Local Thickness of Epicardial Adipose Tissue Surrounding the Left Anterior Descending Artery Is a Simple Predictor of Coronary Artery Disease　― New Prediction Model in Combination With Framingham Risk Score ―

Abstract

Background: Compared with global cardiac adiposity, the local accumulation of fat surrounding coronary arteries might have a more direct impact on coronary artery disease (CAD). Here, we compared the local epicardial adipose tissue (EAT) thickness and global cardiac adiposity volumes for predicting CAD.

Methods and Results: A total of 197 consecutive subjects underwent 320-slice multi-detector computed tomography coronary angiography and were segregated into CAD (≥1 coronary artery branch stenosis ≥50%) and non-CAD groups. EAT thickness was measured at the right coronary artery (EAT_RCA), the left anterior descending artery (EAT_LAD), and the left circumflex artery (EAT_LCX). Although EAT_RCA and EAT_LCX were similar between the 2 groups, EAT_LAD was larger in the CAD group than in the non-CAD group (5.45±2.16 mm vs. 6.86±2.19 mm, P<0.001). EAT_LAD, after correcting for confounding factors, was strongly associated with CAD (r=0.276, P<0.001) and Gensini score (r=0.239, P<0.001). On multiple regression analysis, Framingham risk score combined with EAT_LAD was a strong predictor of CAD (adjusted R²=0.121; P<0.001).

Conclusions: The local fat thickness surrounding the LAD is a simple and useful surrogate marker for estimating the presence, severity, and extent of CAD, independent of classical cardiovascular risk factors.

Epicardial adipose tissue (EAT) is located below the visceral layer of the pericardium and directly contacts the coronary arteries. The proximity of EAT to the coronary arteries prompted us to consider the pathophysiological consequences associated with this tissue. We and others have shown that EAT is a source of multiple inflammatory cytokines.^1–5 Reportedly, EAT volume (EATV), determined using either multi-detector computed tomography (MDCT) or magnetic resonance imaging and echocardiography, correlates with the presence and severity of coronary artery disease (CAD).^6–8 Prospectively, EAT also predicts the incidence of CAD independent of traditional cardiovascular risk factors.⁹

Compared with global cardiac adiposity, such as EATV, the local accumulation of fat surrounding an artery, which can be assumed using local EAT thickness, might have a more direct impact on regional coronary atherosclerosis and predict CAD more efficiently. This idea, however, has not been fully clarified.

The aim of this study was therefore to measure the local EAT thicknesses surrounding the coronary arteries of patients with or without CAD and determined their clinical utility, compared with the global cardiac EATV, for predicting CAD.

Methods

Patients

A total of 482 consecutive subjects who underwent MDCT between April 2012 and August 2015 at Tokushima University Hospital were enrolled in this study. Of these, 25 were excluded due to lack of data for Framingham risk score (FRS) calculation. Also, 95 subjects who could not have Gensini score calculated due to poor coronary CT angiography (CTA) quality (mainly coronary artery calcification) and lack of follow-up coronary angiography (CAG), and 160 who did not have abdominal CT were excluded. Of 202, 5 were excluded because of insufficient image quality for EAT measurements. Finally, we analyzed CTA datasets of 197 patients. The other exclusion criteria were irregular heartbeat during MDCT; serum creatinine >2.0 mg/dL; class III or IV heart failure; known hypersensitivity to iodine-based contrast agents; acute coronary events, stroke, or coronary revascularization within the preceding 3 months; overt liver disease; and hypothyroidism.¹⁰ The ethics committee of Tokushima University Hospital approved the study.

Covariates, including cardiac disease history and risk factors, were obtained from patient electronic medical records. All participants provided written informed consent after they were advised regarding radiation exposure-related risks and the possible complications associated with iodine-containing contrast agents. Hypertension was defined as systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg, or the current use of antihypertensive medication. Diabetes was defined as glycated hemoglobin concentration ≥6.5%, fasting plasma glucose >126 mg/dL, or the current use of anti-diabetic medications. Dyslipidemia was defined as total serum cholesterol ≥220 mg/dL, low-density lipoprotein cholesterol (LDL-C) ≥140 mg/dL, serum triglyceride (TG) ≥150 mg/dL, and serum high-density lipoprotein cholesterol (HDL-C) <40 mg/dL, and/or current use of anti-hyperlipidemic medications. Smoking was defined as past or current smoking; non-smoking was defined as never having smoked. FRS is a multivariable statistical model that uses age, sex, smoking history, blood pressure, cholesterol and HDL-C, and blood glucose level or history of diabetes to estimate the coronary event risk for individuals without previously diagnosed CAD.^11–13 FRS was used to estimate a 10-year risk of CAD, defined as angina pectoris, recognized and unrecognized myocardial infarction, and death due to CAD. According to this score, patients were stratified as follows: low likelihood of CAD, <10%; intermediate likelihood, 10–20%; and high likelihood, >20%.^14,15

CTA Data Acquisition

CTA was performed using a 320-slice CT scanner (Aquilion One; Toshiba Medical Systems, Tokyo, Japan) having 0.275-ms rotation and 0.5/320/0.25 collimation as previously described.¹⁶ Briefly, a plain scan was taken to measure the coronary calcium score, using the standard Agatston method (slice thickness, 3 mm; maximum tube current, 170 mA; tube voltage, 120 KV). For CTA, tube voltage was set at 120 KV, with a maximum tube current of 214 mA. All reconstructed CT image data were transferred to an offline workstation (Synapse Vincent, ver. 4.4, Fuji Film, Tokyo, Japan) for post-processing and image analysis. Contrast agent transmit time, using a bolus tracking technique according to body weight, was measured using a contrast agent (Iopamiron, 370 mg iodine/mL, Bracco Imaging, Milan, Italy) and contrast agent injection syringe (Bracco Imaging) followed by a 50-mL saline flush, both at flow rates of 5 mL/s using a dual-head power injector (Nemoto Kyorindo, Tokyo, Japan). For coronary CTA, 40–50 mL contrast agent was injected into an antecubital vein, directly followed by 40 mL i.v. saline (maximum, 5 mL/s). One nitroglycerine spray (0.3 mg) was given sublingually. If needed, i.v. metoprolol tartrate (12.5 mg) was used to lower heart rate to <60 beats/min. The scan time was 5–8s, and images were collected during an inspiratory breath hold. Measurements were performed during a motionless phase of the cardiac cycle, which was usually a diastolic phase, with retrospective cardiac gating at 70–80% of the RR interval. For image reconstruction, a medium sharp convolution kernel was used, with a 0.5-mm slice thickness and a 0.25-mm increment.

CTA and EAT Volume and Thickness

The presence of stenosis (>50% of the luminal diameter) in at least 1 major epicardial coronary artery or branch was defined as CAD. The coronary tree was further divided into 15 segments, based on the American Heart Association classification.¹⁷ To evaluate atherosclerosis severity, the coronary segments with the most severe luminal diameter stenosis were scored as 0 points, <25% stenosis; 1 point, 25–50%; 2 points, 51–75%; 4 points, 76–90%; 8 points, 91–99%; or 16 points, 100%, as modified from Gensini,¹⁸ and the cumulative score from all segments was recorded as the modified Gensini severity score. We had performed CAG in patients with suspected significant coronary stenosis on MDCT or in whom severity could not be determined because of coronary calcification.

Measurement of EATV was performed as described.¹ Briefly, using Vincent’s volume measurement software, both the EAT thickness and cross-sectional EATV were measured using the computer workstation. To calculate EATV, all axial images (approximately 250 per patient) were loaded into the workstation and the epicardium was manually traced on these images. EATV measurements were performed on axial images by manually tracing the parietal pericardium, from the left main pulmonary artery level to the left ventricular apex, as described.¹ The total EATV was calculated on a highlighted fat map, with a threshold of −190 to −30 HU.²⁰ The EATV index was defined as EATV/body surface area.¹ Local EAT thicknesses were measured on short-axis view (Figure 1). Using an axial, 4-chamber view in which all 4 chambers had been observed proportionally, EAT thickness measurements were performed at the following points: (1) surrounding the left anterior descending artery (LAD; EAT_LAD); (2) surrounding the right coronary artery (RCA; EAT_RCA); and (3) surrounding the left circumflex artery (LCX; EAT_LCX). On the highlighted fat map, the length of lines perpendicular to the lines between the heart surface and the visceral epicardium was determined: (1) EAT_LAD at the anterior interventricular groove (AIVG); (2) EAT_RCA at the right atrioventricular groove (RAVG); and (3) EAT_LCX at the left atrioventricular groove (LAVG). The subcutaneous fat area (SFA) and intra-abdominal visceral fat area (VFA) were measured at the level of the umbilicus as previously described.¹⁶

Figure 1.

Representative measurements of local epicardial adipose tissue (EAT) thickness in the area surrounding right coronary artery (RCA; EAT_RCA), left anterior descending artery (LAD; EAT_LAD) and left circumflex artery (LCX; EAT_LCX). From (A,D) plain axial 4-chamber views, a region of interest (ROI) for EAT measurements was manually placed (B,E) along the visceral pericardium, and (C,F) the EAT area was automatically acquired as the density range between −190 and −30 Hounsfield units. As compared with (C) a 72-year-old man without coronary artery disease (CAD), a (F) 70-year-old woman with obstructive CAD had an increase in EAT_LAD (5.9 mm vs. 10.2 mm), although similar EAT_RCA and EAT_LCX. The presence of ≥1 stenosis >50% on luminal diameter in at least 1 major epicardial coronary artery or branches was defined as CAD. Red arrows, measures of EAT thickness. LA, left atrium; LV, left ventricle; RA, right atrium; RV, right ventricle.

Preliminary to this study, we had determined the location of the EATT measures. Comparison with curved planar reformation (CPR) images of the coronary arteries was used to determine the appropriate views (axial and coronal) to measure EAT thickness. For all samples we had compared, the coronal views were variable, but the axial views were not. Thus, the axial view was selected. Next, we measured the variation of data according to the height in the view. In 8 non-CAD (61±10 years, 5 men) and 8 CAD patients (69±10 years, 4 men), EAT thickness in the axial 4-chamber view (current protocol) had only small differences in mm and %, and hence this measurement was used as the representative value (Table S1).

Statistical Analysis

All statistical analysis was performed using SPSS 21.0 for Windows (SPSS, Chicago, IL, USA). Continuous variables are expressed as mean±SD and were compared using unpaired Student’s t-test. Categorical variables are summarized as frequency and percentage and were compared using chi-squared test. Differences in continuous variables between the 2 groups were compared using independent Student’s t-test. Comparisons between groups stratified by the number of narrowed coronary arteries were done using 1-way analysis of variance and a linear regression model. Statistical significance for all tests was set at P<0.05; Bonferroni correction was applied when necessary. Correlations between CAD (yes or no), EATV, waist circumference (WC), VFA, and EAT thickness were determined using Pearson’s correlation coefficient test. Multivariate regression analysis was used to determine the predictors of CAD, adjusting for clinical risk factors (including age, male sex, diabetes mellitus, hypertension, smoking habits, body mass index [BMI], and WC). Receiver operating characteristic (ROC) curve analysis was used to determine the optimal cut-off points for EAT measurements to predict significant CAD, and to assess whether the dichotomously defined EAT thickness measurements provided added predictive value for significant CAD, in addition to traditional risk factors.

Results

General Characteristics

General subject characteristics are listed according to CAD status in Table 1. In the CAD group, the mean age and prevalence of hypertension were higher, and the mean VFA and number of patients with BMI ≥25 kg/m² tended to be higher. Otherwise, the 2 groups had similar numbers of male patients and similar average BMI, WC, and SFA. The number of patients with type 2 diabetes mellitus, dyslipidemia, and history of smoking was not significantly different between the 2 groups. FRS was higher in the CAD group than in the non-CAD group (9.6±7.2% vs. 12.5±7.2%, P=0.006).

Table 1. Subject Characteristics vs. CAD Status

Parameters	Non-CAD (n=84)	CAD (n=113)	P-value†
Age (years)	64±14	70±11	0.002
Male	48 (57)	74 (66)	0.233
Anthropometry
BMI (kg/m2)	23.8±4.4	24.4±4.0	0.148
BMI ≥25 kg/m2	32 (38)	57 (50)	0.085
WC (cm)	85.1±12.0	86.6±11.5	0.358
VFA (cm2)	102±57	119±66	0.063
SFA (cm2)	135±88	130±86	0.686
Comorbidities
T2DM	21 (25)	38 (34)	0.177
Hypertension	60 (71)	94 (83)	0.048
Dyslipidemia	60 (71)	86 (76)	0.459
History of smoking	33 (39)	59 (52)	0.072
FRS (%)	9.6±7.2	12.5±7.2	0.006
EAT measures
EATV (cm3)	104±45	124±57	0.006
EATV index (cm3/m2)	63±25	75±32	0.004
EATRCA thickness (mm)	15.49±4.73	16.46±4.76	0.160
EATLAD thickness (mm)	5.45±2.16	6.86±2.19	<0.001
EATLCX thickness (mm)	9.80±3.39	10.33±3.61	0.297
Agatston score
Total	116±328	879±1,333	<0.001
In RCA	45±189	349±743	<0.001
In LAD	53±131	309±437	<0.001
In LCX	8±29	169±369	<0.001
Coronary artery stenosis ≥50%
RCA	–	69 (61)	–
LAD	–	72 (64)	–
LCX	–	61 (54)	–
1-vessel disease	–	55 (49)	–
2-vessel disease	–	27 (24)	–
3-vessel disease	–	31 (27)	–

Data are given as mean±SD or n (%). ^†Independent t-test or chi-squared test. BMI, body mass index; BSA, body surface area; CAD, coronary artery disease; EAT, epicardial adipose tissue; EATV, epicardial adipose tissue volume; EATV index, EATV/BSA; FRS, Framingham risk score; LAD, left ascending artery; LCX, left circumflex artery; RCA, right coronary artery; SFA, subcutaneous fat area; T2DM, type 2 diabetes mellitus; VFA, visceral fat area; WC, waist circumference.

EATV and EAT Thickness

Figure 1 shows the representative EAT thickness measurements in the areas surrounding the RCA, LAD, and LCX. EAT_LAD, but not EAT_RCA or EAT_LCX, was larger in CAD patients (10.2 mm) than in non-CAD patients (5.9 mm). Comparing the group means, EATV, EATV indexes, and EAT_LAD were higher in the CAD group (Table 1, Figure 2A,C,E,G). When patients were further divided into subgroups according to number of diseased vessels, that is, those with 0-vessel disease (0VD; no CAD) or 1–3VD (CAD), EATV_LAD increased with the number of diseased vessels, whereas EATV index, EAT_RCA, and EAT_LCX did not differ across the 4 groups (Figure 2B,D,F,H).

Figure 2.

Total EAT volume (EATV) index (EATV/body surface area) and local EAT thickness in the area surrounding the RCA (EAT_RCA), LAD (EAT_LAD) and LCX (EAT_LCX) according to CAD status and number of diseased vessels. 0VD, 0-vessel disease (no CAD); 1–3VD, 1–3-vessel disease (CAD). CAD was defined as ≥1 stenosis >50% on luminal diameter in at least 1 major epicardial coronary artery or branches. Whiskers, mean±SD. Between-group differences for CAD status were analyzed using independent Student’s t-test, and those for 0VD vs. 1–3VD were analyzed using 1-way analysis of variance and a linear regression model, followed by post-hoc Dunnet test. Statistical significance for all tests was set at P<0.05. *P<0.05, ^¶P<0.001 vs. 0VD. Abbreviations as in Figure 1.

Univariate Indicators of CAD and Gensini Score

Univariate regression analysis was used to estimate the presence of CAD and determine the Gensini score (Table 2; Figure 3). CAD was significantly correlated with age, hypertension, and FRS. Although whole body and abdominal adiposity parameters (BMI, BMI ≥25 kg/m², VFA, SFA, and WC) were not correlated with CAD, some EAT measures (EATV, EATV index, and EAT_LAD; but not EAT_RCA and EAT_LCX) were correlated with CAD. EAT_LAD (r=0.306, P<0.001) was more strongly correlated with CAD than was EATV or EATV index. Gensini score was significantly correlated with age, male sex, VFA, smoking history, and FRS. EAT_LAD, but not EATV, EATV index, EAT_RCA, or EAT_LCX, was also correlated with the Gensini score (Table 2; Figure 3A–D). On scatter plot analysis between the Gensini score in each coronary artery and the corresponding EAT measures, significant correlation was found only between Gensini score in the LAD and EAT_LAD, but not between those in the RCA and LCX (Figure S1A).

Table 2. Univariate Indicators of CAD and Gensini Score

Parameters	CAD		Gensini score
	r	P-value	r	P-value
Age (years)	0.236	0.001	0.168	0.018
Male gender (yes or no)	0.085	0.235	0.158	0.026
Anthropometry
BMI (kg/m2)	0.070	0.327	0.085	0.235
BMI ≥25 kg/m2 (yes or no)	0.123	0.086	0.100	0.160
VFA (cm2)	0.133	0.063	0.163	0.022
SFA (cm2)	−0.029	0.686	0.012	0.869
WC (cm)	0.066	0.358	0.087	0.226
Comorbidities
T2DM (yes or no)	0.096	0.179	0.143	0.045
Hypertension (yes or no)	0.141	0.048	0.027	0.710
Dyslipidemia (yes or no)	0.053	0.461	0.123	0.086
History of smoking (yes or no)	0.128	0.073	0.164	0.021
FRS (%)	0.197	0.006	0.200	0.005
EAT measures
EATV (cm3)	0.194	0.006	0.100	0.162
EATV index (cm3/m2)	0.206	0.004	0.100	0.161
EATRCA thickness (mm)	0.101	0.160	0.006	0.935
EATLAD thickness (mm)	0.306	<0.001	0.263	<0.001
EATLCX thickness (mm)	0.075	0.297	0.065	0.367

Abbreviations as in Table 1.

Figure 3.

Linear correlation between Gensini score, total EATV index (EATV/body surface area) and local EAT thickness in the area surrounding the RCA (EAT_RCA), LAD (EAT_LAD) and LCX (EAT_LCX) for (A–D) all patients; (E–H) the LAD group (coronary stenosis >50% in LAD); and (I–L) the non-LAD group (no coronary stenosis >50% in LAD). (○) Non-coronary artery disease (non-CAD); (●) CAD. CAD was defined as ≥1 stenosis >50% on luminal diameter in at least 1 major epicardial coronary artery or branches. Linear regression analysis was carried out for the combined group of non-CAD and CAD subjects. Abbreviations as in Figures 1,2.

To minimize the bias of LAD dominance in the severity of coronary atherosclerosis, we subdivided patients into a LAD group (coronary stenosis >50% in LAD; Figure 3E–H) and a non-LAD group (no coronary stenosis >50% in LAD; Figure 3I–L). We found that Gensini score was smaller, but a borderline correlation was still observed between EAT_LAD and Gensini score even in the non-LAD group (Figure 3C,G,K).

Multivariate Indicators of CAD and Gensini Score

We determined the impact of individual covariates on CAD and Gensini score using multivariate regression models in a standard, sequential (hierarchical) fashion (Table 3).

Table 3. Multivariate Indicators of CAD (A) and Gensini Score (B)

A. CAD
	Model 1		Model 2		Model 3		Model 4		Model 5		Model 6		Model 7
Adjusted R2	0.059		0.065		0.073		0.135		0.034		0.058		0.121
P-value	0.002		0.006		0.004		<0.001		0.006		0.001		<0.001
	r	P-value	r	P-value	r	P-value	r	P-value	r	P-value	r	P-value	r	P-value
Age (years)	0.254	<0.001	0.238	0.001	0.202	0.008	0.189	0.008
Male gender (yes or no)	0.080	0.256	0.044	0.587	0.055	0.500	0.075	0.345
BMI (kg/m2)	0.097	0.170	0.067	0.363	0.015	0.852	0.020	0.774
History of smoking (yes or no)			0.108	0.178	0.103	0.198	0.089	0.25
Hypertension (yes or no)			0.097	0.177	0.096	0.179	0.114	0.098
Dyslipidemia (yes or no)			0.064	0.377	0.049	0.499	0.063	0.366
T2DM (yes or no)			0.054	0.446	0.037	0.603	0.051	0.453
FRS (%)									0.197	0.006	0.163	0.022	0.191	0.005
EATV index (cm3/m2)					0.127	0.117					0.175	0.014
EATLAD thickness (mm)							0.276	<0.001					0.303	<0.001
B. Gensini score
	Model 1		Model 2		Model 3		Model 4		Model 5		Model 6		Model 7
Adjusted R2	0.049		0.074		0.069		0.125		0.035		0.034		0.098
P-value	0.005		0.003		0.006		<0.001		0.005		0.013		<0.001
	r	P-value	r	P-value	r	P-value	r	P-value	r	P-value	r	P-value	r	P-value
Age (years)	0.187	0.009	0.198	0.006	0.201	0.009	0.156	0.029
Male gender (yes or no)	0.152	0.033	0.120	0.142	0.119	0.148	0.146	0.067
BMI (kg/m2)	0.092	0.200	0.046	0.525	0.051	0.527	0.006	0.929
History of smoking (yes or no)			0.105	0.191	0.105	0.19	0.088	0.258
Hypertension (yes or no)			−0.014	0.843	−0.014	0.844	0.001	0.987
Dyslipidemia (yes or no)			0.151	0.037	0.152	0.038	0.15	0.033
T2DM (yes or no)			0.11	0.120	0.111	0.12	0.107	0.118
FRS (%)									0.200	0.005	0.188	0.009	0.196	0.004
EATV index (cm3/m2)					−0.011	0.894					0.064	0.374
EATLAD thickness (mm)							0.239	<0.001					0.260	<0.001

Abbreviations as in Table 1.

Age was a determinant of CAD, after correcting for sex and BMI (model 1; Table 3A). The addition of smoking history, hypertension, dyslipidemia (LDL-C >140 mg/dL or statin use, TG >150 mg/dL, or HDL-C <40 mg/dL), and type 2 diabetes mellitus did not result in appreciable increases in corrected R² (model 2). When added to the combination model using these traditional risk factors (models 1, 2), EATV index (a marker of global cardiac adiposity) did not increase corrected R² (model 3), but the addition of EAT_LAD increased corrected R² (0.135; model 4). The FRS 10-year CAD risk was a weak model for CAD (model 5), but the addition of the EATV index (0.058; model 6) and EAT_LAD increased corrected R² (0.121; model 7).

Age and male sex (model 1) were determinants of Gensini score (Table 3B), after correcting for BMI. The addition of traditional risk factors (smoking history, hypertension, dyslipidemia, and type 2 diabetes mellitus; 0.074; model 2) or EATV index (0.069; model 3) did not result in appreciable increases in corrected R², but addition of EAT_LAD to the model increased corrected R² (0.125; model 4). Compared to only FRS (R²=0.035; model 5), the addition of EAT_LAD also increased corrected R² (0.098; model 7).

EAT Measures and Possible Covariates

To determine the clinical features associated with EAT thickness, we performed univariate regression analysis between the EAT measures and their possible covariates (Table S2). EATV index was correlated with CAD, age, BMI, WC, SFA, and VFA. EAT_RCA and EAT_LCX were not correlated with CAD, but were correlated with age, BMI, WC, SFA, and VFA. In contrast, EAT_LAD was correlated with CAD and age, but not with BMI, WC, SFA, or VFA.

Combined FRS and EAT Prediction Model for CAD

Optimal cut-off points for predicting CAD on ROC curve analysis are shown in Figure 4. FRS, EATV index, and EAT_LAD cut-off points for predicting CAD were >8% (sensitivity, 66%; specificity, 54%), >90 cm³/m² (sensitivity, 33%; specificity, 88%), and >5.7 mm (sensitivity, 72%; specificity, 57%), respectively. Compared with only the FRS cut-off (model 1), the addition of the EAT_LAD cut-off to the FRS (model 1 vs. model 3: P=0.014) resulted in a stronger predictive value than did the model combining the EATV index cut-off and that for the FRS (model 2 vs. model 3: P=0.042).

Figure 4.

Receiver operating characteristic (ROC) curve analysis of the predictive accuracy of (A) EAT thickness in the area surrounding the left anterior descending artery (LAD; EAT_LAD) and (B) multivariate models combining EAT measures and Framingham risk score (FRS), for significant CAD. (B) Cut-offs of FRS and total EATV index (EATV/body surface area) for predicting CAD were >8% (sensitivity, 66%; specificity, 54%) and >90 cm³/m² (sensitivity, 33%; specificity, 88%), respectively. Between-model comparison was done using C statistics, and statistical significance was set at P<0.05. AUC, area under the curve. Other abbreviations as in Figures 1,2.

Discussion

There major observations were noted in the present study. First, after correcting for possible confounding factors, EAT_LAD, but not whole cardiac adiposity (EATV index), was independently associated with CAD and with the severity and extent of coronary atherosclerosis (Gensini score; Tables 2,3). Second, along with the unequal distributions of EAT, we compared local variation in the strength of associations between EAT thickness and coronary atherosclerosis. Of all EAT measurements, EAT_LAD was most significantly increased (Table 1) and associated with CAD and Gensini score (Table 2). Third, EAT_LAD, but no other EATV measures, provided additional prognostic value for predicting CAD, in combination with FRS. Thus, the simple measurement of EAT_LAD can be a useful surrogate clinical marker of CAD.

Local EAT Thickness and Global Cardiac Adiposity

A growing body of evidence suggests that local fat distribution, compared with global cardiac adiposity, plays an important role in the development of an unfavorable metabolic and cardiovascular risk profile.^7,20–22 When added to models using traditional risk factors for CAD and Gensini scores (models 1, 2), EATV index, a marker of global cardiac adiposity, did not increase corrected R² (model 3). Addition of EAT_LAD, however, increased corrected R² (0.135; model 4). Thus, we suggest that only modest associations are seen between global EATV and coronary atherosclerosis, but strong associations are observed between local EAT thicknesses and the presence, severity, and extent of CAD. Of the studies examining the relationship between EAT measures and atherosclerosis of the adjacent arteries, the majority have shown that increases either in EATV or in the local thickness of the coronary perivascular adipose tissue were significantly correlated with obstructive CAD.⁷

The correlation of global EAT adiposity and local EAT thicknesses with regard to prognostic impact has not fully been evaluated. In the present study, local EAT thickness, as compared with EATV, was a sensitive and prognostic measure of CAD. Global cardiac adiposity is well known to be strongly associated with whole-body adiposity.^24,25 In the current study, EATV index correlated well with CAD, as well as with age, BMI, WC, SFA, and VFA (Table S2). In contrast, EAT_RCA and EAT_LCX correlated with age, BMI, WC, SFA, and VFA, but not with CAD; EAT_LAD correlated with CAD and age, but not with BMI, WC, SFA, or VFA. Therefore, EAT_LAD independently reflects local circumstances, relative to whole-body adiposity. EAT, mainly consisting of adipocytes and tissue macrophages, can be a source of endocrine and paracrine cytokines, substrates, and adipokines. Thus, if EAT becomes “sick fat“, it directly contributes to atherosclerosis through an outside-to-inside inflammatory atherogenic signal conveyed through the vasa vasorum.^25–27 In contrast, the progression of atherosclerotic plaque²⁸ and the presence of an oxygen-insufficient microenvironment, the inner layers may provide an inside-to-outside signal, similar to hypoxic conditions, resulting in EAT accumulation.²⁹ This may be supported by our previous observation that pro-inflammatory cytokine profiles are more prominently enhanced in locations where severe atherosclerosis is observed.^1,30

Location-Specific EAT Thickness and CAD

EAT_LAD, but not EAT_RCA or EAT_LCX, was greater in the CAD group. When patients were divided into 0VD or 1–3VD (CAD) subgroups, EAT_LAD increased across the groups, whereas EATV index, EAT_RCA, and EAT_LCX were similar across the 4 groups (Figure 2). Previous reports have described associations between location-specific EAT thickness and obstructive CAD.^20–22 Wang et al showed that EAT thickness in the LAVG, but not EATV, was significantly associated with coronary atherosclerosis;²⁰ also, a meta-analysis confirmed that increases in location-specific EAT thicknesses at the LAVG are associated with obstructive CAD.²¹ Our previous study found that EAT_LAD was greater in the CAD group than in the non-CAD group, and that it had added diagnostic value over other conventional risk factors.²² Ohyama et al reported that the local EATV around the LAD, but not around the RCA or LCX, was significantly increased in patients with LAD spasms, suggesting the involvement of coronary EAT in the pathogenesis of coronary spasms, a form of atherosclerosis.^31,32

Thus, location-specific EAT thickness is more closely associated with the presence of coronary atherosclerosis than is global cardiac adiposity, but the EAT thickness location with the greatest CAD prognostic value remains controversial. Combining data from current and previous studies,^20–22 there are 2 possible explanations for the different impacts of location-specific EAT thicknesses. First, there are deeper gaps in atrioventricular grooves than in the AIVG. The paracrine effects of cytokines from EAT may be small in the RAVG and LAVG. Second, measurement of AIVG may be more accurate than for RAVG and LAVG because of anatomical characteristics, and thus its changes may sensitively reflect the presence of atherosclerosis. On ROC curve analysis for estimating CAD, of the EAT thicknesses at the RCA, LAD and LCX (Figure S1B), EAT_LAD had the largest area under the curve compared with EAT_RCA or EAT_LCX, further supporting this notion. The finding of significant correlation only between Gensini score in the LAD and EAT_LAD, but not between those in RCA and LCX (Figure S1A), may also explain the close association of EAT_LAD with the severity of CAD. Even in the non-LAD group, a borderline correlation was still observed between EAT_LAD and Gensini score (Figure 3I–L). We cannot denote a causality that EAT thickness of LAD affects to atherosclerosis of other 2 branches, however could suggest an interaction.

CAD Prediction Model

In the current study, cut-offs for predicting CAD were determined for FRS (>8%; sensitivity, 66%; specificity, 54%), EATV index (>90 cm³/m²; sensitivity, 33%; specificity, 88%), and EAT_LAD (>5.7 mm; sensitivity, 72%; specificity, 57%). Compared with only FRS >8% (model 1), the addition of EAT_LAD produced a stronger predictive value than the model combining EATV index and FRS (model 1 vs. model 3, P=0.014; model 2 vs. model 3, P=0.042). Therefore, we can assume that EAT_LAD, but not EATV index, provides additional prognostic value for predicting CAD, in combination with FRS. We first showed that the simple measurement of EAT_LAD is a useful surrogate marker of CAD in the clinical setting. Although FRS is a standard prediction model for CAD, its predictive power is limited. By adding EAT_LAD, simply measured on either echocardiography²² or CT (the present study), the model gains satisfactory prognostic power for predicting the presence of CAD.

Study Limitations

This study has potential limitations. First, the design of this clinical study was cross-sectional, and it was conducted at a single center with a relatively small number of subjects. Second, the subjects consisted entirely of Japanese patients, therefore the relevance of this study to other ethnic backgrounds awaits further research. Third, the ROC results cannot be applied broadly to patients with suspected CAD, and the sensitivity and specificity of EAT_LAD for diagnosing CAD need to be confirmed in a future, larger study. Fourth, we did not consider the impact of medication^33–35 or lifestyle³⁶ on EAT.

Conclusions

Local fat thickness in the area surrounding the LAD is a simple and useful measure for estimating the presence, severity, and extent of CAD, independent of classical coronary risk factors. Further investigation, involving a large-scale study, is warranted to confirm the clinical efficacy of EAT thickness in patients with suspected coronary atherosclerosis.

Acknowledgments

This work was supported by JSPS Grant-in-Aid for Scientific Research (no. 20590651, 23591314, 20590542, 16K01823 to M. Shimabukuro, no. 25670390 and 25293184 to M. Sata).

Disclosures

The authors declare no conflicts of interest.

Supplementary Files

Supplementary File 1

Figure S1. (A) Linear correlation between Gensini score in coronary vessels and EAT thickness in the area surrounding the RCA (EAT_RCA), LAD (EAT_LAD) and LCX (EAT_LCX) in patients with (○) non-CAD or (●) CAD.

Table S1. Differences in EAT parameters vs. CAD status

Table S2. EAT measures and possible covariates

Please find supplementary file(s);

http://dx.doi.org/10.1253/circj.CJ-17-1289

References

• 1.   Shimabukuro M, Hirata Y, Tabata M, Dagvasumberel M, Sato H, Kurobe H, et al. Epicardial adipose tissue volume and adipocytokine imbalance are strongly linked to human coronary atherosclerosis. Arterioscler Thromb Vasc Biol 2013; 33: 1077–1084.
• 2.   Hirata Y, Kurobe H, Akaike M, Chikugo F, Hori T, Bando Y, et al. Enhanced inflammation in epicardial fat in patients with coronary artery disease. Int Heart J 2011; 52: 139–142.
• 3.   Sacks HS, Fain JN. Human epicardial fat: What is new and what is missing? Clin Exp Pharmacol Physiol 2011; 38: 879–887.
• 4.   Mazurek T, Zhang L, Zalewski A, Mannion JD, Diehl JT, Arafat H, et al. Human epicardial adipose tissue is a source of inflammatory mediators. Circulation 2003; 108: 2460–2466.
• 5.   Matloch Z, Kotulak T, Haluzik M. The role of epicardial adipose tissue in heart disease. Physiol Res 2016; 65: 23–32.
• 6.   Rosito GA, Massaro JM, Hoffmann U, Ruberg FL, Mahabadi AA, Vasan RS, 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.
• 7.   Verhagen SN, Visseren FL. Perivascular adipose tissue as a cause of atherosclerosis. Atherosclerosis 2011; 214: 3–10.
• 8.   Ahn SG, Lim HS, Joe DY, Kang SJ, Choi BJ, Choi SY, et al. Relationship of epicardial adipose tissue by echocardiography to coronary artery disease. Heart 2008; 94: e7.
• 9.   Albuquerque FN, Somers VK, Blume G, Miranda W, Korenfeld Y, Calvin AD, et al. Usefulness of epicardial adipose tissue as predictor of cardiovascular events in patients with coronary artery disease. Am J Cardiol 2012; 110: 1100–1105.
• 10.   Raff GL, Abidov A, Achenbach S, Berman DS, Boxt LM, Budoff MJ, et al. SCCT guidelines for the interpretation and reporting of coronary computed tomographic angiography. J Cardiovasc Comput Tomogr 2009; 3: 122–136.
• 11.   Wannamethee SG, Shaper AG, Lennon L, Morris RW. Metabolic syndrome vs Framingham Risk Score for prediction of coronary heart disease, stroke, and type 2 diabetes mellitus. Arch Intern Med 2005; 165: 2644–2650.
• 12.   D’Agostino RB Sr, Pencina MJ, Massaro JM, Coady S. Cardiovascular Disease Risk Assessment: Insights from Framingham. Glob Heart 2013; 8: 11–23.
• 13.   Lloyd-Jones DM, Wilson PW, Larson MG, Beiser A, Leip EP, D’Agostino RB, et al. Framingham risk score and prediction of lifetime risk for coronary heart disease. Am J Cardiol 2004; 94: 20–24.
• 14.   Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation 1998; 97: 1837–1847.
• 15.   D’Agostino RB Sr, Grundy S, Sullivan LM, Wilson P. Validation of the Framingham coronary heart disease prediction scores: Results of a multiple ethnic groups investigation. JAMA 2001; 286: 180–187.
• 16.   Dagvasumberel M, Shimabukuro M, Nishiuchi T, Ueno J, Takao S, Fukuda D, et al. Gender disparities in the association between epicardial adipose tissue volume and coronary atherosclerosis: A 3-dimensional cardiac computed tomography imaging study in Japanese subjects. Cardiovasc Diabetol 2012; 11: 106.
• 17.   Austen WG, Edwards JE, Frye RL, Gensini GG, Gott VL, Griffith LS, et al. A reporting system on patients evaluated for coronary artery disease. Report of the Ad Hoc Committee for Grading of Coronary Artery Disease, Council on Cardiovascular Surgery, American Heart Association. Circulation 1975; 51: 5–40.
• 18.   Gensini GG. A more meaningful scoring system for determining the severity of coronary heart disease. Am J Cardiol 1983; 51: 606.
• 19.   Maimaituxun G, Shimabukuro M, Salim HM, Tabata M, Yuji D, Morimoto Y, et al. Gender-linked impact of epicardial adipose tissue volume in patients who underwent coronary artery bypass graft surgery or non-coronary valve surgery. PLoS One 2017; 12: e0177170.
• 20.   Wang TD, Lee WJ, Shih FY, Huang CH, Chen WJ, Lee YT, et al. Association of epicardial adipose tissue with coronary atherosclerosis is region-specific and independent of conventional risk factors and intra-abdominal adiposity. Atherosclerosis 2010; 213: 279–287.
• 21.   Wu FZ, Chou KJ, Huang YL, Wu MT. The relation of location-specific epicardial adipose tissue thickness and obstructive coronary artery disease: Systemic review and meta-analysis of observational studies. BMC Cardiovasc Disord 2014; 14: 62.
• 22.   Hirata Y, Yamada H, Kusunose K, Iwase T, Nishio S, Hayashi S, et al. Clinical utility of measuring epicardial adipose tissue thickness with echocardiography using a high-frequency linear probe in patients with coronary artery disease. J Am Soc Echocardiogr 2015; 28: 1240–1246.e1241.
• 23.   Iacobellis G. Local and systemic effects of the multifaceted epicardial adipose tissue depot. Nat Rev Endocrinol 2015; 11: 363–371.
• 24.   Silaghi A, Piercecchi-Marti MD, Grino M, Leonetti G, Alessi MC, Clement K, et al. Epicardial adipose tissue extent: Relationship with age, body fat distribution, and coronaropathy. Obesity (Silver Spring) 2008; 16: 2424–2430.
• 25.   Bays HE. Adiposopathy is “sick fat” a cardiovascular disease? J Am Coll Cardiol 2011; 57: 2461–2473.
• 26.   Lee HY, Despres JP, Koh KK. Perivascular adipose tissue in the pathogenesis of cardiovascular disease. Atherosclerosis 2013; 230: 177–184.
• 27.   Cherian S, Lopaschuk GD, Carvalho E. Cellular cross-talk between epicardial adipose tissue and myocardium in relation to the pathogenesis of cardiovascular disease. Am J Physiol Endocrinol Metab 2012; 303: E937–E949.
• 28.   Hassan M, Said K, Rizk H, ElMogy F, Donya M, Houseni M, et al. Segmental peri-coronary epicardial adipose tissue volume and coronary plaque characteristics. Eur Heart J Cardiovasc Imaging 2016; 17: 1169–1177.
• 29.   Xu J, Lu X, Shi GP. Vasa vasorum in atherosclerosis and clinical significance. Int J Mol Sci 2015; 16: 11574–11608.
• 30.   Hirata Y, Tabata M, Kurobe H, Motoki T, Akaike M, Nishio C, et al. Coronary atherosclerosis is associated with macrophage polarization in epicardial adipose tissue. J Am Coll Cardiol 2011; 58: 248–255.
• 31.   Ohyama K, Matsumoto Y, Nishimiya K, Hao K, Tsuburaya R, Ota H, et al. Increased coronary perivascular adipose tissue volume in patients with vasospastic angina. Circ J 2016; 80: 1653–1656.
• 32.   Nishio S, Kusunose K, Yamada H, Hirata Y, Ise T, Yamaguchi K, et al. Echocardiographic epicardial adipose tissue thickness is associated with symptomatic coronary vasospasm during provocative testing. J Am Soc Echocardiogr 2017; 30: 1021–1027.e1021.
• 33.   Cho KI, Kim BJ, Cha TJ, Heo JH, Kim HS, Lee JW. Impact of duration and dosage of statin treatment and epicardial fat thickness on the recurrence of atrial fibrillation after electrical cardioversion. Heart Vessels 2015; 30: 490–497.
• 34.   Shimabukuro M, Okawa C, Yamada H, Yanagi S, Uematsu E, Sugasawa N, et al. The pathophysiological role of oxidized cholesterols in epicardial fat accumulation and cardiac dysfunction: A study in swine fed a high caloric diet with an inhibitor of intestinal cholesterol absorption, ezetimibe. J Nutr Biochem 2016; 35: 66–73.
• 35.   Yagi S, Hirata Y, Ise T, Kusunose K, Yamada H, Fukuda D, et al. Canagliflozin reduces epicardial fat in patients with type 2 diabetes mellitus. Diabetol Metab Syndr 2017; 9: 78.
• 36.   Fornieles Gonzalez G, Rosety Rodriguez MA, Rodriguez Pareja MA, Diaz Ordonez A, Rosety Rodriguez J, Pery Bohorquez MT, et al. A home-based treadmill training reduced epicardial and abdominal fat in postmenopausal women with metabolic syndrome. Nutr Hosp 2014; 30: 609–613.