Original article 419
Association between thrombin-activatable fibrinolysis inhibitor gene polymorphisms and venous thrombosis risk: a meta-analysis Wei Wanga,M, He Maa,M, Lili Lub, Guixiang Suna, Dang Liuc, Yunti Zhouc, Yue Tonga and Zhaojun Lua Thrombin-activatable fibrinolysis inhibitor (TAFI) is an important antifibrinolytic factor that has been shown in increased concentrations to be associated with an increased risk for venous thrombosis. However, the effect of TAFI gene polymorphisms on the risk of venous thrombosis remains debatable. The aim of the current study was to evaluate the association of three single nucleotide polymorphisms: 505G>A (rs3742264), 1040 C>T (rs1926447) and S438G>A (rs2146881) with venous thrombosis risk using a meta-analysis. A systematic literature search for eligible studies published before 20 January 2015 was conducted in PubMed, EMBASE, Web of Science, WanFang database and Chinese National Knowledge Infrastructure. We assessed the possible association by pooled odds ratio and its 95% confidence interval. A total of 14 independent case–control studies including 2970 cases and 3049 controls were enrolled in the final meta-analysis. A significant reduction of venous thrombosis risk in the 505G>A polymorphism was observed under allele comparison, homozygote comparison and recessive models, but opposite results were seen in Asians. Likewise, there was a significant decreased susceptibility to venous thrombosis in the 1040C>T polymorphism in homozygote comparison and recessive models. In the subgroup analysis, the nonvenous
thromboembolism disease group showed a significantly increased venous thrombosis risk. Pooled estimates did not show evidence of association between S438G>A and venous thrombosis risk in any genetic model. This meta-analysis suggested that although the S438G>T polymorphism is not correlated with venous thrombosis risk in all models, a trend toward reduced risk still could be observed. The A allele and AA genotype of 505G>A in whites and the TT genotype of 1040C>T were significantly associated with a decreased risk of venous thrombosis, except in the non-venous thromboembolism group. Blood Coagul Fibrinolysis 27:419–430 Copyright ß 2016 Wolters Kluwer Health, Inc. All rights reserved.
Introduction
decades ago. It is a new principal antifibrinolytic factor that plays an important role in the regulation of the fibrinolytic system [4]. A variety of studies have demonstrated that TAFI can be converted to an active carboxypeptidase by enzymes such as thrombin and plasmin. After being activated by thrombin, TAFI suppresses fibrinolysis by removing carboxy-terminal lysine vestiges from partly degraded fibrin; thus, hindering combination of the fibrinolytic parts plasminogen and tissuetype plasminogen activator [5,6]. Increased plasma antigen levels of TAFI have been found to be associated with a risk of venous thrombosis [7]. The TAFI gene, located in chromosome 13 (13q14.11), spans approximately 48 kb and contains 11 exons [8]. In total, 19 single nucleotide polymorphisms (SNPs) have been reported: ten in the 50 -flanking region, six in the coding region and three in the 30 untranslated region [9]. TAFI plasma levels are also determined genetically; genetic factors could explain more than 25% of their variability [4,10–12]. Recently, effects of TAFI polymorphisms affecting antigen levels on venous thrombosis have been reported.
Venous thrombosis is a common and serious disorder that contributes to clotting of the blood in the venous system [1]. Its manifestations not only include two main clinical presentations, deep vein thrombosis (DVT) and pulmonary embolism, but also include some manifestations in which thrombosis is situated in rare venous sites, such as cerebral venous thrombosis or splanchnic vein thrombosis. Venous thrombosis can be seen as a common but complex disease in which environmental and multiple genetic risk factors play a vital role in development. Recently, several emerging candidate gene polymorphisms have been reported, including lipoprotein (a), tissue factor, thrombomodulin, EPCR, and fibrinogen as well as thrombin-activatable fibrinolysis inhibitor (TAFI). TAFI, a 56-kDa zymogen based on 423 amino acids, is primarily synthesized in the liver [2]. Eaton et al. [3] first isolated and characterized the cDNA of TAFI two Wei Wang and He Ma have contributed equally to the manuscript.
0957-5235 Copyright ß 2016 Wolters Kluwer Health, Inc. All rights reserved.
Blood Coagulation and Fibrinolysis 2016, 27:419–430 Keywords: meta-analysis, polymorphism, thrombin-activatable fibrinolysis inhibitor, venous thrombosis a
Department of Public Health, bDepartment of Gastroenterology and Department of General Practice, Xuzhou Medical College, Xuzhou, Jiangsu, China c
Correspondence to Zhaojun Lu, Department of Public Health, Xuzhou Medical College, 209 Tong-shan Road, Xuzhou, Jiangsu, 221004, China Tel: +86 13382654382; e-mail: [email protected] Received 4 August 2015 Revised 29 September 2015 Accepted 3 October 2015
DOI:10.1097/MBC.0000000000000475
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
420 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
Fig. 1
Materials and methods
Flowchart of the selection process for including articles. VT, venous thrombosis.
Search strategy
We carried out a comprehensive literature search in PubMed, EMBASE, Web of Science, WanFang database and Chinese National Knowledge Infrastructure. The last search was performed on January 20, 2015. The keywords included ‘thrombin activatable fibrinolytic inhibitor OR TAFI OR CPB2 OR Carboxypeptidase’ AND ‘polymorphism OR polymorphisms OR variation OR variants’ AND ‘vein Thrombosis OR venous Thrombosis OR vein Thromboses OR Venous Thromboembolic OR pulmonary embolism OR Budd–Chiari syndrome’. There was no restriction on publication years but the languages were limited to English and Chinese in our search. Reference lists of review articles and primary studies were manually searched to identify additional eligible studies. Inclusion and exclusion criteria
All studies included in this meta-analysis were selected based on the following criteria: case–control design for human beings, evaluating the associations between the SNPs 505G>A rs3742264 and/or 1040C>T rs1926447 and/or 438G>A rs2146881 and venous thrombosis disease, supplying useful genotype frequencies and the control groups were healthy persons or patients without a history of venous thrombosis disease. Exclusion criteria included the following: no controls; duplicate publications; reviews, comments or animal studies and studies with no detailed genotype frequencies. If data were published in more than one article, only the study with the largest sample size was included. Data extraction
The following data were independently and carefully extracted by two researchers (W. W., Y. T.) from all
Identification
Several functional variants such as the 505G>A SNP and 1040 C>T SNP in the coding region and the 438G>A SNP in the promoter region are associated with TAFI plasma antigen levels. Because of this, the SNPs have been widely studied and expected to be involved in venous thrombosis disease. A number of articles investigated the association between the three SNPs and venous thrombosis risk [12–25], but the results were mixed and remained inconclusive. Meta-analysis can use a measurable method to combine the results from different articles on an identical theme; thus, supplying more authentic conclusions. Though a meta-analysis published not long ago pooled some of the original studies together, several credible case–control studies [12,15,19,23] have been omitted, and we thought a more comprehensive analysis was necessary. Therefore, we performed this meta-analysis to explore a more accurate estimation of the relationship between TAFI gene variants and venous thrombosis risk.
Records identified and screened through database retrieving (Pubmed = l7, Embase = 64, Web= 65, Wanfang = 5, CNKI= l)
Studies were exclued for improper title/abstract (n = 59) Full-text articles assessed for eligibility (n = 20) Full-text articles were exclued with reason of not for VT (n = 3) not case-control design(n = 3)
Eligibility
Screening
Potentially relevant studies after duplicates removed (n = 79)
Included
Studies included in meta-ananlysis (n = l4)
Article about 505G>A (n = l0)
Article about 11040C>T (n = l0)
Article about -438G>A(n = 6)
eligible studies: last name of first author, year of publication, original country, ethnicity, genotyping method, type of disease, source of control, Hardy–Weinberg equilibrium (HWE) in controls, numbers of cases and controls and allele as well as genotype frequencies for cases and controls. If disagreements existed, the original data were rechecked and consensus was reached through discussion. Study validity assessment
The quality of the studies that were included in this review was independently evaluated by two researchers (W.W., H.M.) according to the Newcastle Ottawa Scale (NOS) [26]. In this scale, three main items were assessed: selection, comparability and exposure. Studies with scores equal to or higher than five were considered to be of high quality. Discrepancies were settled by discussion between the two researchers. Statistical analysis
First, we tested the distribution of the genotypes for HWE by using the goodness-of-fit x2 test in controls; P < 0.05 was considered to be a violation of HWE. The strength of the association between the three SNPs and venous thrombosis risk were estimated by the odds ratio (OR) and its 95% confidence interval (95% CI). The pooled ORs were estimated for five genetic models: allele comparison (505G>A, 438G>A: A vs. G; 1040C>T: T vs. C), heterozygote comparison (505G>A, 438G>A: GA vs. GG; 1040C>T: CT vs. CC), homozygote comparison (505G>A, 438G>A: AA vs. GG; 1040C>T: TT vs. CC), dominant model (505G>A, 438G>A: GA þ AA vs. GG; 1040C>T: CT þ TT vs. CC) and recessive model
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
2001 2001 2003 2006
2006
2006 2006
2007
2007
2008
2010
2010 2012
2013
Franco Morange Zidane Morange
Verdu´
Lichy Martini
Xu
De Bruijne
Verdu´
Zhang
Hoekstra Li
Tokgoz
Turkey
European China
China
Spain
Netherlands
China
Germany Netherlands
Spain
Brazil France Netherlands France
Country
White
White Asian
Asian
White
White
Asian
White White
White
White White White White
Ethnicity
CVT
SVT VTE
VTE
DVT
SVT
Pulmonary embolism
CVT DVT
DVT
DVT VTE Pulmonary embolism VTE
Type of disease
Population-based
Population-based Hospital-based
Hospital-based
Population-based
Population-based
Population-based
Population-based Hospital-based
Population-based
Population-based Hospital-based Population-based Hospital-based
Control source
PCR-RFLP
PCR Allele-specific PCR
ARMS-PCR
Allele-specific PCR
PCR
PCR-RFLP
Allele-specific PCR PCR
PCR
PCR PCR PCR PCR
Genotyping method 438G/A 505G/A 438G/A 505G/A 1040C/T 505G/A 1040C/T 438G/A 505G/A 1040C/T 438G/A 505G/A 1040C/T 505G/A 1040C/T 438G/A 505G/A 1040C/T 505G/A 1040C/T 1040C/T 505G/A 1040C/T 505G/A 1040C/T 438G/A
SNP 388 168 66 127 127 60 61 75 471 471 471 114 114 117 118 114 131 131 339 477 97 80 80 59 59 59
Case 388 145 152 53 53 61 62 197 472 472 472 114 114 116 118 116 100 100 317 509 99 80 80 100 100 100
Control
Sample size (men)
222/201 100/78 42/67 70/22 62/25 10/7 25/9 46/103 236/211 219/215 240/257 56/71 85/79 58/48 40/53 48/60 41/22 67/33 169/180 324/370 43/46 24/28 27/15 28/40 21/45 37/50
Wild type
139/157 64/63 23/72 45/25 55/22 43/38 29/31 26/82 198/210 218/212 202/185 48/36 24/29 54/50 69/52 58/46 66/58 50/51 160/116 142/126 43/41 47/36 42/44 23/48 30/43 18/43
Heterozygous
27/30 4/4 1/13 12/6 10/6 7/16 7/22 3/12 37/51 34/45 29/30 10/7 5/6 5/18 9/13 8/10 24/20 14/16 10/21 11/13 11/12 9/16 11/21 8/12 8/12 4/7
Mutant
Genotype distribution (case/control)
Y(0.932) N(0.035) Y(0.299) Y(0.782) Y(0.730) N(0.033) Y(0.717) Y(0.670) Y(0.907) Y(0.484) Y(0.668) Y(0.406) Y(0.142) Y(0.412) Y(0.963) Y(0.780) Y(0.109) Y(0.615) Y(0.694) Y(0.564) Y(0.544) Y(0.478) Y(0.342) Y(0.677) Y(0.727) Y(0.582)
Y/N(P)
HWE
7
8 6
6
6
6
6
6 7
6
7 7 7 6
NOS
CVT, cerebral venous thrombosis; DVT, deep venous thrombosis; HWE, Hardy–Weinberg equilibrium; NOS, Newcastle–Ottawa scale; PCR, polymerase chain reaction; RFLP-PCR, restriction fragment length polymorphism; SNP, single nucleotide polymorphism; SVT, splanchnic vein thrombosis; TAFI, thrombin activatable fibrinolysis inhibitor; VTE, venous thromboembolism.
Year
Baseline characteristics of studies included in the meta-analysis
Author
Table 1
Association between TAFI gene polymorphisms and VT Wang et al. 421
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
422 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
Fig. 2
Study ID
OR (95% CI)
% weight
Morane (2001)
0.78 (0.19, 3.22)
3.11
Morange (2006)
0.63 (0.21, 187)
55.53
Verdu (2006)
0.31 (0.08, 1.14)
5.59
Martini (2006)
0.65 (0.41, 1.103) 32.56
Xu (2007)
1.81 (0.65, 5.06)
de Buijine (2007)
0.23 (0.08, 0.66) 11.71
Verdu (2008)
0.64 (0.29, 1.42)
11.09
Zhang (2010)
0.51(0.23, 1.11)
13.52
Li (2012)
0.66 (0.25, 1.75)
7.22
Tokgoz (2013)
0.95 (0.34, 2.63)
5.53
Overall (I-squared = 7.5%, P = 0.373)
0.63 (0.48, 0.82) 100.00
0.0795
1
3.94
12.6
Forest plots of meta-analysis for association between 505G>A polymorphism and venous thrombosis (AA vs. GG). A significant reduction susceptibility was observed between the 505G>A polymorphism and risk of venous thrombosis. CI, confidence interval; OR, odds ratio.
(505G>A, 438G>A: AA vs. GA þ GG; 1040C>T: TT vs. CT þ CC). The significance of the pooled ORs was examined by the Z-test. The x2-based Q test and I2 statistics were used to test the heterogeneity between studies; P < 0.10 or I2 > 50% indicated evidence of heterogeneity [27]. The fixed-effects model (Mantel–Haenszel method) was used to estimate the summary ORs when there was no significant heterogeneity [28]; otherwise, the random effect model (DerSimonian and Laird method) was used [29]. Subgroup analyses were conducted according to ethnicity, type of disease, source of control and genotyping method. Meta-regression and Galbraith plot analyses were conducted for further exploration of heterogeneity. The study characteristics included as covariates in the meta-regression were publication year and NOS score. Furthermore, sensitivity analysis was used to assess the stability of the results by deleting one single study at a time to examine the influence of single data points. Potential publication bias was evaluated using the Begg test [30] and Egger regression test [31]; an asymmetric plot and P < 0.05 were considered as statistically significant publication bias. All statistical tests were twosided, and P values A and venous thrombosis in a total of 1666 cases and 1558 controls, 10 studies [15,16,18–25] involving 1735 cases and 1707 controls for 1040C>T and 6 studies [12,14,17,18,20,25] involving 1173 cases and 1425 controls for 438G>A. There were 11 studies carried out in whites, whereas the other studies were conducted in Asians. Detailed genotype distribution and the results of the HWE test in control population are summarized in Table 1; all studies were consistent with HWE except for two studies [13,16] of 505G>A. NOS quality scores for each study ranged from 6 to 8, with all studies being classified as high quality.
Results Study selection and characteristics
Association between the 505G>A (rs3742264) polymorphism and venous thrombosis susceptibility
As shown in Fig. 1, a total of 152 relevant studies were initially identified after a systematic search; 53 studies
Ten studies with a total of 1666 cases and 1558 controls examined the association between the 505G>>A
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
0.286 0.729 0.318 20.1 0.0 12.6 0.58 (0.41, 0.81) 1.30 (0.65, 2.58) 0.61 (0.40, 0.94) 0.903 0.049 0.084 0.0 74.2 59.5 0.76 (0.63, 0.92) 1.22 (0.82, 1.84) 1.14 (0.89, 1.47) 0.387 0.383 0.889 7.5 0.0 0.0 0.54 (0.37, 0.77) 1.31 (0.64, 2.68) 0.59 (0.36,0.95) 0.865 0.046 0.043
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
3
CI, confidence interval; OR, odds ratio; PQ, P-value for heterogeneity test.
0.0 74.8 68.3 0.81 (0.66, 0.99) 1.18 (0.77, 1.82) 1.26 (0.99, 1.63) 0.743 0.095 0.388 0.77 (0.67, 0.89) 1.19 (0.87, 1.63) 0.98 (0.81, 1.18) 5 2
0.0 64.1 0.0
0.503 0.069 0.0 57.6 0.68 (0.50, 0.93) 0.61 (0.40, 0.92) 0.015 0.977 64.6 0.0 0.99 (0.84, 1.16) 0.92 (0.80, 1.06) 0.532 0.209 0.0 33.9 0.69 (0.50, 0.96) 0.49 (0.30, 0.80) 0.009 0.828 67.4 0.0 1.05 (0.89, 1.24) 0.74 (0.53, 1.04) 0.081 0.719 0.93 (0.82, 1.06) 0.74 (0.60, 0.92) 6 4
49.0 0.0
0.282 0.140 19.4 49.1 0.67 (0.50, 0.89) 0.61 (0.38, 1.00) 0.962 0.675 0.0 0.0 0.75 (0.63, 0.89) 1.39 (1.08, 1.78) 0.516 0.145 0.0 48.2 0.58 (0.43, 0.80) 0.76 (0.45, 1.27) 0.907 0.912 0.0 0.0 0.78 (0.65, 0.94) 1.52 (1.18, 1.97) 0.898 0.261 0.78 (0.69, 0.89) 1.13 (0.93, 1.37) 7 3
0.0 25.6
0.484 0.030 0.0 78.9 0.68 (0.52, 0.89) 0.52 (0.27, 1.02) 0.018 0.948 58.4 0.0 0.94 (0.72, 1.22) 0.73 (0.48, 1.09) 0.596 0.058 0.0 72.3 0.66 (0.50, 0.88) 0.46 (0.23, 0.94) 0.011 0.552 61.6 0.0 0.98 (0.74, 1.31) 0.81 (0.53, 1.24) 0.109 0.316 0.90 (0.81, 1.01) 0.73 (0.54, 0.98) 8 2
40.4 0.6
0.244 21.7 0.65 (0.51, 0.84) 0.031 51.0 0.90 (0.72, 1.12) 0.373 7.5 0.63 (0.48, 0.82) 0.022 53.7 0.95 (0.75, 1.21) 0.106 0.88 (0.79, 0.98) 10
Overall Type of disease VTE Non-VTE Ethnicity White Asian Source of control Hospital-based Population-based Genotyping method PCR RFLP-PCR Allele-specific PCR
37.9
OR (95% CI) PQ I2 (%) OR (95% CI) OR (95% CI) N
I2 (%)
PQ
OR (95% CI)
I2 (%)
PQ
OR (95% CI)
I2 (%)
PQ
AAþGA vs. GG AA vs. GG AG vs. GG A vs. G
Category
Similarly, 10 studies involving 1735 cases and 1707 controls were assessed for the association between the 1040C>T polymorphism and venous thrombosis risk. We found a significantly decreased venous thrombosis susceptibility in homozygote comparison (TT vs. CC: OR ¼ 0.63, 95% CI 0.42–0.94, P ¼ 0.022 Fig. 4) and recessive models (TT vs. CT þ CC: OR ¼ 0.67, 95% CI 0.52–0.86, P ¼ 0.002 Fig. 4); no significant association was found in allele comparison (T vs. C: OR ¼ 0.85, 95%
Results of meta-analysis for TAFI 505G> A polymorphism and venous thrombosis risk
Association between the 1040C>T (rs1926447) polymorphism and venous thrombosis susceptibility
Table 2
Subsequently, we performed subgroup analyses to avoid the potential impact of confounding factors. Heterogeneity vanished in the ethnicity subgroup, suggesting that ethnicity might contribute mainly to the heterogeneity. In white populations, we found a decreased venous thrombosis risk in all genetic models (Table 2). In Asian populations, there was a significantly increased venous thrombosis risk observed in the heterozygote comparison (GA vs. GG: OR ¼ 1.52, 95% CI 1.18–1.97, P ¼ 0.001 Fig. 3) and dominant models (GA þ AA vs. GG:OR ¼ 1.39, 95% CI 1.08–1.78, P ¼ 0.009 Fig. 3). When stratifying by type of disease, significantly reduced risk was found among both venous thromboembolism (VTE) (homozygote comparison, recessive model) and the non-VTE group (homozygote comparison, recessive model). When it came to sources of control, the TAFI 505G>A polymorphism was significantly associated with a decreased venous thrombosis risk in both the hospital-based group (homozygote comparison, recessive model) and the population-based group (allele comparison, homozygote comparison, recessive model). For the polymerase chain reaction (PCR) group, significantly reduced risk was found in all genetic models. For the allelespecific PCR group, similar results were observed in homozygote comparison and recessive models. However, no significant association between the 505G>A polymorphism and venous thrombosis risk was seen in the restriction fragment length polymorphism-PCR group. The main results of the subgroup analyses are shown in Table 2.
AA vs. GA þ GG
polymorphism and venous thrombosis risk. There was a significant heterogeneity within studies of heterozygote comparison and dominant models, thus the randomeffects model was used. A significant reduction of venous thrombosis susceptibility was observed in allele comparison (A vs. G: OR ¼ 0.88, 95% CI 0.79–0.98, P ¼ 0.018), homozygote comparison (AA vs. GG: OR ¼ 0.63, 95% CI 0.48–0.82, P ¼ 0.001 Fig. 2) and recessive models (AA vs. GA þ GG: OR ¼ 0.65, 95% CI 0.51–0.84, P ¼ 0.001); no significant association was found in heterozygote comparison (GA vs. GG: OR ¼ 0.95, 95% CI 0.75–1.21, P ¼ 0.681) and dominant models (GA þ AA vs. GG: OR ¼ 0.90, 95% CI 0.72–1.12, P ¼ 0.347).
I2 (%)
PQ
Association between TAFI gene polymorphisms and VT Wang et al. 423
424 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
Fig. 3
(a) Studay ID
OR(95% CI)
% weight
White Morange (2001)
0.79 (0.50, 1.25) 11.96
Morange (2006)
0.57 (0.29, 1.12) 7.82
Verdu (2006)
0.79 (0.27, 2.29) 4.18
Martini (2006)
0.84 (0.64, 1.10) 16.56
de bruijine (2007)
0.89 (0.52, 1.54) 10.19
Verdu (2008)
0.61 (0.33, 1.14) 8.70
Tokgoz (2013)
0.68 (0.34, 1.37) 7.70
Subtotal (I-squared = 0.0%, P = 0.907) . Asian
0.78 (0.65, 0.94) 67.12
Xu (2007)
1.69 (0.97, 2.95) 9.93
Zhang (2010)
1.47 (1.07, 2.02) 15.31
Li (2012)
1.52 (0.76, 3.06) 7.65
Subtotal (I-squared = 0.0%, P = 0.912) .
1.52 (1.18, 1.97) 32.88
Overall (I-squared = 53.7%, P = 0.22)
0.95 (0.75, 1.21) 100.00
NOTE: weights are from random effects analysis
0.274
1
3.64
(b) Studay ID
OR(95% CI)
% weight
White Morange (2001)
0.79 (0.51, 1.24) 11.72
Morange (2006)
0.58 (0.30, 1.11) 7.78
Verdu (2006)
0.65 (0.23, 1.83) 3.87
Martini (2006)
0.81 (0.62, 1.104) 17.04
De bruijine (2007)
0.72 (0.43, 1.20) 10.16
Verdu (2008)
0.62 (0.34, 1.13) 8.58
Tokgoz (2013) Subtotal (I-squared = 0.0%, P = 0.907)
0.75 (0.63, 0.89) 66.95
0.74 (0.39, 1.41) 7.78
. Asian Xu (2007)
1.71 (1.01, 2.90) 9.95
Zhang (2010)
1.32 (0.97, 1.80) 15.54 1.26 (0.65, 2.44) 7.56
Li (2012) Subtotal (I-squared = 0.0%, P = 0.912) .
1.39 (1.108, 1.78) 33.05
Overall (I-squared = 51.0%, P = 0.31)
0.90 (0.75, 1.12) 100.00
NOTE: weights are from random effects analysis
0.229
1
4.36
Forest plot for the subgroup analysis of ethnicity (a) GA vs. GG; (b) GA þ AA vs. GG. The 505G>A polymorphism was found to be associated with increased risk of Asians CI, confidence interval; OR, odds ratio.
CI 0.68–1.05, P ¼ 0.133), heterozygote comparison (CT vs. CC: OR ¼ 0.94, 95% CI 0.72–1.23, P ¼ 0.636) and dominant models (CT þ TT vs. CC: OR ¼ 0.87,
95% CI 0.65–1.15, P ¼ 0.325). Q-test showed significant heterogeneity in all genetic models except the recessive model.
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Association between TAFI gene polymorphisms and VT Wang et al. 425
Fig. 4
(a) Study ID
OR (95% CI)
Morange (2006)
0.67 (0.22, 2.05)
8.01
Verdu (2006)
0.11 (0.04, 0.36)
7.75
Martini (2006)
0.74 (0.46, 1.20)
16.56
Xu (2007)
0.77 (0.23, 2.64)
7.06
de bruijine (2007)
0.92 (0.36, .2.36)
9.75
Verdu (2008)
0.43 (0.19, 0.99)
11.17
Zhang (2010)
0.97 (0.43, 2.19)
11.34
Hoekstra (2010)
0.98 (0.39, 2.46)
10.06
Li (2012)
0.29 (0.11, 0.76)
9.52
Tokgoz (2013)
1.43 (0.51, 4.02)
8.78
Overall (I-squared = 47.9%, P = 0.045)
0.63 (0.42, 0.94)
100.00
% weight
NOTE: weights are from random effects analysis
0.0366
1
27.3
(b) Study ID
OR (95% CI)
% weight
Morange (2006)
0.67 (0.23, 1.95)
5.15
Verdu (2006)
0.24 (0.09, 0.61)
12.76
Martini (2006)
0.74 (0.46, 1.18)
27.55
Xu (2007)
0.83 (0.24, 2.79)
3.79
de bruijine (2007)
0.67 (0.27, 1.63)
7.93
Verdu (2008)
0.63 (0.29, 1.36)
10.70
Zhang (2010)
0.90 (0.40, 2.03)
8.12
Hoekstra (2010)
0.93 (0.39, 2.22)
6.96
Li (2012)
0.45 (0.20, 1.00)
11.96
Tokgoz (2013)
1.15 (0.44, 3.00)
5.08
Overall (I-squared = 0.0%, P = 0.510)
0.67 (0.52, 0.86)
100.00
0.0917
1
10.9
Forest plots of meta-analysis for association between 1040C>T polymorphism and venous thrombosis (a) TT vs. CC; (b) TT vs. CT þ CC. A significant reduction susceptibility was observed between the 1040C>T variant and risk of venous thrombosis.
As shown in Table 3, in whites, no significant association was found for any genetic models except the recessive model. Similarly, no significant association was observed in any genetic model in Asians. Moreover, in the stratified analysis based on type of disease, a significantly decreased risk of venous thrombosis was observed under the allele comparison, homozygote comparison and recessive model in the VTE group. It is worth noting that increased venous thrombosis risk was found in the heterozygote comparison (CT vs. CC:
OR ¼ 1.45, 95% CI 1.02–2.05, P ¼ 0.037) in the nonVTE group. Regarding the source of control, we found a significantly decreased venous thrombosis risk in the hospital-based group under the homozygote comparison and recessive model, whereas only under the recessive model in the population-based group. In the subgroup analyses on genotyping method, no significant association was observed in any genetic model of the three subgroups except for the recessive model in the PCR group.
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
0.250 0.675 0.490 0.64 (0.46, 0.88) 1.01 (0.48, 2.15) 0.63 (0.40, 1.00)
25.7 0.0 0.0
Heterogeneity in the four genetic models mentioned above could still be detected in the subgroup analyses. A meta-regression was used next, but it failed to confirm that the publication year and NOS score were the sources of heterogeneity. Finally, we performed Galbraith plot analysis for further detection of heterogeneity. The results found that, four studies [16,21,22,24] for the allele comparison, one study [16] for the homozygote comparison, three studies [16,20,21] for heterozygote comparison and three studies [16,21,24] for the dominant model contributed mainly to heterogeneity (Fig. 5). After we excluded these studies, heterogeneity can ignored among these genetic models (I2 ¼ 0.0%, P ¼ 0.659; I2 ¼ 0.0%, P ¼ 0.456; I2 ¼ 15.8%, P ¼ 0.309; and I2 ¼ 4.8%, P ¼ 0.390).
69.6 52.8 86.2
Association between the S438G>A (rs2146881) polymorphism and venous thrombosis susceptibility
The association between the 438G>A polymorphism and susceptibility to venous thrombosis was analyzed in six independent studies with 1173 cases and 1425 controls. Significant between-study heterogeneity was observed under allele comparison, heterozygote comparison and dominant models. None of the genetic models produced a significant association between the 438G/A polymorphism and venous thrombosis risk (A vs. G: OR ¼ 0.91, 95% CI 0.80–1.03, P ¼ 0.147; AA vs. GG: OR ¼ 0.80, 95% CI 0.58–1.11, P ¼ 0.185; GA vs. GG: OR ¼ 0.92, 95% CI 0.79–1.09, P ¼ 0.345; GA þ AA vs. GG: OR ¼ 0.91, 95% CI 0.78–1.06, P ¼ 0.219; AA vs. GA þ GG: OR ¼ 0.83, 95% CI 0.60–1.14, P ¼ 0.238). In the subgroup analyses, the same results were obtained (Table 4). Nevertheless, the results still show a trend toward decreased susceptibility to venous thrombosis.
CI, confidence interval; OR, odds ratio; PQ, P-value for heterogeneity test.
57.9 48.6 83.3 73.1 42.6 84.9 5 2 3
0.84 (0.61, 1.16) 1.01 (0.64, 1.59) 0.76 (0.45, 1.28)
0.005 0.187 0.001
1.02 (0.70, 1.49) 1.05 (0.55, 2.01) 0.72 (0.35, 1.51)
0.050 0.163 0.002
0.59 (0.32, 1.12) 1.11 (0.50, 2.44) 0.51 (0.26, 1.02)
61.7 0.0 47.3
0.033 0.454 0.150
0.91 (0.59, 1.39) 1.05 (0.55, 1.98) 0.67 (0.31, 1.46)
0.010 0.146 0.001
0.796 0.167 0.0 38.1 0.70 (0.50, 0.97) 0.63 (0.43, 0.93) 50.8 81.2 35.1 76.7 54.2 80.5 5 5
0.90 (0.72, 1.12) 0.81 (0.52, 1.25)
0.068 0.000
1.00 (0.78, 1.28) 0.90 (0.49, 1.63)
0.188 0.002
0.69 (0.48, 0.97) 0.59 (0.27, 1.30)
0.0 69.9
0.427 0.010
0.93 (0.71, 1.22) 0.81 (0.43, 1.53)
0.087 0.000
0.354 0.455 9.9 0.0 0.67 (0.50, 0.89) 0.66 (0.40, 1.11) 72.0 74.6 65.1 66.6 71.6 76.7
10
Overall Type of disease VTE Non-VTE Ethnicity White Asian Source of control Hospital-based Population-based Genotyping method PCR RFLP-PCR Allele-specific PCR
7 3
0.84 (0.64, 1.11) 0.84 (0.53, 1.34)
0.002 0.014
0.95 (0.66, 1.36) 0.88 (0.51, 1.51)
0.009 0.050
0.64 (0.39, 1.05) 0.61 (0.28, 1.30)
55.9 44.8
0.034 0.163
0.87 (0.60, 1.28) 0.82 (0.45, 1.47)
0.002 0.019
0.396 0.710 4.0 0.0 0.62 (0.46, 0.82) 0.88 (0.52, 1.48) 74.8 0.0 66.1 0.0 76.2 0.0
N
7 3
0.74 (0.56, 0.98) 1.15 (0.90, 1.47)
0.000 0.800
0.79 (0.57, 1.09) 1.45 (1.02, 2.05)
0.007 0.550
0.50 (0.31, 0.82) 1.07 (0.61, 1.86)
53.7 0.0
0.044 0.804
0.71 (0.50, 1.02) 1.37 (0.98, 1.91)
0.001 0.608
0.510 0.0 0.67 (0.52, 0.86) 69.9 61.8 70.3 0.85 (0.68, 1.05)
PQ I (%) OR (95% CI)
0.000
0.94 (0.72, 1.23)
0.005
0.63 (0.42, 0.94)
47.9
0.045
0.87 (0.65, 1.15)
0.000
I2 (%) OR (95% CI) I (%) OR (95% CI) I (%) I (%) OR (95% CI)
PQ
OR (95% CI)
2 2
CT vs. CC T vs. C
2
Category
Table 3
Results of meta-analysis for TAFI 1040C>T polymorphism and venous thrombosis risk
TT vs. CC
PQ
2
TT þ CT vs. CC
PQ
TT vs. CT þ CC
PQ
426 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
Heterogeneity still existed in the subgroup analyses; the publication year and NOS score were not viewed as the reasons for heterogeneity when we carried out a metaregression. However, Galbraith plot analysis successfully distinguished the deviants that might have led to the heterogeneity (Fig. 6). For allele comparison and dominant models, heterogeneity could be ascribed mainly to one study [14]; for heterozygote comparison, another study [20] might have been responsible for the heterogeneity. Homogeneity test I2 values deceased more or less after removing each study from the total dataset (I2 ¼ 37.0%, P ¼ 0.174; I2 ¼ 57.9%, P ¼ 0.050 and I2 ¼ 59.9%, P ¼ 0.041). Sensitivity analysis
Analyzing the three SNPs by deleting each study from the total dataset, we did not notice any significant difference in the pooled estimates. For 505G>A, when excluding the studies that deviated from HWE, the overall results were not materially affected. The results indicated that this meta-analysis provided reliable evidence (data not shown).
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Association between TAFI gene polymorphisms and VT Wang et al. 427
Fig. 5
Fitted values
b/se(b)
Tokgoz
Morange Xu
Tokgoz
Mattini
−2 Li
Fitted values
b/se(b)
2
Bruijine
Hoekstra
0
b/se(b)
(b)
2
b/se(b)
(a)
0
HoekstraZhang Bruijine Xu Morange Matini
−2
Verdu08
Verdu08
Li
−2.98221 Verdu06
0
1/se(b)
2.16849 2
0
10.0556 Fitted values
b/se(b)
(c)
Verdu06
−3.71027
(d)
Bruijine
Bruijine
Zhang
Mattini Hoekstra
Xu Li
−2
Verdu06
0
b/se(b)
0
Hoekstra
0
Morange
Matini
Xu
−2 Li
−2.24125
Verdu08 Verdu06
Verdu08
1 /se(b)
Zhang
Tokgoz
Hoekstra Morange
4.05382 Fitted values
b/sec(b) 2
Tokgoz
b/se(b)
1/se(b)
7.34788
0
1/se(b)
7.65241
Galbraith plots of 1040C>T polymorphism and venous thrombosis risk in different contrast models. (a) Allele comparison (T vs. C); (b) homozygote comparison (TT vs. CC); (c) heterozygote comparison (CT vs. CC); (d) dominant model (TT þ CT vs. CC).
Publication bias
Publication bias was assessed by Begg’s funnel plot and Egger’s test; the results are shown in Table 5. The symmetric funnel plot suggested there was no obvious publication bias, whereas Egger’s test also showed no statistical significance for the detection of publication bias between the three genetic variations and venous thrombosis risk in any of the models.
Discussion In addition to plasminogen activator inhibitor-1, the major physiological inhibitor of plasminogen activation, TAFI is another important influencing factor of impaired fibrinolysis, which can result in thrombotic disorders. Plasma levels of TAFI resulting from the genetic variations have been reported in men with thrombotic diseases [4]. However, there is still no definite evidence concerning the association between TAFI polymorphisms and venous thrombosis risk. Our previous metaanalysis performed in the same way investigated the association between the three TAFI polymorphisms and venous thrombosis risk [32]. However, some defects existed in the literature included; an article that did not report a venous thrombosis case–control study and mainly researched pediatric oncology patients [33] was falsely included. Considering this, we performed this
meta-analysis to produce more credible evidence for or against an association between TAFI genetic variants and venous thrombosis. In this meta-analysis, 14 studies including 2339 cases and 4048 controls were analyzed. As for association between the 505G>A polymorphism and venous thrombosis susceptibility, a protective relationship between them was identified in allele comparison, homozygote comparison and recessive models, suggesting that the A allele and AA genotype were significantly associated with a decreased risk of venous thrombosis, which might result from the role of the 505A allele in the alteration of TAFI levels. An earlier study reported that the A allele was associated with lower TAFI levels than the G allele [20]. Stratified analyses also suggested protective effects on venous thrombosis in overall studies; however, Asians showed an opposite result in the subgroup analysis; this might have resulted from ethnic differences. As for association between 1040C/T and venous thrombosis risk, the meta-analysis results showed that TT genotype carriers had a 37% lower risk of venous thrombosis than those with CC genotype and a 33% lower risk than the combined group of CT and CC carriers. In the subgroup analysis, similar results could be found except that the non-VTE disease group showed a significantly increased venous thrombosis risk. It has been shown that the
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
0.428 0.0 0.007 75.5 0.83 (0.59, 1.17) 0.013
26.1
0.255
0.91 (0.62, 1.34)
Heterogeneity is hard to avoid in a meta-analysis and the key is determining the sources of heterogeneity. Significant heterogeneity was found for the association of 505G>A and venous thrombosis in two genetic models. By conducting stratified analyses, we found no heterogeneity existing in the ethnicity subgroup, suggesting that ethnicity contributed to the source of heterogeneity. However, for the other two SNPs, we failed to find the source of heterogeneity by stratified or meta-regression analyses. We finally discovered the source of heterogeneity by Galbraith plot analysis. For the four genetic models of 1040C>T, all I2 values were less than 50%, and P values were greater than 0.10 after eliminating the studies judged as outliers by the Galbraith plots. Although I2 values were not reduced as was expected in 438G>A, the heterogeneity was acceptable in the end.
0.82 (0.58, 1.16) 0.88 (0.62, 1.26)
0.79 (0.61, 1.03)
0.90 (0.68, 1.20)
3 3
5
4
CI, confidence interval; OR, odds ratio; PQ, P-value for heterogeneity test.
72.3 0.95 (0.66, 1.38) 0.013 72.1
0.86 (0.68, 1.07) 6
Overall Type of disease VTE Non-VTE Source of control Population-based Genotyping method PCR
OR (95% CI)
1040C>T SNP was strongly associated with TAFI levels [11,34], yet increased plasma TAFI levels have been found to be associated with a risk of venous thrombosis [7]. Therefore, it can be concluded from this study that the polymorphisms 505G>A and 1040C>T, which are seen more in normal population groups, may reduce thrombus formation by causing reductions in TAFI activity. As for association between 438G>A and venous thrombosis risk, although the 438A allele has been associated with decreased TAFI antigen levels [14]. The results showed that neither the overall analysis nor the subgroup analysis supported a genetic association between 438G>A and susceptibility to venous thrombosis.
0.83 (0.59, 1.17)
0.609 0.0 0.042 0.70 (0.46, 1.05) 0.079 52.1
0.78 (0.55, 1.11)
26.1
0.235
0.0
0.457
0.76 (0.54, 1.08)
59.6
0.75 (0.51, 1.13)
0.253 0.907 27.3 0.0 0.009 0.053 0.81 (0.56, 1.17) 0.79 (0.40, 1.54) 0.023 1.039 0.010 0.162 78.1 45.1
0.84 (0.56, 1.26) 0.88 (0.47, 1.63)
73.6 69.2
0.80 (0.58, 1.11) 0.025 60.9
0.86 (0.63, 1.17)
PQ I (%)
OR (95% CI)
64.4
0.015
49.9 0.0
0.136 0.787
0.80 (0.52, 1.24) 0.86 (0.49, 1.52)
78.6 66.0
0.84 (0.58, 1.20) 0.79 (0.41, 1.53)
0.700 0.0 0.009 0.0
0.481
0.83 (0.61, 1.13)
67.2
0.83 (0.60, 1.13)
I2 (%) OR (95% CI) PQ I (%) I (%) OR (95% CI) I (%)
PQ
2
N
AG vs. GG
2 2
A vs. G
Category
Table 4
Results of meta-analysis for TAFI -438G> A polymorphism and venous thrombosis risk
AA vs. GG
PQ
OR (95% CI)
2
AA þ GA vs. GG
AA vs. GA þ GG
PQ
428 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
To some extent, our meta-analysis has several advantages compared with the preexisting meta-analyses. First, we utilized more strict inclusion criteria and included more original articles. Second, all included studies were of high quality according to the quality assessment and demonstrated no publication bias. Third, Galbraith plot analysis was used for further assessment of heterogeneity. Fourth, our research included better stratified analyses to avoid more potential confounding factors. In addition, the results may have crucial public health implications for improving the prediction of venous thrombosis risk. Several underlying limitations of our study should also be noted. First, the original articles’ lack of relevant data limited the evaluation of potential gene–gene and gene–environment interactions. Second, two studies of 505G>A did not conform to HWE; fortunately, when we limited the studies to those complying with HWE, the results were not altered. Third, the number of original articles included in our research might not be sufficiently large; therefore, the results should be extrapolated very cautiously. In conclusion, the overall results of our meta-analysis suggest that the 505G>A and 1040C>T polymorphisms contribute to decreased risk of venous thrombosis,
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Association between TAFI gene polymorphisms and VT Wang et al. 429
Fig. 6
(a)
b/se(b)
(b)
Fitted values
b/se(b)
2
2
Martini
0 Tokgoz
de Bruij
Martini
b/se(b)
b/se(b)
de Bruij
Lichy
0
Lichy
Franco
−2
Tokgoz
−2 Zidane
−2.77233
0
Franco
Zidane
−2.56696
1/se(b)
(c)
Fitted values
13.1228
0
b/se(b)
1/se(b)
7.36937
Fitted values
2 de Bruij
b/se(b)
Martini
0
Tokgoz
Lichy
Franco
−2 Zidane
−2.74103 0
1/se(b)
7.66112
Galbraith plots of 438G> A polymorphism and venous thrombosis risk in different contrast models. (a) Allele comparison (A vs. G); (b) heterozygote comparison (GA vs. GG); (c) dominant model (GA þ AA vs. GG).
Table 5
Begg and Egger test to measure the funnel plot asymmetric
SNPs 505G>A Begg test Egger test 1040C>T Begg test Egger test 438G>A Begg test Egger test
Allelic comparison
Homozygote comparison
Heterozygote comparison
Dominant model
Recessive model
0.592 0.732
0.929 0.869
0.858 0.569
0.929 0.551
0.929 0.819
0.107 0.228
0.371 0.526
0.283 0.198
0.107 0.162
0.721 0.781
0.133 0.215
0.060 0.090
0.133 0.356
0.060 0.310
0.060 0.077
SNP, single nucleotide polymorphism.
whereas the 505G>A polymorphism plays a dangerous role in venous thrombosis development in Asians; the non-VTE disease group with the 1040C>T polymorphism showed a significantly increased venous thrombosis risk. In addition, there was insufficient evidence to fully confirm that 438G>A had any influence on the susceptibility to venous thrombosis, but a trend toward reduced risk was still observed. Large-scale well designed studies should be conducted to validate the findings of this study.
Acknowledgements The present study was supported by the National Natural Science Foundation of China (grant number 81172604).
Conflicts of interest
There are no conflicts of interest.
References 1 2
3
4
Bafunno V, Margaglione M.. Genetic basis of thrombosis. Clin Chem Lab Med 2010; 48 (Suppl 1):S41–S51. Bouma BN, Marx PF, Mosnier LO, Meijers JC. Thrombin-activatable fibrinolysis inhibitor (TAFI, plasma procarboxypeptidase B, procarboxypeptidase R, procarboxypeptidase U). Thromb Res 2001; 101:329–354. Eaton DL, Malloy BE, Tsai SP, Henzel W, Drayna D. Isolation, molecular cloning, and partial characterization of a novel carboxypeptidase B from human plasma. J Biol Chem 1991; 266:21833–21838. Boffa MB, Koschinsky ML. Curiouser and curiouser: recent advances in measurement of thrombin-activatable fibrinolysis inhibitor (TAFI) and in understanding its molecular genetics, gene regulation, and biological roles. Clin Biochem 2007; 40:431–442.
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
430 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
5
6
7
8 9
10
11
12
13
14
15
16
17
18
Plow EF, Allampallam K, Redlitz A. The plasma carboxypeptidases and the regulation of the plasminogen system. Trends Cardiovasc Med 1997; 7:71–75. Wang W, Boffa MB, Bajzar L, Walker JB, Nesheim ME. A study of the mechanism of inhibition of fibrinolysis by activated thrombin-activable fibrinolysis inhibitor. J Biol Chem 1998; 273:27176–27181. Van Tilburg NH, Rosendaal FR, Bertina RM. Thrombin activatable fibrinolysis inhibitor and the risk for deep vein thrombosis. Blood 2000; 95:2855–2859. Tsai SP, Drayna D. The gene encoding human plasma carboxypeptidase B (CPB2) resides on chromosome 13. Genomics 1992; 14:549–550. Boffa MB, Maret D, Hamill JD, Bastajian N, Crainich P, Jenny NS, et al. Effect of single nucleotide polymorphisms on expression of the gene encoding thrombin-activatable fibrinolysis inhibitor: a functional analysis. Blood 2008; 111:183–189. Boffa MB, Nesheim ME, Koschinsky ML. Thrombin activable fibrinolysis inhibitor (TAFI): molecular genetics of an emerging potential risk factor for thrombotic disorders. Curr Drug Targets Cardiovasc Haematol Disord 2001; 1:59–74. Frere C, Tregouet DA, Morange PE, Saut N, Kouassi D, Juhan-Vague I, et al. Fine mapping of quantitative trait nucleotides underlying thrombinactivatable fibrinolysis inhibitor antigen levels by a transethnic study. Blood 2006; 108:1562–1568. Franco RF, Fagundes MG, Meijers JC, Reitsma PH, Lourenco D, Morelli V, et al. Identification of polymorphisms in the 50 -untranslated region of the TAFI gene: relationship with plasma TAFI levels and risk of venous thrombosis. Haematologica 2001; 86:510–517. Morange PE, Aillaud MF, Nicaud V, Henry M, Juhan-Vague I. Ala147Thr and Cþ1542G polymorphisms in the TAFI gene are not associated with a higher risk of venous thrombosis in FV Leiden carriers. Thromb Haemost 2001; 86:1583–1584. Zidane M, De Visser MC, Ten Wolde M, Vos HL, De Monye W, Bertina RM, et al. Frequency of the TAFI 438 G/A and factor XIIIA Val34Leu polymorphisms in patients with objectively proven pulmonary embolism. Thromb Haemost 2003; 90:439–445. Procare-GEHT Group. ABO blood group but not haemostasis genetic polymorphisms significantly influence thrombotic risk: a study of 180 homozygotes for the Factor V Leiden mutation. Br J Haematol 2006; 135:697–702. Verdu J, Marco P, Benlloch S, Sanchez J, Lucas J. Thrombin activatable fibrinolysis inhibitor (TAFI) polymorphisms and plasma TAFI levels measured with an ELISA insensitive to isoforms in patients with venous thromboembolic disease (VTD). Thromb Haemost 2006; 95:585–586. Lichy C, Dong-Si T, Reuner K, Genius J, Rickmann H, Hampe T, et al. Risk of cerebral venous thrombosis and novel gene polymorphisms of the coagulation and fibrinolytic systems. J Neurol 2006; 253:316–320. Martini CH, Brandts A, De Bruijne EL, Van Hylckama Vlieg A, Leebeek FW, Lisman T, Rosendaal FR. The effect of genetic variants in the thrombin activatable fibrinolysis inhibitor (TAFI) gene on TAFI-antigen levels, clot lysis time and the risk of venous thrombosis. Br J Haematol 2006; 134: 92–94.
19
20
21
22
23
24
25
26
27 28
29 30 31 32
33
34
Xu X. The correlation between thrombin activatable fibrinolysis inhibitor (TAFI) gene polymorphisms and pulmonary thromboembolism in Chinese Han population [dissertation]. Capital Medical University; 2007. http://d.wanfa ngdata.com.cn/Thesis/Y1069064 (in Chinese). [Accessed 6 Nov 2015] De Bruijne EL, Darwish Murad S, De Maat MP, Tanck MW, Haagsma EB, Van Hoek B, et al. Genetic variation in thrombin-activatable fibrinolysis inhibitor (TAFI) is associated with the risk of splanchnic vein thrombosis. Thromb Haemost 2007; 97:181–185. Verdu J, Marco P, Benlloch S, Lucas J. Association between the Thr325Ile and Ala147Thr polymorphisms of the TAFI gene and the risk of venous thromboembolic disease. Clin Appl Thromb Hemost 2008; 14:494–495. Zhang BC. Correlation between TAFI 505A > G and 1040C > T single nucleic acid polymorphisms and VTE [dissertation]. Capital Medical University; 2010. http://d.wanfangdata.com.cn/Thesis/Y1770631 (in Chinese). [Accessed 6 Nov 2015] Hoekstra J, Guimaraes AH, Leebeek FW, Darwish Murad S, Malfliet JJ, Plessier A, et al. Impaired fibrinolysis as a risk factor for Budd-Chiari syndrome. Blood 2010; 115:388–395. Li C, Hou LH, Liu XE, Huang C, Wei XH. Correlation analysis of thrombinactivatable fibrinolysis inhibitor single nucleotide polymorphism with venous thromboembolism [in Chinese]. Zhongguo Shi Yan Xue Ye Xue Za Zhi 2012; 20:949–953. Tokgoz S, Zamani AG, Durakbasi-Dursun HG, Yilmaz O, Ilhan N, Demirel S, et al. TAFI gene polymorphisms in patients with cerebral venous thrombosis. Acta Neurol Belg 2013; 113:291–297. Stang A. Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol 2010; 25:603–605. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 21:1539–1558. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22:719–748. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7:177–188. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994; 50:1088–1101. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315:629–634. Qian K, Xu J, Wan H, Fu F, Lu J, Lin Z, et al. Impact of genetic polymorphisms in thrombin activatable fibrinolysis inhibitor (TAFI) on venous thrombosis disease: a meta-analysis. Gene 2015; 569:173–181. Knoefler R, Ludwig K, Kostka H, Kuhlisch E, Siegert G, Suttorp M. The impact of single nucleotide polymorphisms of the thrombin activatable fibrinolysis inhibitor (TAFI) gene on TAFI antigen levels in healthy children and pediatric oncology patients. Semin Thromb Hemost 2003; 29:575– 583. Guimaraes AH, Bertina RM, Rijken DC. A new functional assay of thrombin activatable fibrinolysis inhibitor. J Thromb Haemost 2005; 3:1284–1292.
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Association between thrombin-activatable fibrinolysis inhibitor gene polymorphisms and venous thrombosis risk: a meta-analysis Wei Wanga,M, He Maa,M, Lili Lub, Guixiang Suna, Dang Liuc, Yunti Zhouc, Yue Tonga and Zhaojun Lua Thrombin-activatable fibrinolysis inhibitor (TAFI) is an important antifibrinolytic factor that has been shown in increased concentrations to be associated with an increased risk for venous thrombosis. However, the effect of TAFI gene polymorphisms on the risk of venous thrombosis remains debatable. The aim of the current study was to evaluate the association of three single nucleotide polymorphisms: 505G>A (rs3742264), 1040 C>T (rs1926447) and S438G>A (rs2146881) with venous thrombosis risk using a meta-analysis. A systematic literature search for eligible studies published before 20 January 2015 was conducted in PubMed, EMBASE, Web of Science, WanFang database and Chinese National Knowledge Infrastructure. We assessed the possible association by pooled odds ratio and its 95% confidence interval. A total of 14 independent case–control studies including 2970 cases and 3049 controls were enrolled in the final meta-analysis. A significant reduction of venous thrombosis risk in the 505G>A polymorphism was observed under allele comparison, homozygote comparison and recessive models, but opposite results were seen in Asians. Likewise, there was a significant decreased susceptibility to venous thrombosis in the 1040C>T polymorphism in homozygote comparison and recessive models. In the subgroup analysis, the nonvenous
thromboembolism disease group showed a significantly increased venous thrombosis risk. Pooled estimates did not show evidence of association between S438G>A and venous thrombosis risk in any genetic model. This meta-analysis suggested that although the S438G>T polymorphism is not correlated with venous thrombosis risk in all models, a trend toward reduced risk still could be observed. The A allele and AA genotype of 505G>A in whites and the TT genotype of 1040C>T were significantly associated with a decreased risk of venous thrombosis, except in the non-venous thromboembolism group. Blood Coagul Fibrinolysis 27:419–430 Copyright ß 2016 Wolters Kluwer Health, Inc. All rights reserved.
Introduction
decades ago. It is a new principal antifibrinolytic factor that plays an important role in the regulation of the fibrinolytic system [4]. A variety of studies have demonstrated that TAFI can be converted to an active carboxypeptidase by enzymes such as thrombin and plasmin. After being activated by thrombin, TAFI suppresses fibrinolysis by removing carboxy-terminal lysine vestiges from partly degraded fibrin; thus, hindering combination of the fibrinolytic parts plasminogen and tissuetype plasminogen activator [5,6]. Increased plasma antigen levels of TAFI have been found to be associated with a risk of venous thrombosis [7]. The TAFI gene, located in chromosome 13 (13q14.11), spans approximately 48 kb and contains 11 exons [8]. In total, 19 single nucleotide polymorphisms (SNPs) have been reported: ten in the 50 -flanking region, six in the coding region and three in the 30 untranslated region [9]. TAFI plasma levels are also determined genetically; genetic factors could explain more than 25% of their variability [4,10–12]. Recently, effects of TAFI polymorphisms affecting antigen levels on venous thrombosis have been reported.
Venous thrombosis is a common and serious disorder that contributes to clotting of the blood in the venous system [1]. Its manifestations not only include two main clinical presentations, deep vein thrombosis (DVT) and pulmonary embolism, but also include some manifestations in which thrombosis is situated in rare venous sites, such as cerebral venous thrombosis or splanchnic vein thrombosis. Venous thrombosis can be seen as a common but complex disease in which environmental and multiple genetic risk factors play a vital role in development. Recently, several emerging candidate gene polymorphisms have been reported, including lipoprotein (a), tissue factor, thrombomodulin, EPCR, and fibrinogen as well as thrombin-activatable fibrinolysis inhibitor (TAFI). TAFI, a 56-kDa zymogen based on 423 amino acids, is primarily synthesized in the liver [2]. Eaton et al. [3] first isolated and characterized the cDNA of TAFI two Wei Wang and He Ma have contributed equally to the manuscript.
0957-5235 Copyright ß 2016 Wolters Kluwer Health, Inc. All rights reserved.
Blood Coagulation and Fibrinolysis 2016, 27:419–430 Keywords: meta-analysis, polymorphism, thrombin-activatable fibrinolysis inhibitor, venous thrombosis a
Department of Public Health, bDepartment of Gastroenterology and Department of General Practice, Xuzhou Medical College, Xuzhou, Jiangsu, China c
Correspondence to Zhaojun Lu, Department of Public Health, Xuzhou Medical College, 209 Tong-shan Road, Xuzhou, Jiangsu, 221004, China Tel: +86 13382654382; e-mail: [email protected] Received 4 August 2015 Revised 29 September 2015 Accepted 3 October 2015
DOI:10.1097/MBC.0000000000000475
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
420 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
Fig. 1
Materials and methods
Flowchart of the selection process for including articles. VT, venous thrombosis.
Search strategy
We carried out a comprehensive literature search in PubMed, EMBASE, Web of Science, WanFang database and Chinese National Knowledge Infrastructure. The last search was performed on January 20, 2015. The keywords included ‘thrombin activatable fibrinolytic inhibitor OR TAFI OR CPB2 OR Carboxypeptidase’ AND ‘polymorphism OR polymorphisms OR variation OR variants’ AND ‘vein Thrombosis OR venous Thrombosis OR vein Thromboses OR Venous Thromboembolic OR pulmonary embolism OR Budd–Chiari syndrome’. There was no restriction on publication years but the languages were limited to English and Chinese in our search. Reference lists of review articles and primary studies were manually searched to identify additional eligible studies. Inclusion and exclusion criteria
All studies included in this meta-analysis were selected based on the following criteria: case–control design for human beings, evaluating the associations between the SNPs 505G>A rs3742264 and/or 1040C>T rs1926447 and/or 438G>A rs2146881 and venous thrombosis disease, supplying useful genotype frequencies and the control groups were healthy persons or patients without a history of venous thrombosis disease. Exclusion criteria included the following: no controls; duplicate publications; reviews, comments or animal studies and studies with no detailed genotype frequencies. If data were published in more than one article, only the study with the largest sample size was included. Data extraction
The following data were independently and carefully extracted by two researchers (W. W., Y. T.) from all
Identification
Several functional variants such as the 505G>A SNP and 1040 C>T SNP in the coding region and the 438G>A SNP in the promoter region are associated with TAFI plasma antigen levels. Because of this, the SNPs have been widely studied and expected to be involved in venous thrombosis disease. A number of articles investigated the association between the three SNPs and venous thrombosis risk [12–25], but the results were mixed and remained inconclusive. Meta-analysis can use a measurable method to combine the results from different articles on an identical theme; thus, supplying more authentic conclusions. Though a meta-analysis published not long ago pooled some of the original studies together, several credible case–control studies [12,15,19,23] have been omitted, and we thought a more comprehensive analysis was necessary. Therefore, we performed this meta-analysis to explore a more accurate estimation of the relationship between TAFI gene variants and venous thrombosis risk.
Records identified and screened through database retrieving (Pubmed = l7, Embase = 64, Web= 65, Wanfang = 5, CNKI= l)
Studies were exclued for improper title/abstract (n = 59) Full-text articles assessed for eligibility (n = 20) Full-text articles were exclued with reason of not for VT (n = 3) not case-control design(n = 3)
Eligibility
Screening
Potentially relevant studies after duplicates removed (n = 79)
Included
Studies included in meta-ananlysis (n = l4)
Article about 505G>A (n = l0)
Article about 11040C>T (n = l0)
Article about -438G>A(n = 6)
eligible studies: last name of first author, year of publication, original country, ethnicity, genotyping method, type of disease, source of control, Hardy–Weinberg equilibrium (HWE) in controls, numbers of cases and controls and allele as well as genotype frequencies for cases and controls. If disagreements existed, the original data were rechecked and consensus was reached through discussion. Study validity assessment
The quality of the studies that were included in this review was independently evaluated by two researchers (W.W., H.M.) according to the Newcastle Ottawa Scale (NOS) [26]. In this scale, three main items were assessed: selection, comparability and exposure. Studies with scores equal to or higher than five were considered to be of high quality. Discrepancies were settled by discussion between the two researchers. Statistical analysis
First, we tested the distribution of the genotypes for HWE by using the goodness-of-fit x2 test in controls; P < 0.05 was considered to be a violation of HWE. The strength of the association between the three SNPs and venous thrombosis risk were estimated by the odds ratio (OR) and its 95% confidence interval (95% CI). The pooled ORs were estimated for five genetic models: allele comparison (505G>A, 438G>A: A vs. G; 1040C>T: T vs. C), heterozygote comparison (505G>A, 438G>A: GA vs. GG; 1040C>T: CT vs. CC), homozygote comparison (505G>A, 438G>A: AA vs. GG; 1040C>T: TT vs. CC), dominant model (505G>A, 438G>A: GA þ AA vs. GG; 1040C>T: CT þ TT vs. CC) and recessive model
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
2001 2001 2003 2006
2006
2006 2006
2007
2007
2008
2010
2010 2012
2013
Franco Morange Zidane Morange
Verdu´
Lichy Martini
Xu
De Bruijne
Verdu´
Zhang
Hoekstra Li
Tokgoz
Turkey
European China
China
Spain
Netherlands
China
Germany Netherlands
Spain
Brazil France Netherlands France
Country
White
White Asian
Asian
White
White
Asian
White White
White
White White White White
Ethnicity
CVT
SVT VTE
VTE
DVT
SVT
Pulmonary embolism
CVT DVT
DVT
DVT VTE Pulmonary embolism VTE
Type of disease
Population-based
Population-based Hospital-based
Hospital-based
Population-based
Population-based
Population-based
Population-based Hospital-based
Population-based
Population-based Hospital-based Population-based Hospital-based
Control source
PCR-RFLP
PCR Allele-specific PCR
ARMS-PCR
Allele-specific PCR
PCR
PCR-RFLP
Allele-specific PCR PCR
PCR
PCR PCR PCR PCR
Genotyping method 438G/A 505G/A 438G/A 505G/A 1040C/T 505G/A 1040C/T 438G/A 505G/A 1040C/T 438G/A 505G/A 1040C/T 505G/A 1040C/T 438G/A 505G/A 1040C/T 505G/A 1040C/T 1040C/T 505G/A 1040C/T 505G/A 1040C/T 438G/A
SNP 388 168 66 127 127 60 61 75 471 471 471 114 114 117 118 114 131 131 339 477 97 80 80 59 59 59
Case 388 145 152 53 53 61 62 197 472 472 472 114 114 116 118 116 100 100 317 509 99 80 80 100 100 100
Control
Sample size (men)
222/201 100/78 42/67 70/22 62/25 10/7 25/9 46/103 236/211 219/215 240/257 56/71 85/79 58/48 40/53 48/60 41/22 67/33 169/180 324/370 43/46 24/28 27/15 28/40 21/45 37/50
Wild type
139/157 64/63 23/72 45/25 55/22 43/38 29/31 26/82 198/210 218/212 202/185 48/36 24/29 54/50 69/52 58/46 66/58 50/51 160/116 142/126 43/41 47/36 42/44 23/48 30/43 18/43
Heterozygous
27/30 4/4 1/13 12/6 10/6 7/16 7/22 3/12 37/51 34/45 29/30 10/7 5/6 5/18 9/13 8/10 24/20 14/16 10/21 11/13 11/12 9/16 11/21 8/12 8/12 4/7
Mutant
Genotype distribution (case/control)
Y(0.932) N(0.035) Y(0.299) Y(0.782) Y(0.730) N(0.033) Y(0.717) Y(0.670) Y(0.907) Y(0.484) Y(0.668) Y(0.406) Y(0.142) Y(0.412) Y(0.963) Y(0.780) Y(0.109) Y(0.615) Y(0.694) Y(0.564) Y(0.544) Y(0.478) Y(0.342) Y(0.677) Y(0.727) Y(0.582)
Y/N(P)
HWE
7
8 6
6
6
6
6
6 7
6
7 7 7 6
NOS
CVT, cerebral venous thrombosis; DVT, deep venous thrombosis; HWE, Hardy–Weinberg equilibrium; NOS, Newcastle–Ottawa scale; PCR, polymerase chain reaction; RFLP-PCR, restriction fragment length polymorphism; SNP, single nucleotide polymorphism; SVT, splanchnic vein thrombosis; TAFI, thrombin activatable fibrinolysis inhibitor; VTE, venous thromboembolism.
Year
Baseline characteristics of studies included in the meta-analysis
Author
Table 1
Association between TAFI gene polymorphisms and VT Wang et al. 421
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
422 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
Fig. 2
Study ID
OR (95% CI)
% weight
Morane (2001)
0.78 (0.19, 3.22)
3.11
Morange (2006)
0.63 (0.21, 187)
55.53
Verdu (2006)
0.31 (0.08, 1.14)
5.59
Martini (2006)
0.65 (0.41, 1.103) 32.56
Xu (2007)
1.81 (0.65, 5.06)
de Buijine (2007)
0.23 (0.08, 0.66) 11.71
Verdu (2008)
0.64 (0.29, 1.42)
11.09
Zhang (2010)
0.51(0.23, 1.11)
13.52
Li (2012)
0.66 (0.25, 1.75)
7.22
Tokgoz (2013)
0.95 (0.34, 2.63)
5.53
Overall (I-squared = 7.5%, P = 0.373)
0.63 (0.48, 0.82) 100.00
0.0795
1
3.94
12.6
Forest plots of meta-analysis for association between 505G>A polymorphism and venous thrombosis (AA vs. GG). A significant reduction susceptibility was observed between the 505G>A polymorphism and risk of venous thrombosis. CI, confidence interval; OR, odds ratio.
(505G>A, 438G>A: AA vs. GA þ GG; 1040C>T: TT vs. CT þ CC). The significance of the pooled ORs was examined by the Z-test. The x2-based Q test and I2 statistics were used to test the heterogeneity between studies; P < 0.10 or I2 > 50% indicated evidence of heterogeneity [27]. The fixed-effects model (Mantel–Haenszel method) was used to estimate the summary ORs when there was no significant heterogeneity [28]; otherwise, the random effect model (DerSimonian and Laird method) was used [29]. Subgroup analyses were conducted according to ethnicity, type of disease, source of control and genotyping method. Meta-regression and Galbraith plot analyses were conducted for further exploration of heterogeneity. The study characteristics included as covariates in the meta-regression were publication year and NOS score. Furthermore, sensitivity analysis was used to assess the stability of the results by deleting one single study at a time to examine the influence of single data points. Potential publication bias was evaluated using the Begg test [30] and Egger regression test [31]; an asymmetric plot and P < 0.05 were considered as statistically significant publication bias. All statistical tests were twosided, and P values A and venous thrombosis in a total of 1666 cases and 1558 controls, 10 studies [15,16,18–25] involving 1735 cases and 1707 controls for 1040C>T and 6 studies [12,14,17,18,20,25] involving 1173 cases and 1425 controls for 438G>A. There were 11 studies carried out in whites, whereas the other studies were conducted in Asians. Detailed genotype distribution and the results of the HWE test in control population are summarized in Table 1; all studies were consistent with HWE except for two studies [13,16] of 505G>A. NOS quality scores for each study ranged from 6 to 8, with all studies being classified as high quality.
Results Study selection and characteristics
Association between the 505G>A (rs3742264) polymorphism and venous thrombosis susceptibility
As shown in Fig. 1, a total of 152 relevant studies were initially identified after a systematic search; 53 studies
Ten studies with a total of 1666 cases and 1558 controls examined the association between the 505G>>A
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
0.286 0.729 0.318 20.1 0.0 12.6 0.58 (0.41, 0.81) 1.30 (0.65, 2.58) 0.61 (0.40, 0.94) 0.903 0.049 0.084 0.0 74.2 59.5 0.76 (0.63, 0.92) 1.22 (0.82, 1.84) 1.14 (0.89, 1.47) 0.387 0.383 0.889 7.5 0.0 0.0 0.54 (0.37, 0.77) 1.31 (0.64, 2.68) 0.59 (0.36,0.95) 0.865 0.046 0.043
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
3
CI, confidence interval; OR, odds ratio; PQ, P-value for heterogeneity test.
0.0 74.8 68.3 0.81 (0.66, 0.99) 1.18 (0.77, 1.82) 1.26 (0.99, 1.63) 0.743 0.095 0.388 0.77 (0.67, 0.89) 1.19 (0.87, 1.63) 0.98 (0.81, 1.18) 5 2
0.0 64.1 0.0
0.503 0.069 0.0 57.6 0.68 (0.50, 0.93) 0.61 (0.40, 0.92) 0.015 0.977 64.6 0.0 0.99 (0.84, 1.16) 0.92 (0.80, 1.06) 0.532 0.209 0.0 33.9 0.69 (0.50, 0.96) 0.49 (0.30, 0.80) 0.009 0.828 67.4 0.0 1.05 (0.89, 1.24) 0.74 (0.53, 1.04) 0.081 0.719 0.93 (0.82, 1.06) 0.74 (0.60, 0.92) 6 4
49.0 0.0
0.282 0.140 19.4 49.1 0.67 (0.50, 0.89) 0.61 (0.38, 1.00) 0.962 0.675 0.0 0.0 0.75 (0.63, 0.89) 1.39 (1.08, 1.78) 0.516 0.145 0.0 48.2 0.58 (0.43, 0.80) 0.76 (0.45, 1.27) 0.907 0.912 0.0 0.0 0.78 (0.65, 0.94) 1.52 (1.18, 1.97) 0.898 0.261 0.78 (0.69, 0.89) 1.13 (0.93, 1.37) 7 3
0.0 25.6
0.484 0.030 0.0 78.9 0.68 (0.52, 0.89) 0.52 (0.27, 1.02) 0.018 0.948 58.4 0.0 0.94 (0.72, 1.22) 0.73 (0.48, 1.09) 0.596 0.058 0.0 72.3 0.66 (0.50, 0.88) 0.46 (0.23, 0.94) 0.011 0.552 61.6 0.0 0.98 (0.74, 1.31) 0.81 (0.53, 1.24) 0.109 0.316 0.90 (0.81, 1.01) 0.73 (0.54, 0.98) 8 2
40.4 0.6
0.244 21.7 0.65 (0.51, 0.84) 0.031 51.0 0.90 (0.72, 1.12) 0.373 7.5 0.63 (0.48, 0.82) 0.022 53.7 0.95 (0.75, 1.21) 0.106 0.88 (0.79, 0.98) 10
Overall Type of disease VTE Non-VTE Ethnicity White Asian Source of control Hospital-based Population-based Genotyping method PCR RFLP-PCR Allele-specific PCR
37.9
OR (95% CI) PQ I2 (%) OR (95% CI) OR (95% CI) N
I2 (%)
PQ
OR (95% CI)
I2 (%)
PQ
OR (95% CI)
I2 (%)
PQ
AAþGA vs. GG AA vs. GG AG vs. GG A vs. G
Category
Similarly, 10 studies involving 1735 cases and 1707 controls were assessed for the association between the 1040C>T polymorphism and venous thrombosis risk. We found a significantly decreased venous thrombosis susceptibility in homozygote comparison (TT vs. CC: OR ¼ 0.63, 95% CI 0.42–0.94, P ¼ 0.022 Fig. 4) and recessive models (TT vs. CT þ CC: OR ¼ 0.67, 95% CI 0.52–0.86, P ¼ 0.002 Fig. 4); no significant association was found in allele comparison (T vs. C: OR ¼ 0.85, 95%
Results of meta-analysis for TAFI 505G> A polymorphism and venous thrombosis risk
Association between the 1040C>T (rs1926447) polymorphism and venous thrombosis susceptibility
Table 2
Subsequently, we performed subgroup analyses to avoid the potential impact of confounding factors. Heterogeneity vanished in the ethnicity subgroup, suggesting that ethnicity might contribute mainly to the heterogeneity. In white populations, we found a decreased venous thrombosis risk in all genetic models (Table 2). In Asian populations, there was a significantly increased venous thrombosis risk observed in the heterozygote comparison (GA vs. GG: OR ¼ 1.52, 95% CI 1.18–1.97, P ¼ 0.001 Fig. 3) and dominant models (GA þ AA vs. GG:OR ¼ 1.39, 95% CI 1.08–1.78, P ¼ 0.009 Fig. 3). When stratifying by type of disease, significantly reduced risk was found among both venous thromboembolism (VTE) (homozygote comparison, recessive model) and the non-VTE group (homozygote comparison, recessive model). When it came to sources of control, the TAFI 505G>A polymorphism was significantly associated with a decreased venous thrombosis risk in both the hospital-based group (homozygote comparison, recessive model) and the population-based group (allele comparison, homozygote comparison, recessive model). For the polymerase chain reaction (PCR) group, significantly reduced risk was found in all genetic models. For the allelespecific PCR group, similar results were observed in homozygote comparison and recessive models. However, no significant association between the 505G>A polymorphism and venous thrombosis risk was seen in the restriction fragment length polymorphism-PCR group. The main results of the subgroup analyses are shown in Table 2.
AA vs. GA þ GG
polymorphism and venous thrombosis risk. There was a significant heterogeneity within studies of heterozygote comparison and dominant models, thus the randomeffects model was used. A significant reduction of venous thrombosis susceptibility was observed in allele comparison (A vs. G: OR ¼ 0.88, 95% CI 0.79–0.98, P ¼ 0.018), homozygote comparison (AA vs. GG: OR ¼ 0.63, 95% CI 0.48–0.82, P ¼ 0.001 Fig. 2) and recessive models (AA vs. GA þ GG: OR ¼ 0.65, 95% CI 0.51–0.84, P ¼ 0.001); no significant association was found in heterozygote comparison (GA vs. GG: OR ¼ 0.95, 95% CI 0.75–1.21, P ¼ 0.681) and dominant models (GA þ AA vs. GG: OR ¼ 0.90, 95% CI 0.72–1.12, P ¼ 0.347).
I2 (%)
PQ
Association between TAFI gene polymorphisms and VT Wang et al. 423
424 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
Fig. 3
(a) Studay ID
OR(95% CI)
% weight
White Morange (2001)
0.79 (0.50, 1.25) 11.96
Morange (2006)
0.57 (0.29, 1.12) 7.82
Verdu (2006)
0.79 (0.27, 2.29) 4.18
Martini (2006)
0.84 (0.64, 1.10) 16.56
de bruijine (2007)
0.89 (0.52, 1.54) 10.19
Verdu (2008)
0.61 (0.33, 1.14) 8.70
Tokgoz (2013)
0.68 (0.34, 1.37) 7.70
Subtotal (I-squared = 0.0%, P = 0.907) . Asian
0.78 (0.65, 0.94) 67.12
Xu (2007)
1.69 (0.97, 2.95) 9.93
Zhang (2010)
1.47 (1.07, 2.02) 15.31
Li (2012)
1.52 (0.76, 3.06) 7.65
Subtotal (I-squared = 0.0%, P = 0.912) .
1.52 (1.18, 1.97) 32.88
Overall (I-squared = 53.7%, P = 0.22)
0.95 (0.75, 1.21) 100.00
NOTE: weights are from random effects analysis
0.274
1
3.64
(b) Studay ID
OR(95% CI)
% weight
White Morange (2001)
0.79 (0.51, 1.24) 11.72
Morange (2006)
0.58 (0.30, 1.11) 7.78
Verdu (2006)
0.65 (0.23, 1.83) 3.87
Martini (2006)
0.81 (0.62, 1.104) 17.04
De bruijine (2007)
0.72 (0.43, 1.20) 10.16
Verdu (2008)
0.62 (0.34, 1.13) 8.58
Tokgoz (2013) Subtotal (I-squared = 0.0%, P = 0.907)
0.75 (0.63, 0.89) 66.95
0.74 (0.39, 1.41) 7.78
. Asian Xu (2007)
1.71 (1.01, 2.90) 9.95
Zhang (2010)
1.32 (0.97, 1.80) 15.54 1.26 (0.65, 2.44) 7.56
Li (2012) Subtotal (I-squared = 0.0%, P = 0.912) .
1.39 (1.108, 1.78) 33.05
Overall (I-squared = 51.0%, P = 0.31)
0.90 (0.75, 1.12) 100.00
NOTE: weights are from random effects analysis
0.229
1
4.36
Forest plot for the subgroup analysis of ethnicity (a) GA vs. GG; (b) GA þ AA vs. GG. The 505G>A polymorphism was found to be associated with increased risk of Asians CI, confidence interval; OR, odds ratio.
CI 0.68–1.05, P ¼ 0.133), heterozygote comparison (CT vs. CC: OR ¼ 0.94, 95% CI 0.72–1.23, P ¼ 0.636) and dominant models (CT þ TT vs. CC: OR ¼ 0.87,
95% CI 0.65–1.15, P ¼ 0.325). Q-test showed significant heterogeneity in all genetic models except the recessive model.
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Association between TAFI gene polymorphisms and VT Wang et al. 425
Fig. 4
(a) Study ID
OR (95% CI)
Morange (2006)
0.67 (0.22, 2.05)
8.01
Verdu (2006)
0.11 (0.04, 0.36)
7.75
Martini (2006)
0.74 (0.46, 1.20)
16.56
Xu (2007)
0.77 (0.23, 2.64)
7.06
de bruijine (2007)
0.92 (0.36, .2.36)
9.75
Verdu (2008)
0.43 (0.19, 0.99)
11.17
Zhang (2010)
0.97 (0.43, 2.19)
11.34
Hoekstra (2010)
0.98 (0.39, 2.46)
10.06
Li (2012)
0.29 (0.11, 0.76)
9.52
Tokgoz (2013)
1.43 (0.51, 4.02)
8.78
Overall (I-squared = 47.9%, P = 0.045)
0.63 (0.42, 0.94)
100.00
% weight
NOTE: weights are from random effects analysis
0.0366
1
27.3
(b) Study ID
OR (95% CI)
% weight
Morange (2006)
0.67 (0.23, 1.95)
5.15
Verdu (2006)
0.24 (0.09, 0.61)
12.76
Martini (2006)
0.74 (0.46, 1.18)
27.55
Xu (2007)
0.83 (0.24, 2.79)
3.79
de bruijine (2007)
0.67 (0.27, 1.63)
7.93
Verdu (2008)
0.63 (0.29, 1.36)
10.70
Zhang (2010)
0.90 (0.40, 2.03)
8.12
Hoekstra (2010)
0.93 (0.39, 2.22)
6.96
Li (2012)
0.45 (0.20, 1.00)
11.96
Tokgoz (2013)
1.15 (0.44, 3.00)
5.08
Overall (I-squared = 0.0%, P = 0.510)
0.67 (0.52, 0.86)
100.00
0.0917
1
10.9
Forest plots of meta-analysis for association between 1040C>T polymorphism and venous thrombosis (a) TT vs. CC; (b) TT vs. CT þ CC. A significant reduction susceptibility was observed between the 1040C>T variant and risk of venous thrombosis.
As shown in Table 3, in whites, no significant association was found for any genetic models except the recessive model. Similarly, no significant association was observed in any genetic model in Asians. Moreover, in the stratified analysis based on type of disease, a significantly decreased risk of venous thrombosis was observed under the allele comparison, homozygote comparison and recessive model in the VTE group. It is worth noting that increased venous thrombosis risk was found in the heterozygote comparison (CT vs. CC:
OR ¼ 1.45, 95% CI 1.02–2.05, P ¼ 0.037) in the nonVTE group. Regarding the source of control, we found a significantly decreased venous thrombosis risk in the hospital-based group under the homozygote comparison and recessive model, whereas only under the recessive model in the population-based group. In the subgroup analyses on genotyping method, no significant association was observed in any genetic model of the three subgroups except for the recessive model in the PCR group.
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
0.250 0.675 0.490 0.64 (0.46, 0.88) 1.01 (0.48, 2.15) 0.63 (0.40, 1.00)
25.7 0.0 0.0
Heterogeneity in the four genetic models mentioned above could still be detected in the subgroup analyses. A meta-regression was used next, but it failed to confirm that the publication year and NOS score were the sources of heterogeneity. Finally, we performed Galbraith plot analysis for further detection of heterogeneity. The results found that, four studies [16,21,22,24] for the allele comparison, one study [16] for the homozygote comparison, three studies [16,20,21] for heterozygote comparison and three studies [16,21,24] for the dominant model contributed mainly to heterogeneity (Fig. 5). After we excluded these studies, heterogeneity can ignored among these genetic models (I2 ¼ 0.0%, P ¼ 0.659; I2 ¼ 0.0%, P ¼ 0.456; I2 ¼ 15.8%, P ¼ 0.309; and I2 ¼ 4.8%, P ¼ 0.390).
69.6 52.8 86.2
Association between the S438G>A (rs2146881) polymorphism and venous thrombosis susceptibility
The association between the 438G>A polymorphism and susceptibility to venous thrombosis was analyzed in six independent studies with 1173 cases and 1425 controls. Significant between-study heterogeneity was observed under allele comparison, heterozygote comparison and dominant models. None of the genetic models produced a significant association between the 438G/A polymorphism and venous thrombosis risk (A vs. G: OR ¼ 0.91, 95% CI 0.80–1.03, P ¼ 0.147; AA vs. GG: OR ¼ 0.80, 95% CI 0.58–1.11, P ¼ 0.185; GA vs. GG: OR ¼ 0.92, 95% CI 0.79–1.09, P ¼ 0.345; GA þ AA vs. GG: OR ¼ 0.91, 95% CI 0.78–1.06, P ¼ 0.219; AA vs. GA þ GG: OR ¼ 0.83, 95% CI 0.60–1.14, P ¼ 0.238). In the subgroup analyses, the same results were obtained (Table 4). Nevertheless, the results still show a trend toward decreased susceptibility to venous thrombosis.
CI, confidence interval; OR, odds ratio; PQ, P-value for heterogeneity test.
57.9 48.6 83.3 73.1 42.6 84.9 5 2 3
0.84 (0.61, 1.16) 1.01 (0.64, 1.59) 0.76 (0.45, 1.28)
0.005 0.187 0.001
1.02 (0.70, 1.49) 1.05 (0.55, 2.01) 0.72 (0.35, 1.51)
0.050 0.163 0.002
0.59 (0.32, 1.12) 1.11 (0.50, 2.44) 0.51 (0.26, 1.02)
61.7 0.0 47.3
0.033 0.454 0.150
0.91 (0.59, 1.39) 1.05 (0.55, 1.98) 0.67 (0.31, 1.46)
0.010 0.146 0.001
0.796 0.167 0.0 38.1 0.70 (0.50, 0.97) 0.63 (0.43, 0.93) 50.8 81.2 35.1 76.7 54.2 80.5 5 5
0.90 (0.72, 1.12) 0.81 (0.52, 1.25)
0.068 0.000
1.00 (0.78, 1.28) 0.90 (0.49, 1.63)
0.188 0.002
0.69 (0.48, 0.97) 0.59 (0.27, 1.30)
0.0 69.9
0.427 0.010
0.93 (0.71, 1.22) 0.81 (0.43, 1.53)
0.087 0.000
0.354 0.455 9.9 0.0 0.67 (0.50, 0.89) 0.66 (0.40, 1.11) 72.0 74.6 65.1 66.6 71.6 76.7
10
Overall Type of disease VTE Non-VTE Ethnicity White Asian Source of control Hospital-based Population-based Genotyping method PCR RFLP-PCR Allele-specific PCR
7 3
0.84 (0.64, 1.11) 0.84 (0.53, 1.34)
0.002 0.014
0.95 (0.66, 1.36) 0.88 (0.51, 1.51)
0.009 0.050
0.64 (0.39, 1.05) 0.61 (0.28, 1.30)
55.9 44.8
0.034 0.163
0.87 (0.60, 1.28) 0.82 (0.45, 1.47)
0.002 0.019
0.396 0.710 4.0 0.0 0.62 (0.46, 0.82) 0.88 (0.52, 1.48) 74.8 0.0 66.1 0.0 76.2 0.0
N
7 3
0.74 (0.56, 0.98) 1.15 (0.90, 1.47)
0.000 0.800
0.79 (0.57, 1.09) 1.45 (1.02, 2.05)
0.007 0.550
0.50 (0.31, 0.82) 1.07 (0.61, 1.86)
53.7 0.0
0.044 0.804
0.71 (0.50, 1.02) 1.37 (0.98, 1.91)
0.001 0.608
0.510 0.0 0.67 (0.52, 0.86) 69.9 61.8 70.3 0.85 (0.68, 1.05)
PQ I (%) OR (95% CI)
0.000
0.94 (0.72, 1.23)
0.005
0.63 (0.42, 0.94)
47.9
0.045
0.87 (0.65, 1.15)
0.000
I2 (%) OR (95% CI) I (%) OR (95% CI) I (%) I (%) OR (95% CI)
PQ
OR (95% CI)
2 2
CT vs. CC T vs. C
2
Category
Table 3
Results of meta-analysis for TAFI 1040C>T polymorphism and venous thrombosis risk
TT vs. CC
PQ
2
TT þ CT vs. CC
PQ
TT vs. CT þ CC
PQ
426 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
Heterogeneity still existed in the subgroup analyses; the publication year and NOS score were not viewed as the reasons for heterogeneity when we carried out a metaregression. However, Galbraith plot analysis successfully distinguished the deviants that might have led to the heterogeneity (Fig. 6). For allele comparison and dominant models, heterogeneity could be ascribed mainly to one study [14]; for heterozygote comparison, another study [20] might have been responsible for the heterogeneity. Homogeneity test I2 values deceased more or less after removing each study from the total dataset (I2 ¼ 37.0%, P ¼ 0.174; I2 ¼ 57.9%, P ¼ 0.050 and I2 ¼ 59.9%, P ¼ 0.041). Sensitivity analysis
Analyzing the three SNPs by deleting each study from the total dataset, we did not notice any significant difference in the pooled estimates. For 505G>A, when excluding the studies that deviated from HWE, the overall results were not materially affected. The results indicated that this meta-analysis provided reliable evidence (data not shown).
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Association between TAFI gene polymorphisms and VT Wang et al. 427
Fig. 5
Fitted values
b/se(b)
Tokgoz
Morange Xu
Tokgoz
Mattini
−2 Li
Fitted values
b/se(b)
2
Bruijine
Hoekstra
0
b/se(b)
(b)
2
b/se(b)
(a)
0
HoekstraZhang Bruijine Xu Morange Matini
−2
Verdu08
Verdu08
Li
−2.98221 Verdu06
0
1/se(b)
2.16849 2
0
10.0556 Fitted values
b/se(b)
(c)
Verdu06
−3.71027
(d)
Bruijine
Bruijine
Zhang
Mattini Hoekstra
Xu Li
−2
Verdu06
0
b/se(b)
0
Hoekstra
0
Morange
Matini
Xu
−2 Li
−2.24125
Verdu08 Verdu06
Verdu08
1 /se(b)
Zhang
Tokgoz
Hoekstra Morange
4.05382 Fitted values
b/sec(b) 2
Tokgoz
b/se(b)
1/se(b)
7.34788
0
1/se(b)
7.65241
Galbraith plots of 1040C>T polymorphism and venous thrombosis risk in different contrast models. (a) Allele comparison (T vs. C); (b) homozygote comparison (TT vs. CC); (c) heterozygote comparison (CT vs. CC); (d) dominant model (TT þ CT vs. CC).
Publication bias
Publication bias was assessed by Begg’s funnel plot and Egger’s test; the results are shown in Table 5. The symmetric funnel plot suggested there was no obvious publication bias, whereas Egger’s test also showed no statistical significance for the detection of publication bias between the three genetic variations and venous thrombosis risk in any of the models.
Discussion In addition to plasminogen activator inhibitor-1, the major physiological inhibitor of plasminogen activation, TAFI is another important influencing factor of impaired fibrinolysis, which can result in thrombotic disorders. Plasma levels of TAFI resulting from the genetic variations have been reported in men with thrombotic diseases [4]. However, there is still no definite evidence concerning the association between TAFI polymorphisms and venous thrombosis risk. Our previous metaanalysis performed in the same way investigated the association between the three TAFI polymorphisms and venous thrombosis risk [32]. However, some defects existed in the literature included; an article that did not report a venous thrombosis case–control study and mainly researched pediatric oncology patients [33] was falsely included. Considering this, we performed this
meta-analysis to produce more credible evidence for or against an association between TAFI genetic variants and venous thrombosis. In this meta-analysis, 14 studies including 2339 cases and 4048 controls were analyzed. As for association between the 505G>A polymorphism and venous thrombosis susceptibility, a protective relationship between them was identified in allele comparison, homozygote comparison and recessive models, suggesting that the A allele and AA genotype were significantly associated with a decreased risk of venous thrombosis, which might result from the role of the 505A allele in the alteration of TAFI levels. An earlier study reported that the A allele was associated with lower TAFI levels than the G allele [20]. Stratified analyses also suggested protective effects on venous thrombosis in overall studies; however, Asians showed an opposite result in the subgroup analysis; this might have resulted from ethnic differences. As for association between 1040C/T and venous thrombosis risk, the meta-analysis results showed that TT genotype carriers had a 37% lower risk of venous thrombosis than those with CC genotype and a 33% lower risk than the combined group of CT and CC carriers. In the subgroup analysis, similar results could be found except that the non-VTE disease group showed a significantly increased venous thrombosis risk. It has been shown that the
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
0.428 0.0 0.007 75.5 0.83 (0.59, 1.17) 0.013
26.1
0.255
0.91 (0.62, 1.34)
Heterogeneity is hard to avoid in a meta-analysis and the key is determining the sources of heterogeneity. Significant heterogeneity was found for the association of 505G>A and venous thrombosis in two genetic models. By conducting stratified analyses, we found no heterogeneity existing in the ethnicity subgroup, suggesting that ethnicity contributed to the source of heterogeneity. However, for the other two SNPs, we failed to find the source of heterogeneity by stratified or meta-regression analyses. We finally discovered the source of heterogeneity by Galbraith plot analysis. For the four genetic models of 1040C>T, all I2 values were less than 50%, and P values were greater than 0.10 after eliminating the studies judged as outliers by the Galbraith plots. Although I2 values were not reduced as was expected in 438G>A, the heterogeneity was acceptable in the end.
0.82 (0.58, 1.16) 0.88 (0.62, 1.26)
0.79 (0.61, 1.03)
0.90 (0.68, 1.20)
3 3
5
4
CI, confidence interval; OR, odds ratio; PQ, P-value for heterogeneity test.
72.3 0.95 (0.66, 1.38) 0.013 72.1
0.86 (0.68, 1.07) 6
Overall Type of disease VTE Non-VTE Source of control Population-based Genotyping method PCR
OR (95% CI)
1040C>T SNP was strongly associated with TAFI levels [11,34], yet increased plasma TAFI levels have been found to be associated with a risk of venous thrombosis [7]. Therefore, it can be concluded from this study that the polymorphisms 505G>A and 1040C>T, which are seen more in normal population groups, may reduce thrombus formation by causing reductions in TAFI activity. As for association between 438G>A and venous thrombosis risk, although the 438A allele has been associated with decreased TAFI antigen levels [14]. The results showed that neither the overall analysis nor the subgroup analysis supported a genetic association between 438G>A and susceptibility to venous thrombosis.
0.83 (0.59, 1.17)
0.609 0.0 0.042 0.70 (0.46, 1.05) 0.079 52.1
0.78 (0.55, 1.11)
26.1
0.235
0.0
0.457
0.76 (0.54, 1.08)
59.6
0.75 (0.51, 1.13)
0.253 0.907 27.3 0.0 0.009 0.053 0.81 (0.56, 1.17) 0.79 (0.40, 1.54) 0.023 1.039 0.010 0.162 78.1 45.1
0.84 (0.56, 1.26) 0.88 (0.47, 1.63)
73.6 69.2
0.80 (0.58, 1.11) 0.025 60.9
0.86 (0.63, 1.17)
PQ I (%)
OR (95% CI)
64.4
0.015
49.9 0.0
0.136 0.787
0.80 (0.52, 1.24) 0.86 (0.49, 1.52)
78.6 66.0
0.84 (0.58, 1.20) 0.79 (0.41, 1.53)
0.700 0.0 0.009 0.0
0.481
0.83 (0.61, 1.13)
67.2
0.83 (0.60, 1.13)
I2 (%) OR (95% CI) PQ I (%) I (%) OR (95% CI) I (%)
PQ
2
N
AG vs. GG
2 2
A vs. G
Category
Table 4
Results of meta-analysis for TAFI -438G> A polymorphism and venous thrombosis risk
AA vs. GG
PQ
OR (95% CI)
2
AA þ GA vs. GG
AA vs. GA þ GG
PQ
428 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
To some extent, our meta-analysis has several advantages compared with the preexisting meta-analyses. First, we utilized more strict inclusion criteria and included more original articles. Second, all included studies were of high quality according to the quality assessment and demonstrated no publication bias. Third, Galbraith plot analysis was used for further assessment of heterogeneity. Fourth, our research included better stratified analyses to avoid more potential confounding factors. In addition, the results may have crucial public health implications for improving the prediction of venous thrombosis risk. Several underlying limitations of our study should also be noted. First, the original articles’ lack of relevant data limited the evaluation of potential gene–gene and gene–environment interactions. Second, two studies of 505G>A did not conform to HWE; fortunately, when we limited the studies to those complying with HWE, the results were not altered. Third, the number of original articles included in our research might not be sufficiently large; therefore, the results should be extrapolated very cautiously. In conclusion, the overall results of our meta-analysis suggest that the 505G>A and 1040C>T polymorphisms contribute to decreased risk of venous thrombosis,
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Association between TAFI gene polymorphisms and VT Wang et al. 429
Fig. 6
(a)
b/se(b)
(b)
Fitted values
b/se(b)
2
2
Martini
0 Tokgoz
de Bruij
Martini
b/se(b)
b/se(b)
de Bruij
Lichy
0
Lichy
Franco
−2
Tokgoz
−2 Zidane
−2.77233
0
Franco
Zidane
−2.56696
1/se(b)
(c)
Fitted values
13.1228
0
b/se(b)
1/se(b)
7.36937
Fitted values
2 de Bruij
b/se(b)
Martini
0
Tokgoz
Lichy
Franco
−2 Zidane
−2.74103 0
1/se(b)
7.66112
Galbraith plots of 438G> A polymorphism and venous thrombosis risk in different contrast models. (a) Allele comparison (A vs. G); (b) heterozygote comparison (GA vs. GG); (c) dominant model (GA þ AA vs. GG).
Table 5
Begg and Egger test to measure the funnel plot asymmetric
SNPs 505G>A Begg test Egger test 1040C>T Begg test Egger test 438G>A Begg test Egger test
Allelic comparison
Homozygote comparison
Heterozygote comparison
Dominant model
Recessive model
0.592 0.732
0.929 0.869
0.858 0.569
0.929 0.551
0.929 0.819
0.107 0.228
0.371 0.526
0.283 0.198
0.107 0.162
0.721 0.781
0.133 0.215
0.060 0.090
0.133 0.356
0.060 0.310
0.060 0.077
SNP, single nucleotide polymorphism.
whereas the 505G>A polymorphism plays a dangerous role in venous thrombosis development in Asians; the non-VTE disease group with the 1040C>T polymorphism showed a significantly increased venous thrombosis risk. In addition, there was insufficient evidence to fully confirm that 438G>A had any influence on the susceptibility to venous thrombosis, but a trend toward reduced risk was still observed. Large-scale well designed studies should be conducted to validate the findings of this study.
Acknowledgements The present study was supported by the National Natural Science Foundation of China (grant number 81172604).
Conflicts of interest
There are no conflicts of interest.
References 1 2
3
4
Bafunno V, Margaglione M.. Genetic basis of thrombosis. Clin Chem Lab Med 2010; 48 (Suppl 1):S41–S51. Bouma BN, Marx PF, Mosnier LO, Meijers JC. Thrombin-activatable fibrinolysis inhibitor (TAFI, plasma procarboxypeptidase B, procarboxypeptidase R, procarboxypeptidase U). Thromb Res 2001; 101:329–354. Eaton DL, Malloy BE, Tsai SP, Henzel W, Drayna D. Isolation, molecular cloning, and partial characterization of a novel carboxypeptidase B from human plasma. J Biol Chem 1991; 266:21833–21838. Boffa MB, Koschinsky ML. Curiouser and curiouser: recent advances in measurement of thrombin-activatable fibrinolysis inhibitor (TAFI) and in understanding its molecular genetics, gene regulation, and biological roles. Clin Biochem 2007; 40:431–442.
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
430 Blood Coagulation and Fibrinolysis 2016, Vol 27 No 4
5
6
7
8 9
10
11
12
13
14
15
16
17
18
Plow EF, Allampallam K, Redlitz A. The plasma carboxypeptidases and the regulation of the plasminogen system. Trends Cardiovasc Med 1997; 7:71–75. Wang W, Boffa MB, Bajzar L, Walker JB, Nesheim ME. A study of the mechanism of inhibition of fibrinolysis by activated thrombin-activable fibrinolysis inhibitor. J Biol Chem 1998; 273:27176–27181. Van Tilburg NH, Rosendaal FR, Bertina RM. Thrombin activatable fibrinolysis inhibitor and the risk for deep vein thrombosis. Blood 2000; 95:2855–2859. Tsai SP, Drayna D. The gene encoding human plasma carboxypeptidase B (CPB2) resides on chromosome 13. Genomics 1992; 14:549–550. Boffa MB, Maret D, Hamill JD, Bastajian N, Crainich P, Jenny NS, et al. Effect of single nucleotide polymorphisms on expression of the gene encoding thrombin-activatable fibrinolysis inhibitor: a functional analysis. Blood 2008; 111:183–189. Boffa MB, Nesheim ME, Koschinsky ML. Thrombin activable fibrinolysis inhibitor (TAFI): molecular genetics of an emerging potential risk factor for thrombotic disorders. Curr Drug Targets Cardiovasc Haematol Disord 2001; 1:59–74. Frere C, Tregouet DA, Morange PE, Saut N, Kouassi D, Juhan-Vague I, et al. Fine mapping of quantitative trait nucleotides underlying thrombinactivatable fibrinolysis inhibitor antigen levels by a transethnic study. Blood 2006; 108:1562–1568. Franco RF, Fagundes MG, Meijers JC, Reitsma PH, Lourenco D, Morelli V, et al. Identification of polymorphisms in the 50 -untranslated region of the TAFI gene: relationship with plasma TAFI levels and risk of venous thrombosis. Haematologica 2001; 86:510–517. Morange PE, Aillaud MF, Nicaud V, Henry M, Juhan-Vague I. Ala147Thr and Cþ1542G polymorphisms in the TAFI gene are not associated with a higher risk of venous thrombosis in FV Leiden carriers. Thromb Haemost 2001; 86:1583–1584. Zidane M, De Visser MC, Ten Wolde M, Vos HL, De Monye W, Bertina RM, et al. Frequency of the TAFI 438 G/A and factor XIIIA Val34Leu polymorphisms in patients with objectively proven pulmonary embolism. Thromb Haemost 2003; 90:439–445. Procare-GEHT Group. ABO blood group but not haemostasis genetic polymorphisms significantly influence thrombotic risk: a study of 180 homozygotes for the Factor V Leiden mutation. Br J Haematol 2006; 135:697–702. Verdu J, Marco P, Benlloch S, Sanchez J, Lucas J. Thrombin activatable fibrinolysis inhibitor (TAFI) polymorphisms and plasma TAFI levels measured with an ELISA insensitive to isoforms in patients with venous thromboembolic disease (VTD). Thromb Haemost 2006; 95:585–586. Lichy C, Dong-Si T, Reuner K, Genius J, Rickmann H, Hampe T, et al. Risk of cerebral venous thrombosis and novel gene polymorphisms of the coagulation and fibrinolytic systems. J Neurol 2006; 253:316–320. Martini CH, Brandts A, De Bruijne EL, Van Hylckama Vlieg A, Leebeek FW, Lisman T, Rosendaal FR. The effect of genetic variants in the thrombin activatable fibrinolysis inhibitor (TAFI) gene on TAFI-antigen levels, clot lysis time and the risk of venous thrombosis. Br J Haematol 2006; 134: 92–94.
19
20
21
22
23
24
25
26
27 28
29 30 31 32
33
34
Xu X. The correlation between thrombin activatable fibrinolysis inhibitor (TAFI) gene polymorphisms and pulmonary thromboembolism in Chinese Han population [dissertation]. Capital Medical University; 2007. http://d.wanfa ngdata.com.cn/Thesis/Y1069064 (in Chinese). [Accessed 6 Nov 2015] De Bruijne EL, Darwish Murad S, De Maat MP, Tanck MW, Haagsma EB, Van Hoek B, et al. Genetic variation in thrombin-activatable fibrinolysis inhibitor (TAFI) is associated with the risk of splanchnic vein thrombosis. Thromb Haemost 2007; 97:181–185. Verdu J, Marco P, Benlloch S, Lucas J. Association between the Thr325Ile and Ala147Thr polymorphisms of the TAFI gene and the risk of venous thromboembolic disease. Clin Appl Thromb Hemost 2008; 14:494–495. Zhang BC. Correlation between TAFI 505A > G and 1040C > T single nucleic acid polymorphisms and VTE [dissertation]. Capital Medical University; 2010. http://d.wanfangdata.com.cn/Thesis/Y1770631 (in Chinese). [Accessed 6 Nov 2015] Hoekstra J, Guimaraes AH, Leebeek FW, Darwish Murad S, Malfliet JJ, Plessier A, et al. Impaired fibrinolysis as a risk factor for Budd-Chiari syndrome. Blood 2010; 115:388–395. Li C, Hou LH, Liu XE, Huang C, Wei XH. Correlation analysis of thrombinactivatable fibrinolysis inhibitor single nucleotide polymorphism with venous thromboembolism [in Chinese]. Zhongguo Shi Yan Xue Ye Xue Za Zhi 2012; 20:949–953. Tokgoz S, Zamani AG, Durakbasi-Dursun HG, Yilmaz O, Ilhan N, Demirel S, et al. TAFI gene polymorphisms in patients with cerebral venous thrombosis. Acta Neurol Belg 2013; 113:291–297. Stang A. Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol 2010; 25:603–605. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 21:1539–1558. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22:719–748. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7:177–188. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994; 50:1088–1101. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315:629–634. Qian K, Xu J, Wan H, Fu F, Lu J, Lin Z, et al. Impact of genetic polymorphisms in thrombin activatable fibrinolysis inhibitor (TAFI) on venous thrombosis disease: a meta-analysis. Gene 2015; 569:173–181. Knoefler R, Ludwig K, Kostka H, Kuhlisch E, Siegert G, Suttorp M. The impact of single nucleotide polymorphisms of the thrombin activatable fibrinolysis inhibitor (TAFI) gene on TAFI antigen levels in healthy children and pediatric oncology patients. Semin Thromb Hemost 2003; 29:575– 583. Guimaraes AH, Bertina RM, Rijken DC. A new functional assay of thrombin activatable fibrinolysis inhibitor. J Thromb Haemost 2005; 3:1284–1292.
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
