Epicardial adipose tissue (EAT) lies between the myocardium and the epicardium, and is a type of adipose tissue with endocrine and systemic metabolic regulatory functions (1). EAT not only provides mechanical protection for arteries and nerves and energy supply to the myocardium, but also buffers against fatty acid toxicity levels (2). However, because there is no physical barrier separating EAT from the myocardial surface and it shares the same microcirculation, adipose tissue can easily penetrate the myocardium and interact directly with it as EAT expands and infiltrates. EAT secretes large amounts of adipokines and active products such as adrenomedullin and lipocalin (3, 4), leading to myocardial lipotoxicity, increased myocardial fibrosis, insulin resistance, and pathologically pro-inflammatory transformations (5–8), thus participating in heart failure (HF) process.

The pathophysiological mechanism pertaining to EAT in HF with reduced ejection fraction (HFrEF) vs. preserved ejection fraction (HFpEF) is still being studied. But numerous studies have shown that EAT mass or volume indices are lower in HFrEF than healthy controls, while EAT mass is higher in HFpEF (9–12). And patient EAT is associated with impaired exercise capacity (13) and left ventricular dysfunction (10–12, 14). In one study, increased EAT in HFrEF was associated with a lower risk of adverse events (AEs), whereas increased EAT in HFpEF was associated with a higher risk (13). However, it remains controversial whether EAT itself is associated with adverse outcomes in different types of HF, or whether these associations show opposite trends. In addition, differences in EAT measurements protocols, as well as EAT metrics and their thresholds for heterogeneity, are not yet clear.

Therefore, the study aimed at evaluating whether EAT volume or thickness, as an imaging biomarker, can predict poor prognosis in patients with different types of HF and to provide comprehensive data on the heterogeneity of EAT measurement protocols and metrics originating from different studies.

This study was reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines statement (15). The study protocol has been registered in PROSPERO (registry No.: CRD420250653252).

In our systematic review and meta-analysis of nine cohort studies involving 1,939 HF patients and 329 control populations, We analyzed HR data between EAT and adverse outcomes in HF patients. The primary findings indicate that in the analysis of the primary outcome, an increase of EAT exhibited an significant relevance to the primary outcomes in HF patients. In subgroup analyses, EAT was still related to the risk of the primary outcome in patients with EF ≥50%, while did not exhibit any relevance to that in patients with EF <50%. And the AUC value for predicting the primary outcome based on EAT was high in patients with EF >40%. In the analysis of secondary outcomes, EAT presented no correlation with the risk of the composite outcome of cardiovascular death and hospitalization for HF patients with EF >40% and EF ≤40%, while increased EAT in HF patients indicated a higher possibility of readmission for HF.

EAT has a protective brown fat profile and is able to act as a metabolic reservoir for HFrEF (24). In times of high metabolic demand, such as cardiac ischemia, EAT can help improve cardiac function and contractility through free fatty acids as local energy sources (25). And the pathophysiological pathways associated with increased metabolism in patients with HFrEF are upregulated (26). Thus, the thinner EAT may reflect the potential increased energy demand of the heart in patients with HFrEF. Lower EAT in HFrEF patients also correlates significantly with poorer systolic function, myocardial damage, and cardiac remodeling (27). In HFpEF, significant EAT accumulation may lead to myocardial steatosis or adipose tissue infiltration of adjacent myocardium (28), and secretion of a wide range of pathopathological, pro-inflammatory, and atherosclerotic adipogenic factors through paracrine, endocrine, and vascular pathways (29), thereby contributing to hemodynamic derangement (30–32), myocardial fibrosis, and sclerosis (29), leading to cardiac impairment and poor prognosis (9, 33). In addition, large accumulations of EAT may also exert mechanical compression on the heart, thereby limiting encapsulated myocardial expansion, which can lead to diastolic dysfunction and increase cardiac filling pressures (30, 34). Although the mechanism between EAT and impaired prognosis are not fully understood in HFrEF and HFpEF patients, the role of EAT in the pathophysiologic process of HF shows the possibility that quantitative detection of EAT may independently predict HF patients' poor prognosis in clinical practice. In fact, EAT is related to prognosis in patients suffering from coronary artery disease and type 2 diabetes as well as in individuals who do not have cardiovascular disease (35–37).

At present, the prognostic assessment of HF is mainly based on independent and routine risk factors such as demographic characteristics, biomarkers, and imaging parameters. Kenneth et al. (38) used age, New York Heart Association functional class, atrial fibrillation, chronic obstructive pulmonary disease, chronic kidney disease, LVEF, and diabetes mellitus to predict all-cause mortality in HFrEF, and the C-statistics of this model were 0.75 and 0.74 in the development and validation cohorts, respectively. Komajda et al. (39) showed that N-terminal pro-B-type natriuretic peptide, age, diabetes mellitus, and LVEF were the strongest independent factors for all-cause mortality in HFpEF, with a C-statistic of 0.736 for the constructed model. It can be seen that models constructed by these factors have only moderate clinical performance in predicting AEs and that different types of HF may require different clinical predictors to predict prognosis. Our meta-analysis indicated that EAT accumulation could trigger a higher risk of the primary outcome in HF patients. And in the analysis of EAT and HF readmission events, although the results of Liu et al. (20) and Lin et al. (21) were inconsistent, our results found an intense relevance of EAT to HF readmission risk. In line with this, a prospective cohort study showed that pericardial fat hinted a higher risk in HF patients, particularly HFpEF patients (35). This suggests that EAT may be able to be used as a useful parameter for assessing the clinical prognosis of HF, but the predictive value of EAT for poor prognosis in the entire heart failure cohort is not known. In HF patients with EF ≥50%, EAT similarly demonstrated an association with a higher risk of the primary outcome and showed higher pooled AUC values in predicting outcomes. Additionaly, Wang et al. (16) engaged in further validating the incremental predictive value of EAT for baseline model performance by adding it to the baseline multivariate model. The results show that EAT can significantly improve the performance of predicting the primary outcome (net weight classification index improvement by 8.4%). Thus, EAT holds promise as a clinical predictor for HF patients with EF ≥50%, contributing to the development of more effective predictive models that can significantly improve risk prediction.

However, the association between EAT and the primary outcome does not exist for those with EF <50%. This may be due to the fact that the positive effects of EAT thickening on ventricular structure and function are counterbalanced by an increased propensity for ventricular arrhythmias (40). Inconsistent with our findings, Paris et al. (18) included 69 systolic HF patients who received ICD implantation, and showed that an increase of EAT thickness was associated with HF hospitalization risk and HF-related mortality. Pugliese et al. (13) argued that a decrease in EAT thickness was significantly associated with an adverse prognosis in HFrEF patients. All these findings differ from the result of our study, which may be due to the inconsistency of the included patient populations or the inclusion of both HFrEF and HFmrEF patients in our pooled results. Thus, the predictive role of EAT for main outcome in HF patients with EF <50% is debated and further clinical studies are still needed. In addition, This meta-analysis showed no significant correlation between EAT and the composite outcome of cardiovascular death and HF hospitalization, no matter HFrEF or HFpEF patients. In contrast, Both Van woerden et al. (17) and Lin et al. (21) showed a correlation between increased EAT and cardiovascular death and HF hospitalization in HFpEF. However, according to a correlation analysis of cardiovascular mortality outcomes in the study by Lin et al. (21), EAT did not relate to the patient's risk. Therefore, the relevance of EAT to the composite outcomes of cardiovascular death and HF hospitalization needs to be further studied in HFrEF and HFpEF.

Notably, the included studies in the analysis had a high degree of heterogeneity, which may be due to differences in study methodology, population characteristics, follow-up times, and outcome definitions and EAT measurement protocols. EAT may have prognostic differences across geographic regions and imaging modalities. All studies included were from patient populations in different countries and regions, and the patient populations studied also included different HF types. The study methodology was also inconsistent, containing prospective and retrospective studies. Most importantly, differences in EAT measurement protocols may be the main cause of heterogeneity. Some of the included studies used echocardiography to measure EAT thickness, while others used CMR and CCTA to measure EAT volume. And one study calculated the mean thickness over 3 cardiac cycles, while the other one considered 2 cardiac cycles. Some studies used a 1.5T CMR scanner to delineate and measure EAT volumes from the short-axis view, while others used 3T from the short-axis four-chamber view and measured EAT volumes in different units of measure, including ml, cm³, ml/cm². This observation highlights the lack of standardization of measurement techniques and EAT metrics, which challenges the application of EAT as a risk stratification tool. Thus, future studies require standardized measurement techniques and defined quantitative metrics and thresholds for EAT to enable accurate risk stratification.

The systematic review and meta-analysis here indicated the significant association of EAT to the risk of the composite outcome of the primary outcome and HF readmission in people with HF, but not to cardiovascular mortality and HF hospitalization. In addition, although the relevance of EAT to cardiovascular death and HF hospitalization is controversial in patients with HFpEF, their EAT triggers a higher risk of the primary outcome. The association of EAT with ACM, cardiovascular death, HF hospitalization, and their composite outcomes in patients with HFrEF still needs to be confirmed by further research. But It is important to note that there was high heterogeneity among the included studies, due to different EAT measurement protocols, a further step is needed to identify standardized measurement protocols and defined measurement units.