Abstract Body: Introduction: Epicardial adipose tissue (EAT) has emerged as a potential biomarker and therapeutic target in heart failure, but its relationship across HF phenotypes remains incompletely characterized.

Hypothesis: We hypothesized that EAT differs across HFpEF, HFrEF, and HFmrEF phenotypes, reflecting distinct pathophysiological roles.

Aim: To update the Wu et al. 2023 meta-analysis examining EAT across HF phenotypes with newly published data.

Methods: We searched PubMed, Embase, and Scopus till October 2025). Observational studies comparing EAT between HF patients and controls were included, stratified by phenotype. Random-effects meta-analysis with Hartung-Knapp adjustment was performed with subgroup analyses by imaging modality and temporal trends.

Results: Thirty-one studies (4,037 HF patients) were analyzed. EAT was significantly elevated in HFpEF versus controls (SMD 1.17; 95% CI: 0.76-1.59; p<0.0001; k=22) but not in HFrEF (SMD -0.36; p=0.18) or HFmrEF (SMD -0.21; p=0.69). EAT correlated negatively with function in HFpEF (r=-0.55) but positively in HFrEF (r=+0.57). Echocardiography yielded larger effects than CMR (p=0.0007). Findings persisted after trim-and-fill adjustment (SMD 0.74; 95% CI: 0.24-1.23).

Conclusion: EAT is phenotype-specifically elevated in HFpEF with divergent functional correlations across phenotypes, supporting EAT as a biomarker and therapeutic target for metabolic HFpEF.