Journal of Nephrology

, Volume 29, Issue 2, pp 241–250 | Cite as

Epicardial adipose tissue in long-term hemodialysis patients: its association with vascular calcification and long-term development

  • Xoana Barros
  • Timm Dirrichs
  • Ralf Koos
  • Sebastian Reinartz
  • Nadine Kaesler
  • Rafael Kramann
  • Ulrich Gladziwa
  • Markus Ketteler
  • Jürgen Floege
  • Nikolaus Marx
  • José V. Torregrosa
  • András Keszei
  • Vincent M. Brandenburg
Original Article



Epicardial adipose tissue (EAT) is associated with coronary artery disease (CAD) in the general population. EAT is suggested to promote CAD by paracrine mechanisms and local inflammation. We evaluated whether in chronic hemodialysis (HD) patients EAT associates with CAD, how the amount of EAT develops over time, and if EAT independently predicts the mortality risk.


Post-hoc analysis of a prospective study in 59 chronic HD patients who underwent non-enhanced multi-slice computed tomography (MSCT) at baseline. Thirty-seven patients underwent another MSCT after 24 ± 5 months. We measured EAT volume (cm³) and Agatston calcification scores of coronary arteries (CAC) and aortic valves (AVC). All-cause mortality was assessed after a follow-up of 88 months (IQR 52–105).


Baseline EAT was 128.2 ± 60.8 cm³ and significantly higher than in a control group of non-renal patients (94 ± 46 cm³; p < 0.05). Median Agatston score for CAC was 329 (IQR 23–1181) and for AVC was 0 (IQR 0–25.3) in HD patients. We observed significant positive correlations between baseline EAT and age (r = 0.386; p = 0.003), BMI (r = 0.314; p = 0.016), CAC (r = 0.278; p = 0.03), and AVC (r = 0.282; p = 0.03). In multivariate analysis, age, BMI and AVC remained as significant predictors of EAT (p < 0.01). Calcification scores significantly increased over 2 years; in contrast EAT change was not significant (+11 %, IQR −10 to 24 %; p = 0.066). The limited patient number in the present study precludes analysis of the EAT impact upon survival.


EAT correlated significantly with cardiovascular calcification in long-term HD patients. Mean EAT did not significantly change over 2 years.


Epicardial adipose tissue Hemodialysis Coronary artery calcification Survival 


Cardiovascular (CV) mortality is massively increased in patients with end-stage renal disease (ESRD). Non-enhanced multi-slice computed tomography (MSCT) can help identify patients with particularly high cardiovascular risk by quantifying coronary artery calcification (CAC) and aortic valve calcification (AVC) [1]. As vascular disease is a highly prevalent phenomenon in patients with ESRD [2], calcification scoring has been an established method for risk assessment now for almost 20 years [1, 3], whereas computed tomography (CT)-based quantification of epicardial adipose tissue (EAT) and its use as a cardiovascular or mortality risk predictor is a novel and evolving field.

EAT is located within the parietal pericardium and can be accurately measured in terms of volume and mass by CT [4]. EAT shares an embryologic origin with visceral fat and is also metabolically active, producing cytokines and proinflammatory mediators [5]. Given its close anatomical proximity, EAT may exert paracrine effects on adjacent coronary arteries promoting the development of atherosclerotic plaques [6]. In the general population EAT thickness has been related to incident fatal and nonfatal coronary events independently of traditional cardiovascular risk factors [7, 8, 9] and coronary artery disease [9]. Both vascular calcification as well as EAT can be measured with the same diagnostic tool in one single diagnostic procedure. Thus, it can be easily assessed whether measurement of both parameters in parallel can optimize risk stratification in ESRD patients beyond calcification scoring alone [8].

Currently, our knowledge about the associations between EAT, arteriosclerosis and CV risk in ESRD patients is limited to a few studies [10, 11, 12]. Among those studies, only a post hoc analysis of the Renagel in New Dialysis (RIND) patients study [13] has analyzed the effect of EAT on survival in 95 incident hemodialysis (HD) patients [12]. EAT was an independent predictor for mortality in this cohort. All prior studies relied exclusively on cross-sectional measurements of EAT and other CV risk markers and did not provide information about the longitudinal evolution of EAT.

The present data emerge from a post hoc analysis of a prospective, longitudinal heart MSCT study in prevalent HD patients, whose baseline calcification data have been previously published [14]. The aim of this study was to investigate the longitudinal evolution of EAT over time and the potential association of EAT with survival in prevalent, long-term hemodialysis patients with a particularly high CAC burden.

Materials and methods

Patient characteristics

Patient data were extracted from a prospective longitudinal follow-up study in a cohort of chronic HD patients recruited at the Aachen University Hospital and three collaborating outpatient dialysis centers. Baseline data were partly analyzed regarding the association of biomarkers with vascular calcification [14].

Standard bicarbonate dialysis procedures were a thrice-weekly hemodialysis or hemodiafiltration session for 4.5–5.5 h each. Dialysate calcium concentration was 1.25 or 1.5 mmol/l. All adult patients were eligible. Exclusion criteria were: less than three HD sessions per week, anticipated living kidney donation, current atrial fibrillation, body weight exceeding 130 kg, immobilization, severe comorbidities requiring medical assistance for transport and performance of MSCT scanning, history of combined coronary bypass surgery or coronary stent implantation or aortic valve surgery, cancer, ethanol or drug abuse, and claustrophobia. Patients with a previous kidney transplant were only included if they had re-initiated regular HD therapy.

Heart CT was performed in 67 patients at baseline. From those, one patient was excluded from further analysis because CAC evaluation revealed a previously unknown history of coronary stent implantation. Quantification of recorded EAT assessment was not possible in seven patients because of incomplete recording of the area of interest. Therefore, a complete data set (EAT, CAC and AVC) was obtained in 59 patients at baseline.

Patients were interviewed prior to second MSCT regarding de-novo cardiac surgery or percutaneous coronary intervention (PCI) and excluded if such procedures were reported. The reasons for not performing a second MSCT scan were withdrawal of consent (n = 3), kidney transplantation (n = 3), death (n = 3), PCI or aortic valve surgery (n = 6), unknown (n = 3), or other reasons (n = 4). A second MSCT was performed in 37 patients (63 %).

Clinical data and blood samples were obtained prior to the first MSCT. Data were obtained during interview and physical examination as well as from a review of the patient’s files. For patient details regarding the baseline and the follow-up cohort, respectively, please refer to Table 1.
Table 1

Demographic, clinical and laboratory variables at baseline


Entire cohort (n = 59)

Follow-up cohort (n = 37)

Age (years)

59.12 ± 15.05

60 ± 14.6

Women n (%)

29 (49.2)

20 (54.1)

BMI (kg/m2)

25.47 ± 5.72

25.81 ± 6.74

Diabetes n (%)

15 (25.4)

11 (29.7)

Hypertension n (%)

53 (89.9)

34 (91.9)

Current smoker n (%)

8 (13.6)

3 (8.1)

History of CAD n (%)

20 (33.9)

12 (32.4)

History of CVD n (%)

5 (8.5)

2 (5.4)

History of PTX n (%)

6 (10.2)

3 (8.1)

Time in HD (months)

35.3 (15.9–77.4)

44.6 (21.1–79.5)

Hematocrit (%)

34 ± 5

34 ± 6

Hb (g/l)

112.5 ± 15.8

113.24 ± 18.56

Albumin (g/dl)

39 (36–41.25)

39 (37.15–41)

hsCRP (mg/l)

4.76 (1.73–9.37)

4.89 (1.5–9.82)

C-Peptide (pmol/l)

3006.06 ± 1429.3

3117.53 ± 1453.08

Magnesium (mmol/l)

1.02 ± 0.18

1.04 ± 0.18

Calcium (mmol/l)

2.36 ± 0.27

2.37 ± 0.28

Phosphate (mmol/l)

1.83 ± 0.54

1.85 ± 0.55

Intact PTH (ng/l)

102.5 (43.7–325.25)

125.5 (48.75–375)

Sclerostin (ng/ml)

1.44 (0.97–2)

1.4 (0.88–2.06)

ucMGP (nM)

2329.7 ± 746.5

2210.6 ± 708.7

OPG (pmol/l)

5.09 ± 2.67

5 ± 2.35

Fetuin-A (g/l)

0.45 ± 0.1

0.47 ± 0.1

uc/c Osteocalcin

2.25 (0.83–4.81)

2.25 (0.96–5.38)

Adiponectin (µg/ml)

9.5 (6.17–15.5)

10.05 (5.92–15.05)

YKL-40 (ng/ml)

265.26 ± 95.71

257.16 ± 95.5

Data are presented as mean (±SD), median (IQR), number of subjects (%)

BMI body mass index, CAD coronary artery disease, CVD cerebrovascular disease, PTX parathyroidectomy, HD Hemodialysis, Hb Hemoglobin, hsCRP high- sensitive C-reactive protein, PTH parathyroid hormone, ucMGP uncarboxylated matrix-Gla protein, OPG Osteoprotegerin, uc/c osteocalcin ratio uncarboxylated/carboxylated osteocalcin

Fifty-four healthy subjects without kidney failure (28 % males, mean age 58 ± 10 years) from the internal medicine outpatient department at the University Hospital of Aachen, Germany were assigned as controls. Those patients were free of overt cardiovascular disease and selected based on estimated glomerular filtration rate (eGFR) >60 ml/min/1.73 m2. Survival analysis of patients as assessed in June 2014 was performed according to data bank query in the corresponding dialysis centers with a median follow-up of 88 (IQR 52–105) months. Mortality was assessed as all-cause mortality.

The study was approved by the ethical committee of the RWTH Aachen University Hospital.

MSCT imaging procedure

All MSCT examinations were performed on a 64-slice Dual Source CT scanner (SOMATOM Definition, Siemens, Forchheim, Germany). Scan parameters included a collimation of 2 × 32 × 0.6 mm, a rotation time of 330 ms, and a tube voltage of 120 kV. For electrocardiogram (ECG)-synchronization, prospective ECG triggering was applied. Axial images were reconstructed using 70 % of the RR interval with an effective slice thickness of 3 mm and a reconstruction increment of 1.5 mm. A dedicated convolution kernel (B35f), a field of view of 180 × 180 mm2 and a matrix of 512 × 512 were applied. CT scans were started at the level of the pulmonary artery to the level of the diaphragm to include.

Measurement of coronary and aortic valve calcification

CT scans were transferred to an offline workstation (Syngo CaScoring, Wizard, Siemens) to quantify coronary (CAC) and aortic valve calcification (AVC). Both, coronary and aortic valve calcification measurement was performed using dedicated software (Syngo MMWP, VA13A, Siemens). CAC was defined as an area of more than two connected voxels with an attenuation of more than 130 Hounsfield-Units (HU). CAC was identified for the coronaries by calculating the lesion volume in mm3, the calcium mass in mg and the resulting calcium score according to the method described by Agatston et al. [15]. In conformity with CAC-measurement, AVC was considered present with the existence of an area of more than two connected voxels with an attenuation of more than 130 HU. Similar to CAC measurement, the modified Agatston-score as well as Volume and mass Score were computed for the aortic valve.

Measurement of epicardial adipose tissue

Epicardial adipose tissue (EAT) was defined as fat located between the heart and pericardium, enclosed by the visceral pericardium [16]. The epicardial contours were defined as follows: most cranial slice at height of pulmonary trunk bifurcation, most caudal slice just below the posterior descending artery. EAT was measured by freehand, layer-by-layer region of interest (ROI)-based analysis and defined as an area of more than two connected voxels with an attenuation in between −200 and −40 HU. Results were expressed as volume (cm3) for EAT.

Longitudinal follow-up

According to the longitudinal development of vascular calcification and EAT, we computed the absolute rate of change as the difference between the values at follow-up and baseline. Percentage change was calculated as absolute rate divided by the value at baseline, multiplied by 100. Yearly change was presented as absolute rate divided by time between measurements in years. We defined a subgroup with EAT augmentation (progression) as those patients with EAT at baseline <EAT at follow-up, and with EAT decline (regression) as those patients with EAT at baseline >EAT at follow-up. For intergroup comparisons between those with EAT increase versus those with EAT decrease we compared the percentage EAT development in the lowest versus the highest tertile and analyzed the intergroup differences without patients from the intermediate tertile.

Biochemical analysis

Patient serum was collected prior to dialysis, after an overnight fast, after the long dialysis interval according to standard procedure. Material was immediately frozen at −20 °C after centrifuge and transferred to the University Hospital Aachen biobank for storage at -80 °C. The serum parameters creatinine (sCr), urea (sUrea), phosphate (sPO4), calcium (sCa), hemoglobin (Hb), hematocrit (Hct), high sensitive C-reactive protein (hsCRP), magnesium (Mg), and albumin (Alb) were all measured via standard laboratory methods on a daily basis.

The following research parameters were all measured centrally after the end of the study: uncarboxylated matrix-Gla protein (ucMGP), fetuin-A, osteoprotegerin (OPG), the ratio of uncarboxylated/carboxylated osteocalcin (uc/c OC), sclerostin, adiponectin, YKL-40 and intact parathyroid hormone (iPTH).

UcMGP was measured using a competitive enzyme-linked immunosorbent assay (ELISA) as previously described by our group [17]. Commercially available immunoassays were used to determine levels of the following parameters: sclerostin (TECO® Sclerostin EIA Kit (TECOmedical AG, Sissach, Switzerland); fetuin-A (Epitope Diagnostics USA, EIA Kit); osteoprotegerin, OPG (MicroVue Quidel, USA, EIA Kit); carboxylated osteocalcin, cOC (TaKaRa, Japan, EIA Kit); uncarboxylated osteocalcin, ucOC (TaKaRa, Japan, EIA Kit); iPTH (MicroVue, USA, EIA Kit); adiponectin (TECO Adiponektin EIA Kit; TECOmedical AG, Sissach, Switzerland); and YKL-40 (MicroVue YKL-40 EIA by Quidel, USA).

Statistical analysis

Results are expressed as number (%) for categorical variables, as mean ± SD for normally distributed continuous variables and median (interquartile range [IQR]) for continuous variables with skewed distribution. The normal distribution of all variables was tested with the Kolmogorov–Smirnov (KS) test. Categorical variables were compared between groups with the Chi squared test. For differences between groups, Student’s t test or the Mann–Whitney U test was performed for normally and non-normally distributed variables, respectively. For differences between periods, the paired Student’s t test or Wilcoxon test for paired values was used. All p values were two-tailed and a p value of <0.05 was considered significant.

Associations between continuous variables were explored with Pearson correlation if parametric or Spearman correlation if non parametric. All variables significantly associated with EAT volume (p < 0.1) on univariate analysis were entered in a multivariate linear regression model with backward selection (removal alpha level <0.05).

Left truncated analysis was performed using time since ESRD as the time scale. Survival rates for patients with EAT volume above or below the median were estimated by the Kaplan–Meier method. Cox proportional hazards (PH) models were further fitted to assess the association of EAT at baseline (as a continuous variable and categorized at median EAT volume) with all-cause death. Multivariable models were built using the following variables: age, serum PO4 and CAC.

The statistical analysis was performed with SPSS 21 (IBM Corporation, USA) and survival analysis with R (R Core Team, 2013. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL


Baseline results

EAT showed a mean level at baseline of 128 ± 61 cm3 and was significantly higher compared to the control group of non-renal patients without overt cardiovascular disease (94 ± 46 cm3; p < 0.05). Clinical, laboratory, and CT-derived EAT and calcification scores at baseline in the entire and the follow-up cohorts are listed in Tables 1 and 2. EAT at baseline was positively correlated with age, body mass index (BMI), CAC and AVC in the Spearman correlation analysis (Table 3). No significant correlations were observed with time on HD or with any laboratory parameter. There was no significant correlation of EAT volume at baseline with sPO4 (R = 0.149, p = 0.259). No significant differences regarding EAT were observed between males versus females, smokers versus non-smokers, and presence versus absence of diabetes. At multivariate analysis, age, BMI and AVC remained as significant predictors of EAT (p < 0.01). Patients with EAT volume at baseline above the median (cut-off 113 cm3) were older, exhibited higher serum phosphate levels and more coronary and aortic valve calcification than patients below the median (Table 4).
Table 2

MSCT measurement at baseline and at follow-up


Entire cohort baseline (n = 59)

Follow-up cohort baseline (n = 37)

Follow-up cohort at follow-up (n = 37)

p value*

Volume EAT (cm3)

128.21 ± 60.76

122.91 ± 59.26

132.69 ± 57.28


Agatston CAC

329.1 (22.6–1181.3)

266.5 (15.1–1206.2)

477 (32.5–1524.1)

0.000 b

Agatston AVC

0 (0–25.3)

0 (0–20)

0 (0–62.75)

0.002 b

Bold values indicate significant differences between groups (p value < 0.05)

Data are presented as mean (±SD), median (IQR)

MCT multi-slice computed tomography, EAT epicardial adipose tissue, CAC coronary artery calcification, AVC aortic valve calcification

* P value for comparison of parameters between baseline and follow-up in the follow-up cohort (n = 37)

aPaired t test was performed for comparison of continuous variables

bWilcoxon for paired values was performed for comparison of non-continuous variables

Table 3

Spearman correlation with epicardial adipose tissue (EAT) at baseline


Rho Spearman

p value







Agatston score CAC



Agatston score AVC



BMI, body mass index, CAC coronary artery calcification, AVC aortic valve calcification

Table 4

Comparison of patients with baseline epicardial adipose tissue (EAT) above or below the median in entire cohort (nN = 59)


EAT < 113 cm3 (n = 30)

EAT > 113 cm3 (n = 29)

p value

Age (years)

54.7 ± 15.8

63.7 ± 13

0.021 a

Women n (%)

16 (53.3)

13 (44.8)


BMI (kg/m2)

24.5 ± 6.6

26.5 ± 4.4


Diabetes n (%)

6 (20)

9 (31)


Hypertension n (%)

29 (96.7)

24 (82.8)


Current smoker n (%)

4 (13.3)

4 (13.8)


History of CAD n (%)

7 (23.3)

13 (44.8)


History of CVD n (%)

3 (10)

2 (6.9)


History of PTX n (%)

3 (10)

3 (10)


Time in HD (months)

50.1 (26.6–82.4)

32.7 (11–75)


Hematocrit (%)

34 ± 6

34 ± 5


Hb (g/l)

112.6 ± 18.17

112.41 ± 13.36


Albumin (g/dl)

39 (37.5–42.5)

39.5 (35–41)


hsCRP (mg/l)

4.75 (1.85–10.5)

4.79 (1.4–9.12)


C-Peptide (pmol/l)

2948.84 ± 1345.9

3183.28 ± 1526.71


Magnesium (mmol/l)

1.02 ± 0.17

1.03 ± 0.19


Calcium (mmol/l)

2.34 ± 0.22

2.37 ± 0.31


Phosphate (mmol/l)

1.64 ± 0.38

2.02 ± 0.62

0.007 a

Intact PTH (ng/l)

117.5 (56.25–352.5)

94.5 (39.25–294.75)


Sclerostin (ng/ml)

1.31 (0.69–2.03)

1.45 (1.14–1.97)


ucMGP (nM)

2326.89 ± 837.4

2332.92 ± 642.33


OPG (pmol/l)

4.8 ± 2.7

5.38 ± 2.67


Fetuin (g/l)

0.47 ± 0.11

0.44 ± 0.09


uc/c OC

2.03 (0.83–4.05)

2.27 (0.64–5.97)


Adiponectin (µg/ml)

9.3 (5.77–16.35)

9.5 (7.47–13.12)


YKL-40 (ng/ml)

255.26 ± 103.84

275.26 ± 87.58


Agatston CAC

95.8 (6.7–779.9)

514 (182.6–1227.1)

0.038 b

Agatston CAC > 0, n (%)

25 (83.3)

28 (96.6)


Agatston CAC > 100, n (%)

14 (46.7)

25 (86.2)

0.001 c

Agatson AVC

0 (0–1.75)

2.6 (0–144.8)

0.013 b

Agatson AVC > 0, n (%)

7 (23.3)

15 (51.7)

0.024 c

Bold values indicate significant differences between groups (p value < 0.05)

Data are presented as mean (±SD), median (IQR), number of subjects (%)

EAT epicardial adipose tissue, BMI body mass index, CAD coronary artery disease, CVD cerebrovascular disease, PTX parathyroidectomy, HD hemodyalisis, Hb hemoglobin, hsCRP high-sensitive C-reactive protein, PTH parathyroid hormone, ucMGP uncarboxylated matrix-Gla protein, OPG osteoprotegerin, uc/c osteocalcin ratio uncarboxylated/carboxylated osteocalcin, CAC coronary artery calcification, AVC aortic valve calcification

a t-Student’s t test for comparison of continuous variables with normal distribution

bU-Mann–Whitney U test for comparison of continuous variables with non-normal distribution

cχ2 test for comparison of categorical variables

Follow-up results

A second MSCT was performed in 37 HD patients after a mean follow-up of 23.9 ± 4.7 months. During this follow-up time mean EAT levels remained stable with a slight non-significant increase, whereas CAC and AVC increased significantly (Table 2).

The absolute change in EAT was 9.9 cm3 (IQR −9.4 to 26.94 cm3), with a percentage EAT change of 11 % (IQR −10.1 to 24.17 %) and a yearly EAT change of 5.5 cm3 (IQR −4.4 to 15.2 cm3). The absolute Agatston change was 111 (IQR 2.6 to 385) for CAC and 0 (IQR 0–16.5) for AVC, with a yearly Agatston change of 64.5 (IQR 1.2–186.6) and 0 (IQR 0–8.1), respectively. We compared the association between baseline CAC, AVC and EAT with the corresponding absolute change of the same parameter over time. We observed that both CAC and AVC revealed a significant, positive correlation between baseline Agatston score and its absolute change (Rho = 0.708, p < 0.001 and Rho = 0.328, p = 0.047; respectively) (Fig. 1 depicting this association for CAC). In contrast, a negative correlation was present between baseline EAT and absolute EAT change (R = −0.327, p = 0.048) (Fig. 2). Similar results were observed comparing baseline CAC, AVC and EAT with the corresponding yearly change of the same parameter over time. We observed that CAC revealed a significant, positive correlation between baseline Agatston score and its yearly change (Rho = 0.7, p < 0.001), almost significant regarding AVC and its yearly change (Rho = 0.313, p = 0.06). In contrast, a negative correlation was present between baseline EAT and yearly EAT change (R = −0.318, p = 0.055).
Fig. 1

Correlation between baseline CAC and absolute CAC change (Agatston). CAC, coronary artery calcification

Fig. 2

Correlation between baseline EAT volume with absolute EAT change. EAT, epicardial adipose tissue

Table 5 depicts clinical, laboratory and CT data after stratifying patients into those with progression versus those with regression of EAT over time. Those with an increase of EAT over time revealed significantly lower levels of EAT at baseline (88 ± 40 cm3 versus 146 ± 61 cm3, p = 0.01).
Table 5

Comparison patients with progression and regression of epicardial adipose tissue (EAT)


Regression (n = 13, 35 %)

Progression (n = 13, 35 %)

p value

Age (years)

60 ± 11.6

56 ± 14.3


Women n (%)

9 (69.2)

6 (46.2)


BMI (kg/m2)

26.1 ± 4.9

24.4 ± 5.4


Delta BMI (%)

−2.38 ± 9.15

2.78 ± 5.7


Time in HD (months)

42.7 (12.1–77.3)

61.2 (31.5–77.8)


Diabetes n (%)

5 (38.5)

1 (7.7)


Hypertension n (%)

12 (92.3)

12 (92.3)


Current smoker n (%)

1 (7.7)

2 (15.4 %)


History of CAD n (%)

5 (38.5)

3 (23.1)


History of CVD n (%)

0 (0)

2 (15.4)


Previous KT n (%)

4 (30.8)

3 (23.1)


Agatston CAC at baseline

214.7 (15.8–668.45)

512.4 (22.9–1247.7)


Agatston CAC at follow-up

346.7 (25.5–1174.9)

673.5 (83.6–1707.8)


Yearly change CAC

45.96 (0.1–110.39)

116.39 (20.8–244.25)


Agatston AVC at baseline

4 (0–90.55)

0 (0–12.5)


Agatston AVC at follow-up

0 (0–99.15)

9 (0–79.15)


Yearly change AVC

0 (0–7.63)

0.5 (0–19.84)


Volume EAT at baseline (cm3)

145.6 ± 61.24

87.72 ± 40.46

0.01 a

Volume EAT at follow-up (cm3)

124.12 ± 52.73

124.27 ± 52.21


Bold value indicates significant differences between groups (p value < 0.05)

Regression is defined as percentage change of EAT below the lowest tertile (= −1.39 %). Progression is defined as percentage change of EAT above highest tertile (20.5 %). Percentage change was calculated as the difference between the values at follow-up and baseline, divided by the value at baseline, multiplied by 100

BMI body mass index, HD hemodialysis, CAD coronary artery disease, CVD cerebrovascular disease, KT kidney transplant, CAC coronary artery calcification, AVC aortic valve calcification, EAT epicardial adipose tissue

Data are presented as mean (±SD), median (IQR), number of subjects (%)

at-Student’s t-test for comparison of continuous variables with normal distribution

bU-Mann–Whitney U-test for comparison of continuous variables with non-normal distribution

cχ2 test for comparison of categorical variables


The overall median follow-up time was 88 (IQR 52−105) months. Thirty-two patients died (all-cause mortality). At univariate analysis EAT was not associated with all-cause mortality: Cox pH models revealed a hazard ratio (HR) of 1.04 (95 % confidence interval [CI] 0.98–1.10) per 10 cm3 increase of baseline EAT. EAT was significantly higher in patients who died during follow-up compared to survivors: 144.8 ± 58 versus 108.5 ± 59 cm3 (p = 0.021). The median EAT of those who died was 138.9 cm3. Five-year survival rate was 62 % (95 % CI: 39–98) and 80 % (95 % CI: 60–100) in patients with EAT volume above and below the median, respectively. Patients with EAT above the median showed a lower survival rate, although not significant (p = 0.16).


The present longitudinal study is the first to investigate the evolution of EAT over time in long-term HD patients and to describe the relationship between EAT volume and progression of vascular calcification. We also analyzed for the first time the potential ability of EAT to independently predict mortality risk in long-term HD patients.

The major findings of our study are that, firstly, EAT was significantly associated with the severity of both CAC and AVC scores in a cross-sectional analysis. Secondly, EAT had no correlation with chronic kidney disease–mineral and bone disorder (CKD-MBD) biomarkers indicative of disturbances of the bone-vascular axis. Thirdly, in contrast to calcification scores which showed a marked progression in the vast majority of ESRD patients, EAT did not show such a homogeneous increase over time. Lastly, we could not detect a predictive power of EAT on mortality, a finding which may be related to the limited number of patients in the present study.

Our data confirm, however, previous reports, that found EAT volume to be higher in ESRD than in healthy subjects [11, 18] indicating that visceral fat metabolism might be disturbed in these patients. Our results are also consistent with previous studies which describe a positive correlation of EAT volume in ESRD with age, BMI and the amount of vascular calcification [10, 11]. We observed that sPO4 levels were higher in patients in the highest EAT tertile (Table 4), but this potential influence of sPO4 on EAT was not confirmed at the correlation analysis. Previous cross-sectional studies also showed that EAT was associated with arterial stiffness [10], left ventricular hypertrophy [19] and with the presence of malnutrition, inflammation and atherosclerosis/calcification (MIAC) in ESRD patients [11]. These studies fuel speculations whether large amounts of EAT represent an additional cardiovascular risk factor in ESRD patients.

The longitudinal development of EAT has never been evaluated in ESRD patients so far. Two previous studies in non-renal patients described regression of EAT, as measured by echocardiography, following significant weight loss after bariatric surgery [20] or after significantly reduced calorie intake [21]. Another study observed an effect of weight fluctuations upon EAT in subjects not enrolled in an intensive weight-loss program [22]. Statin therapy in postmenopausal women was associated with regression of EAT—with the limitation that this study lacked of a placebo group and assuming that the natural evolution of EAT is to increase because of its positive correlation with age [23].

According to the present study EAT remained stable over 2 years in hemodialysis patients. There was only a modest, insignificant longitudinal increase. We detected a substantial subgroup of hemodialysis patients in whom EAT volume decreased over time. These patients showed a non-significant trend to lose weight and had significantly higher baseline EAT levels. This is remarkably different to the exponential progression of CAC and AVC present in ESRD patients [13, 24], in whom baseline calcification levels strongly predict the amount of calcification increase.

An independent prognostic power of EAT to predict cardiovascular events has been demonstrated in the general population [7]. Moreover, measuring EAT increases the power of CAC scoring in healthy subjects without proven cardiovascular disease to predict coronary events [8]. ESRD patients, however, are remarkably different from the general population because of their large burden of cardiovascular risk factors and their overwhelming cardiovascular mortality. Thus, extrapolating cardiovascular risk assessment by certain biomarkers or factors from the general population to ESRD cohorts needs to be done with caution.

A previous post hoc analysis from a well-designed trial about evolution of vascular calcification in hemodialysis patients (RIND) introduced recently EAT volume as a potential independent risk factor for mortality in ESRD patients: The authors analyzed a cohort of 95 incident dialysis patients with a mean age of 55 ± 15 years and a mean CAC score of 96 ± 0.9 [12]. In that study, at univariate survival analyses, each 10 cc increase in EAT volume was associated with a significant 6 % increase in the risk of death during follow-up (HR: 1.060; 95 % CI: 1.013–1.109; p = 0.012). We did not observe such an independent influence of EAT upon mortality in the present cohort, although we found a similar but not significant HR of 1.04. The main reason for this discrepancy could be the lower number of patients in our cohort compared to the previous larger study. Other reasons may be the different dialysis vintage (incident versus long-term dialysis patients), which in turn is associated with very different levels of coronary artery calcifications and potentially of other factors influencing survival. Indeed, the differences in time on dialysis (incident in RIND versus prevalent patients in the present study), prevalence of diabetes (59 % in RIND versus 25 % in the present cohort) as well the CAC score were clinically meaningful (CAC Agatson in RIND 96 versus 329 in the present study). We cannot assess the biological meaning of the numerical difference regarding the EAT volume in both studies (113 cm3 in RIND versus 128 cm3 in the present cohort), neither can we exclude technical or analytical reasons for this discrepancy.

We acknowledge several limitations of the present study, in particular the low number of patients, which does not allow firm conclusions to be drawn about the impact of EAT on survival. Results should be extrapolated only with caution to presumably healthy HD patients (being transplanted prior to the 2nd examination) and those particularly diseased (with a risk of death during follow-up).

Moreover, our data do not allow to analyze if EAT is associated with atherosclerotic or arteriosclerotic coronary artery disease since only the coronary calcification score was assessed. Furthermore, paracrine effects of EAT might not be limited to the vasculature but might also affect the myocardium directly. However, we cannot provide data about cardiac function since, for example, echocardiography parameters were not obtained systematically. This is a significant limitation taking into account that left-ventricular function is a considerable confounder for the interaction of EAT thickness and vascular disease [25]. Further studies should also focus on the importance of progression versus regression of EAT in terms of mortality, which cannot be analyzed based on the limited data available.

In summary, our study confirms that EAT is measurable in parallel to CAC and AVC by heart MSCT in hemodialysis patients. There is a strong correlation between EAT volume and the coronary as well as valvular calcification burden in these patients. In contrast to the calcification scores, the simple rule “the more at baseline, the higher the progression rate” is refuted by our EAT data. Acknowledging the limited number of patients included in the present study, we could not detect any predictive power of EAT levels for future mortality in our long-term hemodialysis patients. If EAT finally turns out to be a reliable risk prediction tool in only specific subgroups of ESRD such as those with low levels of background calcification needs to be assessed in further studies.



Data were presented in part as an abstract at the ASN Renal Week congress 2014 in Philadelphia. The work performed by Xoana Barros at University Hospital of the RWTH Aachen was possible thanks to the Long-Term ERA-EDTA Fellowship Exchange Programme.

Compliance with the ethical standards

Conflict of interest

The authors declare they have no competing interests.

Ethical approval

The study was approved by the ethical committee of the RWTH Aachen University Hospital.

Informed consent

Patients were included after written and informed consent.


  1. 1.
    Rennenberg RJ, Kessels AG, Schurgers LJ, van Engelshoven JM, de Leeuw PW, Kroon AA (2009) Vascular calcifications as a marker of increased cardiovascular risk: a meta-analysis. Vasc Health Risk Manag 5(1):185–197CrossRefPubMedPubMedCentralGoogle Scholar
  2. 2.
    Stefan G, Capusa C, Stancu S, Petrescu L, Nedelcu ED, Andreiana I, Mirescu G (2014) Abdominal aortic calcification and renal resistive index in patients with chronic kidney disease: is there a connection? J Nephrol 27(2):173–179CrossRefPubMedGoogle Scholar
  3. 3.
    Raggi P, Chertow GM, Torres PU, Csiky B, Naso A, Nossuli K, Moustafa M, Goodman WG, Lopez N, Downey G, Dehmel B, Floege J (2011) The ADVANCE study: a randomized study to evaluate the effects of cinacalcet plus low-dose vitamin D on vascular calcification in patients on hemodialysis. Nephrol Dial Transplant 26(4):1327–1339CrossRefPubMedGoogle Scholar
  4. 4.
    Marwan M, Achenbach S (2013) Quantification of epicardial fat by computed tomography: why, when and how? J Cardiovasc Comput Tomogr 7(1):3–10CrossRefPubMedGoogle Scholar
  5. 5.
    Mazurek T, Zhang L, Zalewski A, Mannion JD, Diehl JT, Arafat H, Sarov-Blat L, O’Brien S, Keiper EA, Johnson AG, Martin J, Goldstein BJ, Shi Y (2003) Human epicardial adipose tissue is a source of inflammatory mediators. Circulation 108(20):2460–2466CrossRefPubMedGoogle Scholar
  6. 6.
    Alexopoulos N, McLean DS, Janik M, Arepalli CD, Stillman AE, Raggi P (2010) Epicardial adipose tissue and coronary artery plaque characteristics. Atherosclerosis 210(1):150–154CrossRefPubMedGoogle Scholar
  7. 7.
    Mahabadi AA, Berg MH, Lehmann N, Kälsch H, Bauer M, Kara K, Dragano N, Moebus S, Jöckel K-H, Erbel R, Möhlenkamp S (2013) Association of epicardial fat with cardiovascular risk factors and incident myocardial infarction in the general population: the heinz nixdorf recall study. J Am Coll Cardiol 61(13):1388–1395CrossRefPubMedGoogle Scholar
  8. 8.
    Kunita E, Yamamoto H, Kitagawa T, Ohashi N, Oka T, Utsunomiya H, Urabe Y, Tsushima H, Awai K, Budoff MJ, Kihara Y (2014) Prognostic value of coronary artery calcium and epicardial adipose tissue assessed by non-contrast cardiac computed tomography. Atherosclerosis 233(2):447–453CrossRefPubMedGoogle Scholar
  9. 9.
    Hajsadeghi F, Nabavi V, Bhandari A, Choi A, Vincent H, Flores F, Budoff M, Ahmadi N (2014) Increased epicardial adipose tissue is associated with coronary artery disease and major adverse cardiovascular events. Atherosclerosis 237(2):486–489CrossRefPubMedGoogle Scholar
  10. 10.
    Turan MN, Gungor O, Asci G, Kircelli F, Acar T, Yaprak M, Ceylan N, Demirci MS, Bayraktaroglu S, Toz H, Ozkahya M, Ok E (2013) Epicardial adipose tissue volume and cardiovascular disease in hemodialysis patients. Atherosclerosis 226(1):129–133CrossRefPubMedGoogle Scholar
  11. 11.
    Turkmen K, Kayikcioglu H, Ozbek O, Solak Y, Kayrak M, Samur C, Anil M, Zeki Tonbul H (2011) The relationship between epicardial adipose tissue and malnutrition, inflammation, atherosclerosis/calcification syndrome in ESRD patients. Clin J Am Soc Nephrol 6(8):1920–1925CrossRefPubMedPubMedCentralGoogle Scholar
  12. 12.
    D’Marco LG, Bellasi A, Kim S, Chen Z, Block GA, Raggi P (2013) Epicardial adipose tissue predicts mortality in incident hemodialysis patients: a substudy of the Renagel in New Dialysis trial. Nephrol Dial Transplant 28(10):2586–2595CrossRefPubMedGoogle Scholar
  13. 13.
    Block GA, Spiegel DM, Ehrlich J, Mehta R, Lindbergh J, Dreisbach A, Raggi P (2005) Effects of sevelamer and calcium on coronary artery calcification in patients new to hemodialysis. Kidney Int 68(4):1815–1824CrossRefPubMedGoogle Scholar
  14. 14.
    Brandenburg VM, Kramann R, Koos R, Kruger T, Schurgers L, Muhlenbruch G, Hubner S, Gladziwa U, Drechsler C, Ketteler M (2013) Relationship between sclerostin and cardiovascular calcification in hemodialysis patients: a cross-sectional study. BMC Nephrol 14:219CrossRefPubMedPubMedCentralGoogle Scholar
  15. 15.
    Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr, Detrano R (1990) Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 15(4):827–832CrossRefPubMedGoogle Scholar
  16. 16.
    Noyes AM, Dua K, Devadoss R, Chhabra L (2014) Cardiac adipose tissue and its relationship to diabetes mellitus and cardiovascular disease. World J Diabetes 5(6):868–876CrossRefPubMedPubMedCentralGoogle Scholar
  17. 17.
    Cranenburg EC, Koos R, Schurgers LJ, Magdeleyns EJ, Schoonbrood TH, Landewe RB, Brandenburg VM, Bekers O, Vermeer C (2010) Characterisation and potential diagnostic value of circulating matrix Gla protein (MGP) species. Thromb Haemost 104(4):811–822CrossRefPubMedGoogle Scholar
  18. 18.
    Altun B, Tasolar H, Eren N, Binnetoglu E, Altun M, Temiz A, Gazi E, Barutcu A, Altunoren O, Colkesen Y, Uysal F (2014) Epicardial adipose tissue thickness in hemodialysis patients. Echocardiography 31(8):941–946CrossRefPubMedGoogle Scholar
  19. 19.
    Ulusal Okyay G, Okyay K, Polattas Solak E, Sahinarslan A, Pasaoglu O, Ayerden Ebinc F, Pasaoglu H, Boztepe Derici U, Sindel S, Arinsoy T (2015) Echocardiographic epicardial adipose tissue measurements provide information about cardiovascular risk in hemodialysis patients. Hemodial Int 19(3):452–462CrossRefPubMedGoogle Scholar
  20. 20.
    Willens HJ, Byers P, Chirinos JA, Labrador E, Hare JM, de Marchena E (2007) Effects of weight loss after bariatric surgery on epicardial fat measured using echocardiography. Am J Cardiol 99(9):1242–1245CrossRefPubMedGoogle Scholar
  21. 21.
    Iacobellis G, Singh N, Wharton S, Sharma AM (2008) Substantial changes in epicardial fat thickness after weight loss in severely obese subjects. Obesity (Silver Spring) 16(7):1693–1697CrossRefGoogle Scholar
  22. 22.
    Nakazato R, Rajani R, Cheng VY, Shmilovich H, Nakanishi R, Otaki Y, Gransar H, Slomka PJ, Hayes SW, Thomson LE, Friedman JD, Wong ND, Shaw LJ, Budoff M, Rozanski A, Berman DS, Dey D (2012) Weight change modulates epicardial fat burden: a 4-year serial study with non-contrast computed tomography. Atherosclerosis 220(1):139–144CrossRefPubMedGoogle Scholar
  23. 23.
    Alexopoulos N, Melek BH, Arepalli CD, Hartlage GR, Chen Z, Kim S, Stillman AE, Raggi P (2013) Effect of intensive versus moderate lipid-lowering therapy on epicardial adipose tissue in hyperlipidemic post-menopausal women: a substudy of the BELLES trial (Beyond Endorsed Lipid Lowering with EBT Scanning). J Am Coll Cardiol 61(19):1956–1961CrossRefPubMedGoogle Scholar
  24. 24.
    Moe SM, O’Neill KD, Reslerova M, Fineberg N, Persohn S, Meyer CA (2004) Natural history of vascular calcification in dialysis and transplant patients. Nephrol Dial Transplant 19(9):2387–2393CrossRefPubMedGoogle Scholar
  25. 25.
    Doesch C, Suselbeck T, Haghi D, Streitner F, Schoenberg SO, Borggrefe M, Papavassiliu T (2012) The relationship between the severity of coronary artery disease and epicardial adipose tissue depends on the left ventricular function. PLoS One 7(11):e48330CrossRefPubMedPubMedCentralGoogle Scholar

Copyright information

© Italian Society of Nephrology 2015

Authors and Affiliations

  • Xoana Barros
    • 1
    • 7
  • Timm Dirrichs
    • 2
  • Ralf Koos
    • 3
  • Sebastian Reinartz
    • 2
  • Nadine Kaesler
    • 1
  • Rafael Kramann
    • 1
  • Ulrich Gladziwa
    • 4
  • Markus Ketteler
    • 5
  • Jürgen Floege
    • 1
  • Nikolaus Marx
    • 6
  • José V. Torregrosa
    • 7
  • András Keszei
    • 8
  • Vincent M. Brandenburg
    • 6
  1. 1.Department of NephrologyUniversity Hospital of the RWTH AachenAachenGermany
  2. 2.Department of RadiologyUniversity Hospital of the RWTH AachenAachenGermany
  3. 3.Department of CardiologyStädtische Kliniken Mönchengladbach GmbH, Elisabeth-Krankenhaus RheydtRheydtGermany
  4. 4.Dialysis CenterKuratorium für Heimdialyse (KfH)WürselenGermany
  5. 5.Department of NephrologyKlinikum CoburgCoburgGermany
  6. 6.Department of CardiologyUniversity Hospital of the RWTH AachenAachenGermany
  7. 7.Department of NephrologyHospital ClinicBarcelonaSpain
  8. 8.Department of Medical InformaticsUniversity Hospital of the RWTH AachenAachenGermany

Personalised recommendations