Abstract
OBJECTIVE: The purpose of our study is to examine whether serial measurements of serum sodium values after diagnosis identify a higher-risk subset of patients with heart failure with preserved ejection fraction. METHODS: We identified 50,932 subjects with HFpEF with 759,577 recorded sNa measurements (mean age 72 ± 11 years) using a validated algorithm in the VA national database from 2002 to 2012. We examined the association of repeated measures of sNa with mortality using a multivariable Cox proportional hazards model. RESULTS: After a median follow-up of 2.9 years (IQR: 1.2-5.4), 19,011 deaths occurred. After adjusting for age, sex, race, BMI, glomerular filtration rate, potassium, coronary artery disease, hypertension, hyperlipidemia, atrial fibrillation, pulmonary disease, diabetes, anemia, and medications, we found J-shaped associations of serum sodium with mortality. HRs for all-cause mortality were 2.48 (95% CI: 2.38-2.60) for the sNA 115.00-133.99 category; and 1.40 (95% CI: 1.35-1.46) for the sNA 143.00-175.00 category compared to the 137.01-140.99 category (ref). We used generalized estimating equation-based negative binomial regression to compute the incidence density ratios (IDR) to examine days hospitalized for heart failure and for all causes. There were a total of 1,275,614 days of all-cause hospitalization and 104,006 days of heart-failure hospitalization. The IDRs for the lowest sNA group were 2.03 (95% CI: 1.90-2.18) for all-cause hospitalization and 1.73 (95% CI: 1.39-2.16) for heart-failure hospitalization. CONCLUSIONS: Our findings suggest that monitoring of serum sodium values during longitudinal follow-up can identify HFpEF patients at risk of adverse outcomes.