Non-Alcoholic Fatty Pancreas Disease is Associated with Increased Epicardial Adipose Tissue and Aortic Intima-Media Thickness

Abstract

Background

We aimed to investigate whether non-alcoholic fatty pancreas disease (NAFPD) is associated with epicardial adipose tissue (EAT), which is a kind of ectopic fat accumulation, and aortic intima-media thickness (aIMT), which is associated with subclinical atherosclerosis.

Methods

Fifty-four patients with NAFPD (22 men; mean age: 52 ± 10 years) and 49 without NAFPD (16 men; mean age: 49 ± 8 years) were included in this study. NAFPD and aIMT were evaluated using transabdominal ultrasonography (TUS). EAT was evaluated with transthoracic echocardiography.

Results

EAT (6.09 ± 1.52 mm and 3.87 ± 1.31 mm, p < 0.001) and aIMT [1.12 (0.60-1.90) mm vs. 0.93 (0.50-1.44) mm, p < 0.001] were significantly higher in the NAFPD positive subjects, compared to the NAFPD negative subjects. Multivariate analysis showed that age (p = 0.016), body mass index (p = 0.004), and presence of NAFPD (p = 0.024) were associated with increased aIMT. In addition, multivariate analysis demonstrated that the presence of NAFPD (p < 0.001) was associated with increased EAT.

Conclusions

The presence of NAFPD on TUS is associated with increased aIMT and EAT. Our study results may suggest that NAFPD may reflect subclinical atherosclerosis and may be a simple warning sign for physicians.

INTRODUCTION

Pancreatic fat deposition, which is a kind of ectopic fat accumulation, is referred to as non-alcoholic fatty pancreatic disease (NAFPD) or steatosis.¹ Ectopic fat deposits are closely association with cardiovascular events including non-alcoholic fatty liver disease (NAFLD) and epicardial adipose tissue (EAT).^2,3 Unlike NAFLD, the clinical implications of pancreatic fat deposition have yet to be determined.

Intima-media thickness (IMT) reflects subclinical atherosclerosis and is associated with well-known risk factors for cardiovascular disease.⁴ Aortic IMT has been reported to have a stronger ability than carotid IMT to detect preclinical atherosclerosis in type 1 diabetes.⁵ In addition, EAT, which describes visceral fat tissue accumulation around the heart and epicardial coronary arteries, has been implicated in the development of coronary artery disease as a result of its paracrine and endocrine functions.^6,7

The aim of this study was to investigate the relationship between NAFPD and aIMT, which is a subclinical atherosclerosis marker, and EAT which is associated with the development of coronary artery disease.

METHODS

Study population

We screened NAFPD in consecutive asymptomatic subjects who presented to the Preventive Cardiology Clinic, University of Health Sciences, Trabzon Ahi Evren Cardiovascular and Thoracic Surgery Research and Application Center, Department of Cardiology for cardiovascular risk assessment and primary prevention of cardiovascular disease. Transabdominal ultrasonography (TUS) was used to determine pancreatic fat deposition, and those with fatty pancreatic tissue were considered to have NAFPD. A total of 103 subjects without any cardiovascular disease were grouped on the basis of the presence of NAFPD on TUS. Of the 103 subjects, 54 (22 men; mean age: 52 ± 10 years) were NAFPD positive and 49 (16 men; mean age: 49 ± 8 years) were NAFPD negative. Demographic, biochemical and cardiovascular risk factors were screened in all subjects. Subjects with a history of cardiovascular disease (prior myocardial infarction, coronary artery bypass surgery, percutaneous coronary intervention, and peripheral arterial occlusive disease), moderate to severe valvular disease or prior valvular surgery, congenital heart disease, bacterial endocarditis, presence of active infection and inflammation, oncological disorders, methyl alcohol use, antihyperlipidemic drug use, history of stroke and transient ischemic attack, congestive heart failure, atrial fibrillation, hypertrophic obstructive cardiomyopathy, and liver or thyroid dysfunction were excluded. All subjects gave informed consent, and the study protocol was approved by the Local Ethics Committee.

Assessment of cardiovascular risk factors

A history of arterial hypertension (HT), diabetes mellitus (DM), hyperlipidemia (HL), and smoking was identified in all patients, as was a family history of coronary artery disease (CAD). Type 2 DM was diagnosed on the basis of a history of treated DM and/or a fasting blood glucose level of ≥ 126 mg/dl. HL was diagnosed as a fasting total cholesterol level of ≥ 200 mg/dl, a fasting low-density lipoprotein level of ≥ 160 mg/dl, a fasting triglyceride level of ≥ 200 mg/dl, or the use of anti-hyperlipidemic medications. HT was diagnosed as a positive history of HT, use of antihypertensive medications, or a mean systolic blood pressure of ≥ 140 mmHg and/or a mean diastolic blood pressure of ≥ 90 mmHg in two blood pressure measurements taken from each arm. A family history of CAD was considered positive when a male first-degree relative aged < 55 years or a female first-degree relative aged < 65 years had suffered sudden cardiac death or had CAD.

Assessment of NAFPD

All subjects were examined by a single radiologist to minimize inter-observer variability. All examinations were carried out using the same sonography device (Toshiba Aplio 300Toshiba Co. Ltd., Tokyo, Japan) with a convex probe of a 3.5-5 MHz transducer, and the patients were asked to fast for at least eight hours prior to the procedure. NAFPD was considered positive when pancreatic echogenicity exceeded renal echogenicity (Figure 1A) and negative when otherwise (Figure 1B). The pancreatic echogenicity was also graded into 4 grades. Ultrasonic assessment of normal pancreatic echogenicity and its relationship to fat deposition was as follows: level 0, pancreatic echogenicity was similar to renal parenchyma; level 1, pancreatic echogenicity slightly exceeded renal echogenicity; level 2, pancreatic echogenicity was significantly increased, albeit to a lesser degree than retroperitoneal fat echogenicity; and level 3, pancreatic echogenicity was similar to or higher than that of retroperitoneal fat. NAFPD was defined as being positive when the pancreatic echogenicity reached level 1, 2 or 3.

Figure 1.

(A) Arrow shows non-alcoholic fatty pancreatic disease (NAFPD). (B) Arrow shows normal pancreas.

Measurement of aIMT

AIMT was quantified by focusing the image on the opposite wall (dorsal arterial wall of the most distal 15 mm of the abdominal aorta), with gain settings adjusted for optimal image quality. A resolution box function was utilized to magnify images that were 15 mm in width. A scanning frequency of 13 MHz was preferred, although scanning frequencies of 11.5 and 10 MHz were also used as necessary (to achieve adequate tissue penetration). Several images of the far wall of the distal most 15 mm of the aorta were captured and stored digitally for subsequent off-line analysis. Two best-quality images were selected for off-line analysis for each subject. IMT was quantified with ultrasonic calipers in each case. A minimum of 4 to 6 measurements covering the entire segment of interest on the far arterial wall were taken for each image (Figure 2A). Subjects were dichotomized according to median aIMT. The group with aIMT above the median was defined as the increased aIMT group.

Figure 2.

(A) Measurement of aortic intima-media thickness (aIMT) by transabdominal ultrasonography. (B) Measurement of epicardial adipose tissue (EAT) by transthorasic echocardiography.

Measurement of EAT

Transthoracic echocardiography was carried out with a Vivid S5 cardiovascular ultrasound system (GE Healthcare, Wauwatosa, Wis., USA) with the patient in the left lateral decubitus position. A cardiologist who was blinded to the demographic and clinical data of the subjects made all measurements. Epicardial fat was defined as a non-echogenic space between the outer wall of the myocardium and the pericardial layer. EAT was defined as the space from the outer edge of myocardium to the pericardium, and was measured perpendicularly to the free wall of the right ventricle at end-diastole during 3 cardiac cycles using M-mode images.⁶ EAT was most accurately measured from the parasternal short-axis and long-axis views, with optimal cursor beam orientation in both views. Maximum EAT was quantified at a point on the right ventricular free wall along the midline of the ultrasound beam, perpendicularly to the aortic annulus, and used as an anatomic reference for this window. EAT was quantified on the right ventricular free wall along the midline of the ultrasound beam, 2 cm from the interventricular septum in the parasternal short-axis view. Three measurements were averaged at each echocardiographic window (Figure 2B). Subjects were dichotomized according to median EAT. The group with EAT above the median was defined as the increased EAT group.

Statistical analysis

The minimum number of subjects required in each group was determined to be 43 so that a difference of 0.18 units between the two group averages could be significant (type I error = 0.05, test power 0.80). SPSS (Statistical Package for Social Sciences) for Windows 10 (SPSS Inc. Chicago, IL, USA) software package was used for all statistical analyses. The Kolmogorov-Smirnov test was used to assess the distribution of continuous data. Continuous variables with normal distribution were expressed as mean ± standard deviation (SD) and those without as median (minimum-maximum); frequency and percentage were used to express categorical variables. Continuous variables with normal distribution were compared with the independent samples t test, and continuous variables with non-normal distribution were compared with the Mann-Whitney U test. The chi-square test was used for comparisons of categorical variables, and Spearman’s test was used for correlation analysis between continuous variables. Multiple logistic regression was performed to assess the independent relationships between supramedian aIMT, EAT and clinical, biochemical and ultrasonographic variables. Statistical significance was set at p < 0.05, and the confidence interval was set at 95%.

RESULTS

The clinical and demographic characteristics are shown in Table 1 and 2. NAFPD was detected in 54% of our subjects. There were no significant differences in age, HT, DM, family history of CAD, body mass index (BMI), or HL between the subjects with and without NAFPD.

Table 1. Clinical and demographic charactheristics with and without NAFPD subjects.

Variables	NAFPD absence (n = 49)	NAFPD presence (n = 54)	p value
Clinical and demographic charactheristics
Age, year	49 ± 8	52 ± 10	0.131
Gender, male, n (%)	16 (33)	22 (40)	0.396
Hypertension, n (%)	19 (39)	29 (54)	0.129
DM, n (%)	11 (22)	21 (39)	0.112
Dyslipidemia, n (%)	21 (43)	23 (43)	0.978
Current smoking, n (%)	18 (37)	11 (20)	0.104
Family history of CAD, n (%)	2 (4)	5 (9)	0.285
BMI (kg/m2)	31.4 ± 4.5	33.1 ± 5.3	0.103
HR (bpm)	79 ± 14	78 ± 14	0.654
SBP, mmHg	120.4 ± 12.7	124.5 ± 16.5	0.367
	120 (90-160)	124 (100-180)
DBP, mmHg	78.6 ± 8.9	81.3 ± 9.6	0.155
	78 (60-100)	81 (60-100)
Biochemical and hematological parameters
Serum glucose (mg/dL)	116.6 ± 49.1	128.7 ± 45.7	0.02
	117 (83-330)	128 (86-292)
Serum creatinine (mg/dL)	0.76 ± 0.15	0.82 ± 0.22	0.173
ALT (IU/L)	24.5 ± 11.4	35.8 ± 26.7	0.234
	24.5 (9-63)	35 (10-115)
LDL-c (mg/dL)	152 ± 37	151 ± 38	0.904
Triglyceride (mg/dL)	189 ± 105	178 ± 83	0.623
WBC (×109/L)	7.32 ± 1.66	6.87 ± 1.31	0.247
Hb (g/dL)	13.6 ± 1.6	13.8 ± 1.6	0.570
PLT (×109/L)	249 ± 50	237 ± 57	0.324

ALT, alanin amino transferase; BMI, body mass index; CAD, coronary artery disease; DBP, diastolic blood pressure; DM, diabetes mellitus; Hb, hemoglobin; HR, heart rate; LDL-c, low-density lipoprotein cholesterol; NAFPD, non-alcoholic fatty pancreatic disease; PLT, platelet; SBP, systolic blood pressure; WBC, white blood cells.

Table 2. Echocardiographic and ultrasonography measurement parameters.

Variables	NAFPD absence	NAFPD presence	p value
aIMT (mm)	0.93 ± 0.2	1.12 ± 0.26	< 0.001
	0.93 (0.50-1.44)	1.12 (0.60-1.90)
AoD (mm)	28.1 ± 3.3	28.3 ± 3.1	0.791
LVEDd (mm)	44.8 ± 4.1	44.6 ± 3.9	0.742
LVESd (mm)	27.2 ± 4.1	27.6 ± 4.1	0.636
IVST (mm)	8.9 ± 1.3	9.4 ± 1.6	0.115
PWT (mm)	9.6 ± 1.7	9.7 ± 1.1	0.733
LAD (mm)	33.1 ± 3.6	34.3 ± 3.1	0.074
EAT tickness (mm)	3.87 ± 1.31	6.09 ± 1.52	< 0.001

aIMT, aort intima-media thickness; AoD, aortic root diameter; cIMT, carotid intima-media thickness; EAT, epicardial adipose tissue; IVST, interventricular septal thickness; LAD, left atrail diameter; LVEDd, left ventricle end-diastolic diameter; NAFPD, non-alcoholic fatty pancreatic disease; PWT, posterior wall tickness.

There were also no significant differences in low-density lipoprotein cholesterol (LDL-c), triglycerides, alanine amino transferase (ALT), white blood cells (WBC), hemoglobin, and platelets between the two groups. Fasting blood glucose was higher in the NAFPD subjects than in those without NAFPD [128 (86-292) mg/dL and 117 (83-330) mg/dL, p = 0.02].

Heart rate, and systolic and diastolic blood pressures were similar in both groups. In addition, there were no significant differences in echocardiographic parameters such as aortic root diameter, interventricular septal thickness (IVST), posterior wall thickness (PWT), left ventricle end diastolic diameter (LVEDd), left ventricle end systolic diameter (LVESd) and left atrial diameter between the two groups.

In the subjects with NAFPD, EAT thickness (NAFPD: 6.09 ± 1.52 mm and without NAFPD: 3.87 ± 1.31 mm, p < 0.001) and aIMT [NAFPD: 1.12 (0.60-1.90) mm without NAFPD: 0.93 (0.50-1.44) mm, p < 0.001] were higher than in those without NAFPD (Figure 3A, Figure 3B).

Figure 3.

(A) EAT thickness in with and without non-alcoholic fatty pancreatic disease (NAFPD) subjects. (B) Aortic intima-media thickness (aIMT) in with and without NAFPD subjects.

Median aIMT and EAT were 1 mm and 4.89 mm, respectively. Multivariate analysis showed that age (p = 0.016), BMI (p = 0.004), and presence of NAFPD (p = 0.024) were associated with increased aIMT (Table 3). In addition, multivariate analysis demonstrated that the presence of NAFPD (p < 0.001) was associated with increased EAT (Table 4).

Table 3. Univariate and multivariate analysis for increased aIMT.

Variables	Univariate analysis		Multivariate analysis
	Odds ratios (95% CI)	p value	Adjusted odds ratios (95% CI)	p value
Age	1.097 (1.040-1.157)	0.001	1.080 (1.015-1.151)	0.016
Body mass index	1.248 (1.114-1.398)	0.000	1.196 (1.059-1.351)	0.004
Presence of NAFPD	3.810 (1.659-8.745)	0.002	3.048 (1.158-8.024)	0.024
Hypertension	2.519 (1.113-5.699)	0.027	1.041 (0.360-3.010)	0.940
DM	4.457 (1.636-12.143)	0.003	2.064 (0.634-6.721)	0.229
Gender	0.772 (0.343-1.738)	0.532
LDL-c	0.995 (0.985-1.009)	0.459
Triglyceride	1.001 (0.995-1.006)	0.837
Dyslipidemia	1.379 (0.619-3.077)	0.432
Smoking	2.231 (0.877-5.677)	0.092
Family history of CAD	0.925 (0.197-4.383)	0.925
Serum urea	1.006 (0.961-1.053)	0.798

CAD, coronary artery disease; CI, confidence interval; DM, diabetes mellitus; LDL-c, low-density lipoprotein cholesterol; NAFPD, non-alcoholic fatty pancreas disease.

Table 4. Univariate and multivariate analysis for increased EAT.

Variables	Univariate analysis		Multivariate analysis
	Odds ratios (95% CI)	p value	Adjusted odds ratios (95% CI)	p value
Age	1.048 (1.002-1.097)	0.042	0.961 (0.907-1.018)	0.176
Body mass index	1.086 (0.998-1.181)	0.054
Presence of NAFPD	17.778 (6.530-48.399)	< 0.001	16.396 (5.867-45.819)	< 0.001
Hypertension	1.696 (0.764-3.762)	0.194
DM	2.006 (0.742-4.772)	0.116
Gender	1.594 (0.702-3.620)	0.266
LDL-c	1.010 (0.996-1.024)	0.180
Triglyceride	1.000 (0.995-1.005)	0.893
Dyslipidemia	1.852 (0.828-4.140)	0.133
Smoking	0.318 (0.123-0.822)	0.018	0.476 (0.148-1.529)	0.212
Family history of CAD	2.182 (0.381-12.505)	0.381
Serum urea	1.007 (0.963-1.053)	0.768

CAD, coronary artery disease; CI, confidence interval; DM, diabetes mellitus; EAT, epicardial adipose tissue; LDL-c, low-density lipoprotein cholesterol; NAFPD, non-alcoholic fatty pancreas disease.

Intraobserver variability was calculated as the difference in two measurements of the same patient by one observer divided by the mean value, and was < 5% for all measurements.

DISCUSSION

To the best of our knowledge, this is the first study to show a relationship between NAFPD and aortic IMT and EAT. The chronic inflammatory process underlying atherosclerosis begins early in life.⁸ Globally, atherosclerosis is a major cause of morbidity and mortality, underlying most cardiovascular disorders.⁹ The noninvasive evaluation of subclinical atherosclerotic disease in humans has enabled clinicians to better assess overall cardiovascular risk. For IMT measurements, the arterial wall is composed of three layers: (I) the tunica intima, (II) the tunica media, and (III) the tunica externa (tunica adventitia). Atherosclerosis develops as structural alterations and thickening of intima-media at the intersection of the tunica intima and tunica media layers. Since separate measurements of the intimal and medial thicknesses have been proven to be unfeasible, measuring both layers combined has emerged as a more accurate and reproducible method.¹⁰ As the atherosclerotic process starts to develop in the distal aorta and coronary arteries, the aorta is a reasonable anatomic site to examine for atherosclerotic alterations during life.¹¹

Wu and Wang et al. reported an association between NAFPD and advanced age, greater BMI, glycated hemoglobin, lipid parameters, and systolic blood pressure.¹² We demonstrated an increase in BMI among the subjects with NAFPD, however that increase failed to reach statistical significance. Age and lipid parameters did not differ significantly between the subjects with and without NAFPD in this study. Moreover, there were no significant differences in systolic and diastolic blood pressures between the subjects with and without NAFPD. In contrast, there was a significant difference in fasting blood glucose between the NAFPD positive and negative subjects. A number of prior studies have indicated that fatty infiltration of the pancreas downgrades the number of β cells and impairs their function.^13,14 A former study showed a positive correlation between fasting blood glucose and cIMT.¹⁵ We found a higher fasting blood glucose level among the subjects with NAFPD, which may represent one mechanism of increased aIMT.

EAT occupies the space between the myocardium and visceral pericardium and wraps three quarters of the surface area of the heart.⁷ An excess amount of EAT has been linked with incident coronary artery disease and major adverse cardiac events.¹⁶ A strong correlation, independent of BMI and other traditional risk factors, has also been reported between EAT and coronary plaque burden and coronary artery disease.^17,18 EAT, as any visceral adipose tissue elsewhere in the body, is regarded to be a risk factor for coronary artery disease.¹⁶ The pathophysiological link between atherosclerosis and EAT has been explained by the endocrinological activity of the latter, which is similar to that of visceral adipose tissue. This causes the release of proinflammatory, atherogenic mediators and cytokines such as interleukin-6, tumor necrosis factor-α, and adiponectin.⁷ EAT has been suggested to be a modifiable factor or a target to modify cardiovascular risk. Weight loss may contribute to reduced EAT and may also affect NAFPD.¹⁹ Ozturk et al. showed that EAT thickness might be related to microalbuminuria in patients with essential hypertension.²⁰ Kocaman et al. revealed an association between EAT and cIMT, a marker of subclinical atherosclerosis.³ We found an increased amount of EAT among subjects with NAFPD. EAT is well known for its contribution to the development of the atherosclerotic process.^3,21 Subjects with NAFPD may have increased EAT, which may in turn be associated with an increase in aIMT, a marker of early and subclinical atherosclerosis.

Ultrasonography can be used to evaluate the pancreas.^21,22 Fatty infiltration of the pancreas causes the emergence of a hyperechogenic appearance compared to the neighboring solid organs, most notably the kidneys and liver.¹⁴ We used TUS to evaluate our subjects. Labropoulos et al. quantified abdominal aIMT in adults with the help of TUS, and reported a greater aIMT in those with atherosclerosis.²³ Autopsy studies have shown that the abdominal aorta is the site where the earliest atherosclerotic lesions emerge.²⁴ Maroules et al. reported the predictive value of MR-guided assessments of aortic atherosclerosis for future adverse cardiovascular events.²⁵ Adipose tissue acts like an endocrinologically active organ, producing considerable amounts of adipokines, including leptin and adiponectin, cytokines such as tumor necrosis factor α, interleukin-6 (IL-6) and monocyte chemotactic protein-1 combined with macrophages producing IL-1b and myeloperoxidase, thereby establishing a proinflammatory state.¹³ Chemokines, when increased in quantity, may contribute to the increase in aIMT in patients with NAFPD.

NAFPD can be readily shown by TUS, an easily performed, inexpensive, and noninvasive method. Our results suggest that NAFPD detected by TUS may be related to subclinical atherosclerosis. Ultrasonography is currently more commonly utilized than any other imaging tool capable of detecting atherosclerosis, such as EAT and aIMT. The presence of NAFPD which is incidentally detected by TUS may be a sign of the existence of subclinical atherosclerosis and may prompt clinicians to take necessary measures to limit or slow down that process.

Study limitations

This study has a number of limitations. The most notable limitation is the small sample volume. Second, the generalizability of our results is questionable since we only enrolled patients who presented to our clinic. Despite a meticulous search of overt cardiovascular disease, we did not employ invasive tests such as coronary angiography to decisively rule it out. We therefore could not refute the possibility of the presence of cardiovascular disease during patient enrollment. In addition, we did not measure HbA1c levels or waist circumference. IMT measurements were performed manually in this study. Finally, we did not use a computerized analysis system which may have reduced errors in measurements.

CONCLUSIONS

The findings of the present study suggest that, NAFPD, which is easily detected by routine TUS, may indicate the existence of subclinical atherosclerosis and may be a simple warning sign for physicians.

DECLARATION OF CONFLICT OF INTEREST

All the authors declare no conflict of interest.