Risk Score Model for Predicting Mortality in Advanced Heart Failure Patients Followed in a Heart Failure Clinic

The prevalence of heart failure (HF) in the population is increasing, concomitant with high incidence of rehospitalizations and mortality. The aim of this study was to characterize a prognostic risk score model for patients with chronic HF. A total of 500 patients followed at the HF clinic were evaluated by clinical, functional, laboratory, imaging, and therapeutic variables that were correlated to mortality during a follow-up period of 25 months. Risk stratification was carried out by applying a risk score model based on multivariate analysis. Predictors correlated with mortality during follow-up were systolic blood pressure <110 mm Hg, male sex, age older than 70 years, 6-minute walk distance <300 m, lack of β-blocker therapy, hyperuricemia (>7.5 mg/dL), hyponatremia, and prolonged QTc interval (>450 ms). Based on these variables, a risk score model (score 0-55) was established and included low risk, score <21 (9% mortality during 2-year follow-up); moderate risk, 21 to 29 (22%); high risk, 30 to 35 (35%), and very high risk: ≥36 points (62% 2-year mortality). The risk model had good discrimination ability (concordance index 0.75), which was better than the performance of the Seattle Heart Failure Model on our cohort (0.69). Simple noninvasive characteristics examined during the initial admission to the HF clinic can serve as prognostic markers for mortality and may help in the process of therapeutic decision-making in patients with HF.