A clinical prediction model integrating cardiovascular-kidney-metabolic biomarkers for the composite outcome of quality-of-life deterioration and rehospitalization in elderly HFpEF patients.

摘要

BACKGROUND: Cardiovascular-kidney-metabolic (CKM) syndrome stage 4a encompasses elderly heart failure with preserved ejection fraction (HFpEF) patients with metabolic risk factors or chronic kidney disease (CKD). No existing HFpEF prediction tool has incorporated glycated haemoglobin A1c (HbA1c) alongside echocardiographic diastolic function parameters within a CKM-anchored architecture targeting quality-of-life (QoL) deterioration. OBJECTIVE: To develop, internally validate, and temporally validate a least absolute shrinkage and selection operator (LASSO)-Cox nomogram for predicting the long-term composite endpoint of quality-of-life (QoL) deterioration and unplanned heart failure rehospitalisation in elderly CKM stage 4a HFpEF patients. METHODS: We retrospectively enrolled 302 consecutive patients aged ≥60 years with HFpEF meeting CKM stage 4a criteria admitted between January 2020 and December 2023, and a separate temporal validation cohort (n = 102) admitted between January 2024 and June 2025. Eligible patients had confirmed HFpEF per 2021 European Society of Cardiology guidelines (symptomatic heart failure [New York Heart Association {NYHA} class II-IV], left ventricular ejection fraction [LVEF] ≥50%, elevated N-terminal pro-B-type natriuretic peptide [NT-proBNP], and echocardiographic diastolic dysfunction) and fulfilled CKM stage 4a criteria (metabolic risk factors or CKD without dialysis dependency). The primary outcome was a composite of QoL deterioration (Minnesota Living with Heart Failure Questionnaire [MLHFQ] increase ≥5 points) and unplanned heart failure rehospitalisation. Feature selection used LASSO-Cox regression integrated with multiple imputation. RESULTS: The primary endpoint occurred in 116 patients (38.4%) over a median follow-up of 22.4 months. LASSO selected eight predictors: age, NYHA class, ln-NT-proBNP, estimated glomerular filtration rate (eGFR), uric acid, E/e' ratio, left atrial volume index (LAVI), and HbA1c; LVEF was compressed to zero. The nomogram achieved C-indices of 0.812 (95% CI 0.770-0.854) in training, 0.790 (0.728-0.852) in internal validation, and 0.783 (0.718-0.848) in temporal validation. In the temporal validation cohort the integrated calibration index was 0.041 at 12 months. Decision curve analysis demonstrated net clinical benefit across threshold probabilities of 10%-85%. An exploratory CKM metabolic burden score stratified 24-month event rates at 21.6%, 41.3%, and 62.5%. CONCLUSIONS: An eight-variable nomogram integrating HbA1c with CKM-axis biomarkers and diastolic function parameters predicted the long-term composite of QoL deterioration and unplanned heart failure rehospitalisation in elderly CKM stage 4a HFpEF patients, with performance preserved in temporal validation. Prospective multicentre validation in geographically distinct populations is warranted.