Abstract
BACKGROUND: Most previous risk prediction models in patients hospitalized for heart failure (HF) are derived from populations in Western countries, and it is unclear whether these models are applicable to Asian populations. This study aimed to construct a risk score system for predicting one-year mortality risk in Asian patients and to compare the applicability of this risk score system with the 3C-HF score system. METHODS: We used the population in the Taiwan Society of Cardiology-Heart Failure with Reduced Ejection Fraction (TSOC-HFrEF) registry, which is a prospective cohort of patients admitted for acute decompensated heart failure (ADHF) in Taiwan. The risk score system was constructed using multivariate Cox-model derived coefficients. A bootstrapping procedure was also used for bias-corrected evaluations. Comparisons between this constructed model and the 3C-HF score prediction model were evaluated using calibration plots and area under curve (AUC)/receiver operating characteristic (ROC) curve. RESULTS: Patients with complete data (n = 1127) in the TSOC-HFrEF registry were analyzed. During one year of follow-up, 14.5% (n = 163) of the patients died. A risk score system was constructed with the following predictors: body mass index, diastolic blood pressure, dyslipidemia, diabetes, aortic regurgitation, QRS duration, hemoglobin concentration, and digoxin usage. Compared to the 3C-HF score system, this risk score system had a similar discriminatory ability (AUC/ROC values of 0.675 and 0.636, p = 0.127) and both were well-calibrated in the Taiwan population. CONCLUSIONS: The proposed risk score system for predicting one-year all-cause mortality in Taiwanese patients with ADHF may facilitate risk stratification in Asian patients with HF.