Incidence and prevalence of venous thromboembolism in chronic liver disease: A systematic review and meta-analysis

Abstract

Background and aims
Historically, bleeding was thought to be a frequent and fatal complication of liver disease. However, thrombosis due to coagulation disorders in cirrhosis remains a real risk. We aim to systematically analyse published articles to evaluate epidemiology of venous thromboembolism (VTE) in chronic liver disease (CLD).

Method
Electronic search was conducted on Ovid Medline, EMBASE and Scopus from inception to November 2021 to identify studies presenting epidemiology VTE (deep vein thrombosis and pulmonary embolism) in CLD in inpatients and/or community settings. Random-effects meta-analysis was performed to determine pooled per-year cumulative incidence, incidence rate and prevalence. Heterogeneity was measured by I² test, and, potential sources of heterogeneity by meta-regression and sensitivity analysis. PROSPERO registration-CRD42021239117.

Results
Twenty-nine studies comprising 19,157,018 participants were included, of which 15,2049 (0.79%) had VTE. None of the included studies were done in the community. In hospitalised patients with CLD: pooled cumulative incidence of VTE was 1.07% (95% CI 0.80,1.38) per-year, incidence rate was 157.15 (95% CI 14.74,445.29) per 10,000 person-years, and period prevalence was 1.10% (95% CI 0.85,1.38) per year. There was significant heterogeneity and publication bias. Pooled relative risk (RR) of studies reporting incidence rate was 2.11 (95% CI 1.35,3.31). CLD patients (n = 1644), who did not receive pharmacological prophylaxis were at 2.78 times (95% CI 1.11, 6.98) increased risk of VTE compared to those receiving prophylaxis.

Conclusion
Hospitalised patients with CLD may be at an increased risk of VTE. For every 1000 hospitalised patients with CLD ten have new, and eleven have pre-existing diagnoses of VTE per-year.

Full text

Thrombosis Research 215 (2022) 19–29
Available online 17 May 2022
0049-3848/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Incidence and prevalence of venous thromboembolism in chronic liver
disease: A systematic review and meta-analysis
Mohsan Subhani
a
,
b
,
*
, Abhishek Sheth
a
,
b
, Jamal Ahmed
c
, Pramudi Wijayasiri
a
,
b
,
Syed A. Gardezi
c
, Doyo Enki
d
, Joanne R. Morling
a
,
b
,
e
, Guruprasad P. Aithal
a
,
b
,
Stephen D. Ryder
a
,
b
, Aloysious D. Aravinthan
a
,
b
a
Nottingham Digestive Diseases Centre (NDDC), Translational Medical Sciences, School of Medicine, University of Nottingham, NG7 2UH, UK
b
NIHR Nottingham Biomedical Research Centre, Nottingham University Hospitals NHS Trust and the University of Nottingham, Nottingham, NG7 2UH, UK
c
Royal Gwent Unit, Royal Gwent Hospital, Newport, NP20 2UB, UK
d
School of Medicine, University of Nottingham, UK
e
Division of Epidemiology and Public Health, University of Nottingham, Nottingham, NG5 1PB, UK
ARTICLE INFO
Keywords:
Chronic liver disease
Venous thromboembolism
Deep vein thrombosis
Pulmonary embolism
Epidemiology
Incidence
Prevalence
Systematic review
ABSTRACT
Background and aims: Historically, bleeding was thought to be a frequent and fatal complication of liver disease.
However, thrombosis due to coagulation disorders in cirrhosis remains a real risk. We aim to systematically
analyse published articles to evaluate epidemiology of venous thromboembolism (VTE) in chronic liver disease
(CLD).
Method: Electronic search was conducted on Ovid Medline, EMBASE and Scopus from inception to November
2021 to identify studies presenting epidemiology VTE (deep vein thrombosis and pulmonary embolism) in CLD in
inpatients and/or community settings. Random-effects meta-analysis was performed to determine pooled per-
year cumulative incidence, incidence rate and prevalence. Heterogeneity was measured by I
2
test, and, poten-
tial sources of heterogeneity by meta-regression and sensitivity analysis. PROSPERO registration-
CRD42021239117.
Results: Twenty-nine studies comprising 19,157,018 participants were included, of which 15,2049 (0.79%) had
VTE. None of the included studies were done in the community. In hospitalised patients with CLD: pooled cu-
mulative incidence of VTE was 1.07% (95% CI 0.80,1.38) per-year, incidence rate was 157.15 (95% CI
14.74,445.29) per 10,000 person-years, and period prevalence was 1.10% (95% CI 0.85,1.38) per year. There
was signicant heterogeneity and publication bias. Pooled relative risk (RR) of studies reporting incidence rate
was 2.11 (95% CI 1.35,3.31). CLD patients (n =1644), who did not receive pharmacological prophylaxis were at
2.78 times (95% CI 1.11, 6.98) increased risk of VTE compared to those receiving prophylaxis.
Conclusion: Hospitalised patients with CLD may be at an increased risk of VTE. For every 1000 hospitalised
patients with CLD ten have new, and eleven have pre-existing diagnoses of VTE per-year.
1. Introduction
Venous thromboembolism (VTE) in chronic liver disease (CLD) is an
increasingly encountered complication [1]. Although, traditionally,
bleeding was thought to be a frequent and fatal complication of CLD,
evidence now supports that an auto-anticoagulatory state in CLD can
predispose to thrombosis [2,3]. Liver plays a pivotal role in the
regulation of coagulation pathways by producing both pro-coagulant
and anti-coagulant factors [4]. Signicant impairment in liver syn-
thetic function causes a state of dynamic disequilibrium in haemostasis,
which may increase the risk of both bleeding and thromboembolic
events [5]. CLD-associated coagulopathy is due to complex alterations in
liver and vascular endothelial haemostatic factors, such as reduced
protein C, protein S and antithrombin, and increased Von Willebrand
factor and factor VIII [6]. These pathophysiological haemostatic
* Corresponding author at: Queens Medical Centre, Derby Road, Nottingham NG72UH, UK.
E-mail addresses: mohsan.subhani@nottingham.ac.uk (M. Subhani), abhishek.sheth@nottingham.ac.uk (A. Sheth), jamal.ahmed@wales.nhs.uk (J. Ahmed),
pramudi.wijayasiri@nottingham.ac.uk (P. Wijayasiri), syed.gardezi@wales.nhs.uk (S.A. Gardezi), doyo.enki@nottingham.ac.uk (D. Enki), joanne.morling@
nottingham.ac.uk (J.R. Morling), guru.aithal@nottingham.ac.uk (G.P. Aithal), stephen.ryder@nuh.nhs.uk (S.D. Ryder), Aloysious.aravinthan@nottingham.ac.uk
(A.D. Aravinthan).
Contents lists available at ScienceDirect
Thrombosis Research
journal homepage: www.elsevier.com/locate/thromres
https://doi.org/10.1016/j.thromres.2022.05.004
Received 9 March 2022; Received in revised form 22 April 2022; Accepted 7 May 2022

Thrombosis Research 215 (2022) 19–29
20
changes in turn can promote a procoagulant state translating to an
increased risk of thrombosis [7,8].
According to 2017 Global Burden of Disease study, CLD is attributed
to two million deaths per year worldwide, over half of these deaths are
due to complications of cirrhosis [9,10]. Chronic changes in liver ar-
chitecture can eventually lead to signicant hepatic dysfunction and
result in fatal complications including coagulation disorders, hep-
atorenal syndrome, spontaneous bacterial peritonitis, hepatic encepha-
lopathy, and hepatocellular carcinoma [11]. Management of these
complications often requires recurrent hospital admissions. Hospital-
isation, burden of comorbidities, hepatic synthetic dysfunction and pe-
riods of immobility signicantly increase the risk of VTE; deep vein
thrombosis (DVT) and pulmonary embolism (PE) [12].
Liver focused interventions in hospital have helped to reduce in-
hospital mortality in patients with CLD [13], whereas, the in-hospital
mortality in CLD after a thrombotic event remains a signicant
concern [14,15]. A Danish nationwide cohort study reported a 7% (95%
CI 5, 10%) 30-day mortality in patients with DVT and cirrhosis
compared to 3% (95% CI 2,3%) in patients with DVT but without
cirrhosis. The 30-day mortality was 35% (95% CI 29, 42%) in patients
with PE and cirrhosis compared to 16% (95% CI 14–19%) in patients
with PE but without cirrhosis [15]. Hospitalised patients with cirrhosis
who develop VTE during hospital stay are at twice increased risk (OR
2.16, 95% CI 1.96, 2.38) of in-hospital death compared to those without
VTE [14].
Incidence rate of VTE (DVT and PE) in all hospitalised patients has
been reported to be as high as 960 per 10,000 person years, whereas the
incidence rate is at least 100 times lower in community residents (7.1
per 10,000 person years) [16]. In hospitalised patients 4 to 12% have
VTE events, and VTE contributes to 7 to 10% of all in-hospital deaths
[17,18]. There is signicant heterogeneity in the reported incidence of
VTE in CLD with some researchers reporting a lower risk and others a
higher risk. Reported incidences of VTE in hospitalised patients with
cirrhosis varies between 0.33 and 6.32% [19,20]. Results from previous
systematic reviews have been conicting. Qi et al.; (2014) reported a
cumulative incidence of 1.0% [20], while, Ambrosino et al.; (2017) re-
ported a 3.7% of VTE (DVT and PE) in hospitalised patients with CLD
[2]. Moreover, since the publication of the above systematic reviews
[2,20], there has been substantial new data examining epidemiology of
VTE in CLD. Nine new studies [21–29] that were not included in the
previous systematic reviews, comprising over seven million participants
and sixty thousand VTE events (DVT and/or PE) have been published.
We aim to systematically analyse published articles to evaluate the
epidemiology of VTE in patients with CLD, and to compare relative risk
of VTE in patients with CLD to those without CLD.
2. Methods
Joanna Briggs Institute's methodological guidelines for systematic
reviews of observational studies reporting prevalence and incidence
[30] and Meta-analyse Of Observational Studies in Epidemiology
(MOOSE) guidelines were followed [31]. Protocol was registered on
Prospero (CRD42021239117).
2.1. Search strategy
We searched Ovid Medline, EMBASE and Scopus from databases
inception to 2nd November 2021. Search for grey literature was con-
ducted using Open Grey, Ethos, Google scholar, and clinical.trials.gov.
Reference list of included articles was manually searched. Condition,
Context, population (CoCoPO) model-based search strategy was devel-
oped [30,32]. The search strategy was developed in consultation with
expert an expert librarian.
Different combinations of following search terms were used;
“Chronic liver disease”, or “Cirrhosis” or “Cirrhotic”, or “Cirrhosis” or
“end stage liver disease”, or “CLD” or “advanced liver brosis” or
“decompensated liver disease” Or (CLD and liver), “liver cirrhosis” or
“Cirrho$” or “CLD adj2 VTE”, or “Cirrhosis adj3 VTE”, or “CLD adj3
venous, “(‘liver’ or ‘hepatic’)”, or thromboembolism”, or “Cirrhosis adj3
venous thromboembolism”, AND “Venous thromboembolism”, or “deep
vein thrombosis”, or “deep venous thrombosis”, or “pulmonary embo-
lism”, or “VTE” or “DVT”, or “PE” or, “venous thrombosis”, or “throm-
botic disease”, or “thromboembolic disease”, AND, “Inpatient”, or
“hospitalised”, or “Hospital”, or “Hospitali$ed”, or “secondary care”, or
“in patient”, OR, “outpatient”, or “out-patient”, or “community”, or “out
of hospital”, or “primary care”, AND (“incidence” or “prevalence” or
“epidemiology” or “risk” or “predictor” or “predictive”).
In cases of missing data, abstract only publication and data enquiry,
the corresponding author of the study was contacted.
2.2. Eligibility
2.2.1. Inclusion criteria
Studies reporting incidence and/or prevalence of VTE (DVT and PE),
in adult patients with CLD including cirrhosis in hospital or community
setting were included. CLD was diagnosed based on clinical presenta-
tion, laboratory tests, radiological imaging, or liver biopsy, and VTE was
diagnosed on radiological imaging.
2.2.2. Exclusion criteria
•Systematic reviews, literature reviews, and editorials
•Studies reporting incidence of VTE post liver transplant, liver
resection, and in patients with hepatocellular carcinoma (HCC), or
after treatment for HCC.
•Studies which had insufcient data to determine pooled incidence or
prevalence
2.3. Outcomes
2.3.1. Primary outcome
Primary outcome was to describe weighted average (pooled) of cu-
mulative incidence (per year), incidence rate (person-years), and period
prevalence (per year) of VTE (combined DVT and PE) in patients with
CLD.
2.3.2. Secondary outcomes
Secondary outcomes were:
•DVT and PE analysis separately
•Case versus control analysis for studies included control cohort
without CLD
•Case versus control analysis for VTE prophylaxis versus no VTE
prophylaxis
Based on descriptions in the included studies the following deni-
tions were adopted for the current systematic review.
Nomenclature
CLD Chronic liver disease
DVT Deep vein thrombosis
HCC Hepatocellular carcinoma
INR International normalised ratio
NAFLD Non-alcoholic fatty liver disease
PE Pulmonary embolism
RR Relative risk
VTE Venous thromboembolism
M. Subhani et al.

Thrombosis Research 215 (2022) 19–29
21
2.3.3. Cumulative incidence (per year)
Number of new cases of VTE in hospitalised patients with CLD over 1
year per total number of CLD patients hospitalised in that year.
2.3.4. Incidence rate (person-years)
New cases of VTE in hospitalised patients with CLD per 10,000
person years.
2.3.5. Period prevalence (per year)
Number of cases who had pre-existing diagnosis of VTE in hospital-
ised patients with CLD in 1 year per total number of CLD patients hos-
pitalised in that year.
2.4. Screening and data extraction
Two reviewers (MS and SAG) independently screened the titles and
abstracts for eligibility, removed duplicate entries, and recorded re-
viewers' decisions using Rayyan-QRCI systematic review software,
Endnote (version-X9) and Microsoft Excel. Third reviewer (PW) oversaw
the process and resolved any conicts in discussion with senior author
(ADA). A three-stage data extraction approach was adopted. First, the
review team met at the start to nalise the data extraction proforma and
conducted a pilot data extraction to ensure consistency of data extrac-
tion and resolved any queries. Second, four reviewers in two pairs (SAG
and JA, MS and AS) independently extracted data from the included
studies, fth reviewer (PW) cross checked extracted data for any in-
consistencies or errors. Third, the nal data extraction was indepen-
dently reviewed by the senior author (ADA) to resolve any conicts.
2.5. Quality assessment
Quality and risk of bias assessment for included studies was carried
out using Critical Appraisal Skills Programme (CASP) tool [33].
2.6. Data synthesis and analysis
Statistical analysis was carried out using RStudio version 4.0.2
(2020-06-22).
Descriptive statistics were calculated to determine number of events
(VTE) and sample size. Due to the presence of signicant heterogeneity
among included studies a random effects meta-analysis was conducted.
Where heterogeneity was non-signicant a xed effects meta-analysis
was conducted. Point estimates (per year), incidence rate (person-
years) and period prevalence (per year) with 95% condence interval
(CI) from individual studies were pooled using DerSimonian-Laird
random effects methods [34], with variations in raw proportion dealt
with using the Freeman–Tukey double arcsine transformation [35]. CI
for variance between studies were calculated using Jackson method for
condence interval of tau [2] and tau [36]. Whereas Clopper-Pearson
method was used to determine condence interval for individual
studies [37]. Pooled cumulative incidence and period prevalence were
reported per 100 patients per year with 95% CI. Pooled incidence rate
was reported as per 10,000 person years with 95% CI. Heterogeneity
between studies was calculated using ‘I
2
’. An I
2
>50% indicated sig-
nicant heterogeneity. Differences between subgroups was assessed
using Cochran's Q (chi-square). Forest plots were used for graphical
display of estimated study results and funnel plots for publication bias.
Signicance of publication bias was conrmed using contour enhanced
funnel plot and by Egger's test [38,39].
Where data was not suitable for meta-analysis, a narrative descrip-
tion was performed. Data from included studies were analysed for the
intended primary outcome. Where available, the main data were ana-
lysed per protocol secondary outcomes.
In priori subgroup analysis was done for study region, sample size,
study quality, and sex. Data was insufcient for separate analysis for
comorbidity (history of recent surgery, diabetes, cardiovascular
diseases), severity of CLD, and aetiology of CLD.
Given the signicant variability in pooled effect size and heteroge-
neity between studies a sensitivity analysis was undertaken by restrict-
ing the meta-analysis to (a) year of publication (before 2010 or after
2010), (b) sample size (>10,000 or <10,000), (c) publication type
(abstract only or full text), and (d) information on associated malig-
nancy or missing. The cut-off for year of publication was decided based
the fact that most literature discussing the safety and efcacy of VTE
prophylaxis in CLD were published after 2010 [21,40,41]. The studen-
tized residual tests and leave-one-out analysis were done to determine
the impact of outlier and individual studies on effect size [42]. Baujat
[43] and r diagnostic tests for inuential studies [42] plot were used to
graphically display the results. To further ascertain source of heteroge-
neity a meta-regression analysis was conducted to evaluate the impact of
baseline covariates (age, sex, quality of included studies, sample size,
publication type, study region) on heterogeneity. Covariates which were
signicant in univariable analysis were included in multivariable meta-
regression.
3. Results
After screening of titles and abstracts, 84 studies were selected for
further search, 29 studies [3,12,14,17,21–29,44–59] reporting inci-
dence or prevalence of VTE in CLD were included (Fig. 1). All included
studies were of observational design; 18 were conducted in the USA,
four in Europe and seven in Asia. None of included the studies were done
in the community. Of the included studies, 16 reported incidence, 11
prevalence, and two reported both. In 20 studies included participants
had a diagnosis of cirrhosis, and in nine studies CLD, Enger et al. [57]
included both. Characteristics of included studies are summarised in
Table 1.
3.1. Participants
A total of 19,157,018 participants were included; of which 152,049
(0.79%) had VTE. Mean age of participants was 56.1 years (SD ±4.6),
60.1% were males and 65.3% were Caucasians (Table 1). Of the studies
providing details of comorbidities, 0.10% (n =14,003/13991921) had
malignancy [3,17,21,28,44,47,48,54,56,58,59]; 0.84% (n =11,832/
1404948) hypertension [22,25,57,58], 0.59% (n =82,109/13952904)
diabetes [17,21,25,28,48,54,58,59], and 0.001% (n =89/11225377)
had history of major surgery in the past 3 months [3,17,21,54]. Singh
et al. only included patients with a diagnosis of type 2 diabetes mellitus
and concomitant non-alcohol fatty liver disease (NAFLD) [25].
3.2. Incidence
Eighteen studies reported cumulative incidence per year
[3,17,21,22,24,25,27,44–48,51,52,54,55,57–59], comprising a total of
1,628,164 participants, and 17,424 new VTE events in hospitalised CLD
patients. Three studies reported incidence rate (person-years)
[51,57,58]. Enger et al. [57] reported incidence separately for cirrhosis
and CLD.
The per year cumulative incidence of VTE varied from 0.20% (95%
CI 0.05, 0.51) to 8.59% (95% CI 4.78, 13.99) across the studies.
Weighted average: cumulative incidence of VTE in hospitalised CLD
patients by random effects meta-analysis was 1.07% (95% CI 0.80, 1.38)
per year (Fig. 2a), and incidence rate was 157.15 (95% CI 14.74, 445.29)
per 10,000 person-years (Fig. 2b). There was a statistically signicant
heterogeneity (I
2
99%, p <0.01), and publication bias towards studies
reporting signicant results or increased risk (p <0.0001) (Fig. 3).
Weighted average cumulative incidence of DVT was 0.76% (95% CI
0.41, 1.20) per year, varying from 0.10% (95% CI 0.01, 0.36) to 7.36%
(95% CI 3.86, 12.51) across the nine studies. (SP-Fig. 1).
Weighted average cumulative incidence of pulmonary embolism
(PE) was 0.31 (95% CI 0.06,0.72) per year, varying from 0.10% (95% CI
M. Subhani et al.

Thrombosis Research 215 (2022) 19–29
22
0.001, 0.36) to 1.23% (95% CI 0.15, 4.36) across the seven studies. (SP-
Fig. 2).
3.3. Prevalence
Thirteen studies reported period prevalence per year
[12,14,22,23,26,28,29,48–50,55,56,59], comprising a total of
17,533,466 hospitalised CLD patients had 134,646 pre-existing VTE
events. Period prevalence of VTE in CLD varied from 0.33% (95% CI
0.23, 0.46) to 4.69% (95% CI 2.45, 8.04) per year.
Weighted average period prevalence of VTE in hospitalised patients
with CLD was 1.10% (95% CI 0.85, 1.38) per year. There was statisti-
cally signicant heterogeneity between studies (I
2
100%, p <0.01)
(Fig. 4) and publication bias (p ≤0.04) (Fig. 5).
Weighted average period prevalence of DVT was 1.44% (95% CI
0.79, 2.27) per year, with a range of 0.35 (95% CI 0.14, 0.72) to 4.69
(95% CI 2.45, 8.04) (SP-Fig. 3).
Weighted average period prevalence of PE was 0.24% (95% CI 0.10,
0.44) per year (SP-Fig. 4).
3.4. Cases versus control
Eight studies reporting cumulative per year incidence were included
in cases versus control meta-analysis. Weighted average RR of cumula-
tive per year incidence of VTE in hospitalised participants with CLD
compared to those without CLD was 1.55 (95% CI 0.98, 2.45) (SP-
Fig. 5a).
Three studies reporting incident rate were included in cases versus
control meta-analysis. Weighted average RR of incidence rate of VTE in
chronic liver disease to those without CLD was 2.11 (95% CI 1.35, 3.31)
(SP-Fig. 5b).
Two studies [21,54] comprising 1644 participants, provided details
on VTE events in hospitalised patients with CLD on VTE pharmacolog-
ical prophylaxis (n =441) versus those not on VTE pharmacological
prophylaxis (n =1203). 1.13% (n =5/441) on VTE prophylaxis had a
VTE event compared to 3.08% not on VTE prophylaxis. On conducting
xed effects meta-analysis, hospitalised patients with CLD not on VTE
prophylaxis were 2.78 times (RR 2.78, 95% CI 1.11, 6.98) more likely to
have VTE during hospital stay compared to those on VTE prophylaxis.
Heterogeneity between studies was not signicant (I
2
0%, p =0.67).
Data was insufcient to conduct meta-analysis on incidence and
prevalence of VTE based on study setting (community versus in-
hospital).
3.5. Subgroup analysis
A summary of subgroup analysis is provided in Table 2.
3.5.1. Study region
Weighted average cumulative incidence of VTE in studies under-
taken in Europe/USA (n =14) was 1.12% (95% CI 0.82, 1.47) per year,
and in studies undertaken in Asia (n =4) was 0.89% (95% CI 0.36, 1/61)
per year. Weighted average period prevalence of VTE in Europe/USA
was 1.04% (95% CI 0.77, 1.35), and in Asia was 1.39% (95% CI 0.81,
2.11) per year. Difference in incidence or prevalence of VTE between
Europe/USA and Asia was non-signicant (p =0.58 and p =0.21,
respectively).
Records identified through
database search
(n =13920)
Additional record identified
through other sources
(n = 30)
Records screened after
duplicate removed
(n =9018)
Records excluded
(n = 8934)
Total record screened (n=13950)
Records assessed for
eligibility
(n = 84)
Records excluded, with reasons
(n=55)
Duplicate data (n=3)
Insufficient data (n=14)
Literature review (n=2)
Mechanistic studies (n=5)
VTE included Non DVT and PE (n=3)
Only included Specific cohortsa
(n=12)
Due study eligibility criteria not
possible to determine incidence or
prevalence (n=8)
VTE prophylaxis (n=3)
Only included participants who had
VTE (n=5)
Articles included in
synthesis
(n =29)
noitacifitnedI
Screening
Included
Eligibility
Duplicate removed
(n = 4932)
Fig. 1. PRISMA ow diagram for
studies selection.
(VTE-venous thromboembolism, DVT-
deep vein thrombosis, PE-pulmonary
embolism, HCV-hepatitis c virus, HBV,
hepatitis b virus, ITU-intensive care
unit).
a
Post trauma, post knee arthroplasty,
post hip replacement, post infection,
over age of 65 years, patient dies of PE,
patient on chemotherapy, patient with
any alcohol related health condition,
ITU only patients, Patient with HCV/
HBV infection and on treatment.
M. Subhani et al.

Thrombosis Research 215 (2022) 19–29
23
3.5.2. Sample size
Weighted average cumulative incidence of VTE in studies of sample
size greater than 10,000 was 0.74% (95% CI 0.41, 1.15) per year, and in
studies recruited less than 10,000 participants was 1.33% (0.95, 1.76)
per year. Weighted average period prevalence of VTE in studies of
cohort size greater than 10,000 was 1.04% (95% CI 0.76, 1.36) per year,
and in studies recruited less than 10,000 participants was 1.29% (95%
CI 0.80, 1.88) per year. Weighted average cumulative incidence of VTE
was signicantly (p =0.02) lower in studied of cohort size greater than
10,000, whereas there was no signicant (p0.29) difference in preva-
lence between subgroups.
3.5.3. Study quality
Weighted average cumulative incidence of VTE in studies of low
quality was 0.70% (95% CI 0.26, 1.33), in medium quality was 2.18%
(95% CI 1.18, 3.45), and in high quality study was 1.14% (95% CI 0.61,
1.83) per year. Difference between subgroups based on study quality
was statistically signicant (p =0.03). Weighted average period prev-
alence of VTE in studies of low quality was 1.15% (95% CI 0.86, 1.47)
per year. Data was insufcient to calculate pooled period prevalence for
high and medium quality studies.
3.5.4. Sex
Weighted average cumulative incidence of VTE in CLD in female
participants was 4.99% (95% CI 1.84, 9.35), and in male participants
was 4.60% (95% CI 3.02, 6.47) per year. As studies included to pool
cumulative incidence based on sex had reported higher risk incidence of
VTE in CLD hence the effect size was higher.
Weighted average period prevalence of VTE in CLD in female par-
ticipants was 1.08% (95% CI 0.38, 2.07) and in male participants was
1.25% (95% CI 0.53, 2.23) per year. Difference in incidence or preva-
lence based on sex was non-signicant (p =0.30, p =0.87).
3.6. Sensitivity analysis
The results of sensitivity analysis are summarised in Table 3.
Despite undertaking multiple restricted analysis based on; studies
published up to 2010 or after 2010, studies with a sample size >10,000
or <10,000, abstract only publications or full text publications, and
included studies provided information on associated malignancy or
missing information. It was not possible to remove residual
heterogeneity.
3.6.1. Outlier studies
Studentized residual tests, Buajat plot and r diagnostic tests for
inuential studies conrmed three studies [3,21,25] signicantly
inuenced the incidence meta-analysis results (SP-Fig. 6), and two
studies [28,48] prevalence meta-analysis results (SP-Fig. 7).
For incidence, on excluding Singh et al. (2019) [25] from meta-
analysis, weighted average cumulative incidence per year of VTE in
hospitalised CLD patients was 0.92% (95% CI 0.67, 1.21) (SP-Fig. 8a).
On excluding Singh et al. (2019) and Bogari et al. (2014) [21,25]
weighted average cumulative per year incidence of VTE was 0.85%
(95% CI 0.61, 0.97) (SP-Fig. 8b). On excluding all three studies
Table 1
Characteristics of included studies.
Study ID Country Recruitment period Cohorts Incidence/Period prevalence Events
a
(n)/Cohort
b
(n) Sex
(male)
Age
(years)
Ethnicity
(white)
Northup 2006 USA 1993–2001 Cirrhotic Cum incidence
c
113/21000 – – –
Gracia-Fuster 2008 Spain 1992–2007 Cirrhotic Cum incidence 17/2074 – – –
Gulley 2008 USA 1995–2005 Cirrhotic Cum incidence 18/963 655 50.5 578
Lizarraga 2010 USA 2004–2008 Cirrhotic Cum incidence 108/14790 – – –
Dabbagh 2010 USA 2000–2007 CLD Cum incidence 12/190 121 50.7 –
Lesmana 2010 Indonesia 2004–2007 Cirrhotic Period prevalence
c
12/256 164 60.5 –
Gagan 2010* USA 2007–2007 Cirrhotic Period prevalence 2915/560503 – – –
Wu 2010 USA 1998–2006 Cirrhotic Period prevalence 5288/649879 397,926 57.9 433,002
Aldawood 2011 KSA 2009–2009 Cirrhotic Cum incidence 6/226 140 63 0
Ali 2011 USA 2005–2005 Cirrhotic Period prevalence 8248/449799 275,079 – 230,536
Saleh 2011 USA 1979–2006 CLD Period prevalence 72,000/9492000 5,678,000 56 6,250,240
Ahmed 2012* USA 2000–2009 CLD Incidence rate
d
149/47391 28,908 – –
Girleanu 2012 Romania 2010–2011 Cirrhotic Cum incidence 31/3108 – – –
Kohsaka 2012* Japan – Cirrhotic Cum incidence 10/719 215 58.9 0
Barclay 2013 USA 2008–2011 CLD Cum incidence 12/1518 1074 49.8 –
Walsh 2013 USA 2006–2010 CLD Cum incidence 17/2606 – – –
Period prevalence 27/2606 – – –
Ponziani 2013* Italy 1982–2012 Cirrhotic Period prevalence 34/10359 – – –
Bogari 2014 USA 2010–2013 CLD Cum incidence 14/163 106 54 83
Enger 2014 USA 2000–2006 Cirrhotic Incidence rate 76/15158 9102 56.5 6043
CLD (HCV) 68/22733 14,191 49 9983
Ng 2015 Taiwan 2007–2010 Cirrhotic Incidence rate 26/2779 1836 59 0
Yang 2015 Singapore 2004–2011 CLD Period prevalence 102/6372 2288 53.4 0
Zang 2016 China 2011–2013 Cirrhotic Cum incidence 4/2006 1330 56.2 0
Cirrhotic Period prevalence 9/2006
Tak 2017* India 2016–2017 Cirrhotic Period prevalence 6/365 58.5 296 0
Barba 2018 Spain 2005–2014 CLD Cum incidence 5623/324076 224,359 65.2
Kasarala 2018* USA 2005–2014 Cirrhotic Period prevalence 14,422/1030164 – – –
Singh 2019 USA 2000–2015 CLD (NAFLD) Cum incidence 71/1295 454 55.2 1102
Greenberg 2019* USA 2003–2014 Cirrhotic Cum incidence 11,049/1165369 – – –
Yassine 2020* USA 2015–2019 Cirrhotic Period prevalence 5179/157400 86,220 124,020
Elkafrawy 2020* USA 2004–2014 Cirrhotic Period prevalence 26,404/5171757 – – –
Mean age, n-number, CLD-chronic liver disease, HCV-hepatitis c virus, NAFLD-non-alcoholic fatty liver disease, KSA- Kingdom of Saudi Arabia.
All included studies were of observational design and done in hospital setting. Sing et al.; (2019) did not specify the study setting.
a
Events- number of cases had venous thromboembolism (Deep vein thrombosis (DVT) and/or Pulmonary embolism (PE)).
b
Cohort -total number of participants.
c
Cumulative incidence and period prevalence per year.
d
Incidence rate (person years).
*
Abstract only publications.
M. Subhani et al.

Thrombosis Research 215 (2022) 19–29
24
[3,21,25] weighted average cumulative per year incidence of VTE was
0.79% (95% CI 0.55, 1.06) (SP-Fig. 8c). The residual heterogeneity
remained signicant.
For prevalence meta-analysis, on excluding Yassine et al. (2020) [28]
weighted average period prevalence was 0.93% (95% CI 0.72, 1.16) per
year (SP-Fig. 9a). On excluding both studies [28,48] weighted average
period prevalence was 0.88% (95% CI 0.68, 1.11) per year (SP-Fig. 9b).
The residual heterogeneity remained signicant.
3.7. Meta-regression
For incidence: on univariable meta-regression analysis, quality of
included study (low, medium, high), and sample size (greater than
10,000 versus less than 10,000) signicantly (p =0.03, p =0.02,
respectively) inuenced the results of meta-analysis. The association for
age, sex, publication type and study region (Europe and USA versus
Asia) was non-signicant (p =0.90, p =0.88, p =0.303, p =0.57,
respectively). On multivariable meta-regression analysis age, quality,
and sample size of included studies accounted for (R
2
) 67.40% of het-
erogeneity (p <0.001) whereas on including age, quality, sample size,
and study region into model accounted (R
2
) 78.5% of heterogeneity (p
<0.001). Quality and sample size combined were unable to account for
any residual heterogeneity.
For prevalence: on univariable meta-regression analysis for all
covariates was non-signicant (Table 4).
3.8. Risk of bias assessment
Quality assessment stratied eight studies as high quality, ve as
medium, and sixteen as low quality. The main area of concerns were
inconsistencies and errors in reported data. (SP-Table 1).
a) Forest plot for cumulave per year incidence of VTE in chronic liver disease
b) Forest plot for incidence rate of VTE in chronic liver disease
Fig. 2. Forest plots for incidence metanalysis a) Cumulative per year incidence of venous thromboembolism in chronic liver disease b) Incidence rate of venous
thromboembolism in chronic liver disease.
a) Funnel plot for publicaon bias for incidence metanalysis
b) Contour enhanced funnel plot for publicaon bias for incidence metanalysis
Fig. 3. Publication bias for incidence metanalysis a) Funnel plot b) Contour
enhanced funnel plot.
M. Subhani et al.

Thrombosis Research 215 (2022) 19–29
25
4. Discussion
This systematic review conrms VTE risk is signicant in hospital-
ised patients with CLD. Twenty-nine included studies summarised the
epidemiology of VTE in over 19 million CLD participants, spanning 30
years of research across 3 continents. Hospitalised CLD patients were at
double the risk (RR 2.11, 95% CI 1.35, 3.31) of VTE compared to those
without CLD. Meta-analysis estimates that for every 1000 hospitalised
CLD patients, 10 will develop a new, and 11 a pre-existing diagnosis of
VTE per year. The incidence rate was 157.15 per 10,000 person-years.
Historically, it has been argued that CLD does not increase the risk of
Fig. 4. Forest plot for period prevalence (per year) of venous thromboembolism in chronic liver disease.
a) Funnel plot for publicaon bias for prevalence metanalysis
c) Contour enhanced funnel plot for prevalence metanalysis
Fig. 5. Publication bias for prevalence metanalysis a) Funnel plot b) Contour
enhanced funnel plot.
Table 2
Subgroup analysis.
Subgroup Studies Cumulative incidence
(%) (per year)
Subgroup Heterogeneity
P
a
I
2
(%) p
Study region
Europe,
USA
14 1.12(0.92, 1.47) 0.58 99.0 <0.01
Asia 4 0.89 (0.36, 1.61) 87.0 <0.01
Sample size
>10,000 5 0.74 (0.41, 1.15) 0.02 100.0 <0.01
<10,000 13 1.33 (0.95, 1.76) 1.0 <0.01
Study quality
Low 6 0.70 (0.26, 1.33) 0.03 98.0 <0.01
Medium 5 2.18 (1.18, 3.45) 98.0 <0.01
High 7 1.14 (0.61, 1.83) 99.0 <0.01
Sex
b
Female 2 4.99 (1.84, 9.35) 0.30 0.0 0.96
Male 2 4.60 (3.02, 6.47) 83.0 0.02
Subgroup Studies Period prevalence (%)
(per year)
P
a
I
2
(%) p
Study region
Europe,
USA
9 1.04 (0.77, 1.35) 0.21 100.0 <0.01
Asia 4 1.39 (0.81, 2.11) 91.0 <0.01
Sample size
>10,000 8 1.04 (0.76, 1.36) 0.29 100.0 <0.01
<10,000 5 1.29 (0.80, 1.88) 100.0 <0.01
Study quality
Low 11 1.15 (0.86, 1.47) 0.72 100 <0.01
Medium 1 1.04 (0.29, 2.23)
High 1 0.81 (0.22, 1.79)
Sex
Female 4 1.08 (0.38, 2.07) 0.87 100.0 <0.01
Male 4 1.25 (0.53, 2.23) 100.0 <0.01
% (95% condence interval), (VTE-venous thromboembolism).
Europe n =4, United states (USA) n =18.
a
p for signicance of difference between subgroups.
b
Fixed effects metanalysis.
M. Subhani et al.

Thrombosis Research 215 (2022) 19–29
26
VTE and CLD patients may be less likely to develop VTE [60,61]. Our
ndings are consistent with most recent nationwide studies conrming
CLD considerably increases the risk of VTE [15,58]. Estimated cumu-
lative incidence of VTE in CLD was higher than that reported by Qi et al.
(2014), but lower than the values reported by Ambrosino et al. [2,20].
In-hospital VTE in CLD constitutes up to a tenth of the burden of in-
hospital VTE [17,18]. Moreover, hospitalisation, periods of immobility
and hepatic synthetic dysfunction signicantly increase the risk of VTE
in cirrhosis [12,62].
Our results show that hospitalised CLD patients without VTE phar-
macological prophylaxis were twice (RR 2.78, 95% CI 1.11, 6.98) as
likely to develop VTE compared to those receiving prophylaxis. Special
attention should be paid while generalising these ndings as the search
strategy was not customised to search for studies discussing the role of
VTE prophylaxis in CLD. Moreover, there is paucity in available evi-
dence on the effectiveness of VTE prophylaxis in CLD, and the results are
conicting, although most researchers agrees it does not increase the
risk of bleeding [21,40,54,63].
Subgroup analysis demonstrated that though there was no signicant
difference in the cumulative incidence of VTE in hospitalised CLD pa-
tients across Europe, USA, and Asia, the period prevalence was signi-
cantly higher in Asia. Studies with a smaller cohort (<10,000) reported
higher estimated cumulative incidence (1.33/100) compared to studies
with a larger cohort (0.74/100). There was no signicant difference in
estimated cumulative incidence and period prevalence of VTE in CLD
between male and female participants. Subgroup analysis ndings were
consistent with previous systematic reviews [20].
Pulmonary embolism (PE) compared to deep vein thrombosis carries
signicantly higher risk of mortality in cirrhosis [15]. 30-day mortality
in hospitalised patients with PE and cirrhosis has been reported as high
as 35%, compared to 7% in patients with DVT and cirrhosis. PE is among
the leading causes of preventable death in hospitalised patients [64].
This review conrms cirrhotic patients are at increased risk of both DVT
and PE. Clinicians treating such patients need to be alert to the VTE risk
in liver disease to prevent avoidable deaths.
We followed a robust methodology in study selection criteria and
data analysis to strengthen our ndings. Three studies [15,65,66]
included in previous systematic reviews [2,20] were excluded due to
missing data or due to inclusion of very specic populations, such as,
only trauma patients or patients with international normalise ratio (INR)
>1.4. For the cases versus control analysis, we paid special attention to
only include studies with similar control conditions to minimise the
impact of confounders and bias. Furthermore, we included nine addi-
tional studies [21–29], comprising over seven million participants and
sixty thousand VTE events, that were not included in any of previous
reviews [2,20]. Two of these studies [21,24] were of high, and one [25]
of medium quality.
Signicant heterogeneity between studies was noted in most of our
meta-analyses, which is a limitation, and hence a random effects method
was used. Subgroup analyses and meta regression were undertaken to
ascertain the sources of heterogeneity. Cohort size, study region and
Table 3
Sensitivity analysis.
Cumulative incidence per 100 patients per year
No. of
studies
Cumulative
incidence
(95% CI)
I
2
(%)
Tau
[2]
p
(heterogeneity)
Year of publication
Before 2010 5 1.15 (0.62,
1.82)
91.0 0.001 <0.01
After 2010 13 1.06 (0.75,
1.41)
99.0 0.001 <0.01
Sample size
>10,000 5 0.74 (0.41,
1.15)
100.0 0.001 <0.01
<10,000 13 1.33 (0.95,
1.76)
95.0 0.001 <0.01
Publication type
Full text 15 1.26 (0.84,
1.76)
99.0 0.002 <0.01
Abstract only 3 0.78 (0.18,
1.77)
99.0 0.002 <0.01
Malignancy
Information
provided
8 0.58 (0.37,
0.84)
91.0 0.003 <0.01
Information
missing
11 1.38 (0.99,
1.83)
99.0 0.001 <0.01
Period prevalence per year
No. of
Studies
Period
prevalence
(95% CI)
I
2
(%)
Tau
[2]
p
(heterogeneity)
Year of publication
Before 2010 3 0.97 (0.47,
1.63)
100.0 0.001 <0.01
After 2010 10 1.14 (0.85,
1.48)
100.0 0.001 <0.01
Sample size
>10,000 8 1.04 (0.76,
1.36)
100.0 0.001 <0.01
<10,000 5 1.29 (0.80,
1.88)
88.0 0.001 <0.01
Publication type
Full text 7 1.21 (0.74,
1.80)
100.0 0.001 <0.01
Abstract only 6 1.06 (0.59,
1.65)
100.0 0.001 <0.01
Malignancy
Information
provided
5 1.61 (0.36,
3.68)
99.0 0.005 <0.01
Information
missing
8 0.95 (0.70,
1.22)
100.0 0.003 <0.01
Table 4
Meta-regression analysis.
Variables I
2
tau Tests of moderator p (heterogeneity)
Coefcient p
Cumulative incidence meta-regression
Age 96.00% 0.043 0.013 0.908 <0.001
Sex
Male (%) 96.80% 0.036 0.019 0.889 <0.001
Female (%) 96.90% 0.036 0.100 0.751 <0.001
Quality of studies 98.70% 0.038 6.981 0.031 <0.001
Sample size 99.10% 0.027 5.502 0.020 <0.001
Publication type 86.70% 0.038 1.058 0.303 <0.001
Study region 99.20% 0.027 0.308 0.578 <0.001
Period prevalence meta-regression
Age 94.40% 0.017 0.756 0.385 <0.001
Sex
Male (%) 99.88% 0.041 1.22 0.269 <0.001
Female (%) 99.88% 0.041 1.19 0.274 <0.001
Quality of studies 99.95% 0.022 0.645 0.724 <0.001
Sample size 99.94% 0.021 1.101 0.293 <0.001
Publication type 99.94% 0.031 0.192 0.661 <0.001
Study region 99.94% 0.021 1.592 0.207 <0.001
Liver disease included cirrhosis or chronic liver disease unspecied of unspeci-
ed stage.
Publication type included full text and abstract only.
M. Subhani et al.

Thrombosis Research 215 (2022) 19–29
27
type of CLD (cirrhosis vs unclassied) were found to be the sources of
signicant heterogeneity. Secondly, as per protocol, we aimed to
compare epidemiology of VTE in CLD in community versus in-hospital
settings. Despite adopting an inclusive search strategy, none of the
included studies were done in a community setting. This is a recurring
issue that has been noted in the previous systematic reviews [2,18,20].
Furthermore, a signicant proportion of CLD patients remain undiag-
nosed until their rst hospital admission [67], which makes it harder to
determine the epidemiology of VTE in CLD in community settings.
Thirdly, none of the included studies were from Africa, Australia, or
South America, which limits the generalisability of this review to these
populations. All studies included were of observational design, which
inherently increases the risk of bias, such as selection and information
bias, which might have under- or overestimated the results [68]. To
address any potential source of uncertainty in the results a sensitivity
analysis involving subgroup analyses and meta-regression was under-
taken. Several studies included in the current meta-analysis had par-
ticipants with a history of malignancy. As malignancies can signicantly
increase the risk of VTE, this may have inuenced the estimated inci-
dence of VTE in liver disease [69]. Considering malignancy is a pro-
thrombotic condition [70] a meta-analysis on studies providing or
missing details on associated malignancies was conducted. Only seven
studies reporting cumulative incidence, and ve studies reporting
period prevalence provided information on associated malignancies.
Due to limited information on associated malignancies, and with none of
the included studies investigating incidence or prevalence of VTE in CLD
in the absence or presence of malignancy, it was not possible to precisely
ascertain the effect of malignancy in CLD on risk of VTE. HCC inde-
pendent of cirrhosis or stage of liver disease increases the risk of venous
thromboembolism [71,72]. Whereas the focus of the current systematic
review was to evaluate the increased risk of VTE due to the systemic
effect of cirrhosis. In this view, the studies where the cohort was purely
consistent of patients with HCC were excluded. Moreover, most of the
included studies were missing details on other comorbidities which
could increase the risk of VTE. Keeping this in view a dedicated cases
(patient with CLD and VTE) versus control (patient without CLD but had
VTE) analysis was conducted. Lastly, due to insufcient data it was not
possible to ascertain risk of VTE based on underlying aetiology or stage
of CLD. As cirrhosis and its association with portal vein thrombosis
(PVT) is well described and studied, factors affecting PVT are in part
explained by changes in portal circulation [73]. Increased intrahepatic
vascular resistance and sluggish blood ow in portal vein in cirrhosis are
two important contributory factors in pathophysiology of PVT [73].
While focus of the current systematic review was to study the systemic
effect of cirrhosis, hence we excluded PVT from analysis. Lastly, due to
insufcient data, it was not possible to conducted separate subgroup
analysis to report the risk of VTE based on severity (MELD or Child Pugh
score) of liver disease.
The ndings of this review have clinical implications. A VTE event
complicating CLD is likely to increase the risk of morbidity, mortality
and prolong hospital stay [14,17]. Healthcare services are already
spending over 1% of their budget in treating CLD, any further in hospital
events signicantly increase the economic burden [74,75]. Despite a
signicant risk of VTE in CLD, only 26.8% of hospitalised patients were
on pharmacological prophylaxis. A recent literature review reported
over 76% of hospitalised patients with cirrhosis did not receive either
pharmacological or mechanical VTE prophylaxis [76]. It highlights the
importance of thrombosis in hospitalised patients with cirrhosis and
strongly suggests that universal VTE prophylaxis should be prescribed.
This will require a “cultural shift” given the ingrained views of doctors
caring for CLD patients that the major risk is haemorrhage. Studies have
shown pharmacological VTE prophylaxis to be safe in CLD and reduce
the incidence of VTE [41,77]. The most recent Baveno consensus
(Baveno VII), European Association for the Study of the Liver (EASL) and
American Gastroenterology Association (AGA) guidelines emphasise on
personalising the care in portal hypertension and advocate for
thromboprophylaxis to reduce the risk of venous thrombosis in cirrhosis
[8,78,79].
In conclusion, our results show hospitalised patients with underlying
CLD may exhibit an increased risk of developing VTE (DVT and/or PE);
for every 1000 hospitalised patients with CLD, 10 will develop a new,
and 11 will have a pre-existing diagnosis of VTE per year. An auto-
anticoagulatory state in cirrhosis does not always protect against
thrombosis. It is beyond the scope of this study to determine current
practice but, anecdotally, prescribing of low molecular weight heparin
(LMWH) is often withheld in cirrhotic patients. Educating healthcare
professionals providing direct care to these patients will be a key driver
to inuence clinical practice.
CRediT authorship contribution statement
Mohsan Subhani: Contributed to protocol writing, scoping search,
drafting nal Search Strategy, literature search, abstract screening, data
extraction including critical appraisal, meta-analysis, report writing,
and proofreading nal manuscript.
Abhishek Sheth: Contributed to data extraction including critical
appraisal, meta-analysis, report writing, and proofreading nal
manuscript.
Jamal Ahmed: Contributed to data extraction including critical
appraisal, report writing and proofreading nal manuscript.
Pramudi Wijayasiri Contributed as the additional reviewer in
literature search, abstract screening, data extraction, and proof readying
manuscript.
Syed Anjum Gardezi: Contributed as the second reviewer in liter-
ature search, abstract screening, data extraction including critical
appraisal, meta-analysis, report writing, and proofreading nal
manuscript.
Doyo Enki: Statistical support for meta-analysis, proof-reading
results.
Joanne R Morling: Contributed to nalising research question,
proofreading, and nalising the manuscript.
Guruprasad Aithal: Contributed to nalising research question,
proofreading, and nalising the manuscript.
Stephen D Ryder: Contributed to nalising research question,
proofreading, and nalising the manuscript.
Dr. Aloysious D Aravinthan: Senior Author: Contributed to nal-
ising research question, proof-reading, developing the search strategy
and protocol, statistical support, and nalising the manuscript. He also
acted as 3rd reviewer in case of disagreement in primary reviewers.
Financial support
JRM (co-author) receives salary support from a Medical Research
Council Clinician Scientist Fellowship [grant number MR/P008348/1].
Prospero Registration number: CRD42021239117.
Data availability
The data that support the ndings of this study are available on
request from the corresponding author.
Declaration of competing interest
No declared conict interest from any authors.
Acknowledgment
Alison Ashmore, senior research librarian (Nottingham University
Libraries) contributed to nalising the search strategy.
M. Subhani et al.

Thrombosis Research 215 (2022) 19–29
28
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.thromres.2022.05.004.
References
[1] T. Khoury, A.R. Ayman, J. Cohen, S. Daher, C. Shmuel, M. Mizrahi, The complex
role of anticoagulation in cirrhosis: an updated review of where we are and where
we are going, Digestion 93 (2) (2016) 149–159.
[2] P. Ambrosino, L. Tarantino, G. Di Minno, et al., The risk of venous
thromboembolism in patients with cirrhosis. A systematic review and meta-
analysis, Thromb. Haemost. 117 (1) (2017) 139–148.
[3] O. Dabbagh, A. Oza, S. Prakash, R. Sunna, T.M. Saettele, Coagulopathy does not
protect against venous thromboembolism in hospitalized patients with chronic
liver disease, Chest 137 (5) (2010) 1145–1149.
[4] L. Amitrano, M.A. Guardascione, V. Brancaccio, A. Balzano, Coagulation disorders
in liver disease, Semin. Liver Dis. 22 (1) (2002) 83–96.
[5] A. Tripodi, Hemostasis abnormalities in cirrhosis, Curr. Opin. Hematol. 22 (5)
(2015) 406–412.
[6] A. Tripodi, P.M. Mannucci, The coagulopathy of chronic liver disease, N. Engl. J.
Med. 365 (2) (2011) 147–156.
[7] S. Caldwell, L.E. Carlini, Coagulation homeostasis in liver disease, Clin. Liver Dis.
16 (4) (2020) 137–141.
[8] R.S. O'Shea, P. Davitkov, C.W. Ko, et al., AGA clinical practice guideline on the
management of coagulation disorders in patients with cirrhosis, Gastroenterology
161 (5) (2021) 1615–1627, e1611.
[9] A.A. Mokdad, A.D. Lopez, S. Shahraz, et al., Liver cirrhosis mortality in 187
countries between 1980 and 2010: a systematic analysis, BMC Med. 12 (1) (2014)
145.
[10] J.M. Paik, P. Golabi, Y. Younossi, A. Mishra, Z.M. Younossi, Changes in the global
burden of chronic liver diseases from 2012 to 2017: the growing impact of NAFLD,
Hepatology 72 (5) (2020) 1605–1616.
[11] S. Nusrat, M.S. Khan, J. Fazili, M.F. Madhoun, Cirrhosis and its complications:
evidence based treatment, World J. Gastroenterol. 20 (18) (2014) 5442–5460.
[12] M. Ali, A.N. Ananthakrishnan, E.L. McGinley, K. Saeian, Deep vein thrombosis and
pulmonary embolism in hospitalized patients with cirrhosis: a nationwide analysis,
Dig. Dis. Sci. 56 (7) (2011) 2152–2159.
[13] M.L. Schmidt, A.S. Barritt, E.S. Orman, P.H. Hayashi, Decreasing mortality among
patients hospitalized with cirrhosis in the United States from 2002 through 2010,
Gastroenterology 148 (5) (2015) 967–977, e962.
[14] H. Wu, G.C. Nguyen, Liver cirrhosis is associated with venous thromboembolism
among hospitalized patients in a nationwide US study, Clin. Gastroenterol.
Hepatol. 8 (9) (2010) 800–805.
[15] K.K. Søgaard, E. Horv´
ath-Puh´
o, H. Grønbaek, P. Jepsen, H. Vilstrup, H.T. Sørensen,
Risk of venous thromboembolism in patients with liver disease: a nationwide
population-based case-control study, Am. J. Gastroenterol. 104 (1) (2009) 96–101.
[16] J.A. Heit, L.J. Melton III, C.M. Lohse, et al., Incidence of venous thromboembolism
in hospitalized patients vs community residents, Mayo Clin. Proc. 76 (11) (2001)
1102–1110.
[17] A. Aldawood, Y. Arabi, A. Aljumah, et al., The incidence of venous
thromboembolism and practice of deep venous thrombosis prophylaxis in
hospitalized cirrhotic patients, Thromb. J. 9 (1) (2011), 1-1.
[18] A. Aggarwal, K. Puri, S. Liangpunsakul, Deep vein thrombosis and pulmonary
embolism in cirrhotic patients: systematic review, World J. Gastroenterol. 20 (19)
(2014) 5737–5745.
[19] M. Buresi, R. Hull, C.S. Cofn, Venous thromboembolism in cirrhosis: a review of
the literature, Can. J. Gastroenterol. 26 (12) (2012) 905–908.
[20] X. Qi, W. Ren, X. Guo, D. Fan, Epidemiology of venous thromboembolism in
patients with liver diseases: a systematic review and meta-analysis, Intern. Emerg.
Med. 10 (2) (2015) 205–217.
[21] H. Bogari, A.E. Patanwala, R. Cosgrove, M. Katz, Risk-assessment and
pharmacological prophylaxis of venous thromboembolism in hospitalized patients
with chronic liver disease, Thromb. Res. 134 (6) (2014) 1220–1223.
[22] X. Zhang, X. Qi, V. De Stefano, et al., Epidemiology, risk factors, and in-hospital
mortality of venous thromboembolism in liver cirrhosis: a single-center
retrospective observational study, Med. Sci. Monit. 22 (2016) 969–976.
[23] V.G. Tak, Gaurav, Sandeep Nijhawan, Vasudha Goyal, Prevalence of deep vein
thrombosis in patients with advanced liver cirrhosis: rare entity in a tertiary care
centre, Indian J. Gastroenterol. 36 (1) (2017) 1–105.
[24] R. Barba, A. Gonzalvez-Gasch, D. Joya Seijo, et al., Venous thromboembolism in
patients with liver diseases, J. Thromb. Haemost. 16 (10) (2018) 2003–2007.
[25] A. Singh, A. ElBoraie, N. Lan, M. Gupta, Factors associated with deep vein
thrombosis in type 2 diabetics with biopsy-proven non-alcoholic fatty liver disease,
SN Compr. Clin. Med. 2 (1) (2020) 32–41.
[26] G.R. Kasarala, H.L. Tillmann, Tu1567 - cirrhosis per se is not associated with
increased risk of venous thromboembolism, Gastroenterology 154 (6, Supplement
1) (2018). S-1253-S-1254.
[27] I. Greenberg, L.D. Quintero, A. Tafur, 1086 Venous thromboembolism trends and
impact in patients with cirrhosis: analysis of the national inpatient sample database
2003-2014, Off. J. Am. Coll. Gastroenterol. (2019) 114.
[28] A.A. Yassine, M. Abureesh, M. Alkhayyat, L. Deeb, S1081 Predictors of Deep
Venous Thrombosis in the Hospitalized Patients With Liver Cirrhosis in the U.S.
Ofcial Journal of the American College of Gastroenterology | ACG 115, 2020.
S548.
[29] A.S. Elkafrawy, Muhammad, Wael Al-Yaman, Carlos Romero-Marrero,
Mohamed Ahmed, Temporal trends and predictors of venous thromboembolism in
hospitalized patients with liver cirrhosis: a nationwide analysis, Hepatology 72
(S1) (2020), 1160A-1225A.
[30] Z. Munn, S. Moola, K. Lisy, D. Riitano, C. Tufanaru, Methodological guidance for
systematic reviews of observational epidemiological studies reporting prevalence
and cumulative incidence data, Int. J. Evid. Based Healthc. 13 (3) (2015) 147–153.
[31] B.S. Brooke, T.A. Schwartz, T.M. Pawlik, MOOSE reporting guidelines for meta-
analyses of observational studies, JAMA Surg. 156 (8) (2021) 787–788.
[32] Z. Munn, C. Stern, E. Aromataris, C. Lockwood, Z. Jordan, What kind of systematic
review should I conduct? A proposed typology and guidance for systematic
reviewers in the medical and health sciences, BMC Med. Res. Methodol. 18 (1)
(2018) 5.
[33] CASP, Critical Appraisal Skills Programme, 2020.
[34] R. DerSimonian, N. Laird, Meta-analysis in clinical trials, Control. Clin. Trials 7 (3)
(1986) 177–188.
[35] F.F. Murray, W.T. John, Transformations related to the angular and the square
root, Ann. Math. Stat. 21 (4) (1950) 607–611.
[36] D. Jackson, J. Bowden, Condence intervals for the between-study variance in
random-effects meta-analysis using generalised heterogeneity statistics: should we
use unequal tails? BMC Med. Res. Methodol. 16 (1) (2016) 118.
[37] B.H. Willis, D. Coomar, M. Baragilly, Tailored meta-analysis: an investigation of
the correlation between the test positive rate and prevalence, J. Clin. Epidemiol.
106 (2019) 1–9.
[38] J.L. Peters, A.J. Sutton, D.R. Jones, K.R. Abrams, L. Rushton, Contour-enhanced
meta-analysis funnel plots help distinguish publication bias from other causes of
asymmetry, J. Clin. Epidemiol. 61 (10) (2008) 991–996.
[39] M. Egger, G. Davey Smith, M. Schneider, C. Minder, Bias in meta-analysis detected
by a simple, graphical test, BMJ (Clinical Research ed) 315 (7109) (1997)
629–634.
[40] J. Yerke, S.R. Bauer, S. Bass, et al., Effectiveness of venous thromboembolism
prophylaxis in patients with liver disease, World J. Hepatol. 11 (4) (2019)
379–390.
[41] C. G´
omez Cuervo, O. Bisbal Pardo, M.A. P´
erez-Jacoiste Asín, Efcacy and safety of
the use of heparin as thromboprophylaxis in patients with liver cirrhosis: a
systematic review and meta-analysis, Thromb. Res. 132 (4) (2013) 414–419.
[42] W. Viechtbauer, M.W. Cheung, Outlier and inuence diagnostics for meta-analysis,
Res. Synth. Methods 1 (2) (2010) 112–125.
[43] B. Baujat, C. Mah´
e, J.P. Pignon, C. Hill, A graphical method for exploring
heterogeneity in meta-analyses: application to a meta-analysis of 65 trials, Stat.
Med. 21 (18) (2002) 2641–2652.
[44] P.G. Northup, M.M. McMahon, A.P. Ruhl, et al., Coagulopathy does not fully
protect hospitalized cirrhosis patients from peripheral venous thromboembolism,
Am. J. Gastroenterol. 101 (7) (2006) 1524–1528, quiz 1680.
[45] M.J. García-Fuster, N. Abdilla, M.J. Fabi´
a, C. Fern´
andez, V. Oliver, M.J. Forner,
Venous thromboembolism and liver cirrhosis, Rev. Esp. Enferm. Dig. 100 (5)
(2008) 259–262.
[46] D. Gulley, E. Teal, A. Suvannasankha, N. Chalasani, S. Liangpunsakul, Deep vein
thrombosis and pulmonary embolism in cirrhosis patients, Dig. Dis. Sci. 53 (11)
(2008) 3012–3017.
[47] W. Anthony Lizarraga, S. Dalia, S.E. Reinert, F.J. Schiffman, Venous thrombosis in
patients with chronic liver disease, Blood Coagul. Fibrinolysis 21 (5) (2010)
431–435.
[48] C.R.A. Lesmana, S. Inggriani, L. Cahyadinata, L.A. Lesmana, Deep vein thrombosis
in patients with advanced liver cirrhosis: a rare condition? Hepatol. Int. 4 (1)
(2010) 433–438.
[49] G. Kumar, N. Kumar, A. Deshmukh, Is cirrhosis protective for venous
thromboembolism? Analysis from a National Inpatient Sample, Chest 138 (4,
Supplement) (2010), 935A.
[50] T. Saleh, F. Matta, F. Alali, P.D. Stein, Venous thromboembolism with chronic liver
disease, Am. J. Med. 124 (1) (2011) 64–68.
[51] S. Ahmed, V. Mehta, N. Vendetti, et al., Risk of venous thromboembolism (VTE) in
patients with chronic hepatitis C (CHC): 200, Pharmacoepidemiol. Drug Saf. 21
(2012).
[52] I. Gîrleanu, A. Trifan, C. Cojocariu, A.M. Sîngeap, C. Sfarti, C. Stanciu, The risk of
thrombotic events in patients with liver cirrhosis, Rev. Med. Chir. Soc. Med. Nat.
Iasi 116 (4) (2012) 991–996.
[53] S. Kohsaka, T. Nagai, M. Yaegashi, K. Fukuda, Pulmonary embolism and deep
venous thrombosis in hospitalized patients with liver cirrhosis, Hepatol. Res. 42 (4)
(2012) 433–434.
[54] S.M. Barclay, M.N. Jeffres, K. Nguyen, T. Nguyen, Evaluation of pharmacologic
prophylaxis for venous thromboembolism in patients with chronic liver disease,
Pharmacotherapy 33 (4) (2013) 375–382.
[55] K.A. Walsh, D.A. Lewis, T.M. Clifford, et al., Risk factors for venous
thromboembolism in patients with chronic liver disease, Ann. Pharmacother. 47
(3) (2013) 333–339.
[56] F. Ponziani, M. Zocco, M. Garcovich, D. Roccarina, A. Gasbarrini, F-34
epidemiology of venous thrombotic events and pulmonary embolism among
hospitalized cirrhotic patients: a single center experience, Dig. Liver Dis. 45 (2013)
S41.
[57] C. Enger, U.M. Forssen, D. Bennett, D. Theodore, S. Shantakumar, A. McAfee,
Thromboembolic events among patients with hepatitis C virus infection and
cirrhosis: a matched-cohort study, Adv. Ther. 31 (8) (2014) 891–903.
M. Subhani et al.

Thrombosis Research 215 (2022) 19–29
29
[58] K.J. Ng, Y.K. Lee, M.Y. Huang, C.Y. Hsu, Y.C. Su, Risks of venous thromboembolism
in patients with liver cirrhosis: a nationwide cohort study in Taiwan, J. Thromb.
Haemost. 13 (2) (2015) 206–213.
[59] Y. Yang, X.Z. Zhang, H.S. Ng, J.C. Fong, L.H. Lee, The effect of chronic liver disease
on venous thromboembolism among medically managed patients in Singapore
General Hospital, Thromb. Res. 136 (3) (2015) 548–551.
[60] J.A. Heit, M.D. Silverstein, D.N. Mohr, T.M. Petterson, W.M. O'Fallon, L.
J. Melton III, Risk factors for deep vein thrombosis and pulmonary embolism: a
population-based case-control study, Arch. Intern. Med. 160 (6) (2000) 809–815.
[61] A.A. Shaheen, D. Tanyingoh, E. Dixon, G. Kaplan, R. Myers, Medical injuries
among hospitalized patients with cirrhosis: a population-based study: 1913,
Hepatology 56 (2012).
[62] J. Edelsberg, M. Hagiwara, C. Taneja, G. Oster, Risk of venous thromboembolism
among hospitalized medically ill patients, Am. J. Health Syst. Pharm. 63 (20 Suppl
6) (2006) S16–S22.
[63] K.J. Moorehead, M.N. Jeffres, S.W. Mueller, A retrospective cohort analysis of
pharmacologic VTE prophylaxis and Padua prediction score in hospitalized
patients with chronic liver disease, J. Pharm. Pract. 30 (1) (2017) 58–63.
[64] G. Agnelli, Prevention of venous thromboembolism in surgical patients,
Circulation. 110 (24 Suppl 1) (2004), Iv4-12.
[65] H.M. Al-Dorzi, H.M. Tamim, A.S. Aldawood, Y.M. Arabi, Venous thromboembolism
in critically ill cirrhotic patients: practices of prophylaxis and incidence,
Thrombosis 2013 (2013), 807526-807526.
[66] P. Talving, T. Lustenberger, O.T. Okoye, et al., The impact of liver cirrhosis on
outcomes in trauma patients: a prospective study, J. Trauma Acute Care Surg. 75
(4) (2013) 699–703.
[67] R. Williams, R. Aspinall, M. Bellis, et al., Addressing liver disease in the UK: a
blueprint for attaining excellence in health care and reducing premature mortality
from lifestyle issues of excess consumption of alcohol, obesity, and viral hepatitis,
Lancet 384 (9958) (2014) 1953–1997.
[68] D.A. Grimes, K.F. Schulz, Bias and causal associations in observational research,
Lancet (London, England) 359 (9302) (2002) 248–252.
[69] A. Mullard, H. Innes, Venous thromboembolism in malignancy, Clin. Med. (Lond.)
14 (5) (2014) 532–534.
[70] G.J. Caine, P.S. Stonelake, G.Y.H. Lip, S.T. Kehoe, The hypercoagulable state of
malignancy: pathogenesis and current debate, Neoplasia 4 (6) (2002) 465–473.
[71] A. Zanetto, E. Campello, C. Bulato, et al., More pronounced hypercoagulable state
and hypobrinolysis in patients with cirrhosis with versus without HCC, Hepatol.
Commun. 5 (12) (2021) 1987–2000.
[72] G.J. Caine, P.S. Stonelake, G.Y. Lip, S.T. Kehoe, The hypercoagulable state of
malignancy: pathogenesis and current debate, Neoplasia 4 (6) (2002) 465–473.
[73] A. Mantaka, A. Augoustaki, E.A. Kouroumalis, D.N. Samonakis, Portal vein
thrombosis in cirrhosis: diagnosis, natural history, and therapeutic challenges,
Ann. Gastroenterol. 31 (3) (2018) 315–329.
[74] W.R. Kim, R.S. Brown Jr., N.A. Terrault, H. El-Serag, Burden of liver disease in the
United States: summary of a workshop, Hepatology 36 (1) (2002) 227–242.
[75] G.W. Neff, C.W. Duncan, E.R. Schiff, The current economic burden of cirrhosis,
Gastroenterol. Hepatol. 7 (10) (2011) 661–671.
[76] M. Buresi, R. Hull, C.S. Cofn, Venous thromboembolism in cirrhosis: a review of
the literature, Can. J. Gastroenterol. 26 (12) (2012) 905–908.
[77] S.M. Barclay, M.N. Jeffres, K. Nguyen, T. Nguyen, Evaluation of pharmacologic
prophylaxis for venous thromboembolism in patients with chronic liver disease,
Pharmacotherapy 33 (4) (2013) 375–382.
[78] European Association for the Study of the Live, EASL Clinical Practice Guidelines
on prevention and management of bleeding and thrombosis in patients with
cirrhosis, J. Hepatol. 76 (5) (2022) 1151–1184, https://doi.org/10.1016/j.
jhep.2021.09.003.
[79] R. de Franchis, J. Bosch, G. Garcia-Tsao, T. Reiberger, C. Ripoll, Baveno VII -
renewing consensus in portal hypertension, J. Hepatol. 76 (4) (2022) 959–974.
M. Subhani et al.