Author Response to: “Epicardial and Pericardial Fat: Separated but Under the Same Roof”

We appreciated the letter by Dr. Iacobellis in response to our study that assessed the clinical significance and genetic determinants of epicardial and pericardial adiposity in UK Biobank¹. Dr. Iacobellis raised important points that are relevant for interpreting our findings.

First, we would note that there is considerable variation in terminology for adipose tissue surrounding the heart. We therefore adopted the recommendations of the 2023 European Society of Cardiology Working Group on Coronary Pathophysiology and Micro-circulation consensus statement².

As Dr. Iacobellis emphasizes, epicardial adipose tissue (EAT) and pericardial adipose tissue (PAT) are distinct adipose tissue compartments. PAT is located external to the pericardium, whereas EAT is in immediate proximity to the myocardium, surrounds the coronary arteries, and may have distinct physiological functions².

Accurate differentiation between EAT and PAT is challenging in large magnetic resonance imaging datasets. To avoid bias—particularly in individuals with small amounts of adipose tissue—we segmented the combined area of EAT and PAT (termed EPAT for clarity) and evaluated its clinical and genetic associations. We highlighted our inability to differentiate between EAT and PAT as the primary limitation of our study. We note that a similar approach has been used in other large cohort studies^3-5, one of which reported a high correlation (r>0.9) between EAT and EPAT volumes³.

These earlier studies have reported significant associations between EPAT quantity and common cardiovascular conditions^3-5, and such findings are cited in reviews discussing the clinical relevance of EAT and PAT⁶. When analyzing EPAT without accounting for VAT, we observed similar findings. However, we then performed a systematic assessment of EPAT’s cardiovascular disease associations before and after adjustment for VAT. We found that the patterns of cardiovascular disease associations for EPAT resembled those for VAT and were no longer significant after adjustment for VAT. We believe that accounting for VAT adds useful context for interpreting these observational findings.

The genetic associations observed for EPAT paint a concordant but more complex picture. We found that EPAT was highly genetically correlated with VAT when considering all variants across the genome. Interestingly, we found relatively stronger genetic signals for EPAT than for VAT at the loci near WARS2, TRIB2 and EBF1. Whether these differences in association patterns were driven by EAT—with the PAT component of EPAT effectively acting as noise for the EAT signal—is an intriguing possibility. However, the specific roles of these loci with respect to disease risk remain to be determined.

We fully agree with Dr. Iacobellis that precise phenotyping is essential to understand the roles of different adipose tissue compartments in cardiovascular disease pathophysiology. We anticipate that future datasets or analytical approaches may enable follow-up studies of distinct EAT and PAT measurements in large and deeply phenotyped populations. It is possible that clinical or genetic associations could diverge between these two adipose tissue compartments. Our findings imply that it will be important for future studies to report association estimates before and after adjustment for VAT to distinguish local from global adipose relationships.