Association of Epicardial Adipose Tissue With Left Ventricular Strain and MR Myocardial Perfusion in Patients With Known Coronary Artery Disease

Epicardial adipose tissue (EAT) may have a paracrine effect on coronary microcirculation and myocardium. However, it is unclear whether EAT is linked to cardiac function and perfusion.

metabolically active visceral fat that is embedded in the myocardium and shares the same microcirculation with the heart. 3It affects the myocardium through the production of various bioactive molecules, as well as inflammatory and pro-inflammatory factors via the paracrine and vascocrine pathways. 3It has been reported that an increase in EAT underlies cardiac diastolic dysfunction in obesity and diabetes, implying that EAT may be a biomarker of diastolic dysfunction. 5,6However, existing research has mainly focused on the effect of EAT on diastolic function or coronary atherosclerosis, such as plaque characteristics. 2,3,5The impact of EAT on left ventricular (LV) subclinical systolic function in CAD patients with preserved ejection fraction (EF) remains to be studied.
Reduced myocardial perfusion has been found to predict adverse cardiovascular events in CAD. 7][10] However, these methods have limitations, related to radiation exposure or low spatial resolution. 11Furthermore, the associations between EAT and myocardial perfusion remain to be fully elucidated.Cardiac MRI could allow to evaluate both ischemic and nonischemic cardiomyopathies and detect hemodynamically relevant CAD.And cardiac MRI perfusion imaging has been increasingly implemented for the noninvasive assessment of myocardial microcirculation function, with high accuracy and repeatability. 12,13gainst this background, we aimed to investigate the association of EAT with LV systolic function and myocardial perfusion by utilizing cardiac MRI in patients with known CAD and normal EF.In addition, we also aimed to evaluate the relationship between LV strain and myocardial perfusion.

Materials and Methods
This retrospective study was approved by the institutional review board (IRB) of our hospital and conducted following the Declaration of Helsinki.The requirement for written informed consent was waived due to this study's retrospective design.

Patients and Controls
This retrospective study included patients with known CAD (according to cardiac computed tomography angiography [CTA] or coronary angiography revealing obstructive CAD), who underwent cardiac MRI from January 2018 to October 2022 at our institution.We excluded patients with valvular disease, cardiomyopathy, arrhythmia, symptoms of heart failure, a history of myocardial infarction, LVEF <50%, poor image quality, and serious hepatic (Child-Turcotte-Pugh score ≥ 10) or renal (estimated glomerular filtration rate [eGFR] < 60 mL/min/1.73m 2 ) disease.Finally, a total of 78 CAD patients were included in this study and these patients were further divided into two groups according to median EAT volume.Additionally, we included 20 age and sex-matched healthy subjects as controls.Demographic and clinical information, including age, gender, height, weight, systolic/diastolic blood pressure, were collected at the time of the cardiac MRI examinations, according to the medical records of the patients and controls.The body mass index (BMI) was calculated as weight (kg) divided by the square of height (m).Body surface area (BSA) was computed as 0.0057 Â height (cm) + 0.0121 Â weight (kg) + 0.0882 for men and 0.0073 Â height (cm) + 0.0127 Â weight (kg) À 0.2106 for women.Blood pressure was recorded as an average of three measurements in the right arm in a sitting position after a 10 minutes resting period.Data regarding smoking, diabetes, hypertension, dyslipidemia, the family of CAD, and medication were collected from hospital records.

Cardiac MRI Protocol
All cardiac MRI examinations were performed using a 1.5-T MRI scanner (Siemens Avanto, Erlangen, Germany).Electrocardiography (ECG) gating (Siemens AG, Willelsbacherplalz 2, DE-80333 Muenchen, Germany) and the respiratory gating technique were applied to monitor dynamic changes over the entire scanning process.Data were acquired during the breath-holding period following the end of inspiration.
A balanced steady-state free precession (bSSFP) sequence was used to obtain two-, three-, and four-chamber cine images in the long-axis view, as well as 8-12 continuous cine images from the apex to the base in the short-axis view.The acquisition parameters were as follows: repetition time (TR) = 3.2 msec, echo time (TE) = 1.1 msec, flip angle = 58 , field of view = 300 Â 300 mm 2 , slice thickness = 6 mm, interslice gap = 0 mm, and matrix size = 250 Â 200.A 0.2 mL/kg dose of gadopentetate meglumine (Consun Pharmaceutical Group Ltd, Guangzhou, China) was injected intravenously through the cubital vein at a rate of 2.5-3.0 mL/sec with a fully automatic high-pressure syringe (Antmed Co. Ltd, Shenzhen, China).
Rest first-pass perfusion images were acquired in three standard short-axis slices (basal, middle, and apical) with dynamic inversion recovery prepared echo-planar image sequence.The acquisition parameters were as follows: TR/TE = 187/1.0msec, flip angle = 50 , slice thickness = 8 mm, and matrix size = 270 Â 210.To rule out myocardial infarction, late gadolinium enhancement (LGE) images were acquired with a segmented-turbo fast low-angle shot (FLASH) phase-sensitive inversion recovery (PSIR) sequence about 10-15 minutes after contrast administration.The acquisition parameters were as follows: TR/TE = 700/3.2msec, flip angle = 25 , and slice thickness = 8 mm.

Cardiac MRI Data Analysis
All cardiac MRI data were analyzed using commercial software (cvi42, version 5.13.7;Circle Cardiovascular Imaging Inc., Calgary, AB, Canada) by a radiologist with 5 years of experience (Z.J.), who was blinded to the clinical data.
The endocardial and epicardial borders of the LV myocardium in all of the short-axis cine images were manually traced at the end-diastolic and end-systolic phases.Besides, the papillary muscles and trabeculae were excluded in all series.The LV structural and functional parameters were automatically calculated, including mass, end-systolic volume (ESV), end-diastolic volume (EDV), stroke volume (SV), and EF.Specifically, mass, EDV, and ESV indexed for BSA (mass index [MI], EDV index [EDVI], and ESV index [ESVI]) were calculated using the Mosteller formula. 14lobal LV strain parameter analysis was performed by manually delineating the endocardium and epicardium at the end-diastolic phase and automatically tracking myocardial voxel points on shortaxis, two-, and four-chamber long-axis cine images (Fig. 1).The LV strain index included global radial peak strain (GRS), global circumferential peak strain (GCS), and global longitudinal peak strain (GLS).
Semiquantitative perfusion analysis was performed as previously described. 14First, a phase with good contrast that endocardium, epicardium and blood pool contour could be displayed most clearly was set.Then, the endocardium contour, epicardium contour, blood pool contour at the basal, middle, and apical views were manually delineated in the first-pass perfusion images of the perfusion module.The anterior and inferior segmentation reference points were set and contour propagation was performed.Contours for each phase within the analysis range were checked and adjusted.Finally, signal intensity-time curves were corrected and derived for the blood pool and 16 myocardial segments, according to the American Heart Association (AHA) conventions (Fig. 1).Semiquantitative segmental perfusion parameters were automatically generated and global perfusion parameters were calculated as mean values for the 16 myocardial segments. 14Specifically, LV perfusion parameters included upslope, time-to-maximum signal intensity (TTM), and maximum signal intensity (MaxSI).
The tissue characteristics technique of the cvi42 software was used to quantify EAT volume.The epicardial border and the visceral layer of the pericardium were manually outlined on contiguous enddiastolic short-axis cine images from the base to the apex (Fig. 2).We calculated the total EAT volume as the sum of the areas of all slices multiplied by the slice thickness.

INTRAOBSERVER AND INTEROBSERVER AGREEMENT.
Intraobserver and interobserver agreement for LV strain variables, perfusion parameters and EAT volume were measured in 30 randomly selected patients.One investigator (Z.J., 5 years of experience) analyzed these patients on two separate occasions with a 1-month interval to evaluate intraobserver agreement.For interobserver agreement, a second observer (X.Z., 3 years of experience) blinded to the clinical and cardiac MRI results independently measured the these cardiac MRI parameters.

Statistical Analysis
All analyses were performed with SPSS software (version 22.0; IBM Corp., Armonk, NY, USA).Categorical variables were expressed as numbers (percentages) and were compared using Chi-squared or Fisher exact tests, as applicable.Continuous variables were presented as mean AE standard deviation (SD) for normal distributions or median and interquartile range (IQR) for non-normal distributions.
One-way analysis of variance (ANOVA) or Kruskal-Wallis tests were used to compare baseline characteristics including demographic and clinical information, and cardiac MRI findings among the controls and CAD patients with high and low EAT volume groups as appropriate, followed by Bonferroni's post hoc test.Multivariate linear regression analyses were performed to determine the associations between EAT volume, LV strains, and myocardial perfusion indices in CAD patients after adjustment for age, sex, BMI, diabetes, hypertension, dyslipidemia and LVMI.Multicollinearity was assessed by collinearity diagnostics (i.e.tolerance < 0.1 and/or variance inflation factor > 10).Interobserver and intraobserver agreement was determined by the evaluation of intraclass correlation coefficients (ICCs).A P value <0.05 (two-tailed) was considered statistically significant.

Results
The baseline characteristics for the total sample are presented in Table 1.The median EAT volumes were 104.5 cm 3 (IQR 73.1, 116.6) for the controls and 105.6 cm 3 (IQR 84.9, 118.2) and 159.6 cm 3 (IQR 147.9, 175.2) for the patientswith low and high EAT volume groups, respectively, with a significant difference in the three groups.Individuals in the three groups were comparable in age (P = 0.062), systolic blood pressure (P = 0.147), and diastolic blood pressure (P = 0.074).Also, the proportions for sex (P = 0.896) and smoking (P = 0.984) were similar (all P > 0.05).Patients with high EAT volumes exhibited significantly higher BMI values as compared to controls (28.08 AE 3.02 kg/m 2 .25.54 AE 2.84 kg/m 2 ).Patients with high EAT volumes and patients with low EAT volumes showed significantly higher proportion for the family of CAD (62% vs. 54% vs. 5%), diabetes (67% vs. 62% vs. 0%), hypertension (74% vs. 64% vs. 0%), and dyslipidemia (69% vs. 72% vs. 0%) than controls.Table 2 presents a summary of LV structure, function, and perfusion parameters of the study population.The LV end-diastolic volume index (LVEDVI) (P = 0.681), LV endsystolic volume index (LVESVI) (P = 0.090), stroke volume index (SVI) (P = 0.973), and LVEF (P = 0.137) were not significantly different among the patients with high and low EAT volume groups and control group (all P > 0.05), whereas LV mass index (LVMI) was significantly higher in patients with high EAT volume than in the control group (59.34 AE 6.63 g/m 2 vs. 52.30AE 14.25 g/m 2 ).Compared with the control group, GRS, GCS, and GLS were significantly lower in both groups of patients.Moreover, patients with high EAT volume exhibited significantly lower GRS, GCS, and GLS than patients with low EAT volume.The upslope, perfusion index, and MaxSI of the two groups of patients were significantly lower than in the control group, with more significant differences in the high EAT volume group.Regarding TTM, the value was significantly higher in the high EAT volume group than in the low EAT volume group, which exhibited similar levels as the control group.Multivariate linear regression analysis demonstrated that the EAT volume was independently and significantly associated with upslope (β = À0.47),perfusion index (β = À0.27),TTM (β = 0.28), MaxSI (β = À0.36),GRS (β = À0.54),GCS (β = 0.38), and GLS (β = 0.25).(Tables 3 and 4).Additionally, when EAT and perfusion parameters were included in the multivariate linear regression analysis, EAT (β = À0.32) and upslope (β = 0.16) were significantly associated with GRS.Furthermore, EAT (β = 0.30) and perfusion index (β = À0.12) were independently and significantly associated with GCS.EAT (β = 0.13) and perfusion index (β = À0.29) were significantly associated with GLS.

Discussion
In this study, we used cardiac MRI to investigate the relationship of EAT with LV strain and perfusion in patients with known CAD.We found that EAT volume was significantly higher and LV strain and perfusion parameters were significantly lower in patients with known CAD than in healthy controls.Moreover, EAT was significantly and negatively associated with LV strain and perfusion in patients with known CAD.Furthermore, myocardial perfusion was independently associated with LV strain in patients with known CAD.
The EAT is an endocrine organ and the source of bioactive adipocytokines, which potentially affect the myocardium and coronary arteries. 15EAT is believed to be associated with CAD development, independent of the existence of cardiovascular risk factors or the extent of other fat depots. 16dditionally, previous CT-based research suggested that EAT was a useful marker for CAD in asymptomatic patients with noncalcified plaques and zero calcium scores. 3Besides, studies have proposed that it is independently associated with coronary calcification, indicating that EAT may be involved in all stages of CAD. 3,15The LV is generally composed of three myocardial layers: the inner oblique, middle circular, and outer oblique layers.During systole, the inner oblique layer undergoes a maximal dimensional shift. 17During ischemic myocardial disease, subendocardial fibers are more susceptible to damage and exhibit impaired longitudinal contractile function. 18According to the results of a large prospective study, GLS was strongly associated with incident CAD in older adults after 10 years of follow-up examinations. 19This may indicate that early detection of asymptomatic myocardial dysfunction may provide an opportunity for prevention and early intervention. 19In contrast to evidence regarding the impact of EAT on LV diastolic function, the association between EAT and LV systolic function has not been studied in detail.A previous study that utilized the two-dimensional speckle tracking echocardiography strain technique independently of BMI reported that an increase in EAT was only linked to reduced GLS. 20Further research suggested that the accumulation of EAT is more strongly related to subclinical LV dysfunction than whole-body adiposity. 21Consistent with previous research, our study demonstrated that besides longitudinal strain, circumferential and radial strains were also lower.Furthermore, decreases in GRS, GCS, and GLS were further aggravated by EAT, and all three strains were independently associated with EAT in patients with known CAD and preserved LVEF.Several potential mechanisms that could influence the impact of EAT on LV systolic function have been proposed in recent years.One hypothesis is that EAT may exert a direct local effect on the myocardium via the paracrine action of EAT-derived cytokines. 18,22,23Furthermore, excess amounts of EAT may induce an increase in intramyocardial triacylglycerol and interstitial myocardial fibrosis, thereby leading to further dysfunction. 24Additionally, the mechanical restriction of myocardial expansion from EAT during diastole may reduce cardiac output. 258][29] Besides, in CAD cases, endothelial dysfunction, vascular remodeling, and microvascular rarefaction could impact both macro-and microvasculature. 27Resting myocardial perfusion reflecting autoregulated blood flow is associated with myocardial oxygen consumption and could be used to noninvasively evaluate myocardial microcirculation function. 14,30In a retrospective study of 221 patients with suspected CAD who underwent coronary CTA, EAT showed an independent association with myocardial ischemia evaluated by CT-FFR, suggesting that EAT was likely to be involved in myocardial ischemia. 2In this study, we observed a decrease in myocardial perfusion parameters and showed a negative association between EAT and myocardial perfusion in patients with known CAD.Due to the proximity of EAT, the myocardium, and the coronary arteries, EAT may have local effects on myocardial microcirculation and could induce the development of myocardial ischemia.Previous evidence indicates that activation of the renin-angiotensin-aldosterone system is associated with adipose tissue-derived angiotensinogen, free fatty acids, and leptin. 27,31Additionally, perivascular adipose tissue-derived adipokines are pro-inflammatory molecules that contribute to oxidative stress in the endothelium and could damage the endothelial function and nitric oxide bioavailability. 27,31This promotes coronary arterial vasoconstriction and vascular smooth muscle cell proliferation, resulting in microvascular dysfunction and perfusion impairment. 27,31evious studies suggested the possibility that myocardial microcirculation was correlated with cardiac systolic function, further supporting the importance of efficient energy production in normal myocardial contraction. 14,32Based on existing evidence, the effect of microcirculation on LV strain is interpreted as a disruption of interstitial and cellular myocardial microcirculation. 33,344][35] Previous studies contended that EAT induced LV dysfunction by damaging small vessel blood flow. 36Our study revealed an association between subclinical LV systolic dysfunction and impaired myocardial perfusion, and that EAT was associated with myocardial perfusion in patients with known CAD.These findings may provide useful information for increasing awareness of the pathophysiological mechanisms of systolic dysfunction.However, whether EAT can serve as an effective therapeutic target for improving myocardial microcirculation and function requires further investigation.

Limitations
This was a retrospective single-center study with a small sample size and we could not assess causality.Besides, all of the patients involved in this study were Chinese, thus investigations into the relevance of this study to other ethnic groups are required.Furthermore, stress examinations were not performed on our subjects, thus we were not able to evaluate the cardiac systolic function and perfusion reserve.Finally, there was no prognostic information regarding EAT, LV strain, and perfusion.Thus, there should be further research into whether pharmacological intervention on EAT could relieve LV strain and enhance myocardial perfusion.

Conclusion
The results from this study may indicate that EAT volume was independently and negatively associated with LV strain and myocardial perfusion in patients with known CAD.Besides, myocardial perfusion was found to be associated with LV strain.

FIGURE 1 :
FIGURE 1: Representative cardiac MRI pseudocolor images at end-diastole, cardiac MRI-derived peak strain curves, first-pass myocardial perfusion images and signal intensity-time curves in a patient with CAD.(a1) Left ventricle pseudocolor images in short axis; (a2) LV global peak strain curve in circumferential direction; (b1) left ventricle pseudocolor images in horizontal four-chamber long axis; B2 LV global peak strain curves in longitudinal direction; (c1) first-pass myocardial perfusion MR images from left basalventricular slice; (c2) signal intensity-time curves, orange curve represents blood-pooled time-signal intensity curve and other colour curves represent time-signal intensity curves in each myocardial segment.

FIGURE 2 :
FIGURE 2: Representative EAT quantification images of the short axis from basal to apical slice at end-diastole.The epicardial border is shown in red, the visceral layer of the pericardium in green, and the adipose tissue in yellow.

TABLE 1 .
Baseline Characteristics of the Study Cohort Values are given as mean AE standard deviation, numbers in brackets are percentages.CAD = coronary artery disease; EAT = epicardial adipose tissue; BMI = body mass index; ACEI = angiotensin converting enzyme inhibitor; ARB = angiotensin receptor blocker.*P < 0.05 vs. controls.† P < 0.05 vs. CAD with low EAT volume.

TABLE 2 .
Comparison of Cardiac MRI Findings Among CAD Patients With Low/High EAT Volume and Controls CAD = coronary artery disease; EAT = epicardial adipose tissue; LV = left ventricular; EDVI = end-diastolic volume index; ESVI = end-systolic volume index; SVI = stroke volume index; MI = mass index; EF = ejection fraction; GRS = global radial peak strain; GCS = global circumferential peak strain; GLS = global longitudinal peak strain; TTM = time-to-maximum signal intensity; Max SI = max signal intensity.*P < 0.05 vs. controls.† P < 0.05 vs. CAD with low EAT volume.Volume 58, No. 5 Journal of Magnetic Resonance Imaging

TABLE 4 .
Multivariate Linear Regression Analysis of EAT Volume and Myocardial Perfusion With Myocardial Perfusion in Patients With CAD After Adjustment for Age, Sex, BMI, Diabetes, Hypertension, Dyslipidemia, and LVMI Model 1: association of EAT volume with LV strain.Model 2: association of LV strain with EAT volume and myocardial perfusion.CAD = coronary artery disease; EAT = epicardial adipose tissue; BMI = body mass index; LVMI = left ventricular mass index; GRS = global radial peak strain; GCS = global circumferential peak strain; GLS = global longitudinal peak strain; TTM = time-tomaximum signal intensity; Max SI = max signal intensity.

TABLE 3 .
Multivariate Linear Regression Analysis of EAT Volume With Myocardial Perfusion in Patients With CAD After Adjustment for Age, Sex, BMI, Diabetes, Hypertension, Dyslipidemia and LVMI