Abstract

Background

We evaluated the association of cardiac adipose tissue including epicardial adipose tissue and pericardial adipose tissue with incident cardiovascular disease and mortality, coronary artery calcium, carotid intima media thickness and inflammatory markers.

Design

A prospective study of 200 patients with type 2 diabetes and elevated urinary albumin excretion rate (UAER).

Methods

Cardiac adipose tissue was measured from baseline echocardiography. The composite endpoint comprised incident cardiovascular disease and all-cause mortality. Coronary artery calcium, carotid intima media thickness and inflammatory markers were measured at baseline. Cardiac adipose tissue was investigated as continuous and binary variable. Analyses were performed unadjusted (model 1), and adjusted for age, sex (model 2), body mass index, low-density lipoprotein cholesterol, smoking, glycated haemoglobin, and systolic blood pressure (model 3).

Results

Patients were followed-up after 6.1 years for non-fatal cardiovascular disease (n = 29) or mortality (n = 23). Cardiac adipose tissue (p = 0.049) and epicardial adipose tissue (p = 0.029) were associated with cardiovascular disease and mortality in model 1. When split by the median, patients with high cardiac adipose tissue had a higher risk of cardiovascular disease and mortality than patients with low cardiac adipose tissue in unadjusted (hazard ratio 1.9, confidence interval: 1.1; 3.4, p = 0.027) and adjusted (hazard ratio 2.0, confidence interval: 1.1; 3.7, p = 0.017) models. Cardiac adipose tissue (p =  0.033) was associated with baseline coronary artery calcium (model 1) and interleukin-8 (models 1–3, all p < 0.039).

Conclusions

In type 2 diabetes patients without coronary artery disease, high cardiac adipose tissue levels were associated with increased risk of incident cardiovascular disease or all-cause mortality even after accounting for traditional cardiovascular disease risk factors. High cardiac adipose tissue amounts were associated with subclinical atherosclerosis (coronary artery calcium) and with the pro-atherogenic inflammatory marker interleukin-8.

Introduction

Patients with type 2 diabetes mellitus (T2D) are at high risk of developing cardiovascular disease (CVD) and mortality.1 Cardiac adipose tissue (CAT) is the visceral fat surrounding the heart, and consists of epicardial and pericardial adipose tissue (EAT and PAT). Patients with T2D have more and inflamed CAT compared to non-diabetic individuals.2 CAT has been shown to secrete inflammatory substances, which are hypothesized to act locally on the adjacent vasculature and promote atherosclerosis.3 Furthermore, CAT from CVD patients has elevated expression of inflammatory cytokines such as interleukin-6 (IL-6), tumour necrosis factor α (TNFα), and interleukin-1β (IL-1β).4 These cytokines have all independently been implicated in the development of CVD.5

To date, mainly cross-sectional studies support this inflammatory role of CAT in the pathogenesis of CVD: Both EAT and PAT have been associated with coronary arterial calcium (CAC), and carotid intima media thickness (CIMT).6,7 However, few observational population-based studies have investigated EAT and PAT and their association with cardiovascular incidents.8,9

In this study, we hypothesize that incident CVD and mortality in patients with T2D is promoted by excessive CAT. The primary aim is to examine the association of CAT with incident CVD and all-cause mortality. The secondary aim is to examine the association of CAT with subclinical atherosclerosis and low-grade inflammation.

Methods

Patients

Two hundred T2D patients with elevated urinary albumin excretion rate (UAER) but without known coronary artery disease (CAD) were identified at the Steno Diabetes Center from January 2007–February 2008 for a study of risk factors for incident and prevalent silent CVD.10 The design and eligibility of patients have been described previously.10 In brief, written information was sent out to 613 T2D patients treated in a secondary setting at Steno Diabetes Center. A total of 72 patients refused to participate, and 341 were excluded due to prevalent symptomatic heart disease evaluated from patient records, interviews and questionnaires or prevalent symptoms from the heart including Q waves in 12-lead ECG, or a normal UAER <30 mg/24-h. The final study population comprised 200 patients, with elevated UAER (two of three UAER measurements >30 mg/24-h) who underwent baseline echocardiography (n = 194), CAC (n = 195) and CIMT (n = 182) scans and measurements of markers of low-grade inflammation (n = 200).

This study complies with the Declaration of Helsinki, the research protocol was approved by the local ethics committee and all patients gave written informed consent.

Echocardiography of EAT and PAT and other measurements

The thickness of EAT was measured as the echo-free space above the free wall of the right ventricle to the epicardium at end-systole in the parasternal long axis view perpendicular to the aortic annulus, and PAT as the hypoechoic space in front of EAT and on the external pericardium.11,12 The analysis was performed by a trained medical doctor blinded for the study outcomes. All measurements were read twice and the intra-reader coefficient was 0.93 (R = 0.86). A random sample of 10 recordings was re-read by a blinded experienced cardiologist and the inter-reader coefficient was 0.96 (R = 0.90). EAT and PAT were combined to obtain CAT which was of primary interest since it was hypothesized that both EAT and PAT indirectly could promote CVD and all-cause mortality.

Cardiac CT for measurements of CAC score was performed as previously described and intimal and medial calcification in the left main, left anterior descending artery, circumflex artery and right coronary artery were summed to obtain a total CAC score.10

CIMT was measured by an experienced sonographer at the posterior wall 20 mm proximal to the bifurcation of the common carotid artery bilaterally as previously described.13

The plasma biomarkers TNFα, high-sensitivity C-reactive protein (hsCRP), IL-1β, IL-6 and IL-8 were measured as described previously.14

Follow-up

All patients still alive (n = 174) were invited for a five-year reassessment by email and telephone. CAC reassessment was performed in 146 patients and 28 declined the invitation. CIMT was reassessed in 133 patients and 41 declined. Follow-up in relation to incident CVD and death was performed after 6.1 years (5th–95th percentile 2.9; 7.1) from 2012–2013.14 In brief, all patients were traced through the Danish National Death Register and the Danish National Health Register, and no patients were lost to follow-up. Cause of death was registered for deceased patients and information about non-fatal CVD was obtained. All deaths were classified as CVD unless other causes of death were registered. For patients who experienced multiple endpoints the analysis only included the first.

Statistical analysis

Statistical analyses were performed using SPSS (version 20.0, Chicago, Illinois, USA) and SAS software (version 9.3, SAS Institute, Cary, North Carolina, USA). Continuous variables are reported as means ± standard deviation (SD) and categorical data as total numbers with corresponding percentages. Non-normally distributed parameters (CAC and biomarkers of low-grade inflammation) are reported as medians with interquartile range (IQR). EAT, PAT and CAT were measured both as continuous variables and split by medians.

Linear regression models were performed to assess the association of cardiac fat with subclinical atherosclerosis and biomarkers of low-grade inflammation.

Estimates of log-transformed variables were back-transformed and results given in percentage change in dependent variable per unit change in independent variable.

Cox regression analysis was performed to obtain hazard ratios (HRs) with 95% confidence intervals (CIs) for the risk of the composite endpoint per mm increment of cardiac fat or per pg/ml increment of IL-8. Non-linearity was tested by adding cardiac fat squared as a covariate to the models.

Adjustment of unadjusted models (model 1) included age and sex (model 2) plus traditional risk factors: age, sex, low-density lipoprotein (LDL)-cholesterol, smoking, body mass index (BMI), glycated haemoglobin (HbA1c) and systolic blood pressure (model 3). A two-tailed p value of <0.05 was considered significant.

Results

Patient characteristics

Of the total population 152 (76%) were men, with mean (±SD) age of 59 ± 9 years, known diabetes duration of 13 ± 7 years, and 158 (79%) had micro- and 42 (21%) macroalbuminuria. Categorized according to the pre-specified composite endpoint, patients who died or had a CVD event (n = 52) were older (p = 0.001), predominantly men (p = 0.031) and had a longer diabetes duration (p = 0.004) compared to patients without the endpoint (n = 148) at follow-up (Table 1).

Table 1.

Patient characteristics at baseline according to event status.

Composite endpoint (CVD or mortality)
Characteristic Total No Yes p Value
Participants, n (%) 200 148 (74) 52 (26)
Male, n (%) 151 (76) 106 (72) 45 (87) 0.031
Age, years 59 (9) 58 (9) 62 (7) 0.001
Current smoker, (%) 59 (30) 39 (26) 20 (39) 0.072
Diabetes duration, (years) 13 (7) 12 (7) 15 (7) 0.004
Systolic blood pressure (mm Hg) 130 (16) 129 (15) 130 (16) 0.524
Body mass index (kg/m2) 32.6 (5.8) 32.7 (5.7) 32.0 (5.9) 0.461
Cholesterol (mmol/l) 3.9 (0.9) 3.9 (0.9) 4.0 (1.0) 0.504
LDL (mmol/l) 1.9 (0.8) 1.8 (0.8) 2.0 (0.8) 0.227
HDL (mmol/l) 1.2 (0.4) 1.2 (0.4) 1.2 (0.4) 0.602
HbA1C (%) 7.9 (1.3) 7.9 (1.4) 7.7 (1.3) 0.580
HbA1C (mmol/l) 63 (14) 63 (15) 61 (14) 0.580
P-creatinine (µmol/l) 76.5 (18.3) 75.3 (18.3) 79.8 (18.1) 0.125
eGFR (ml/min/1.73 m2) 89 (17) 90.7 (17.5) 85.9 (16.6) 0.075
UAER (mg/24-h) 102 (39–229) 96 (38–195) 132 (44–487) 0.181
Microalbuminuria (%) 158 (79) 122 (82) 36 (69) 0.050
Macroalbuminuria (%) 42 (21) 26 (18) 16 (31) 0.050
Physical activity (h/week) 0.5 (0–3.5) 1 (0–4) 0 (0–2) 0.009
CAT (mm) 9.2 (3.4) 9.0 (3.5) 10.0 (3.4) 0.048
EAT (mm) 3.2 (1.6) 3.0 (1.5) 3.6 (1.9) 0.025
PAT (mm) 6.0 (2.9) 5.9 (2.8) 6.4 (2.9) 0.260
CAC baseline 183 (6–604) 91 (1–393) 626 (268–1831) <0.0001
CAC follow-up 781 (111–1792) 450 (82–1503) 1732 (1156–3705) 0.001
ΔCAC 509 (102–1096) 336 (70–948) 1071 (520–1725) 0.001
CIMT baseline (mm) 0.7 (0.2) 0.7 (0.1) 0.8 (0.2) 0.016
CIMT follow-up (mm) 0.8 (0.2) 0.8 (0.2) 0.8 (0.2) 0.165
ΔCIMT (mm) 0.1 (0.1) 0.1 (0.1) 0.1 (0.1) 0.084
Treatment with:
 Statin, n (%) 189 (95) 139 (94) 50 (96) 0.420
 Aspirin, n (%) 183 (92) 135 (91) 48 (92) 0.808
 Antihypertensive, n (%) 200 (100) 148 (100) 52 (100) 0.800
 Oral antidiabetic, n (%) 170 (85) 128 (87) 42 (81) 0.321
 Insulin, n (%) 124 (62) 87 (59) 37 (71) 0.114
Composite endpoint (CVD or mortality)
Characteristic Total No Yes p Value
Participants, n (%) 200 148 (74) 52 (26)
Male, n (%) 151 (76) 106 (72) 45 (87) 0.031
Age, years 59 (9) 58 (9) 62 (7) 0.001
Current smoker, (%) 59 (30) 39 (26) 20 (39) 0.072
Diabetes duration, (years) 13 (7) 12 (7) 15 (7) 0.004
Systolic blood pressure (mm Hg) 130 (16) 129 (15) 130 (16) 0.524
Body mass index (kg/m2) 32.6 (5.8) 32.7 (5.7) 32.0 (5.9) 0.461
Cholesterol (mmol/l) 3.9 (0.9) 3.9 (0.9) 4.0 (1.0) 0.504
LDL (mmol/l) 1.9 (0.8) 1.8 (0.8) 2.0 (0.8) 0.227
HDL (mmol/l) 1.2 (0.4) 1.2 (0.4) 1.2 (0.4) 0.602
HbA1C (%) 7.9 (1.3) 7.9 (1.4) 7.7 (1.3) 0.580
HbA1C (mmol/l) 63 (14) 63 (15) 61 (14) 0.580
P-creatinine (µmol/l) 76.5 (18.3) 75.3 (18.3) 79.8 (18.1) 0.125
eGFR (ml/min/1.73 m2) 89 (17) 90.7 (17.5) 85.9 (16.6) 0.075
UAER (mg/24-h) 102 (39–229) 96 (38–195) 132 (44–487) 0.181
Microalbuminuria (%) 158 (79) 122 (82) 36 (69) 0.050
Macroalbuminuria (%) 42 (21) 26 (18) 16 (31) 0.050
Physical activity (h/week) 0.5 (0–3.5) 1 (0–4) 0 (0–2) 0.009
CAT (mm) 9.2 (3.4) 9.0 (3.5) 10.0 (3.4) 0.048
EAT (mm) 3.2 (1.6) 3.0 (1.5) 3.6 (1.9) 0.025
PAT (mm) 6.0 (2.9) 5.9 (2.8) 6.4 (2.9) 0.260
CAC baseline 183 (6–604) 91 (1–393) 626 (268–1831) <0.0001
CAC follow-up 781 (111–1792) 450 (82–1503) 1732 (1156–3705) 0.001
ΔCAC 509 (102–1096) 336 (70–948) 1071 (520–1725) 0.001
CIMT baseline (mm) 0.7 (0.2) 0.7 (0.1) 0.8 (0.2) 0.016
CIMT follow-up (mm) 0.8 (0.2) 0.8 (0.2) 0.8 (0.2) 0.165
ΔCIMT (mm) 0.1 (0.1) 0.1 (0.1) 0.1 (0.1) 0.084
Treatment with:
 Statin, n (%) 189 (95) 139 (94) 50 (96) 0.420
 Aspirin, n (%) 183 (92) 135 (91) 48 (92) 0.808
 Antihypertensive, n (%) 200 (100) 148 (100) 52 (100) 0.800
 Oral antidiabetic, n (%) 170 (85) 128 (87) 42 (81) 0.321
 Insulin, n (%) 124 (62) 87 (59) 37 (71) 0.114

BMI: body mass index; CAC: coronary artery calcium; CAT: cardiac adipose tissue; CIMT: carotid intima media thickness; EAT: epicardial adipose tissue; eGFR: estimated glomerular filtration rate; HbA1C: glycated haemoglobin; HDL: high-density lipoprotein; IQR: interquartile range; LDL: low-density lipoprotein; PAT: pericardial adipose tissue; SD: standard deviation; UAER: urinary albumin excretion rate.

Data are expressed as means (SD), medians (IQR) and number (n) and percentage (%). Unpaired t test was used for parametric means (SD) or log-transformed non-parametric data (median (IQR) values). Chi2 test was used for categorical data.

Table 1.

Patient characteristics at baseline according to event status.

Composite endpoint (CVD or mortality)
Characteristic Total No Yes p Value
Participants, n (%) 200 148 (74) 52 (26)
Male, n (%) 151 (76) 106 (72) 45 (87) 0.031
Age, years 59 (9) 58 (9) 62 (7) 0.001
Current smoker, (%) 59 (30) 39 (26) 20 (39) 0.072
Diabetes duration, (years) 13 (7) 12 (7) 15 (7) 0.004
Systolic blood pressure (mm Hg) 130 (16) 129 (15) 130 (16) 0.524
Body mass index (kg/m2) 32.6 (5.8) 32.7 (5.7) 32.0 (5.9) 0.461
Cholesterol (mmol/l) 3.9 (0.9) 3.9 (0.9) 4.0 (1.0) 0.504
LDL (mmol/l) 1.9 (0.8) 1.8 (0.8) 2.0 (0.8) 0.227
HDL (mmol/l) 1.2 (0.4) 1.2 (0.4) 1.2 (0.4) 0.602
HbA1C (%) 7.9 (1.3) 7.9 (1.4) 7.7 (1.3) 0.580
HbA1C (mmol/l) 63 (14) 63 (15) 61 (14) 0.580
P-creatinine (µmol/l) 76.5 (18.3) 75.3 (18.3) 79.8 (18.1) 0.125
eGFR (ml/min/1.73 m2) 89 (17) 90.7 (17.5) 85.9 (16.6) 0.075
UAER (mg/24-h) 102 (39–229) 96 (38–195) 132 (44–487) 0.181
Microalbuminuria (%) 158 (79) 122 (82) 36 (69) 0.050
Macroalbuminuria (%) 42 (21) 26 (18) 16 (31) 0.050
Physical activity (h/week) 0.5 (0–3.5) 1 (0–4) 0 (0–2) 0.009
CAT (mm) 9.2 (3.4) 9.0 (3.5) 10.0 (3.4) 0.048
EAT (mm) 3.2 (1.6) 3.0 (1.5) 3.6 (1.9) 0.025
PAT (mm) 6.0 (2.9) 5.9 (2.8) 6.4 (2.9) 0.260
CAC baseline 183 (6–604) 91 (1–393) 626 (268–1831) <0.0001
CAC follow-up 781 (111–1792) 450 (82–1503) 1732 (1156–3705) 0.001
ΔCAC 509 (102–1096) 336 (70–948) 1071 (520–1725) 0.001
CIMT baseline (mm) 0.7 (0.2) 0.7 (0.1) 0.8 (0.2) 0.016
CIMT follow-up (mm) 0.8 (0.2) 0.8 (0.2) 0.8 (0.2) 0.165
ΔCIMT (mm) 0.1 (0.1) 0.1 (0.1) 0.1 (0.1) 0.084
Treatment with:
 Statin, n (%) 189 (95) 139 (94) 50 (96) 0.420
 Aspirin, n (%) 183 (92) 135 (91) 48 (92) 0.808
 Antihypertensive, n (%) 200 (100) 148 (100) 52 (100) 0.800
 Oral antidiabetic, n (%) 170 (85) 128 (87) 42 (81) 0.321
 Insulin, n (%) 124 (62) 87 (59) 37 (71) 0.114
Composite endpoint (CVD or mortality)
Characteristic Total No Yes p Value
Participants, n (%) 200 148 (74) 52 (26)
Male, n (%) 151 (76) 106 (72) 45 (87) 0.031
Age, years 59 (9) 58 (9) 62 (7) 0.001
Current smoker, (%) 59 (30) 39 (26) 20 (39) 0.072
Diabetes duration, (years) 13 (7) 12 (7) 15 (7) 0.004
Systolic blood pressure (mm Hg) 130 (16) 129 (15) 130 (16) 0.524
Body mass index (kg/m2) 32.6 (5.8) 32.7 (5.7) 32.0 (5.9) 0.461
Cholesterol (mmol/l) 3.9 (0.9) 3.9 (0.9) 4.0 (1.0) 0.504
LDL (mmol/l) 1.9 (0.8) 1.8 (0.8) 2.0 (0.8) 0.227
HDL (mmol/l) 1.2 (0.4) 1.2 (0.4) 1.2 (0.4) 0.602
HbA1C (%) 7.9 (1.3) 7.9 (1.4) 7.7 (1.3) 0.580
HbA1C (mmol/l) 63 (14) 63 (15) 61 (14) 0.580
P-creatinine (µmol/l) 76.5 (18.3) 75.3 (18.3) 79.8 (18.1) 0.125
eGFR (ml/min/1.73 m2) 89 (17) 90.7 (17.5) 85.9 (16.6) 0.075
UAER (mg/24-h) 102 (39–229) 96 (38–195) 132 (44–487) 0.181
Microalbuminuria (%) 158 (79) 122 (82) 36 (69) 0.050
Macroalbuminuria (%) 42 (21) 26 (18) 16 (31) 0.050
Physical activity (h/week) 0.5 (0–3.5) 1 (0–4) 0 (0–2) 0.009
CAT (mm) 9.2 (3.4) 9.0 (3.5) 10.0 (3.4) 0.048
EAT (mm) 3.2 (1.6) 3.0 (1.5) 3.6 (1.9) 0.025
PAT (mm) 6.0 (2.9) 5.9 (2.8) 6.4 (2.9) 0.260
CAC baseline 183 (6–604) 91 (1–393) 626 (268–1831) <0.0001
CAC follow-up 781 (111–1792) 450 (82–1503) 1732 (1156–3705) 0.001
ΔCAC 509 (102–1096) 336 (70–948) 1071 (520–1725) 0.001
CIMT baseline (mm) 0.7 (0.2) 0.7 (0.1) 0.8 (0.2) 0.016
CIMT follow-up (mm) 0.8 (0.2) 0.8 (0.2) 0.8 (0.2) 0.165
ΔCIMT (mm) 0.1 (0.1) 0.1 (0.1) 0.1 (0.1) 0.084
Treatment with:
 Statin, n (%) 189 (95) 139 (94) 50 (96) 0.420
 Aspirin, n (%) 183 (92) 135 (91) 48 (92) 0.808
 Antihypertensive, n (%) 200 (100) 148 (100) 52 (100) 0.800
 Oral antidiabetic, n (%) 170 (85) 128 (87) 42 (81) 0.321
 Insulin, n (%) 124 (62) 87 (59) 37 (71) 0.114

BMI: body mass index; CAC: coronary artery calcium; CAT: cardiac adipose tissue; CIMT: carotid intima media thickness; EAT: epicardial adipose tissue; eGFR: estimated glomerular filtration rate; HbA1C: glycated haemoglobin; HDL: high-density lipoprotein; IQR: interquartile range; LDL: low-density lipoprotein; PAT: pericardial adipose tissue; SD: standard deviation; UAER: urinary albumin excretion rate.

Data are expressed as means (SD), medians (IQR) and number (n) and percentage (%). Unpaired t test was used for parametric means (SD) or log-transformed non-parametric data (median (IQR) values). Chi2 test was used for categorical data.

Patients with CVD or mortality had more EAT (p = 0.025) and CAT (p = 0.048), higher levels of CAC (p < 0.0001) and CIMT (p = 0.016), compared to patients without an event. Progression in CAC was more pronounced (p = 0.001) in patients with vs without the composite endpoint (Table 1).

Follow-up

Median follow-up was 6.1 years (IQR 5.9–6.6 years), and during follow-up 52 events had occurred including non-fatal CVD (n = 29) or death without previous non-fatal CVD (n = 23).

Non-fatal CVD events included myocardial infarction, stroke, ischaemic cardiovascular disease, and heart failure.14 Of the 23 deaths, 11 were classified as CVD-related, nine as cancer-related, and three as related to other causes.

Cardiac fat and cardiovascular morbidity and mortality

Total CAT

In model 1, CAT was associated with the composite endpoint including CVD and all-cause mortality (HR 1.1, CI 1.0–1.2, p = 0.049) (Table 2). The association did not remain in model 2 (HR 1.1, CI 1.0–1.2, p = 0.122) when adjusting for age and sex or after further adjustment of risk factors in model 3 (HR 1.1, CI 1.0–1.3, p = 0.065). CAT was not associated with the composite endpoint in a non-linear manner in model 1 (HR 1.2, CI 0.8–1.7, p = 0.307). The median CAT level was 8.75 mm and in a sub-analysis when patients were split into two groups by the median, the risk of the composite endpoint was higher in patients with CAT levels above vs below the median (HR 1.9, CI 1.1–3.4, p = 0.027) in model 1. In model 2 (HR 2.0, CI 1.1–3.5, p = 0.022) and model 3 (HR 2.0, CI 1.1–3.7, p = 0.017), patients with higher CAT levels had a doubled risk of the composite endpoint vs patients with CAT levels below the median (Figure 1). The elevated risk for patients with high CAT levels was maintained when a combined CVD morbidity/mortality endpoint was used including only non-fatal and fatal CVD in the fully adjusted model (HR 1.92, CI 1.0–3.7, p = 0.049). Regarding subclinical atherosclerosis, CAT was associated with CAC score at baseline (1.12, CI 1.00–1.25, p = 0.033) in linear regression model 1 but not in model 2 and 3 (all p ≥ 0.236) or to progression of CAC at follow-up adjusted for baseline CAC (p > 0.671). In models 1–3, there was no correlation of CAT and CIMT at baseline (all p ≥ 0.478) or progression at follow-up (all p ≥ 0.756).

Table 2.

Cox regression of the association of cardiac adipose tissue (CAT), epicardial adipose tissue (EAT) and pericardial adipose tissue (PAT) and incident cardiovascular disease (CVD) and all-cause mortality.

Fatal and non-fatal CVD event (n = 40)
Composite (n = 52)
Adipose tissue Model HR (95% CI) p Value HR (95% CI) p Value
1 1.1 (1.0–1.2) 0.208 1.1 (1.0–1.2) 0.049
CAT 2 1.0 (1.0–1.1) 0.446 1.1 (1.0–1.2) 0.122
3 1.0 (1.0–1.1) 0.251 1.1 (1.0–1.2) 0.065
1 1.2 (1.0–1.4) 0.115 1.2 (1.0–1.4) 0.028
EAT 2 1.1 (0.9–1.3) 0.362 1.1 (1.0–1.3) 0.102
3 1.1 (0.9–1.3) 0.316 1.1 (1.0–1.3) 0.097
1 1.0 (0.9–1.2) 0.511 1.1 (1.0–1.2) 0.245
PAT 2 1.0 (0.9–1.1) 0.685 1.0 (1.0–1.2) 0.350
3 1.0 (0.9–1.2) 0.425 1.1 (1.0–1.6) 0.213
Fatal and non-fatal CVD event (n = 40)
Composite (n = 52)
Adipose tissue Model HR (95% CI) p Value HR (95% CI) p Value
1 1.1 (1.0–1.2) 0.208 1.1 (1.0–1.2) 0.049
CAT 2 1.0 (1.0–1.1) 0.446 1.1 (1.0–1.2) 0.122
3 1.0 (1.0–1.1) 0.251 1.1 (1.0–1.2) 0.065
1 1.2 (1.0–1.4) 0.115 1.2 (1.0–1.4) 0.028
EAT 2 1.1 (0.9–1.3) 0.362 1.1 (1.0–1.3) 0.102
3 1.1 (0.9–1.3) 0.316 1.1 (1.0–1.3) 0.097
1 1.0 (0.9–1.2) 0.511 1.1 (1.0–1.2) 0.245
PAT 2 1.0 (0.9–1.1) 0.685 1.0 (1.0–1.2) 0.350
3 1.0 (0.9–1.2) 0.425 1.1 (1.0–1.6) 0.213

BMI: body mass index; CI confidence interval; HbA1c: glycated haemoglobin; HR: hazard ratio; LDL: low-density lipoprotein.

HR values are presented per mm increase in adipose tissue thickness. Model 1 is unadjusted; model 2 is adjusted for age and sex; model 3 is adjusted for age, sex, LDL-cholesterol, HbA1c, systolic blood pressure, smoking and BMI.

Table 2.

Cox regression of the association of cardiac adipose tissue (CAT), epicardial adipose tissue (EAT) and pericardial adipose tissue (PAT) and incident cardiovascular disease (CVD) and all-cause mortality.

Fatal and non-fatal CVD event (n = 40)
Composite (n = 52)
Adipose tissue Model HR (95% CI) p Value HR (95% CI) p Value
1 1.1 (1.0–1.2) 0.208 1.1 (1.0–1.2) 0.049
CAT 2 1.0 (1.0–1.1) 0.446 1.1 (1.0–1.2) 0.122
3 1.0 (1.0–1.1) 0.251 1.1 (1.0–1.2) 0.065
1 1.2 (1.0–1.4) 0.115 1.2 (1.0–1.4) 0.028
EAT 2 1.1 (0.9–1.3) 0.362 1.1 (1.0–1.3) 0.102
3 1.1 (0.9–1.3) 0.316 1.1 (1.0–1.3) 0.097
1 1.0 (0.9–1.2) 0.511 1.1 (1.0–1.2) 0.245
PAT 2 1.0 (0.9–1.1) 0.685 1.0 (1.0–1.2) 0.350
3 1.0 (0.9–1.2) 0.425 1.1 (1.0–1.6) 0.213
Fatal and non-fatal CVD event (n = 40)
Composite (n = 52)
Adipose tissue Model HR (95% CI) p Value HR (95% CI) p Value
1 1.1 (1.0–1.2) 0.208 1.1 (1.0–1.2) 0.049
CAT 2 1.0 (1.0–1.1) 0.446 1.1 (1.0–1.2) 0.122
3 1.0 (1.0–1.1) 0.251 1.1 (1.0–1.2) 0.065
1 1.2 (1.0–1.4) 0.115 1.2 (1.0–1.4) 0.028
EAT 2 1.1 (0.9–1.3) 0.362 1.1 (1.0–1.3) 0.102
3 1.1 (0.9–1.3) 0.316 1.1 (1.0–1.3) 0.097
1 1.0 (0.9–1.2) 0.511 1.1 (1.0–1.2) 0.245
PAT 2 1.0 (0.9–1.1) 0.685 1.0 (1.0–1.2) 0.350
3 1.0 (0.9–1.2) 0.425 1.1 (1.0–1.6) 0.213

BMI: body mass index; CI confidence interval; HbA1c: glycated haemoglobin; HR: hazard ratio; LDL: low-density lipoprotein.

HR values are presented per mm increase in adipose tissue thickness. Model 1 is unadjusted; model 2 is adjusted for age and sex; model 3 is adjusted for age, sex, LDL-cholesterol, HbA1c, systolic blood pressure, smoking and BMI.

Figure 1.

Kaplan-Meyer proportional hazards plot of cardiac adipose tissue (CAT) and risk of cardiovascular disease (CVD) or mortality. CVD or mortality in patients with CAT amounts below the median (low CAT) group (blue) and above the median (high CAT) group (green). Hazard ratio (HR) 2.0, confidence interval (CI): 1.1; 3.7, p = 0.017.

EAT

In model 1, EAT was associated with the composite endpoint including CVD and all-cause mortality (HR 1.2, CI 1.0–1.4, p = 0.028) (Table 2). The association did not remain in model 2 (HR 1.1, CI 1.0–1.3, p = 0.102) or model 3 (HR 1.1, CI 1.0–1.3, p = 0.097). There was no non-linear association of EAT and CVD and all-cause mortality (HR 1.7, CI 0.6–4.6, p = 0.288) in model 1, and no association when patients were split into two groups according to the median amount of EAT (2.95 mm) (HR 1.17, CI 0.7–2.0, p = 0.576). EAT was not associated with CAC (p = 0.130) or CIMT (p = 0.427) at baseline or at progression (all p ≥ 0.431).

PAT

PAT was not associated linearly or non-linearly with the composite endpoint in models 1–3 (all p > 0.245) (Table 2) or when split by the median (5.5 mm) PAT level (HR 1.47, CI 0.8–2.6, p = 0.176). PAT was not associated with unadjusted CAC (p = 0.089) or CIMT (p = 0.491) at baseline or at progression (all p ≥ 0.851).

Cardiac adipose tissue and markers of low-grade inflammation

Total cardiac fat

CAT was not linearly associated to any marker of low-grade inflammation in models 1–3 (p > 0.075). When split by the median, patients with high CAT levels had higher IL-8 levels in model 1 (14%, CI 1–30%, p = 0.039) compared with patients with CAT levels below the median. In model 2 (14%, CI 1–30%, p = 0.038) and model 3 (14%, CI 1–30% p = 0.026) the IL-8 levels remained 14% higher for patients with CAT levels above vs. below the median. High CAT levels were not associated with any other marker of low-grade inflammation in models 1–3 (hsCRP, TNFα, IL-1β, IL-6) (all p ≥ 0.264) (Supplementary Material, Table 1).

EAT

EAT was associated with IL-8 in model 1 (4%, CI 0.1–7.8%, p = 0.049), model 2 (4%, CI 0.2–8.0%, p = 0.046) and model 3 (4%, CI 0.18–8.37, p = 0.047) but not with any other markers of low-grade inflammation (p ≥ 0.079). When split by the median, high EAT levels were not associated with markers of low-grade inflammation in models 1–3 (all p > 0.073).

PAT

There were no linear or binary association of PAT to any markers of low-grade inflammation (hsCRP, TNFα, IL-1β, IL-6) in models 1–3 (all p > 0.493).

Low-grade inflammation, CVD and CAT

A sub-analysis revealed that IL-8 was positively associated with the unadjusted composite endpoint of CVD and all-cause mortality (HR 1.01, CI 1.00–1.02, p = 0.004). Furthermore, IL-8 was linearly correlated to progression of CAC when baseline CAC levels were accounted for: When IL-8 was doubled the follow-up CAC levels were increased by 50% (50.0% (12.5–99.9%), p = 0.006). The association remained in model 2 (45.2% (9.4–92.7%), p = 0.012) and model 3 (40.0% (5.5–85.7%), p = 0.026). When added separately to model 3, plasma IL-8 (HR 2.0, CI 1.1–3.7, p = 0.018) did not influence the strength of the association of high CAT levels with CVD and all-cause mortality. Other inflammatory variables including TNFα (HR 2.0, CI 1.1–3.6, p = 0.020) and physical activity (HR 2.0, CI 1.1–3.7, p = 0.019) did not change the CAT associated risk of the composite outcome.

Discussion

In this study of 200 T2D patients with elevated urinary albumin excretion rate and without clinical features of CAD, we found no evidence that continuous measures of CAT were associated with CVD. However, when we explored binary CAT levels we found them to be associated with a composite endpoint of incident CVD and all-cause mortality after 6.1 years of follow-up. When patients, according to proposed threshold values, were split into two groups by the median amount of CAT, the patients with high CAT levels had a doubled risk of incident CVD or death compared to patients with lower CAT levels, even after adjusting for traditional CVD risk factors.11,15

The knowledge of risk factors and biomarkers of diabetic complications has increased over recent years and pose promise for future diagnostic and prognostic markers and therapeutic targets.14 Despite this, type 2 diabetic patients still have an increased risk of CVD compared to individuals without diabetes.1 Thus, additional biomarkers which can lead to early identification of high-risk patients are needed, and a focus on using cardiac fat as a promising imaging biomarker is increasing.12 Recent cross-sectional studies have reported that increased cardiac fat is associated with subclinical atherosclerosis.6,7 Yet, only one prospective community-based case control and one case cohort study have demonstrated that higher levels of cardiac fat were associated with future CVD.8,9 Our study extends these findings by demonstrating that high total cardiac fat levels were independently associated with increased risk of CVD and all-cause mortality in a high risk T2D population after six years of follow-up.

In this study, we did not find a significant linear or non-linear association between thickness of cardiac fat and the composite endpoint. Yet, biological threshold values of cardiac fat above 7–9 mm have been associated with CAD previously.11,15 Thus, when we, as an initial approach, used the median CAT thickness of 8.75 mm as a biological threshold our results showed that high levels of CAT were associated to both the composite endpoint of CVD and all-cause mortality and also the CVD endpoint of non-fatal and fatal CVD. This exclusively binary association may indicate that cardiac fat at physiological levels predominantly displays cardio-protective functions (e.g. brown fat functions, energy reserve, mechanical and structural support).1618 Conversely, when cardiac fat is in excess it may play an adverse paracrine pro-atherogenic role.18

Total cardiac fat consists of EAT and PAT, which have different embryonic origin, vascularization, location and possibly functions.19 Despite this, the literature often inconsistently discriminates between the two fat layers.19 In this study, EAT but not PAT was associated with CVD and all-cause mortality in unadjusted models. EAT is located adjacent to the coronary arteries and may promote CVD via paracrine signalling of pro-atherogenic substances.4 However, PAT lacks direct myocardial contact and may only indirectly influence the process, as suggested by the association of total CAT but not PAT with CVD and all-cause mortality in this study.

Mazurek et al. hypothesized an inflammatory pro-atherogenic mechanism of EAT after demonstrating that epicardial fat biopsies from patients with CVD exhibited infiltration of chronic inflammatory cells and increased inflammatory gene expression compared to subcutaneous fat.4 Chatterjee et al. found higher levels of IL-6, IL-8 and MCP-1 gene expression in EAT from organ donors, and demonstrated that cultured perivascular adipocytes secreted IL-8.3 Interestingly, we found that plasma IL-8, but none of the other inflammatory markers, was positively associated with EAT and with increased CAT levels in T2D patients. IL-8 is a powerful chemokine, which is found in atherosclerotic plaques and can attract monocytes for firm adhesion and promote CVD.20 Mirroring this, we found IL-8 to be associated with our primary endpoint of CVD and all-cause mortality. However, IL-8 did not change the CAT-associated risk of CVD and all-cause mortality.

Strengths and limitations

The strength of the study is the long follow-up, where no patients were lost to follow-up. Longitudinal studies in this field are lacking,18 and to the best of our knowledge this is the first study to assess cardiac fat layers separately in a high-risk population of T2D. Extensive risk factor adjustment was possible in this study, and solidified the association of CAT and CVD and all-cause mortality. Sensitivity analyses including not only the pre-specified variables but also the variables found specifically to be associated with the outcome in this population (IL8, TNFα and physical activity) did not change the association between CAT levels and outcome.

The study sample size and event rate resulted in limited statistical power, and no correction for multiple testing was performed. No non-diabetic control group was available. CAT is closely associated to total visceral fat,21 and thus fat depots in other regions may contribute to the association with CVD and death. Unfortunately, we did not have a measurement of visceral fat in this study. We utilized echocardiography to obtain cardiac fat measurements, and not the gold-standard magnetic resonance imaging (MRI) or computed tomography (CT), and thus were not able to investigate CAT radiodensity, which has been associated with CVD.22,23 However, volumetric recordings from echocardiography correlate well with cardiac MRI.21 Echocardiography was used in this study, as it is readily available, fast, inexpensive and the recommended method for potential future clinical screening of cardiac fat to assess CVD risk.12

Clinical implications

Larger prospective studies are necessary to clarify whether a linear or nonlinear association of CAT with CVD and all-cause mortality exists. Furthermore, studies are needed to explore whether CAT is a risk factor of CVD. Subsequently, randomized clinical trials may clarify whether anti-inflammatory drugs or lifestyle interventions could reduce the cardiac fat and its potential inflammatory role in CVD progression in patients with T2D.

Conclusions

In patients with T2D and elevated urinary albumin excretion rate and no clinical features of CAD, a higher than the median amount of total cardiac fat, rather than EAT or PAT alone, was associated with increased risk of incident CVD and all-cause mortality. This association remained even after adjusting for traditional CVD risk factors. Higher levels of CAT were associated with subclinical atherosclerosis (CAC) at baseline, and both CAT and EAT were associated with the pro-atherogenic inflammatory marker IL-8. Large-scale longitudinal studies are warranted to confirm and expand the findings of this study.

Author contribution

BJvS, CSH, HR, HHP, PKJ, MEJ, BKP and PR conceived and designed the research; RCH, BJvS, CSH, HR, JBR, SEH and PR analysed and interpreted the data; RHC, BJvS, CSH performed the statistical analysis; RHC wrote the manuscript; BJvS, CSH, HR, HHP, PKJ, MEJ, BKP, SEH, UBA, PH critically revised the manuscript for key intellectual content; and CSH, BJvS, MEJ supervised the study. All authors approved the final version of the manuscript. RHC is responsible for the integrity of the work as a whole.

Acknowledgements

The authors wish to thank all participants and acknowledge the work of L Jelstrup, UM Smidt, AG Lundgaard, BR Jensen, TR Juhl and JA Hermann, Steno Diabetes Center, Gentofte, Denmark.

Declaration of conflicting interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: The authors declare no conflict of interest associated with this manuscript. PR reports having given lectures for Astra Zeneca, BMS and Boehringer Ingelheim, and has served as a consultant for AbbVie, Astra Zeneca, BMS, Eli Lilly, Boehringer Ingelheim, Astellas, Janssen and Novo Nordisk, all fees given to Steno Diabetes Center, and has equity interest in Novo Nordisk.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by internal funding by Steno Diabetes Center and Center of Inflammation and Metabolism/Center for Physical Activity Research. This work was also supported by the Danish Heart foundation (grant number 16-R107-A6704-22970).

References

1

Cavender
MA
,
Steg
PG
,
Smith
SC
et al.   
Impact of diabetes mellitus on hospitalization for heart failure, cardiovascular events, and death: Outcomes at 4 years from the Reduction of Atherothrombosis for Continued Health (REACH) registry
.
Circulation
 
2015
;
132
:
923
931
.

2

Yang
F-S
,
Yun
C-H
,
Wu
T-H
et al.   
High pericardial and peri-aortic adipose tissue burden in pre-diabetic and diabetic subjects
.
BMC Cardiovasc Disord
 
2013
;
13
:
98
98
.

3

Chatterjee
TK
,
Stoll
LS
,
Denning
GD
et al.   
Pro-inflammatory phenotype of perivascular adipocytes: Influence of high fat feeding
.
Circ Res
 
2009
;
104
:
541
549
.

4

Mazurek
T
,
Zhang
L
,
Zalewski
A
et al.   
Human epicardial adipose tissue is a source of inflammatory mediators
.
Circulation
 
2003
;
108
:
2460
2466
.

5

Zakynthinos
E
,
Pappa
N
.
Inflammatory biomarkers in coronary artery disease
.
J Cardiol
 
2009
;
53
:
317
333
.

6

Bos
D
,
Shahzad
R
,
Walsum
T
et al.   
Epicardial fat volume is related to atherosclerotic calcification in multiple vessel beds
.
Eur Heart J Cardiovasc Imaging
 
2015
;
16
:
1264
1269
.

7

Sengul
C
,
Cevik
C
,
Ozveren
O
et al.   
Echocardiographic epicardial fat thickness is associated with carotid intima-media thickness in patients with metabolic syndrome
.
Echocardiography
 
2011
;
28
:
853
858
.

8

Ding
J
,
Hsu
F-C
,
Harris
TB
et al.   
The association of pericardial fat with incident coronary heart disease: The Multi-Ethnic Study of Atherosclerosis (MESA) 1–3
.
Am J Clin Nutr
 
2009
;
90
:
499
504
.

9

Cheng
VY
,
Dey
D
,
Tamarappoo
B
et al.   
Pericardial fat burden on ECG-Gated noncontrast CT in asymptomatic patients who subsequently experience adverse cardiovascular events
.
JACC Cardiovasc Imaging
 
2010
;
3
:
352
360
.

10

Reinhard
H
,
Hansen
PR
,
Persson
F
et al.   
Elevated NT-proBNP and coronary calcium score in relation to coronary artery disease in asymptomatic type 2 diabetic patients with elevated urinary albumin excretion rate
.
Nephrol Dial Transplant
 
2011
;
26
:
3242
3249
.

11

Iacobellis
G
,
Willens
HJ
.
Echocardiographic epicardial fat: A review of research and clinical applications
.
J Am Soc Echocardiogr
 
2009
;
22
:
1311
1319
.

12

Davidovich
D
,
Gastaldelli
A
,
Sicari
R
.
Imaging cardiac fat
.
Eur Heart J Cardiovasc Imaging
 
2013
;
14
:
625
630
.

13

Reinhard
H
,
Wiinberg
N
,
Hansen
PR
et al.   
NT-proBNP levels, atherosclerosis and vascular function in asymptomatic type 2 diabetic patients with microalbuminuria: Peripheral reactive hyperaemia index but not NT-proBNP is an independent predictor of coronary atherosclerosis
.
Cardiovasc Diabetol
 
2011
;
10
:
71
71
.

14

Von Scholten
BJ
,
Reinhard
H
,
Hansen
TW
et al.   
Markers of inflammation and endothelial dysfunction are associated with incident cardiovascular disease, all-cause mortality, and progression of coronary calcification in type 2 diabetic patients with microalbuminuria
.
J Diabetes Complications
 
2016
;
30
:
248
255
.

15

Iacobellis
G
,
Willens
HJ
,
Barbaro
G
et al.   
Threshold values of high-risk echocardiographic epicardial fat thickness
.
Obesity (Silver Spring)
 
2008
;
16
:
887
892
.

16

Sacks
HS
,
Fain
JN
,
Bahouth
SW
et al.   
Adult epicardial fat exhibits beige features
.
J Clin Endocrinol Metab
 
2013
;
98
:
1448
1455
.

17

Marchington
JM
,
Mattacks
CA
,
Pond
CM
.
Adipose tissue in the mammalian heart and pericardium; structure, fetal development and biochemical properties
.
Comp Biochem Physiol
 
1989
;
94
:
225
232
.

18

Sacks
HS
,
Fain
JN
.
Human epicardial fat: What is new and what is missing?
 
Clin Exp Pharmacol Physiol
 
2011
;
38
:
879
887
.

19

Iacobellis
G
.
Epicardial and pericardial fat: Close, but very different
.
Obesity
 
2009
;
17
:
625
625
.

20

Gerszten
RE
,
Garcia-Zepeda
EA
,
Lim
Y-C
et al.   
MCP-1 and IL-8 trigger firm adhesion of monocytes to vascular endothelium under flow conditions
.
Nature
 
1999
;
398
:
718
723
.

21

Iacobellis
G
,
Assael
F
,
Ribaudo
MC
et al.   
Epicardial fat from echocardiography: A new method for visceral adipose tissue prediction
.
Obes Res
 
2003
;
11
:
304
310
.

22

Franssens
BT
,
Nathoe
HM
,
Leiner
T
et al.   
Relation between cardiovascular disease risk factors and epicardial adipose tissue density on cardiac computed tomography in patients at high risk of cardiovascular events
.
Eur J Prev Cardiol
 
2017
;
24
:
660
670
.

23

Dutour
A
.
Looking beyond ectopic fat amount: A SMART method to quantify epicardial adipose tissue density
.
Eur J Prev Cardiol
 
2017
;
24
:
657
659
.

This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Comments

0 Comments
Submit a comment
You have entered an invalid code
Thank you for submitting a comment on this article. Your comment will be reviewed and published at the journal's discretion. Please check for further notifications by email.