Intern Emerg Med DOI 10.1007/s11739-014-1163-7
CE-ORIGINAL
Epidemiology of venous thromboembolism in patients with liver diseases: a systematic review and meta-analysis Xingshun Qi • Weirong Ren • Xiaozhong Guo Daiming Fan
•
Received: 12 October 2014 / Accepted: 20 November 2014 Ó SIMI 2014
Abstract The risk of venous thromboembolism (VTE) may be increased in patients with liver diseases. A systematic review and meta-analysis were conducted to analyze the epidemiology of VTE in such patients. All relevant studies were searched via the PubMed, EMBASE, and Cochrane Library databases. The incidence and prevalence of VTE were pooled using random-effect models. Subgroup analyses were conducted according to the type of VTE [deep vein thrombosis (DVT), pulmonary embolism (PE)], type of liver disease (liver cirrhosis alone/unclassified liver diseases or non-cirrhotics), region in which the study was performed (USA/Europe/Asia), number of total observed patients with liver diseases ([1,000/\1,000 patients), study quality (high/low), and methods for identifying the cases (ICD codes/clinical charts). Of 4,843 Electronic supplementary material The online version of this article (doi:10.1007/s11739-014-1163-7) contains supplementary material, which is available to authorized users. X. Qi (&) X. Guo (&) Department of Gastroenterology, General Hospital of Shenyang Military Area, No. 83 Wehua Road, Shenyang 110840, China e-mail: [email protected] X. Guo e-mail: [email protected] X. Qi W. Ren D. Fan (&) Xijing Hospital of Digestive Diseases, Fourth Military Medical University, No. 127 Changle West Road, Xi’an 710032, China e-mail: [email protected] W. Ren e-mail: [email protected] W. Ren Department of Digestive Diseases, Sanmenxia Central Hospital, Henan University of Science and Technology, Xiaoshan Road, Sanmenxia 472000, China
papers initially identified, 20 were included. The incidence of VTE varied from 0.33 to 6.32 % in 14 studies with a pooled value of 1 % (95 % confidence interval (CI) 0.7–1.3 %). The pooled incidence of DVT and PE was 0.6 % (95 % CI 0.4–0.8 %) and 0.28 % (95 % CI 0.13–0.49 %), respectively. The prevalence of VTE varied from 0.6 to 4.69 % in six studies with a pooled value of 1.0 % (95 % CI 0.7–1.2 %). The pooled prevalence of DVT and PE was 0.7 % (95 % CI 0.6–0.9 %) and 0.36 % (95 % CI 0.13–0.7 %), respectively. The heterogeneity was statistically significant in the main and subgroup metaanalyses. In conclusion, about 1 % of patients with liver diseases develop or are diagnosed with VTE during their hospitalizations. However, the epidemiological data are very heterogeneous among studies. Keywords Deep vein thrombosis Pulmonary embolism Prevalence Incidence Liver cirrhosis Abbreviations CI Confidence interval VTE Venous thromboembolism DVT Deep vein thrombosis PE Pulmonary embolism
Introduction Nowadays, the emerging evidence confirms a new paradigm shift from the bleeding tendency to the rebalanced, but unstable, hemostasis in patients with liver diseases [1– 4]. Epidemiological studies have also shown that the risk of venous thromboembolism (VTE) may be increased in these
123
Intern Emerg Med
patients [5, 6]. Laboratory research studies further demonstrate that both increased levels of factor VIII and remarkably decreased levels of protein C contribute to the presence of hypercoagulability of plasma from patients with liver diseases [7–9]. These findings suggest that the patients with liver diseases are not protected against the development of VTE, despite often having a prolonged internationalized normalized ratio, and being considered to be in an auto-anticoagulation status [10, 11]. However, despite the significance of VTE in patients with liver diseases, and even with an elevated internationalized normalized ratio gradually having been recognized, no obvious improvement has been made in translating the information into clinical practice due to the absence of clear recommendations from practice guidelines and fear of bleeding risks [12]. VTE, including deep vein thrombosis (DVT) and pulmonary embolism (PE), represent major global public health problems [13]. Recently, several studies report upon the incidence or prevalence of VTE in patients with liver diseases [6, 10, 11, 14–18]. The present study aims to systematically collect the epidemiological data regarding VTE in patients with liver diseases, and to combine the individual data into a generalizable result. The sources of heterogeneity among different studies are also explored.
cirrhotics. Additionally, several exclusion criteria were as follows: 1. 2. 3. 4. 5.
6.
Methods
Reviews, meta-analyses, comments, and editorials were excluded. Case reports (number of patients less than ten) were excluded. Experimental and animal studies were excluded. Studies unrelated to our topics were excluded. Portal vein thrombosis was not considered in the present study. These exclusions were chosen primarily due to the potential discrepancy in the pathogenesis of venous thrombosis at the different sites. The mechanism of portal vein thrombosis in liver diseases was not only attributed to the coagulation abnormalities [20], but also to the reduced portal vein velocity [21]. If the studies reported the number of portal vein thrombosis and DVT/PE separately, we only collected the data regarding the proportion of DVT/PE. Otherwise, they were excluded. If all observed patients with liver diseases were simultaneously accompanied by other conditions affecting the development of VTE, studies would be excluded. The conditions affecting the development of VTE included: liver malignancy, liver surgery alone (i.e., liver transplantation or resection), those undergoing endoscopic cyanoacrylate injection, those receiving erythropoietin or erythropoiesis-stimulating agents during hepatitis C virus treatment alone, prophylactic or therapeutic anticoagulation, and other specific drugs (i.e., recombinant activated factor VII.) No publication date, publication language, or publication status were excluded.
This work was performed using the guidelines for the reporting of meta-analysis of observational studies, which were published by the Meta-analysis of Observational Studies in Epidemiology (MOOSE) Group [19].
7.
Search strategy
Data collection
The PubMed, EMBASE, and Cochrane Library databases were employed for searching for relevant references. Search items were as follows: (‘‘cirrhosis’’ or ‘‘cirrhotic’’ or ‘‘disease’’) AND (‘‘liver’’ or ‘‘hepatic’’) AND (‘‘venous thromboembolism’’ or ‘‘VTE’’ or ‘‘venous thrombosis’’ or ‘‘deep vein thrombosis’’ or ‘‘pulmonary embolism’’) AND (‘‘incidence’’ or ‘‘prevalence’’ or ‘‘epidemiology’’ or ‘‘risk’’ or ‘‘predictor’’ or ‘‘predictive’’). The last search was performed on November 5, 2013. We manually searched the bibliographies of retrieved papers for additional references.
The primary items included: the first author, journal, publication year, publication type, region where the study was conducted, study design, period of enrollment, source of patients, methods to identify the cases, type of liver diseases, type of VTE, number of total observed patients with liver diseases, incidence or prevalence of VTE, eligibility criteria, and diagnosis of liver disease and VTE. Notably, the term ‘‘incidence’’ was defined by all new cases of VTE diagnosed in patients with liver diseases during some study period; by comparison, the term ‘‘prevalence’’ was defined by the accumulation of previous and new cases of VTE diagnosed in patients with liver diseases.
Eligibility criteria Study quality assessment The eligible papers should report the epidemiology of VTE in patients with liver diseases. VTE refers to DVT and PE. Liver diseases are divided into liver cirrhosis and non-
123
No score system was well established to assess the study quality in the proportion meta-analysis. Thus, we assessed
Intern Emerg Med
the study quality according to the eight following representative and important questions. Q1. Q2. Q3. Q4 Q5. Q6. Q7. Q8.
Was the study prospective? Were patients’ inclusion criteria available? Was the diagnosis of liver disease provided? Was the diagnosis of VTE provided? Was information regarding the VTE prophylaxis described? Was information regarding the presence of cancer described? Was information regarding the patients undergoing surgery described? Was information regarding the causes of liver diseases described?
If the answers to C5 questions were ‘‘Yes’’, the study was considered to be of high quality. Otherwise, it was considered to be of poor quality. Data analysis First, we drew the bar figures regarding the incidence or prevalence of VTE using SPSS version 16.0 (Chicago, Illinois, USA). Second, we calculated a pooled incidence or prevalence of VTE with 95 % confidence interval (CI) using StatsDirect statistical software version 2.7.8 (StatsDirect Ltd, Sale, Cheshire, UK.) We used a random-effect model to generate a more conservative estimate of the incidence or prevalence. The heterogeneity among studies was assessed using the Chi squared test (P \ 0.10 was considered to represent significant statistical heterogeneity), and the I2 statistic (I2 [ 50 % was considered as having substantial heterogeneity). We also performed the Egger test to evaluate the presence of publication bias. To explore the potential heterogeneity among studies, we conducted subgroup analyses according to the type of VTE (DVT/PE), type of liver disease (liver cirrhosis alone/ unclassified liver diseases or non-cirrhotics), region in which the study was performed (USA/European countries/ Asian countries), number of total observed patients with liver diseases ([1,000/\1,000 patients), study quality (high/low), and methods for identification of the cases (ICD codes/clinical charts).
Results Overall, 4,843 papers were retrieved by the search strategy. After exclusion, 20 papers were included in our study [6, 10, 11, 14–18, 22–33]. Notably, no additional eligible papers were manually identified. The causes for exclusion are shown in Fig. 1.
The characteristics of included studies are described in Table 1. Among them, 14 studies report the incidence of total VTE, DVT or PE in patients with liver diseases alone, 6 studies report the prevalence of total VTE, DVT or PE alone, and one study reports both prevalence and incidence (Fig. 2). A majority of studies (n = 13) were conducted by the investigators from the USA, three studies from European countries, and four studies from Asian countries. Of these included studies, three employed the USA Nationwide Inpatient Sample database from different years, one employed the USA National Hospital Discharge Survey database, one employed the USA claims database, one employed the Regenstrief Medical Record system, and 14 employed their respective hospital databases. The number of patients with liver diseases included in each study varied from 79 to 9,492,000. Eligibility criteria and diagnostic methods are described in Supplementary Tables 1 and 2, respectively. According to the above-mentioned criteria for assessing the study quality, 10 and 10 studies were considered to be of relatively high and poor quality (Supplementary Table 3). Incidence The total incidence of VTE varies from 0.33 to 6.32 % in 14 studies. The pooled incidence of VTE is 1 % (95 % CI 0.7–1.3 %) with a statistically significant heterogeneity among studies (I2 = 89.9 %; P \ 0.0001) (Fig. 3a). The Egger test shows a proof of publication bias (P = 0.0001). The incidence of DVT varies from 0.14 to 4.71 % in 12 studies. The pooled incidence of DVT is 0.6 % (95 % CI 0.4–0.8 %) with a statistically significant heterogeneity among studies (I2 = 85.7 %; P \ 0.0001) (Fig. 3b). The Egger test shows a proof of publication bias (P = 0.0081). The incidence of PE varies from 0 to 0.83 % in nine studies. The pooled incidence of PE is 0.28 % (95 % CI 0.13–0.49 %) with a statistically significant heterogeneity among studies (I2 = 90.4 %; P \ 0.0001) (Fig. 3c). There is not any significant publication bias by the Egger test (P = 0.2839). The heterogeneity remained statistically significant in all but the subgroup analysis regarding the incidence of VTE in Asian countries (Table 2); the incidence of VTE appears to be higher in studies including fewer than 1,000 patients than in those including more than 1,000 patients, because their respective 95 % CI values do not overlap. Prevalence The total prevalence of VTE varies from 0.6 to 4.69 % in six studies. Notably, two of these studies analyze the outcome of patients with two different types of liver diseases
123
Intern Emerg Med Fig. 1 Flowchart of study inclusion
(liver cirrhosis and unclassified liver diseases.) The pooled prevalence of VTE is 1.0 % (95 % CI 0.7–1.2 %) with a statistically significant heterogeneity among studies (I2 = 99.9 %; P \ 0.0001) (Fig. 4a). There is not any significant publication bias by the Egger test (P = 0.412).
123
The prevalence of DVT varies from 0.5 to 4.69 % in three studies. The pooled prevalence of DVT is 0.7 % (95 % CI 0.6–0.9 %) with a statistically significant heterogeneity among studies (I2 = 99.8 %; P \ 0.0001) (Fig. 4b). There is not any significant publication bias by the Egger test (P = 0.3966).
Chiba, Japan
Tennessee, USA
Virginia, USA
Rome, Italy
Virginia, USA
Tennessee, USA
Kentucky, USA
Kohsaka [28]
Lizarraga [25]
Northup [11]
Ponziani [30]
Shah [31]
Smith
Walsh [33]
[32]
Indiana, USA
Valencia, Spain
Gulley [16]
Missouri, USA
Dabbagh [10] Garcı´aFuster [14].
Iasi, Roumania
Retrospective
Nevada, USA
Barclay [26]
Girleanu [27]
Retrospective
Riyadh, Saudi Arabia Riyadh, Saudi Arabia
Aldawood [23] Al-Dorzi [24]
Retrospective
Retrospective
Prospective
Retrospective
Retrospective
Retrospective
Retrospective
Retrospective
Retrospective
2006.10–2010.7
2009.8.1–2011.7.31
2009.12–2010.1
1982.1–2012.2
1993–2001
2004.1.1–2008.1.1
NA
1995.1.1–2005.12.31
2010.1–2011.12
1992.1–2007.12
2000.1.1–2007.1.31
2008.4–2011.8
NA
Retrospective
Retrospective
2009.1.1–2009.12.31
2000–2009
Enrollment period
Retrospective
Retrospective
Pennsylvania, USA
Ahmed [22]
Study design
Region
References
Table 1 Characteristics of included studies
The University HealthSystem Consortium database in University of Kentucky-Chandler Hospital
A tertiary care academic medical center (Methodist University Hospital)
Inpatient Hepatology Service at the University of Virginia
The Agostino Gemelli Hospital
The University of Virginia Health System
The Rhode Island Hospital and the Miriam Hospital
The Keio University School of Medicine
The Regenstrief Medical Record system
The Center of Gastroenterology and Hepatology Iasi
A tertiary care Hospital of University of Missouri-Columbia The Hospital Clı´nico Universitario of Valencia
The University Medical Center of Southern Nevada
A tertiary teaching hospital (King Abdulaziz Medical City) The adult intensive care unit of a tertiary care center (King Abdulaziz Medical City)
The US claims database
Source of cases
Database
ICD-9-CM
Admission of patients to the inpatient service
Clinical records
ICD-9
Inpatients admissions by ICD-9
Medical records
ICD-9 & current procedural terminology codes
Clinical records
Clinical records
ICD-9
ICD-9
Clinical records
ICD-9
ICD-9
Identification of cases
CLD CLD
Incidence
CLD or LC
Decompensated LC
LC
Prevalence
Incidence
Incidence
Incidence
LC
CLD
Incidence
Incidence
LC
LC
LC
LC
CLD
CLD
LC
LC
Chronic hepatitis C virus
Type of cases
Incidence
Incidence
Incidence
Incidence
Incidence
Incidence
Incidence
Incidence
Incidence
Incidence/ prevalence
2,606
2,606
410
85
10,359
21,000
14,790
719
963
3,108
2,074
190
1,581
79
226
47,391
Cases with liver diseases
1.04
0.65
0.73
4.71
0.33
0.54
N/A
1.39
1.87
1.00
0.82
6.32
0.79
2.53
2.65
0.33
VTE incidence or prevalence (%)
Intern Emerg Med
123
0.81 Compensated LC Prevalence
408,253
0.82 Decompensated LC Prevalence
241,626
The prevalence of PE varies from 0.1 to 0.87 % in two studies. The pooled prevalence of PE is 0.36 % (95 % CI 0.13–0.7 %) with a statistically significant heterogeneity among studies (I2 = 100 %; P \ 0.0001) (Fig. 4C). An Egger test was not available due to the small number of studies. All subgroup analyses still had a statistically significant heterogeneity (Table 3). Notably, there are five USA studies with more than 1,000 patients in each study, but only one Asian study with fewer than 1,000 patients. Thus, the results of the subgroup analyses according to the country where the study was conducted were the same as those according to the number of patients included. Additionally, the prevalence of VTE appears to be higher in studies with fewer than 1,000 patients than in those with more than 1,000 patients, because their respective 95 % CI values do not overlap.
ICD-9-CM
Discussion
CLD chronic liver disease, LC liver cirrhosis, NA not available, VTE venous thromboembolism
The Nationwide Inpatient Sample 1998-2006 1998–2006 Maryland, USA Wu [6]
Retrospective
0.60
0.90 Non-alcoholic CLD Prevalence
4,927,000 Alcoholic CLD Prevalence
Michigan, USA Saleh [18]
Retrospective
1979–2006
The National Hospital Discharge Survey
ICD-9-CM
4,565,000
4.69 Jakarta, Indonesia
123
Kumar [29]
Lesmana [17]
Retrospective
2004.8–2007.7
The Medistra hospital in Jakarta
LC Prevalence Medical records
256
0.52 Retrospective
2007
The Nationwide Inpatient Sample 2007
LC Prevalence ICD-9-CM
560,503
1.83 449,798 LC Prevalence ICD-9-CM The Nationwide Inpatient Sample 2005 2005
Wisconsin, USA Wisconsin, USA Ali [15]
Retrospective
Region References
Table 1 continued
Study design
Enrollment period
Source of cases
Type of cases Incidence/ prevalence Identification of cases
Cases with liver diseases
VTE incidence or prevalence (%)
Intern Emerg Med
Increasing evidence suggests that the presence of liver disease should increase the development of VTE. When the present systematic review regarding the epidemiology of VTE in liver cirrhosis was being performed, we paid some additional attention to the issue regarding the association between liver diseases and VTE. Seven relevant papers were identified by our search strategy. Among them, 5 papers support that the risk of VTE is elevated by the presence of liver diseases. In a Saudi single-center study, AlGahtani et al. [34] find that the prevalence of chronic liver diseases is significantly higher in patients with PE than in those without PE (8.8 % [6/68] versus 2.9 % [8/273], p = 0.040). In a UK prospective cohort study, Huerta and colleagues report that chronic liver diseases are more frequently observed in patients with PE or DVT than in controls (0.67 [20/3,006], 0.54 % [19/3,544] versus 0.29 % [29/10,000]) [35]. But the difference is marginally significant (odds ratio = 1.65, 95 % CI 0.97–2.82). In another UK study, Ramagopalan et al. confirm that the risk of VTE is elevated by chronic active hepatitis in either the Oxford Record Linkage Study area (year 1963–1998: rate ratio = 2.24, 95 % CI 1.25–3.70, p = 0.003; year 1999–2008: rate ratio = 3.99, 95 % CI 1.46–8.72, p = 0.001), or the whole of England (year 1999–2008: rate ratio = 2.04, 95 % CI 1.65–2.50, p \ 0.001) [36]. In a large study from the Netherlands (MEGA study), Ocak et al. [37] also observe a significantly higher prevalence of liver diseases in patients with VTE than in controls (27/4,311 versus 22/5,768; OR = 1.7, 95 % CI 1.0–2.9). Recently, in a nationwide Danish case– control study with 99,444 patients with VTE and 496,872 population controls, Sogaard and colleagues suggest that
Intern Emerg Med
Fig. 2 Incidence (a) and prevalence (b) of venous thromboembolism, deep vein thrombosis, and pulmonary embolism
the patients with liver disease have a substantially increased risk of VTE (liver cirrhosis: risk ratio = 1.74, 95 % CI 1.54–1.95; non-cirrhotic liver disease: risk
ratio = 1.87, 95 % CI 1.73–2.03) [5]. By contrast, the remaining two studies do not support the association between VTE and liver diseases. In a US population-based
123
Intern Emerg Med Fig. 3 Pooled incidence of venous thromboembolism (a), deep vein thrombosis (b), and pulmonary embolism (c)
123
Intern Emerg Med Table 2 Incidence of VTE in patients with liver diseases: results of subgroup analyses Subgroups
No. of studies
Pooled proportion using randomeffects model
Heterogeneity I
P
Egger:bias
Liver cirrhosis
9
0.011 (95 % CI 0.007–0.015)
87.6 % (95 % CI 78.4–91.8 %)
P \ 0.0001
2.147 (95 % CI 0.732 to 3.561) P = 0.009
Unclassified liver diseases
5
0.010 (95 % CI 0.004–0.017)
90.8 % (95 % CI 81–94.4 %)
P \ 0.0001
2.107 (95 % CI 0.475–3.739) P = 0.026
Number of patients 7 [1,000
0.006 (95 % CI 0.004–0.008)
87.4 % (95 % CI 75.4–92.2 %)
P \ 0.0001
2.781 (95 % CI 0.516–5.046) P = 0.025
Number of patients 7 \1,000 Countries
0.025 (95 % CI 0.014–0.038)
69.3 % (95 % CI 7.7–84.2 %)
P = 0.0033
1.793 (95 % CI -0.094 to 3.681) P = 0.059
2
Publication bias
Type of liver diseases
Number of patients
USA
8
0.010 (95 % CI 0.006–0.014)
91.6 % (95 % CI 86.4–94.2 %)
P \ 0.0001
2.401 (95 % CI 0.929–3.872) P = 0.007
European countries
3
0.007 (95 % CI 0.003–0.013)
90.9 % (95 % CI 71.7–95.3 %)
P \ 0.0001
N/A
Asian countries
3
0.019 (95 % CI 0.011–0.030)
10.1 % (95 % CI 0–75.5 %)
P = 0.329
N/A
High quality
8
0.012 (95 % CI 0.008–0.018)
87.1 % (95 % CI 76–91.7 %)
P \ 0.0001
2.090 (95 % CI 0.915–3.264) P = 0.005
Poor quality
6
0.007 (95 % CI 0.004–0.01)
85.7 % (95 % CI 67.9–91.7 %)
P \ 0.0001
1.759 (95 % CI 0.668–2.849) P = 0.011
ICD codes
7
0.011 (95 % CI 0.007–0.016)
92.6 % (95 % CI 87.9–95 %)
P \ 0.0001
2.513 (95 % CI 0.744–4.282) P = 0.015
Clinical charts
7
0.0096 (95 % CI 0.005–0.015)
85.5 % (95 % CI 69.9–91.2 %)
P \ 0.0001
2.290 (95 % CI 0.828–3.751) P = 0.01
Study quality
Case identification
CI confidence interval, N/A not available
case–control study, Heit et al. [38] show that the prevalence of serious liver diseases is similar between patients with and without PE (5/625 versus 6/625). The investigators suggest that patients with serious liver diseases are 90 % less likely to develop VTE. In another populationbased study using the Nationwide Inpatient Sample database, Shaheen et al. [39] report that patients with cirrhosis are less likely to develop postoperative VTE (age/sexadjusted incidence rate ratio = 0.52, 95 % CI 0.48–0.56). Because the data are expressed in different ways, and the selection of case and control groups is inconsistent between studies, we could not perform any meta-analyses. However, on the basis of these aforementioned data, it appears that the risk of VTE is increased by the presence of liver diseases. In the present systematic review, we tried our best to collect all available evidence to analyze the epidemiology of VTE in patients with liver diseases. The major finding of our systematic review and meta-analysis is that about 1 % of patients with liver diseases develop or are diagnosed
with VTE during their hospitalizations. Compared to previous narrative reviews and every single study with a different sample size and enrollment period, the results of the systematic review may be more generalizable to patients with liver diseases. Another finding of our study is that the incidence of DVT and PE is 0.6 and 0.28 %, respectively; and the prevalence of DVT and PE is 0.7 and 0.36 %, respectively. Considering that their respective 95 % CI values overlap, we can not arbitrarily suggest that DVT is more common than PE. The results of the subgroup analyses also indicate a slightly higher prevalence in studies involving fewer than 1,000 patients than in those involving more than 1,000 patients. But it should be noted that the studies with a small sample size may bring in more selection bias than those with a relatively large sample size. More notably, the subgroup analyses are designed to explore the source of heterogeneity. Therefore, these findings of the subgroup analyses must be interpreted with caution, because they were not confirmed by any statistical analyses.
123
Intern Emerg Med Fig. 4 Pooled prevalence of venous thromboembolism (a), deep vein thrombosis (b), and pulmonary embolism (c)
123
Intern Emerg Med Table 3 Prevalence of VTE in patients with liver diseases: Results of subgroup analyses Subgroups
No. of studies
Pooled proportion using random-effects model
Heterogeneity
Liver cirrhosis
5
0.012 (95 % CI 0.007–0.019)
Unclassified liver diseases
3
Number of patients 7 [1,000
2
Publication bias P
Egger:bias
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
23.698 (95 % CI -60.156 to 107.553) P = 0.435
0.008 (95 % CI 0.006–0.012)
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
N/A
0.009 (95 % CI 0.007–0.011)
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
17.499 (95 % CI -32.658 to 67.656) P = 0.411
1
0.047 (95 % CI 0.024–0.080)
N/A
N/A
N/A
USA
7
0.009 (95 % CI 0.007–0.011)
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
17.499 (95 % CI -32.658 to 67.656) P = 0.411
Asian countries
1
0.047 (95 % CI 0.024–0.080)
N/A
N/A
N/A
High quality
2
0.025 (95 % CI 0.002–0.074)
N/A
P = 0.0003
N/A
Poor quality
6
0.009 (95 % CI 0.006–0.011)
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
25.304 (95 % CI -46.135 to 96.743) P = 0.3811
Case identification ICD codes
6
0.009 (95 % CI 0.006–0.011)
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
25.304 (95 % CI -46.135 to 96.743) P = 0.3811
Clinical charts
2
0.025 (95 % CI 0.002–0.074)
N/A
P = 0.0003
N/A
I
Type of liver diseases
Number of patients
Number of patients \ 1,000 Countries
Study quality
CI confidence interval, N/A not available
To the best of our knowledge, this study is the first systematic review and meta-analysis to explore the epidemiology of VTE in liver diseases. Compared to a previous narrative review by Tripodi et al. [9], the strengths of our systematic review are as follows: (1) we performed an extensive literature search via three major databases to maximize the number of retrieved papers; (2) except for the description of data from different studies, we conducted the data synthesis of all available evidence by means of metaanalysis. Additional strengths include: (1) the eligibility criteria of our study are rigorous, some specific conditions affecting the development of VTE are excluded; (2) only a random-effect model is employed to produce a conservative result; and (3) a detailed questionnaire involving eight critical issues is used to assess the study quality. Our study has several limitations. First, a statistically significant heterogeneity among studies is found in nearly all meta-analyses. Although the subgroup analyses were conducted according to different variables, the potential reasons remain obscure. Indeed, only the subgroup analysis regarding the incidence of VTE in liver diseases in Asian studies does not demonstrate any significant heterogeneity. However, it should be noted that statistically significant heterogeneity was commonly observed in nearly all
previous proportion meta-analyses [40–42]. Second, five studies were published in abstract or letter forms. Thus, much of the information regarding the characteristics of these studies was often missing. Third, only one study with a relatively small sample size was prospective, suggesting that it had a prior study design to observe the incidence of VTE. But the remaining studies were retrospective, suggesting the presence of recall bias. Fourth, enrollment spanned more than 10 years in five studies. During a relatively long period of time, the diagnostic methods of VTE were evolving, which potentially increased the heterogeneity of the results. Indeed, the study by Wu et al. [6] demonstrates an increasing trend in the prevalence of VTE in patients with liver diseases. Fifth, the level of statistical significance may be heavily weighted by the sample size. Thus, the results of meta-analyses are subdivided according to a sample size of fewer than or more than 1,000 patients. The epidemiology of VTE is higher in studies with \1,000 patients than in those with [1,000 patients (incidence: 2.5 versus 0.6 %; prevalence: 4.7 versus 0.9 %), suggesting the potential chance in studies with a small sample size. Sixth, the results may be influenced by the methods for identifying the VTE cases. The subgroup meta-analyses are separately conducted in the studies using ICD codes and
123
Intern Emerg Med
clinical charts. However, the incidence of VTE is similar between the two groups (1.1 versus 0.96 %). Although the prevalence of VTE is lower in the studies using ICD codes than in those using clinical charts (0.9 versus 2.5 %), their 95 % CIs overlap. In conclusion, our study suggests that VTE will occur in 1 % of patients with liver diseases during their hospitalizations. Considering an elevated risk of VTE in liver cirrhosis, further prospective studies might be helpful to explore the role of VTE prophylaxis. Conflict of interest
16.
17.
18. 19.
None. 20.
References 21. 1. Tripodi A, Mannucci PM (2011) The coagulopathy of chronic liver disease. N Engl J Med 365:147–156 2. Lisman T, Caldwell SH, Burroughs AK, Northup PG, Senzolo M, Stravitz RT et al (2010) Hemostasis and thrombosis in patients with liver disease: the ups and downs. J Hepatol 53:362–371 3. Northup PG, Caldwell SH (2013) Coagulation in liver disease: a guide for the clinician. Clin Gastroenterol Hepatol 11:1064–1074 4. Amitrano L, Guardascione MA, Brancaccio V, Balzano A (2002) Coagulation disorders in liver disease. Semin Liver Dis 22:83–96 5. Sogaard KK, Horvath-Puho E, Gronbaek H, Jepsen P, Vilstrup H, Sorensen HT (2009) Risk of venous thromboembolism in patients with liver disease: a nationwide population-based case-control study. Am J Gastroenterol 104:96–101 6. Wu H, Nguyen GC (2010) Liver cirrhosis is associated with venous thromboembolism among hospitalized patients in a nationwide US study. Clin Gastroenterol Hepatol 8:800–805 7. Tripodi A, Primignani M, Chantarangkul V, Dell’Era A, Clerici M, de Franchis R et al (2009) An imbalance of pro- vs anticoagulation factors in plasma from patients with cirrhosis. Gastroenterology 137:2105–2111 8. Tripodi A, Primignani M, Lemma L, Chantarangkul V, Mannucci PM (2013) Evidence that low protein C contributes to the procoagulant imbalance in cirrhosis. J Hepatol 59:265–270 9. Tripodi A, Anstee QM, Sogaard KK, Primignani M, Valla DC (2011) Hypercoagulability in cirrhosis: causes and consequences. J Thromb Haemost 9:1713–1723 10. Dabbagh O, Oza A, Prakash S, Sunna R, Saettele TM (2010) Coagulopathy does not protect against venous thromboembolism in hospitalized patients with chronic liver disease. Chest 137:1145–1149 11. Northup PG, McMahon MM, Ruhl AP, Altschuler SE, VolkBednarz A, Caldwell SH et al (2006) Coagulopathy does not fully protect hospitalized cirrhosis patients from peripheral venous thromboembolism. Am J Gastroenterol 101:1524–1528 quiz 1680 12. Gopal K, Broder A, Benias P, Min AD (2010) Practice patterns among medicine housestaff in prevention of venous thromboembolism among hospitalized patients with end-stage liver disease. Gastroenterology 138:S457 13. Goldhaber SZ, Bounameaux H (2012) Pulmonary embolism and deep vein thrombosis. Lancet 379:1835–1846 14. Garcia-Fuster MJ, Abdilla N, Fabia MJ, Fernandez C, Oliver V (2008) Forner MJ [Venous thromboembolism and liver cirrhosis]. Rev Esp Enferm Dig 100:259–262 15. Ali M, Ananthakrishnan AN, McGinley EL, Saeian K (2011) Deep vein thrombosis and pulmonary embolism in hospitalized
123
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
patients with cirrhosis: a nationwide analysis. Dig Dis Sci 56:2152–2159 Gulley D, Teal E, Suvannasankha A, Chalasani N, Liangpunsakul S (2008) Deep vein thrombosis and pulmonary embolism in cirrhosis patients. Dig Dis Sci 53:3012–3017 Lesmana CR, Inggriani S, Cahyadinata L, Lesmana LA (2010) Deep vein thrombosis in patients with advanced liver cirrhosis: a rare condition? Hepatol Int 4:433–438 Saleh T, Matta F, Alali F, Stein PD (2011) Venous thromboembolism with chronic liver disease. Am J Med 124:64–68 Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D et al (2000) Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of observational studies in epidemiology (MOOSE) group. JAMA 283:2008–2012 Amitrano L, Brancaccio V, Guardascione MA, Margaglione M, Iannaccone L, D’Andrea G et al (2000) Inherited coagulation disorders in cirrhotic patients with portal vein thrombosis. Hepatology 31:345–348 Zocco MA, Di Stasio E, De Cristofaro R, Novi M, Ainora ME, Ponziani F et al (2009) Thrombotic risk factors in patients with liver cirrhosis: correlation with MELD scoring system and portal vein thrombosis development. J Hepatol 51:682–689 Ahmed S, Mehta V, Vendetti N, Ann Phang K, Su Y, Kim H et al (2012) Risk of venous thromboembolism (VTE) in patients with chronic hepatitis C (CHC. Pharmacoepidemiol Drug Saf 21:94 Aldawood A, Arabi Y, Aljumah A, Alsaadi A, Rishu A, Aldorzi H et al (2011) The incidence of venous thromboembolism and practice of deep venous thrombosis prophylaxis in hospitalized cirrhotic patients. Thromb J 9:1 Al-Dorzi HM, Bhat S, Tamim H, Khokar S, Hussain S, Al-Dawood A et al (2010) Practices of venous thromboembolism prophylaxis and incidence in critically ill cirrhotic patients. Am J Respir Crit Care Med 181 Anthony Lizarraga W, Dalia S, Reinert SE, Schiffman FJ (2010) Venous thrombosis in patients with chronic liver disease. Blood Coagul Fibrinolysis 21:431–435 Barclay SM, Jeffres MN, Nguyen K, Nguyen T (2013) Evaluation of pharmacologic prophylaxis for venous thromboembolism in patients with chronic liver disease. Pharmacotherapy 33:375–382 Girleanu I, Trifan A, Cojocariu C, Singeap AM, Sfarti C, Stanciu C (2012) The risk of thrombotic events in patients with liver cirrhosis. Rev Med Chir Soc Med Nat Iasi 116:991–996 Kohsaka S, Nagai T, Yaegashi M, Fukuda K (2012) Pulmonary embolism and deep venous thrombosis in hospitalized patients with liver cirrhosis. Hepatology Research 42:433–434 Kumar G, Kumar N, Deshmukh A, Sakhuja A, Taneja A, Padala K et al (2010) Is cirrhosis protective for venous thromboembolism? Analysis from a national inpatient sample. Chest 138 Ponziani FR, Zocco MA, Garcovich M, Roccarina D, Gasbarrini A (2013) Epidemiology of venous thrombotic events and pulmonary embolism among hospitalized cirrhotic patients: a single center experience. Digestive and Liver Disease 45:S41 Shah NL, Northup PG, Caldwell SH (2012) A clinical survey of bleeding, thrombosis, and blood product use in decompensated cirrhosis patients. Ann Hepatol 11:686–690 Smith CB, Hurdle AC, Kemp LO, Sands C, Twilla JD (2013) Evaluation of venous thromboembolism prophylaxis in patients with chronic liver disease. J Hosp Med 8:569–573 Walsh KA, Lewis DA, Clifford TM, Hundley JC, Gokun Y, Angulo P et al (2013) Risk factors for venous thromboembolism in patients with chronic liver disease. Ann Pharmacother 47:333–339 Algahtani FH, Bayoumi N, Abdelgadir A, Al-Nakshabandi N, Al Aseri Z, Al Ghamdi M et al (2013) Clinical characteristics and
Intern Emerg Med
35.
36.
37.
38.
risk factors of pulmonary embolism: data from a Saudi tertiarycare center. J Thromb and Haemost 11:388–390 Huerta C, Johansson S, Wallander MA, Garcia Rodriguez LA (2007) Risk factors and short-term mortality of venous thromboembolism diagnosed in the primary care setting in the United Kingdom. Arch Intern Med 167:935–943 Ramagopalan SV, Wotton CJ, Handel AE, Yeates D, Goldacre MJ (2011) Risk of venous thromboembolism in people admitted to hospital with selected immune-mediated diseases: recordlinkage study. BMC Medicine 9 Ocak G, Vossen CY, Verduijn M, Dekker FW, Rosendaal FR, Cannegieter SC et al (2013) Risk of venous thrombosis in patients with major illnesses: results from the MEGA study. J Thromb Haemost 11:116–123 Heit JA, Silverstein MD, Mohr DN, Petterson TM, O’Fallon WM, Melton LJ 3rd (2000) Risk factors for deep vein thrombosis and pulmonary embolism: a population-based case-control study. Arch Intern Med 160:809–815
39. Shaheen AAM, Tanyingoh D, Dixon E, Kaplan GG, Myers RP (2012) Medical injuries among hospitalized patients with cirrhosis: a population-based study. Hepatology 56:1083A–1084A 40. Ford AC, Marwaha A, Lim A, Moayyedi P (2010) What is the prevalence of clinically significant endoscopic findings in subjects with dyspepsia? Systematic review and meta-analysis. Clin Gastroenterol Hepatol 8:830–837 837 e831–832 41. Ford AC, Marwaha A, Lim A, Moayyedi P (2010) Systematic review and meta-analysis of the prevalence of irritable bowel syndrome in individuals with dyspepsia. Clin Gastroenterol Hepatol 8:401–409 42. Larney S, Kopinski H, Beckwith CG, Zaller ND, Jarlais DD, Hagan H et al (2013) Incidence and prevalence of hepatitis C in prisons and other closed settings: results of a systematic review and meta-analysis. Hepatology 58:1215–1224
123
CE-ORIGINAL
Epidemiology of venous thromboembolism in patients with liver diseases: a systematic review and meta-analysis Xingshun Qi • Weirong Ren • Xiaozhong Guo Daiming Fan
•
Received: 12 October 2014 / Accepted: 20 November 2014 Ó SIMI 2014
Abstract The risk of venous thromboembolism (VTE) may be increased in patients with liver diseases. A systematic review and meta-analysis were conducted to analyze the epidemiology of VTE in such patients. All relevant studies were searched via the PubMed, EMBASE, and Cochrane Library databases. The incidence and prevalence of VTE were pooled using random-effect models. Subgroup analyses were conducted according to the type of VTE [deep vein thrombosis (DVT), pulmonary embolism (PE)], type of liver disease (liver cirrhosis alone/unclassified liver diseases or non-cirrhotics), region in which the study was performed (USA/Europe/Asia), number of total observed patients with liver diseases ([1,000/\1,000 patients), study quality (high/low), and methods for identifying the cases (ICD codes/clinical charts). Of 4,843 Electronic supplementary material The online version of this article (doi:10.1007/s11739-014-1163-7) contains supplementary material, which is available to authorized users. X. Qi (&) X. Guo (&) Department of Gastroenterology, General Hospital of Shenyang Military Area, No. 83 Wehua Road, Shenyang 110840, China e-mail: [email protected] X. Guo e-mail: [email protected] X. Qi W. Ren D. Fan (&) Xijing Hospital of Digestive Diseases, Fourth Military Medical University, No. 127 Changle West Road, Xi’an 710032, China e-mail: [email protected] W. Ren e-mail: [email protected] W. Ren Department of Digestive Diseases, Sanmenxia Central Hospital, Henan University of Science and Technology, Xiaoshan Road, Sanmenxia 472000, China
papers initially identified, 20 were included. The incidence of VTE varied from 0.33 to 6.32 % in 14 studies with a pooled value of 1 % (95 % confidence interval (CI) 0.7–1.3 %). The pooled incidence of DVT and PE was 0.6 % (95 % CI 0.4–0.8 %) and 0.28 % (95 % CI 0.13–0.49 %), respectively. The prevalence of VTE varied from 0.6 to 4.69 % in six studies with a pooled value of 1.0 % (95 % CI 0.7–1.2 %). The pooled prevalence of DVT and PE was 0.7 % (95 % CI 0.6–0.9 %) and 0.36 % (95 % CI 0.13–0.7 %), respectively. The heterogeneity was statistically significant in the main and subgroup metaanalyses. In conclusion, about 1 % of patients with liver diseases develop or are diagnosed with VTE during their hospitalizations. However, the epidemiological data are very heterogeneous among studies. Keywords Deep vein thrombosis Pulmonary embolism Prevalence Incidence Liver cirrhosis Abbreviations CI Confidence interval VTE Venous thromboembolism DVT Deep vein thrombosis PE Pulmonary embolism
Introduction Nowadays, the emerging evidence confirms a new paradigm shift from the bleeding tendency to the rebalanced, but unstable, hemostasis in patients with liver diseases [1– 4]. Epidemiological studies have also shown that the risk of venous thromboembolism (VTE) may be increased in these
123
Intern Emerg Med
patients [5, 6]. Laboratory research studies further demonstrate that both increased levels of factor VIII and remarkably decreased levels of protein C contribute to the presence of hypercoagulability of plasma from patients with liver diseases [7–9]. These findings suggest that the patients with liver diseases are not protected against the development of VTE, despite often having a prolonged internationalized normalized ratio, and being considered to be in an auto-anticoagulation status [10, 11]. However, despite the significance of VTE in patients with liver diseases, and even with an elevated internationalized normalized ratio gradually having been recognized, no obvious improvement has been made in translating the information into clinical practice due to the absence of clear recommendations from practice guidelines and fear of bleeding risks [12]. VTE, including deep vein thrombosis (DVT) and pulmonary embolism (PE), represent major global public health problems [13]. Recently, several studies report upon the incidence or prevalence of VTE in patients with liver diseases [6, 10, 11, 14–18]. The present study aims to systematically collect the epidemiological data regarding VTE in patients with liver diseases, and to combine the individual data into a generalizable result. The sources of heterogeneity among different studies are also explored.
cirrhotics. Additionally, several exclusion criteria were as follows: 1. 2. 3. 4. 5.
6.
Methods
Reviews, meta-analyses, comments, and editorials were excluded. Case reports (number of patients less than ten) were excluded. Experimental and animal studies were excluded. Studies unrelated to our topics were excluded. Portal vein thrombosis was not considered in the present study. These exclusions were chosen primarily due to the potential discrepancy in the pathogenesis of venous thrombosis at the different sites. The mechanism of portal vein thrombosis in liver diseases was not only attributed to the coagulation abnormalities [20], but also to the reduced portal vein velocity [21]. If the studies reported the number of portal vein thrombosis and DVT/PE separately, we only collected the data regarding the proportion of DVT/PE. Otherwise, they were excluded. If all observed patients with liver diseases were simultaneously accompanied by other conditions affecting the development of VTE, studies would be excluded. The conditions affecting the development of VTE included: liver malignancy, liver surgery alone (i.e., liver transplantation or resection), those undergoing endoscopic cyanoacrylate injection, those receiving erythropoietin or erythropoiesis-stimulating agents during hepatitis C virus treatment alone, prophylactic or therapeutic anticoagulation, and other specific drugs (i.e., recombinant activated factor VII.) No publication date, publication language, or publication status were excluded.
This work was performed using the guidelines for the reporting of meta-analysis of observational studies, which were published by the Meta-analysis of Observational Studies in Epidemiology (MOOSE) Group [19].
7.
Search strategy
Data collection
The PubMed, EMBASE, and Cochrane Library databases were employed for searching for relevant references. Search items were as follows: (‘‘cirrhosis’’ or ‘‘cirrhotic’’ or ‘‘disease’’) AND (‘‘liver’’ or ‘‘hepatic’’) AND (‘‘venous thromboembolism’’ or ‘‘VTE’’ or ‘‘venous thrombosis’’ or ‘‘deep vein thrombosis’’ or ‘‘pulmonary embolism’’) AND (‘‘incidence’’ or ‘‘prevalence’’ or ‘‘epidemiology’’ or ‘‘risk’’ or ‘‘predictor’’ or ‘‘predictive’’). The last search was performed on November 5, 2013. We manually searched the bibliographies of retrieved papers for additional references.
The primary items included: the first author, journal, publication year, publication type, region where the study was conducted, study design, period of enrollment, source of patients, methods to identify the cases, type of liver diseases, type of VTE, number of total observed patients with liver diseases, incidence or prevalence of VTE, eligibility criteria, and diagnosis of liver disease and VTE. Notably, the term ‘‘incidence’’ was defined by all new cases of VTE diagnosed in patients with liver diseases during some study period; by comparison, the term ‘‘prevalence’’ was defined by the accumulation of previous and new cases of VTE diagnosed in patients with liver diseases.
Eligibility criteria Study quality assessment The eligible papers should report the epidemiology of VTE in patients with liver diseases. VTE refers to DVT and PE. Liver diseases are divided into liver cirrhosis and non-
123
No score system was well established to assess the study quality in the proportion meta-analysis. Thus, we assessed
Intern Emerg Med
the study quality according to the eight following representative and important questions. Q1. Q2. Q3. Q4 Q5. Q6. Q7. Q8.
Was the study prospective? Were patients’ inclusion criteria available? Was the diagnosis of liver disease provided? Was the diagnosis of VTE provided? Was information regarding the VTE prophylaxis described? Was information regarding the presence of cancer described? Was information regarding the patients undergoing surgery described? Was information regarding the causes of liver diseases described?
If the answers to C5 questions were ‘‘Yes’’, the study was considered to be of high quality. Otherwise, it was considered to be of poor quality. Data analysis First, we drew the bar figures regarding the incidence or prevalence of VTE using SPSS version 16.0 (Chicago, Illinois, USA). Second, we calculated a pooled incidence or prevalence of VTE with 95 % confidence interval (CI) using StatsDirect statistical software version 2.7.8 (StatsDirect Ltd, Sale, Cheshire, UK.) We used a random-effect model to generate a more conservative estimate of the incidence or prevalence. The heterogeneity among studies was assessed using the Chi squared test (P \ 0.10 was considered to represent significant statistical heterogeneity), and the I2 statistic (I2 [ 50 % was considered as having substantial heterogeneity). We also performed the Egger test to evaluate the presence of publication bias. To explore the potential heterogeneity among studies, we conducted subgroup analyses according to the type of VTE (DVT/PE), type of liver disease (liver cirrhosis alone/ unclassified liver diseases or non-cirrhotics), region in which the study was performed (USA/European countries/ Asian countries), number of total observed patients with liver diseases ([1,000/\1,000 patients), study quality (high/low), and methods for identification of the cases (ICD codes/clinical charts).
Results Overall, 4,843 papers were retrieved by the search strategy. After exclusion, 20 papers were included in our study [6, 10, 11, 14–18, 22–33]. Notably, no additional eligible papers were manually identified. The causes for exclusion are shown in Fig. 1.
The characteristics of included studies are described in Table 1. Among them, 14 studies report the incidence of total VTE, DVT or PE in patients with liver diseases alone, 6 studies report the prevalence of total VTE, DVT or PE alone, and one study reports both prevalence and incidence (Fig. 2). A majority of studies (n = 13) were conducted by the investigators from the USA, three studies from European countries, and four studies from Asian countries. Of these included studies, three employed the USA Nationwide Inpatient Sample database from different years, one employed the USA National Hospital Discharge Survey database, one employed the USA claims database, one employed the Regenstrief Medical Record system, and 14 employed their respective hospital databases. The number of patients with liver diseases included in each study varied from 79 to 9,492,000. Eligibility criteria and diagnostic methods are described in Supplementary Tables 1 and 2, respectively. According to the above-mentioned criteria for assessing the study quality, 10 and 10 studies were considered to be of relatively high and poor quality (Supplementary Table 3). Incidence The total incidence of VTE varies from 0.33 to 6.32 % in 14 studies. The pooled incidence of VTE is 1 % (95 % CI 0.7–1.3 %) with a statistically significant heterogeneity among studies (I2 = 89.9 %; P \ 0.0001) (Fig. 3a). The Egger test shows a proof of publication bias (P = 0.0001). The incidence of DVT varies from 0.14 to 4.71 % in 12 studies. The pooled incidence of DVT is 0.6 % (95 % CI 0.4–0.8 %) with a statistically significant heterogeneity among studies (I2 = 85.7 %; P \ 0.0001) (Fig. 3b). The Egger test shows a proof of publication bias (P = 0.0081). The incidence of PE varies from 0 to 0.83 % in nine studies. The pooled incidence of PE is 0.28 % (95 % CI 0.13–0.49 %) with a statistically significant heterogeneity among studies (I2 = 90.4 %; P \ 0.0001) (Fig. 3c). There is not any significant publication bias by the Egger test (P = 0.2839). The heterogeneity remained statistically significant in all but the subgroup analysis regarding the incidence of VTE in Asian countries (Table 2); the incidence of VTE appears to be higher in studies including fewer than 1,000 patients than in those including more than 1,000 patients, because their respective 95 % CI values do not overlap. Prevalence The total prevalence of VTE varies from 0.6 to 4.69 % in six studies. Notably, two of these studies analyze the outcome of patients with two different types of liver diseases
123
Intern Emerg Med Fig. 1 Flowchart of study inclusion
(liver cirrhosis and unclassified liver diseases.) The pooled prevalence of VTE is 1.0 % (95 % CI 0.7–1.2 %) with a statistically significant heterogeneity among studies (I2 = 99.9 %; P \ 0.0001) (Fig. 4a). There is not any significant publication bias by the Egger test (P = 0.412).
123
The prevalence of DVT varies from 0.5 to 4.69 % in three studies. The pooled prevalence of DVT is 0.7 % (95 % CI 0.6–0.9 %) with a statistically significant heterogeneity among studies (I2 = 99.8 %; P \ 0.0001) (Fig. 4b). There is not any significant publication bias by the Egger test (P = 0.3966).
Chiba, Japan
Tennessee, USA
Virginia, USA
Rome, Italy
Virginia, USA
Tennessee, USA
Kentucky, USA
Kohsaka [28]
Lizarraga [25]
Northup [11]
Ponziani [30]
Shah [31]
Smith
Walsh [33]
[32]
Indiana, USA
Valencia, Spain
Gulley [16]
Missouri, USA
Dabbagh [10] Garcı´aFuster [14].
Iasi, Roumania
Retrospective
Nevada, USA
Barclay [26]
Girleanu [27]
Retrospective
Riyadh, Saudi Arabia Riyadh, Saudi Arabia
Aldawood [23] Al-Dorzi [24]
Retrospective
Retrospective
Prospective
Retrospective
Retrospective
Retrospective
Retrospective
Retrospective
Retrospective
2006.10–2010.7
2009.8.1–2011.7.31
2009.12–2010.1
1982.1–2012.2
1993–2001
2004.1.1–2008.1.1
NA
1995.1.1–2005.12.31
2010.1–2011.12
1992.1–2007.12
2000.1.1–2007.1.31
2008.4–2011.8
NA
Retrospective
Retrospective
2009.1.1–2009.12.31
2000–2009
Enrollment period
Retrospective
Retrospective
Pennsylvania, USA
Ahmed [22]
Study design
Region
References
Table 1 Characteristics of included studies
The University HealthSystem Consortium database in University of Kentucky-Chandler Hospital
A tertiary care academic medical center (Methodist University Hospital)
Inpatient Hepatology Service at the University of Virginia
The Agostino Gemelli Hospital
The University of Virginia Health System
The Rhode Island Hospital and the Miriam Hospital
The Keio University School of Medicine
The Regenstrief Medical Record system
The Center of Gastroenterology and Hepatology Iasi
A tertiary care Hospital of University of Missouri-Columbia The Hospital Clı´nico Universitario of Valencia
The University Medical Center of Southern Nevada
A tertiary teaching hospital (King Abdulaziz Medical City) The adult intensive care unit of a tertiary care center (King Abdulaziz Medical City)
The US claims database
Source of cases
Database
ICD-9-CM
Admission of patients to the inpatient service
Clinical records
ICD-9
Inpatients admissions by ICD-9
Medical records
ICD-9 & current procedural terminology codes
Clinical records
Clinical records
ICD-9
ICD-9
Clinical records
ICD-9
ICD-9
Identification of cases
CLD CLD
Incidence
CLD or LC
Decompensated LC
LC
Prevalence
Incidence
Incidence
Incidence
LC
CLD
Incidence
Incidence
LC
LC
LC
LC
CLD
CLD
LC
LC
Chronic hepatitis C virus
Type of cases
Incidence
Incidence
Incidence
Incidence
Incidence
Incidence
Incidence
Incidence
Incidence
Incidence/ prevalence
2,606
2,606
410
85
10,359
21,000
14,790
719
963
3,108
2,074
190
1,581
79
226
47,391
Cases with liver diseases
1.04
0.65
0.73
4.71
0.33
0.54
N/A
1.39
1.87
1.00
0.82
6.32
0.79
2.53
2.65
0.33
VTE incidence or prevalence (%)
Intern Emerg Med
123
0.81 Compensated LC Prevalence
408,253
0.82 Decompensated LC Prevalence
241,626
The prevalence of PE varies from 0.1 to 0.87 % in two studies. The pooled prevalence of PE is 0.36 % (95 % CI 0.13–0.7 %) with a statistically significant heterogeneity among studies (I2 = 100 %; P \ 0.0001) (Fig. 4C). An Egger test was not available due to the small number of studies. All subgroup analyses still had a statistically significant heterogeneity (Table 3). Notably, there are five USA studies with more than 1,000 patients in each study, but only one Asian study with fewer than 1,000 patients. Thus, the results of the subgroup analyses according to the country where the study was conducted were the same as those according to the number of patients included. Additionally, the prevalence of VTE appears to be higher in studies with fewer than 1,000 patients than in those with more than 1,000 patients, because their respective 95 % CI values do not overlap.
ICD-9-CM
Discussion
CLD chronic liver disease, LC liver cirrhosis, NA not available, VTE venous thromboembolism
The Nationwide Inpatient Sample 1998-2006 1998–2006 Maryland, USA Wu [6]
Retrospective
0.60
0.90 Non-alcoholic CLD Prevalence
4,927,000 Alcoholic CLD Prevalence
Michigan, USA Saleh [18]
Retrospective
1979–2006
The National Hospital Discharge Survey
ICD-9-CM
4,565,000
4.69 Jakarta, Indonesia
123
Kumar [29]
Lesmana [17]
Retrospective
2004.8–2007.7
The Medistra hospital in Jakarta
LC Prevalence Medical records
256
0.52 Retrospective
2007
The Nationwide Inpatient Sample 2007
LC Prevalence ICD-9-CM
560,503
1.83 449,798 LC Prevalence ICD-9-CM The Nationwide Inpatient Sample 2005 2005
Wisconsin, USA Wisconsin, USA Ali [15]
Retrospective
Region References
Table 1 continued
Study design
Enrollment period
Source of cases
Type of cases Incidence/ prevalence Identification of cases
Cases with liver diseases
VTE incidence or prevalence (%)
Intern Emerg Med
Increasing evidence suggests that the presence of liver disease should increase the development of VTE. When the present systematic review regarding the epidemiology of VTE in liver cirrhosis was being performed, we paid some additional attention to the issue regarding the association between liver diseases and VTE. Seven relevant papers were identified by our search strategy. Among them, 5 papers support that the risk of VTE is elevated by the presence of liver diseases. In a Saudi single-center study, AlGahtani et al. [34] find that the prevalence of chronic liver diseases is significantly higher in patients with PE than in those without PE (8.8 % [6/68] versus 2.9 % [8/273], p = 0.040). In a UK prospective cohort study, Huerta and colleagues report that chronic liver diseases are more frequently observed in patients with PE or DVT than in controls (0.67 [20/3,006], 0.54 % [19/3,544] versus 0.29 % [29/10,000]) [35]. But the difference is marginally significant (odds ratio = 1.65, 95 % CI 0.97–2.82). In another UK study, Ramagopalan et al. confirm that the risk of VTE is elevated by chronic active hepatitis in either the Oxford Record Linkage Study area (year 1963–1998: rate ratio = 2.24, 95 % CI 1.25–3.70, p = 0.003; year 1999–2008: rate ratio = 3.99, 95 % CI 1.46–8.72, p = 0.001), or the whole of England (year 1999–2008: rate ratio = 2.04, 95 % CI 1.65–2.50, p \ 0.001) [36]. In a large study from the Netherlands (MEGA study), Ocak et al. [37] also observe a significantly higher prevalence of liver diseases in patients with VTE than in controls (27/4,311 versus 22/5,768; OR = 1.7, 95 % CI 1.0–2.9). Recently, in a nationwide Danish case– control study with 99,444 patients with VTE and 496,872 population controls, Sogaard and colleagues suggest that
Intern Emerg Med
Fig. 2 Incidence (a) and prevalence (b) of venous thromboembolism, deep vein thrombosis, and pulmonary embolism
the patients with liver disease have a substantially increased risk of VTE (liver cirrhosis: risk ratio = 1.74, 95 % CI 1.54–1.95; non-cirrhotic liver disease: risk
ratio = 1.87, 95 % CI 1.73–2.03) [5]. By contrast, the remaining two studies do not support the association between VTE and liver diseases. In a US population-based
123
Intern Emerg Med Fig. 3 Pooled incidence of venous thromboembolism (a), deep vein thrombosis (b), and pulmonary embolism (c)
123
Intern Emerg Med Table 2 Incidence of VTE in patients with liver diseases: results of subgroup analyses Subgroups
No. of studies
Pooled proportion using randomeffects model
Heterogeneity I
P
Egger:bias
Liver cirrhosis
9
0.011 (95 % CI 0.007–0.015)
87.6 % (95 % CI 78.4–91.8 %)
P \ 0.0001
2.147 (95 % CI 0.732 to 3.561) P = 0.009
Unclassified liver diseases
5
0.010 (95 % CI 0.004–0.017)
90.8 % (95 % CI 81–94.4 %)
P \ 0.0001
2.107 (95 % CI 0.475–3.739) P = 0.026
Number of patients 7 [1,000
0.006 (95 % CI 0.004–0.008)
87.4 % (95 % CI 75.4–92.2 %)
P \ 0.0001
2.781 (95 % CI 0.516–5.046) P = 0.025
Number of patients 7 \1,000 Countries
0.025 (95 % CI 0.014–0.038)
69.3 % (95 % CI 7.7–84.2 %)
P = 0.0033
1.793 (95 % CI -0.094 to 3.681) P = 0.059
2
Publication bias
Type of liver diseases
Number of patients
USA
8
0.010 (95 % CI 0.006–0.014)
91.6 % (95 % CI 86.4–94.2 %)
P \ 0.0001
2.401 (95 % CI 0.929–3.872) P = 0.007
European countries
3
0.007 (95 % CI 0.003–0.013)
90.9 % (95 % CI 71.7–95.3 %)
P \ 0.0001
N/A
Asian countries
3
0.019 (95 % CI 0.011–0.030)
10.1 % (95 % CI 0–75.5 %)
P = 0.329
N/A
High quality
8
0.012 (95 % CI 0.008–0.018)
87.1 % (95 % CI 76–91.7 %)
P \ 0.0001
2.090 (95 % CI 0.915–3.264) P = 0.005
Poor quality
6
0.007 (95 % CI 0.004–0.01)
85.7 % (95 % CI 67.9–91.7 %)
P \ 0.0001
1.759 (95 % CI 0.668–2.849) P = 0.011
ICD codes
7
0.011 (95 % CI 0.007–0.016)
92.6 % (95 % CI 87.9–95 %)
P \ 0.0001
2.513 (95 % CI 0.744–4.282) P = 0.015
Clinical charts
7
0.0096 (95 % CI 0.005–0.015)
85.5 % (95 % CI 69.9–91.2 %)
P \ 0.0001
2.290 (95 % CI 0.828–3.751) P = 0.01
Study quality
Case identification
CI confidence interval, N/A not available
case–control study, Heit et al. [38] show that the prevalence of serious liver diseases is similar between patients with and without PE (5/625 versus 6/625). The investigators suggest that patients with serious liver diseases are 90 % less likely to develop VTE. In another populationbased study using the Nationwide Inpatient Sample database, Shaheen et al. [39] report that patients with cirrhosis are less likely to develop postoperative VTE (age/sexadjusted incidence rate ratio = 0.52, 95 % CI 0.48–0.56). Because the data are expressed in different ways, and the selection of case and control groups is inconsistent between studies, we could not perform any meta-analyses. However, on the basis of these aforementioned data, it appears that the risk of VTE is increased by the presence of liver diseases. In the present systematic review, we tried our best to collect all available evidence to analyze the epidemiology of VTE in patients with liver diseases. The major finding of our systematic review and meta-analysis is that about 1 % of patients with liver diseases develop or are diagnosed
with VTE during their hospitalizations. Compared to previous narrative reviews and every single study with a different sample size and enrollment period, the results of the systematic review may be more generalizable to patients with liver diseases. Another finding of our study is that the incidence of DVT and PE is 0.6 and 0.28 %, respectively; and the prevalence of DVT and PE is 0.7 and 0.36 %, respectively. Considering that their respective 95 % CI values overlap, we can not arbitrarily suggest that DVT is more common than PE. The results of the subgroup analyses also indicate a slightly higher prevalence in studies involving fewer than 1,000 patients than in those involving more than 1,000 patients. But it should be noted that the studies with a small sample size may bring in more selection bias than those with a relatively large sample size. More notably, the subgroup analyses are designed to explore the source of heterogeneity. Therefore, these findings of the subgroup analyses must be interpreted with caution, because they were not confirmed by any statistical analyses.
123
Intern Emerg Med Fig. 4 Pooled prevalence of venous thromboembolism (a), deep vein thrombosis (b), and pulmonary embolism (c)
123
Intern Emerg Med Table 3 Prevalence of VTE in patients with liver diseases: Results of subgroup analyses Subgroups
No. of studies
Pooled proportion using random-effects model
Heterogeneity
Liver cirrhosis
5
0.012 (95 % CI 0.007–0.019)
Unclassified liver diseases
3
Number of patients 7 [1,000
2
Publication bias P
Egger:bias
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
23.698 (95 % CI -60.156 to 107.553) P = 0.435
0.008 (95 % CI 0.006–0.012)
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
N/A
0.009 (95 % CI 0.007–0.011)
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
17.499 (95 % CI -32.658 to 67.656) P = 0.411
1
0.047 (95 % CI 0.024–0.080)
N/A
N/A
N/A
USA
7
0.009 (95 % CI 0.007–0.011)
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
17.499 (95 % CI -32.658 to 67.656) P = 0.411
Asian countries
1
0.047 (95 % CI 0.024–0.080)
N/A
N/A
N/A
High quality
2
0.025 (95 % CI 0.002–0.074)
N/A
P = 0.0003
N/A
Poor quality
6
0.009 (95 % CI 0.006–0.011)
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
25.304 (95 % CI -46.135 to 96.743) P = 0.3811
Case identification ICD codes
6
0.009 (95 % CI 0.006–0.011)
99.9 % (95 % CI 99.9–99.9 %)
P \ 0.0001
25.304 (95 % CI -46.135 to 96.743) P = 0.3811
Clinical charts
2
0.025 (95 % CI 0.002–0.074)
N/A
P = 0.0003
N/A
I
Type of liver diseases
Number of patients
Number of patients \ 1,000 Countries
Study quality
CI confidence interval, N/A not available
To the best of our knowledge, this study is the first systematic review and meta-analysis to explore the epidemiology of VTE in liver diseases. Compared to a previous narrative review by Tripodi et al. [9], the strengths of our systematic review are as follows: (1) we performed an extensive literature search via three major databases to maximize the number of retrieved papers; (2) except for the description of data from different studies, we conducted the data synthesis of all available evidence by means of metaanalysis. Additional strengths include: (1) the eligibility criteria of our study are rigorous, some specific conditions affecting the development of VTE are excluded; (2) only a random-effect model is employed to produce a conservative result; and (3) a detailed questionnaire involving eight critical issues is used to assess the study quality. Our study has several limitations. First, a statistically significant heterogeneity among studies is found in nearly all meta-analyses. Although the subgroup analyses were conducted according to different variables, the potential reasons remain obscure. Indeed, only the subgroup analysis regarding the incidence of VTE in liver diseases in Asian studies does not demonstrate any significant heterogeneity. However, it should be noted that statistically significant heterogeneity was commonly observed in nearly all
previous proportion meta-analyses [40–42]. Second, five studies were published in abstract or letter forms. Thus, much of the information regarding the characteristics of these studies was often missing. Third, only one study with a relatively small sample size was prospective, suggesting that it had a prior study design to observe the incidence of VTE. But the remaining studies were retrospective, suggesting the presence of recall bias. Fourth, enrollment spanned more than 10 years in five studies. During a relatively long period of time, the diagnostic methods of VTE were evolving, which potentially increased the heterogeneity of the results. Indeed, the study by Wu et al. [6] demonstrates an increasing trend in the prevalence of VTE in patients with liver diseases. Fifth, the level of statistical significance may be heavily weighted by the sample size. Thus, the results of meta-analyses are subdivided according to a sample size of fewer than or more than 1,000 patients. The epidemiology of VTE is higher in studies with \1,000 patients than in those with [1,000 patients (incidence: 2.5 versus 0.6 %; prevalence: 4.7 versus 0.9 %), suggesting the potential chance in studies with a small sample size. Sixth, the results may be influenced by the methods for identifying the VTE cases. The subgroup meta-analyses are separately conducted in the studies using ICD codes and
123
Intern Emerg Med
clinical charts. However, the incidence of VTE is similar between the two groups (1.1 versus 0.96 %). Although the prevalence of VTE is lower in the studies using ICD codes than in those using clinical charts (0.9 versus 2.5 %), their 95 % CIs overlap. In conclusion, our study suggests that VTE will occur in 1 % of patients with liver diseases during their hospitalizations. Considering an elevated risk of VTE in liver cirrhosis, further prospective studies might be helpful to explore the role of VTE prophylaxis. Conflict of interest
16.
17.
18. 19.
None. 20.
References 21. 1. Tripodi A, Mannucci PM (2011) The coagulopathy of chronic liver disease. N Engl J Med 365:147–156 2. Lisman T, Caldwell SH, Burroughs AK, Northup PG, Senzolo M, Stravitz RT et al (2010) Hemostasis and thrombosis in patients with liver disease: the ups and downs. J Hepatol 53:362–371 3. Northup PG, Caldwell SH (2013) Coagulation in liver disease: a guide for the clinician. Clin Gastroenterol Hepatol 11:1064–1074 4. Amitrano L, Guardascione MA, Brancaccio V, Balzano A (2002) Coagulation disorders in liver disease. Semin Liver Dis 22:83–96 5. Sogaard KK, Horvath-Puho E, Gronbaek H, Jepsen P, Vilstrup H, Sorensen HT (2009) Risk of venous thromboembolism in patients with liver disease: a nationwide population-based case-control study. Am J Gastroenterol 104:96–101 6. Wu H, Nguyen GC (2010) Liver cirrhosis is associated with venous thromboembolism among hospitalized patients in a nationwide US study. Clin Gastroenterol Hepatol 8:800–805 7. Tripodi A, Primignani M, Chantarangkul V, Dell’Era A, Clerici M, de Franchis R et al (2009) An imbalance of pro- vs anticoagulation factors in plasma from patients with cirrhosis. Gastroenterology 137:2105–2111 8. Tripodi A, Primignani M, Lemma L, Chantarangkul V, Mannucci PM (2013) Evidence that low protein C contributes to the procoagulant imbalance in cirrhosis. J Hepatol 59:265–270 9. Tripodi A, Anstee QM, Sogaard KK, Primignani M, Valla DC (2011) Hypercoagulability in cirrhosis: causes and consequences. J Thromb Haemost 9:1713–1723 10. Dabbagh O, Oza A, Prakash S, Sunna R, Saettele TM (2010) Coagulopathy does not protect against venous thromboembolism in hospitalized patients with chronic liver disease. Chest 137:1145–1149 11. Northup PG, McMahon MM, Ruhl AP, Altschuler SE, VolkBednarz A, Caldwell SH et al (2006) Coagulopathy does not fully protect hospitalized cirrhosis patients from peripheral venous thromboembolism. Am J Gastroenterol 101:1524–1528 quiz 1680 12. Gopal K, Broder A, Benias P, Min AD (2010) Practice patterns among medicine housestaff in prevention of venous thromboembolism among hospitalized patients with end-stage liver disease. Gastroenterology 138:S457 13. Goldhaber SZ, Bounameaux H (2012) Pulmonary embolism and deep vein thrombosis. Lancet 379:1835–1846 14. Garcia-Fuster MJ, Abdilla N, Fabia MJ, Fernandez C, Oliver V (2008) Forner MJ [Venous thromboembolism and liver cirrhosis]. Rev Esp Enferm Dig 100:259–262 15. Ali M, Ananthakrishnan AN, McGinley EL, Saeian K (2011) Deep vein thrombosis and pulmonary embolism in hospitalized
123
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
patients with cirrhosis: a nationwide analysis. Dig Dis Sci 56:2152–2159 Gulley D, Teal E, Suvannasankha A, Chalasani N, Liangpunsakul S (2008) Deep vein thrombosis and pulmonary embolism in cirrhosis patients. Dig Dis Sci 53:3012–3017 Lesmana CR, Inggriani S, Cahyadinata L, Lesmana LA (2010) Deep vein thrombosis in patients with advanced liver cirrhosis: a rare condition? Hepatol Int 4:433–438 Saleh T, Matta F, Alali F, Stein PD (2011) Venous thromboembolism with chronic liver disease. Am J Med 124:64–68 Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D et al (2000) Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of observational studies in epidemiology (MOOSE) group. JAMA 283:2008–2012 Amitrano L, Brancaccio V, Guardascione MA, Margaglione M, Iannaccone L, D’Andrea G et al (2000) Inherited coagulation disorders in cirrhotic patients with portal vein thrombosis. Hepatology 31:345–348 Zocco MA, Di Stasio E, De Cristofaro R, Novi M, Ainora ME, Ponziani F et al (2009) Thrombotic risk factors in patients with liver cirrhosis: correlation with MELD scoring system and portal vein thrombosis development. J Hepatol 51:682–689 Ahmed S, Mehta V, Vendetti N, Ann Phang K, Su Y, Kim H et al (2012) Risk of venous thromboembolism (VTE) in patients with chronic hepatitis C (CHC. Pharmacoepidemiol Drug Saf 21:94 Aldawood A, Arabi Y, Aljumah A, Alsaadi A, Rishu A, Aldorzi H et al (2011) The incidence of venous thromboembolism and practice of deep venous thrombosis prophylaxis in hospitalized cirrhotic patients. Thromb J 9:1 Al-Dorzi HM, Bhat S, Tamim H, Khokar S, Hussain S, Al-Dawood A et al (2010) Practices of venous thromboembolism prophylaxis and incidence in critically ill cirrhotic patients. Am J Respir Crit Care Med 181 Anthony Lizarraga W, Dalia S, Reinert SE, Schiffman FJ (2010) Venous thrombosis in patients with chronic liver disease. Blood Coagul Fibrinolysis 21:431–435 Barclay SM, Jeffres MN, Nguyen K, Nguyen T (2013) Evaluation of pharmacologic prophylaxis for venous thromboembolism in patients with chronic liver disease. Pharmacotherapy 33:375–382 Girleanu I, Trifan A, Cojocariu C, Singeap AM, Sfarti C, Stanciu C (2012) The risk of thrombotic events in patients with liver cirrhosis. Rev Med Chir Soc Med Nat Iasi 116:991–996 Kohsaka S, Nagai T, Yaegashi M, Fukuda K (2012) Pulmonary embolism and deep venous thrombosis in hospitalized patients with liver cirrhosis. Hepatology Research 42:433–434 Kumar G, Kumar N, Deshmukh A, Sakhuja A, Taneja A, Padala K et al (2010) Is cirrhosis protective for venous thromboembolism? Analysis from a national inpatient sample. Chest 138 Ponziani FR, Zocco MA, Garcovich M, Roccarina D, Gasbarrini A (2013) Epidemiology of venous thrombotic events and pulmonary embolism among hospitalized cirrhotic patients: a single center experience. Digestive and Liver Disease 45:S41 Shah NL, Northup PG, Caldwell SH (2012) A clinical survey of bleeding, thrombosis, and blood product use in decompensated cirrhosis patients. Ann Hepatol 11:686–690 Smith CB, Hurdle AC, Kemp LO, Sands C, Twilla JD (2013) Evaluation of venous thromboembolism prophylaxis in patients with chronic liver disease. J Hosp Med 8:569–573 Walsh KA, Lewis DA, Clifford TM, Hundley JC, Gokun Y, Angulo P et al (2013) Risk factors for venous thromboembolism in patients with chronic liver disease. Ann Pharmacother 47:333–339 Algahtani FH, Bayoumi N, Abdelgadir A, Al-Nakshabandi N, Al Aseri Z, Al Ghamdi M et al (2013) Clinical characteristics and
Intern Emerg Med
35.
36.
37.
38.
risk factors of pulmonary embolism: data from a Saudi tertiarycare center. J Thromb and Haemost 11:388–390 Huerta C, Johansson S, Wallander MA, Garcia Rodriguez LA (2007) Risk factors and short-term mortality of venous thromboembolism diagnosed in the primary care setting in the United Kingdom. Arch Intern Med 167:935–943 Ramagopalan SV, Wotton CJ, Handel AE, Yeates D, Goldacre MJ (2011) Risk of venous thromboembolism in people admitted to hospital with selected immune-mediated diseases: recordlinkage study. BMC Medicine 9 Ocak G, Vossen CY, Verduijn M, Dekker FW, Rosendaal FR, Cannegieter SC et al (2013) Risk of venous thrombosis in patients with major illnesses: results from the MEGA study. J Thromb Haemost 11:116–123 Heit JA, Silverstein MD, Mohr DN, Petterson TM, O’Fallon WM, Melton LJ 3rd (2000) Risk factors for deep vein thrombosis and pulmonary embolism: a population-based case-control study. Arch Intern Med 160:809–815
39. Shaheen AAM, Tanyingoh D, Dixon E, Kaplan GG, Myers RP (2012) Medical injuries among hospitalized patients with cirrhosis: a population-based study. Hepatology 56:1083A–1084A 40. Ford AC, Marwaha A, Lim A, Moayyedi P (2010) What is the prevalence of clinically significant endoscopic findings in subjects with dyspepsia? Systematic review and meta-analysis. Clin Gastroenterol Hepatol 8:830–837 837 e831–832 41. Ford AC, Marwaha A, Lim A, Moayyedi P (2010) Systematic review and meta-analysis of the prevalence of irritable bowel syndrome in individuals with dyspepsia. Clin Gastroenterol Hepatol 8:401–409 42. Larney S, Kopinski H, Beckwith CG, Zaller ND, Jarlais DD, Hagan H et al (2013) Incidence and prevalence of hepatitis C in prisons and other closed settings: results of a systematic review and meta-analysis. Hepatology 58:1215–1224
123
