Background: Few studies have incorporated echocardiography and laboratory data to predict clinical outcomes in heart failure (HF) with preserved ejection fraction (HFpEF).

Objectives: This study aims to use machine learning to find predictors of HF hospitalization and cardiovascular (CV) death in HFpEF.

Methods: From the Chang Gung Research Database in Taiwan, 6,092 HFpEF patients (2,898 derivation, 3,194 validation) identified between 2008-2017 were followed until 2019. A random survival forest (RSF) model, using 58 variables, was developed to predict the composite outcome of HF hospitalization and CV death.

Results: During 2.9-year follow-up, 37.7% of derivation and 36.0% of validation cohort patients experienced HF hospitalization or CV death. The study identified 15 predictive indicators, including age ≥ 65 years, B-type natriuretic peptide level ≥ 600 pg/mL, left atrium size ≥ 46 mm, atrial fibrillation, frequency of HF hospitalization within 3 years, body mass index < 30 kg/m2, moderate or severe mitral regurgitation, left ventricular (LV) posterior wall thickness of <10 or ≥13 mm, dysnatremia, LV end-diastolic dimension of <40 or ≥56 mm, uric acid level ≥ 7 mg/dL, triglyceride level of <70 or ≥200 mg/dL, blood urea nitrogen level ≥ 20 mg/dL, interventricular septum thickness of <11 or ≥20 mm, and glycated hemoglobin level of <6% or ≥8%. The RSF model demonstrated robust external generalizability with an 86.9% area under curve in validation.

Conclusions: Machine learning identified 15 predictors of HF hospitalization and CV death in HFpEF patients, helping doctors identify high-risk individuals for tailored treatment.

JACC: Asia Editor-in-Chief 

Jian’an Wang, MD, PhD, FACC

CME Editor 

Kenneth A. Ellenbogen, MD

Author
Megan Pelter, MD

Important Dates

Date of Release: December 3, 2024
Term of Approval/Date of CME/MOC Expiration: December 2, 2025