Abstract
AIMS: Increased red cell distribution width (RDW) is a poor prognostic factor in patients with heart failure (HF). However, only a few large-scale studies have identified the clinical utility of RDW after adjusting for covariates affecting RDW. METHODS AND RESULTS: From January 2010 to April 2021, we retrospectively enrolled patients diagnosed with HF from three referral hospitals with available RDW data (taken within 3 months of HF diagnosis) using an integrated clinical data system. Patients with an ejection fraction (EF) < 50% or HFA-PEFF (Heart Failure Association Pre-test assessment, Echocardiography and natriuretic peptide, Functional testing, Final aetiology) score ≥ 2 without severe valvular heart disease or coronary revascularization were enrolled. The primary endpoint was all-cause mortality, and cardiovascular mortality was also collected. Multivariable Cox regression analysis and stabilized inverse probability of treatment weighting (IPTW) were used to identify any association between RDW and all-cause death by balancing covariates or compounding factors. The global χ(2) score was calculated and discrimination analysis was performed to evaluate the incremental value of RDW in predicting prognosis. Among the 6599 participants enrolled in this study, 1256 (19.0%) cases of all-cause death occurred, and the median duration of follow-up was 887 (interquartile range 351-1589) days. Elevated RDW at the initial diagnosis was associated with poor prognosis [cumulative incidence: 819 (30.2%) vs. 437 (11.2%), relative risk 1.58, 95% confidence interval (CI) 1.51-1.67, log-rank P < 0.001]. Multivariable Cox analysis showed that elevated RDW was a poor prognostic factor for the primary endpoint [hazard ratio (HR) 1.11, 95% CI 1.06-1.16, P < 0.001], independent of clinical risk factors, N-terminal pro-brain natriuretic peptide (NT-proBNP), and EF, which was concordant with the stabilized IPTW (HR 1.29, 95% CI 1.10-1.49, P < 0.001). Adding RDW to model composed of traditional risk factors, NT-proBNP, and echocardiographic parameters showed incremental prognostic value for predicting poor prognosis (area under the receiver operating characteristic curve, 0.799-0.826; P < 0.001). CONCLUSIONS: Increased RDW at the time of diagnosis is associated with poor prognosis in patients with HF, independent of clinical risk factors, such as NT-proBNP, and echocardiographic parameters. Therefore, RDW may aid in the management of these patients beyond traditional risk factors.