+Model

CLINRE-736; No. of Pages 9 ARTICLE IN PRESS

Clinics and Research in Hepatology and Gastroenterology (2015) xxx, xxxxxx

Available online at

ScienceDirect
www.sciencedirect.com

ORIGINAL ARTICLE

Association between portal vein thrombosis

and risk of bleeding in liver cirrhosis:
A systematic review of the literature
Xingshun Qi a,b,, Chunping Su c, Weirong Ren b,d, Man Yang b,e,
Jia Jia b,f, Junna Dai a, Wenda Xu a, Xiaozhong Guo a,

a
Department of Gastroenterology, General Hospital of Shenyang Military Area, Shenyang 110840, China
b
Xijing Hospital of Digestive Diseases, Fourth Military Medical University, Xian 710032, China
c
Library of Fourth Military Medical University, Xian 710032, China
d
Department of Digestive Diseases, Sanmenxia Central Hospital, Henan University of Science and
Technology, Xiaoshan Road, Sanmenxia 472000, China
e
Department of Gastroenterology, Songgang Peoples Hospital, Shenzhen 518105, China
f
Department of Digestive Diseases, Shaanxi Provincial Peoples Hospital, Xian 710068, China

Summary
Aims: A systematic review of the literature was conducted to explore the association of portal
vein thrombosis (PVT) with the risk of bleeding in liver cirrhosis.
Methods: PubMed, EMBASE, and Cochrane library databases were searched for all relevant
papers, which compared the prevalence of bleeding at baseline and/or incidence of bleeding
during follow-up between cirrhotic patients with and without PVT.
Results: Eighteen papers were eligible for this systematic review. The heterogeneity among
studies was marked with regards to the treatment modalities, sources of bleeding, lengths
of follow-up, and ways of data expression. But most of their ndings were homozygous and
suggested that the cirrhotic patients with PVT were more likely to have previous histories of
bleeding at their admission and to develop de novo bleeding and/or rebleeding during the
short- and long-term follow-up. The association of PVT with the risk of bleeding might be
weakened in the multivariate analyses. Additionally, as for the cirrhotic patients with gastric
variceal bleeding treated with medical/endoscopic therapy, the association of PVT with the
risk of rebleeding remained controversial in 2 studies; as for the cirrhotic patients undergoing
transjugular intrahepatic portosystemic shunts for the management of variceal bleeding, a
pre-existing PVT was not associated with the risk of rebleeding.

Corresponding authors. Department of Gastroenterology, General Hospital of Shenyang Military Area, No. 83 Wenhua Road, Shenyang

110840 China. Tel.: +86 24 288 976 03; fax: +86 24 288 511 13.
E-mail addresses: xingshunqi@126.com (X. Qi), guo xiao zhong@126.com (X. Guo).

http://dx.doi.org/10.1016/j.clinre.2015.02.012
2210-7401/ 2015 Published by Elsevier Masson SAS.

Please cite this article in press as: Qi X, et al. Association between portal vein thrombosis and risk of
bleeding in liver cirrhosis: A systematic review of the literature. Clin Res Hepatol Gastroenterol (2015),
http://dx.doi.org/10.1016/j.clinre.2015.02.012
+Model
CLINRE-736; No. of Pages 9 ARTICLE IN PRESS
2 X. Qi et al.

Conclusions: Based on a systematic review of the literature, there was a positive association
between the presence of PVT and risk of bleeding in liver cirrhosis in most of clinical conditions.
However, whether PVT aggravated the development of bleeding during follow-up needed to be
further explored.
2015 Published by Elsevier Masson SAS.

Introduction number of observed patients, and number of patients with

Gastroesophageal varices can be found in approximately
50% of cirrhotic patients at the time of diagnosis [1,2].
The development of varices is primarily attributed to an
increased portal pressure caused by brosis and regener-
ative nodules (as the hepatic venous pressure gradient is
more than 10 mmHg, varcies will develop [3]). Once the
varices are ruptured, the mortality is up to 1520% within
6 weeks and as high as 40% within 1 year [4,5]. The most
common predictors for the rst occurrence or recurrence
of variceal bleeding include the diameter of varices, red
signs on endoscopy, and Child-Pugh score [6]. Recently, the
researchers also have cast more attention to the effect of
portal vein thrombosis (PVT) on the development of variceal
bleeding in liver cirrhosis [7,8], because it can further ele-
vate the resistance to portal inow. The important topic may
inuence the risk stratication of variceal bleeding, thereby
improving the algorithm for the management of variceal
bleeding in liver cirrhosis. Herein, we systematically review
the relevant literature to clarify the association between
PVT and risk of bleeding in liver cirrhosis.
PVT. Additionally, we collected the data regarding the pro-
portion of bleeding in cirrhotic patients with and without
PVT. If the original data were not reported, we collected
the odds or hazard ratios to express the difference in the
proportion of bleeding between the two groups. Data were
not synthesized, because they were expressed in different
ways.
Grade of evidence
The evidence was classied into high- and low-grade. The
evidence was of high-grade, if any one of the 2 following
points was met:
a multivariate analysis was performed to explore the sta-
tistically signicant difference;
if only a univariate analysis was performed, the base-
line Child-Pugh class or MELD score should be matched
between patients with and without PVT.
Otherwise, the evidence was of low-grade.

Methods
Results
Search strategy and study selection
Characteristics of studies
As previously described, we retrieved all papers regarding
Initially, 10,936 papers regarding PVT were identied.
PVT via the PubMed, EMBASE, and Cochrane library
Among them, 14 papers were eligible for this systematic
databases [9,10]. After this systematic search, more
review [1124] (Fig. 1). Another 4 eligible papers, which
recently published publications were also hand-searched.
were published after the systematic search, were also iden-
Among the clinical studies with more than 10 patients, we
tied by hand searching [2528]. Thus, 18 studies were
further identied the studies that evaluated the association
nally included. The characteristics of included studies were
between PVT and risk of bleeding in liver cirrhosis. Exclusion
summarized in Table 1. According to the regions, 5 studies
criteria were as follows:
were performed in China Taiwan, 5 studies in Italy, 3 stud-
ies in USA, 2 studies in France, 1 study in Canada, 1 study in
only malignancy was enrolled;
Switzerland, and 1 study in France and Belgium. According to
PVT developed after surgery, therapeutic endoscopy, or
the enrolment periods, 3 studies were launched before 1990,
interventional treatments;
5 studies between 1990 and 2000, and 10 studies after 2000.
PVT developed in non-cirrhotic patients;
According to the publication forms, 2 studies were published
the control group (i.e., patients without PVT) was missing;
in abstracts, and 16 studies in full-texts. Hepatocellular car-
the association between PVT and risk of bleeding was not
cinoma was excluded in 7 studies, but not in 9 studies. The
evaluated.
information regarding the exclusion of hepatocellular carci-
noma was not reported in 2 other studies. The prevalence
Data extraction of PVT in liver cirrhosis was 725%.
Multivariate analyses were performed in 8 studies
We extracted the following characteristics of the included [11,13,14,17,18,23,24,28], and only univariate analyses
studies: rst author, publication year, study design, enrol- were performed in 10 others. Of these studies with-
ment period, target population, treatment modalities, total out multivariate analyses, 5 had similar proportions of

Please cite this article in press as: Qi X, et al. Association between portal vein thrombosis and risk of
bleeding in liver cirrhosis: A systematic review of the literature. Clin Res Hepatol Gastroenterol (2015),
http://dx.doi.org/10.1016/j.clinre.2015.02.012
+Model
CLINRE-736; No. of Pages 9 ARTICLE IN PRESS
PVT and bleeding in liver cirrhosis 3

Figure 1 Flowchart of study selection.

Child-Pugh classes between patients with and without PVT
[12,15,20,25,27], 1 had similar MELD scores between the
two groups [26], 3 had signicantly different proportions of
Child-Pugh classes between the two groups [16,19,21], and
1 did not clearly report any relevant information [22]. Thus,
14 studies were considered to have relatively high-grade
evidence.
The results regarding the association between PVT and
the risk of bleeding were shown in Table 2.
Prior bleeding
Three studies showed that the cirrhotic patients with PVT
had a higher proportion of history of variceal bleeding than
those without PVT [15,16,22]. First, Schmassmann et al.
enrolled 60 cirrhotic patients with and without previous
hemorrhage [22]. In the hemorrhage group, 10 of 30 patients
had PVT; by contrast, in the non-hemorrhage group, only
3 of 30 patients had PVT. Second, Francoz et al. enrolled
251 cirrhotic patients listed for the rst liver transplanta-
tion [16]. Of the patients with splanchnic vein thrombosis,
47% (18/38) had a previous history of variceal bleeding; by
comparison, 23% (48/213) of the patients without splanchnic
vein thrombosis had a previous history of variceal bleeding.
Third, Doumit et al. also reported that signicantly more cir-
rhotic patients with PVT had a history of variceal bleeding at
admission (the detailed data were missing in the abstract)
[15].
By contrast, one study by Amitrano et al., in which 387
patients presenting with esophageal variceal bleeding within
recent 4 weeks were included, reported a similar proportion
of previous bleeding between patients with and without PVT
(10/67, 14.9% versus 60/316, 19.0%; P = 0.435) [25]. But the
source of previous bleeding was not shown.
All bleeding (de novo or rebleeding) during the
total follow-up
Two studies reported a higher rate of bleeding in cirrhotic
patients with PVT than in those without PVT. John et al.
prospectively collected 290 cirrhotic patients evaluated for
liver transplantation [26]. Among them, 47 patients were

Please cite this article in press as: Qi X, et al. Association between portal vein thrombosis and risk of
bleeding in liver cirrhosis: A systematic review of the literature. Clin Res Hepatol Gastroenterol (2015),
http://dx.doi.org/10.1016/j.clinre.2015.02.012
+Model
CLINRE-736; No. of Pages 9 ARTICLE IN PRESS
4 X. Qi et al.

Table 1 Characteristics of included studies.

First author Country Design Period Target population Total number No. PVT
(year) of patients (%)

Amitrano Italy Prospective January 2010 LC with acute EVB 185 32 (17%)
et al., 2012 cohort study July 2011 vasoactive therapy,
[11] (full-text) antibiotics, and EVL
Amitrano Italy Cohort study January 2001 LC and recent EVB 387 67 (17.3%)
et al., 2012 (full-text) December 2010 treated by band ligation
[25] (HCC was excluded)
Attili et al., Italy Prospective February 2000 LC without HCC 129 25 (19%)
2012 [12] longitudinal July 2005
study
(abstract)
Chen et al., Taiwan Retrospective July 2005 LC with acute EVB treated 101 25 (25%)
2012 [13] study December 2009 with somatostatin,
(full-text) antibiotics, EVL
DAmico et al., Italy Multicenter, October LC with hematemesis 291a 37 (13%)a
2003 [14] prospective, 1997January and/or melena
cohort study 1998a (treatments include
(full-text) vasoactive drugs,
endoscopic therapy,
combination of
endoscopic and
vasoactive therapy,
balloon tamponade alone,
or none)
DellEra et al., Italy Retrospective February 1995 LC treated with EVL in the 214 44 (11%)
2014 [27] study (full-text) February 2009 primary and secondary
prophylaxis of variceal
bleeding (advanced HCC
was excluded)
Doumit et al., Canada NA 2002 LC who underwent portal 398 44 (11%)
2009 [15] (abstract) vein doppler US and who
had no pre-existing HCC,
TIPS or surgical shunt
in-situ, or prior LT
Francoz et al., France NA Jan 1996 LC listed for a rst LT with 251 38 (15%)
2005 [16] (full-text) December 2001 or without hemorrhage
Hung et al., Taiwan RCT April 2007 LC with acute GVB after 95 13 (14%)
2012 [17] (full-text) March 2011 primary hemostasis using
gastric variceal
obturation therapy
(randomized to repeated
gastric variceal
obturation alone or in
combination with
non-selective -blockers)
John, et al. USA Prospective July 2004 LC evaluated for LT (HCC 290 70b (24%)
2013 [26] cohort study June 2009 was excluded, follow-up
(full-text) > 6 months)
Lee et al., 2009 Taiwan Retrospective Decemeber 2005 LC after the cessation of 97 19 (20%)
[18] study February 2008 acute EVB
(full-text)
Nery, et al. France, Satellite study of June 2000 LC without HCC 1243 118c (9%)
2015 [28] Belgium a multicenter March 2006
RCT (full-text)

Please cite this article in press as: Qi X, et al. Association between portal vein thrombosis and risk of
bleeding in liver cirrhosis: A systematic review of the literature. Clin Res Hepatol Gastroenterol (2015),
http://dx.doi.org/10.1016/j.clinre.2015.02.012
+Model
CLINRE-736; No. of Pages 9 ARTICLE IN PRESS
PVT and bleeding in liver cirrhosis 5

Table 1 (Continued)

First author Country Design Period Target population Total number No. PVT
(year) of patients (%)

Orloff et al., USA Prospective 19581991 LC with acutely bleeding 1300 85 (7%)
1997 [19] cohort study esophagogastric varices
(full-text) or a previous episode of
bleeding esophagogastric
varices treated with
emergency or elective
portacaval shunt
Perarnau et al., France Retrospective 19902004 LC undergoing TIPS for 273 27 (10%)
2010 [20] study emergency bleeding
(full-text) hemostasis or rebleeding
prevention
Sarfeh, 1979 USA Retrospective 19721978 Biopsy-proved LC treated 86 18 (21%)
[21] study with portal
(full-text) decompression surgery
for active or previous
variceal hemorrhage
Schmassmann Switzerland NA January 1989 Biopsy-proved LC with or 60 13 (22%)
et al., 1993 (full-text) Decemeber 1990 without hemorrhage
[12]
Wu et al., 2002 Taiwan Retrospective November 1992 LC with acute GVB 83 15 (18%)
[23] study October 1998 treated with endoscopic
(full-text) N-butyl-2-cyanoacrylate
injection
Yang et al., Taiwan NA May 2002 LC with esophageal 96 9 (9%)
2007 [24] (full-text) October 2004 varices treated with
elective or emergent EVL
EVB: esophageal variceal bleeding; EVL: endoscopic variceal ligation; GVB: gastric variceal bleeding; HCC: hepatocellular carcinoma;
LC: liver cirrhosis; LT: liver transplantation; NA: not available; PVT: portal vein thrombosis; RCT: randomized controlled trial; TIPS:
transjugular intrahepatic portosystemic shunt.
a The data in the training set.
b Forty-seven patients were diagnosed with PVT at baseline, and 23 patients developed PVT during follow-up.
c One hundred and eighteen patients developed PVT during follow-up.

diagnosed with PVT at baseline, and another 23 patients
developed PVT during the follow-up period. The incidence
of upper gastrointestinal bleeding was 21.3% (10/47) in
patients with PVT at baseline, 17.4% (4/23) in those who
developed PVT, and 11.4% (25/220) in those without PVT.
A trend toward a higher incidence of bleeding in patients
with PVT was observed (14/70 versus 25/220, P = 0.065).
DellEra et al. performed a retrospective analysis to explore
the impact of PVT on the efcacy of endosopic variceal
band ligation in 214 cirrhotic patients (22 with PVT and 192
without PVT) [27]. The rate of bleeding was higher in PVT
patients than in non-PVT patients (2/22, 9.1% versus 15/192,
7.8%), but no signicant difference was achieved. Notably,
one bleeding episode in non-PVT patients was attributed to
endoscopic treatment-related ulcer.
In addition, a multicenter prospective study by Nevy
et al. compared the risk of decompensation between
patients who developed PVT during follow-up and those
who did not [28]. Variceal bleeding was one of 4 features
for decompensation. In the univariate analysis, the risk of
decompensation was signicantly associated with the devel-
opment of PVT. However, the signicance disappeared in the
multivariate analysis.
De novo bleeding during the total follow-up
Two follow-up studies reported a higher incidence of de novo
bleeding in cirrhotic patients with PVT than in those with-
out PVT [12,15]. Notably, the detailed treatment modalities
were not reported. First, in a 5-year follow-up study of 398
cirrhotic patients, the incidence of de novo variceal bleed-
ing was 39% and 20% in PVT (n = 44) and no PVT (n = 354)
groups, respectively [15]. Second, Attili et al. also reported
a signicantly higher cumulative incidence of gastrointesti-
nal bleeding in cirrhotic patients who developed PVT during
the follow-up than in those who did not (P < 0.00001, by
log-rank test) [12].
De novo bleeding within 14 days after endoscopic
therapy
One study by Yang et al. analyzed the risk factors associated
with early bleeding in cirrhotic patients undergoing elective
or emergent endoscopic variceal band ligation [24]. Early
bleeding was dened as the bleeding episode or rebleed-
ing within 14 days after endoscopic therapy. The proportion

Please cite this article in press as: Qi X, et al. Association between portal vein thrombosis and risk of
bleeding in liver cirrhosis: A systematic review of the literature. Clin Res Hepatol Gastroenterol (2015),
http://dx.doi.org/10.1016/j.clinre.2015.02.012
+Model
CLINRE-736; No. of Pages 9
6 X. Qi et al.
Table 2 Results of included studies.
First author (year)
Amitrano et al., 2012
[11] AJG
Amitrano et al., 2012
[11] EJGH
Attili et al., 2012 [12]
Chen et al., 2012 [13]
DAmico et al., 2003 [14]
DellEra et al., 2014 [27]
Doumit et al., 2009 [15]
Francoz et al., 2005 [16]
Hung et al., 2012 [17]
John et al., 2013 [26]
Lee et al., 2009 [18]
Nery et al., 2014 [28]
Orloff et al., 1997 [19]
Perarnau et al., 2010
[20]
Sarfeh, 1979 [21]
Schmassmann et al.,
1993 [12]
Wu et al., 2002 [22]
Yang et al., 2007 [23]
CI: condence interval; HR: hazard ratio; OR: odds ratio; LC: liver cirrhosis; LT: liver transplantation; NA: not available; PVT: portal
vein thrombosis.
Please cite this article in press as: Qi X, et al. Association between portal vein thrombosis and risk of
bleeding in liver cirrhosis: A systematic review of the literature. Clin Res Hepatol Gastroenterol (2015),
http://dx.doi.org/10.1016/j.clinre.2015.02.012
ARTICLE IN PRESS
Outcomes observed
Proportion of PVT in patients with and
without 5-day failure after acute variceal
bleeding
Rebleeding at the 1-year follow-up in
patients with and without PVT
Rebleeding at the nal follow-up in patients
with and without PVT
Gastrointestinal bleeding from any sources
in patients with and without PVT
PVT for predicting the 6-week rebleeding
PVT for predicting the 5-day failure after
acute upper digestive bleeding
Bleeding from esophageal varices in
patients with and without PVT
Percentage of variceal bleeding during
follow-up in patients with and without PVT
History of variceal bleed at baseline in
patients with and without PVT
History of variceal bleed in patients with
and without PVT
PVT for predicting the gastric variceal
rebleeding
Upper gastrointestinal bleeding in patients
with and without PVT
Percentage of PVT in patients with and
without rebleeding
PVT for predicting the decompensation
(including variceal bleeding)
Percentage of variceal rebleeding in
patients with and without PVT
Percentage of rebleeding in patients with
and without rebleeding
Proportion of rebleeding in PVT and
non-PVT groups
Proportion of PVT in hemorrhage and
non-hemorrhage groups
PVT for predicting the gastric variceal
rebleeding
Proportion of PVT in bleeding and
non-bleeding groups
Comparative data (statistical signicance)
8/31 versus 24/154 (P = 0.047)
8/67 versus 30/316 (P = 0.543)
14/67 versus 52/316 (P = 0.382)
No detailed data were reported. (P < 0.00001
by log-rank test)
Univariate: HR = 2.62, 95% CI: 1.185.79
(P = 0.018)
Multivariate: HR = 2.734, 95% CI: 1.2286.088
(P = 0.014)
For any sources of bleeding: multivariate:
OR = 3.19, 95% CI: 1.536.67 (P = 0.002)
For variceal bleeding: multivariate: OR = 3.06,
95% CI: 1.396.68 (P = 0.005)
2/22 versus 15/192 (P = 0.761)
39% versus 20% (P < 0.05)
Data were not reported (P = 0.05)
18/38 versus 48/165 (P = 0.001)
Univariate: HR = 3.344, 95% CI: 1.6136.897
(P = 0.001)
Multivariate: HR = 4.049, 95% CI: 1.9088.621
(P < 0.001)
14/70 versus 25/220 (P = 0.065)
5/14 versus 14/83 (P = 0.141)
For partial PVT: univariate: HR = 1.77, 95% CI:
1.072.92 (P = 0.027)
For partial PVT: multivariate: HR = 1.60, 95%
CI: 0.693.74 (P = 0.28)
For partial or complete PVT: univariate:
HR = 1.61, 95% CI: 0.982.62 (P = 0.058)
For partial or complete PVT: multivariate:
HR = 1.37, 95% CI: 0.623.03 (P = 0.44)
5% versus 1% (emergency) or 0.3% (elective)
2/29 versus 18/402 (not signicant)
8/18 versus 4/68 (P < 0.01)
10/30 versus 3/30 (not signicant)
Univariate: OR = 0.96, 95% CI: 0.248.89 (not
signicant)
Multivariate: OR = 0.17, 95% CI: 0.021.69 (not
signicant)
4/19 versus 5/77 (not signicant)
+Model
CLINRE-736; No. of Pages 9 ARTICLE IN PRESS
PVT and bleeding in liver cirrhosis
of PVT was higher in bleeding group than in non-bleeding
group (4/19, 21% versus 5/77, 7%), but the difference was
not statistically signicant.
Five-day failure after medical/endoscopic therapy
Two prospective cohort studies demonstrated a higher rate
of 5-day failure in cirrhotic patients with PVT than in those
without PVT [11,14]. Five-day failure was dened as the fail-
ure to control bleeding, rebleeding, or death within 5 days.
First, DAmico et al. evaluated the short-term outcomes and
prognostic indicators of cirrhotic patients with upper diges-
tive bleeding [14]. In the training cohort, 37 of 251 cirrhotic
patients had PVT. In the multivariate logistic regression
analysis, the presence of PVT was the independent pre-
dictor for 5-day failure (for any sources of bleeding: odds
ratio = 3.19, 95% condence interval: 1.536.67, P = 0.002;
for variceal bleeding: odds ratio = 3.06, 95% condence
interval: 1.396.68, P = 0.005). Second, Amitrano et al.
collected the data from 185 cirrhotic patients with acute
variceal bleeding [11]. In the univariate analysis, PVT was
more prevalent in patients with 5-day failure than in
those without (26%, 8/31 versus 16%, 24/154, P = 0.047).
But only a trend was showed without any statistical
signicance in a multivariate logistic regression analysis
(odds ratio = 2.942, 95% condence interval: 0.8849.790,
P = 0.079).
Six-week rebleeding after medical/endoscopic
therapy
Two retrospective studies found a higher rate of 6-week
rebleeding in cirrhotic patients with PVT than in those with-
out PVT [13,18]. First, Chen et al. enrolled 101 cirrhotic
patients with endoscopy-proven active esophageal variceal
bleeding who underwent esophageal variceal ligation [13].
Among them, 25 patients were diagnosed with PVT. Uni-
variate Cox regression analysis demonstrated a signicant
relationship between PVT and 6-week rebleeding (hazard
ratio = 2.62, 95% condence interval: 1.185.79, P = 0.018).
The statistical signicance was conrmed by a multivari-
ate analysis (hazard ratio = 2.734, 95% condence interval:
1.2286.088, P = 0.014). Lee et al. also showed a higher pro-
portion of PVT in patients who developed variceal rebleeding
within 6 weeks after the cessation of initial esophageal
variceal bleeding than in those who did not (5/14, 36% ver-
sus 14/83, 17%) [18]. But the difference between the two
groups was not statistically signicant.
Rebleeding after medical/endoscopic therapy
during the total follow-up
One study by Amitrano et al. showed a higher incidence of
rebleeding from varices in patients with PVT than in those
without PVT at the 1-year follow-up (8/67, 11.9% versus
30/316, 9.5%; P = 0.543) and at the nal follow-up (14/67,
20.9% versus 52/316, 16.5%; P = 0.382) [25].
Please cite this article in press as: Qi X, et al. Association between portal vein thrombosis and risk of
bleeding in liver cirrhosis: A systematic review of the literature. Clin Res Hepatol Gastroenterol (2015),
http://dx.doi.org/10.1016/j.clinre.2015.02.012
7
Gastric variceal rebleeding after
medical/endoscopic therapy during the total
follow-up
In 2 studies, the ndings regarding the association between
gastric variceal rebleeding and PVT in liver cirrhosis were
opposite [17,23]. First, Wu et al. retrospectively analyzed
the risk factors of gastric variceal rebleeding in 83 patients
undergoing endoscopic N-butyl-2-cyanoacrylate injection
[23]. Fifteen of them had PVT. The odds ratios for PVT were
0.96 (95% condence interval: 0.248.89) and 0.17 (95%
condence interval: 0.021.69) in univariate and multivari-
ate logistic regression analyses, respectively. This suggested
a similar incidence of gastric variceal rebleeding between
cirrhotic patients with and without PVT. Second, a random-
ized controlled trial by Hung et al. evaluated the predictors
for rebleeding in 95 cirrhotic patients with acute gastric
variceal bleeding after gastric variceal obturation in com-
bination with and without non-selective beta-blockers [17].
Cox regression analysis identied the presence of main por-
tal vein thrombosis as a major predictor for rebleeding
(univariate analysis: hazard ratio = 3.344, 95% condence
interval: 1.6136.897, P = 0.001; multivariate analysis: haz-
ard ratio = 4.049, 95% condence interval: 1.9088.621,
P = 0.001).
Rebleeding after shunt surgery during the total
follow-up
Two studies demonstrated a higher rate of rebleeding after
shunt surgery in cirrhotic patients with PVT than in those
without [19,21]. First, in an early report, Sarfeh observed
the probability of rebleeding in cirrhotic patients who had
undergone portal decompression surgery for variceal hem-
orrhage [21]. Seven of 68 patients with the patent portal
veins rebled from varices, whereas 9 of 18 patients with
the thrombosed portal veins rebled (P < 0.01). Second, in a
prospective cohort study, Orloff et al. enrolled 1300 cirrhotic
patients who underwent emergency (n = 400) and elective
(n = 900) portacaval shunt surgery [19]. The incidence of
variceal rebleeding was 5% in 85 patients with PVT, 1% in
335 patients without PVT undergoing emergency surgery,
and 0.3% in 880 patients without PVT undergoing elective
surgery.
Rebleeding after transjugular intrahepatic
portosystemic shunt during the total follow-up
In a retrospective study by Perarnau et al., the incidence
of rebleeding after transjugular intrahepatic portosystemic
shunt was similar between cirrhotic patients with and with-
out PVT (8% versus 7%) [20].
Discussion
This is the rst systematic review of the literature to explore
whether or not the presence of PVT can increase the risk of
bleeding in liver cirrhosis. In this study, the relevant papers
were identied from an extensive search strategy. Addition-
ally, the reliability of their ndings was assessed according
+Model
CLINRE-736; No. of Pages 9 ARTICLE IN PRESS
8 X. Qi et al.
to the pre-specied criteria. A majority of studies reported
relatively high-grade evidence.
Several positive correlations were summarized as fol-
lows:
at admission, a higher proportion of cirrhotic patients
with PVT had experienced bleeding;
regardless of treatment modalities, previous histories of
bleeding, or sources of bleeding, PVT increased the risk
of de novo bleeding;
as for the cirrhotic patients with acute variceal bleeding
treated with medical/endoscopic therapy, PVT increased
the risk of 5-day failure and 6-week rebleeding;
as for the cirrhotic patients undergoing surgical shunts for
the management of variceal bleeding, a pre-existing PVT
increased the risk of rebleeding.
Unexpectedly, as for the cirrhotic patients with gastric
variceal bleeding treated with medical/endoscopic therapy,
whether or not PVT would increase the risk of rebleeding
remained controversial in 2 studies. This might be explained
by the fact that gastric variceal bleeding was more likely to
develop at a relatively low portal pressure than esophageal
variceal bleeding [29,30]. In addition, as for the cirrhotic
patients undergoing transjugular intrahepatic portosystemic
shunts for the management of variceal bleeding, a pre-
existing PVT was not associated with the risk of rebleeding.
However, it should be noticeable that a transjugular intra-
hepatic portosystemic shunt becomes technically complex
in the setting of complete PVT and even impossible when an
occluded portal vein progresses into brotic cord [31,32].
Generally, these ndings suggested that PVT should increase
the risk of bleeding in liver cirrhosis in nearly all clinical
conditions but after transjugular intrahepatic portosystemic
shunt.
Given the detrimental effect of PVT on the risk of bleed-
ing, the necessity of recanalizing the thrombosed portal
vein in a timely fashion might be considered. Several case
series have recently showed a relatively high rate of por-
tal vein recanalization after anticoagulation in cirrhotic
patients with PVT with or without gastroesophageal varices
[16,3335]. And few serious drug-related adverse events
were reported [36]. On the other hand, the researchers have
suggested the possibility of spontaneous portal vein recanal-
ization in cirrhotic patients with partial PVT [26,37,38].
However, the probability of disease progression appeared
to be higher in the absence of anticoagulation therapy
[36]. More importantly, once PVT worsened, the progno-
sis of liver cirrhosis would be further deteriorated [38].
Unfortunately, no predictors for the thrombus extension in
untreated patients are well established [39]. Thus, antico-
agulation should be recommended to cirrhotic patients with
PVT, if its contraindications are excluded [40].
A major limitation of this study was that the treatment
modalities (medical/endoscopic therapy, shunt surgery, or
transjugular intrahepatic portosystemic shunt), sources of
bleeding (esophageal varices, gastric varices, or unknown),
lengths of observation (short-term or total follow-up), and
ways of data expression (number of events, hazard ratio,
or odd ratios) were very heterogeneous among studies.
Because only a few studies were eligible for every out-
come, the meta-analyses might be inappropriate. Instead,
Please cite this article in press as: Qi X, et al. Association between portal vein thrombosis and risk of
bleeding in liver cirrhosis: A systematic review of the literature. Clin Res Hepatol Gastroenterol (2015),
http://dx.doi.org/10.1016/j.clinre.2015.02.012
we just described the relevant outcomes from every indi-
vidual study. Another limitation was that most of included
studies (9/14) did not exclude the patients with hepatocel-
lular carcinoma. Thus, we could not identify whether the
nature of PVT should be attributed to malignancy or not. If
PVT originated from the tumor invasion, the patients would
be considered to have advanced HCC (BCLCC stage) with a
very short survival time [41], which might preclude from the
development of bleeding. Finally, the degree of PVT was not
reported in any included studies. Therefore, we could not
clarify the risk of bleeding in the subgroups with complete
versus partial PVT [42,43].
In conclusions, based on the present systematic review of
the literature, the presence of PVT should be positively asso-
ciated with the risk of portal hypertension-related bleeding
in liver cirrhosis. However, this issue regarding whether or
not PVT aggravated the development of bleeding during
follow-up needed to be further explored by well-designed
observational studies. If the answer was Yes, an early
diagnosis and treatment of PVT would be further encouraged
to minimize the risk of bleeding. Otherwise, a wait-and-see
strategy would be considered.
Author contributions
Xingshun Qi conceived and drafted the manuscript. Xing-
shun Qi, Chunping Su, Weirong Ren, Man Yang, Jia Jia, Junna
Dai, and Wenda Xu performed the literature search and
selection, and data extraction; Xiaozhong Guo gave criti-
cal comments and revised the manuscript. All authors have
made an intellectual contribution to the manuscript and
approved the submission.
Disclosure of interest
The authors declare that they have no conicts of interest
concerning this article.
References
[1] Garcia-Tsao G, Sanyal AJ, Grace ND, Carey W. Prevention and
management of gastroesophageal varices and variceal hemor-
rhage in cirrhosis. Hepatology 2007;46:92238.
[2] Garcia-Tsao G, Bosch J. Management of varices and variceal
hemorrhage in cirrhosis. N Engl J Med 2010;362:82332.
[3] Groszmann RJ, Garcia-Tsao G, Bosch J, et al. Beta-blockers to
prevent gastroesophageal varices in patients with cirrhosis. N
Engl J Med 2005;353:225461.
[4] de Franchis R. Revising consensus in portal hypertension:
report of the Baveno V consensus workshop on methodology
of diagnosis and therapy in portal hypertension. J Hepatol
2010;53:7628.
[5] DAmico G, Pagliaro L, Bosch J. The treatment of portal hyper-
tension: a meta-analytic review. Hepatology 1995;22:33254.
[6] Prediction of the rst variceal hemorrhage in patients with
cirrhosis of the liver and esophageal varices. A prospective
multicenter study. N Engl J Med 1988;319:9839.
[7] Qi X, Bai M, Yang Z, et al. Occlusive portal vein thrombosis
as a new marker of decompensated cirrhosis. Med Hypotheses
2011;76:5226.
[8] Garcia-Pagan JC, Valla DC. Portal vein thrombosis: a pre-
dictable milestone in cirrhosis? J Hepatol 2009;51:6324.
+Model
CLINRE-736; No. of Pages 9 ARTICLE IN PRESS
PVT and bleeding in liver cirrhosis
[9] Qi X, Yang M, Ren W, et al. Find duplicates among the PubMed,
EMBASE, and Cochrane Library Databases in systematic review.
PLoS One 2013;8:e71838.
[10] Qi X, Jia J, Ren W, et al. Scientic publications on portal vein
thrombosis and Budd-Chiari syndrome: a global survey of the
literature. J Gastrointestin Liver Dis 2014;23:6571.
[11] Amitrano L, Guardascione MA, Manguso F, et al. The effective-
ness of current acute variceal bleed treatments in unselected
cirrhotic patients: rening short-term prognosis and risk fac-
tors. Am J Gastroenterol 2012;107:18728.
[12] Attili AF, Lupo M, Gigliotti F, Trapani S, Attili F, De Santis A. Por-
tal vein thrombosis is a risk factor for death and complications
in liver cirrhosis. Dig Liver Dis 2012;44:S107.
[13] Chen PH, Chen WC, Hou MC, et al. Delayed endoscopy increases
re-bleeding and mortality in patients with hematemesis and
active esophageal variceal bleeding: acohort study. J Hepatol
2012;57:120713.
[14] DAmico G, De Franchis R. Upper digestive bleeding in cirrhosis.
Post-therapeutic outcome and prognostic indicators. Hepatol-
ogy 2003;38:599612.
[15] Doumit M, Bolano S, Yeung E, Khalili K, Wong F. The prevalence
of non-tumor related portal vein thrombosis. Can J Gastroen-
terol 2009:23.
[16] Francoz C, Belghiti J, Vilgrain V, et al. Splanchnic vein throm-
bosis in candidates for liver transplantation: usefulness of
screening and anticoagulation. Gut 2005;54:6917.
[17] Hung HH, Chang CJ, Hou MC, et al. Efcacy of non-selective
beta-blockers as adjunct to endoscopic prophylactic treatment
for gastric variceal bleeding: a randomized controlled trial. J
Hepatol 2012;56:102532.
[18] Lee SW, Lee TY, Chang CS. Independent factors associated
with recurrent bleeding in cirrhotic patients with esophageal
variceal hemorrhage. Dig Dis Sci 2009;54:112834.
[19] Orloff MJ, Orloff MS, Orloff SL, Girard B. Portal vein thrombo-
sis in cirrhosis with variceal hemorrhage. J Gastrointest Surg
1997;1:12330 [discussion 301].
[20] Perarnau JM, Baju A, DAlteroche L, Viguier J, Ayoub J.
Feasibility and long-term evolution of TIPS in cirrhotic
patients with portal thrombosis. Eur J Gastroenterol Hepatol
2010;22:10938.
[21] Sarfeh IJ. Portal vein thrombosis associated with cirrhosis: clin-
ical importance. Arch Surg 1979;114:9025.
[22] Schmassmann A, Zuber M, Livers M, Jager K, Jenzer HR, Fehr
HF. Recurrent bleeding after variceal hemorrhage: predictive
value of portal venous duplex sonography. AJR Am J Roentgenol
1993;160:417.
[23] Wu CY, Yeh HZ, Chen GH. Pharmacologic efcacy in gastric
variceal rebleeding and survival: including multivariate anal-
ysis. J Clin Gastroenterol 2002;35:12732.
[24] Yang MT, Chen HS, Lee HC, Lin CL. Risk factors and survival of
early bleeding after esophageal variceal ligation. Hepatogas-
troenterology 2007;54:17059.
[25] Amitrano L, Guardascione MA, Scaglione M, et al. Splanch-
nic vein thrombosis and variceal rebleeding in patients with
cirrhosis. Eur J Gastroenterol Hepatol 2012;24:13815.
[26] John BV, Konjeti R, Aggarwal A, et al. Impact of untreated por-
tal vein thrombosis on pre and post liver transplant outcomes
in cirrhosis. Ann Hepatol 2013;12:9528.
Please cite this article in press as: Qi X, et al. Association between portal vein thrombosis and risk of
bleeding in liver cirrhosis: A systematic review of the literature. Clin Res Hepatol Gastroenterol (2015),
http://dx.doi.org/10.1016/j.clinre.2015.02.012
9
[27] DellEra A, Iannuzzi F, Fabris FM, et al. Impact of portal vein
thrombosis on the efcacy of endoscopic variceal band ligation.
Dig Liver Dis 2014;46:1526.
[28] Nery F, Chevret S, Condat B, et al. Causes and consequences of
portal vein thrombosis in 1243 patients with cirrhosis: results
of a longitudinal study. Hepatology 2015;61:6607.
[29] Huang YH, Yeh HZ, Chen GH, et al. Endoscopic treatment
of bleeding gastric varices by N-butyl-2-cyanoacrylate (His-
toacryl) injection: long-term efcacy and safety. Gastrointest
Endosc 2000;52:1607.
[30] Seewald S, Ang TL, Imazu H, et al. A standardized injec-
tion technique and regimen ensures success and safety of
N-butyl-2-cyanoacrylate injection for the treatment of gas-
tric fundal varices (with videos). Gastrointest Endosc 2008;68:
44754.
[31] Qi X, Han G, Yin Z, et al. Transjugular intrahepatic por-
tosystemic shunt for portal cavernoma with symptomatic
portal hypertension in non-cirrhotic patients. Dig Dis Sci
2012;57:107282.
[32] Han G, Qi X, He C, et al. Transjugular intrahepatic por-
tosystemic shunt for portal vein thrombosis with symptomatic
portal hypertension in liver cirrhosis. J Hepatol 2011;54:
7888.
[33] Senzolo MT, Rossetto MSV, et al. Prospective evaluation of anti-
coagulation and transjugular intrahepatic portosystemic shunt
for the management of portal vein thrombosis in cirrhosis. Liver
Int 2012;32:91927.
[34] Delgado MG, Seijo S, Yepes I, et al. Efcacy and safety
of anticoagulation on patients with cirrhosis and por-
tal vein thrombosis. Clin Gastroenterol Hepatol 2012;10:
77683.
[35] Amitrano L, Guardascione MA, Menchise A, et al. Safety and
efcacy of anticoagulation therapy with low molecular weight
heparin for portal vein thrombosis in patients with liver cirrho-
sis. J Clin Gastroenterol 2010;44:44851.
[36] Qi X, De Stefano V, Li H, Dai J, Guo X, Fan D. Anticoagulation
for the treatment of portal vein thrombosis in liver cirrhosis: a
systematic review and meta-analysis of observational studies.
Eur J Intern Med 2015;26:1545.
[37] Maruyama H, Okugawa H, Takahashi M, Yokosuka O. De
novo portal vein thrombosis in virus-related cirrhosis: pre-
dictive factors and long-term outcomes. Am J Gastroenterol
2013;108:56874.
[38] Luca A, Caruso S, Milazzo M, et al. Natural course of extrahe-
patic non-malignant partial portal vein thrombosis in patients
with cirrhosis. Radiology 2012;265:12432.
[39] Qi X, Yang Z, Fan D. Spontaneous resolution of portal vein
thrombosis in cirrhosis: where do we stand, and where will
we go? Saudi J Gastroenterol 2014;20:2656.
[40] Qi X, Han G, Fan D. Management of portal vein thrombosis in
liver cirrhosis. Nat Rev Gastroenterol Hepatol 2014;11:43546.
[41] EASL-EORTC clinical practice guidelines: management of hepa-
tocellular carcinoma. J Hepatol 2012;56:90843.
[42] Qi X, Han G, Fan D. Degree of portal vein thrombosis might
be associated with recanalization during anticoagulation. Clin
Gastroenterol Hepatol 2012;10:820.
[43] Qi X, Han G, Wang J, Wu K, Fan D. Degree of portal vein throm-
bosis. Hepatology 2010;51:108990.