Genetic association between plasminogen activator inhibitor‐1 rs1799889 polymorphism and venous thromboembolism: Evidence from a comprehensive meta‐analysis

Abstract Background Association between plasminogen activator inhibitor‐1 (PAI‐1) rs1799889 polymorphism and venous thromboembolism (VTE) were explored by many previous studies, yet the findings of these studies were conflicting. Hypothesis PAI‐1 rs1799889 polymorphism may serve as a genetic marker of VTE. We aimed to better clarify the relationship between PAI‐1 rs1799889 polymorphism and VTE in a larger combined population by performing a meta‐analysis. Methods Literatures were searched in Pubmed, Embase, Web of Science, and China National Knowledge Infrastructure (CNKI). We used Review Manager to combine the results of individual studies. Results Forty‐eight studies involving 14 806 participants were eligible for inclusion. Combined results revealed that PAI‐1 rs1799889 polymorphism was significantly associated with VTE in Caucasians (dominant comparison: odds ratio [OR] 1.20, 95% confidence interval [CI] 1.09‐1.32; recessive comparison: OR 0.84, 95% CI 0.76‐0.94; allele comparison: OR 1.08, 95% CI 1.02‐1.15) and East Asians (dominant comparison: OR 1.60, 95% CI 1.17‐2.19; allele comparison: OR 1.53, 95% CI 1.21‐1.93). Further analyses obtained similar significant associations in these with deep vein thrombosis (DVT) and these with Factor V Leiden mutation. Conclusions Our findings supported that PAI‐1 rs1799889 polymorphism may serve as one of the predisposing factors of VTE in both Caucasians and East Asians, especially in these with DVT and these with Factor V Leiden mutation.

thrombosis via inhibiting biological activities of tissue plasminogen activator (t-PA) and urokinase plasminogen activator (u-PA). 6 Previous basic researches showed that blockage of PAI-1 could lead to thrombus degradation, whereas activation of PAI-1 could accelerate thrombus formation. 7,8 So if a gene polymorphism could alter the expression level or protein structure of PAI-1, it is possible that this polymorphism may also affect individual susceptibility to thrombotic disorders like VTE.
In this meta-analysis, we aimed to better clarify the relationship between PAI-1 rs1799889 A/G polymorphism and VTE. We will also perform comprehensive analyses to investigate the effects of ethnic background, type of disease, and established risk factors of VTE (Factor V Leiden mutation, cancer status, and recent major surgery) on genetic association between PAI-1 rs1799889 A/G polymorphism and VTE.

| MATERIALS AND METHODS
This meta-analysis was written in accordance with PRISMA guideline. 13 We also created an Open Science Framework (osf.io) account to make this meta-analysis more publicly available.
To be included in this meta-analysis, some criteria must be met: (b) studies that were not carried out in humans; (c) case reports or case series; (d) reviews and comments. If we found repeated publications during literature searching, only the most comprehensive study was included for analyses. The authors used Newcastle-Ottawa scale (NOS) to assess the quality of eligible studies. 14 The score range of NOS is between zero and nine, when a study got a score of seven or more, we considered that the methodology of this study is good.

| Data extraction and quality assessment
Two authors extracted data and assessed quality of eligible publications. The authors wrote to the leadings authors for additional information if essential information was found to be incomplete.

| Statistical analyses
Review Manager was used to combine the results of eligible studies.
Z test was employed to assess whether PAI-1 rs1799889 A/G polymorphism was significantly associated with VTE, with the statistical significance P level set at .05. We used I 2 statistics to assess betweenstudy heterogeneities. We used Random-effect models (DerSimonian-Laird method) to combine the results if I 2 is larger than 50%. Otherwise, fixed-effect models (Mantel-Haenszel method) were used to combine the results. We also conducted subgroup analyses by ethnic- of irrelevant and duplicate articles. Another eighty-five articles were further excluded by us because these articles did not meet the inclusion criteria that were set forth for this meta-analysis. Totally fortheight studies containing 5731 cases and 9075 controls were ultimately included in this meta-analysis (see Figure 1). Table 1 presented essential data extracted from included studies.  Table 2).

| Meta-analysis results
We also performed stratified analyses to explore the effects of established risk factors of VTE on observed genetic associations between PAI-1 rs1799889 A/G polymorphism and VTE, and we found positive results in these with Factor V Leiden mutation, whereas no any significant associations were detected in these with cancer or these who recently had a major surgery operation.

| Sensitivity analyses
We examined the stability of combined results by deleting one study each time and combining the results of the remaining studies. The trends of associations remained consistent in sensitivity analyses, which indicated that the combined results were statistically stable.

| Publication biases
Funnels plots were employed to estimate whether our combined results may be influenced by publication biases. Funnel plots of every comparison were symmetrical, which indicated that the combined results were unlikely to be seriously impacted by overt publication biases (see Figure S1).

| DISCUSSION
The A/G variant of rs1799889 polymorphism is associated with a guanosine insertion at the −675 site of the PAI. 15 Past pre-clinical studies also demonstrated that the transcriptional activity of A allele was significantly higher than that of the G allele. 16 Besides, more stratified analyses should also be conducted by future meta-analyses if there are sufficient data to support additional analyses for other established risk factors of VTE. Fifthly, although we also conducted subgroup analyses by type of disease in this meta-analysis, it is noteworthy that studies only focused on PE were scare, so the results of subgroup analyses by type of disease should also be taken as exploratory. Future studies are still needed to confirm these findings.
Some limitations of this meta-analysis should also be mentioned.
Firstly, the results regarding associations between polymorphisms in PAI-1 rs1799889 polymorphism and VTE were based on combining unadjusted findings of eligible publications due to lack of raw data. 21 Secondly, gray literatures were not searched. So although funnel plots of every comparison were symmetrical, it is still possible that the combined results may be affected by publication biases. 22,23

| CONCLUSIONS
In summary, the combined results of this meta-analysis proved that PAI-1 rs1799889 A/G polymorphism may serve as one of the predisposing factors of VTE in both Caucasians and East Asians, especially in these with DVT and these with Factor V Leiden mutation.
Further studies with larger sample sizes still need to verify our findings. Besides, given that the pathogenesis of VTE is complex, despite our comprehensive analyses, we still recommend further studies to explore gene-gene interactions and gene-environmental interactions in the development of VTE.