ORIGINAL
PAPER
Associations of Angiotensin-Converting Enzyme Insertion/Deletion, Angiotensin II Receptor A1166C, and Endothelial Nitric Oxide Synthase 4b/a Gene Polymorphisms With Pregnancy Hypertensive Disorders: A Meta-Analysis Feng-Feng Gong, MS;1,2,* Cai-Yun Hu, MS;1,2,* Shan-Shan Lu, MS;1,2 Zhen-Zhong Qian, MS;1,2 Fang Feng, MS;1,2 Yi-Le Wu, MS;1,2 Hui-Yun Yang, MS;1,2 Ye-Huan Sun, MS, PhD1,2 From the Department of Epidemiology and Health Statistics, School of Public Health, Anhui Medical University, Hefei, Anhui, China;1 and Center for Evidence-Based Practice, Anhui Medical University, Hefei, Anhui, China2
There have been numerous studies concerning the associations of angiotensin-converting enzyme (ACE) gene insertion/ deletion (I/D), angiotensin II receptor 1 (AT1R) gene A1166C, and endothelial nitric oxide synthase (eNOS) gene 4b/a polymorphisms with risk of pregnancy hypertensive disorders (PHDs). However, the results are inconsistent. A total of 83 eligible studies (10,354/18,446 cases/controls) were included in this meta-analysis. Pooled odds ratios with corresponding 95% confidence intervals were used to calculate these associations. The effects of ethnicity and types of PHDs were
also considered. Results showed significant associations between the ACE gene polymorphism and PHDs in all of the populations except that in Africa. The associations also existed in AT1R, eNOS gene polymorphism and PHDs in part of the gene models in the overall population. These results indicated the ACE gene polymorphism was associated with an increased risk of PHDs, whereas the eNOS and AT1R gene polymorphism only have increased susceptibility to PHDs in part of the gene models. J Clin Hypertens (Greenwich). 2015;17:954–962. ª 2015 Wiley Periodicals, Inc.
Pregnancy hypertensive disorders (PHDs) are the most common obstetrical complications of pregnancy, as well as a major cause of maternal and perinatal morbidity and mortality. PHDs including gestational hypertension (GH), preeclampsia (PE), and eclampsia, as well as chronic hypertension complicated by preeclampsia, have an incidence rate of 7% to 12%.1 Despite ample researches in this field, the etiology and pathogenesis of PHDs are still unclear. In addition, it has been widely recognized that PHDs were a series of multifactorial, pregnancy-specific vascular disorders, marked by hypertension and proteinuria compared with women of formerly normal blood pressure (BP).2 Furthermore, more studies have directly shown that PHDs are diseases with complex genetic susceptibility3,4 and are thought to have a polygenic basis. In women with PHDs, increasing vascular resistance causes a failure to develop physiological hypervolemia, a mechanism leading to the pathophysiology of PHDs. In this mechanism, angiotensin-converting enzyme (ACE) plays an irreplaceable role.5 The ACE gene encodes an enzyme which could catalyze the conversion of angiotensin I to a physiologically active peptide
angiotensin (Ang) II. Ang II is an aldosterone-stimulating peptide and potent vasopressor that stabilizes BP and water balance. The enzyme converts Ang I to Ang II through the removal of the terminal His-Leu, which results in the increase of the vasoconstrictor activity of angiotensin.6,7 Furthermore, the majority of the biological effects of Ang II are mediated by Ang II receptor 1 (AT1R), which plays a key role in the physiological effects, including mediating vasoconstriction, water and salt metabolism, vascular smooth muscle proliferation and regulation of its function. In addition, a vasoactive substance called endothelial nitric oxide also plays an important role in the regulation of BP.8 Nitric oxide is the precursor of endothelial nitric oxide, which is a reactive free radical and serves as a biological mediator in some processes, including antimicrobial, neurotransmission and antitumoral activities. Endothelial nitric oxide synthase (eNOS), a kind of chemical product of endothelial nitric oxide, catalyzes the generation of L-citrulline and nitric oxide from L-arginine and molecular oxygen.9 In addition, the human gene encoding eNOS is located in chromosome 7q35-36, which contains 26 exons and 25 introns.10 eNOS is a critical mediator of cardiovascular homeostasis by regulation of the diameter of blood vessels and stabilizes an antiapoptotic and antiproliferative environment in the vasculature. A number of studies have investigated the relationships between ACE, AT1R and eNOS gene polymorphisms and PHDs. The inconsistency of these results has been attributed to inadequate statistical power caused by small sample size and geographical differences.
*These authors contributed equally to this work. Address for correspondence: Ye-Huan Sun, MS, PhD, Department of Epidemiology and Statistics, School of Public Health, Anhui Medical University, No. 81 Meishan Road, Hefei, Anhui 230032, China E-mail: [email protected] Manuscript received: February 5, 2015; revised: April 30, 2015; accepted: May 1, 2015 DOI: 10.1111/jch.12606
954
The Journal of Clinical Hypertension
Vol 17 | No 12 | December 2015
ACE, AT1R and eNOS variants and PHDs | Gong et al.
Meta-analysis, an approach that able to enhance statistical power and stability by integrating the results of individual studies, could offer a powerful means to solve these problems. Therefore, we conducted a comprehensive meta-analysis of all eligible papers regarding the association between ACE, AT1R, and eNOS gene polymorphisms and PHDs risk to calculate the strength of genetic association.
largest study was preferred); (3) did not report the genotype frequencies; and (4) if the participants enrolled in the studies had other diseases (heart disease, asthma, diabetes, kidney disease, urinary tract infection, hematological disorder, autoimmune disease). The data included in this meta-analysis were obtained from existing publications, so further written consent by the participants and approval by an ethics committee were not needed for this study.
MATERIALS AND METHODS Search Strategy The following electronic databases were searched: PubMed, the Cochrane Library, Elsevier ScienceDirect, EMBASE, SpringerLink, Google Scholar, the Chinese National Knowledge Infrastructure, the Chinese Biomedical Database, VIP (Chinese), and the Wanfang Database (Chinese). All of the studies were limited to English or Chinese language, the search was updated in May 2014. The following keyword strings were used in the search strategy: “pregnancy hypertensive disorders” or “pregnancy-induced hypertension” or “gestational hypertension” or “preeclampsia” or “eclampsia” and “angiotensin-converting enzyme” or “ACE” or “angiotensin II receptor 1” or “AT1R” or “A1166C” or “endothelial nitric oxide synthase” or “eNOS” or “4b/a.” Both strings were used in combination with “variant,” “mutation,” “polymorphism,” or “polymorphisms.” The search was conducted on human subjects. Furthermore, the references of all of the relevant articles were used to identify supplementary eligible studies. Inclusion and Exclusion Criteria Inclusion criteria were set as follows: (1) case-control studies; (2) provided complete data on genotype frequencies of gene polymorphisms for calculating odds ratios (ORs) with a 95% confidence interval (95% CI); (3) genotype frequencies were reported in both case and control groups, and methods used for genotypic were effective and individual subject was unrelated; and (4) PHD was defined as BP ≥140/90 mm Hg on at least two occasions 6 hours apart that generally appeared after 20 weeks of gestation and returned to normal 12 weeks postpartum, GH was defined as BP ≥140/90 mm Hg after 20 weeks of gestation at two occasions at least 6 hours apart and no proteinuria in earlier pregnant women, PE was defined as BP ≥140/90 mm Hg along with proteinuria ≥2+ on dipstick twice at least 6 hours apart, and eclampsia was diagnosed with tonic-clonic seizures in pregnant women with BP ≥140/90 mm Hg and proteinuria. BP was measured with the patient in a right position after a 10-minute rest with a standard mercury sphygmomanometer. The right arm was placed in a horizontal position at the heart level, which was used for the measurement. Exclusion criteria were set as follows: (1) was a review or a case report; (2) was a duplicated publication (in studies with overlapping cases and/or controls, the
Data Extraction The following characteristics were extracted from all eligible studies: the first author’s name, year of publication, country, ethnicity of the study population, types of PHDs, and sample size and distributions of genotype and allele frequencies in both case and control groups. Different ethnicities were categorized as Asian, Caucasian, and other. All information was checked and collected by two researchers independently (Gong and Hu), and the inconsistencies were examined and discussed until consensus was achieved for all data. Statistical Analysis The strength of associations between ACE, AT1R, and eNOS polymorphisms and PHDs were assessed by ORs and 95% CIs. A chi-square test was used for HardyWeinberg equilibrium, and a P value 50% indicated obvious between-study heterogeneity,11 and ORs and 95% CIs were calculated by randomeffects model,12 otherwise, the fixed-effects model was selected.13 Sensitivity analyses consisting of sequential omission of individual studies were performed to assess the stability of the results. Subgroup analyses were mainly performed by ethnicity (Caucasian, Asian, and others) and types of PHDs (PHD, GH, PE, and eclampsia). Asymmetry of the funnel plot, Begg’s and Egger’s tests, and funnel plots were used to estimate evidence for potential publication bias.14 Statistical analyses were conducted using Stata 9.0 (Stata Corp, College Station, TX).
RESULTS Eligible Studies for Meta-Analysis From 3759 articles retrieved in the primary search, 2873 articles were reviewed through title or abstract after duplicates were omitted, 2147 records did not meet the inclusion criteria, leaving 726 potentially relevant studies evaluated in full-text. Finally, 83 independent studies researching the effect of three different genetic polymorphism on PHDs were identified after the search and screening process. Among the eligible studies, a total of 70 studies of the ACE I/D polymorphism (5596/8113 cases/controls), 24 studies of the AT1R A1166C polymorphism (2039/2462 cases/ controls), and 28 studies of the eNOS 4b/a polymorThe Journal of Clinical Hypertension
Vol 17 | No 12 | December 2015
955
ACE, AT1R and eNOS variants and PHDs | Gong et al.
phism (2719/7871 cases/controls) were pooled and managed together (Table S1). Quantitative Synthesis Association Between the ACE I/D Gene Polymorphism and PHDs. As for the ACE I/D polymorphism, a significant association was observed between the ACE gene polymorphism and PHDs in all genetic models of overall populations (D vs I: OR=1.47, 95% CI=1.29– 1.68; DD vs II: OR=1.90, 95% CI=1.52–2.37; ID vs II: OR=1.21, 95% CI=1.05–1.40; ID+DD vs II: OR=1.47, 95% CI=1.25–1.74; DD vs ID+II: OR=1.69, 95% CI=1.42–2.01). The subgroup analysis by ethnicity revealed the association between that the ACE polymorphism and PHDs existed in the Asian population (D vs I: OR=1.59, 95% CI=1.30–1.95; DD vs II: OR=2.15, 95% CI=1.55–2.99; ID vs II: OR=1.28, 95% CI=1.07– 1.55; ID+DD vs II: OR=1.60, 95% CI=1.28–2.01; DD vs ID+II: OR=1.86, 95% CI=1.40–2.47) and the Caucasian population (D vs I: OR=1.31, 95% CI=1.14– 1.50; DD vs II: OR=1.50, 95% CI=1.17–1.93; ID+DD vs II: OR=1.23, 95% CI=0.97–1.55; DD vs ID+II: OR=1.51, 95% CI=1.26–1.80; except ID vs II: OR=1.06, 95% CI=0.83–1.36), but not the African population (D vs I: OR=1.13, 95% CI=0.94–1.35; DD vs II: OR=1.37, 95% CI=0.91–2.05; ID vs II: OR=1.36, 95% CI=0.85–2.16; ID+DD vs II: OR=1.35, 95% CI=0.90–2.02; DD vs ID+II: OR=1.09, 95% CI=0.86– 1.37). In addition, subgroup analyses by types of PHDs revealed that the association between the ACE polymorphism and PHDs existed in PE (all genetic models except ID vs II) and PHDs (all genetic models). However, no association was detected in GH (the overall group and subgroup) and eclampsia (overall group and subgroup), despite the different strengths. Association Between the AT1R A1166C Gene Polymorphism and PHDs. The relationship between the AT1R A1166C polymorphism and PHDs was observed in three genetic models of overall populations (C vs A: OR=1.68, 95% CI=1.30–2.17; AC vs AA: OR=1.82, 95% CI=1.38–2.40; AC+CC vs AA: OR=1.81, 95% CI=1.37–2.40; except CC vs AA: OR=1.17, 95% CI=0.81–1.69; CC vs AC+AA: OR=1.13, 95% CI=0.79–1.63). The subgroup analysis by ethnicity revealed that the association between the AT1R polymorphism and PHDs existed in the Asian population (for C vs A: OR=2.11, 95% CI=1.55–2.87; for CC vs AA: OR=1.86, 95% CI=1.06–3.26; for AC vs AA: OR=2.22, 95% CI=1.57–3.15; for AC+CC vs AA: OR=2.26, 95% CI=1.60–3.19; for CC vs AC+AA: OR=1.79, 95% CI=1.02–3.12), but not in the Caucasian population (for C vs A: OR=1.00, 95% CI=0.79–1.25; for CC vs AA: OR=0.80, 95% CI=0.48– 1.32; for AC vs AA: OR=1.11, 95% CI=0.84–1.45; for AC+CC vs AA: OR=1.05, 95% CI=0.81–1.35; for CC vs AC+AA: OR=0.79, 95% CI=0.48–1.29). 956
The Journal of Clinical Hypertension
Vol 17 | No 12 | December 2015
Subgroup analyses by types of PHDs revealed that the association between the polymorphism and PHDs existed in PHDs (all genetic models in the subgroup) and PE (for C vs A, for CC vs AA and for AC+CC vs AA). Nevertheless, no association could be detected in any of the remaining gene models. Association Between the eNOS 4b/a Gene Polymorphism and PHDs. An association was detected between the eNOS 4b/a polymorphism and the susceptibility to PHDs in two genetic models of overall populations (aa vs bb: OR=1.47, 95% CI=1.12–1.93; aa vs ba+bb: OR=1.52, 95% CI=1.16–1.99; except a vs b: OR=1.02, 95% CI=0.92–1.12; ba+aa vs bb: OR=0.96, 95% CI=0.85–1.07). The subgroup analysis by ethnicity revealed that the association between the ACE polymorphism and PHDs existed in the Caucasian population in two genetic models (aa vs bb: OR=1.49, 95% CI=1.11–2.00; aa vs ba+bb: OR=1.56, 95% CI=1.16– 2.09; except a vs b: OR=0.99, 95% CI=0.89–1.10; ba+aa vs bb: OR=0.92, 95% CI=0.81–1.04), but not in the Asian population (for a vs b: OR=1.20, 95% CI=0.93–1.55; for aa vs bb: OR=1.35, 95% CI=0.65– 2.84; for ba+aa vs bb: OR=1.23, 95% CI=0.91–1.65; for aa vs ba+bb: OR=1.30, 95% CI=0.62–2.71). In addition, subgroup analyses by types of PHDs revealed that the association between the eNOS polymorphism and PHDs existed in PE (two genetic models: for aa vs bb and for aa vs ba+bb). However, we were unable to detect an association in all of the other genetic models (Figures 1A–C and S1). Publication Bias Diagnostics Funnel plot, Begg’s test and Egger’s test were performed to assess the publication bias of these studies. The funnel plot was symmetrical, indicating that there was no evidence of obvious publication bias in this sample. However, the results of Begg’s and Egger’s tests showed publication bias observed in a few models, which could be attributed to very small sample sizes or geographical differences. The majority of P values were ≥.05 (Table). Funnel plots are shown in Figure 2. Sensitivity Analysis and Test of Heterogeneity To estimate the sensitivity of this meta-analysis, a leaveone-out sensitivity analysis was performed. Results were not substantially altered when any single study was deleted, showing that the results of this meta-analysis were generally robust. Since some heterogeneity was found across the studies, we conducted a meta-regression analysis of subgroups based on ethnicity, types of PHDs, and sample size to explore the source of heterogeneity. As a result, none of these factors obviously contributed to the heterogeneity. However, the effects of factors such as gene-environmental interactions, lifestyle, or subject feature on the association between polymorphisms and PHDs and their contribution to heterogeneity could not be obtained for they were not available in the original data. More original
ACE, AT1R and eNOS variants and PHDs | Gong et al.
(A)
FIGURE 1. (A) Forest plots for the association between angiotensin-converting enzyme insertion/deletion gene polymorphisms and pregnancy hypertensive disorders (PHDs) of genotype models: for D vs I. (B) Forest plots for the association between angiotensin II receptor 1 A1166C gene polymorphisms and PHDs of genotype models: for C vs A. (C) Forest plots for the association between endothelial nitric oxide synthase 4b/a gene polymorphisms and PHDs of genotype models: for A vs B. OR indicates odds ratio; CI, confidence interval.
The Journal of Clinical Hypertension
Vol 17 | No 12 | December 2015
957
ACE, AT1R and eNOS variants and PHDs | Gong et al.
(B)
(C)
FIGURE 1. Continued. 958
The Journal of Clinical Hypertension
Vol 17 | No 12 | December 2015
ACE, AT1R and eNOS variants and PHDs | Gong et al.
TABLE. Summary of Studies Included in the Meta-Analysis Genotype ACE
Studies All studies
Asian
Caucasian
African
AT1R
All studies
Asian
Caucasian
eNOS
All studies
Asian
Caucasian
Comparison Model
Studies (Cases/Controls)
OR (95% CI)
POR
M*
I2, %
Pbegg
Pegger
D vs I
70 (5596/8113)
1.47 (1.29–1.68)
PAPER
Associations of Angiotensin-Converting Enzyme Insertion/Deletion, Angiotensin II Receptor A1166C, and Endothelial Nitric Oxide Synthase 4b/a Gene Polymorphisms With Pregnancy Hypertensive Disorders: A Meta-Analysis Feng-Feng Gong, MS;1,2,* Cai-Yun Hu, MS;1,2,* Shan-Shan Lu, MS;1,2 Zhen-Zhong Qian, MS;1,2 Fang Feng, MS;1,2 Yi-Le Wu, MS;1,2 Hui-Yun Yang, MS;1,2 Ye-Huan Sun, MS, PhD1,2 From the Department of Epidemiology and Health Statistics, School of Public Health, Anhui Medical University, Hefei, Anhui, China;1 and Center for Evidence-Based Practice, Anhui Medical University, Hefei, Anhui, China2
There have been numerous studies concerning the associations of angiotensin-converting enzyme (ACE) gene insertion/ deletion (I/D), angiotensin II receptor 1 (AT1R) gene A1166C, and endothelial nitric oxide synthase (eNOS) gene 4b/a polymorphisms with risk of pregnancy hypertensive disorders (PHDs). However, the results are inconsistent. A total of 83 eligible studies (10,354/18,446 cases/controls) were included in this meta-analysis. Pooled odds ratios with corresponding 95% confidence intervals were used to calculate these associations. The effects of ethnicity and types of PHDs were
also considered. Results showed significant associations between the ACE gene polymorphism and PHDs in all of the populations except that in Africa. The associations also existed in AT1R, eNOS gene polymorphism and PHDs in part of the gene models in the overall population. These results indicated the ACE gene polymorphism was associated with an increased risk of PHDs, whereas the eNOS and AT1R gene polymorphism only have increased susceptibility to PHDs in part of the gene models. J Clin Hypertens (Greenwich). 2015;17:954–962. ª 2015 Wiley Periodicals, Inc.
Pregnancy hypertensive disorders (PHDs) are the most common obstetrical complications of pregnancy, as well as a major cause of maternal and perinatal morbidity and mortality. PHDs including gestational hypertension (GH), preeclampsia (PE), and eclampsia, as well as chronic hypertension complicated by preeclampsia, have an incidence rate of 7% to 12%.1 Despite ample researches in this field, the etiology and pathogenesis of PHDs are still unclear. In addition, it has been widely recognized that PHDs were a series of multifactorial, pregnancy-specific vascular disorders, marked by hypertension and proteinuria compared with women of formerly normal blood pressure (BP).2 Furthermore, more studies have directly shown that PHDs are diseases with complex genetic susceptibility3,4 and are thought to have a polygenic basis. In women with PHDs, increasing vascular resistance causes a failure to develop physiological hypervolemia, a mechanism leading to the pathophysiology of PHDs. In this mechanism, angiotensin-converting enzyme (ACE) plays an irreplaceable role.5 The ACE gene encodes an enzyme which could catalyze the conversion of angiotensin I to a physiologically active peptide
angiotensin (Ang) II. Ang II is an aldosterone-stimulating peptide and potent vasopressor that stabilizes BP and water balance. The enzyme converts Ang I to Ang II through the removal of the terminal His-Leu, which results in the increase of the vasoconstrictor activity of angiotensin.6,7 Furthermore, the majority of the biological effects of Ang II are mediated by Ang II receptor 1 (AT1R), which plays a key role in the physiological effects, including mediating vasoconstriction, water and salt metabolism, vascular smooth muscle proliferation and regulation of its function. In addition, a vasoactive substance called endothelial nitric oxide also plays an important role in the regulation of BP.8 Nitric oxide is the precursor of endothelial nitric oxide, which is a reactive free radical and serves as a biological mediator in some processes, including antimicrobial, neurotransmission and antitumoral activities. Endothelial nitric oxide synthase (eNOS), a kind of chemical product of endothelial nitric oxide, catalyzes the generation of L-citrulline and nitric oxide from L-arginine and molecular oxygen.9 In addition, the human gene encoding eNOS is located in chromosome 7q35-36, which contains 26 exons and 25 introns.10 eNOS is a critical mediator of cardiovascular homeostasis by regulation of the diameter of blood vessels and stabilizes an antiapoptotic and antiproliferative environment in the vasculature. A number of studies have investigated the relationships between ACE, AT1R and eNOS gene polymorphisms and PHDs. The inconsistency of these results has been attributed to inadequate statistical power caused by small sample size and geographical differences.
*These authors contributed equally to this work. Address for correspondence: Ye-Huan Sun, MS, PhD, Department of Epidemiology and Statistics, School of Public Health, Anhui Medical University, No. 81 Meishan Road, Hefei, Anhui 230032, China E-mail: [email protected] Manuscript received: February 5, 2015; revised: April 30, 2015; accepted: May 1, 2015 DOI: 10.1111/jch.12606
954
The Journal of Clinical Hypertension
Vol 17 | No 12 | December 2015
ACE, AT1R and eNOS variants and PHDs | Gong et al.
Meta-analysis, an approach that able to enhance statistical power and stability by integrating the results of individual studies, could offer a powerful means to solve these problems. Therefore, we conducted a comprehensive meta-analysis of all eligible papers regarding the association between ACE, AT1R, and eNOS gene polymorphisms and PHDs risk to calculate the strength of genetic association.
largest study was preferred); (3) did not report the genotype frequencies; and (4) if the participants enrolled in the studies had other diseases (heart disease, asthma, diabetes, kidney disease, urinary tract infection, hematological disorder, autoimmune disease). The data included in this meta-analysis were obtained from existing publications, so further written consent by the participants and approval by an ethics committee were not needed for this study.
MATERIALS AND METHODS Search Strategy The following electronic databases were searched: PubMed, the Cochrane Library, Elsevier ScienceDirect, EMBASE, SpringerLink, Google Scholar, the Chinese National Knowledge Infrastructure, the Chinese Biomedical Database, VIP (Chinese), and the Wanfang Database (Chinese). All of the studies were limited to English or Chinese language, the search was updated in May 2014. The following keyword strings were used in the search strategy: “pregnancy hypertensive disorders” or “pregnancy-induced hypertension” or “gestational hypertension” or “preeclampsia” or “eclampsia” and “angiotensin-converting enzyme” or “ACE” or “angiotensin II receptor 1” or “AT1R” or “A1166C” or “endothelial nitric oxide synthase” or “eNOS” or “4b/a.” Both strings were used in combination with “variant,” “mutation,” “polymorphism,” or “polymorphisms.” The search was conducted on human subjects. Furthermore, the references of all of the relevant articles were used to identify supplementary eligible studies. Inclusion and Exclusion Criteria Inclusion criteria were set as follows: (1) case-control studies; (2) provided complete data on genotype frequencies of gene polymorphisms for calculating odds ratios (ORs) with a 95% confidence interval (95% CI); (3) genotype frequencies were reported in both case and control groups, and methods used for genotypic were effective and individual subject was unrelated; and (4) PHD was defined as BP ≥140/90 mm Hg on at least two occasions 6 hours apart that generally appeared after 20 weeks of gestation and returned to normal 12 weeks postpartum, GH was defined as BP ≥140/90 mm Hg after 20 weeks of gestation at two occasions at least 6 hours apart and no proteinuria in earlier pregnant women, PE was defined as BP ≥140/90 mm Hg along with proteinuria ≥2+ on dipstick twice at least 6 hours apart, and eclampsia was diagnosed with tonic-clonic seizures in pregnant women with BP ≥140/90 mm Hg and proteinuria. BP was measured with the patient in a right position after a 10-minute rest with a standard mercury sphygmomanometer. The right arm was placed in a horizontal position at the heart level, which was used for the measurement. Exclusion criteria were set as follows: (1) was a review or a case report; (2) was a duplicated publication (in studies with overlapping cases and/or controls, the
Data Extraction The following characteristics were extracted from all eligible studies: the first author’s name, year of publication, country, ethnicity of the study population, types of PHDs, and sample size and distributions of genotype and allele frequencies in both case and control groups. Different ethnicities were categorized as Asian, Caucasian, and other. All information was checked and collected by two researchers independently (Gong and Hu), and the inconsistencies were examined and discussed until consensus was achieved for all data. Statistical Analysis The strength of associations between ACE, AT1R, and eNOS polymorphisms and PHDs were assessed by ORs and 95% CIs. A chi-square test was used for HardyWeinberg equilibrium, and a P value 50% indicated obvious between-study heterogeneity,11 and ORs and 95% CIs were calculated by randomeffects model,12 otherwise, the fixed-effects model was selected.13 Sensitivity analyses consisting of sequential omission of individual studies were performed to assess the stability of the results. Subgroup analyses were mainly performed by ethnicity (Caucasian, Asian, and others) and types of PHDs (PHD, GH, PE, and eclampsia). Asymmetry of the funnel plot, Begg’s and Egger’s tests, and funnel plots were used to estimate evidence for potential publication bias.14 Statistical analyses were conducted using Stata 9.0 (Stata Corp, College Station, TX).
RESULTS Eligible Studies for Meta-Analysis From 3759 articles retrieved in the primary search, 2873 articles were reviewed through title or abstract after duplicates were omitted, 2147 records did not meet the inclusion criteria, leaving 726 potentially relevant studies evaluated in full-text. Finally, 83 independent studies researching the effect of three different genetic polymorphism on PHDs were identified after the search and screening process. Among the eligible studies, a total of 70 studies of the ACE I/D polymorphism (5596/8113 cases/controls), 24 studies of the AT1R A1166C polymorphism (2039/2462 cases/ controls), and 28 studies of the eNOS 4b/a polymorThe Journal of Clinical Hypertension
Vol 17 | No 12 | December 2015
955
ACE, AT1R and eNOS variants and PHDs | Gong et al.
phism (2719/7871 cases/controls) were pooled and managed together (Table S1). Quantitative Synthesis Association Between the ACE I/D Gene Polymorphism and PHDs. As for the ACE I/D polymorphism, a significant association was observed between the ACE gene polymorphism and PHDs in all genetic models of overall populations (D vs I: OR=1.47, 95% CI=1.29– 1.68; DD vs II: OR=1.90, 95% CI=1.52–2.37; ID vs II: OR=1.21, 95% CI=1.05–1.40; ID+DD vs II: OR=1.47, 95% CI=1.25–1.74; DD vs ID+II: OR=1.69, 95% CI=1.42–2.01). The subgroup analysis by ethnicity revealed the association between that the ACE polymorphism and PHDs existed in the Asian population (D vs I: OR=1.59, 95% CI=1.30–1.95; DD vs II: OR=2.15, 95% CI=1.55–2.99; ID vs II: OR=1.28, 95% CI=1.07– 1.55; ID+DD vs II: OR=1.60, 95% CI=1.28–2.01; DD vs ID+II: OR=1.86, 95% CI=1.40–2.47) and the Caucasian population (D vs I: OR=1.31, 95% CI=1.14– 1.50; DD vs II: OR=1.50, 95% CI=1.17–1.93; ID+DD vs II: OR=1.23, 95% CI=0.97–1.55; DD vs ID+II: OR=1.51, 95% CI=1.26–1.80; except ID vs II: OR=1.06, 95% CI=0.83–1.36), but not the African population (D vs I: OR=1.13, 95% CI=0.94–1.35; DD vs II: OR=1.37, 95% CI=0.91–2.05; ID vs II: OR=1.36, 95% CI=0.85–2.16; ID+DD vs II: OR=1.35, 95% CI=0.90–2.02; DD vs ID+II: OR=1.09, 95% CI=0.86– 1.37). In addition, subgroup analyses by types of PHDs revealed that the association between the ACE polymorphism and PHDs existed in PE (all genetic models except ID vs II) and PHDs (all genetic models). However, no association was detected in GH (the overall group and subgroup) and eclampsia (overall group and subgroup), despite the different strengths. Association Between the AT1R A1166C Gene Polymorphism and PHDs. The relationship between the AT1R A1166C polymorphism and PHDs was observed in three genetic models of overall populations (C vs A: OR=1.68, 95% CI=1.30–2.17; AC vs AA: OR=1.82, 95% CI=1.38–2.40; AC+CC vs AA: OR=1.81, 95% CI=1.37–2.40; except CC vs AA: OR=1.17, 95% CI=0.81–1.69; CC vs AC+AA: OR=1.13, 95% CI=0.79–1.63). The subgroup analysis by ethnicity revealed that the association between the AT1R polymorphism and PHDs existed in the Asian population (for C vs A: OR=2.11, 95% CI=1.55–2.87; for CC vs AA: OR=1.86, 95% CI=1.06–3.26; for AC vs AA: OR=2.22, 95% CI=1.57–3.15; for AC+CC vs AA: OR=2.26, 95% CI=1.60–3.19; for CC vs AC+AA: OR=1.79, 95% CI=1.02–3.12), but not in the Caucasian population (for C vs A: OR=1.00, 95% CI=0.79–1.25; for CC vs AA: OR=0.80, 95% CI=0.48– 1.32; for AC vs AA: OR=1.11, 95% CI=0.84–1.45; for AC+CC vs AA: OR=1.05, 95% CI=0.81–1.35; for CC vs AC+AA: OR=0.79, 95% CI=0.48–1.29). 956
The Journal of Clinical Hypertension
Vol 17 | No 12 | December 2015
Subgroup analyses by types of PHDs revealed that the association between the polymorphism and PHDs existed in PHDs (all genetic models in the subgroup) and PE (for C vs A, for CC vs AA and for AC+CC vs AA). Nevertheless, no association could be detected in any of the remaining gene models. Association Between the eNOS 4b/a Gene Polymorphism and PHDs. An association was detected between the eNOS 4b/a polymorphism and the susceptibility to PHDs in two genetic models of overall populations (aa vs bb: OR=1.47, 95% CI=1.12–1.93; aa vs ba+bb: OR=1.52, 95% CI=1.16–1.99; except a vs b: OR=1.02, 95% CI=0.92–1.12; ba+aa vs bb: OR=0.96, 95% CI=0.85–1.07). The subgroup analysis by ethnicity revealed that the association between the ACE polymorphism and PHDs existed in the Caucasian population in two genetic models (aa vs bb: OR=1.49, 95% CI=1.11–2.00; aa vs ba+bb: OR=1.56, 95% CI=1.16– 2.09; except a vs b: OR=0.99, 95% CI=0.89–1.10; ba+aa vs bb: OR=0.92, 95% CI=0.81–1.04), but not in the Asian population (for a vs b: OR=1.20, 95% CI=0.93–1.55; for aa vs bb: OR=1.35, 95% CI=0.65– 2.84; for ba+aa vs bb: OR=1.23, 95% CI=0.91–1.65; for aa vs ba+bb: OR=1.30, 95% CI=0.62–2.71). In addition, subgroup analyses by types of PHDs revealed that the association between the eNOS polymorphism and PHDs existed in PE (two genetic models: for aa vs bb and for aa vs ba+bb). However, we were unable to detect an association in all of the other genetic models (Figures 1A–C and S1). Publication Bias Diagnostics Funnel plot, Begg’s test and Egger’s test were performed to assess the publication bias of these studies. The funnel plot was symmetrical, indicating that there was no evidence of obvious publication bias in this sample. However, the results of Begg’s and Egger’s tests showed publication bias observed in a few models, which could be attributed to very small sample sizes or geographical differences. The majority of P values were ≥.05 (Table). Funnel plots are shown in Figure 2. Sensitivity Analysis and Test of Heterogeneity To estimate the sensitivity of this meta-analysis, a leaveone-out sensitivity analysis was performed. Results were not substantially altered when any single study was deleted, showing that the results of this meta-analysis were generally robust. Since some heterogeneity was found across the studies, we conducted a meta-regression analysis of subgroups based on ethnicity, types of PHDs, and sample size to explore the source of heterogeneity. As a result, none of these factors obviously contributed to the heterogeneity. However, the effects of factors such as gene-environmental interactions, lifestyle, or subject feature on the association between polymorphisms and PHDs and their contribution to heterogeneity could not be obtained for they were not available in the original data. More original
ACE, AT1R and eNOS variants and PHDs | Gong et al.
(A)
FIGURE 1. (A) Forest plots for the association between angiotensin-converting enzyme insertion/deletion gene polymorphisms and pregnancy hypertensive disorders (PHDs) of genotype models: for D vs I. (B) Forest plots for the association between angiotensin II receptor 1 A1166C gene polymorphisms and PHDs of genotype models: for C vs A. (C) Forest plots for the association between endothelial nitric oxide synthase 4b/a gene polymorphisms and PHDs of genotype models: for A vs B. OR indicates odds ratio; CI, confidence interval.
The Journal of Clinical Hypertension
Vol 17 | No 12 | December 2015
957
ACE, AT1R and eNOS variants and PHDs | Gong et al.
(B)
(C)
FIGURE 1. Continued. 958
The Journal of Clinical Hypertension
Vol 17 | No 12 | December 2015
ACE, AT1R and eNOS variants and PHDs | Gong et al.
TABLE. Summary of Studies Included in the Meta-Analysis Genotype ACE
Studies All studies
Asian
Caucasian
African
AT1R
All studies
Asian
Caucasian
eNOS
All studies
Asian
Caucasian
Comparison Model
Studies (Cases/Controls)
OR (95% CI)
POR
M*
I2, %
Pbegg
Pegger
D vs I
70 (5596/8113)
1.47 (1.29–1.68)
