Bedside ultrasound congestion phenotypes in acute heart failure
A Torrelles, M Domingo, E Zamora, J Lupon, M Soler, G Romero-Gonzalez, A Caballero, M Maymi, R Vila, G Guix, E Barcelo, B Ochoa, E Revuelta, C Tural, A Bayes-GenisAbstract
Background
Congestion is the main cause of hospitalization for acute decompensated heart failure (HF). The traditional standardized decongestive approach, largely based on loop diuretics, often fails to achieve adequate decongestion, underscoring the need for individualized strategies. Congestion phenotypes may differ in clinical presentation, therapeutic response, and prognosis. Characterizing these penotypes at admission may guide personalized decongestive strategies.
Purpose
To describe ultrasound-defined congestion profiles at hospital admission and evaluate their association with clinical characteristics, congestion biomarkers, and in-hospital outcomes.
Methods
This prospective, single-centre study included consecutive patients hospitalized for acute HF. At admission, congestion was assessed within the first 12 hours using a multiparametric approach: ultrasound [lung and VExUS (Venous Excess Ultrasound) protocols], clinical scores [Composite Congestive Score (CCS) and EVEREST score], and biomarkers (NT-proBNP, CA125). Patients were classified into three predefined phenotypes: intravascular (VExUS ≥ 1); tissue [≥ 3 B-lines in at least 2 zones per hemithorax (among 8 anterolateral zones scanned)]; and mixed (intravascular + tissue). ANOVA or Kruskal–Wallis were used for continuous variables, χ² for categorical variables, and linear-by-linear association for trends.
Results
259 patients were included (71±13 years, 64.9% male). Most patients (59.5%) were admitted with de novo HF. Mixed phenotype was the most prevalent (73.4%), followed by tissue (18.5%) and intravascular (8.1%) phenotypes. Significant differences were observed across phenotypes regarding clinical and biochemical severity at admission (Table). Clinical congestion scores correlated significantly with ultrasound phenotypes (EVEREST p=0.002; CCS p<0.001), reflecting higher congestion burden in the mixed group. Additionally, significant differences were found in glomerular filtration rate (p=0.017) and hemoglobin levels (p=0.001). No significant differences were found based on sex (p=0.124), atrial fibrillation (p=0.213) or left ventricular ejection fraction categories (p=0.983). Patients with the mixed phenotype exhibited the highest NT-proBNP and CA125 values (Kruskal-Wallis p=0.003 and p=0.005, respectively). Post-hoc analyses (Dum+Bonferroni correction) showed significant differences specifically between the intravascular and mixed groups (p=0.002 and p=0.02, respectively). In-hospital mortality was 7.7% and showed a significant linear association with the complexity of the congestion phenotype (p=0.027).
Conclusion
Early multiparametric ultrasound assessment identifies distinct congestion phenotypes in acute HF patients. Mixed phenotype is the most frequent and is associated with greater clinical and biochemical severity. This approach also aids in stratifying in-hospital mortality risk.For image description, please refer to the figure legend and surrounding text.