If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Some studies have reported that vascular endothelial growth factor (VEGF) genetic polymorphisms are associated with recurrent pregnancy loss (RPL), but the results are controversial. This study is aimed to quantify the strength of this association.
Methods
A systematic review of the published literature from Medline, Springer, and China National Knowledge Infra structure (CNKI) databases was conducted and investigations of VEGF genetic polymorphisms in RPL were selected. We estimated the pooled odds ratio (OR) to assess this possible association.
Results
Fifteen case-control studies comprising 2702 cases and 2667 controls and including five genetic polymorphisms (rs3025039, rs833061, rs15703060, rs2010963 and rs699947) were eligible for this meta-analysis. The overall analysis suggested that only two genetic polymorphisms (rs1570360, rs3025039) were associated with increased risk of RPL. A significant increased risk between VEGF rs1570360 polymorphism and RPL was only found under the dominant model in Caucasians (OR = 1.70, 95% CI 1.02-2.82, P = 0.04). Whereas, we found that VEGF rs3025039 polymorphism was significantly associated with RPL both under the dominant and recessive model in East Asians, and their summary odd ratios and 95% CIs were 1.26, 1.04-1.53, P = 0.02 and 2.94, 1.80-4.83, P = 0, respectively.
Conclusions
This meta-analysis showed that only rs1570360 (especially in Caucasians) and rs3025039 (especially in East Asians) may be risk factors for RPL.
Recurrent pregnancy loss (RPL), defined as a pregnancy failure occurring before 24 weeks of gestation more than two or three times according to most definitions, affects 1–5% of reproductive age women [
]. Various factors have been identified as being related to pregnancy loss, such as parental chromosomal anomalies, thrombophilia, metabolic disorders, uterine abnormalities, immune factors, maternal infections, and lifestyle factors [
Genetic association studies of angiogenesis-and vasoconstrictionrelated genes in women with recurrent pregnancy loss: a systematic reviewand meta-analysis.
The vascular endothelial growth factor (VEGF) family is comprised of VEGF, VEGF-B, VEGF-C, VEGF-D and placental growth factor. Of these factors, VEGF is a major factor in angiogenesis and a prime regulator of endothelial cell proliferation [
]. Over-expression of VEGF has been observed in a variety of tissues in physiological and pathological conditions and is believed to be involved in various diseases, including abnormalities of embryo development, cancers, and cerebrovascular or cardiovascular diseases. During early gestation, VEGF is essential for the maturation of oocytes, the proliferation of trophoblasts, the implantation and development of the embryo, the angiogenesis of the placenta, and the growth of maternal and fetal blood vessels in the uterus [
]. Several VEGF polymorphisms VEGFA −2578C/A (rs699947), −460T/C (rs833061), −1154G/A (rs15703060), −634G/C (rs2010963), 398G/A (rs833068), 497G/A (rs833070), −583T/C (rs3025020) and 936C/T (rs3025039) have been reported to affect VEGF activity and expression. However, the genetic association is inconclusive.
Considering the importance of VEGF in human pregnancy, we investigated the role of VEGF genetic polymorphisms in RPL. More than six studies were reported in five genetic polymorphisms (rs1570360, rs699947, rs2010963, rs3025039 and rs833061). Therefore, we chose these five candidate genetic polymorphisms for a systematic review of their association with RPL.
Materials and methods
Search strategy
The timeline of a comprehensive literature search of databases from Medline, Springer, and China National Knowledge Infra structure (CNKI) was up to September 2015 without language restrictions. The key search terms were used: (“VEGF” or “vascular endothelial growth factor”) and (“Abortion, Habitual” or “recurrent miscarriage” or “recurrent pregnancy loss”). Eligible reports were identified by two reviewers and controversial studies were adjudicated by the third reviewer. The search results were limited to humans.
Selection criteria
The studies included for this meta-analysis must meet the following requirements: (1) case-control studies; (2) RPL defined as two or more losses in the first two trimesters of pregnancy; (3) RPL with known causes was excluded; (4) genotypes identified by DNA analysis; (5) information on genotype frequency was provided. In studies with overlapping cases or controls, the latest or the largest study with extractable data was included in the meta-analysis.
Data acquisition
Two reviewers independently used a standardized form with eligibility criteria to extract data. Disagreements were resolved by discussion or the third reviewer. The information included: name of the first author, year of publication, country of origin, number of cases and controls, mean age (if possible), and genotype distribution.
Quality assessment
The Newcastle-ottawa quality assessment scale (NOS) was used to assess the quality of case-control studies. This scale judged the studies on three broad perspectives: selection, comparability and exposure. Scores were ranged from 0 to 9. Studies with a score ≥7 were considered to be of high quality [
Chi-squared analysis with exact probability was used to test the departure from Hardy-Weinberg equilibrium (HWE) for VEGF genetic polymorphism in the control group, and P < 0.05 was considered representative of departure from HWE. We evaluated five genetic polymorphisms (rs1570360, rs699947, rs2010963, rs3025039 and rs833061). Subgroup analysis was performed based on ethnicity (East Asian, Caucasian, and other population). Odds ratio (OR) represents the effect size, and 95% confidence intervals (CI) was assessed for interval estimates. Heterogeneity was tested using I2 statistic and Cochran’s Q test to quantify the proportion of the total variation. The corresponding p-value of the Q statistic below 0.05 or I2 > 50% was considered significant heterogeneity. In the presence of heterogeneity, the random-effect model was adopted as the pooling the data, otherwise, the fixed-effect model was used. The sensitivity analysis was performed by omitting each individual study and an individual study was suspected of excessive sensitivity if the point estimate of its omitted analysis lay outside the 95% CI of the combined analysis. The potential publication bias was estimated by visual inspection of Begg’s funnel plot and the Egger’s test were used to provide evidences. To consider potential geographic variation, we performed subgroup analysis stratified by geographic position. All statistical analysis was performed by using STATA version 12.0 (Stata Corp, College Station, TX, USA). All the P values were two-sided.
Identification of specific vascular endothelial growth factor susceptible and protective haplotypes associated with recurrent spontaneous miscarriages.
] involving 2702 cases and 2667 controls were eligible for this meta-analysis on the relation of VEGF genetic polymorphism to RPL risk. Of these eligible studies, all were case-control designs and conducted in eight countries: five in China, two each in the USA and Bahrain, and one each in other six countries (Table 1). All eligible studies were transcribed in English, except three in Chinese (16, 19, and 21). Six relational studies were found on rs833061 polymorphism (17–22), twelve studies on rs3025039 (9, 10, 12, 13, 15, 17–23), nine studies on rs2010963 (9, 10, 12, 13, 17, 18, 21–23), fourteen studies on rs1570360 (9, 11–23), and eleven studies on rs699947 (9, 10, 12, 13, 15, 17–22). The number of cases varied from 38 to 339, and the number of controls varied from 30 to 371. Two or more pregnancy losses were used to define RPL in 8 studies, and three or more pregnancy loss were found in the remaining 7 studies. Seven studies recruited cases before 20 weeks, but eight studies did not described the gestational age of pregnancy loss. At least one live birth and no history of miscarriage in nine studies were selected as controls and more than two live births without abortion were in six studies. All included studies excluded ‘known’ causes for RPL. General characteristics in the published studies included in this meta-analysis were listed in Table 1, Table 2. The Table 3 showed the results of this meta-analysis.
Table 1Characteristics of studies association between VEGF genetic polymorphisms and RPL.
Author (year)
Country (ethnicity)
Case
Controls
NOS scores
n
Mean age (years)
Number of miscarriages
Gestational age
n
Mean age (years)
Number of live births
Papazolou 2005
Greek Caucasian
52
33
3
<20 weeks
82
NA
2
8
Lee 2010
Korea East Asian
215
32.8
2
NA
113
31.1
1
7
Xing 2011
China East Asian
339
29.7
3
NA
291
30.8
1
8
Eller 2011
USA Caucasian
93
30
3
<20 weeks
178
28
1
8
Magdoud 2012
Bahrain Other
304
32.4
3
NA
371
31.9
2
8
Coulam 2008
USA Caucasian
152
NA
2
NA
65
NA
1
7
Aggarwal 2011
India Other
200
28
3
<20 weeks
200
31.9
2
8
Li 2010
China East Asian
271
29
2
<20 weeks
250
30
1
8
Su 2011
Taiwan East Asian
115
31
2
NA
170
29.8
1
7
Almawi 2013
Bahrain Other
52
31.6
3
NA
82
31.6
2
8
Zhu 2014
China East Asian
203
27.6
2
<20 weeks
190
27.3
1
7
Samli 2012
Turkey Other
38
30.1
2
<20 weeks
25
36.8
1
7
Huang 2013
China East Asian
120
29.9
2
<20 weeks
110
28.8
1
7
Luo 2013
China East Asian
227
30.6
3
NA
232
30.8
2
7
Traina 2011
Brazil Other
77
30.4
3
NA
85
33.2
2
8
NA: not available. NOS: Newcastle-ottawa quality assessment scale.
The most commonly studied allelic variant in VEGF was the rs1570360 polymorphism. Fourteen eligible studies included a total of 1855 cases and 2014 controls for analysis. Under the recessive genetic model (AA vs. GG + GA), significant heterogeneity among the study was observed (Cochran’s Q test: χ2 = 28.40, p-value = 0.008; Higgins statistics: I2 = 54.2%), so random-effect model was used. The summary ORs and 95% CI for fixed-effect model showed significant association between the rs1570360 polymorphism and RPL both was only observed in Caucasian (OR = 1.70, 95% CI 1.02–2.82) (Fig. 1). Under the dominant genetic model, no significant among-study heterogeneity was observed (Cochran’s Q test: χ2 = 13.68, p-value = 0.39; Higgins statistics: I2 = 5%). However, the association between rs1570360 polymorphism and RPL was not observed in any ethnicity (Table 3). Publication bias was not detected in the included studies (Egger’s P-value = 0.88). No individual study was found to have an excessive effect on the overall results (data not shown).
Fig. 1Association of rs1570360 polymorphism and RPL risk under the recessive genetic model.
Eleven studies with a total of 1866 RPL women and 1980 normal controls investigated the association between rs699947 polymorphism and RPL. The heterogeneity among each study was not significant in both the dominant (AA + CA vs. CC) and recessive model (AA vs. CA + CC) (Cochran’s Q test: χ2 = 6.06, p-value = 0.21; Higgins statistics: I2 = 0, fixed-effects model in the recessive model). Meta-analysis showed no significant association between VEGF rs699947 polymorphism and RPL in any ethnicity (Table 3). Tests for publication bias did not detect a significant bias (Egger’s P-value = 0.26). No individual study was found to have an excessive effect on the overall results (data not shown).
rs2010963 polymorphism and RPL
Nine studies comprising 1502 cases and 1647 controls were included for this meta-analysis. No significant heterogeneity was observed among these included studies both in the dominant (CC + CG vs. GG) and recessive mode (CC vs. GG + CG) (Cochran’s Q test: χ2 = 3.97, p-value = 0.86; Higgins statistics: I2 = 0, fixed-model effects in the recessive model). Meta-analysis showed no significant association between rs2010963 polymorphism and RPL in any genetic model. Publication bias was not detected in the included studies (Egger’ s P-value = 0.85). No individual study was found to have an excessive effect on the overall results (data not shown).
rs833061 polymorphism and RPL
Six studies, without anyone being excluded from HWE, included a total of 999 patients and 1037 controls for analysis. The heterogeneity among each study was significant in recessive genetic model (CC vs. CT + TT) (Cochran’s Q test: χ2 = 16.92,p-value = 0.01; Higgins statistics: I2 = 70.4%, random-effects model in the recessive model). Meta-analysis showed no significant association between rs833061 polymorphism and RPL in any ethnicity (Table 3). Tests for publication bias did not detect a significant bias (Egger’ s P-value = 0.89). Two studies (Almawi 2013, Zhu 2014) were found to have an excessive effect on the overall results. After excluding the two articles, the meta-analysis showed an increased association between rs833061 genetic polymorphism and RPL, but this association was not significant (OR = 1.18, recessive model).
rs3025039 polymorphism and RPL
Twelve studies with one being excluded from HWE (Zhu 2014) included 1947 RPL patients and 2110 controls in this meta-analysis. The overall results showed significant association between rs3025039 polymorphism and RPL both under dominant (TT + TC vs. CC) and recessive genetic models (TT vs. TC + CC). The heterogeneity among these included studies did not show significance (Cochran’s Q test: χ2 = 18.64, p-value = 0.07; Higgins statistics: I2 = 41%, fixed-model effects in the recessive model). Subgroup analysis showed that the association between rs3025039 polymorphism and RPL was only observed in East Asian.Under the dominant and recessive models, respectively, the overall ORs and 95% CIs were 1.27, 1.12–1.46, P = 0 (1.26, 1.04–1.53, P = 0.02, in East Asians) and 2.11, 1.47–3.03, p = 0.02 (2.94, 1.80–4.83, P = 0, in East Asians) (Fig. 2). Publication bias was not detected in the included studies (Egger’s P-value = 0.40). No individual study was found to have an excessive effect on the overall results (data not shown).
Fig. 2Association of rs3025039 polymorphism and RPL risk under the dominant and recessive genetic model. (A) Dominant model. (B) Recessive model.
Vascular endothelial growth factor (VEGF) plays an important role in inadequate vascular formation and oocyte maturation, decidualized endometrial vascularization, and embryo implantation. In human early pregnancy, the state of chorionic villi vascularization is close to embryonic development, and diminished placental trophoblastic VEGF has been described in the decidual endothelium of recurrent miscarriages [
]. Many studies have reported the association between VEGF genetic polymorphisms and RPL, but the results are controversial, so this meta-analysis was performed to quantify this association.
In this systematic review of studies on the association between five VEGF genetic polymorphisms (rs1570360, rs699947, rs2010963, rs3025039 and rs833061) and RPL risk. The present meta-analysis included 2702 cases and 2667 controls from 15 independent case-controls studies. More than six studies were included for this meta-analysis in each genetic polymorphism. Our results showed that rs1570360 and rs3025039 were significantly associated with RPL. The summary ORs of rs699947, rs2010963 and rs833061 failed to show significance in RPL after meta-analysis. However, when subgroup (East Asian, Caucasian and other population) analysis by ethnicity was performed, our results showed that significant association between rs1570360 and RPL risk was observed in Caucasian. However, this significant association was observed in East Asian between rs3025039 and RPL risk.
In the present meta-analysis, attempts were made to avoid general limitations such as selection bias and publication bias that might exist in the original papers. The most obvious difference would be the definition of RPL. Only the patients of RPL in three studies [
] met three or more miscarriages before 20 weeks of gestation age. Many chosen groups include women with two consecutive miscarriages and did not describe the precise pregnancy loss of gestation age. By combining these datas, it could indicate that AA genotype of rs1570360 polymorphism and TT genotype of rs3025039 polymorphism in two pregnancy losses might be an important risk factor for RPL.
Another potential difference would be the gestational age at pregnancy loss. Only seven studies described the gestational age of pregnancy loss (all were before 20 weeks), and the gestational age at pregnancy loss in remaining eight studies were unknown. Some other potential less obvious biases would be the maternal age. The mean age ranged from 27.6 to 33 years in cases and 27.3–37 years in controls (Table 1). Some of the studies did not provide the information about this factor and therefore we could not integrate this variable into this meta-analysis. Although allele/genotype are similar in different studies in the same ethnicity, different allele/genotype frequencies of the five genetic polymorphisms exist in different ethnic groups. This would be an another factor that affect the results. Besides, our study only included articles published in English and Chinese from the three selected databases, and the number of studies and the sample sizes were relatively small for analysis of each gene polymorphism thereby having insufficient power to estimate the association between VEGF genetic polymorphism and RPL risk.
We used sensitivity analysis to attempt to explore the potential important causes of the between-study heterogeneity for the studies. Our results found that significant between-study heterogeneity was observed only in one genetic polymorphism (rs833061). After excluding the two articles, lower heterogeneity (I2 = 0, and the previous I2 was 70.4%) was found in the allele model (C vs. T), and the meta-analysis showed an increased association between rs833061 genetic polymorphism and RPL, but this association was not significant (OR = 1.18, recessive model). Moreover, in our meta-analysis, we could not find any significant publication bias in any inheritance model by Begg’ s funnel plot and the Egger’ s test (data not shown).
Conclusions
Only two of five selected VEGF genetic polymorphisms (rs1570360 and rs3025039) were found to show significant association with the occurrence of RPL. Since we could not completely preclude the potential bias and confounders in this meta-analysis, further studies with more samples and standardized unbiased genotyping methods are needed to confirm our findings.
Funding
None.
Conflict of interest
The authors declare that they have no conflict of interest.
Acknowledgments
Wen Zhang provided language help.
References
Vaiman D.
Genetic regulation of recurrent spontaneous abortion in humans.
Genetic association studies of angiogenesis-and vasoconstrictionrelated genes in women with recurrent pregnancy loss: a systematic reviewand meta-analysis.
Identification of specific vascular endothelial growth factor susceptible and protective haplotypes associated with recurrent spontaneous miscarriages.
30107-0&id=gr1.jpg){kind=link}
30107-0&id=gr2.jpg){kind=link}