Volumetric Assessment of Epicardial Adipose Tissue With Cardiovascular Magnetic Resonance Imaging
Abstract
Objective: Previous studies determined the amount of epicardial fat by measuring the right ventricular epicardial fat thickness. However, it is not proven whether this one-dimensional method correlates well with the absolute amount of epicardial fat. In this prospective study, a new cardiovascular magnetic resonance imaging (CMR) method using the three-dimensional summation of slices method was introduced to assess the total amount of epicardial fat.
Research Methods and Procedures: CMR was performed in 43 patients with congestive heart failure and in 28 healthy controls. The absolute amount of epicardial fat was assessed volumetrically in consecutive short-axis views by means of the modified Simpson's rule. Additionally, the right ventricular epicardial fat thickness was measured in two different imaging planes: long-axis view (EFT-4CV) and consecutive short-axis views (EFT-SAX).
Results: Using the volumetric approach, patients with congestive heart failure had less epicardial fat mass than controls (51 g vs. 65 g, p = 0.01). This finding was supported by EFT-SAX (2.9 mm vs. 4.3 mm, p < 0.0001) but not by EFT-4CV (3.5 mm vs. 3.8 mm, p = not significant). Epicardial fat mass correlated moderately with EFT-SAX in both groups (r = 0.466, p = 0.012 in controls and r = 0.590, p < 0.0001 in patients) and with EFT-4CV in controls (r = 0.387, p = 0.042). There were no significant differences between EFT-4CV and EFT-SAX in controls (4.3 mm vs. 3.8 mm, p = 0.240). However, in the heart failure group, EFT-4CV was significantly higher compared with EFT-SAX (3.5 mm vs. 2.9 mm, p = 0.003). Interobserver variability and reproducibility were superior for the volumetric approach compared with thickness measurements.
Discussion: Quantitative assessment of epicardial fat mass using the CMR-based volumetric approach is feasible and yields superior reproducibility compared with conventional methods.
Introduction
Visceral obesity plays a key role in the development of an unfavorable metabolic and cardiovascular risk profile (1, 2, 3, 4, 5). The epicardial adipose tissue is the visceral fat deposited around the heart between the pericardium and the myocardium. This small, visceral fat depot is now recognized as a rich source of free fatty acids and a number of bioactive molecules, such as adiponectin, resistin, and inflammatory cytokines, which might significantly affect cardiac pathology (6, 7). Furthermore, epicardial fat correlates strongly with intra-abdominal visceral fat (8, 9), left ventricular mass (10, 11), and several metabolic syndrome parameters (9). Therefore, non-invasive assessment of epicardial fat could be a simple and reliable indicator of the amount of visceral adipose tissue and cardiovascular risk (1).
Recent studies proposed the use of echocardiography for direct assessment of epicardial adipose tissue (8, 9). In these studies, epicardial fat thickness was measured on the right ventricular free wall from both parasternal long-axis and short-axis views in healthy subjects. All subjects also underwent cardiovascular magnetic resonance (CMR)1 imaging. Echocardiographic assessment of adipose tissue on the free wall of the right ventricle showed excellent correlation with CMR measurements of epicardial fat (8). However, it is well known from anatomical studies that the epicardial adipose tissue is not uniformly distributed around the heart. Epicardial fat is prominent near the interventricular and the atrioventricular grooves. The greatest amount of epicardial fat can be found at the lateral right ventricular wall, followed by the anterior wall (12, 13). Moreover, the distribution of epicardial fat seems to show considerable inter-individual differences. Therefore, it is unclear whether thickness of right ventricular epicardial fat measured at single points correlates well with the absolute amount of epicardial fat.
Magnetic resonance imaging can be regarded as the gold standard for the estimation of whole body adipose tissue (14, 15, 16). Furthermore, it has proven useful for the detection of fatty tissue in and around the heart (17), and it is an established reference standard for the assessment of left ventricular volumes and mass, using the summation of slices method (18, 19, 20, 21). It can assess the entire heart in consecutive layers and is, therefore, not limited by the position and orientation of the imaging planes.
The aim of our study was to introduce a new CMR method using the three-dimensional summation of slices method for the assessment of the amount of epicardial fat in controls and in patients with congestive heart failure and to correlate these results with measurements of the right ventricular epicardial fat thickness.
Research Methods and Procedures
Study Population
A total of 43 patients (35 males and 8 females; mean age, 62 ± 12 years) with congestive heart failure (19 patients with dilated cardiomyopathy: 4 females; mean age, 57 ± 15 years, 24 patients with coronary artery disease: 4 females; mean age, 65 ± 9 years), and 28 healthy controls (22 males and 6 females; mean age, 57 ± 11 years) were included in the study. All patients and volunteers underwent CMR examination using the same protocol. Informed consent for the CMR protocol was obtained from all subjects.
Image Acquisition
All studies were performed using a 1.5-Tesla whole-body imaging system (Magnetom Sonata, Siemens Medical Systems, Erlangen, Germany). A dedicated 4-element, phased-array cardiac coil was used. Images were acquired during repeated end-expiratory breath-holds. Scout images (coronal, sagittal, and axial planes) were obtained for planning of the final double-oblique long-axis and short-axis views. To evaluate functional parameters, electrocardiogram-gated cine images were then acquired using a segmented steady-state free precession [fast imaging with steady-state precession (true-FISP)] sequence (time to echo/time of repetition 1.6/3.2 ms, temporal resolution 35 ms, in-plane spatial resolution 1.4 × 1.8 mm, slice thickness 5 mm, interslice gap 5 mm). Seven to 12 short-axis views covering the whole left and right ventricle were obtained.
For the assessment of the epicardial adipose tissue, we used a dark blood prepared T1-weighted multislice turbo spin-echo pulse sequence with a water suppression prepulse to obtain a transversal 4-chamber view and short-axis images in the same orientations used for the cine short-axis images. Imaging parameters were as follows: time of repetition = 800 ms, time to echo = 24 ms, slice thickness = 4 mm, interslice gap = 2 mm, and field of view = 30 to 34 cm.
Image Analysis and Determination of Ventricular Parameters
Image analysis and quantitative analysis was performed off-line using dedicated software (ARGUS, Siemens). Each study was examined for abnormalities in the morphology of the right and left ventricle. End-diastolic and end-systolic volumes and left ventricular mass was analyzed with the serial short-axis true-FISP cine loops, using manual segmentation. Stroke volumes and ejection fractions were calculated. Additionally, left- and right-ventricular diameters were measured.
Measurement of Right Ventricular Epicardial Fat Thickness
Maximum epicardial fat thickness (EFT) at the right ventricular free wall was measured in different views: in a transversal 4-chamber view (EFT-4CV; Figure 1A) and in consecutive short-axis views (EFT-SAX) covering the whole ventricle (Figure 1B). Fat thickness in all short-axis images was averaged to give one final value for EFT-SAX in each patient. Additionally, in each patient, mean epicardial fat thickness (EFT-4CV/SAX) was calculated by averaging results from the long-axis and short-axis measurements. Epicardial fat in and near the atrioventricular and interventricular grooves was not considered.
Epicardial fat thickness of the right ventricular free wall measured in the four-chamber view (A) and in a representative short-axis view (B).
Volumetric Assessment of the Absolute Mass of Epicardial Adipose Tissue
The amount of epicardial fat was calculated by using the modified Simpson's rule with integration over the image slices: The contours of epicardial adipose tissue were outlined at end diastole in the short-axis views covering the entire left and right ventricle (Figure 2).
Volumetric measurement of epicardial fat mass. The contours of epicardial adipose tissue were outlined in end-diastolic images of short-axis views covering the whole left and right ventricle.
For epicardial fat mass determination, the area subtended by the manual tracings was determined in each slice and multiplied by the slice thickness to yield the fat volume. Total epicardial fat volume was obtained after the summation of data of all slices. To obtain epicardial fat mass, the volume of epicardial fat was multiplied by the specific weight of fat (0.92 g/cm3) (22).
To assess interobserver reproducibility, a second independent observer (T.P.) repeated epicardial fat measurements in each dataset using the same conventions. The observer was blinded to patient details and to the findings of the first observer.
Statistical Analysis
The data are presented as mean value ± standard deviation (SD). BMI was calculated by the common formula: BMI (kg/m2) = weight (kg)/height (m)2. Body surface area (BSA) was assessed by a variation of the DuBois and DuBois formula: BSA (m2) = [weight (kg)0.425 × height (cm)0.725] × 0.007184 (23). The Student t test for paired collectives was used to compare epicardial fat thickness measurements in long- and short-axis views. The correlation between measurements of epicardial fat thickness and fat mass, as well as the correlation between epicardial fat and body composition data, and ejection fraction were assessed by the Spearman's rank correlation coefficient. The interobserver reproducibility was assessed by Bland-Altman analysis, providing the mean difference (d) and its limits of agreement (d ±1.96 SD, where SD = SD of the differences) (24). In addition, the coefficient of variability (CV; equal to the SD of the difference between two measurements over the mean of the two measurements, expressed as a percentage) was calculated. A p-value <0.05 was considered significant for all comparisons.
Results
Left and right ventricular parameters in patients with congestive heart failure and in controls are outlined in Table 1. Heart failure patients revealed significantly lower left and right ventricular ejection fractions, as well as significantly higher left and right ventricular volumes. Table 2 lists body composition data and epicardial fat measurements in patients and controls. There were no significant differences regarding age, BMI, and BSA between the two groups. Heart failure patients revealed significantly lower epicardial fat mass and EFT-SAX. In healthy controls, there were no significant differences between EFT-4CV and EFT-SAX (3.8 mm vs. 4.3 mm; p = 0.2246). However, in the group of heart failure patients, EFT-4CV was significantly higher than EFT-SAX (3.5 mm vs. 2.9 mm; p = 0.003). Table 3 displays the correlation of epicardial fat measurements with body composition data and ejection fraction.
| Parameter | Healthy controls (n = 28) | Patients (n = 43) | p |
|---|---|---|---|
| LV-EF (%) | 58.8 ± 4.2 (51.0 to 67.0) | 25.9 ± 6.8 (10.0 to 35.0) | <0.001 |
| LV-M (g) | 128 ± 33.1 (76.0 to 197.0) | 202.3 ± 51.3 (77.0 to 370.0) | <0.001 |
| LV-EDV (mL) | 142.6 ± 36.3 (50.0 to 221.0) | 290.9 ± 80.8 (175.0 to 515.0) | <0.001 |
| LV-ESV (mL) | 60.0 ± 15.0 (30.0 to 103.0) | 221.0 ± 74.7 (108.0 to 460.0) | <0.001 |
| LV-SV (mL) | 84.4 ± 19.9 (50.0 to 132.0) | 71.0 ± 21.1 (32.0 to 118.0) | 0.009 |
| RV-EF (%) | 58.6 ± 5.5 (43.0 to 68.0) | 43.5 ± 10.9 (19.0 to 69.0) | <0.001 |
| RV-EDV (mL) | 137.9 ± 31.2 (74.0 to 200.0) | 163.3 ± 63.9 (53.0 to 347.0) | 0.030 |
| RV-ESV (mL) | 54.5 ± 13.8 (32.0 to 81.0) | 96.7 ± 52.4 (32.0 to 255.0) | <0.001 |
| RV-SV (mL) | 80.7 ± 21.3 (43.0 to 131.0) | 68.9 ± 21.5 (19.0 to 113.0) | 0.026 |
- CMR, cardiovascular magnetic resonance; LV-EF, left ventricular ejection fraction; LV-M, left ventricular mass; LV-EDV, left ventricular end-diastolic volume; LV-ESV, left ventricular end-systolic volume; LV-SV, left ventricular stroke volume; RV-EF, right ventricular ejection fraction; RV-EDV, right ventricular end-diastolic volume; RV-ESV, right ventricular end-systolic volume; RV-SV, right ventricular stroke volume; NS, not significant. Values are mean ± standard deviation (range).
| Parameter | Controls | Patients | p |
|---|---|---|---|
| No. (male/female) | 28 (22/6) | 43 (35/8) | |
| Age (yrs) | 56.6 ± 10.9 (42 to 80) | 61.9 ± 12.4 (36 to 84) | NS |
| BSA (m²) | 1.96 ± 0.20 (1.64 to 2.46) | 1.91 ± 0.23 (1.48 to 2.40) | NS |
| BMI (kg/m²) | 27.5 ± 4.2 (20.0 to 36.6) | 27.0 ± 4.4 (18.8 to 38.7) | NS |
| Epicardial thickness (mm) | |||
| Measured in 4CV (EFT-4CV) | 3.8 ± 1.5 (1.2 to 6.4) | 3.5 ± 1.5 (0.0 to 10.8) | NS |
| Measured in SAX (EFT-SAX) | 4.3 ± 1.3 (1.9 to 7.9) | 2.9 ± 1.3 (0.5 to 6.3) | <0.0001 |
| Middle value 4CV/SAX (EFT-4CV/SAX) | 4.1 ± 1.1 (2.1 to 7.1) | 3.2 ± 1.2 (0.3 to 7.7) | 0.002 |
| Epicardial fat mass (g) | 64.6 ± 21.2 (28.2 to 112.8) | 51.0 ± 20.9 (19.7 to 96.8) | 0.01 |
- CMR, cardiovascular magnetic resonance; BSA, body surface area; EFT-4CV, epicardial fat thickness measured in the 4-chamber view; EFT-SAX, epicardial fat thickness measured in the short-axis view; EFT-4CV/SAX, mean value of epicardial fat thickness measured in the 4-chamber and short-axis views; NS, not significant. Values are mean ± standard deviation (range).
| EFT-4CV | EFT-SAX | EFT-4CV/SAX | Epicardial mass | |
|---|---|---|---|---|
| BSA | r = 0.21 | r = 0.28 | r = 0.32 | r = 0.45 |
| NS | p = 0.0166 | p = 0.0059 | p = <0.0001 | |
| BMI | r = 0.11 | r = 0.27 | r = 0.29 | r = 0.35 |
| NS | p = 0.0210 | p = 0.0154 | p = 0.0025 | |
| EF | r = 0.04 | r = 0.38 | r = 0.36 | r = 0.31 |
| NS | p = 0.0010 | p = 0.0019 | p = 0.0086 |
- BSA, body surface area; EFT-4CV, epicardial fat thickness measured in the 4-chamber view; EFT-SAX, epicardial fat thickness measured in the short-axis view; EFT-4CV/SAX, mean value of epicardial fat thickness measured in the 4-chamber and short-axis views; EF, epicardial fat; NS, not significant (p > 0.05).
In the subgroup of healthy controls, epicardial fat mass was higher in men than in women (Table 4). In the subgroup of heart failure patients, there was a trend toward higher epicardial fat mass in men, but this trend did not reach statistical significance. Epicardial fat thickness tended toward higher values in men in both controls and patients, but these differences were not statistically significant.
| Men | Women | p | |
|---|---|---|---|
| Healthy controls | |||
| Age | 57.6 ± 10.5 | 57.7 ± 13.3 | NS |
| BMI | 26.9 ± 4.5 | 29.4 ± 2.3 | NS |
| EFT-4CV (mm) | 3.9 ± 1.4 | 3.8 ± 1.8 | NS |
| EFT-SAX(mm) | 4.5 ± 1.3 | 3.6 ± 0.9 | NS |
| EFT-4CV/SAX (mm) | 4.2 ± 1.1 | 3.7 ± 1.0 | NS |
| Epicardial mass (g) | 69.2 ± 21.6 | 47.5 ± 5.2 | 0.023 |
| Patients | |||
| Age | 61.1 ± 12.7 | 65.3 ± 10.9 | NS |
| BMI | 27.5 ± 4.3 | 24.3 ± 4.1 | NS |
| EFT-4CV (mm) | 3.6 ± 1.7 | 3.2 ± 0.5 | NS |
| EFT-SAX(mm) | 2.9 ± 1.4 | 2.7 ± 0.9 | NS |
| EFT-4CV/SAX (mm) | 3.2 ± 1.3 | 3.0 ± 0.5 | NS |
| Epicardial mass (g) | 52.5 ± 19.3 | 45.4 ± 26.4 | NS |
- EFT-4CV, epicardial fat thickness measured in the 4-chamber view; EFT-SAX, epicardial fat thickness measured in the short-axis view; EFT-4CV/SAX, mean value of epicardial fat thickness measured in the 4-chamber and short-axis views; NS, not significant (p > 0.05). Values are mean ± standard deviation.
Table 5 displays Spearman's rank correlation analysis of volumetrically measured epicardial fat mass with EFT-4CV, EFT-SAX, and EFT-4CV/SAX. Epicardial fat mass correlated significantly with EFT-SAX in both groups and with EFT-4CV in controls. However, in the heart failure group, no correlation was found between epicardial fat mass and EFT-4CV.
| EFT-4CV | EFT-SAX | EFT-4CV/SAX | |
|---|---|---|---|
| Epicardial fat mass in controls | r = 0.387 | r = 0.466 | r = 0.610 |
| p = 0.042 | p = 0.012 | p = 0.001 | |
| Epicardial fat mass in patients | r = 0.256 | r = 0.590 | r = 0.582 |
| p = 0.098 | p < 0.0001 | p < 0.0001 |
- EFT-4CV, epicardial fat thickness measured in the 4-chamber view; EFT-SAX, epicardial fat thickness measured in the short-axis view; EFT-4CV/SAX, mean value of epicardial fat thickness measured in the 4-chamber and short-axis views; NS, not significant. p Values <0.05 are considered significant.
Bland-Altman plots of interobserver measurements of epicardial fat thickness and mass are displayed in Figure 3. The CV was 5.9% for the volumetric method, 13.6% for EFT-4CV, 18.1% for EFT-SAX, and 10.3% for EFT-4CV/SAX, respectively.
Scattergram showing Bland-Altman plots of individual differences. (A) Results for EFT-4CV. (B) Results for EFT-SAX. (C) Results for EFT-4CV/SAX. (D) Results for epicardial fat mass. Mean bias (solid line) and limits of agreement (striped lines) are indicated in each panel.
Discussion
The present study describes a new volumetric method of quantitative epicardial fat assessment by CMR. Epicardial fat mass determined by this new method correlated with conventional indices of epicardial fat thickness measured at the right ventricular free wall. Epicardial fat is a true visceral fat tissue deposited around the heart with characteristics of a high-insulin resistant tissue and is a source of several inflammatory mediators. Although there have been relatively few studies on epicardial fat, the evidence reported so far suggests that epicardial adipose tissue is anatomically and clinically related to cardiac function and morphology. For example, Iacobellis et al. recently investigated the relationship between epicardial adipose tissue and measurements of visceral adipose tissue, left ventricular mass, and clinical parameters of the metabolic syndrome and suggested echocardiography as a tool for direct assessment of epicardial fat (10). However, the subjects included in these studies were mainly healthy individuals with a wide range of adiposity. The thickness of epicardial fat was measured at the right ventricular free wall both from parasternal long-axis and short-axis views. These echocardiographic measurements could be validated by CMR, which is considered the gold standard for visceral adipose tissue estimation. However, given the considerable inter-individual differences in the distribution of epicardial fat, it is unclear whether right ventricular epicardial fat thickness measured at single points correlates well with the absolute amount of epicardial fat, particularly in patients with congestive heart failure. Therefore, in the present study, we used a volumetric method to estimate the total amount of epicardial fat to obtain a parameter that is independent of individual differences of fat distribution.
Epicardial Fat Thickness Measurements
The results of epicardial fat thickness measurements at the free wall of the right ventricle accord well with results reported by Schejbal (13), who measured epicardial fat thickness in an unselected autopsy collective. This author found a thickness of 4.12 ± 1.4 mm at the ventrolateral edge of the right ventricle. Different from our results, Iacobellis et al. (8) measured a mean epicardial fat thickness of 7.6 ± 3.5 mm in male and 6.9 ± 3.7 mm in female subjects by echocardiography in a group of 72 individuals. These differences can be explained by differences in anthropometric and metabolic profiles of the study subjects: 1) BMI was 33.1 ± 13.8 in males and 33.7 ± 14.9 in females in the study by Iacobellis et al. as compared with 26.9 ± 4.5 in males and 29.4 ± 9.1 in females in our study; and 2) many of the subjects in the Iacobellis et al. study had metabolic syndrome. Interestingly, in the subgroup of subjects with predominant peripheral fat accumulation and no signs of metabolic syndrome, epicardial fat thickness was 4.12 ± 1.67 mm in men and 3.13 ± 1.87 mm in women, which is almost identical to our results in healthy controls.
In the present study, we measured epicardial fat thickness in the 4-chamber view and in short-axis views. Theoretically, one could assume that EFT-SAX and EFT-4CV should lead to similar results. Indeed, this was true in the subgroup of controls. However, in the group of patients with congestive heart failure, EFT-4CV was significantly higher than EFT-SAX. One potential explanation for this finding might be redistribution of epicardial fat due to remodeling of ventricular geometry in patients with congestive heart failure. This assumption is strengthened by the observation that healthy men had significantly more epicardial fat mass than women, whereas this gender-related difference could not be found in heart failure patients.
Volumetric Epicardial Mass Measurements
To our knowledge, this is the first in vivo measurement of epicardial fat volume and mass using CMR. Our measurements of epicardial fat mass showed excellent agreements with findings of Corradi et al. in a pathological study, who found a total epicardial fat weight of 54 ± 23 g in normal hearts and 51 ± 16 g in ischemic hearts (11).
The present study has demonstrated that quantitative assessment of the amount of epicardial fat is feasible and may provide a precise method, which is less susceptible to the individual cardiac anatomy and fat distribution than measurements of epicardial fat thickness at single points. However, simple epicardial fat thickness measurements are less time-consuming and easier to perform.
Correlation of Epicardial Thickness to Epicardial Fat Mass
In our study, we could demonstrate that EFT-SAX correlated better to volumetric measurements of epicardial fat mass than EFT-4CV. The same holds true for the correlation between EFT-4CV/SAX and epicardial fat mass. In heart failure patients, the correlation between EFT-4CV and volumetrically determined epicardial fat mass did not reach statistical significance. This finding is not unexpected. EFT-SAX is averaged from multiple consecutive planes. Thus, it is less susceptible to the individual differences in fat distribution and anatomical conditions than EFT-4CV. The discrepancy between patients and controls affirm our assumption that fat distribution in healthy individuals differs from fat distribution in patients with distorted ventricles. It seems that, during development of heart failure, there is not only a global loss of epicardial fat tissue but also a redistribution of the epicardial fat deposits.
In terms of variability, interobserver reproducibility of volumetric fat measurements was superior to the reproducibility of measurements of epicardial fat thickness. There are mainly two reasons for these differences: 1) Because of its small size, epicardial fat tissue is often difficult to discriminate from the right ventricle wall. Thus, it is not always possible to precisely measure its size. 2) Choosing the point at the right ventricular wall that reveals the maximum epicardial fat thickness can be very challenging at times, and choices may vary considerably, even with experienced observers.
Another main finding of our study was the observation that patients with congestive heart failure had significantly less epicardial fat mass than healthy subjects with comparable age, BMI, and BSA. Similar results were reported in a postmortem study by Schejbal (13) who found significantly less epicardial fat in human hearts with signs of congestive heart failure than in hearts without signs of heart failure. Metabolic abnormalities in patients with heart failure or anatomic alterations associated with disturbed cardiac function and geometry are possible explanations for this finding. However, further studies are needed, before firm conclusions can be drawn.
In summary, quantitative assessment of epicardial fat using the CMR-based volumetric approach is feasible and reproducible. We propose this new method as the in vivo reference standard, as it is less susceptible to the individual cardiac anatomy and fat distribution than measurements of epicardial fat thickness at single points. Additionally, the superior interobserver reproducibility of volumetrically measured epicardial fat speaks in favor of this method.
Acknowledgments
There was no funding/outside support for this study.
