Review Article
Herz 2016 · 41:421–427
DOI 10.1007/s00059-015-4387-z
Received: 8 September 2015
Revised: 22 September 2015
Accepted: 3 November 2015
Published online: 10 December 2015
© Springer Medizin Verlag 2015
Epicardial adipose tissue (EAT) is the
adipose tissue accumulated between the
myocardium and visceral pericardium,
which is a source of inflammatory media-
tors and quantifiable indicators for assess-
ing the cardiometabolic risk [1, 2]. Cellu-
lar and molecular inflammatory cascades
triggered in the EAT, or cytokines and
hormones secreted in the EAT, may be
the potential mechanisms underlying sev-
eral cardiovascular diseases [3, 4]. Earli-
er meta-analyses have been published re-
garding the association of EAT thickness
or volume with metabolic syndrome [5]
and coronary artery disease [6]. Recent-
ly, several studies have also shown that
the total-EAT volume may have a signif-
icant relationship with the increased risk
of atrial fibrillation (AF) [7–13]. More in-
terestingly, there is significant difference
in EAT volume between patients with
persistent AF (PeAF) and paroxysmal AF
(PAF) [14]. Whether EAT volume is tru-
ly different according to the various types
of AF is still unknown. Additionally, apart
from the total-EAT volume, the EAT vol-
ume surrounding the left atrium (LA-
EAT volume) may also have a strong rela-
tionship with an increased risk of AF [10,
11]. It is still not clear whether a reduction
in the LA-EAT volume is sufficient to de-
crease the risk of AF. More importantly,
several studies also showed that the EAT
volume was not strongly associated with
the incidence of AF [14, 15]. Therefore,
our current meta-analysis aimed to evalu-
W. Zhu and H. Zhang contributed equally to
this study.
W. Zhu1 · H. Zhang1 · L. Guo1 · K. Hong1,2
1 Cardiology Department, The Second Affiliated Hospital of Nanchang University, Jiangxi, China
2 Jiangxi Key Laboratory of Molecular Medicine, Jiangxi, China
Relationship between
epicardial adipose tissue
volume and atrial fibrillation
A systematic review and meta-analysis
ate the relationship between EAT volume
and AF comprehensively.
Methods
Eligibility criteria
The following criteria were used in select-
ing studies:
a) Study design: studies evaluating the
relationship between EAT volume
and AF
b) Study type: case-control study
c) Study participants: adult patients with
AF documented by electrocardiogra-
phy
d) Diagnostic tools: EAT volume mea-
sured by multidetector computed to-
mography (MDCT)
e) Study data: sample size, mean, and
standard deviation data available
The following exclusion criteria were ap-
plied: (a) certain publication types (e.g.,
reviews, letters, case reports, comments);
and (b) studies with insufficient data or
duplicate data.
Information sources and
search strategy
The data were systematically retrieved
form the Cochrane Library, PubMed,
Medline, EBSCO, and Embase databas-
es. The search strategy was designed and
guided by an experienced librarian. The
following keywords were used in our
search strategies to identify relevant lit-
eratures published in English: “epicardial
adipose tissue,” “epicardial fat,” “pericar-
dial adipose tissue,” “pericardial fat,” “left
atrium,” “atrial fibrillation,” and “multi-
detector computed tomography”. Elec-
tronic retrieval was performed for arti-
cles published from 1 January 2010 to 1
September 2015. Further manual retriev-
al was performed using reference lists, rel-
evant journals, and conference abstracts.
All studies that met the inclusion crite-
ria were considered eligible for this me-
ta-analysis.
Study selection and data collection
The titles and abstracts of studies retrieved
electronically and manually were screened
independently by two reviewers (W.Z. and
H.Z.). If the necessary information was
not apparent, we performed a comprehen-
sive review of the full text (W.Z. H.Z., and
L.G). In situations of discrepancies, issues
were resolved through discussion or su-
pervised by a third reviewer (H.K.). Rel-
evant data were extracted including the
basic data of the reference (e.g., first au-
thor, publication year, country, diagnostic
tool, type of study, mean age of patients),
number of individuals with and without
AF, and the corresponding EAT volume.
Risk of bias and consistency
test of individual studies
Two independent reviewers (W.Z. and
H.Z.) evaluated the risk of bias accord-
ing to the risk of publication bias assess-
ment tool developed by the Cochrane
Collaboration. We evaluated the consis-
Herz 5 · 2016  | 421
 Review Article

Data synthesis
Records identified through Additional records identified
Identification

database searching through other sources Comparison of EAT volume

(n=850) (n=7) between non-AF and AF subgroups
Total-EAT volume between non-AF and
AF subgroups. Six studies [7, 9, 11–13,
15] were included to clarify the relation-
Records after duplicates removed
(n=788) ship between the total-EAT volume and
AF. There was no significant heteroge-
Screening

neity in the global effect of the included

samples (I2 = 47 %), and therefore a fixed-
Records excluded based
Records screened
on titles/abstracts
effects model was used for this analysis.
(n=788) The overall effect of total-EAT volume
screening
(n=719) was significant between AF and non-
AF subgroups (24.23 ml, 95 %CI: 19.40–
29.06, p < 0.00001; .  Fig.  3). We per-
Eligibility

Full-text articles
formed a sensitivity analysis by exclud-
Full-text articles
assessed for eligibility excluded, with reasons ing each study one at a time, and dem-
(n=69) (n=59) onstrated these results above mentioned
were stable.
LA-EAT volume between non-AF and
AF subgroups. Similarly, three studies
were included to compare the relation-
Included

Studies included in
quantitative synthesis ship between LA-EAT volume and AF
(meta-analysis) [11, 12, 16]. There was obvious hetero-
(n=10) [7-16]
geneity in the global effect of the sam-
ples (I2 = 70 %). To explore the source of
Fig. 1 8 Flow chart of study selection heterogeneity, we performed a sensitiv-
ity analysis and found that the study of
Tsao et al. included a high proportion of
tency test of individual studies using Co- Results coronary artery disease cases [16]. The
chran’s Q test complemented by the I2 sta- EAT might secrete the proinflammatory
tistic. I2 ≤ 25 % indicated low heterogene- Study characteristics hormones and cytokines related to cor-
ity, 25 % < I2 ≤ 50 % indicated intermedi- onary artery disease itself, which poten-
ate heterogeneity, and I2 > 50 % indicated The process of the literature search is tially incurred certain biases for this anal-
high heterogeneity. If the I2 statistic was shown in . Fig. 1. We initially retrieved ysis. Thus, we excluded the study of Ts-
≤ 50 %, a Mantel–Haenszel continuous- 857 articles. Finally, ten articles [7–16] ao et al. and repeated the analysis. With
weighted fixed-effects model analysis was met all the inclusion criteria. The basic this adjustment, the pooled results were
chosen. Any disagreement was resolved characteristics of the included studies are not significantly changed, but the corre-
by consensus. presented in . Table 1. Of the included sponding I2 value of the consistency test
studies, six studies [7, 9, 11–13, 15] re- markedly dropped to 0 %. Therefore, we
Statistical analysis ported on the relationship between to- could reasonably conclude that the study
tal-EAT volume and AF, and three stud- of Tsao et al. was the main source of the
We performed all the statistical analyses ies [11, 12, 16] described the relationship high heterogeneity. The pooled statistics
using Review Manager (RevMan) version between LA-EAT volume and AF; fur- showed that the LA-EAT volume was
5.3 from the Cochrane Collaboration (The thermore, five studies [7, 8, 10, 11, 13, 14] significantly increased in patients with
Nordic Cochrane Center, Rigshospitalet, divided AF patients into PAF and PeAF AF compared with the controls with-
Denmark, http://ims.cochrane.org/rev- subgroups. We determined the presence out AF (16.35 ml, 95 % CI: 12.73–19.98,
man). For each trial, we calculated and of publication bias with funnel plots. The p < 0.00001; .  Fig. 3). Additionally, there
pooled the sample size, mean, and stan- results of the publication bias analysis are was a statistically significant difference
dard deviation data for a comparative shown in . Fig. 2. between the total-EAT and LA-EAT vol-
analysis of the EAT volume. Subgroup or ume subgroups (χ2 = 6.54, p = 0.01).
sensitivity analyses were performed where
appropriate. Statistical significance was set
at p ≤ 0.05.

422 |  Herz 5 · 2016


Comparison of EAT volume
between PAF and PeAF
Total-EAT volume between PAF and PeAF
subgroups. In five of the examined studies
involving the total-EAT volume [7, 8, 11,
13, 14], AF individuals were assigned to
two subgroups: PAF (n = 228) and PeAF
(n = 171). Owing to the I2 value of = 25 %,
a fixed-effects model analysis was used to
evaluate the mean differences. Analysis of
the overall effects on the total-EAT vol-
ume revealed a significant difference be-
tween patients with PAF and those with
PeAF (19.38  ml, 95 % CI: 11.45–27.31,
p < 0.0001). The pooled statistics showed
that the total-EAT volume of PeAF pa-
tients was much larger than that of PAF
patients(. Fig. 4). After performing a sen-
sitivity analysis, we found the pooled re-
sults were not significantly changed.
LA-EAT volume between PAF and
PeAF subgroups. Three studies were in-
cluded in the analysis of the relation-
ship between different types of AF and
LA-EAT volume. Because the I2 val-
ue for the individual studies was 0 %, a
fixed-effects model analysis was used in
this part. The pooled results showed that
there was a significant difference in the
LA-EAT volume between PeAF and PAF
patients (17.91 ml, 95 % CI: 15.13–20.69,
p < 0.00001; . Fig. 4). The sensitivity anal-
ysis demonstrated that the pooled results
were not significantly changed. Addition-
ally, there was no statistically significant
difference between the total-EAT and
LA-EAT volume subgroups (χ2 = 0.12,
p = 0.70).
Discussion
AF is the most common cardiac arrhyth-
mia in clinical practice, and its prevalence
has been predicted to increase significant-
ly in the future. Although AF is general-
ly not immediately life-threatening and
several risk scoring systems have demon-
strated a modest predictive ability for car-
diovascular events [17], AF remains a sig-
nificant cause of high disability and mor-
tality. In recent clinical studies, the EAT
volume was shown to be associated with
AF and future AF events [15, 18]. Wong
and coworkers considered that the EAT
volume was associated with the pres-
ence and severity of AF, or with a poor
Abstract · Zusammenfassung
Herz 2016 · 41:421–427  DOI 10.1007/s00059-015-4387-z
© Springer Medizin Verlag 2015
W. Zhu · H. Zhang · L. Guo · K. Hong
Relationship between epicardial adipose tissue volume and
atrial fibrillation. A systematic review and meta-analysis
Abstract
Background.  Several studies have suggest-
ed that epicardial adipose tissue (EAT) vol-
ume may be associated with the risk of atri-
al fibrillation (AF). However, these studies
have reported conflicting results. We there-
fore aimed to investigate the relationship be-
tween EAT volume and AF.
Methods.  We systematically retrieved the
relevant studies reporting on the relationship
between EAT volume and AF using the Co-
chrane Library, PubMed, Medline, EBSCO, and
Embase databases. Data were extracted from
applicable articles, and mean differences
were pooled using the RevMan 5.3 software.
Results.  Ten case-control studies were iden-
tified. With regard to the relationship be-
tween EAT volume and AF, both total-EAT
volume (24.23 ml, 95 % CI: 19.40–29.06,
p < 0.00001) and EAT volume surrounding the
left atrium (LA-EAT; 16.35 ml, 95 %CI: 12.73–
19.98, p < 0.00001) were significantly in-
Zusammenhang zwischen epikardialem Fettgewebsvolumen
und Vorhofflimmern. Systematische Übersicht und Metaanalyse
Zusammenfassung
Hintergrund.  Verschiedenen Studien zu-
folge ist das epikardiale Fettgewebsvolumen
(„epicardial adipose tissue“, EAT) möglicher-
weise mit dem Risiko des Auftretens von
Vorhofflimmern (VF) assoziiert. Allerdings
werden in diesen Studien widersprüchliche
Ergebnisse beschrieben. Daher war es Ziel
der Autoren, den Zusammenhang zwischen
dem EAT-Volumen und VF zu untersuchen.
Methoden.  Die relevanten Studien zum Zu-
sammenhang zwischen EAT-Volumen und VF
wurden von den Autoren anhand der Daten-
banken der Cochrane Library, von PubMed,
Medline, EBSCO und Embase erfasst. Aus ge-
eigneten Artikeln wurden Daten extrahiert
und die Werte für die mittlere Differenz
mittels der Software RevMan 5.3 gepoolt.
Ergebnisse.  Es fanden sich 10 Fall-Kontroll-
Studien. Im Hinblick auf den Zusammen-
hang zwischen EAT-Volumen und VF waren
sowohl das Gesamt-EAT-Volumen (24,23 ml;
95 %-Konfidenzintervall, 95%-KI: 19,40–
29,06; p < 0,00001) als auch das EAT-Volumen
um den linken Vorhof herum (LA-EAT-
Volumen; 16,35 ml; 95 %-KI: 12,73–19,98;
creased in patients with AF. With regard to
the relationship between the different types
of AF and EAT volume, there was a significant
difference in the total-EAT volume subgroup
(19.38 ml, 95 % CI: 11.45–27.31, p < 0.0001)
and in the LA-EAT volume subgroup
(17.91 ml, 95 % CI: 15.13–20.69, p < 0.00001)
between patients with persistent AF (PeAF)
and paroxysmal AF (PAF). However, there was
no significant difference between the total-
EAT and LA-EAT volume subgroups (χ2 = 0.12,
p = 0.70).
Conclusion.  EAT volume may be associated
with an increased risk of AF. Additionally, the
EAT volume in patients with PeAF was larger
than that in PAF patients, independent of the
location of EAT.
Keywords
Atrial fibrillation · Epicardial adipose tissue ·
Persistent · Paroxysmal · Meta-analysis
p < 0,00001) bei Patienten mit VF signi-
fikant erhöht. Hinsichtlich des Zusammen-
hangs zwischen den verschiedenen Arten
von AF und EAT-Volumen bestand ein signi-
fikanter Unterschied in der Gesamt-EAT-
Volumen-Untergruppe (19,38 ml; 95 %-KI:
11,45–27,31; p < 0,0001) oder in der LA-
EAT-Volumen-Untergruppe (17,91 ml;
95 %-KI:15,13–20,69; p < 0,00001) zwischen
Patienten mit persistierendem VF (PeAF) und
mit paroxysmalem VF (PAF). Es gab jedoch
keinen signifikanten Unterschied zwischen
den Gesamt-EAT- und den LA-EAT-Volumen-
Untergruppen (χ2 = 0,12; p = 0,70).
Schlussfolgerung.  Das EAT-Volumen ist
möglicherweise mit einem erhöhten Risiko
für VF assoziiert. Außerdem war das EAT-
Volumen bei PeAF-Patienten größer als
bei PAF-Patienten, unabhängig von der
Lokalisierung des EAT.
Schlüsselwörter
Vorhofflimmern · Epikardiales Fettgewebe ·
Persistierend · Paroxysmal · Metaanalyse
Herz 5 · 2016  | 423
 Review Article
SE(MD)
0 0 SE(MD)
4 10
8 20
12
16
20
-100 -50 0 50
Subgroups
Total-EAT LA-EAT
a b
outcome after AF radiofrequency abla-
tion [19]. EAT may be an important bio-
marker of metabolic syndrome or coro-
nary artery disease [5, 6, 20], and our cur-
rent study aimed to evaluate the associa-
tion between EAT volume and AF.
In the present study, we found that in-
creased total-EAT volume might be asso-
ciated with AF. In addition, when we fo-
cused on the LA-EAT volume, it was sig-
nificantly increased in patients with AF
compared with individuals without AF.
Therefore, an increased EAT volume
was significantly associated with the oc-
currence of AF. Several previously pub-
lished reports also indicated that there
was a close relationship between to-
tal-EAT volume and AF, and that EAT
might promote the pathogenic process
of AF [21, 22]. These results are also in
agreement with those reported by sup-
ported by Wong et al. [19], although EAT
volume was measured by magnetic reso-
nance imaging in this study. Additionally,
there was a statistically significant differ-
ence between the total-EAT and LA-EAT
volume subgroups (χ2 = 6.54, p = 0.01), in-
dicating that LA-EAT volume might be
sufficient to measure the risk of AF, and
the reduction of LA-EAT volume might
be sufficient to decrease the risk of AF in
clinical practice. However, more research
is still needed to explain the association
between EAT volume and AF.
Additionally, with regard to the rela-
tionship between the different types of
AF and EAT volume, there was a signif-
icant difference in the total-EAT volume
subgroup and in the LA-EAT volume
subgroup between PeAF and PAF pa-
424 |  Herz 5 · 2016
30
40
MD
50
100 -100 -50 0 50
Subgroups
Total-EAT LA-EAT
tients. Noteworthy, there were statistical-
ly significant differences between the to-
tal-EAT and LA-EAT volume subgroups
(χ2 = 0.12, p = 0.70). Therefore, we tempo-
rarily could not assess whether total-EAT
or LA-EAT volume was superior in dis-
tinguishing the risk of different types of
AF. Some studies reasonably supported
our finding that increased EAT volumes
were noted in PeAF patients rather than
PAF patients, and EAT might be involved
in the maintenance of AF [10, 19, 22].
Patients with risk factors are more in-
clined to suffer from AF. Of the studies
included here, six [7, 9, 11–13, 15] pre-
formed multiple logistic regression anal-
ysis to assess the relationship between
EAT and AF after correction for the base-
line variables,. EAT was associated with
AF and this relationship remained signif-
icant after adjusting for AF risk factors.
For example, Al et al. [7] examined the
relationship between risk factors (includ-
ing EAT) and AF, and found EAT vol-
ume was highly associated with AF com-
pletely independent of the presence of
other risk factors. EAT was also associ-
ated with both PeAF and PAF, indepen-
dent of all other risk factors. Although
four studies included in our meta-anal-
ysis [8, 10, 14, 16] did not adjust for AF
risk factors, the sensitivity analysis indi-
cated the pooled results were not signif-
icantly changed. For example, we found
that the study of Tsao et al. included a
high proportion of coronary artery dis-
ease cases [16]. The EAT itself may se-
crete pro-inflammatory hormones and
cytokines related to coronary artery dis-
ease, which potentially incurred certain
Fig. 2 9 Funnel plot of
studies included in bias
analysis with regard to
a the relationship between
EAT volume and AF, b the
relationship between EAT
MD
volume and the types of
100 AF. EAT epicardial adipose
tissue, AF atrial fibrillation,
LA left atrium, SE standard
error, MD mean difference
biases. We therefore excluded this study
and repeated the analysis. With this ad-
justment, the pooled results were not sig-
nificantly changed. Thus, we could rea-
sonably conclude that there was a close
relationship between EAT and AF, inde-
pendent of traditional risk factors.
Moreover, recent studies have regard-
ed EAT volume as a risk factor for the re-
currence of AF after radiofrequency ab-
lation [19, 21, 23]. In these studies, we
found that increased EAT volume was
significantly associated with recurrence of
AF after radiofrequency ablation. More-
over, EAT volume could better predict re-
currence of PeAF after ablation than PAF
could, and it was independent of old-
er age, greater left atrial dimension, and
higher body mass index (BMI) [23]. Ow-
ing to the importance of predicting re-
currence of AF after radiofrequency ab-
lation, EAT volume could become an im-
portant clinical indicator for AF therapy
in the future.
It is worth noting that several stud-
ies describing the association between
pericardial adipose tissue (PAT) and AF
were included in our meta-analysis [7–
9]. EAT is the adipose tissue accumulat-
ed between the myocardium and viscer-
al pericardium, whereas PAT is the adi-
pose tissue accumulated between the vis-
ceral and parietal pericardium, or locat-
ed on the external surface of the parietal
pericardium. However, although the dis-
tinction between EAT and PAT is clear
in both functional and anatomic assess-
ments (e.g., embryological origins, vas-
cular supplies, and potential pathogenic
mechanisms), it is far from evident when
 AF Non-AF Mean Difference Mean Difference
Study or Subgroup Mean SD Total Mean SD Total Weight IV. Fixed. 95% CI IV. Fixed. 95% CI
Total-EAT volume
AI Chekakie (2010) [7] 101.6 44.1 197 76.1 36.3 76 8.0% 25.50 [15.28, 35.72]
Greif (2013) [9] 284.8 139.2 354 255.7 127.1 934 3.0% 29.10 [12.47, 45.73]
Mahabadi (2014) [15] 131 64.9 50 92.7 46.1 3,809 2.6% 38.30 [20.25, 56.35]
Nagashima (2011) [11] 185.6 76.1 40 138.3 45.2 37 1.1% 47.30 [19.58, 75.02]
Nakanishi (2012) [12] 80 22 17 53 20 262 7.3% 27.00 [16.27, 37.73]
Shin (2011) [13] 83.8 26.8 80 67.2 23.1 80 14.0% 16.60 [8.85, 24.35]
Subtotal (95% CI) 738 5,198 36.0% 24.23 [19.40, 29.06]
Heterogeneity: Chi2 = 9.36, df = 5 (P = 0.10); I2 = 47%
Test for overall effect: Z = 9.83 (P < 0.00001)

LA-EAT volume
Nagashima (2011) [11] 51.5 27.8 40 32.9 14.5 37 8.7% 18.60 [8.80, 28.40]
Nakanishi (2012) [12] 34 8 17 18 7 262 55.3% 16.00 [12.10, 19.90]
Subtotal (95% CI) 57 299 64.0% 16.35 [12.73, 19.98]
Heterogeneity: Chi2 = 0.23, df = 1(P = 0.63); I2 = 0%
Test for overall effect: Z = 8.85 (P < 0.00001)

Total (95% CI) 795 5,497 100.0% 19.19 [16.29, 22.09]

-100 -50 0 50 100

Non-AF AF

Fig. 3 8 Fixed-effects analysis of the relationship between EAT volume and AF. EAT epicardial adipose tissue, LA left atrium,
AF atrial fibrillation, CI confidence interval, SD standard deviation, IV inverse variance

PeAF PAF Mean Difference Mean Difference

Study or Subgroup Mean SD Total Mean SD Total Weight IV. Fixed. 95% CI IV. Fixed. 95% CI
Total-EAT volume
AI Chekakie (2010) [7] 115.4 49.3 71 93.9 39.1 126 3.9% 21.50 [8.15, 34.85]
Girerd (2013) [8] 122.8 46.4 24 105.7 41.7 25 1.1% 17.10 [-7.63, 41.83]
Nagashima (2011) [11] 226.4 93.3 16 158.3 47.2 24 0.3% 68.10 [18.64, 117.56]
Park (2014) [14] 163.3 71.3 20 125.3 53.1 13 0.4% 38.00 [-4.54, 80.54]
Shin (2011) [13] 91 26 40 76.6 26 40 5.3% 14.40 [3.01, 25.79]
Subtotal (95% CI) 171 228 11.0% 19.38 [11.45, 27.31]
Heterogeneity: Chi2 = 5.33, df = 4 (P = 0.26); I2 = 25%
Test for overall effect: Z = 4.79 (P < 0.00001)

LA-EAT volume
Girerd (2013) [8] 54.5 24.8 24 43.2 17.6 25 4.7% 11.30 [-0.78, 23.38]
Nagashima (2011) [11] 66.8 35.1 16 41.3 15.3 24 2.1% 25.50 [7.24, 43.76]
Nagashima (2012) [10] 52.9 4.4 18 34.8 4.2 16 82.3% 18.10 [15.21, 20.99]
Subtotal (95% CI) 58 65 89.0% 17.91 [15.13, 20.69]
Heterogeneity: Chi2 = 1.83, df = 2 (P = 0.40); I2= 0%
Test for overall effect: Z = 12.63 (P < 0.00001)

Total (95% CI) 229 293 100.0% 18.07 [15.45, 20.70]

-100 -50 0 50 100

PAF PeAF

Fig. 4 8 Fixed-effects analysis of the relationship between EAT volume with the types of AF. EAT epicardial adipose tissue,
LA left atrium, AF atrial fibrillation, PeAF persistent atrial fibrillation, PAF paroxysmal atrial fibrillation, CI confidence interval,
SD standard deviation, IV inverse variance

their thickness and volume can be quan-
tified by echocardiography, computed to-
mography, or magnetic resonance imag-
ing, respectively [21, 24]. More specifical-
ly, the term “pericardial fat” has been used
interchangeably to evaluate the relation- Implications for further
ship with AF [18], and three studies defin- clinical study
ing epicardial fat as pericardial fat are in-
cluded in the meta-analysis [7–9]. EAT may play an important role in the
development of AF. At present, whether

Herz 5 · 2016  | 425

 Review Article

Table 1  Basic characteristics of studies included in the meta-analysis

Study ID (author, year) Origin MDCT Case- Adjusted for risk factorsa Control sample size Atrial fibrillation
control age (years) sample size age
study (years)
Al Chekakie et al. (2010) [7] USA Yes Yes Age, sex, hypertension, valvular N = 76 55.9 ± 14.9 N = 197 58.1 ± 9.8
heart disease, LVEF, diabetes melli-
tus, left atrial enlargement and BMIa
Girerd et al. (2013) [8] USA Yes Yes No N = 23 59.3 ± 8.5 N = 49 59.3 ± 8.5
Greif et al. (2013) [9] Germany Yes Yes Age, hypertension, valvular disease, N = 934 62.5 ± 11.5 N = 354 62.5 ± 11.5
heart failure and BMIa
Nagashima et al. (2012) [10] Japan Yes Yes No N = 34 56.0 ± 4.5 N = 34 55.5 ± 7.5
Nagashima et al. (2011) [11] Japan Yes Yes Age, BMI, triglyceride levela N = 37 59.1 ± 11.1 N = 40 58.0 ± 10.2
Nakanishi et al. (2012) [12] Japan Yes Yes Age, sex, hypertension, diabetes N = 262 64.0 ± 10.0 N = 17 74.0 ± 4.0
mellitus, hypercholesterolemia, BMI,
serum C-reactive protein, LVMIa
Shin et al. (2011) [13] Korea Yes Yes Age, BMI, LVMI, LVEFa N = 80 51.0 ± 11.9 N = 80 52.2 ± 12.2
Park et al. (2014) [14] Korea Yes Yes No N = 33 56.9 ± 9.0 N = 33 57.0 ± 9.2
Mahabadi et al. (2014) [15] Germany Yes Yes Age, gender, BMI, systolic blood N = 3,809 58.8 ± 7.5 N = 50 64.5 ± 7.4
pressure, and antihypertensive
treatmenta
Tsao et al. (2011) [16] China Yes Yes No N = 34 54.1 ± 9.0 N = 68 54.7 ± 8.5
MDCT multidetector computed tomography, LVEF left ventricular ejection fraction, BMI body mass index, LVMI left ventricular mass index.
aEpicardial adipose tissue was associated with atrial fibrillation and this relationship remained significant after adjusting for atrial fibrillation risk factors.

EAT is a new biomarker and therapeu-
tic target of AF has been in the spotlight,
raising great concern [21, 25]. It provides
new avenues for us to further explore the
potential pathogenesis and therapy of AF.
Abed and coworkers [26] indicated that
AF may be a reversible disease, and for
patients with AF who are overweight and
obese it is necessary to first reduce their
weight. More interestingly, during the
process of weight loss in obese patients,
the reduction in EAT volume is most ob-
vious [27]. We therefore have reason to
deduce that the reduction in EAT volume
(especially the LA-EAT volume) could
be beneficial for AF. The role of EAT in
the pathogenesis of AF is just starting to
be explained, and these potential mech-
anisms should be evaluated more exten-
sively in the future.
Study limitations
The current study had several potential
limitations. Firstly, EAT was influenced
by age, sex, BMI, waist circumference, hy-
perlipidemia, and diabetes, but these fac-
tors could not be excluded in this meta-
analysis. Secondly, although the high het-
erogeneity of the individual studies was
not obvious, there were limited studies
of LA-EAT subgroups and not all ethnic
groups were represented in the selected
studies. Thirdly, there was a strong publi-
cation bias in this field, making the com-
pilation of positive reports much more
likely than negative reports. Finally, there
were also limited studies considering the
relationship between EAT and recurrence
of AF after radiofrequency ablation, al-
though EAT volume might be an effec-
tive way to predict the recurrence of AF
after radiofrequency ablation.
Conclusion
In summary, our meta-analysis indicates
that EAT volume could be associated
with AF. Interestingly, we also found that
EAT volume in patients with PeAF was
larger than that in patients with PAF, in-
dependent of the location of EAT.
Corresponding address
Dr. K. Hong MD, PhD
Jiangxi Key Laboratory of Molecular Medicine
330006 Jiangxi
hongkui88@163.com
Acknowledgements.  We gratefully thank Y-HQ
(School of Public Health, Nanchang University,
Jiangxi, 330006, China) for help us design the search
strategy and assisting in the consistency testing of
the included studies. The authors wish to acknowl-
edge support from the Ministry of Chinese Education
Innovation Team Development Plan [IRT1141, HK],
the National Basic Research Program of China [973
Program: 2013CB531103], and the National Natural
Science Foundation of China [81160023, 81370288].
Compliance with
ethical guidelines
Conflict of interest.  W. Zhu, H. Zhang, L. Guo, and
K. Hong state that there are no conflicts of interest.
References
  1. Parisi V, Rengo G, Pagano G, D’Esposito V, Passaret-
ti F, Caruso A, Grimaldi MG, Lonobile T, Baldasci-
no F, De Bellis A, Formisano P, Ferrara N, Leosco D
(2015) Epicardial adipose tissue has an increased
thickness and is a source of inflammatory medi-
ators in patients with calcific aortic stenosis. Int J
Cardiol 186:167–169
  2. Iacobellis G, Bianco AC (2011) Epicardial adipose
tissue: emerging physiological, pathophysiologi-
cal and clinical features. Trends Endocrinol Metab
22(11):450–457
  3. Talman AH, Psaltis PJ, Cameron JD, Meredith IT,
Seneviratne SK, Wong DT (2014) Epicardial adipose
tissue: far more than a fat depot. Cardiovasc Diagn
Ther 4(6):416–429
  4. Venteclef N, Guglielmi V, Balse E, Gaborit B, Co-
tillard A, Atassi F, Amour J, Leprince P, Dutour A,
Clement K, Hatem SN (2015) Human epicardial ad-
ipose tissue induces fibrosis of the atrial myocardi-
um through the secretion of adipo-fibrokines. Eur
Heart J 36(13):795–805

426 |  Herz 5 · 2016


  5. Pierdomenico SD, Pierdomenico AM, Cuccurullo F,
Iacobellis G (2013) Meta-analysis of the relation of
echocardiographic epicardial adipose tissue thick-
ness and the metabolic syndrome. Am J Cardiol
111(1):73–78
  6. Xu Y, Cheng X, Hong K, Huang C, Wan L (2012)
How to interpret epicardial adipose tissue as a
cause of coronary artery disease: a meta-analysis.
Coron Artery Dis 23(4):227–233
  7. Al CM, Welles CC, Metoyer R, Ibrahim A, Shap-
ira AR, Cytron J, Santucci P, Wilber DJ, Akar JG
(2010) Pericardial fat is independently associat-
ed with human atrial fibrillation. J Am Coll Cardiol
56(10):784–788
  8. Girerd N, Scridon A, Bessiere F, Chauveau S, Geloen
A, Boussel L, Morel E, Chevalier P (2013) Periatri-
al epicardial fat is associated with markers of en-
dothelial dysfunction in patients with atrial fibrilla-
tion. PLoS One 8(10):e77167
  9. Greif M, von Ziegler F, Wakili R, Tittus J, Becker C,
Helbig S, Laubender RP, Schwarz W, D’Anastasi M,
Schenzle J, Leber AW, Becker A (2013) Increased
pericardial adipose tissue is correlated with atrial
fibrillation and left atrial dilatation. Clin Res Cardiol
102(8):555–562
10. Nagashima K, Okumura Y, Watanabe I, Nakai T, Oh-
kubo K, Kofune M, Mano H, Sonoda K, Hiro T, Nikai-
do M, Hirayama A (2012) Does location of epicar-
dial adipose tissue correspond to endocardial high
dominant frequency or complex fractionated atrial
electrogram sites during atrial fibrillation? Circ Ar-
rhythm Electrophysiol 5(4):676–683
11. Nagashima K, Okumura Y, Watanabe I, Nakai T, Oh-
kubo K, Kofune T, Kofune M, Mano H, Sonoda K,
Hirayama A (2011) Association between epicardi-
al adipose tissue volumes on 3-dimensional recon-
structed CT images and recurrence of atrial fibril-
lation after catheter ablation. Circ J 75(11):2559–
2565
12. Nakanishi K, Fukuda S, Tanaka A, Otsuka K, Saka-
moto M, Taguchi H, Yoshikawa J, Shimada K, Yoshi-
yama M (2012) Peri-atrial epicardial adipose tissue
is associated with new-onset nonvalvular atrial fi-
brillation. Circ J 76(12):2748–2754
13. Shin SY, Yong HS, Lim HE, Na JO, Choi CU, Choi JI,
Kim SH, Kim JW, Kim EJ, Park SW, Rha SW, Park CG,
Seo HS, Oh DJ, Kim YH (2011) Total and interatri-
al epicardial adipose tissues are independently as-
sociated with left atrial remodeling in patients
with atrial fibrillation. J Cardiovasc Electrophysiol
22(6):647–655
14. Park J, Park CH, Lee HJ, Wi J, Uhm JS, Pak HN, Lee
M, Kim YJ, Joung B (2014) Left atrial wall thickness
rather than epicardial fat thickness is related to
complex fractionated atrial electrogram. Int J Car-
diol 172(3):e411–e413
15. Mahabadi AA, Lehmann N, Kalsch H, Bauer M,
Dykun I, Kara K, Moebus S, Jockel KH, Erbel R,
Mohlenkamp S (2014) Association of epicardial
adipose tissue and left atrial size on non-contrast
CT with atrial fibrillation: the Heinz Nixdorf Recall
Study. Eur Heart J Cardiovasc Imaging 15(8):863–
869
16. Tsao HM, Hu WC, Wu MH, Tai CT, Lin YJ, Chang SL,
Lo LW, Hu YF, Tuan TC, Wu TJ, Sheu MH, Chang CY,
Chen SA (2011) Quantitative analysis of quanti-
ty and distribution of epicardial adipose tissue sur-
rounding the left atrium in patients with atrial fi-
brillation and effect of recurrence after ablation.
Am J Cardiol 107(10):1498–1503
17. Zhu W, Xiong Q, Hong K (2015) Meta-Analysis of
CHADS(2) versus CHA(2)DS(2)-VASc. Tex Heart Inst
J 42(1):6–15
18. Thanassoulis G, Massaro JM, O’Donnell CJ, Hoff-
mann U, Levy D, Ellinor PT, Wang TJ, Schnabel RB,
Vasan RS, Fox CS, Benjamin EJ (2010) Pericardial fat
is associated with prevalent atrial fibrillation: the
Framingham Heart Study. Circ Arrhythm Electro-
physiol 3(4):345–350
19. Wong CX, Abed HS, Molaee P, Nelson AJ, Brooks
AG, Sharma G, Leong DP, Lau DH, Middeldorp ME,
Roberts-Thomson KC, Wittert GA, Abhayaratna WP,
Worthley SG, Sanders P (2011) Pericardial fat is as-
sociated with atrial fibrillation severity and abla-
tion outcome. J Am Coll Cardiol 57(17):1745–1751
20. de Feyter PJ (2011) Epicardial adipose tissue: an
emerging role for the development of coronary
atherosclerosis. Clin Cardiol 34(3):143–144
21. Hatem SN, Sanders P (2014) Epicardial adi-
pose tissue and atrial fibrillation. Cardiovasc Res
102(2):205–213
22. Kanazawa H, Yamabe H, Enomoto K, Koyama
J, Morihisa K, Hoshiyama T, Matsui K, Ogawa H
(2014) Importance of pericardial fat in the forma-
tion of complex fractionated atrial electrogram re-
gion in atrial fibrillation. Int J Cardiol 174(3):557–
564
23. Kim TH, Park J, Park JK, Uhm JS, Joung B, Lee MH,
Pak HN (2014) Pericardial fat volume is associat-
ed with clinical recurrence after catheter ablation
for persistent atrial fibrillation, but not paroxysmal
atrial fibrillation: an analysis of over 600-patients.
Int J Cardiol 176(3):841–846
24. Mahajan R, Kuklik P, Grover S, Brooks AG, Wong CX,
Sanders P, Selvanayagam JB (2013) Cardiovascular
magnetic resonance of total and atrial pericardi-
al adipose tissue: a validation study and develop-
ment of a 3 dimensional pericardial adipose tissue
model. J Cardiovasc Magn Reson 15:73
25. Fang M, Chen Y, Wang J (2014) Epicardial adipose
tissue: the accomplice implicated in the genesis
and maintenance of atrial fibrillation. Chin Med J
(Engl) 127(14):2700–2706
26. Abed HS, Wittert GA, Leong DP, Shirazi MG, Bah-
rami B, Middeldorp ME, Lorimer MF, Lau DH, Antic
NA, Brooks AG, Abhayaratna WP, Kalman JM, Sand-
ers P (2013) Effect of weight reduction and car-
diometabolic risk factor management on symp-
tom burden and severity in patients with atri-
al fibrillation: a randomized clinical trial. JAMA
310(19):2050–2060
27. Kim MK, Tanaka K, Kim MJ, Matuso T, Endo T, Tomi-
ta T, Maeda S, Ajisaka R (2009) Comparison of epi-
cardial, abdominal and regional fat compartments
in response to weight loss. Nutr Metab Cardiovasc
Dis 19(11):760–766
Fachnachrichten
Oszillierenden Signale
bestimmen das Schicksal
neuer Blutgefäße
Wenn neue Blutgefäße wachsen, müssen
sie sich entscheiden, ob sie neue Seitenäste
bilden oder ihren Durchmesser vergrößern.
Das Hormon VEGFA spielt eine Hauptrolle
beim Wachstum der Adern. Bei niedrigem
VEGFA-Spiegel schaltet es die Gefäßzellen
in den Verzweigungsmodus – dem Gefäß
wachsen neue Seitenäste. Ist es höher
konzentriert, lässt es die Gefäße an Durch-
messer zulegen. Der zu Grunde liegende
Mechanismus war bisher nicht bekannt.
Ein internationales Forscherteam hat nun
herausgefunden, dass Gefäßzellen sich
verbünden und sich gemeinsam be-
wegen können. Die Zellen kommunizieren
untereinander mit oszillierenden Signalen,
wie mithilfe von Computersimulationen
und Experimenten herausgefunden
wurde. Der VEGFA-Spiegel beeinflusst den
Notch-Signalweg, über den benachbarte
Gefäßzellen miteinander kommunizieren. In
der Signalkette werden bestimmte Proteine
in der Zelle periodisch hergestellt und gleich
wieder abgebaut, was zu einer oszillieren-
den Aktivität des Notch-Signalwegs in den
Gefäßzellen führt.
Der neu entdeckte Mechanismus ist für
die Therapie von Krankheiten relevant.
So konnte z. B. gezeigt werden, dass
dieser Vorgang für die Gefäßverdickung
in Krankheitsmodellen für diabetische
Retinopathie oder Krebs verantwortlich
ist. Die Forschungsergebnisse sind somit
für Therapien von Bedeutung, die Gefäße
wieder normalisieren oder ihr Wachstum
hemmen.
Literatur
Ubezio B, Blanco R, Geudens I et al
(2016) Synchronization of endothelial
Dll4-Notch dynamics switches blood
vessels from branching to expansion. eLife
doi:10.7554/eLife.12167
Quelle: Max-Delbrück-Centrum für
Molekulare Medizin (MDC) Berlin, mdc-
berlin.de
Herz 5 · 2016  | 427