AimsRed cell distribution width (RDW) is a strong prognostic marker in patients with heart failure (HF) and reduced ejection fraction and other conditions. However, very little is known about its prognostic significance in HF with preserved ejection fraction. We examined the relationship between RDW and outcomes and the effect of sacubitril/valsartan, compared with valsartan, on RDW and clinical outcomes in PARAGON‐HF.Methods and resultsPARAGON‐HF enrolled patients with a left ventricular ejection fraction of ≥45%, structural heart disease, and elevated N‐terminal pro‐B‐type natriuretic peptide (NT‐proBNP). The primary endpoint was a composite of total HF hospitalizations and cardiovascular deaths. Median RDW at randomization was 14.1% (interquartile range 13.5–15.0%). Patients with higher RDW levels were more often men and had more comorbidity, a higher heart rate and NT‐proBNP concentration, more advanced New York Heart Association class, and worse Kansas City Cardiomyopathy Questionnaire scores. There was a graded relationship between quartiles of RDW at randomization and the primary endpoint, with a significantly higher risk associated with increasing RDW, even after adjustment for NT‐proBNP and other prognostic variables {Quartile 1, reference; Quartile 2, rate ratio 1.03 [95% confidence interval (CI) 0.83 to 1.28]; Quartile 3, 1.25 [1.01 to 1.54]; Quartile 4, 1.70 [1.39 to 2.08]}. This association was seen for each of the secondary outcomes, including cardiovascular and all‐cause death. Compared with valsartan, sacubitril/valsartan reduced RDW at 48 weeks [mean change −0.09 (95% CI −0.15 to −0.02)]. The effect of sacubitril/valsartan vs. valsartan was not significantly modified by RDW levels at randomization.ConclusionsRDW, a routinely available and inexpensive biomarker, provides incremental prognostic information when added to established predictors. Compared with valsartan, sacubitril/valsartan led to a small reduction in RDW.