Congestion-based phenotyping of acute decompensated heart failure with preserved ejection fraction and in-hospital mortality

Background Data on phenotyping acute heart failure with preserved ejection fraction (AHFpEF) remain limited, and clinically applicable congestion-based models are insufficiently explored. Aim To identify phenotypes of AHFpEF based on laboratory and instrumental parameters and to evaluate their association with adverse in-hospital outcomes. Methods A prospective observational registry included 240 patients urgently admitted to the intensive care unit with AHFpEF. Inclusion criteria were: left ventricular EF ≥50%, evidence of left ventricular diastolic dysfunction, and NT-proBNP levels >300 pg/mL in sinus rhythm or >600 pg/mL in atrial fibrillation (AF). All patients underwent physical examination, laboratory testing, chest radiography, echocardiography, lung ultrasound with B-line quantification, venous congestion assessment using the VExUS protocol, and bioimpedance hydration analysis. A two-step cluster analysis was performed using venous congestion grade (VExUS) and bioimpedance phase angle to identify distinct phenotypes (silhouette measure=0.8). Results Three phenotypes were identified (n=88, n=86, and n=60). Phenotype 1 was characterized by the lowest body mass index (BMI) (27.7±4.5 kg/m²), minimal congestion (peripheral edema 36.4%), preserved inferior vena cava collapsibility >50% in 79.5%, rare right ventricular (RV) dilation (9.1%) and dysfunction (3.4%), and the lowest E/e′ values (12.2±2.1). Phenotype 2 demonstrated intermediate congestion severity, with peripheral edema in 80.2%, pleural effusion in 50.0%, AF in 62.8%, and type 2 diabetes in 41.9%. Phenotype 3 exhibited the most pronounced systemic venous overload: peripheral edema in 100%, pleural effusion in 85%, ascites in 46.7%, higher BMI (35.3±7.1 kg/m²), elevated E/e′ (16.1±2.2), RV dilation (51.7%) and reduced RV systolic function (50.0%). Bioimpedance analysis showed a 43.6% increase in total body water and a 54.0% increase in extracellular water (p<0.001). Phenotype 3 patients required higher initial furosemide doses (median 170 mg) and longer diuretic therapy duration (6 days; p<0.001). In-hospital mortality significantly differed across phenotypes: 2.3% in phenotype 1 (2/88), 2.3% in phenotype 2 (2/86), and 21.7% in phenotype 3 (13/60). Kaplan–Meier analysis confirmed worse survival in phenotype 3 compared with phenotypes 1 and 2 (Log-rank p=0.013 and p=0.012, respectively), with most deaths occurring within the first few days of hospitalization. Conclusions Three clinically distinct phenotypes of AHFpEF were identified, differing in congestion severity and in-hospital prognosis. The most unfavorable phenotype was characterized by marked systemic venous overload and nearly tenfold higher in-hospital mortality. Integrated ultrasound-based congestion assessment may improve phenotype-oriented risk stratification and management in AHFpEF.