Associations of Estrogen Receptor Alpha and Beta Gene Polymorphisms with Lipid Levels and Insulin Resistance in Men Zoe A. Efstathiadou, Christina Sakka, Stergios A. Polyzos, Maria Goutou, Nikolaos Stakias, Alexandra Bargiota, George N. Koukoulis PII: DOI: Reference:
S0026-0495(15)00021-9 doi: 10.1016/j.metabol.2015.01.006 YMETA 53160
To appear in:
Metabolism
Received date: Revised date: Accepted date:
5 September 2014 29 December 2014 16 January 2015
Please cite this article as: Efstathiadou Zoe A., Sakka Christina, Polyzos Stergios A., Goutou Maria, Stakias Nikolaos, Bargiota Alexandra, Koukoulis George N., Associations of Estrogen Receptor Alpha and Beta Gene Polymorphisms with Lipid Levels and Insulin Resistance in Men, Metabolism (2015), doi: 10.1016/j.metabol.2015.01.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Associations of Estrogen Receptor Alpha and Beta Gene Polymorphisms
T
with Lipid Levels and Insulin Resistance in Men
Stakias2, Alexandra Bargiota2 and George N. Koukoulis2 1
RI P
Zoe A. Efstathiadou1*, Christina Sakka2*, Stergios A. Polyzos1, Maria Goutou2, Nikolaos
SC
Endocrine Clinic, “Hippokration” General Hospital of Thessaloniki, Thessaloniki, Greece
2
MA NU
Department of Endocrinology, University of Thessaly School of Medicine, Larissa, Greece
Short title: ESR1 and ESR2 genes association with lipids and IR in men
Key words: estrogen receptors, genetic polymorphisms, serum lipids, insulin resistance,
ED
HOMA
AC
CE
Correspondence:
PT
Word count: Full article: 3670, Abstract: 248, Tables: 5, Figures: 2
Dr. Zoe Efstathiadou Department of Endocrinology
“Hippokration” General Hospital of Thessaloniki Konstantinoupoleos 49 54642 Thessaloniki, Greece Tel.: (+30) 6978481331 Fax.: (+30) 2310532372 e-mail: [email protected]
*The authors Z.A. Efstathiadou and C. Sakka have contributed equally to the present work
ACCEPTED MANUSCRIPT Abstract Objective: The association of four single nucleotide polymorphisms in estrogen receptor
T
alpha (ESR1) and beta (ESR2) genes with lipid levels and insulin resistance in men.
RI P
Design and Methods: Lipids, glucose, insulin and HOMA-IR were determined, in a population-based, cross-sectional, cohort of 170 apparently healthy middle-aged Greek men,
SC
along with body mass index (BMI), waist circumference (WC) and percentage of body fat content (%fat). Genotyping of ESR1 for PvuII and XbaI and ESR2 for RsaI and AluI
MA NU
polymorphisms was performed.
Results: Associations of AluI with LDL-Chol (mean±SD, aa 4.3±1.1 vs. Aa 3.7±1.0 and ΑΑ 4.2±1.1, p=0.023) and RsaI with HOMA-IR [median(IQR) RR 1.55(0.88-2.49) vs. Rr/rr 1.69(0.72-2.29), p=0.032] were found. Synergistic effects of RsaI and AluI of ESR2 gene on LDL-Chol levels, %fat and WC, as well as a synergistic effect of both ESR1 and ESR2 genes on
ED
levels of T-Chol (p=0.01) and LDL-Chol (p=0.027) were also shown. These findings remained significant after adjustment for potential confounders.
PT
Significant independent associations of PvuII with %fat (mean±SD, pp 24.6±5.3 vs Pp 22.4±5.2 and PP 21.2±6.7, p=0.044), and RsaI with % fat (RR 22.6 5.5 vs. Rr/rr 25.2 6.3,
CE
p=0.015) and WC (mean±SD RR 97.4 10.4 vs. Rr/rr 102.6 12.6, p=0.013) were found. Synergistic effects on %fat, between the ESR1 polymorphisms (p=0.004), between the ESR2
AC
polymorphisms and among all four ESR polymorphisms studied were also present.
Conclusions: ESR2 is associated with LDL-Chol levels and HOMA-IR in men independently of confounders. Body fat is affected by both genes. Furthermore, a synergistic effect of ESR1 and ESR2 on T-Chol, LDL-Chol and %fat, was shown.
ACCEPTED MANUSCRIPT 1. Introduction There is evidence that genetic variation in estrogen receptor alpha (ESR1) and beta (ESR2) genes is associated with increased cardiovascular risk [1,2], premature ischemic heart
T
disease [3], and unfavorable angiographic changes [4] in both sexes. However, there are
RI P
data contrasting these observations [5].
Through their receptors, estrogens can modulate cardiovascular risk indirectly via
SC
alterations in lipid and glucose metabolism, and directly via their vascular and endothelial effects. Functionally, competent estrogen receptors have been identified in vascular smooth muscle cells, and specific binding sites have been demonstrated in endothelium [6].
MA NU
Furthermore, apart from the classical genomic effect involving gene expression, nontranscriptional mechanisms for estrogen receptor signaling have also been described [7, 8]. On the other hand, genetic association studies have previously shown that ESR1 [911] and ESR2 [12] gene polymorphisms are associated with lipid levels and interact with hormone replacement therapy [12]. An effect on insulin sensitivity and incidence of type 2
ED
diabetes has also been described [13, 14]. However, the effect has not been consistent across different ethnic groups [15, 16], across ages [16] and, finally, it was more prominent
PT
in females than in males [16, 17].
Based on the facts that ESR2 gene is not included in most association studies, data
CE
on males are scarce and inconclusive and there exists significant ethnicity-related variability, we tested the hypothesis that polymorphisms in both genes encoding estrogen receptors
AC
(ESR1 and ESR2) might have an impact on somatometric parameters, serum lipid levels and insulin resistance in Caucasian apparently healthy males. 2. Materials and methods 2.1. Sample selection A total of 170 apparently healthy men, Caucasians, ethnic Greeks, aged 22 to 59 years (mean age ±SD, 42±8.9), were included in the study. Subjects with diagnosed diabetes mellitus, arterial hypertension, cardiovascular disease, renal or hepatic disease, androgen deficiency and those using any medication were excluded from the study. The men enrolled in the study were randomly selected from a larger cluster random sample that was drawn from the adult population of the region of Thessaly, Central Greece (18). The original sample was composed of adults, aged 18 to 79 years, who had been residents of the region of Thessaly for at least one year. Briefly, the total population of
ACCEPTED MANUSCRIPT adults was divided into geographical clusters corresponding to the four provinces of the region. The sample of this study was subsequently selected randomly from a) the capital of each province and b) randomly selected smaller towns and villages. The sample size was
T
proportional to the population of the capital and of each village and town according to the
RI P
2001 census conducted by the National Statistical Service. Only one adult was selected randomly from each household. An interview in a public health office was arranged for each participant, who was instructed to report in the following morning at 9.00 a.m. after an
SC
overnight fast. 2.2. Study protocol
MA NU
Each participant completed a questionnaire inquiring about health status and health behaviors (smoking and alcohol intake) under a doctor’s supervision. Subsequently, each man underwent a thorough physical examination. Anthropometric measurements, such as height, weight and waist circumference (WC), systolic (SBP) and diastolic blood pressure (DBP) were measured in all study participants. Waist circumference was measured using a
ED
plastic non-stretchable tailors measuring tape, at a level midway between the lower rib margin and iliac crest to the nearest centimeter. The body mass index (BMI) was calculated
height in meters.
PT
according to Quetelet’s formula as body weight in kilograms divided by the square of the
CE
The body fat content (% fat mass) was assessed by means of bioelectrical impedance analysis, using a commercially bio-impedance analyzer (Omron HBF 302 Body Fat Analyzer,
AC
Omron Healthcare Inc., Vernon Hills, IL, USA) with the patient in standing position. Finally, plasma glucose and serum insulin, total cholesterol, HDL-cholesterol and triglycerides were measured in early morning samples drawn at the fasting state. LDLcholesterol was estimated with Friedwald method (for triglyceride values not exceeding 400mg/dl). Homeostasis Model Assessment was used for the calculation of HOMA-IR with the formula: fasting insulin (μU/ml)*fasting glucose (mmol/L)/22.5 [19]. All participants gave written informed consent and the study protocol was reviewed and approved by the Ethical Committee of the University Hospital. 2.3. Assays All plasma and serum parameters –except insulin- were measured with standard commercially available assays in an Olympus AU600 analyzer (Olympus Optical, Tokyo, Japan). Serum insulin levels were measured using IRMA (Immunotech Beckman Counter
ACCEPTED MANUSCRIPT Company, Marseille, France, PW). The assay sensitivity for serum insulin measurements was 0.5 IU⁄mL and the interassay coefficient of variation 3.4%.
T
2.4. Genotyping
RI P
All study participants were genotyped for the PvuII (rs2234693) and XbaI (rs9340799) polymorphisms of the ESR1 gene and for the AluI (rs4986938) and RsaI
SC
(rs1256049) polymorphisms of the ESR2 gene. Genomic DNA was extracted from peripheral leucocytes by standard procedures.
For ESR1 gene, 1.3kb DNA fragment (which contained intron 1 and a part of exon 2)
MA NU
was amplified using the following primers: (5’-CTGCCACCCTATCTGTATCTTTTCCTATTCTCC-3’ and 5’-TCTTTCTCTGCCACCCTGGGGTCGATTATCTGA-3’).
Polymerase chain reaction (PCR) amplification was performed for 35 cycles at an annealing temperature of 63C. Amplified DNA was digested with PvuII or XbaI (New England Biolabs Inc Ipswich, MA) at 37C for 3 hours and then electrophoresed in a 1.5%
ED
ethidium bromide agarose gel. PvuII digestion reveals genotypes denoted PP(1.3kb), Pp(1.3kb, 850bp, 450bp) and pp(850bp, 450bp), and XbaI digestion results in genotypes
PT
XX(1.3kb), Xx(1.3kb, 910bp, 390bp) and xx(910bp, 390bp). The restriction fragment length polymorphisms (RLFPs) were presented by P and p (PvuII), and X and x (XbaI), with capital
CE
letters denoting the absence and small letters the presence of restriction sites, respectively. For ESR2 gene, the ligand binding the domain of exon 5 and the 3’ untranslated
AC
region of exon 8 were amplified using the following primers: A, 5’-TCTTGCTTTCCCCAGGCTTT3’; B, 5’-ACCTGTCCAGAACAAGATCT-3’; C, 5’-TTTTTGTCCCCATAGTAACA-3’; and D, 5’AATGAGGGACCACAGCA-3’. PCR amplification was performed for 40 cycles at an annealing temperature of 52C.
Αmplification of exons 5 and 8 yielded the desired products of 156bp and 307bp, respectively. Amplified DNA was digested with RsaI and AluI (New England Biolabs Inc, Ipswich, MA) at 37C for 12 hours and then electrophoresed in a 2.0% ethidium bromide agarose gel. RsaI digestion reveals genotypes denoted RR(156bp), Rr(156bp, 125bp, 31bp) and rr(125bp, 31bp), and AluI digestion results in Aa(307bp), Aa(307bp, 240bp, 67bp) and aa (240bp, 67bp). The RLFPs were presented by R and r (RsaI), and A and a (AluI), with capital letters denoting the absence and small ones the presence of restriction sites, respectively.
ACCEPTED MANUSCRIPT 2.5. Statistical analysis Continuous variables are presented as mean ± standard deviation (SD) or median (interquartile range; IQR) in cases of skewed distributions. Categorical variables are
T
presented as frequencies and/or percentages. Hardy-Weinberg equilibrium was tested with
RI P
the chi-square test for each considered polymorphism. The chi-square test was also used to check for linkage disequilibrium between the two polymorphisms of ESR1 and ESR2 genes. Differences in continuous variables between different genotypes were performed
SC
with one-way analysis of variance for each polymorphism separately. In case of significant differences between genotypes within a certain polymorphism the least significance
MA NU
difference test was used for post-hoc analysis (multiple comparisons). Multivariate linear regression was subsequently performed to evaluate the potential independence of different genotype effects. In particular, comparisons of lipid levels and HOMA-IR across genotypes [three groups; for RsaI only two groups were formed: Rr and rr were studied as one group due to small number of men carrying the rr genotype (n=2)] were
ED
adjusted for age, smoking, alcohol consumption, blood pressure, waist circumference, % body fat and HOMA-IR (for the study of lipids). The possibility of synergism between the two polymorphisms of ESR1 gene or ESR2 gene was also tested. Finally, the synergistic effect of
PT
all four polymorphisms was studied. Only genotypic groups consisting of 5 or more subjects were included in the analyses.
CE
P values
S0026-0495(15)00021-9 doi: 10.1016/j.metabol.2015.01.006 YMETA 53160
To appear in:
Metabolism
Received date: Revised date: Accepted date:
5 September 2014 29 December 2014 16 January 2015
Please cite this article as: Efstathiadou Zoe A., Sakka Christina, Polyzos Stergios A., Goutou Maria, Stakias Nikolaos, Bargiota Alexandra, Koukoulis George N., Associations of Estrogen Receptor Alpha and Beta Gene Polymorphisms with Lipid Levels and Insulin Resistance in Men, Metabolism (2015), doi: 10.1016/j.metabol.2015.01.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Associations of Estrogen Receptor Alpha and Beta Gene Polymorphisms
T
with Lipid Levels and Insulin Resistance in Men
Stakias2, Alexandra Bargiota2 and George N. Koukoulis2 1
RI P
Zoe A. Efstathiadou1*, Christina Sakka2*, Stergios A. Polyzos1, Maria Goutou2, Nikolaos
SC
Endocrine Clinic, “Hippokration” General Hospital of Thessaloniki, Thessaloniki, Greece
2
MA NU
Department of Endocrinology, University of Thessaly School of Medicine, Larissa, Greece
Short title: ESR1 and ESR2 genes association with lipids and IR in men
Key words: estrogen receptors, genetic polymorphisms, serum lipids, insulin resistance,
ED
HOMA
AC
CE
Correspondence:
PT
Word count: Full article: 3670, Abstract: 248, Tables: 5, Figures: 2
Dr. Zoe Efstathiadou Department of Endocrinology
“Hippokration” General Hospital of Thessaloniki Konstantinoupoleos 49 54642 Thessaloniki, Greece Tel.: (+30) 6978481331 Fax.: (+30) 2310532372 e-mail: [email protected]
*The authors Z.A. Efstathiadou and C. Sakka have contributed equally to the present work
ACCEPTED MANUSCRIPT Abstract Objective: The association of four single nucleotide polymorphisms in estrogen receptor
T
alpha (ESR1) and beta (ESR2) genes with lipid levels and insulin resistance in men.
RI P
Design and Methods: Lipids, glucose, insulin and HOMA-IR were determined, in a population-based, cross-sectional, cohort of 170 apparently healthy middle-aged Greek men,
SC
along with body mass index (BMI), waist circumference (WC) and percentage of body fat content (%fat). Genotyping of ESR1 for PvuII and XbaI and ESR2 for RsaI and AluI
MA NU
polymorphisms was performed.
Results: Associations of AluI with LDL-Chol (mean±SD, aa 4.3±1.1 vs. Aa 3.7±1.0 and ΑΑ 4.2±1.1, p=0.023) and RsaI with HOMA-IR [median(IQR) RR 1.55(0.88-2.49) vs. Rr/rr 1.69(0.72-2.29), p=0.032] were found. Synergistic effects of RsaI and AluI of ESR2 gene on LDL-Chol levels, %fat and WC, as well as a synergistic effect of both ESR1 and ESR2 genes on
ED
levels of T-Chol (p=0.01) and LDL-Chol (p=0.027) were also shown. These findings remained significant after adjustment for potential confounders.
PT
Significant independent associations of PvuII with %fat (mean±SD, pp 24.6±5.3 vs Pp 22.4±5.2 and PP 21.2±6.7, p=0.044), and RsaI with % fat (RR 22.6 5.5 vs. Rr/rr 25.2 6.3,
CE
p=0.015) and WC (mean±SD RR 97.4 10.4 vs. Rr/rr 102.6 12.6, p=0.013) were found. Synergistic effects on %fat, between the ESR1 polymorphisms (p=0.004), between the ESR2
AC
polymorphisms and among all four ESR polymorphisms studied were also present.
Conclusions: ESR2 is associated with LDL-Chol levels and HOMA-IR in men independently of confounders. Body fat is affected by both genes. Furthermore, a synergistic effect of ESR1 and ESR2 on T-Chol, LDL-Chol and %fat, was shown.
ACCEPTED MANUSCRIPT 1. Introduction There is evidence that genetic variation in estrogen receptor alpha (ESR1) and beta (ESR2) genes is associated with increased cardiovascular risk [1,2], premature ischemic heart
T
disease [3], and unfavorable angiographic changes [4] in both sexes. However, there are
RI P
data contrasting these observations [5].
Through their receptors, estrogens can modulate cardiovascular risk indirectly via
SC
alterations in lipid and glucose metabolism, and directly via their vascular and endothelial effects. Functionally, competent estrogen receptors have been identified in vascular smooth muscle cells, and specific binding sites have been demonstrated in endothelium [6].
MA NU
Furthermore, apart from the classical genomic effect involving gene expression, nontranscriptional mechanisms for estrogen receptor signaling have also been described [7, 8]. On the other hand, genetic association studies have previously shown that ESR1 [911] and ESR2 [12] gene polymorphisms are associated with lipid levels and interact with hormone replacement therapy [12]. An effect on insulin sensitivity and incidence of type 2
ED
diabetes has also been described [13, 14]. However, the effect has not been consistent across different ethnic groups [15, 16], across ages [16] and, finally, it was more prominent
PT
in females than in males [16, 17].
Based on the facts that ESR2 gene is not included in most association studies, data
CE
on males are scarce and inconclusive and there exists significant ethnicity-related variability, we tested the hypothesis that polymorphisms in both genes encoding estrogen receptors
AC
(ESR1 and ESR2) might have an impact on somatometric parameters, serum lipid levels and insulin resistance in Caucasian apparently healthy males. 2. Materials and methods 2.1. Sample selection A total of 170 apparently healthy men, Caucasians, ethnic Greeks, aged 22 to 59 years (mean age ±SD, 42±8.9), were included in the study. Subjects with diagnosed diabetes mellitus, arterial hypertension, cardiovascular disease, renal or hepatic disease, androgen deficiency and those using any medication were excluded from the study. The men enrolled in the study were randomly selected from a larger cluster random sample that was drawn from the adult population of the region of Thessaly, Central Greece (18). The original sample was composed of adults, aged 18 to 79 years, who had been residents of the region of Thessaly for at least one year. Briefly, the total population of
ACCEPTED MANUSCRIPT adults was divided into geographical clusters corresponding to the four provinces of the region. The sample of this study was subsequently selected randomly from a) the capital of each province and b) randomly selected smaller towns and villages. The sample size was
T
proportional to the population of the capital and of each village and town according to the
RI P
2001 census conducted by the National Statistical Service. Only one adult was selected randomly from each household. An interview in a public health office was arranged for each participant, who was instructed to report in the following morning at 9.00 a.m. after an
SC
overnight fast. 2.2. Study protocol
MA NU
Each participant completed a questionnaire inquiring about health status and health behaviors (smoking and alcohol intake) under a doctor’s supervision. Subsequently, each man underwent a thorough physical examination. Anthropometric measurements, such as height, weight and waist circumference (WC), systolic (SBP) and diastolic blood pressure (DBP) were measured in all study participants. Waist circumference was measured using a
ED
plastic non-stretchable tailors measuring tape, at a level midway between the lower rib margin and iliac crest to the nearest centimeter. The body mass index (BMI) was calculated
height in meters.
PT
according to Quetelet’s formula as body weight in kilograms divided by the square of the
CE
The body fat content (% fat mass) was assessed by means of bioelectrical impedance analysis, using a commercially bio-impedance analyzer (Omron HBF 302 Body Fat Analyzer,
AC
Omron Healthcare Inc., Vernon Hills, IL, USA) with the patient in standing position. Finally, plasma glucose and serum insulin, total cholesterol, HDL-cholesterol and triglycerides were measured in early morning samples drawn at the fasting state. LDLcholesterol was estimated with Friedwald method (for triglyceride values not exceeding 400mg/dl). Homeostasis Model Assessment was used for the calculation of HOMA-IR with the formula: fasting insulin (μU/ml)*fasting glucose (mmol/L)/22.5 [19]. All participants gave written informed consent and the study protocol was reviewed and approved by the Ethical Committee of the University Hospital. 2.3. Assays All plasma and serum parameters –except insulin- were measured with standard commercially available assays in an Olympus AU600 analyzer (Olympus Optical, Tokyo, Japan). Serum insulin levels were measured using IRMA (Immunotech Beckman Counter
ACCEPTED MANUSCRIPT Company, Marseille, France, PW). The assay sensitivity for serum insulin measurements was 0.5 IU⁄mL and the interassay coefficient of variation 3.4%.
T
2.4. Genotyping
RI P
All study participants were genotyped for the PvuII (rs2234693) and XbaI (rs9340799) polymorphisms of the ESR1 gene and for the AluI (rs4986938) and RsaI
SC
(rs1256049) polymorphisms of the ESR2 gene. Genomic DNA was extracted from peripheral leucocytes by standard procedures.
For ESR1 gene, 1.3kb DNA fragment (which contained intron 1 and a part of exon 2)
MA NU
was amplified using the following primers: (5’-CTGCCACCCTATCTGTATCTTTTCCTATTCTCC-3’ and 5’-TCTTTCTCTGCCACCCTGGGGTCGATTATCTGA-3’).
Polymerase chain reaction (PCR) amplification was performed for 35 cycles at an annealing temperature of 63C. Amplified DNA was digested with PvuII or XbaI (New England Biolabs Inc Ipswich, MA) at 37C for 3 hours and then electrophoresed in a 1.5%
ED
ethidium bromide agarose gel. PvuII digestion reveals genotypes denoted PP(1.3kb), Pp(1.3kb, 850bp, 450bp) and pp(850bp, 450bp), and XbaI digestion results in genotypes
PT
XX(1.3kb), Xx(1.3kb, 910bp, 390bp) and xx(910bp, 390bp). The restriction fragment length polymorphisms (RLFPs) were presented by P and p (PvuII), and X and x (XbaI), with capital
CE
letters denoting the absence and small letters the presence of restriction sites, respectively. For ESR2 gene, the ligand binding the domain of exon 5 and the 3’ untranslated
AC
region of exon 8 were amplified using the following primers: A, 5’-TCTTGCTTTCCCCAGGCTTT3’; B, 5’-ACCTGTCCAGAACAAGATCT-3’; C, 5’-TTTTTGTCCCCATAGTAACA-3’; and D, 5’AATGAGGGACCACAGCA-3’. PCR amplification was performed for 40 cycles at an annealing temperature of 52C.
Αmplification of exons 5 and 8 yielded the desired products of 156bp and 307bp, respectively. Amplified DNA was digested with RsaI and AluI (New England Biolabs Inc, Ipswich, MA) at 37C for 12 hours and then electrophoresed in a 2.0% ethidium bromide agarose gel. RsaI digestion reveals genotypes denoted RR(156bp), Rr(156bp, 125bp, 31bp) and rr(125bp, 31bp), and AluI digestion results in Aa(307bp), Aa(307bp, 240bp, 67bp) and aa (240bp, 67bp). The RLFPs were presented by R and r (RsaI), and A and a (AluI), with capital letters denoting the absence and small ones the presence of restriction sites, respectively.
ACCEPTED MANUSCRIPT 2.5. Statistical analysis Continuous variables are presented as mean ± standard deviation (SD) or median (interquartile range; IQR) in cases of skewed distributions. Categorical variables are
T
presented as frequencies and/or percentages. Hardy-Weinberg equilibrium was tested with
RI P
the chi-square test for each considered polymorphism. The chi-square test was also used to check for linkage disequilibrium between the two polymorphisms of ESR1 and ESR2 genes. Differences in continuous variables between different genotypes were performed
SC
with one-way analysis of variance for each polymorphism separately. In case of significant differences between genotypes within a certain polymorphism the least significance
MA NU
difference test was used for post-hoc analysis (multiple comparisons). Multivariate linear regression was subsequently performed to evaluate the potential independence of different genotype effects. In particular, comparisons of lipid levels and HOMA-IR across genotypes [three groups; for RsaI only two groups were formed: Rr and rr were studied as one group due to small number of men carrying the rr genotype (n=2)] were
ED
adjusted for age, smoking, alcohol consumption, blood pressure, waist circumference, % body fat and HOMA-IR (for the study of lipids). The possibility of synergism between the two polymorphisms of ESR1 gene or ESR2 gene was also tested. Finally, the synergistic effect of
PT
all four polymorphisms was studied. Only genotypic groups consisting of 5 or more subjects were included in the analyses.
CE
P values
