Several studies have suggested that epicardial adipose tissue (EAT) is associated with coronary artery disease (CAD). However, the role of EAT as a potential risk factor for, and predictor of, long-term cardiovascular outcomes in patients with CAD requires additional investigation. We investigated the relation among EAT, cardiovascular events, and measures of adiposity in patients with CAD. The study was a prospective cohort study of 194 consecutive patients with CAD who entered a phase II cardiac rehabilitation program at the Mayo Clinic. EAT was measured using echocardiography. The primary outcome was the long-term recurrence of major adverse cardiovascular events (MACE). The outcomes were assessed using the Mayo Clinic electronic medical records. The mean age was 59.4 ± 10.8 years, the body mass index was 28.7 ± 4.6 kg/m 2 , 80% were men, and 21% of the patients underwent coronary artery bypass grafting. The mean follow-up period was 3.6 ± 1.3 years, and 52 MACE occurred. EAT was not a predictor of MACE (hazard ratio 1.32, 95% confidence interval 0.75 to 2.31; p = 0.33) when used as a continuous variable and correlated poorly with the measures of adiposity. However, a nonsignificant trend was seen for a greater incidence of cardiovascular events when EAT was stratified by tertile (hazard ratio for third tertile 1.77, 95% confidence interval 0.84 to 3.32; p = 0.11), after statistical adjustments for age, gender, body mass index, and other covariates. In conclusion, the results of the present longitudinal study suggest that EAT, as measured using echocardiography, does not strongly predict for MACE and is poorly associated with measures of obesity in patients with CAD.
The role of epicardial adipose tissue (EAT) as a potential causal factor for the development of coronary artery disease (CAD) remains unclear. EAT could have a role as a source of proinflammatory cytokines, which might promote atherosclerosis, or through secretion of protective substances with insulin sensitizing and anti-inflammatory properties, such as adiponectin and adrenomedulin. Mostly cross-sectional or case-control studies have suggested an association between EAT and the risk of subclinical atherosclerosis and CAD, but other studies have questioned this association. Few longitudinal studies have tested the association between EAT and the incidence of CAD or the progression of subclinical CAD, and no studies have assessed the prognostic value of EAT in patients with CAD using clinical outcomes. Therefore, our specific aim was to investigate the relationships between EAT and the prognosis, recurrence of cardiovascular events, and the association with measures of adiposity in a cohort of patients with established CAD.
Methods
The study population consisted of 235 consecutive patients with CAD who entered the phase II cardiac rehabilitation program at the Mayo Cardiovascular Health Clinic from March 2004 to June 2009. The patients were enrolled in the study on the first day they entered the cardiac rehabilitation program, when they underwent body composition measurements using air displacement plethysmography (Bod Pod, Life Measurement, Concord, California).
A history of CAD was defined as previous myocardial infarction (ST- or non–ST-segment elevation myocardial infarction), unstable angina that required hospitalization, or previous revascularization using either coronary artery bypass grafting or percutaneous coronary intervention. Hypertension was defined as a mean systolic blood pressure of ≥140 mm Hg or a diastolic blood pressure of ≥90 mm Hg according to the Joint National Committee VII guidelines or the requirement for hypertensive medication. Dyslipidemia was defined as the use of a lipid-lowering agent, low-density lipoprotein cholesterol of ≥130 mg/dl, or total cholesterol of ≥200 mg/dl. Diabetes mellitus was defined as a fasting glucose level of ≥126 mg/dl. A history of heart failure, atrial fibrillation, and stroke were defined by evidence of the diagnosis in the clinical notes, and smoking was defined by a self-reported history of current smoking. All subjects were clinically stable without any evidence of uncontrolled hypertension, unstable or accelerated angina, or decompensated heart failure at enrollment. The institutional review board of the Mayo Clinic (Rochester, Minnesota) approved the study protocol.
Each patient underwent transthoracic echocardiography within 6 months of entering the rehabilitation program, and the measurements of EAT were calculated. EAT was measured as the echo-free area between the outer wall of the myocardium and the visceral layer of the pericardium on the free wall of the right ventricle at the parasternal long-axis view. We used a line that would intersect the aortic annulus as the reference that might or might not cross the thickest segment of EAT. The measurements were performed at end-systole in 3 cardiac cycles to avoid possible variations in the longitudinal thickness of the EAT, as previously described, and calculated the mean average of 3 measurements. The echocardiographic measurements of EAT were performed by a cardiologist who was unaware of the clinical and outcomes data, and a sample of 10 patients were measured in duplicate, in a blinded manner, to assess the intraobserver correlation of EAT measurements.
Each subject underwent body composition measurements using the Bod-Pod (Life Measurement), a validated air displacement plethysmography that measures the body volume by the pressure/volume relation (Boyle’s and Poisson’s laws) under isothermic conditions. It has been shown to reliably correlate with other body composition measurement techniques, such as dual x-ray absorptiometry and hydrodensitometry.
The primary outcome in the present study was the interval to develop a major adverse cardiovascular event (MACE). A MACE was defined as the diagnosis of acute coronary syndrome (i.e., ST- and non–ST-segment elevation myocardial infarction and unstable angina that required hospitalization), coronary revascularization (percutaneous coronary intervention and coronary artery bypass grafting), stroke (i.e., nontraumatic brain hemorrhage or infarction), ventricular arrhythmias requiring hospitalization, and death from any cause.
Patients were prospectively followed up through a review of the Mayo Clinic electronic medical records and the records-linkage system of the Rochester Epidemiology Project, which includes records from the Mayo Clinic and all other inpatient and outpatient medical providers in Olmstead County, Minnesota. The investigators who ascertained the follow-up outcomes were unaware of the patients’ clinical characteristics and EAT measurements. The outcomes data were abstracted by 1 investigator, and a sample of 20 subjects was reviewed in duplicate to assess the interobserver agreement for outcomes.
The baseline patient characteristics are summarized as frequencies and the mean ± SD. The EAT measurements were analyzed both as continuous variables and as an ordinal variable using the median and tertile values as cutoff points. Patients were considered at risk of an event only after the date of their enrollment in the study. We produced Kaplan-Meier survival curves with the MACE composite end point as the outcome of interest and performed log-rank tests to assess the difference in time free of MACE between patients with EAT values greater than and less than the median and using tertiles of EAT, with the first tertile as the reference. We constructed Cox proportional hazards models to test the association between EAT and MACE as either continuous variables or in tertiles of EAT and included the covariates of interest that might be potential confounders, such as age, gender, body mass index (BMI), and current smoking history. A second model included the same covariates plus dyslipidemia, diabetes mellitus, hypertension, a history of atrial fibrillation, a history of heart failure, and current smoking history. Spearman’s rank correlation coefficient was used to assess the correlation between EAT and measures of adiposity and to assess the intraobserver correlation of the EAT measurements. Weighted kappa statistics were used to assess the interobserver agreement for the outcomes. A receiver operating characteristics curve was performed to assess the optimal cutoff of EAT to predict the outcome. Analyses were performed with JMP, version 7 (SAS Institute, Cary, North Carolina) and R software (R Foundation for Statistical Computing, Vienna, Austria). For all comparisons, 2-tailed p <0.05 was considered statistically significant, and the findings were summarized using hazard ratios (HRs) and corresponding 95% confidence intervals (CIs).
Results
The baseline demographic and clinical characteristics of the study group are summarized in Table 1 . A total of 41 patients were excluded from the analysis because of a lack of appropriate echocardiographic images for the EAT measurement. The mean age and BMI of the 194 patients included in the present study was 59.4 ± 10.8 years and 28.7 ± 4.6 kg/m 2 , respectively; 80% of the patients were men, and 95% were non-Hispanic whites. The baseline characteristics of the population, prevalence of co-morbid conditions, and use of cardiac medications are listed in Table 1 . The interobserver agreement for outcomes was 85%, and the intraobserver correlation of EAT measurements was strong (Spearman ρ = 0.93; R 2 = 0.90; p <0.001).
| Variable | Value |
|---|---|
| Age (years) | 59.4 ± 10.8 |
| Men | 80% |
| Body weight (kg) | 86.0 ± 16 |
| Height (cm) | 172.6 ± 8.7 |
| Body mass index (kg/m 2 ) | 28.7 ± 4.6 |
| Body fat (%) | 33.5 ± 8.2 |
| Fat mass (kg) | 29.4 ± 10.7 |
| Lean mass (kg) | 56.7 ± 10.6 |
| Epicardial adipose tissue (cm) | 0.93 ± 0.25 |
| Race | |
| Non-Hispanic white | 95% |
| Other | 5% |
| Systolic blood pressure (mm Hg) | 116.6 ± 15.6 |
| Diastolic blood pressure (mm Hg) | 68.7 ± 8.7 |
| Fasting glucose (mg/dl) | 118.2 ± 40 |
| Total cholesterol (mg/dl) | 160.8 ± 43.7 |
| Triglycerides (mg/dl) | 134.8 ± 74 |
| High-density lipoprotein (mg/dl) | 47.6 ± 13.4 |
| Low-density lipoprotein (mg/dl) | 86.9 ± 36 |
| Left ventricular ejection fraction (%) | 55.5 ± 10.8 |
| Diabetes mellitus | 20% |
| Dyslipidemia | 93% |
| Hypertension | 65% |
| History of atrial fibrillation | 5% |
| History of heart failure | 5% |
| History of stroke | 1% |
| History of current smoking | 12% |
| Aspirin | 95% |
| Clopidogrel | 64% |
| Diuretics | 20% |
| β Blockers | 91% |
| Angiotensin-converting enzyme inhibitors | 52% |
| Statins | 94% |
| Nitrates | 14% |
On stratification of our cohort into 2 groups using the median value of EAT as the cutoff, the patients in the greater than median group for EAT were older and had a larger fat mass; however, no significant differences were found in BMI, lean mass, or waist circumference. The mean systolic and diastolic blood pressure, fasting glucose, triglycerides, cholesterol levels, prevalence of diabetes mellitus, hypertension, dyslipidemia, atrial fibrillation, and heart failure were also not significantly different between the 2 groups ( Table 2 ).
| Variable | EAT (cm) | p Value | |
|---|---|---|---|
| <0.94 (n = 96) | >0.94 (n = 98) | ||
| Age (years) | 57.1 ± 1.1 | 61.8 ± 1.1 | 0.003 |
| Men | 47.4% | 52.6% | 0.26 |
| Body mass index (kg/m 2 ) | 28.1 ± 0.4 | 29.0 ± 0.4 | 0.18 |
| Body fat (%) | 32.8 ± 0.8 | 34.8 ± 0.8 | 0.09 |
| Fat mass (kg) | 27.5 ± 1.1 | 31.0 ± 1.1 | 0.02 |
| Lean mass (kg) | 54.7 ± 1.1 | 56.8 ± 1.1 | 0.16 |
| Waist (cm) | 97.1 ± 1.4 | 100.2 ± 1.3 | 0.1 |
| Non-Hispanic white | 48% | 52% | 0.34 |
| Systolic blood pressure (mm Hg) | 116.4 ± 1.6 | 116.7 ± 1.6 | 0.88 |
| Diastolic blood pressure (mm Hg) | 69.2 ± 0.9 | 67.9 ± 0.9 | 0.33 |
| Fasting glucose (mg/dl) | 117.7 ± 4.6 | 111.6 ± 5.2 | 0.38 |
| Total cholesterol (mg/dl) | 161.3 ± 4.7 | 158.2 ± 4.8 | 0.65 |
| High-density lipoprotein (mg/dl) | 47.1 ± 1.7 | 48.2 ± 1.7 | 0.65 |
| Low-density lipoprotein (mg/dl) | 85.5 ± 4.0 | 84.7 ± 4.0 | 0.88 |
| Triglycerides (mg/dl) | 130.4 ± 8.4 | 132.6 ± 8.5 | 0.85 |
| Diabetes mellitus | 43% | 57% | 0.32 |
| Dyslipidemia | 51% | 49% | 0.1 |
| Hypertension | 53.7% | 46.3% | 0.13 |
| History of atrial fibrillation | 50% | 50% | 0.97 |
| History of heart failure | 29% | 71% | 0.25 |
After a median follow-up of 3.9 ± 1.3 years, 52 patients had a MACE. Of these patients, 35 (67%) had had an acute coronary syndrome, 12 (23%) had died from any cause, 3 (6%) had had a stroke, and 2 (4%) had had ventricular arrhythmia that required hospitalization. The prevalence of cardiovascular events according to the EAT tertiles was 15 (29% in the first tertile), 14 (27% in the second tertile), and 23 (44% in the third tertile). On univariate analysis, using EAT as a continuous variable, EAT was not a predictor of MACE (HR 1.32, 95% CI 0.75 to 2.31, p = 0.33; Table 3 ). When we stratified the participants into 2 groups using the EAT median value as the cutoff, no significant difference was found in MACE between the 2 groups (HR for EAT greater than the median 1.23, 95% CI 0.73 to 2.08; p = 0.43; Figure 1 ). When stratified in tertiles of EAT, a nonstatistically significant trend was seen toward greater cardiovascular events in patients in the third EAT tertile (HR 1.41, 95% CI 0.75 to 2.64, p = 0.17; Figure 2 ). On multivarariate Cox proportional hazards regression model analysis, adjusting for age, gender, BMI, and current smoking history, EAT as a continuous variable was not a predictor of MACE (HR 2.53, 95% CI 0.84 to 7.60; p = 0.10). In the second multivariate model, adjusting also for dyslipidemia, diabetes mellitus, hypertension, a history of atrial fibrillation, and a history of heart failure, the nonstatistically significant trend toward a greater incidence of cardiovascular events persisted in the patients in the greatest tertile of EAT (HR for the third EAT tertile 1.77, 95% confidence interval 0.88 to 3.52; p = 0.11; Table 3 ). Moreover, the area under the receiver operating characteristics curve was so low at 0.58 that no optimal cutoff of EAT was found to predict MACE.
| Variable | HR (95% CI) Unadjusted | p Value | HR (95% CI) Adjusted ⁎ | p Value |
|---|---|---|---|---|
| Age (years) | 1.00 (0.97–1.03) | 0.96 | 1.0 (0.97–1.03) | 0.91 |
| BMI (kg/m 2 ) | 1.03 (0.95–1.11) | 0.48 | 1.0 (0.94–1.07) | 0.98 |
| Gender (male) | 1.47 (0.67–3.23) | 0.34 | 1.58 (0.71–3.50) | 0.26 |
| Dyslipidemia | 1.10 (0.38–3.20) | 0.86 | 0.94 (0.31–2.83) | 0.91 |
| Diabetes mellitus | 1.32 (0.70–2.48) | 0.39 | 1.13 (0.59–2.16) | 0.71 |
| Hypertension | 1.78 (1.03–3.10) | 0.04 | 1.86 (0.98–3.54) | 0.06 |
| History of atrial fibrillation | 2.00 (0.74–5.41) | 0.17 | 1.84 (0.69–4.95) | 0.22 |
| History of heart failure | 0.44 (0.10–2.02) | 0.29 | 0.35 (0.07–1.72) | 0.20 |
| History of current smoking | 2.66 (1.10–6.40) | 0.03 | 2.56 (1.10–5.88) | 0.03 |
| Epicardial adipose tissue (cm) | 1.32 (0.75–2.31) | 0.33 | 2.37 (0.73–7.71) | 0.15 |
| Epicardial adipose tissue tertiles (cm) | ||||
| First tertile | Reference | Reference | ||
| Second tertile | 0.81 (0.40–1.64) | 0.56 | 0.86 (0.39–1.90) | 0.71 |
| Third tertile | 1.41 (0.75–2.65) | 0.17 | 1.77 (0.84–3.32) | 0.11 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
