan-Meier curvesFor image description, please refer to the figure legend and surrounding text.Figure 2:Restricted cubic spline analysFor image description, please refer to the figure legend and surrounding text.