Congestion-Perfusion Phenotypes and In-Hospital Mortality in Acute Heart Failure: Phenotype-Specific Prognostic Markers and Right-Heart Involvement.

摘要

Background and Objectives: Acute heart failure (AHF) has heterogeneous clinical profiles and considerable short-term mortality. Bedside evaluation of congestion and peripheral perfusion may help identify clinically relevant risk groups. This study aimed to evaluate the relationship between a non-invasive congestion-perfusion classification and in-hospital mortality in patients hospitalized with AHF and to explore the prognostic relevance of clinical, biological, and right-heart variables within the phenotypes. Materials and Methods: We performed an observational study that analyzed 790 patients hospitalized with AHF. Patients were classified at presentation into four predefined congestion-perfusion phenotypes: non-congested/preserved perfusion, congested/preserved perfusion, non-congested/impaired perfusion, and congested/impaired perfusion. Congestion was defined using pulmonary and systemic markers, whereas impaired perfusion was defined by the presence of at least two predefined criteria of hypoperfusion. Clinical, biological, and echocardiographic parameters were analyzed according to phenotype and in-hospital mortality. Discriminative performance was evaluated using ROC curve analysis, and logistic regression models were constructed within the congestive phenotypes to investigate the prognostic value of combined clinical and biological markers, with bootstrap internal validation in the combined models. Results: 78 patients died during hospitalization, with different rates across phenotypes; 2.9% in non-congested/preserved perfusion, 3.0% in non-congested/impaired perfusion, 8.0% in congested/preserved perfusion, and 19.5% in congested/impaired perfusion. In congestive phenotypes, several adverse markers were identified. Exploratory combined models demonstrated discriminatory performance in the congested/preserved perfusion phenotype (AUC 0.731) and in the congested/impaired perfusion phenotype (AUC 0.838); bootstrap optimism-corrected AUCs were 0.693 and 0.785, respectively. Right ventricular parameters were more strongly associated with mortality in the congested/impaired perfusion phenotype. LVEF showed limited discrimination within individual phenotypes. Conclusions: The highest in-hospital mortality was observed in the congestive/impaired perfusion phenotype, with a more adverse clinical, biological, and echocardiographic profile. Within the congestive phenotypes, exploratory prognostic marker patterns were partially overlapping, while LVEF showed limited discrimination.