Clinical phenotypes of heart failure patients with supranormal ejection fraction.

ABSTRACT

AIMS: Heart failure (HF) with supranormal ejection fraction (HFsnEF) represents a distinct clinical entity characterized by limited treatment options and an unfavourable prognosis. Revealing its phenotypic diversity is crucial for understanding disease mechanism and optimizing patient management. We aim to identify phenotypic subgroups in HFsnEF using unsupervised clustering analysis. METHODS: Consecutive hospitalized patients with a diagnosis of HF and a left ventricular ejection fraction ≥65% at baseline echocardiographic evaluations were included for analysis. We conducted unsupervised hierarchical clustering analysis on principal components (HCPC) to identify HFsnEF phenogroups using mixed data variables including demographics, HF duration, vital signs, anthropometrics, smoking/drinking status, HF aetiology, comorbid diseases, laboratory tests and echocardiographic parameters. We then employed decision tree modelling to identify parameters capable of distinguishing distinct clusters. Clinical outcomes, including all-cause death, cardiovascular (CV) death and CV readmission for different clusters, were examined. RESULTS: Three mutually exclusive clusters were identified from the cohort of 221 HFsnEF patients. Cluster 1 (52.5%) predominantly consisted of patients with valvular heart disease, who had larger cardiac chambers and a higher prevalence of atrial fibrillation/atrial flutter. Cluster 2 (26.2%) primarily comprised older ischaemic patients with a higher prevalence of metabolic comorbidities. Cluster 3 (21.3%) were mainly hypertrophic cardiomyopathy patients. Two clinical variables were identified that could be used to group all HFsnEF patients into one of the clusters; they were HF aetiology and comorbid diabetes. During the median follow-up of 53.4 months, 46 (20.8%) all-cause deaths occurred, among them 39 of CV causes. Seventy (31.7%) patients experienced CV readmissions. Three clusters showed distinct differences in mortality outcomes, with Cluster 1 exhibiting the highest risk of all-cause mortality [Cluster 1 vs. Cluster 2 adjusted hazard ratio (aHR) = 3.32, P = 0.022; Cluster 1 vs. Cluster 3 aHR = 3.81, P = 0.036; Cluster 2 vs. Cluster 3 aHR = 1.15, P = 0.865] and CV mortality (Cluster 1 vs. Cluster 2 aHR = 3.73, P = 0.022; Cluster 1 vs. Cluster 3 aHR = 4.27, P = 0.020; Cluster 2 vs. Cluster 3 aHR = 1.15, P = 0.870). CV readmission risk was comparable among the three clusters (Cluster 1 vs. Cluster 2 aHR = 0.82, P = 0.590; Cluster 1 vs. Cluster 3 aHR = 1.04, P = 0.900; Cluster 2 vs. Cluster 3 aHR = 1.28, P = 0.580). CONCLUSIONS: In a heterogeneous HFsnEF cohort, three clusters were identified by unsupervised HCPC with distinct clinical characteristics and outcomes.