Objectives/Background: In heart failure (HF), pulse pressure (PP) may reflect both vascular stiffness and left ventricular function, but its prognostic role in relation to ejection fraction (EF) is poorly understood. Methods: In the Swedish Heart Failure Registry, we investigated the association between PP and 1-year mortality in patients with HF and reduced (HFrEF, amp;lt;40%), mid-range (HFmrEF, 40-49%) and preserved EF (HFpEF, amp;gt;= 50%), using multivariable logistic regression and restricted cubic splines. Results: Among 36,770 patients discharged alive or enrolled as out-patients with 1-year follow-up (mean age 74 +/- 12 years, 63% men, 56% HFrEF, 21% HFmrEF, 23% HFpEF), crude one-year mortality was 18%. Mean PP increased across EF groups: 51 +/- 16 in HFrEF, 57 +/- 18 in HFmrEF, 60 +/- 19 mm Hg in HFpEF. In crude regression splines, the association between PP and mortality was U-shaped in HFmrEF and HFpEF, but curvilinear with only low PP associated with mortality in HFrEF. In multivariable analyses, a significant interaction by EF group and PP was observed (p(interaction) = 0.015): low PP was associated with higher mortality in HFrEF (adjusted OR [1st vs. 4th quintile] = 1.40, 95% CI 1.18-1.67) and HFpEF (1.43, 1.14-1.81) but only by trend in HFmrEF; high PP had a trend towards higher mortality in HFmrEF (5th vs. 3rd quintile = 1.30, 1.00-1.69) and HFpEF (1.25, 0.98-1.61). Conclusions: The association between PP and mortality in HF was influenced by EF. Low PP was independently associated with mortality in HFrEF and HFpEF and by trend in HFmrEF. High PP was independently associated with mortality by trend in HFmrEF and HFpEF. (C) 2017 Elsevier B.V. All rights reserved.