**Background:** Epicardial adipose tissue (EAT) may play a role in the pathophysiology of heart failure with preserved ejection fraction (HFpEF), potentially through systemic inflammation, microcirculatory dysfunction, and local mechanical or paracrine effects. Obesity-related HFpEF is a recognized phenotype, but the specific contribution of EAT beyond general obesity remains unclear. This study aimed to evaluate associations between EAT and proteomics, coronary microvascular dysfunction, cardiac structure and function, and quality of life in the PROMIS-HFpEF cohort.
**Methods:** EAT was measured by echocardiography in 182 patients (90% of 202 who underwent successful CFR testing) from the prospective, multinational PROMIS-HFpEF study. EAT was assessed around the right ventricular free wall in the parasternal long-axis view; increased EAT was defined as ≥9 mm. Proteins (266 unique biomarkers) were measured using three 92-plex high-throughput proximity extension assays (Olink Proseek Multiplex CVD II and III, and inflammation). Coronary microvascular dysfunction was evaluated with Doppler-based adenosine coronary flow reserve (CFR). Cardiac structure and function were assessed by comprehensive echocardiography, and quality of life by Kansas City Cardiomyopathy Questionnaire overall summary score (KCCQ-OSS). Statistical analyses included Mann-Whitney U tests, Fisher's exact test, Pearson's correlation, linear regression (crude and BMI-adjusted), and orthogonal partial least squares discriminant analysis (OPLS-DA) on 579 variables from 172 participants.
**Key Results:** Median EAT was 7 (IQR 6–9) mm; 54 patients (30%) had increased EAT (≥9 mm). EAT correlated moderately with BMI (r=0.49, p<0.001). Patients with increased EAT were younger (73 vs. 76 years, p=0.026), had higher BMI (32 vs. 27 kg/m², p<0.001), larger waist circumference (108 vs. 96 cm, p<0.001), higher HbA1c (44 vs. 40 mmol/mol, p=0.049), higher triglycerides (1.3 vs. 1.0 mmol/L, p=0.008), lower HDL (1.2 vs. 1.4 mmol/L, p=0.006), and lower NT-proBNP (466 vs. 1120 pg/ml, p<0.001). On echocardiography, increased EAT patients had smaller indexed LV end-diastolic volume (37 vs. 43 ml/m², p=0.002), smaller indexed LA volume (33 vs. 39 ml/m², p=0.004), higher EF (62% vs. 58%, p=0.019), higher stroke volume (78 vs. 67 ml, p=0.001), and longer mitral deceleration time (200 vs. 172 ms, p<0.001). Non-indexed LV/LA volumes did not differ. After BMI adjustment, only septal wall thickness (coefficient 1.02, 95% CI 1.00–1.04, p=0.018) and mitral E wave deceleration time (coefficient 1.03, 95% CI 1.01–1.05, p=0.005) remained significantly associated with EAT. CFR did not differ between groups (2.0 vs. 2.1, p=0.49). KCCQ-OSS tended to be lower in the increased EAT group (62 vs. 71, p=0.07), and was significantly lower in the EF ≥50% sub-analysis (p=0.04). OPLS-DA identified the top variables associated with EAT ≥9 mm as BMI, weight, waist circumference, and leptin; top variables associated with EAT <9 mm were IGFBP-1, BNP, IGFBP2, and NT-proBNP. Proteomic markers associated with increased EAT included adipokines (IL1-RA, leptin, FABP4, chemerin), markers of insulin resistance and endothelial dysfunction (lower IGFBP1 and IGFBP2), and dyslipidaemia (higher LDL receptor, lower HDL). Proteins more selectively associated with EAT ≥9 mm (vs. general obesity) included higher GLO1, 4E-BP1, ADAM-TS13, STAMPB, and AXIN1, and lower MMP-2 and Notch 3.
**Clinical Implications:** Increased EAT is common in HFpEF (30%) and is associated with an obese, pre-diabetic, dyslipidaemic phenotype with cardiac structural alterations (LV hypertrophy, diastolic dysfunction) and a distinct proteomic profile reflecting adiposity, inflammation, insulin resistance, and endothelial dysfunction. The lack of association with CFR suggests coronary microvascular dysfunction may not be the primary mechanism linking EAT to HFpEF. The attenuation of most associations after BMI adjustment indicates that EAT effects are largely mediated by general obesity, though potential local mechanical or paracrine effects cannot be excluded. Lower NT-proBNP despite greater diuretic use in the high EAT group raises concern for natriuretic peptide deficiency and potential underestimation of congestion in obese HFpEF patients. These findings support EAT as a marker of the obesity-related HFpEF phenotype but do not establish a causal role independent of obesity.