Red blood cell distribution width-derived indces predicts long-term prognosis in acute heart failure

BackgroundRed blood cell distribution width (RDW), a routine indicator of erythrocyte volume variability, has shown potential in recent years in the prognostic assessment of a variety of diseases, including acute heart failure (AHF). The predictive ability of RDW-derived indices, namely the hemoglobin-to-RDW ratio (HRR), the RDW-to-lymphocyte ratio (RLR), the RDW-to-platelet ratio (RPR), the RDW-to-albumin ratio (RAR), and the product of RDW and mean corpuscular volume (RDWxMCV), for the prognosis of AHF remains unclear. MethodsThe study included 4090 eligible patients in the MIMIC 3.0 database, screened variables using Lasso regression, assessed whether these derivatives independently predicted 1-year mortality from AHF by Cox proportional hazards model, and plotted survival curves and time-related ROC curves. Finally, the relationship between each indicator and outcome was analyzed by restricted cubic spline. ResultsPositive events occurred in 2085 (51%) patients with decreased HRR and increased RLR, RPR, RAR, and RDWxMCV (P<0.001).After Cox model adjustment, HRR, RPR, RAR, and RDWxMCV remained independent predictors of 1-year prognosis for AHF (RLR was not) and the relationship may be more linear. Of these, RPR had the greatest impact on survival time (HR = 1.89, 95%CI: 1.33-2.67) and RDWxMCV had the highest predictive efficacy (AUC=0.612). ConclusionRDW-derived indices HRR, RPR, RAR and RDWxMCV independently predicted 1-year mortality in AHF, and RLR had no independent predictive value.