Introduction

Preeclampsia (PE) is a common complication of pregnancy characterized by hypertension and proteinuria after 20 weeks of gestation [1]; it is one of major causes of maternal-fetal and neonatal morbidity and mortality worldwide [2]. Knowing the risk factors for preeclampsia is critical for its prevention and treatment. Genetic factors play an important role in the genesis and development of PE and the genetic susceptibility to preeclampsia has generated great attention; the T allele of AGT may play a role in the pathogenesis of PE reported by Aung et al. [3],which indicated the gene polymorphisms in the renin-angiotensin-aldosterone system (RAAS) may be risk factors to PE.

During normal pregnancy, the upregulation of renin and aldosterone triggered by the stimulation of the RAAS system maintains the balance of blood volume and blood pressure [4]; however, for PE subjects, depression of the RAAS system with increased vascular resistance was observed, suggesting its crucial role in the pathogenesis of PE [5]. Angiotensin (AGT), angiotensin converting enzyme (ACE), and angiotensin II type 1 receptor (AT1R) are the three pivotal nodes in the RAAS system. The cleavage of AGT by renin contributes to the generation of angiotensin I, then ACE catalyzes the conversion of angiotensin I to a physiologically active angiotensin II. Finally, by binding to AT1R, angiotensin II regulates blood pressure by controlling sodium excretion [6]. Therefore, studies regarding the associations between single nucleotide polymorphisms in RAAS genes and PE risk are essential.

Associations between the polymorphisms of AGT T704C (the substitution of C to T at exon 2), ACE I/D (the insertion or deletion of an Alu 289 base pair sequence at intron 16), and AT1R A1166C (the change from C to A at 3’UTR) have been widely studied with conflicting results. To our best knowledge, differences in the geographic regions, ethnicity, and sample size could be reasons for the inconsistency. Moreover, the number of gestational weeks and the severity of PE have been reported to be associated with RAAS susceptibility gene polymorphisms [7,8,9], but these were not discussed in previous meta-analyses. Therefore, we conducted a comprehensive meta-analysis with trial sequential analysis to investigate the associations between the polymorphisms AGT T704C, ACE I/D, AT1R A1166C, and PE risk.

Methods

Literature search

PubMed, Embase, Google scholar, China National Knowledge Internet (CNKI), Baidu Scholar, Wan Fang, and VIP databases were comprehensively searched for studies regarding the associations between ACE insertion/deletion, AGT T704, and AT1R A1166C polymorphisms and preeclampsia susceptibility up to May 13, 2018. No language limitation was set. The following key words were used to discover relevant articles: “angiotensin-converting enzyme,” “angiotensin,” “angiotensin II type 1 receptor,” “ACE,” “AGT,” “AT1R,” “polymorphism,” “variant,” “single nucleotide polymorphism,” “SNP,” “preeclampsia,” “PE,” “hypertension,” and “pregnancy-induced hypertension syndrome.” The references of relevant studies were also screened by hand to identify potential studies. Our work was based on the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement [10] (Fig. 1).

Fig. 1
figure 1

PRISMA 2009 flow diagram

Full size image

Inclusion and exclusion criteria

The inclusion criteria for studies were as follows: (1) case-control studies discussing the relationship between ACE I/D, AGT T704C, AT1R A1166C polymorphisms, and preeclampsia risk; (2) the diagnostic criteria for preeclampsia were defined as gestational hypertension, assessed as SBP > 140 mmHg, DBP > 90 mmHg, and/or rise in SBP > 30 mmHg or DBP > 15 mmHg on at least two occasions 6 h apart, following 20 weeks of gestation, with marked proteinuria (> 300 mg/24 h), or > 2+ proteinuria as tested by the dipstick method [5, 11, 12]; (3) the frequencies of the related polymorphisms in patients and controls could be retrieved to calculate odds ratio with 95% confidence intervals and to assess Hardy-Weinberg equilibrium. The exclusion criteria were (1) reviews or case reports or animal studies; (2) studies without reporting detailed genotype data; and (3) duplicated studies.

Data extraction and quality assessment

The following information from eligible studies were extracted by the first two authors: the first author’ name, publication year, country, geography, ethnicity, PE maternal age, gestational weeks, PE degree, the genotype distributions and alleles in the patient and control groups, the result of the Hardy-Weinberg equilibrium, and the scores for quality assessment. For gestational weeks, early-onset PE was defined as gestational age (GA) between 20 and 33 weeks and 6 days, and late-onset PE was defined as GA 34 weeks and above. Severe PE was defined as severe hypertension (blood pressure ≥ 160/110 mmHg at least twice in a 24-h period) and/or severe proteinuria (5 g/24 h), or as hypertension with multiorgan involvement including fetal growth restriction or HELLP syndrome (hemolysis, elevated liver enzymes, and low platelet count) [13]. Any disagreement was resolved by group discussion with the corresponding author. The qualities of included studies were assessed by all the authors in accordance with the modified Newcastle-Ottawa Scale (NOS) (Table S1) [14]. Studies with scores of 7 points or higher were considered to be of high quality.

Statistical analysis

The odds ratio (OR) and 95% confidence interval (95%CI) were calculated to investigate the effect strength of the associations between ACE I/D, AGT T704C, AT1R A1166C polymorphisms, and preeclampsia risk. The following genetic models were used: allelic genetic model (ACE I/D: D VS I; AGT T704C: C VS T; AT1R A1166C: C VS A), dominant genetic model (ACE I/D: DD + DI VS II; AGT T704C: CC + CT VS TT; AT1R A1166C: CC + CA VS AA), recessive genetic model (ACE I/D: DD VS DI + II; AGT T704C: CC VS CT + TT; AT1R A1166C: CC VS CA + AA), heterozygote genetic model (ACE I/D: DI VS II; AGT T704C: CT VS TT; AT1R A1166C: CA VS AA), and homozygote genetic model (ACE I/D: DD VS II; AGT T704C: CC VS TT; AT1R A1166C: CC VS AA). The Hardy-Weinberg equilibrium was assessed by the chi-squared test for every study in the control group. Heterogeneity in the meta-analysis was determined by the Cochrane’s Q-statistic test, and the inconsistency was quantified with the I2 statistic (I2 value more than 50% or P value less than 0.10 was considered significant heterogeneity and the random effect model was used, otherwise, the fixed-effect model was used). Sensitivity analysis was performed by omitting one study at a time to assess the influence of each study on the pooled results. Subgroup analysis was conducted, stratifying by geography (Asian, Europe, Africa, America and Australia), ethnicity (Caucasoid, Mongoloid, Black, Mixed race), gestational week (early-onset, late-onset, mixed), PE degree (severe, mild, not mentioned), and patient sample size (less than 100, between 100 and 200, more than 200). Publication bias was evaluated by a visual inspection of funnel plot and Egger’s test [15]. If publication bias existed, the “trim and fill” method was used; this method conservatively imputes hypothetical negative unpublished studies to mirror the positive studies that cause funnel plot asymmetry to further assess the possible effect of publication bias [16, 17]. All analyses were performed by Review Manager 5.3 and STATA 12.0 software packages and P < 0.5 was considered statistically significant.

Trial sequential analysis

TSA (trial sequential analysis) (The Copenhagen Trial Unit, Center for Clinical Intervention Research, Denmark) is a methodology that combines an information size calculation (accumulated sample sizes of all included trials) to reduce type I error and type II error for a meta-analysis with the threshold of statistical significance (http://www.ctu.dk/tsa). TSA was introduced into our meta-analysis. The required information size was calculated based on an overall type I error of 5%, a power of 90%, and a relative risk reduction (RRR) assumption of 10%.

Results

The characteristics of eligible studies

Table 1 and Fig. 1 show the main characteristics of the included studies and the study selection flow chart, respectively. A total of forty studies were finally included in our meta-analysis [1, 5, 7,8,9, 18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52], among which thirty-four studies involving 3977 patients and 7065 controls regarded the ACE I/D polymorphism, eighteen studies involving 1814 patients and 2892 controls regarded associations with AGT T704C polymorphism, and twelve studies involving 2391 cases and 6604 controls regarded the AT1R A1166C polymorphism.

Table 1 Characteristic of included studies regarding the associations between ACE insertion/deletion, AGT T704C, AT1R A1166C polymorphims and PE risk
Full size table

Meta-analysis results

Table 2 summarizes the overall and subgroup results regarding the associations between the ACE I/D, AGT T704C, and AT1R A1166C polymorphisms and PE risk. Extensive significant associations were observed for ACE I/D and AGT T704C polymorphisms; however, for the AT1R A1166C polymorphism, no association was detected.

Table 2 Overall and subgroup analysis of associations between ACE insertion/deletion, AGT T704C, AT1R A1166C polymorphisms, and PE risk
Full size table

AGT T704C polymorphism

As summarized in Table 2, the overall analysis indicated that the AGT T704C polymorphism was associated with PE risk in three genetic models (dominant genetic model: CC+CT VS TT: OR = 1.33, 95%CI = 1.12–1.59 (Fig. 3); heterozygote genetic model: OR = 1.26, 95%CI = 1.05–1.52: homozygote genetic model: OR = 1.44, 95%CI = 1.14–1.83). No heterogeneity was observed in the three genetic models. For subgroup analysis by geography, no significant association was detected (Fig. 4b). As stratified by ethnicity, the AGT T704C polymorphism was associated with PE risk both in Caucasoid and Mongoloid populations (Caucasoid: dominant genetic model: CC+CT VS TT: OR = 1.30, 95%CI = 1.05–1.60 (Fig. 5b); heterozygote genetic model: CT VS TT: OR = 1.28, 95%CI = 1.05–1.56. Mongoloid: allelic genetic model: C VS T: OR = 1.60, 95%CI = 1.04–.44; recessive genetic model: CCVS CT+TT: OR = 4.43, 95%CI = 2.57–7.62). No associations were also observed in the severe or the mild subgroup either. In the subgroup analysis by patient sample size, significant associations were detected in the dominant (CC+CT VS TT: OR = 1.60, 95%CI = 1.18–2.19), recessive (CC VS CT+TT: OR = 2.01, 95%CI = 1.50–2.71), and heterozygote (CT VS TT: OR = 1.46, 95%CI = 1.05–2.02) genetic model in more than 200 subgroups.

ACE I/D polymorphism

In the overall analysis, significant associations with significant heterogeneity were observed in the allelic genetic model (D VS I: OR = 1.29, 95%CI = 1.16–1.44), the dominant genetic model (DD+DI VS II: OR = 1.17, 95%CI = 1.05–1.31), the recessive genetic model (DD VS DI+II: OR = 1.52, 95%CI = 1.18–1.94), and the homozygote genetic model (DD VS II: OR = 1.55, 95%CI = 1.26–1.91) (Fig. 2). Galbraith plot analyses were performed to further explore the sources of heterogeneity, and the figure showed that the studies performed by Mello et al. [46], Gonzalez et al. [1], Choi et al. [45], Atalay et al. [26], Zhan1 et al. [32], Jiang et al. [34], and Ma et al. [18] primarily contributed to the heterogeneity. After excluding these studies, the heterogeneity decreased significantly (I2 = 21% and PHeterogeneity = 0.14 for D VS I; I2 = 6% and PHeterogeneity = 0.37 for DD+DI VS II; I2 = 14% and PHeterogeneity = 0.25 for DD VS DI+II; I2 = 0 and PHeterogeneity = 0.58 for DD VS II). For subgroup analysis stratified by geography, the ACE ID polymorphism was similarly associated with PE risk in three genetic models in the Asian population (allelic genetic model: D VS I: OR = 1.31, 95%CI = 1.13–1.53; recessive genetic model: DD VS DI+II: OR = 1.80, 95%CI = 1.33–2.43; homozygote genetic model: DD VS II: OR = 1.53, 95%CI = 1.16–2.01 (Fig. 4a)). Regarding the ethnicity subgroup analysis, significant associations were only observed in allelic (D VS I: OR = 1.39, 95%CI = 1.21–1.60) and homozygote genetic models (DD VS II: OR = 1.68, 95%CI = 1.30–2.17) in Caucasoid. However, for the subgroup analysis of gestational weeks, no significant association was detected in both early-onset and late-onset subgroups. In the severe PE subgroup, the ACE I/D polymorphism was associated with PE in allelic genetic (D VS I: OR = 1.53, 95%CI = 1.28–1.83), dominant (DD+DI VS II: OR = 1.50, 95%CI = 1.11–2.04), and homozygote (DD VS II: OR = 2.14, 95%CI = 1.49–3.09) genetic models. For the subgroup of patient sample size less than 100, wide associations with PE risk were observed in allelic (D VS I: OR = 1.41, 95%CI = 1.19–1.66), dominant (DD+DI VS II: OR = 1.37, 95%CI = 1.09–1.73), recessive (DD VS DI+II: OR = 1.50, 95%CI = 1.05–2.15), and homozygote (DD VS II: OR = 1.85, 95%CI = 1.37–2.51 (Fig. 5a)) genetic models..

Fig. 2
figure 2

Overall analysis of ACE I/D polymorphism and PE risk

Full size image
Fig. 3
figure 3

Overall analysis of AGT T704C polymorphism and PE risk

Full size image

AT1R A1166C polymorphism

As shown in Table 2, significant associations were observed in mixed race, early-onset, late-onset, and more than 200 subgroups; however, only one study was analyzed in these subgroups and the results required interpretation with caution (Figs. 4 and 5).

Fig. 4
figure 4

Subgroup analysis (stratified by geography) of ACE I/D and AGT T704C polymorphisms and PE risk

Full size image
Fig. 5
figure 5

Subgroup analysis (stratified by ethnicity) of ACE I/D and AGT T704C polymorphisms and PE risk

Full size image

Sensitivity analysis and publication bias

Sensitivity analysis was performed, and every study was omitted one a time, without any effect on our overall statistical results, indicating that the results were stable and reliable (Fig. 6). Begg’s and Egger’s test were conducted to analyze publication bias (P = 0.015 for ACE I/D polymorphism; P = 0.627 for AGT T704C polymorphism) (Fig. 7). Our results indicated that publication bias was existed in ACE I/D polymorphism; therefore, we applied a sensitivity analysis using the trim and fill method [16], which conservatively imputed hypothetical negative unpublished studies to mirror the positive studies that cause funnel plot asymmetry; the imputed studies of ACE I/D polymorphism produced a symmetrical funnel plot [53] (Fig. 7a). The shape of funnel plot was symmetrical for the AGT T704C polymorphism (Fig. 7b), implying that there was no publication bias for this polymorphism.

Fig. 6
figure 6

Sensitivity analysis of ACE I/D and AGT T704C polymorphisms and PE risk

Full size image
Fig. 7
figure 7

Begg’s and filled funnel plot of ACE I/D and AGT T704C polymorphisms and PE risk

Full size image

Trial sequential analysis

We performed a TSA for the homozygote genetic model of ACE I/D polymorphism and dominant genetic model of AGT T704C polymorphism (Fig. 8). The results of the two polymorphisms showed that the blue line of the cumulative z-curve crossed the TSA monitoring boundary and the cumulative sample size was reached, indicating that no further studies were essential to confirm the associations.

Fig. 8
figure 8

Trial sequential analysis of ACE I/D and AGT T704C polymorphisms and PE risk

Full size image

Discussion

In pregnant women with PE, downregulated renin-angiotensin system (RAS) activity is observed, resulting in increased vascular responsiveness to angiotensin II [4]. The increased plasm levels of angiotensin (AGT) and angiotensin converting enzyme (ACE) in PE subjects lead to the augmentation of angiotensin II [5, 54]; moreover, the pathophysiological effects of angiotensin II are enhanced by the upregulation of angiotensin II type 1 receptor (AT1R) [9], causing the dysregulation of blood pressure. Gene polymorphisms were reported to be associated with the abnormal expression of mRNA and protein [55, 56]. Our meta-analysis demonstrated that the polymorphisms of AGT T704C and ACE I/D were significantly associated with an increased risk of preeclampsia (PE) and weak associations of the AT1R A1166C polymorphism with PE were observed.

Previous meta-analyses indicated an increased PE risk with high heterogeneity of ACE I/D and AGT T704C polymorphisms, but no association was observed for the AT1R A1166C polymorphism [57,58,59,60]. However, the latest meta-analysis was performed in 2012, and in subsequent years, several studies conducted in different regions and ethnicities were published. An increased frequency of AT1R AC + CC genotypes in mild preeclamptic women was reported by Rahimi et al [9]. An interaction between the AGT T704C and ACE I/D polymorphisms and the risk of severe preeclampsia or the time onset of PE were observed [7, 8], but these were not analyzed in any former meta-analysis. Drawbacks in terms of high heterogeneity, slack inclusion criteria for subjects from different regions and ethnicities, the lack of evaluation of type 1 error and sample size on significant associations, the vague associations between these polymorphisms and the risk of severe PE, and different onset times of PE greatly aroused our interest. Therefore, we performed an updated meta-analysis with trial sequential analysis to consider the undiscussed above-mentioned issue. Regarding the AT1R A1166C polymorphism, significant associations in mixed race, early-onset, late-onset, and more than 200 patient sample size were discovered; however, only one study was analyzed in these subgroups, implying low representativeness of the AT1R A1166C polymorphism and further studies are essential.

In the overall analysis of the AGT T704C polymorphism, a 33% increased PE risk of CC + CT genotypes was observed. The 1.26-fold and 1.44-fold increased risk of PE in CT genotypes and CC genotypes, respectively, were also detected compared to TT genotypes. No heterogeneity in the genetic models and the positive results from the trial sequential analysis ensured the stability and reliability of our result. In the subgroup analysis stratified for geography, no significant association was detected; however, increased risks were observed in Caucasoid (the 1.30-fold and 1.28-fold increased risk of CC + CT genotype and CT genotype compared to TT genotype) and Mongoloid (the 60% increased of C allele in allelic genetic model; the 4.43-fold increased risk of DD genotype in recessive genetic model). In the severe PE degree subgroup analysis, no association was observed both in either severe or mild PE populations, possibly due to the small sample size, more studies are required. In the more than 200 patient sample size, increased risks were observed in the dominant, recessive, and heterozygote genetic models; however, the relatively small number of included studies in the subgroup indicated that these associations need to be interpreted with caution.

For the ACE I/D polymorphism, the D allele increased the risk of PE compared to I allele by 1.29-fold; moreover, the DD + DI, DD and DD genotypes increased risk by 17%, 52%, and 55% compared to II, DI + II, and II genotypes, respectively. Significant heterogeneity was observed in the overall analysis. We performed a Galbraith plot analysis to study potential heterogeneity analysis, and after excluding these studies [1, 18, 26, 32, 34, 45, 46], high heterogeneity was significant reduced. We did a comprehensive literature reviewed in these excluded studies; the mixed ethnicities, differences in geography, and patient sample size may be the reasons for the high heterogeneity. Therefore, a full subgroup analysis was conducted. In Asian populations including subjects from China, South Korea, Turkey, Iran, India, and Japan, the increased risk of PE in D allele (allelic genetic model), DD genotype (recessive genetic model), and DD genotype was 1.31-fold, 1.80-fold, and 1.53-fold, respectively. Regarding subjects from Europe (UK, Italy, Greece, and Norway), a 33% increased risk of PE in D allele (allelic genetic model) and a 68% increased risk of PE in DD genotypes (homozygote genetic model) were detected, appearing as though the Europeans had more risk of PE than did to Asians. In the subgroup analysis by ethnicity, increased risk of PE was only discovered in Caucasoid population, consistent with results of previous studies [57, 59, 61, 62]. We introduced PE degree and gestational week as subgroups to assess the potential relationships between the ACE I/D polymorphism and severe PE degree and onset time of PE. In the severe PE population, widely increased risks were observed, and we also detected a greater risk of PE than in the mild PE population. However, no significant association was detected for early-onset or late-onset of PE. For the patient sample subgroup analysis, increased risks were also observed.

There were several limitations in this meta-analysis. Firstly, language bias existed in our results; although no language limitation was set, only English and Chinese articles were included. Secondly, the sample size of included studies in the subgroup analysis of PE degree and onset time of PE were relatively small in some groups, implying that our results should be explained with caution. Finally, the potential influence of environment factors on genotype-PE associations is worthy of consideration.

Our results indicated that the AGT T704C and ACE I/D polymorphisms were associated with an increased risk of PE. Increased risks were also observed for the two polymorphisms in subgroups including Asians, Europeans, Caucasoid, and Mongoloid. Furthermore, an increased PE risk with the ACE I/D polymorphism in the severe PE population was also detected. Regarding the AT1R A1166C polymorphism, weak associations were observed and further studies are required.