Association of epicardial adipose tissue with metabolic risk factors on cardiovascular outcomes: serial coronary computed tomography angiography study

Article information

Korean J Intern Med. 2024;39(2):283-294
Publication date (electronic) : 2024 February 14
doi : https://doi.org/10.3904/kjim.2023.389
1Department of Internal Medicine, Seoul National University Hospital, Seoul, Korea
2Seoul National University College of Medicine, Seoul, Korea
3Department of Internal Medicine, Seoul National University College of Medicine, Seoul, Korea
4Research and Science Division, Research and Development Center, MEDICALIP Co. Ltd., Seoul, Korea
5Division of Radiology, Seoul National University Bundang Hospital, Seongnam, Korea
Correspondence to Hae Young Lee, M.D., Ph.D., Division of Cardiology, Department of Internal Medicine, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul 03080, Korea Tel: +82-2-2072-0698, Fax: +82-2-3674-0805 E-mail: hylee612@snu.ac.kr
*

These authors contributed equally to this manuscript.

Received 2023 September 18; Revised 2023 November 26; Accepted 2023 November 29.

Abstract

Background/Aims

Epicardial adipose tissue (EAT) shares pathophysiological properties with other visceral fats and potentially triggers local inflammation. However, the association of EAT with cardiovascular disease (CVD) is still debatable. The study aimed to observe the changes and associations in EAT and risk factors over time, as well as to investigate whether EAT was associated with CVD.

Methods

A total of 762 participants from Seoul National University Hospital (SNUH) and SNUH Gangnam Center were included in this study. EAT was measured using coronary computed tomography angiography.

Results

Baseline EAT level was positively associated with body mass index (BMI), calcium score, atherosclerotic cardiovascular disease (ASCVD) 10-year risk score, glucose, triglycerides (TG)/high-density lipoprotein (HDL), but not with total cholesterol, low-density lipoprotein (LDL). At follow-up, EAT levels increased in all groups, with low EAT groups demonstrating a significant increase in EAT per year. Change in EAT was associated with a change in BMI, TG/HDL, and glucose, while changes in LDL, calcium score, and ASCVD 10-year risk score were not associated. Although calcium score and ASCVD 10-year risk score were associated with CVD events, baseline information of EAT, baseline EAT/body surface area, or EAT change was not available.

Conclusions

Metabolic risks, e.g., BMI, TG/HDL, and glucose, were associated with EAT change per year, whereas classical CVD risks, e.g., LDL, calcium score, and ASCVD 10-year risk score, were not. The actual CVD event was not associated with EAT volume, warranting future studies combining qualitative assessments with quantitative ones.

Graphical abstract

INTRODUCTION

Accumulation of adipose tissue, especially in the visceral cavity, plays an essential role in the development of several cardiovascular diseases (CVDs), including metabolic syndrome, hypertension [1], dyslipidemia, insulin resistance, as well as type 2 diabetes [2,3]. Specifically, visceral fat tissue is recognized as an independent risk factor for cardiovascular complications [4]. The amount of visceral fat has been reported to be associated with inflammatory-mediated cytokines such as plasminogen activator inhibitor-1, tumor necrosis factor-α, and interleukin-6 [5]. Moreover, visceral fat increases due to stimulated lipid uptake and esterification in visceral fat during the aging process [6].

Given its clinical significance, research has been directed towards quantifying fat amounts in various areas. Epicardial adipose tissue (EAT), a deposit of visceral fat located between the heart and the pericardium, is primarily composed of white adipose tissue with small amounts of brown adipose tissue [7]. EAT plays a multifaceted role for the heart, including providing mechanical protection for the coronary artery [8,9], contributing to myocardium energy through a high metabolism rate of free fatty acid [10], and performing thermogenic functions [11]. Additionally, EAT produces active biomarkers [12] that can influence heart morphology and function, and also releases micro-ribonucleic acids to the vascular wall, exerting paracrine effects [7].

Due to its proximity to the cardiovascular system, EAT might more be susceptible to pronounced inflammatory responses compared to typical visceral fats. Adipokines released from EAT can locally affect inflammation in the nearby coronary vessel wall and stimulate atherosclerosis progression through outside-to-inside signaling [13,14]. Furthermore, EAT has also been reported to be associated with coronary atherosclerosis and subclinical coronary artery disease [15-17]. Although research has suggested EAT’s involvement in metabolic risk factors and coronary calcification [18-20], the precise contribution of epicardial fat to CVD remains uncertain. Some studies demonstrated that anti-atherosclerotic cytokines, such as adiponectin, secreted from EAT promote cardioprotection [21,22]. Additionally, other studies highlight that pericardial fat, not EAT, is strongly associated with cardiometabolic risk and coronary atherosclerosis [23,24].

Coronary computed tomography angiography (CCTA) is a sensitive imaging modality that allows the accurate measurement of EAT volume with high spatial resolution [25]. However, many investigations focus solely on the cross-sectional view of EAT rather than the longitudinal changes, which can be measured using CCTA. Thus, our study aims to investigate EAT volume and its serial changes, in conjunction with other conventional risk factors, to determine its association with CVD events.

METHODS

Study population

The analysis included participants who had undergone CCTA at least twice, with intervals of 2 years or more, between January 2006 and December 2019 at Seoul National University Hospital (SNUH) and SNUH Gangnam Center. During this period, a total of 4,431 participants were extracted from the SNUH Patients Research Environment, which serves as a clinical data warehouse. Of the total participants, 814 individuals with analyzable images were selected for further analysis, while 52 were excluded for the following reasons: inadequate clinical information (n = 21), insufficient image quality (n = 18), history of percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG) (n = 11), and indistinct EAT delineation due to severe motion artifacts or pericardial effects (n = 2). Ultimately, a total of 762 participants were included in the study (Fig. 1).

Figure 1.

The flow of participants through the study. CABG, coronary artery bypass grafting; CCTA, coronary computed tomography angiography; EAT, epicardial adipose tissue; PCI, percutaneous coronary intervention.

CCTA scan protocol and evaluation

Baseline and follow-up CCTA images were retrieved from the Picture Archiving and Communication System of SNUH, stored in the Digital Imaging and Communications in Medicine format. These images were then downloaded and transferred to the Department of Radiology at Seoul National University Bundang Hospital, which served as the core lab for EAT volume measurements. Additionally, CCTA images were separately analyzed by two experienced radiology researchers using a dedicated workstation (SyngoVia; Siemens Healthcare, Erlangen, Germany). If discrepancies exceeding 10% of EAT volume were detected, a third measurement was performed. The analysis was conducted independently, without interference from the other researchers involved. All CCTA images were measured at the same cardiac cycle to ensure consistency, as the volume of EAT could vary between systolic and diastolic periods. Experienced researchers manually marked an epicardial line on CCTA images, setting thresholds at -190 Hounsfield unit (HU; lower) and -30 HU (upper). This marking was conducted on axial scans of 0.9 mm thickness, spanning carina to heart base. Moreover, EAT measurements, highlighted in purple, were semi-automated. Researchers checked and adjusted the outlined area for accuracy. Cases with unclear EAT lines caused by severe motion artifacts or pronounced pericardial effects (pericardial effusion exceeding 10 mm) were excluded. Additionally, images with insufficient scan range -starting below the carina level or not reaching the heart base-were also excluded, as they could impact EAT volume measurements. The EAT volume was computed by summing EAT measurements from each scanned image, as determined through the aforementioned methodology. To counter the influence of height and weight variations, EAT was normalized by dividing it by the body surface area (BSA) for further analysis. The annual EAT change was computed as (follow-up EAT - baseline EAT) / baseline EAT × 100 / year. For participants who underwent CCTA imaging more than three times within the study period, the initial and final CCTA images were chosen for analysis. A written signed informed consent was obtained from each patient and the research was carried out with approval from Institutional Review Board of SNUH (IRB no H-2022-174-1107).

Clinical data acquisition

Clinical data collection for these participants was manually secured through Electronic Medical Records (EMR). Utilizing EMR at SNUH, various clinical details were gathered, including information from CCTA and laboratory data such as complete blood count (CBC) and medication history. This comprehensive data compilation was aimed at assessing each participant’s cardiovascular risk. Additional medical records were collected, encompassing sex, age, body mass index (BMI), BSA, systolic blood pressure, diastolic blood pressure, pulse rate, and lifestyle information such as smoking and drinking histories. Medical conditions, including diagnosed hypertension and diabetes, were also included. Moreover, participants’ cardiovascular family histories, comprising coronary artery disease, angina, and stroke, were gathered. Cases of hypertension and diabetes were restricted to those diagnosed or under medication while also excluding individuals with a family history of hypertension and diabetes. Smoking history was categorized as never smoked, ex-smoker, or current smoker while drinking history was classified as none, social drinker, or heavy drinker. Measurements for CBC, glycated hemoglobin (HbA1c), total cholesterol, triglyceride (TG), high-density lipoprotein (HDL), and low-density lipoprotein (LDL) were obtained. Data points obtained within a month prior to the timing of the two CCTA scans were chosen. In instances where this data was not available, the previous data was employed using the principle of last observation carried forward. In cases where no data was present, those instances were treated as missing values.

Lastly, a review was conducted to determine whether coronary artery disease had manifested during the follow-up period. CVD events and new cases of coronary artery disease were identified based on the presence of 50% stenosis in the CCTA findings at the follow-up assessment or if coronary artery lesions exhibited progression during the follow-up period and necessitated either PCI or CABG.

Statistical analysis

Data are presented as mean ± standard deviation, or a number with percentages, or as an odds ratio (OR) with a 95% confidence interval (CI). The Agatston score for calcium was categorized as follows: 0 HU, 0.1–100 HU, 100.1–400 HU, and > 400 HU [26]. The atherosclerotic cardiovascular disease (ASCVD) 10-year risk score percentages were classified into the following categories: 0–5%, 5–7.5%, 7.5–20%, and > 20% [27].

To analyze the relationship between EAT (divided into quartiles) and risk factors, t-tests and Mann–Whitney U-tests were employed. The association between changes in EAT and risk factors over time was examined using simple regression models, receiver operating characteristic curve (ROC) curve analysis, and Pearson’s chi-square test. Statistical significance was determined as p < 0.05. All analyses were performed using the SPSS software package (version 26.0; IBM Corp., Armonk, NY, USA).

RESULTS

Baseline analysis

A total of 762 participants (age 58.7 ± 6.6 yr, 91.5% males) were included (Table 1). The interval duration between the two computed tomography (CT) scans was 2,237 ± 984 days. Participants were grouped into quartiles according to the baseline EAT volume, < 84.00 mL, 84.00–107.77 mL, 107.77–135.39 mL, and > 135.39 mL, respectively. Approximately, 42.7% of the participants were obese with a BMI of 24.5 ± 2.6 kg/m2, and the prevalence of hypertension and diabetes mellitus was 42.3% (p < 0.001) and 17.8% (p = 0.033), respectively. A significant difference was observed between Q1 and Q4 for hypertension (28.8% vs. 51.8%, respectively) and diabetes (11.5% vs. 19.9%, respectively). Height, weight, and BMI exhibited a positive association with EAT volume at baseline. The association between smoking and EAT displayed no significance (p = 0.302), while for alcohol consumption (p = 0.012), heavy drinkers were prevalent in Q4.

Clinical characteristics of participants at baseline

Classical risk indicators, such as coronary calcium score (p = 0.013) and the ASCVD 10-year risk score, exhibited a significant correlation with EAT. From laboratory results, HbA1c, glucose, TGs, HDL, and TG/HDL were associated with EAT (all p < 0.001), while total cholesterol (p = 0.294) and LDL (p = 0.569) were not. Among these, except for HDL, which demonstrated a negative association, HbA1c, glucose, and TGs, TG/HDL exhibited a positive association with EAT volume.

EAT characteristics at baseline and follow-up

At baseline, the average EAT was 113.1 ± 39.2 mL, with EAT/BSA at 62.1 ± 19.7 mL/m2 (Table 2). During follow-up, both EAT and EAT/BSA increased to 124.6 ± 44.9 mL and 68.6 ± 22.5 mL/m2, respectively. Weight and BMI displayed no significant change between the baseline and follow-up measurement (weight: from 70.1 ± 9.3 kg to 69.8 ± 9.5 kg; BMI: from 24.5 ± 2.6 kg/m2 to 24.5 ± 2.5 kg/m2). Unlike weight, BMI, and BSA remained relatively unchanged (BSA: from 1.8 ± 0.2 m2 to 1.8 ± 0.1 m2). Furthermore, all groups exhibited an increased mean EAT and EAT/BSA at follow-up. Significant differences in BSA were observed among each quartile at both baseline and follow-up, except for between Q2 and Q3 in both instances.

Changes of EAT volume over time

Groups with initially low EAT levels exhibited a pronounced yearly increase in EAT. Specifically, Q1 demonstrated a change of 2.7 ± 4.5% per year, while Q4 displayed a change of 1.3 ± 2.9% per year. Upon comparison among the groups, statistically significant differences were observed solely in EAT changes between Q1 and Q3/4. The p-value for the comparison between Q1 and Q3 was 0.012, and between Q1 and Q4, the value was 0.003.

Association between the changes in EAT and risk factors over time

The participant population was divided according to the EAT change per year; +5%/yr and -5%/yr were selected arbitrarily as criteria for clinical and statistical importance. Clinical characteristics of participants according to EAT change per year were compared (Table 3). EAT increased in parallel with BMI, TG/HDL, and glucose over time (all p < 0.001). Other risk factors did not exhibit a significant difference. Moreover, BMI, TG/HDL, and glucose were considered significant risk factors, and LDL, calcium score, and ASCVD 10-year risk score were selected as comparative risk factors for simple regression analysis. The association and correlation coefficients (R2) between EAT change per year and BMI, TG/HDL, glucose, LDL, calcium score, and ASCVD 10-year risk score are illustrated by scatter plots (Fig. 2). Although BMI, TG/HDL, and glucose demonstrated significant association with EAT change, R2 of the single parameter was very low.

Clinical characteristics of participants according to EAT change per year

Figure 2.

Impact of EAT change per year on change of clinical characteristics (Scatter plots). Scatter plots display associations between EAT change per year and BMI (A), TG/HDL (C), and glucose change (E). In comparison, they did not display a correlation between EAT change per year and LDL (B), calcium score (D), and ASCVD 10-year risk score change (F). R2 and p values were calculated through simple regression analysis. ASCVD, atherosclerotic cardiovascular disease; BMI, body mass index; EAT, epicardial adipose tissue; HDL, high-density lipoprotein; HU, Hounsfield unit; LDL, low-density lipoprotein; TG, triglyceride.

Association with CVD event

We analyzed the association between baseline EAT, follow-up EAT, EAT/BSA, and CVD events stratified by statin use. However, no statistically significant relationship was observed between changes in EAT and the occurrence of CVD events based on statin use (Table 4). To compare the risk-predicting power of EAT, the association between CVD event and calcium score, ASCVD 10-year risk score, baseline EAT, baseline EAT/BSA, and change of EAT per year was plotted as ROC curve (Fig. 3). Only the calcium score and ASCVD 10-year risk score demonstrated significant predictability (area under the curve: 0.753 vs. 0.574, p < 0.001 for calcium score and p = 0.007 for ASCVD 10-year risk score, respectively). Neither baseline EAT nor EAT change over time displayed any significant predictability of CVD events. According to the baseline characteristics, an increase in EAT quartile was associated with an increase in calcium score. Subsequently, we conducted multiple regression analyses based on the change in calcium score, confirming that HbA1c, weight, baseline TG, and age were independent predictors of calcium score change (β > 0, all p < 0.05) (Supplementary Table 1).

Association between EAT and cardiovascular disease events in terms of statin usage

Figure 3.

ROC curve for the impact of clinical characteristics on CVD event. ASCVD, atherosclerotic cardiovascular disease; AUC, area under the curve; BSA, body surface area; CVD, cardiovascular disease; CI, confidence interval; EAT, epicardial adipose tissue; ROC, receiver operating characteristic.

Lastly, to determine if adding EAT to an existing risk prediction engine has additional effects, the CVD event rate was re-analyzed by dividing the change of EAT into three groups according to calcium score and ASCVD 10-year risk score (Fig. 4). The re-analyzed group with the lowest calcium score displayed a p value of 0.508, followed by > 0.999 and 0.673 for the intermediate and highest groups, respectively. Similarly, for the ASCVD 10-year risk score, the lowest, intermediate, and highest re-analyzed groups had p values of 0.585, > 0.999, and > 0.999, respectively. In summary, no additive effect of EAT change per year on top of the conventional risk prediction markers was observed in any group.

Figure 4.

CVD event rate with EAT change per year grouped by risk prediction markers. Analyzed CVD event and EAT change per year grouped by calcium score (A) and ASCVD 10-year risk score (B). ASCVD, atherosclerotic cardiovascular disease; CVD, cardiovascular disease; EAT, epicardial adipose tissue; HU, Hounsfield units.

DISCUSSION

The current investigation validated the precise measurement of EAT volume and its sequential changes through CCTA. Notably, the study demonstrated the association with other conventional clinical parameters. The key conclusion can be summarized as follows: 1) Individuals with unfavorable risk factors displayed a larger EAT volume in comparison to their counterparts. 2) Variations in BMI, TG/HDL ratio, and glucose levels were associated with annual changes in EAT. In contrast, no evident connection was observed between LDL levels, calcium score, ASCVD 10-year risk score, and the aforementioned changes. 3) No established link existed between CVD and parameters related to EAT, emphasizing the necessity for future research integrating in-depth assessments of EAT’s qualitative attributes. 4) The calcium score emerged as a predictive factor for CVD events, and HbA1c, weight, baseline TG, and age exhibited positive correlations with both EAT quartile and calcium score changes.

At baseline analysis, BMI, calcium score, and ASCVD 10-year risk score demonstrated a positive association with EAT volume. Among the laboratory results, HbA1c, glucose, TGs, and TG/HDL displayed a positive association with EAT volume, while HDL exhibited a negative association, and LDL displayed no correlation. Participants with unfavorable metabolic risk factors were classified into the high EAT quartiles. In this study, we conducted a detailed re-analysis. EAT was normalized by dividing it by BSA, thus minimizing variations related to height and weight. According to quartiles, EAT/BSA demonstrated a positive correlation with EAT volumes. In the follow-up phase, the patterns persisted, with those in the low EAT quartiles still displaying the lowest EAT levels. However, when assessing EAT change as a percentage per year relative to baseline, all groups exhibited positive mean changes. Remarkably, the groups with the highest EAT had the smallest increase. In EAT change analysis, risk factors like BMI, TG/HDL, and glucose decreased in groups exceeding 5% EAT loss but increased in those with over 5% EAT gain. Simple regression analysis confirmed these patterns. However, LDL, calcium score, and ASCVD 10-year risk score exhibited no such associations. Additionally, BMI, TG/HDL, and glucose emerged as meaningful indicators of EAT change, hinting at their role in reflecting the impact of insulin resistance on EAT volume. The lack of correlation with LDL suggests additional mechanisms influencing EAT changes. This influence emphasizes the connection between EAT and medical conditions such as diabetes, obesity, and elevated TG levels, suggesting an increased likelihood of metabolic syndrome. The potential role of EAT in predicting metabolic syndrome is noteworthy, especially in patients undergoing CCTA for health screening. Therefore, the measurement of EAT in CCTA could function as an additional benefit for evaluating the risk of developing the metabolic syndrome.

In this study, classical risk indicators, calcium score, and ASCVD 10-year risk score exhibited significant associations with CVD events, with the calcium score demonstrating the most notable link, followed by the ASCVD 10-year risk score. Notably, traits related to EAT, such as baseline EAT, baseline EAT/BSA, and annual EAT change, displayed no connection to CVD events. However, changes in calcium scores demonstrated positive correlations with HbA1c, weight, baseline TG, and age. Recognizing these factors as substantial contributors to CVD events, it can be inferred that conventional risk factors—HbA1c, weight, baseline TG, and age—play a more significant role in CVD through changes in calcium rather than EAT volume. Previous research indicated a link between visceral fat and obesity, underscoring its involvement in CVD and metabolic disorders like hypertension, abnormal lipid levels, insulin resistance, and type 2 diabetes [2]. Some studies link EAT to metabolic risk factors, CVD, and coronary calcification [18,19], but the exact association of EAT with CVD remains uncertain. In our study, the relationship between EAT change and CVD events underwent further examination by categorizing participants into three groups based on calcium score and ASCVD 10-year risk score. Within each group, participants were once more subdivided based on EAT change per year using criteria of ± 5% per year. However, none of the groups exhibited significant outcomes concerning CVD events, and the CVD ratio displayed no discernible trend when compared to EAT change. A recent study suggested no such correlation between perivascular adipose tissue and calcified plaque volume [28]. Other studies also reported that epicardial and pericardial fat lack association with coronary artery calcification [29,30]. Therefore, deep perspectives are needed to examine the relationship between the histologic quality and density of components as well as the natural course of EAT rather than the quantitative volume.

EAT measurement in many studies often uses CCTA, which is considered the standard method, but the recent focus is on EAT’s cross-sectional view due to ongoing controversies. The longitudinal perspective of EAT and its relation to CVD needs further exploration. This study examines EAT changes and risk factors over time, investigating their correlation and whether EAT directly links to CVD, contrasting with classical indicators like calcium score and ASCVD 10-year risk score.

This study had certain limitations. First, the fact that this study collected data from two centers in South Korea may render the analysis underpowered to demonstrate a significant association between EAT and CVD, posing a risk of type 2 error. For a total of 762 participants, the power of three EAT-related variables for predicting CVD, baseline EAT, baseline EAT/BSA, and change of EAT per year were 0.05, 0.13, and 0.18, respectively. Had the retrospective study included data from multiple centers and a large population, along with high-quality data, the study might have yielded more meaningful results.

Although the results should be interpreted with caution, it is worth noting that our study had the largest sample size among research conducted with a paired set. Second, the participants of this study do not represent the general population. The majority of the individuals who visited SNUH or SNUH Gangnam Center for examinations resided mostly in the urban area. Additionally, the fact that over 90% of the participants were male could introduce some limitations in the interpretation despite the lack of statistical significance. Third, the criteria of -5% and 5%/year for a change of EAT were defined arbitrarily in this study for analysis. Considering the distribution of participants, it was determined that only those with more than 5% change had clinically significant changes. The participants who demonstrated less than 5% of EAT change were few, which might have led to statistically insignificant results. Fourth, when the population was grouped according to calcium score and ASCVD 10- year risk score, the number of people was not divided by tertile or quartile but based on clinical cutoff criteria. Due to a shortage of participants in the greatest criterion, (> 400 HU and > 20%), they were combined with the following group: 100–400 HU and 7.5–20%. Lastly, as a retrospective study, the intervals between baseline and follow-up examination varied among participants. Moreover, obtaining accurate statistical data on variables like BMI, TG/HDL, LDL, calcium score, etc., at fixed time intervals poses a significant challenge. The prolonged time gap between the two CT scans, coupled with limited time-based adjustments, necessitated the consideration of aging, even when focusing on patients for health screenings.

In conclusion, some metabolic risks, e.g., BMI, TG/HDL, and glucose, were associated with volume and EAT change per year. However, classical CVD risks, e.g., LDL, calcium score, ASCVD 10-year risk score, did not demonstrate an association. The actual CVD event was not associated with EAT volume, warranting future studies combining qualitative assessments with quantitative ones. As some risk factors such as TG/HDL and glucose related to insulin resistance were associated with EAT, future studies with extended observation periods focused on insulin resistance are needed.

KEY MESSAGE

1. Participants with worse risk factors demonstrated more volume of EAT than others.

2. BMI, TG/HDL, and glucose were associated with the change of EAT per year, compared to LDL, calcium score, and ASCVD 10-year risk score, which were not.

3. CVD was not associated with factors related to EAT, warranting future studies with qualitative assessments of EAT.

Supplementary Information

Acknowledgments

We thank Soo Yeon Shin and Kyung Min Jung for their technical assistance in the radiological assessment of EAT.

Notes

CRedit authorship contributions

Sungjoon Park: conceptualization, investigation, writing - original draft, writing - review & editing; Dong Eun Kim: conceptualization, investigation, writing - original draft, visualization; Su Min Kim: investigation, formal analysis, writing - original draft; JungMin Choi: investigation, data curation, formal analysis; Sang Joon Park: formal analysis, validation; Hae-Young Lee: investigation, writing - review & editing, visualization, supervision; Eun Ju Chun: investigation, writing - review & editing, visualization, supervision

Conflicts of interest

The authors disclose no conflicts.

Funding

None

References

1. Chang S, Lee W, Kwag B, Chung YH, Kang IS. Maximal pericoronary adipose tissue thickness is associated with hypertension in nonobese patients with acute or chronic illness. Korean J Intern Med 2017;32:668–674.
2. Fujioka S, Matsuzawa Y, Tokunaga K, Tarui S. Contribution of intra-abdominal fat accumulation to the impairment of glucose and lipid metabolism in human obesity. Metabolism 1987;36:54–59.
3. Cai X, Luo D, Liu S, et al. Four-year changes in central fatness, risk of diabetes, and metabolic control in older adults: a cohort study with mediation analysis. Korean J Intern Med 2022;37:230–240.
4. Lee HY, Després JP, Koh KK. Perivascular adipose tissue in the pathogenesis of cardiovascular disease. Atherosclerosis 2013;230:177–184.
5. Bertaso AG, Bertol D, Duncan BB, Foppa M. Epicardial fat: definition, measurements and systematic review of main outcomes. Arq Bras Cardiol 2013;101:e18–e28.
6. Park SE, Park CY, Choi JM, et al. Depot-specific changes in fat metabolism with aging in a type 2 diabetic animal model. PLoS One 2016;11e0148141.
7. Alexopoulos N, Raggi P. The importance of targeting epicardial adipose tissue to improve cardiovascular outcomes in diabetes mellitus. Atherosclerosis 2022;363:69–70.
8. Prati F, Arbustini E, Labellarte A, et al. Eccentric atherosclerotic plaques with positive remodelling have a pericardial distribution: a permissive role of epicardial fat? A three-dimensional intravascular ultrasound study of left anterior descending artery lesions. Eur Heart J 2003;24:329–336.
9. Hogea T, Suciu BA, Ivănescu AD, et al. Increased epicardial adipose tissue (EAT), left coronary artery plaque morphology, and valvular atherosclerosis as risks factors for sudden cardiac death from a forensic perspective. Diagnostics (Basel) 2023;13:142.
10. Pezeshkian M, Noori M, Najjarpour-Jabbari H, et al. Fatty acid composition of epicardial and subcutaneous human adipose tissue. Metab Syndr Relat Disord 2009;7:125–131.
11. Sacks HS, Fain JN, Holman B, et al. Uncoupling protein-1 and related messenger ribonucleic acids in human epicardial and other adipose tissues: epicardial fat functioning as brown fat. J Clin Endocrinol Metab 2009;94:3611–3615.
12. Iacobellis G, Corradi D, Sharma AM. Epicardial adipose tissue: anatomic, biomolecular and clinical relationships with the heart. Nat Clin Pract Cardiovasc Med 2005;2:536–543.
13. Mazurek T, Zhang L, Zalewski A, et al. Human epicardial adipose tissue is a source of inflammatory mediators. Circulation 2003;108:2460–2466.
14. Yudkin JS, Eringa E, Stehouwer CD. “Vasocrine” signalling from perivascular fat: a mechanism linking insulin resistance to vascular disease. Lancet 2005;365:1817–1820.
15. Lu MT, Park J, Ghemigian K, et al. Epicardial and paracardial adipose tissue volume and attenuation - Association with high-risk coronary plaque on computed tomographic angiography in the ROMICAT II trial. Atherosclerosis 2016;251:47–54.
16. Kim TH, Yu SH, Choi SH, et al. Pericardial fat amount is an independent risk factor of coronary artery stenosis assessed by multidetector-row computed tomography: the Korean Atherosclerosis Study 2. Obesity (Silver Spring) 2011;19:1028–1034.
17. Villasante Fricke AC, Iacobellis G. Epicardial adipose tissue: clinical biomarker of cardio-metabolic risk. Int J Mol Sci 2019;20:5989.
18. Rosito GA, Massaro JM, Hoffmann U, et al. Pericardial fat, visceral abdominal fat, cardiovascular disease risk factors, and vascular calcification in a community-based sample: the Framingham Heart Study. Circulation 2008;117:605–613.
19. Mahabadi AA, Massaro JM, Rosito GA, et al. Association of pericardial fat, intrathoracic fat, and visceral abdominal fat with cardiovascular disease burden: the Framingham Heart Study. Eur Heart J 2009;30:850–856.
20. Tesche C, Bauer MJ, Straube F, et al. Association of epicardial adipose tissue with coronary CT angiography plaque parameters on cardiovascular outcome in patients with and without diabetes mellitus. Atherosclerosis 2022;363:78–84.
21. Turer AT, Scherer PE. Adiponectin: mechanistic insights and clinical implications. Diabetologia 2012;55:2319–2326.
22. Wu Y, Zhang A, Hamilton DJ, Deng T. Epicardial fat in the maintenance of cardiovascular health. Methodist Debakey Cardiovasc J 2017;13:20–24.
23. Sicari R, Sironi AM, Petz R, et al. Pericardial rather than epicardial fat is a cardiometabolic risk marker: an MRI vs echo study. J Am Soc Echocardiogr 2011;24:1156–1162.
24. Greif M, Becker A, von Ziegler F, et al. Pericardial adipose tissue determined by dual source CT is a risk factor for coronary atherosclerosis. Arterioscler Thromb Vasc Biol 2009;29:781–786.
25. Gorter PM, van Lindert AS, de Vos AM, et al. Quantification of epicardial and peri-coronary fat using cardiac computed tomography; reproducibility and relation with obesity and metabolic syndrome in patients suspected of coronary artery disease. Atherosclerosis 2008;197:896–903.
26. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr, Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 1990;15:827–832.
27. Arnett DK, Blumenthal RS, Albert MA, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation 2019;140:e563–e595.
28. Lee SE, Sung JM, Andreini D, et al. Association between changes in perivascular adipose tissue density and plaque progression. JACC Cardiovasc Imaging 2022;15:1760–1767.
29. Wassel CL, Laughlin GA, Araneta MR, et al. Associations of pericardial and intrathoracic fat with coronary calcium presence and progression in a multiethnic study. Obesity (Silver Spring) 2013;21:1704–1712.
30. Tanami Y, Jinzaki M, Kishi S, et al. Lack of association between epicardial fat volume and extent of coronary artery calcification, severity of coronary artery disease, or presence of myocardial perfusion abnormalities in a diverse, symptomatic patient population: results from the CORE320 multicenter study. Circ Cardiovasc Imaging 2015;8e002676.

Article information Continued

Figure 1.

The flow of participants through the study. CABG, coronary artery bypass grafting; CCTA, coronary computed tomography angiography; EAT, epicardial adipose tissue; PCI, percutaneous coronary intervention.

Figure 2.

Impact of EAT change per year on change of clinical characteristics (Scatter plots). Scatter plots display associations between EAT change per year and BMI (A), TG/HDL (C), and glucose change (E). In comparison, they did not display a correlation between EAT change per year and LDL (B), calcium score (D), and ASCVD 10-year risk score change (F). R2 and p values were calculated through simple regression analysis. ASCVD, atherosclerotic cardiovascular disease; BMI, body mass index; EAT, epicardial adipose tissue; HDL, high-density lipoprotein; HU, Hounsfield unit; LDL, low-density lipoprotein; TG, triglyceride.

Figure 3.

ROC curve for the impact of clinical characteristics on CVD event. ASCVD, atherosclerotic cardiovascular disease; AUC, area under the curve; BSA, body surface area; CVD, cardiovascular disease; CI, confidence interval; EAT, epicardial adipose tissue; ROC, receiver operating characteristic.

Figure 4.

CVD event rate with EAT change per year grouped by risk prediction markers. Analyzed CVD event and EAT change per year grouped by calcium score (A) and ASCVD 10-year risk score (B). ASCVD, atherosclerotic cardiovascular disease; CVD, cardiovascular disease; EAT, epicardial adipose tissue; HU, Hounsfield units.

Table 1.

Clinical characteristics of participants at baseline

Variable EAT group (mL)
Total p value
Q1 (< 84.00) Q2 (84.00–107.77) Q3 (107.77–135.39) Q4 (> 135.39)
Total 191 (25.1) 190 (24.9) 190 (24.9) 191 (25.1) 762 (100.0)
Sex, male 168 (88.0) 172 (90.5) 175 (92.1) 182 (95.3) 697 (91.5) 0.075
Age (yr) 57.5 ± 7.1 59.6 ± 6.1 58.2 ± 6.1 59.5 ± 6.8 58.7 ± 6.6 0.029
Height (cm) 166.2 ± 6.4 168.2 ± 6.7 169.0 ± 6.5 171.4 ± 6.5 168.7 ± 6.8 < 0.001
Weight (kg) 63.9 ± 7.3 69.1 ± 8.1 71.1 ± 8.7 76.4 ± 8.3 70.1 ± 9.3 < 0.001
BMI (kg/m2) 23.1 ± 2.2 24.2 ± 2.7 24.8 ± 2.3 26.0 ± 2.3 24.5 ± 2.6 < 0.001
 Obesity (BMI > 25) 33 (17.3) 74 (38.9) 92 (48.4) 126 (66.0) 325 (42.7) < 0.001
Smoking 0.302
 Never 75 (39.3) 59 (31.1) 53 (27.9) 59 (30.9) 246 (32.3)
 Ex-smoker 77 (40.3) 91 (47.9) 94 (49.5) 94 (49.2) 356 (46.7)
 Current smoker 36 (18.8) 37 (19.5) 42 (22.1) 36 (18.8) 151 (19.8)
Alcohol 0.012
 None 43 (22.5) 46 (24.2) 41 (21.6) 41 (21.5) 171 (22.4)
 Social 98 (51.3) 100 (52.6) 91 (47.9) 73 (38.2) 362 (47.5)
 Heavy 47 (24.6) 42 (22.1) 57 (30.0) 74 (38.7) 220 (28.9)
Comorbidity
 HTN 55 (28.8) 88 (46.3) 80 (42.1) 99 (51.8) 322 (42.3) < 0.001
 DM 22 (11.5) 33 (17.4) 43 (22.6) 38 (19.9) 136 (17.8) 0.033
 Dyslipidemia 63 (33.0) 87 (45.8) 82 (43.2) 76 (39.8) 308 (40.4) 0.097
 Statin 54 (28.3) 71 (37.4) 72 (37.9) 60 (31.4) 257 (33.7) 0.942
Calcium score (HU) 122.9 ± 279.5 124.6 ± 230.0 131.3 ± 282.9 227.5 ± 560.8 152.3 ± 366.2 0.013
ASCVD 10-year risk score (%) 8.2 ± 6.4 10.6 ± 7.0 9.9 ± 6.7 11.6 ± 7.9 10.1 ± 7.1 < 0.001
SBP (mmHg) 118.2 ± 14.8 121.1 ± 14.3 119.1 ± 15.3 120.8 ± 16.2 119.8 ± 15.2 0.236
DBP (mmHg) 75.5 ± 10.5 77.8 ± 10.8 77.1 ± 11.2 78.1 ± 11.0 77.1 ± 10.9 0.039
Laboratory results
 HbA1c (%) 5.8 ± 0.7 6.0 ± 0.6 6.1 ± 0.8 6.1 ± 0.8 6.0 ± 0.7 < 0.001
 Glucose (mg/dL) 101.9 ± 21.5 106.2 ± 19.1 110.4 ± 30.1 111.7 ± 24.9 107.5 ± 24.5 < 0.001
 Total cholesterol (mg/dL) 198.1 ± 32.8 190.4 ± 33.9 187.9 ± 34.0 195.0 ± 37.1 192.9 ± 34.7 0.294
 TG (mg/dL) 116.7 ± 62.6 124.0 ± 64.8 135.7 ± 70.7 144.7 ± 70.5 130.3 ± 68.0 < 0.001
 HDL (mg/dL) 53.4 ± 11.8 50.4 ± 10.8 49.5 ± 10.2 49.8 ± 10.5 50.7 ± 10.9 < 0.001
 LDL (mg/dL) 123.9 ± 30.1 120.2 ± 28.4 118.1 ± 31.0 122.7 ± 34.0 121.2 ± 30.9 0.569
 TG/HDL 2.4 ± 1.8 2.6 ± 1.7 3.0 ± 1.9 3.1 ± 1.8 2.8 ± 1.8 < 0.001
 HR (bpm) 67.0 ± 9.8 67.5 ± 10.8 69.1 ± 11.6 67.4 ± 10.6 67.8 ± 10.7 0.442

Values are presented as number (%) or mean ± standard deviation.

ASCVD, atherosclerotic cardiovascular disease; BMI, body mass index; DBP, diastolic blood pressure; DM, diabetes mellitus; EAT, epicardial adipose tissue; HDL, high density lipoprotein; HR, heart rate; HTN, hypertension; HU, Hounsfield unit; LDL, low density lipoprotein; SBP, systolic blood pressure; TG, triglycerides.

Table 2.

Changes of EAT volume over time

Variable Q1 Q2 Q3 Q4 Total p value
Baseline
 EAT (mL) 68.5 ± 11.3 96.1 ± 6.6 120.8 ± 7.7 166.8 ± 26.1 113.1 ± 39.2 < 0.001
 BSA (m2) 1.7 ± 0.1 1.8 ± 0.2 1.8 ± 0.1 1.9 ± 0.1 1.8 ± 0.2 < 0.001
 EAT/BSA (mL/m2) 40.0 ± 6.6 53.8 ± 5.2 66.6 ± 6.7 87.8 ± 13.9 62.1 ± 19.7 < 0.001
 Weight (kg) 63.9 ± 7.3 69.1 ± 8.1 71.1 ± 8.7 76.4 ± 8.3 70.1 ± 9.3 < 0.001
 BMI (kg/m2) 23.1 ± 2.2 24.2 ± 2.3 24.8 ± 2.3 26.0 ± 2.3 24.5 ± 2.6 < 0.001
Follow-up
 EAT (mL) 78.7 ± 18.9 107.7 ± 16.6 130.7 ± 18.0 181.2 ± 38.0 124.6 ± 44.9 < 0.001
 BSA (m2) 1.7 ± 0.1 1.8 ± 0.1a) 1.8 ± 0.1a) 1.9 ± 0.1 1.8 ± 0.1 < 0.001
 EAT/BSA (mL/m2) 45.8 ± 10.4 60.4 ± 10.0 72.2 ± 10.2 95.7 ± 19.2 68.6 ± 22.5 < 0.001
 Weight (kg) 64.1 ± 7.4 68.5 ± 8.3 70.7 ± 8.9 75.7 ± 9.4 69.8 ± 9.5 < 0.001
 BMI (kg/m2) 23.2 ± 2.1 24.2 ± 2.1 24.7 ± 2.4 25.8 ± 2.5 24.5 ± 2.5 < 0.001
EAT change per year (%/yr) 2.7 ± 4.5b,c) 2.5 ± 3.8 1.2 ± 3.7b) 1.3 ± 2.9c) 1.9 ± 3.8 0.001

Values are presented as mean ± standard deviation.

BMI, body mass index; BSA, body surface area; EAT, epicardial adipose tissue.

a)

There were significant differences among each quartile of BSA at baseline and follow-up except between Q2 and Q3 at baseline and follow-up BSA (Adjusted p for baseline BSA = 0.130, adjusted p for follow-up BSA = 0.179 by post-hoc comparison).

b)

Adjusted p = 0.012.

c)

Adjusted p = 0.003.

EAT change per year = (follow-up EAT - baseline EAT) / baseline EAT × 100 / year.

Table 3.

Clinical characteristics of participants according to EAT change per year

Variable EAT change (%/yr)
p value (calculated between < -5 and > 5 groups)
< -5 (n = 24) -5 to 5 (n = 637) > 5 (n = 101)
BMI (kg/m2) -4.9 ± 4.6 -0.7 ± 4.1 3.1 ± 4.8 < 0.001
Calcium score (HU) 351.9 ± 616.9 391.8 ± 711.3 118.8 ± 174.7 0.100
ASCVD 10-year risk score (%) 3.0 ± 7.9 5.1 ± 6.4 4.2 ± 6.3 0.241
SBP (mmHg) 0.9 ± 14.0 1.1 ± 17.2 0.8 ± 17.4 0.810
DBP (mmHg) -2.6 ± 11.5 0.1 ± 12.7 -1.4 ± 14.3 0.617
Laboratory results
 HbA1c (%) -0.2 ± 0.4 0.3 ± 7.2 0.1 ± 0.4 < 0.001
 Glucose (mg/dL) -4.8 ± 18.3 3.1 ± 22.0 7.9 ± 18.0 < 0.001
 Total cholesterol (mg/dL) -17.5 ± 40.1 -17.5 ± 41.3 -16.9 ± 34.6 0.866
 TG (mg/dL) -33.8 ± 62.0 -13.7 ± 70.2 -2.3 ± 54.0 0.010
 HDL (mg/dL) 3.2 ± 6.2 1.0 ± 8.9 -2.1 ± 9.4 0.001
 LDL (mg/dL) -14.1 ± 22.8 -7.7 ± 29.4 -6.7 ± 24.3 0.217
 TG/HDL (mg/dL) -19.8 ± 46.3 2.1 ± 65.2 18.6 ± 70.0 0.001

Values are presented as mean ± standard deviation.

ASCVD, atherosclerotic cardiovascular disease; BMI, body mass index; DBP, diastolic blood pressure; EAT indicates epicardial adipose tissue; HDL, high-density lipoprotein; HU, Hounsfield unit; LDL, low-density lipoprotein; SBP, systolic blood pressure; TG, triglyceride.

Table 4.

Association between EAT and cardiovascular disease events in terms of statin usage

Variable Q1 Q2 Q3 Q4 Total p value
Baseline EAT (mL) < 84.00 84.00–107.77 107.77–135.39 > 135.39
 Statin (n = 506) 16 (11.7) 13 (10.8) 16 (13.6) 10 (7.6) 55 (10.9) 0.496
 No statin (n = 256) 5 (9.3) 15 (21.4) 14 (19.4) 8 (13.3) 42 (16.4) 0.241
Follow-up EAT (mL) < 92.44 92.44–118.40 118.40–149.70 > 149.70
 Statin (n = 506) 17 (12.7) 16 (12.2) 11 (9.5) 11 (8.8) 55 (10.9) 0.686
 No statin (n = 256) 10 (18.2) 12 (19.7) 11 (14.9) 9 (13.6) 42 (16.4) 0.777
Baseline EAT/BSA (mL/m2) < 47.59 47.59–59.91 59.91–74.20 > 74.20
 Statin (n = 506) 16 (11.9) 13 (10.6) 16 (13.4) 10 (7.7) 55 (10.9) 0.505
 No statin (n = 256) 9 (16.1) 12 (17.6) 13 (18.6) 7 (17.1) 41 (16.1) 0.705

Values are presented as number (%).

BSA, body surface area; EAT, epicardial adipose tissue.