Abstract
We investigated the relationship between epicardial fat volume (EFV) measured by multislice computed tomography (MDCT) and long-term major adverse cardiac events (MACEs). Consecutive patients (n = 564) were enrolled in this retrospective study. Patients were divided into tertiles according to EFV. Patients were followed up for an average of 18 months. Patients in each tertile were similar in terms of gender and risk factors. Patients with greater EFV in the third group were more likely to be overweight (P = .001) and older (P = .001). High-density lipoprotein cholesterol levels were relatively lower in the third tertile (45 ± 9, 45 ± 11, and 43 ± 9 mg/dL, respectively; P = .018). The third group had a significantly higher rate of myocardial infarction (0.6%, 1.1%, and 3.7%, respectively; P = .043). The incidence of MACEs during the follow-up period was highest in the third group 15.9% (4.1%, 7.7%, and 15.9%, respectively; P = .001). Epicardial fat volume measured by MDCT was associated with increased long-term cardiovascular risk.
Introduction
Epicardial fat tissue (EFT) functions as a visceral fat deposit and shares the same embryonic origin as intra-abdominal adipose tissue.1–3 Epicardial fat tissue is located between the myocardium and the pericardium and is supplied by the same microcirculatory system that perfuses the myocardium.2 Epicardial fat tissue protects the myocardium by producing anti-inflammatory agents including adrenomedullin, adiponectin, and antiatherogenic adipokines. However, this tissue may also exert harmful effects on the myocardium by releasing a number of proinflammatory and proatherogenic cytokines.4–7 Several studies have reported associations between metabolic syndrome, cardiovascular (CV) disease, and coronary artery disease (CAD) with the amount of visceral adipose tissue and EFT mass.8–15
Several techniques can quantify the volume of EFT; each technique has its own advantages and disadvantages.16–18 It is difficult to clearly distinguish epicardial and pericardial adipose tissues via echocardiography, and the sensitivity of this modality is especially limited in obese patients.16 Epicardial fat tissue is heterogeneously distributed on the heart. For example, the anterior aspect of the right ventricular surface is a relatively small collection of EFT, which is difficult to accurately measure by echocardiography.16 Nevertheless, echocardiography compared with other EFT measurement techniques is relatively cost effective and more accessible, and it provides information about heart structure and function.16–18
Magnetic resonance imaging (MRI) may also be used to quantify EFT, and unlike measurements made with multislice computed tomography (MSCT), MRI has high sensitivity and specificity without exposing patients to radiation.17 Yet, MRI is more time consuming, expensive, and not practical for all patients. Multislice computed tomography is a high-resolution modality that provides the most accurate EFT measurements.18 In addition to EFT quantification, MSCT can also evaluate coronary artery health. In our study, we investigated whether there is an association between epicardial fat volume (EFV), as measured by MSCT, and the incidence of long-term major adverse cardiac events (MACEs) in patients at moderate risk for CAD.
Methods
Study Population
Between May 2009 and December 2012, a total of 750 patients at moderate risk for CAD were admitted to the cardiology service and underwent MSCT. Patients with lost or inaccessible laboratory results (n = 20), with no follow-up appointment records (n = 34), and patients who could not be contacted (n = 51) were excluded. Moreover, patients presenting with signs and symptoms of acute coronary syndrome (n = 24), atrial fibrillation (n = 17), active chronic obstructive pulmonary disease (n = 19), and decompensated heart failure (n = 21) were also excluded. Data from the remaining 564 patients were evaluated; 485 were males and the mean age was 54.7 ± 12.7 years. Clinical presentation upon admission, CAD risk factors, and medications were recorded for each patient. Body mass index was calculated by dividing patient weight (kg) by the square of their height (m). Laboratory results were obtained via the Sysmex KX-21N autoanalyzer (Sysmex Corporation, Lincolnshire, IL). The ethics committee approved the experimental protocol.
Multislice Computed Tomography
Multislice Computed Tomography was performed with a Somatom Sensation 64 (Siemens, Forchheim, Germany). Adipose tissue demonstrates an attenuation of between −200 and −50 HU. Epicardial fat volume was measured in cm3 by volumetric analysis using a cardiac workstation software tool (Siemens, Leonardo) and manual region of interest drawings. Patients were divided into tertiles according to EFV: the first group had an EFV <114.8 cm3, the second group had an EFV between 114.8 and 174.5 cm3, and the third group had an EFV >174.5 cm3.
Long-Term Major Adverse Cardiac Events and Clinical Follow-Up
Follow-up data were obtained from hospital records or by conducting interviews in person or over the telephone with patients, families, or the patient’s primary care physician. A MACE was defined as cardiac death, nonfatal myocardial infarction (MI), heart failure, a revascularization procedure (either coronary artery bypass grafting or percutaneous coronary intervention), and stroke following MSCT during the follow-up period. Cardiac death was defined as mortality due to a heart-related cause. A nonfatal MI was characterized by the development of new electrocardiographic changes and/or recurrent chest pain accompanied by a new increase in cardiac markers by at least 20%. Heart failure was diagnosed if the patient met the appropriate Framingham criteria.19
Statistical Analysis
The Kolmogorov-Smirnov test was used to determine whether continuous variables conformed to a normal distribution. Data between groups were compared via 1-way analysis of variance. Categorical variables were expressed as percentages and were analyzed with the chi-square test. A 2-sided P < .05 was considered significant. The receiver–operating characteristic (ROC) curve was used to determine the optimal cutoff EFT volume for predicting of MACEs during the follow-up period. All statistical analyses were performed with SPSS version 15.0 (SPSS Inc, Chicago, Illinois).
Results
The first group consisted of 163 males (mean age 51.5 ± 11.3 years), the second group was comprised of 159 males (average age 53.1 ± 12.1 years), and the third group included 163 males (mean age 59.5 ± 13.4 years). No differences were observed in gender and CAD risk factors between each group (Table 1). Although there were no differences between the groups in terms of history of peripheral arterial disease, the third group had a significantly higher rate of MI (P = .043). Patients with greater EFV in the third group were more likely to be overweight (P = .001) and older (P = .001). Total cholesterol, low-density lipoprotein cholesterol, and triglyceride levels were similar between the groups, while high-density lipoprotein (HDL) cholesterol levels in the third group were relatively lower (P = .018). Patients in the third group had significantly higher fasting blood glucose levels as well (P = .043). Other hematological and biochemical markers were similar between each group (Table 2).
Table 1. Baseline Characteristics.a
| Variable | Epicardial Fat Tissue | |||
|---|---|---|---|---|
| Tertile 1 (<115; n = 194) | Tertile 2 (115–175; n = 181) | Tertile 3 (>175; n = 189) | P | |
| Age (years) | 51.5 ± 11.3 | 53.1 ± 12.1 | 59.5 ± 13.4 | .001 |
| Sex (male) | 163 (84%) | 159 (88%) | 163 (87%) | .543 |
| Body mass index | 26.4 ± 2.4 | 27.1 ± 3.1 | 30.6 ± 5.5 | .001 |
| Coronary risk factors | ||||
| Hypertension | 57 (29%) | 47 (26%) | 62 (33%) | .319 |
| Diabetes mellitus | 30 (16%) | 24 (13%) | 40 (21%) | .101 |
| Current smoking | 110 (57%) | 107 (59%) | 116 (62%) | .610 |
| Hyperlipidemia | 48 (25%) | 49 (27%) | 55 (29%) | .610 |
| Heredity | 109 (56%) | 110 (61%) | 114 (61%) | .610 |
| Triglycerides, mg/dL | 157 ± 68 | 159 ± 76 | 172 ± 65 | .066 |
| Low-density lipoprotein cholesterol, mg/dL | 120 ± 32 | 119 ± 30 | 120 ± 31 | .945 |
| High-density lipoprotein cholesterol, mg/dL | 45 ± 9 | 45 ± 11 | 43 ± 9 | .018 |
| Total cholesterol, mg/dL | 200 ± 38 | 194 ± 42 | 199 ± 48 | .962 |
| Medications | ||||
| ACE inhibitors | 16 (8%) | 12 (7%) | 15 (8%) | .821 |
| b-Blocker | 24 (12%) | 27 (15%) | 35 (19%) | .388 |
| Statin | 39 (20%) | 42 (23%) | 50 (26%) | .339 |
| Aspirin | 24 (12%) | 24 (13%) | 27 (14%) | .849 |
| Antiplatelet agent | 18 (9%) | 17 (9%) | 22 (12%) | .679 |
| Oral antidiabetic agents | 28 (14%) | 22 (12%) | 22 (12%) | .693 |
| Angiotensin receptor blocker | 21 (11%) | 23 (13%) | 29 (15%) | .946 |
| Calcium antagonist | 10 (5%) | 12 (7%) | 18 (10%) | .233 |
| Peripheral artery disease | 2 (1.0%) | 1 (0.6%) | 3 (1.6%) | .384 |
| Myocardial infarction history | 0 (0.0%) | 2 (1.1%) | 7 (3.7%) | .012 |
| CABG history | 1 (0.6%) | 2 (1.1%) | 7 (3.7%) | .043 |
Abbreviations: ACE, angiotensin converting enzyme inhibitors; CABG, coronary artery bypass graft operation.
aData are expressed as mean ± standard deviation for normally distributed data and percentage (%) for categorical variables. The P values lower than 0 .05 was considered significant and made boldface.
Table 2. Laboratory Parameters of the Study Populations.a
| Variable | Epicardial Fat Tissue | |||
|---|---|---|---|---|
| Tertile 1 (<115; n = 194) | Tertile 2 (115–175; n = 181) | Tertile 3 (>175; n = 189) | P | |
| Hemoglobin, g/L | 14.3 ± 1.4 | 15.2 ± 1.3 | 14.9 ± 1.5 | .238 |
| Platelet count, 103/mm3 | 243 ± 59 | 242 ± 53 | 244 ± 65 | .964 |
| White blood cell count, 103/μL | 7.7 ± 1.3 | 7.5 ± 1.4 | 7.58 ± 1.66 | .625 |
| Serum glucose, mg/dL | 104 ± 22 | 106 ± 23 | 111 ± 39 | .043 |
| Creatinine, mg/dL | 0.87 ± 0.41 | 0.81 ± 0.17 | 0.86 ± 0.20 | .101 |
| Alanine aminotransferase, U/L | 33 ± 18 | 32 ± 21 | 31 ± 20 | .834 |
| C-reactive protein, mg/L | 5.4 ± 8.2 | 6.8 ± 5.8 | 6.1 ± 4.5 | .455 |
| Hematocrit, % | 43 ± 3.7 | 45 ± 3.3 | 44 ± 5.3 | .950 |
| Triglycerides, mg/dL | 157 ± 68 | 159 ± 76 | 172 ± 65 | .066 |
| Low-density lipoprotein cholesterol, mg/dL | 120 ± 32 | 119 ± 30 | 120 ± 31 | .945 |
| High-density lipoprotein cholesterol, mg/dL | 45 ± 9 | 45 ± 11 | 43 ± 9 | .018 |
| Total cholesterol, mg/dL | 200 ± 38 | 194 ± 42 | 199 ± 48 | .962 |
aData are expressed as mean ± standard deviation for normally distributed data and percentage (%) for categorical variable. The P values lower than 0 .05 was considered significant and made boldface.
OPEN IN VIEWER
Patients received follow-up care after MSCT for an average of 18 months, and the shortest follow-up period was 10 months while the longest was 52 months. In terms of CV mortality (P = .193), congestive heart failure (P = .963), and cerebrovascular events (P = .346), each group did not differ significantly from one another. The rates of CABG were 1.0%, 0.6%, and 3.7% for the first, second, and third groups, respectively; only the third group had a significantly higher rate of CABG (P = .045). In addition, the number of patients who sustained nonfatal MIs was 0.5% in the first group, 2.2% in the second group, and 4.8% in the third group (P = .027). Patients who subsequently required percutaneous coronary intervention were 2.1%, 4.4%, and 7.4% in the first, second, and third groups, respectively (P = .044).
The incidence of MACEs during the follow-up period was determined to be highest in the third group at 15.9%, while the first and second group incidences were 4.1% and 7.7%, respectively (P = .001) (Table 3). From the ROC analysis, we determined that the cutoff EFV was >180 cm3, and this value was found to have a 74% sensitivity and a 65% specificity in predicting MACEs during the follow-up period (Figure 1).
Table 3. Long-Term Follow-Up for Major Adverse Cardiac Events Stratified by Epicardial Fat Tissue.
| Epicardial Fat Tissue | ||||
|---|---|---|---|---|
| Tertile 1 (<115; n = 194) | Tertile 2 (115-175; n = 181) | Tertile 3 (>175; n = 189) | P | |
| Long-term MACE | 8 (4.1%) | 14 (7.7%) | 30 (15.9%) | .001 |
| Nonfatal MI | 1 (0.5%) | 4 (2.2%) | 9 (4.8%) | .027 |
| Congestive heart failure | 6 (3.1%) | 5 (2.8%) | 5 (2.6%) | .963 |
| Cerebrovascular event | 0 (0.0%) | 1 (0.6%) | 0 (0.0%) | .346 |
| Mortality | 1 (0.5%) | 2 (1.1%) | 5 (2.6%) | .193 |
| PCI | 4 (2.1%) | 8 (4.4%) | 14 (7.4%) | .044 |
| CABG | 2 (1.0%) | 1 (0.6%) | 7 (3.7%) | .045 |
Abbreviations: MACE, major adverse cardiac event; MI, myocardial infarction; PCI, percutaneous coronary intervention; CABG, coronary artery bypass graft operation.
OPEN IN VIEWER
Figure 1. The ROC curve of EFV levels for predicting MACEs during the follow-up period. EFV indicates epicardial fat volume; MACEs, major adverse cardiac events; ROC, receiver–operating characteristic.
OPEN IN VIEWER
Discussion
We investigated whether there is a relationship between EFV as measured by MSCT and the incidence of MACEs in patients at moderate risk of CAD. We demonstrated that over the long-term, patients with higher EFV in the third group had significantly higher rates of MACEs. Specifically, an EFV >180 cm3 had a 74% sensitivity and a 65% specificity in predicting the occurrence of an MACE.
Epicardial fat tissue has been suggested to play a role in the development of oxidative stress and atherosclerosis via an inflammatory-mediated process.8,14,20–23 Nabati and colleagues investigated the relationship between EFT measured by echocardiography and coronary atherosclerosis and showed EFT thickness3 7 mm may identify an individual with higher probability of having coronary atherosclerosis.13 den Dekker et al reported CAD is correlated with EFT but not with mediastinal adipose tissue and suggested that EFT has a local effect on coronary atherosclerosis, apart from the endocrine effect of visceral fat.14 Similarly, a meta-analysis suggested that significantly elevated location-specific EFT thickness appears to be a good predictor in obstructive CAD.24 In another study performed with 202 patients, a significant correlation between EFT and CAD severity suggested that EFT might be used in risk stratifying these patients.25 Others measured EFT by echocardiography in 527 patients who received their first coronary angiography procedure. They reported that increased EFT thickness is associated with an increased likelihood of having clinical and laboratory symptoms and signs of CAD, insulin resistance, and inflammation.12 In another study, 286 patients with moderate CAD, pericardial fat volumes were measured by MSCT, and a strong correlation was found between greater pericardial fat volumes and the presence of CAD. It has been suggested that >300 cm3 of pericardial fat is more indicative of the development of CAD than the Framingham score.18 Beyond all these, Park et al reported that echocardiographically measured EFT thickness is related to restenosis after coronary stenting and suggested EFT thickness may provide additional information for future restenosis.26 Our study is consistent with previous studies as we also demonstrated that an increase in EFV is significantly associated with higher rates of MACEs in the long term.
Previous studies have investigated the relationship between age, obesity, and EFV.16,25,27,28 In our study, we demonstrated that EFV increased with increasing age and obesity. Moreover, we found that serum lipids were dysregulated as triglycerides and HDL cholesterol levels were elevated and decreased, respectively, in patients with greater EFV, which is consistent with increased CV risk. Alexopoulos et al reported statin therapy reduced EFT regression, although intensive therapy was more effective than moderate-intensity therapy in postmenopausal women with hyperlipidemia.29 The quantity of visceral fat is an indicator of adverse cardiac outcomes and the presence of metabolic syndrome.10,18 Iacobellis and Leonetti conducted an investigation with 30 obese patients who showed a significant relationship between insulin resistance and epicardial fat thickness measured by echocardiography.20 In another study of 60 healthy volunteers, MRI and echocardiography were used to measure visceral and EFV. In addition, waist circumferences were taken to estimate the quantity visceral fat. It was determined that more precise methods to measure visceral fat improve risk predictions for metabolic syndrome, diabetes mellitus, and CV disease.21
The amount of EFT is usually measured by echocardiography, MRI, or MSCT; each modality has its own advantages and disadvantages.8,16–18 The sensitivity of echocardiography is reduced when distinguishing between epicardial and pericardial fat, especially for patients with obesity and chronic obstructive pulmonary disease. Epicardial fat tissue is heterogeneously distributed on the heart and is especially sparse on the anterior aspect of the right ventricle.8,16 As fat is heterogeneously distributed, echocardiography may provide inaccurate results if measurements are performed in only one location.8,16 Yet the advantages of echocardiography include its accessibility in terms of cost and the cardiac function data that it provides. Magnetic resonance imaging can also measure EFV, and this modality has a high sensitivity and specificity without exposing the patient to radiation.17 Unfortunately, MRI is time consuming and expensive and so is not feasible or practical for many patients such as those with obesity or claustrophobia.
Multislice computed tomography is another imaging modality that measures EFV with high resolution and provides the most accurate results such that it is the gold standard.18 In addition, MSCT simultaneously evaluates the coronary vessels, heart, and noncardiac pathology. The results of our study suggest that the benefits of obtaining an accurate EFV may outweigh the risk of radiation exposure, as MSCT provides clinical insight into CV health risks.
Conclusion
Epicardial fat volume measured by MSCT may predict the likelihood of MACEs for patients who are at moderate risk of CAD. Thus, measuring the EFV may be an additional clinical indicator of long-term CV risk for patients.