Epicardial adipose tissue: When good fat goes bad

Epicardial adipose tissue (EAT) has rapidly emerged as a focus of cardiovascular research, with growing recognition that this metabolically active fat depot plays an important role in cardiac physiology and pathophysiology. Unlike other visceral adipose stores, EAT is uniquely positioned between the visceral pericardium and the myocardium (Figure), lacking a fascial barrier between and sharing a microcirculation with the myocardium. This anatomic proximity facilitates direct biochemical crosstalk between adipocytes and cardiomyocytes. Under normal physiological conditions, EAT has a supportive role, acting as an energy source for the myocardium and secreting adipokines that exert anti-inflammatory and anti-fibrotic effects [1]. However, under systemic stressors such as obesity, diabetes, or chronic inflammation, EAT undergoes maladaptive remodeling characterized by both volumetric expansion and a shift toward a pro-inflammatory, pro-fibrotic phenotype [1]. These changes promote myocardial inflammation, interstitial fibrosis, and impaired relaxation—processes that have been increasingly linked to the pathogenesis of heart failure with preserved ejection fraction (HFpEF) [1].

Fig.

Epicardial adipose tissue (EAT) has rapidly emerged as a focus of cardiovascular research, with growing recognition that this metabolically active fat depot plays an important role in cardiac physiology and pathophysiology. Unlike other visceral adipose stores, EAT is uniquely positioned between the visceral pericardium and the myocardium (see figure), lacking a fascial barrier between and sharing a microcirculation with the myocardium.

In recent years, interest in EAT has expanded beyond mere recognition and description to mechanistic hypotheses that position it as a driver, rather than a bystander, in the development of multiple cardiovascular diseases. Observational studies have linked greater EAT volumes with impaired exercise capacity in HFpEF and worse clinical outcomes independent of traditional cardiovascular risk factors [2]. In the setting of obesity, patients with HFpEF consistently demonstrate greater EAT volumes compared with both healthy controls and patients with heart failure with reduced ejection fraction (HFrEF). This excess EAT appears to contribute to impaired ventricular compliance by exerting direct pericardial restraint, amplifying ventricular interdependence, and worsening hemodynamic responses to stress. Invasive hemodynamic studies have further demonstrated associations between increased EAT and elevated right atrial and pulmonary capillary wedge pressures, findings that help explain the profound exertional intolerance that typifies obesity-related HFpEF [3]. Importantly, the contribution of EAT extends beyond passive mechanical effects. A growing body of experimental evidence demonstrates that EAT secretes a broad range of pro-inflammatory cytokines, adipokines, and pro-fibrotic mediators which diffuse directly into adjacent myocardium via shared microcirculation [1]. This paracrine signaling shifts the myocardial environment from one that is metabolically supportive to one that is inflammatory and fibrotic, thereby driving diastolic dysfunction and adverse remodeling [1]. Beyond its relationship with HFpEF, EAT has been implicated in a variety of other cardiovascular pathologies. Meta-analyses have shown that EAT volume is independently associated with coronary artery disease and with high-risk plaque features such as low attenuation plaque [4]. Similarly, EAT has been linked to atrial fibrillation (AF) with larger EAT volumes associated with both the incidence of AF and higher recurrence rates following catheter ablation [5]. Ventricular arrhythmias have also been associated with greater EAT burden, both in patients with structurally normal hearts and in those with established systolic dysfunction, further highlighting its arrhythmogenic potential [6]. Collectively, these findings reinforce the concept that EAT is not simply an epiphenomenon of systemic adiposity, but a pathogenic depot capable of modulating cardiac structure, function, and electrophysiology. This evolving body of work has created momentum for EAT to be considered not just a biomarker of systemic metabolic health, but a potential therapeutic target in its own right. There is emerging evidence that pharmacotherapies for obesity and diabetes—namely, glucagon-like peptide 1 receptor agonists and sodium-glucose cotransporter 2 inhibitors that have shown benefits in cardiovascular diseases—can reduce epicardial and pericardial adipose tissue (PeAT) volumes [7], [8]. Whether these structural changes mediate part of their cardiovascular benefit remains speculative but represents an exciting avenue for ongoing investigation in the field.

While the body of knowledge continues to grow rapidly, much remains to be learned about not only EAT’s role in these pathologic processes but also in how best to evaluate and quantify it—questions that remain unanswered at this time. The study by Song et al. in this issue of JCMR [9] leverages cardiovascular magnetic resonance (CMR) to interrogate the interplay between EAT and biventricular function across the major heart failure phenotypes. In this study, EAT burden was compared across HFpEF, HFrEF, and control populations using CMR, while also interrogating the relationship between EAT and biventricular function assessed by myocardial strain. Over 500 patients were included, all of whom underwent standardized 3.0T CMR. EAT was segmented from short-axis cine sequences at end-diastole with careful delineation of the visceral pericardial boundary from apex to atria. This approach allowed for volumetric quantification of EAT, as well as the visceral adipose tissue external to the pericardium termed PeAT. These two depots were then summed to calculate the paracardiac adipose tissue. Biventricular strain parameters were then derived using feature-tracking.

The results reinforce several emerging themes. As expected, patients with HFpEF exhibited significantly greater EAT volumes compared to both HFrEF and control groups (51 ± 21 mL vs 32 ± 14 mL and 33 ± 19 mL, respectively). This difference was notable even though the mean body mass index (BMI) in the HFpEF group was only 25.6 kg/m², underscoring the limitation of BMI as a surrogate for visceral adiposity which is near universally present in HFpEF. In contrast, EAT volumes in HFrEF were not only lower than in HFpEF but also trended slightly below controls, consistent with prior reports suggesting that advanced HFrEF may be characterized by systemic cachexia, muscle wasting, and fat loss [10]. When related to function, higher EAT volumes in HFpEF were modestly but significantly associated with worse left ventricle (LV) and right ventricle (RV) global longitudinal strain, reinforcing the concept that EAT expansion in this population exerts adverse mechanical and biochemical effects on both ventricles. Including the RV is important, as RV dysfunction appears to be a key determinant of symptoms and outcomes in HFpEF, and emerging data suggest that pericardial and epicardial fat adjacent to the RV may further modulate right-sided mechanics and pulmonary vascular interactions. By showing correlations between increasing EAT and worsening RV global longitudinal strain in HFpEF, this study strengthens the argument that pericardiac adiposity exerts biventricular effects and that focusing exclusively on left-sided measures risks missing clinically relevant biology. Interestingly, in HFrEF, the inverse of this relationship was found: greater EAT correlated with relatively better LV and RV strain. This divergence supports the hypothesis that EAT can have both supportive and pathologic functions depending on the surrounding substrate and may differ across heart failure phenotypes.

While the study provides important data, it also lends several opportunities for further discussion. First, accurate delineation and segmentation of EAT can be very challenging, introducing the opportunity for interobserver variability as evidenced by the widely varying EAT volumes in the literature. The thin visceral pericardium separating EAT from PeAT can be difficult to visualize even under favorable imaging conditions. Its visualization can be further complicated by artifacts from cardiac or respiratory motion, implanted cardiac devices, and sternotomy wires, among others. Beyond acquisition, differences in post-processing software, fat–water signal separation, and inclusion or exclusion of adjacent and perivascular fat can all affect the measured volumes, making direct comparisons across cohorts or imaging platforms challenging. Next, the optimal method for measuring EAT remains unsettled. Measurements taken at end-diastole when there is the least cardiac motion are the most widely reported method. However, these require measuring the already small target during the phase when it is the smallest, whereas end-systolic measurements allow for a larger target area at the risk of more blurring from motion. Additionally, there are advantages and disadvantages to both short- and long-axis measurements—both of which have been utilized in published studies. Third, it requires significant time and effort to manually segment accurately. While post-processing software is quickly adapting with new tools, many of the studies thus far have had to employ tools built for myocardial volumetric analysis which further increases the time and difficulty of accurate segmentation. Ideally, in the future, more widely available pericardial segmentation tools will be available. Automated deep learning–based segmentation tools are in development that could reduce the analytic burden and improve reproducibility, hopefully paving the way for EAT quantification to move beyond research settings into routine clinical practice [11]. Despite these challenges, the authors in this study report EAT volumes similar to those of prior studies where EAT was segmented accurately with careful exclusion of fat and fluid external to the visceral pericardium [7].

The findings from this paper should be viewed as both an important advance and a call to action for future research. A first priority is methodological standardization. At present, there is no universally accepted approach to defining and quantifying EAT. CMR offers radiation-free volumetric assessment but technical variability in sequence choice, segmentation boundaries, and lack of efficient and accurate post-processing algorithms continues to limit comparability across studies. Additionally, as discussed recently by Lobeek et al. [12], the nomenclature for these adipose depots varies widely across literature with some studies using epicardial, pericardiac, paracardiac, and mediastinal adipose tissue interchangeably. While much of this nomenclature remains to be agreed upon, one fact is certain: EAT should only be reported (as it was in this study) as the adipose tissue within the visceral pericardium. A consensus on standardized nomenclature, acquisition protocols, segmentation strategies, and reporting standards for these depots will be important to improve our ability to accurately assess, report, and compare results across future studies. Establishing a shared foundation will be essential to ensure that future research can build effectively. Already, investigations are moving beyond simple quantification of EAT toward more detailed assessments of its quality and tissue characteristics, with the goal of better identifying where EAT is along the spectrum from physiologic to pathologic [13], [14], [15].

In summary, the study by Song et al. provides timely evidence that EAT is more than a passive marker of systemic obesity: it is an active participant in the pathophysiology of heart failure, exerting divergent effects across phenotypes. In HFpEF, EAT appears to be a maladaptive depot, expanding in volume and secreting inflammatory mediators that impair both LV and RV function. In HFrEF, in contrast, diminished EAT volume may signal metabolic exhaustion and correlate with more advanced ventricular dysfunction, raising the possibility that preserved EAT reflects resilience rather than risk. Next steps should focus on unifying our measurements, nomenclature, and reporting on EAT so that future investigations can easily be compared and continue to build on our understanding of this important contributor to cardiovascular disease.

Funding

J.C. is supported by NIH grant T32EB003841. A.R.P. is supported by NIH/NHLBI R01HL181257 and has research grants from GE Healthcare and Area19. He has research support from Neosoft, Siemens Healthineers, and CircleCVI. C.W.T. is supported by NIH/NHLBI R01HL155717.

Declaration of competing interests

Jamey Cutts reports a relationship with the National Institutes of Health (NIH) that includes funding grants. Amit Patel reports a relationship with NIH that includes funding grants. Connie Tsao reports a relationship with NIH that includes funding grants. The other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.