Abstract Body (Do not enter title and authors here): Background
Pulse pressure (PP), a marker of aortic stiffness, is associated with risk for mortality and cardiovascular (CV) events, but these associations have not been extensively explored in patients with heart failure (HF). We examined the association between PP and CV outcomes in a pooled dataset of 4 large, well-characterized cohorts of patients with HF with moderately reduced (mrEF) or preserved EF (pEF).

Methods
I-PRESERVE, TOPCAT, PARAGON-HF and DELIVER were global, multicenter, randomized clinical trials testing irbesartan, spironolactone, sacubitril/valsartan and dapagliflozin respectively, against either a placebo or an active control (valsartan in PARAGON-HF), in patients with HF and a LVEF >40% (in DELIVER) or ≥45% (in the other trials). Predictors of higher PP were identified using forward stepwise regression analysis. The relationship between baseline continuous PP and the incidence rates (IR) of first HF hospitalization (HFH) or CV death was analyzed with restricted cubic splines. To assess the prognostic significance of incorporating systolic blood pressure (SBP) along with PP, we categorized patients into 4 groups based on median values of PP and SBP.

Results
A total of 18,627 patients (age 71±9 years, 49% men, PP 55±14 mmHg, SBP 131±15 mmHg) were included. Over a median follow-up of 2.9 years, 3,852 patients experienced a first HFH or CV death event (7.4 events per 100py). Predictors of higher PP values included higher SBP, older age, female sex, Asian race, higher EF, use of diuretics, higher N-terminal pro B-type natriuretic peptide (NT-proBNP) levels, and diabetes. Conversely, factors associated with lower PP included atrial fibrillation, higher heart rate, worse New York Heart Association (NYHA) Functional Class, use of beta-blockers, and higher estimated glomerular filtration rate. After adjusting for predictors of PP and other relevant variables, higher PP was associated with a higher risk for HFH or CV death. This relationship was nonlinear (p<0.001), with risk increasing markedly at higher PP values (Figure 1). After multivariable adjustment, higher PP was associated with higher CV risk regardless of baseline SBP, compared to lower PP and lower SBP (Figure 2).

Conclusions
Our analysis of a pooled dataset from 4 clinical trials including patients with HFmrEF or pEF suggests that PP was an independent predictor of CV events in this population, regardless of baseline SBP.