Abstract: Objective To perform latent profile analysis (LPA) on the clinical phenotypes of elderly emergency patients with heart failure with preserved ejection fraction (HFpEF) and explore their associations with clinical outcomes. Methods A total of 300 elderly emergency patients with HFpEF admitted to the Affiliated Hospital of Qinghai University from January 2023 to December 2024 were selected. Core variables were chosen based on the pathophysiological mechanisms of HFpEF. LPA was conducted to determine the optimal classification. Patients were followed for 30 days and divided into adverse or favorable outcome groups based on their disease progression. Binary multivariate logistic regression was used to identify independent factors influencing outcomes. Results The 4-class model was identified as optimal. The probability of latent categories being assigned to specific types was 96.24%～99.02%, indicating high reliability. The four clinical phenotypes were designated as metabolic-hypertensive type (35.00%), multimorbidity coexistence type (30.00%), inflammatory-volume overload type (25.00%), and isolated diastolic dysfunction type (10.00%). Multivariate logistic regression revealed that advanced age (OR=1.919, 95%CI: 1.278-2.882), comorbid atrial fibrillation (OR=2.272, 95%CI: 1.464-3.526), elevated left atrial volume index (LAVI) (OR=2.979, 95%CI: 1.294-6.858), and the inflammatory-volume overload phenotype (OR=3.201, 95%CI: 1.423-7.201)were independent risk factors for adverse outcomes. Conclusions LPA classifies elderly emergency HFpEF patients into four clinical phenotypes: metabolic-hypertensive, multimorbidity coexistence, inflammatory-volume overload, and isolated diastolic dysfunction. Patients with the inflammatory-volume overload phenotype have the worst prognosis, identifying it as a significant risk factor for adverse clinical outcomes.

Key words: aged, heart failure, latent profile analysis, prognosis prediction