Heart failure & cardiomyopathiesCohort study

Predictors of 1-Year Mortality Among Patients With Heart Failure With Preserved Ejection Fraction

JACC AdvPublication date not resolved

  • retrospective cohort study
  • development, comparison and internal validation of prediction models
  • electronic medical records from a large academic health system
  • training (80%) and test (20%) sets stratified by outcome
  • multiple imputation by chained equations

Authors' conclusion

Our transparent risk tool using routinely available admission data appears feasible, allowing for patient-level, precision health risk assessment.

Study context

Population
Adults with ejection fraction (EF) ≥50% admitted for a first-time heart failure (HF) exacerbation
Participants
7,840 admissions
Exposure
Prediction models for 1-year all-cause mortality after first hospitalization for decompensated HF with preserved EF, compared by AUROC, concordance index and calibration

Main finding

As reported in the abstract

One-year mortality · 1 year

One-year mortality

31.5%

Observational data: associations, not effects.

Further findings

OutcomeValues
Test-set AUROC (classifiers) · acceptable calibrationExtreme Gradient Boosting AUROC 0.751 (95% CI 0.727–0.775) · Light Gradient Boosting AUROC 0.749 (95% CI 0.721–0.776)
Concordance index (survival models) · GBS achieved the highest concordance indexGBS concordance index 0.718 (95% CI 0.696–0.740) · RSF concordance index 0.711 (95% CI 0.690–0.734) · Cox concordance index 0.704 (95% CI 0.680–0.728)
Time-dependent AUROC (survival models) · 12 monthsGBS time-dependent AUROC 0.759 (95% CI 0.716–0.799) · RSF time-dependent AUROC 0.750 (95% CI 0.708–0.789) · Cox time-dependent AUROC 0.735 (95% CI 0.692–0.777)

Baseline characteristics

mean age
78 years
women
55.6%

Most consistent risk signals

Lower albumin, older age, higher N-terminal pro-B-type natriuretic peptide, renal dysfunction, and lower hemoglobin were the most consistent risk signals

Original publication

Predictors of 1-Year Mortality Among Patients With Heart Failure With Preserved Ejection Fraction. JACC Adv. doi:10.1016/j.jacadv.2026.102847

Summarised from the published abstract; every number is checked against it