Abstract
Objective
Sex steroid hormones have been postulated to involve in blood pressure (BP) regulation. We examine the association of endogenous sex hormone levels with longitudinal change of BP and risk of developing hypertension in initially normotensive postmenopausal women.
Methods
We conducted prospective analysis among 619 postmenopausal women free of hypertension at baseline in the Multi-Ethnic Study of Atherosclerosis (MESA). Change of BP and development of incident hypertension were assessed during a mean of 4.8 years follow-up.
Results
After adjusting for age, race/ethnicity, and lifestyle factors, baseline serum estradiol (E2), total and bioavailable testosterone (T), dehydroepiandrosterone (DHEA) were each positively and sex- hormone binding globulin (SHBG) was inversely associated with risk of hypertension. Additional adjustment for body mass index eliminated the associations for E2 and T but only attenuated the associations for DHEA and SHBG. The corresponding multivariable hazard ratios (95% CIs) in the highest quartile were 1.28 (0.83–1.97) for E2, 1.38 (0.89–2.14) for total T, 1.42 (0.90–2.23) for bioavailable T, 1.54 (1.02–2.31) for DHEA, and 0.48 (0.30–0.76) for SHBG. Adjustment for fasting glucose, insulin, and C-reactive protein further attenuated the association for DHEA but not SHBG. Associations of sex hormones with longitudinal BP change were similar.
Conclusion
In postmenopausal women, higher endogenous E2, T, and DHEA and lower SHBG were associated with higher incidence of hypertension and greater longitudinal rise in BP. The associations for E2, T, and DHEA were mostly explained by adiposity, while the association for SHBG was independent of measures of adiposity, insulin resistance, and systemic inflammation.
Keywords: sex steroid hormones, hypertension, blood pressure, postmenopausal women, prospective study, epidemiology
Introduction
High blood pressure (BP) is one of the largest contributors to death in the U.S. and worldwide [1]. BP differs by sex [2], which may be explained partly by endogenous sex steroid hormones. Evidence from laboratory studies shows that both estrogens and androgens can affect BP. Estrogens can promote vasodilation and lower BP [3]; whereas estrogens may also cause impairment in insulin signaling pathway [4], affect normal immune function [5], and induce cytokine activities [6], subsequently raise BP. Androgens also have been shown to activate both vasoconstricting [7] and vasorelaxing [8] mechanisms.
Epidemiologic studies that examined the relations between circulating sex hormone levels and BP or hypertension status are largely confined to cross-sectional studies and retrospective case-control studies, with inconsistent results. In women, higher estradiol (E2) levels were found among those with elevated BP or hypertension than those with normal BP in some [9], but not all [10, 11] studies. Similarly, total or free testosterone (T) concentrations were positively correlated with BP among postmenopausal women in some studies [11, 12] but not in others [13]. Sex-hormone binding globulin (SHBG), a serum protein that affects free circulating sex hormone levels [14], has been shown either not associated [9, 10] or inversely associated [12] with BP in women.
To further evaluate the possible role of sex steroid hormones in BP regulation and development of hypertension, we examined the prospective associations of baseline serum concentrations of endogenous sex hormones, including E2, T, dehydroepiandrosterone (DHEA, a pre-androgen), and SHBG, with the subsequent change of BP and the incidence of hypertension among postmenopausal women free of baseline hypertension.
Materials and Methods
Study Subjects
The Multi-Ethnic Study of Atherosclerosis (MESA) is a multicenter, prospective cohort study of subclinical cardiovascular disease (CVD). Study design has been described in detail elsewhere [15]. In brief, 6,814 men and women from 4 ethnic origins (Caucasian, African American, Hispanic, and Asian American), aged 45–84 years and free of clinically diagnosed CVD, were enrolled from 6 US field centers. The baseline, second, third, and fourth clinic examination took place from 2000 through 2002, 2002 through 2004, 2004 through 2005, and 2005 through 2007, respectively. The study procedures were approved by Institutional Review Boards at all participating sites. All participants provided written consent. For the current study, we considered 3199 postmenopausal female participants in the MESA, excluding those with baseline hypertension (n=1,556) and those who had ever used hormone replacement therapy (n=1,526), had no sex hormone measurements (n=303), or had insufficient information to ascertain incident hypertension status during follow-up (n=222). As a result, 619 postmenopausal women remained for analyses, of whom 595, 570, and 558 attended the second, third, and fourth examination, respectively.
Measurement of Blood Pressure and Ascertainment of Hypertension
At baseline and each follow-up examination, physician diagnosis of hypertension was obtained from self-report, use of BP-lowering medications was assessed by reviewing participants’ medications brought to clinic, and resting seated BP was measured three times in the right arm using a Dinamap model Pro 100 automated sphygmomanometer (Critikon, Tampa, FL), with average of the last two readings used for analysis. Hypertension was defined as measured systolic BP (SBP) ≥140 mmHg, or diastolic BP (DBP) ≥90 mmHg, or using medications for physician diagnosed hypertension. Incident hypertension was defined as newly developed hypertension during follow-up among those free of hypertension at baseline. Time of event was assigned to the mid-point between the last visit without hypertension and the first visit with hypertension.
Measurement of Sex Hormones
Fasting blood samples were obtained from participants between 7:30 and 10:30 AM at baseline examination. Serum samples were extracted by centrifugation at 2000 × g for 15 min or 3000 × g for 10 min, immediately frozen at −70°C, and then shipped to the University of Vermont for long-term storage and future analyses. Serum sex hormone concentrations were measured in the Sex Hormone Laboratory at the University of Massachusetts Medical Center in Worcester, MA. Total E2 was measured using an ultrasensitive RIA kit from Diagnostic Systems Laboratory (Webster, TX). Total T and DHEA were measured directly using RIA kits. SHBG was measured by chemiluminescent enzyme immunometric assay using Immulite kits from Diagnostic Products Corp. (Los Angeles, CA). Concentrations of free T, SHBG-bound T, and albumin-bound T were calculated according to the method of Södergard et al [16]. Bioavailable T was determined as total T minus SHBG-bound T. Calculated bioavailable T has been shown to be comparable to free T concentration obtained by equilibrium dialysis [17]. Assay quality was blindly monitored by assessing 5% randomly selected duplicate participant samples and quality control sera included with the kits. The overall coefficients of variation for total E2, total T, DHEA, and SHBG were 10.5, 12.3, 11.2 and 9.0%, respectively.
Assessment of Other Covariates at Baseline
Participants came to clinic examinations after an 8- to 12-h overnight fast. Information on demographics, lifestyle factors, reproductive history, and medical history were collected from standardized questionnaires. Physical activity was assessed using the MESA Typical Week Physical Activity Survey, which was adapted from the Cross-Culture Activity Participation Study[18] and designed to assess the time spent in and frequency of various physical activities during a typical week in the past month.[19] Smoking and alcohol consumption were self-reported. Anthropometrics including weight, height, and waist circumference were measured using standard protocols as described previously [15]. Body mass index (BMI) was calculated. Serum concentrations of fasting glucose (by the glucose oxidase method using a Vitros 950 analyzer, Johnson & Johnson Ortho-Clinical Diagnostics, Rochester, NY), fasting insulin (by a radioimmunoassay method using the Linco Human Insulin Specific RIA kit, Linco Research, Inc., St. Charles, MO), and C-reactive protein (CRP) (using the BNII nephelometer, N High-Sensitivity CRP, Dade Behring Inc., Deerfield, IL) were measured at central laboratory.
Statistical Analyses
Analyses were conducted using SAS version 9.1 (SAS Institute, Cary, NC). Distribution of each biomarker was checked for outliers, and logarithmic transformation was used to normalize skewed distributions for computation of means. Hypertension risk factors and sex hormone concentrations at baseline were compared between women who developed incident hypertension and those who did not. Cox regression was performed to calculate hazard ratios (HRs) and 95% CIs of incident hypertension according to quartiles of sex hormones. Covariates, including age, race/ethnicity, age at menopause, smoking status, alcohol use, physical activity, and BMI, were sequentially controlled for. Fasting glucose, fasting insulin, and CRP were added lastly as potential explanatory factors. Analyses were also stratified by race/ethnicity, baseline BMI, baseline BP, and interactions by these factors were tested.
In addition, longitudinal change of BP from baseline to each follow-up examination was compared across quartiles of sex hormone variables, using PROC MIXED models with an unstructured covariance matrix for repeated measures. For women who started anti-hypertensive treatment, we imputed the untreated BP at the first visit when hypertension was detected by adding 15 mmHg and 10 mmHg to the measured SBP and DBP, respectively, and assigned BP at all following examinations to missing. Previous simulation studies have demonstrated that addition of a sensible constant to the observed BP is an appropriate method to adjust for the underlying BP in treated subjects [20].
Sensitivity analyses were performed. First, we included women who had previously used hormone replacement therapy but had stopped for ≥1 year; second, we excluded women with baseline diabetes; third, we substituted waist circumference for BMI as a measure of adiposity; fourth, we adjusted for BMI or waist circumference with quadratic terms; fifth, we added a constant of other values to impute untreated BP; finally, we additionally adjusted for baseline BP. All sensitivity analyses obtained results generally similar to main analyses, and therefore are not presented.
Results
The mean (SD) age of 619 postmenopausal women free of hypertension at baseline was 63.5 (9.1) years. The time after menopause was ≥10 years for 63.5% and ≥20 years for 34.1% of women. Over a mean of 4.8 (maximum: 6.7) years follow-up, 194 women developed incident hypertension. As expected, women who developed hypertension were significantly older, more likely in the late postmenopausal period, and had higher BP at baseline than those who remained free of hypertension.(Table 1) Hypertension cases also had higher baseline BMI, waist circumference, fasting glucose, fasting insulin, and CRP. In univariate analyses, compared with women who did not develop incident hypertension, those who developed hypertension had similar baseline concentrations of E2 and DHEA, significantly higher total T and bioavailable T, and significantly lower SHBG.(Table 1)
Table 1.
Baseline characteristics of postmenopausal women who developed incident hypertension compared to those who remained free of hypertension, MESA 2000–2007.
| Characteristics | Incident HTN (N=194) | No HTN (N=425) | Pa |
|---|---|---|---|
| Systolic blood pressure, mmHg | 125.2±0.77b | 111.3±0.61 | < 0.0001 |
| Diastolic blood pressure, mmHg | 68.6±0.57 | 63.7±0.39 | < 0.0001 |
| Age, years | 66.0±0.62 | 62.3±0.44 | < 0.0001 |
| Age at menopause, years | 47.8±0.47 | 48.1±0.29 | 0.59 |
| Years after menopause, % | < 0.0001 | ||
| <5 years | 10.8 | 24.7 | |
| 5-<10 years | 15.0 | 16.7 | |
| 10-<20 years | 27.8 | 30.1 | |
| ≥2 0 years | 46.4 | 28.5 | |
| Race/ethnicity, % | 0.35 | ||
| Caucasian | 32 | 32 | |
| African American | 25 | 20 | |
| Hispanic | 28 | 28 | |
| Asian American | 15 | 20 | |
| Cigarette smoking, % | 0.97 | ||
| Current | 12.4 | 12.5 | |
| Former | 24.7 | 23.8 | |
| Never | 62.9 | 63.7 | |
| Alcohol use, % | 0.60 | ||
| Current | 43.8 | 45.3 | |
| Former | 22.2 | 18.7 | |
| Never | 34.0 | 36.0 | |
| Physical activity, MET-min/week Adiposity | 1208±116.2 | 1354±100.1 | 0.34 |
| Body mass index, kg/m2 | 29.2±0.43 | 27.4±0.27 | 0.0002 |
| Waist circumference, cm | 98.3±1.07 | 93.9±0.69 | 0.0005 |
| Metabolic and inflammatory markersc | |||
| Fasting glucose, mg/dLd | 97.3 (85, 101) | 90.7 (81, 94) | 0.001 |
| Fasting insulin, mU/L | 5.90 (3.90, 8.60) | 4.95 (3.40, 7.40) | 0.0006 |
| C-reactive protein, mg/Le | 2.39 (1.09, 5.58) | 1.76 (0.72, 3.97) | 0.003 |
| Sex hormone variablesc | |||
| E2, nmol/L | 0.053 (0.040, 0.081) | 0.054 (0.037, 0.081) | 0.75 |
| Total T, nmol/L | 0.98 (0.73, 1.35) | 0.88 (0.59, 1.28) | 0.03 |
| Bioavailable T, nmol/L | 0.28 (0.17, 0.42) | 0.23 (0.14, 0.38) | 0.002 |
| DHEA, nmol/L | 11.2 (7.84, 16.2) | 10.9 (7.95, 15.5) | 0.57 |
| SHBG, nmol/L | 46.8 (35.1–62.3) | 52.5 (36.7, 72.4) | 0.005 |
P values are derived from t-test for continuous variables and χ2 test for categorical variables.
Mean±standard error and all such values
Geometric means (interquartile range) are shown for all biomarkers because of the skewed distribution.
The characteristics of baseline hypertension risk factors according to quartiles of sex hormone variables are shown in Supplement Table. Women were on average younger, presumably closer to menopause, in higher quartiles of E2, bioavailable T, and DHEA, but were older in higher quartiles of SHBG. BMI was greater with higher E2, total T, and bioavailable T, while lower with higher SHBG. E2, total T and DHEA seem to have a graded, positive relation with smoking; a similar trend was seen for alcohol in relation to E2. No other consistent relations were found for smoking, alcohol, and physical activity. Without adjustment, E2, bioavailable T, and DHEA were positively related to fasting glucose, fasting insulin, and CRP, while SHBG was strongly and inversely related to these biomarkers.
After adjusting for age and race/ethnicity, E2, total T, bioavailable T, and DHEA were each positively associated, while SHBG was inversely associated with risk of hypertension.(Table 2) Adjustment for hypertension-related lifestyle factors including cigarette smoking, alcohol use, and physical activity (multivariable model 1) did not materially change these associations. Additional adjustment for BMI eliminated the associations for E2, total T, and bioavailable T, but not for DHEA and SHBG. The corresponding multivariable-adjusted HRs and 95% CIs in the highest versus the lowest quartile were 1.28 (0.83–1.97) for E2, 1.38 (0.89–2.14) for total T, 1.42 (0.90–2.23) for bioavailable T, 1.54 (1.02–2.31) for DHEA, and 0.48 (0.30–0.76) for SHBG. Adjustment for fasting glucose, fasting insulin, and CRP (multivariable model 2) further attenuated the HR in the highest quartile of DHEA, while the inverse association for SHBG remained strong, graded, and significant.
Table 2.
Hazard ratios and 95% confidence intervals (CI) of incident hypertension according to quartiles of baseline sex hormone variables, MESA 2000–2007.
| Sex Hormones | N of case/total | Hazard ratios (95% CI)
|
|||
|---|---|---|---|---|---|
| Age, race/ethnicity adjusted | Multivariable model 1a | Multivariable model 1a + BMI | Multivariable model 2b | ||
| E2 | |||||
| Q1 | 46/121 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Q2 | 50/103 | 1.28 (0.86–1.92) | 1.27 (0.85–1.92) | 1.28 (0.85–1.92) | 1.20 (0.79–1.81) |
| Q3 | 47/89 | 1.51 (1.00–2.29) | 1.53 (1.01–2.32) | 1.39 (0.91–2.13) | 1.34 (0.87–2.05) |
| Q4 | 51/112 | 1.50 (1.00–2.25) | 1.51 (0.99–2.29) | 1.28 (0.83–1.97) | 1.14 (0.74–1.77) |
| P, trendc | 0.047 | 0.046 | 0.29 | 0.58 | |
| Total T | |||||
| Q1 | 36/120 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Q2 | 46/100 | 1.48 (0.96–2.30) | 1.37 (0.88–2.14) | 1.26 (0.81–1.99) | 1.29 (0.82–2.03) |
| Q3 | 61/105 | 1.87 (1.24–2.84) | 1.74 (1.15–2.64) | 1.62 (1.06–2.46) | 1.64 (1.08–2.51) |
| Q4 | 51/100 | 1.60 (1.04–2.47) | 1.51 (0.97–2.33) | 1.38 (0.89–2.14) | 1.33 (0.86–2.08) |
| P, trendc | 0.04 | 0.07 | 0.16 | 0.22 | |
| Bioavailable T | |||||
| Q1 | 36/113 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Q2 | 50/119 | 1.25 (0.82–1.92) | 1.21 (0.78–1.87) | 1.12 (0.72–1.75) | 1.08 (0.69–1.69) |
| Q3 | 57/102 | 1.71 (1.13–2.61) | 1.66 (1.08–2.55) | 1.49 (0.97–2.31) | 1.42 (0.92–2.20) |
| Q4 | 51/91 | 1.76 (1.15–2.71) | 1.67 (1.08–2.58) | 1.42 (0.90–2.23) | 1.27 (0.80–2.02) |
| P, trendc | 0.007 | 0.01 | 0.11 | 0.27 | |
| DHEA | |||||
| Q1 | 50/104 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Q2 | 46/110 | 1.04 (0.69–1.56) | 0.94 (0.62–1.43) | 0.91 (0.60–1.38) | 0.83 (0.54–1.26) |
| Q3 | 41/112 | 1.07 (0.69–1.64) | 1.00 (0.65–1.55) | 0.93 (0.60–1.44) | 0.87 (0.56–1.34) |
| Q4 | 57/99 | 1.68 (1.12–2.51) | 1.59 (1.06–2.38) | 1.54 (1.02–2.31) | 1.38 (0.92–2.08) |
| P, trendc | 0.01 | 0.02 | 0.02 | 0.07 | |
| SHBG | |||||
| Q1 | 54/101 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Q2 | 63/92 | 1.05 (0.72–1.51) | 1.03 (0.71–1.50) | 1.05 (0.72–1.52) | 1.16 (0.79–1.70) |
| Q3 | 42/111 | 0.59 (0.39–0.89) | 0.57 (0.37–0.86) | 0.60 (0.39–0.91) | 0.67 (0.43–1.03) |
| Q4 | 35/121 | 0.43 (0.28–0.67) | 0.42 (0.27–0.66) | 0.48 (0.30–0.76) | 0.54 (0.33–0.88) |
| P, trendc | < 0.0001 | < 0.0001 | 0.0003 | 0.002 | |
Multivariable model 1 adjusted for age (continuous), race/ethnicity (Caucasian, African American, Hispanic, Asian American), age at menopause (continuous), cigarette smoking (current, former, never), alcohol use (current, former, never), physical activity (quartiles).
Multivariable model 2 adjusted for variables in model 1 plus BMI and serum concentrations of fasting glucose, fasting insulin, and C-reactive protein (all continuous).
P, trend was tested using the median value in each quartile of the sex hormone variables as ordinal variables.
We also examined the longitudinal change of BP from baseline through the last follow-up in relation to sex hormone levels.(Table 3) After adjusting for age and race/ethnicity, the increment of SBP and/or DBP over time tended to be greater in the higher quartiles of E2, total T, bioavailable T, and DHEA while the BP increase was smaller across increasing quartiles of SHBG. Adjustment for BMI eliminated the associations for E2, total T, and bioavailable T, but not for DHEA and SHBG. Further adjustment for fasting glucose, fasting insulin, and CRP continued attenuating the association for DHEA, while the association for SHBG remained strong and highly significant. When BP levels at baseline and at each follow-up examination were compared between quartiles of SHBG, a consistent, graded pattern was observed, i.e. the increment of SBP and DBP over time was smaller with higher concentrations of baseline SHBG.(Figure 1)
Table 3.
Mean±standard error of systolic blood pressure (SBP) / diastolic blood pressure (DBP) change from baseline to the last follow-up visit according to quartiles of baseline sex hormone variables, MESA 2000–2007
| Sex Hormones | SBP changea
|
DBP changea
|
||||
|---|---|---|---|---|---|---|
| Age, race/ethnicity adjusted | Multivariable model 1b + BMI | Multivariable model 2c | Age, race/ethnicity adjusted | Multivariable model 1b + BMI | Multivariable model 2c | |
| E2 | ||||||
| Q1 | 5.71±1.34 | 6.40±1.38 | 6.71±1.36 | 0.68±0.62 | 0.85±0.64 | 0.95±0.64 |
| Q2 | 8.19±1.43 | 8.22±1.44 | 8.10±1.42 | 1.76±0.66 | 1.75±0.67 | 1.71±0.66 |
| Q3 | 7.09±1.51 | 7.23±1.52 | 7.11±1.50 | 2.49±0.70 | 2.53±0.71 | 2.53±0.70 |
| Q4 | 7.18±1.38 | 6.67±1.44 | 6.33±1.42 | 1.78±0.64 | 1.62±0.67 | 1.47±0.66 |
| P, trendd | 0.59 | 0.97 | 0.73 | 0.23 | 0.40 | 0.57 |
| Total T | ||||||
| Q1 | 5.76±1.34 | 6.35±1.39 | 6.25±1.36 | 0.69±0.62 | 0.86±0.64 | 0.88±0.64 |
| Q2 | 7.49±1.45 | 7.39±1.48 | 7.31±1.46 | 1.82±0.68 | 1.81±0.69 | 1.79±0.68 |
| Q3 | 8.04±1.39 | 8.14±1.40 | 8.08±1.38 | 2.19±0.65 | 2.22±0.65 | 2.20±0.65 |
| Q4 | 6.77±1.44 | 6.54±1.45 | 6.50±1.44 | 1.85±0.67 | 1.67±0.68 | 1.63±0.68 |
| P, trendd | 0.66 | 0.95 | 0.92 | 0.23 | 0.43 | 0.46 |
| Bioavailable T | ||||||
| Q1 | 5.58±1.40 | 6.66±1.47 | 6.77±1.45 | 0.68±0.65 | 1.09±0.68 | 1.18±0.68 |
| Q2 | 7.43±1.32 | 7.51±1.33 | 7.44±1.31 | 2.06±0.61 | 1.99±0.62 | 1.99±0.61 |
| Q3 | 7.24±1.43 | 7.22±1.45 | 6.94±1.43 | 1.50±0.67 | 1.46±0.68 | 1.30±0.67 |
| Q4 | 7.75±1.48 | 6.98±1.50 | 6.96±1.50 | 2.20±0.69 | 1.93±0.70 | 1.96±0.70 |
| P, trendd | 0.37 | 0.98 | 0.97 | 0.21 | 0.56 | 0.60 |
| DHEA | ||||||
| Q1 | 5.21±1.46 | 5.44±1.50 | 5.76±1.48 | 0.31±0.68 | 0.49±0.69 | 0.63±0.69 |
| Q2 | 7.13±1.40 | 7.25±1.43 | 7.00±1.41 | 1.59±0.65 | 1.48±0.66 | 1.36±0.66 |
| Q3 | 5.73±1.37 | 5.84±1.38 | 5.74±1.36 | 1.02±0.63 | 0.99±0.64 | 0.97±0.64 |
| Q4 | 9.66±1.40 | 9.52±1.40 | 9.35±1.38 | 3.48±0.65 | 3.44±0.65 | 3.37±0.65 |
| P, trendd | 0.05 | 0.08 | 0.11 | 0.002 | 0.004 | 0.006 |
| SHBG | ||||||
| Q1 | 11.36±1.41 | 10.74±1.46 | 10.40±1.47 | 3.39±0.66 | 3.15±0.69 | 3.10±0.69 |
| Q2 | 6.66±1.44 | 6.80±1.47 | 6.73±1.45 | 1.52±0.68 | 1.62±0.69 | 1.60±0.69 |
| Q3 | 7.15±1.34 | 7.21±1.35 | 7.49±1.34 | 1.54±0.63 | 1.53±0.63 | 1.58±0.63 |
| Q4 | 2.67±1.35 | 3.50±1.43 | 3.44±1.43 | 0.085±0.63 | 0.28±0.67 | 0.22±0.68 |
| P, trendd | < 0.0001 | 0.001 | 0.002 | 0.0009 | 0.006 | 0.007 |
Longitudinal change of BP from baseline to each follow-up visit was computed using PROC MIXED models with an unstructured covariance matrix for repeated measures. For women who developed incident hypertension and were on anti-hypertensive treatment, the untreated BP at the first visit when hypertension was detected was imputed by adding 15 mmHg for SBP and 10 mmHg for DBP to the measured BP, and BP at all following visits were assigned to missing.
Multivariable model 1 in Table 2 was used.
Multivariable model 2 in Table 2 was used.
P, trend was tested using the median value in each quartile of the sex hormone variables as ordinal variables.
Figure 1.
Adjusted mean systolic and diastolic blood pressure (BP) at baseline (1) and follow-up (2, 3, 4) visits by quartiles of sex hormone binding globulin (SHBG) were computed using PROC MIXED models with an unstructured covariance matrix for repeated measures. For women who started anti-hypertensive treatment, the untreated BP at the first visit when hypertension was detected was imputed by adding 15 mmHg and 10 mmHg to the measured systolic and diastolic BP, respectively, and BPs at all following visits were assigned to missing. Model adjusted for age, race/ethnicity, age at menopause, cigarette smoking, alcohol use, physical activity, and body mass index.
When we stratified analyses by race/ethnicity, we found only modest differences in concentrations of sex hormones between 4 race/ethnicity groups (data not shown). The inverse association between SHBG and risk of hypertension was generally consistent in all race/ethnicity groups, except for Asian Americans for whom the number of cases was small.(Figure 2) The association of E2 with risk of hypertension appeared to differ by race/ethnicity, but the test for interaction was not statistically significant after considering multiple comparisons. No association differed significantly by baseline BMI or baseline BP (all p for interaction > 0.05).
Figure 2.
Hazard ratio and 95% CI of hypertension across quartiles of sex hormone binding globulin (SHBG) in 4 race/ethnicity groups. N shown in parentheses indicates the numbers of cases/total participants at risk. P, trend was tested using the median value in each quartile of SHBG as ordinal variable. Multivariable model in Figure 1 without race/ethnicity was used.
Discussion
In this prospective study among postmenopausal women without baseline hypertension, higher baseline concentrations of endogenous E2, total and bioavailable T, and DHEA and lower concentration of SHBG were associated with a higher incidence of hypertension and a greater increase in BP during follow-up. The associations for E2 and T were eliminated and the association for DHEA was attenuated by adjustment for BMI, whereas the association for SHBG remained strong and significant after adjusting for BMI, fasting glucose, fasting insulin, and CRP.
The observation of BP variation during episodes of estrogen level change in women has led to the hypothesis that estrogens are involved in BP regulation. Data from in vivo and in vitro studies have provided evidence for E2 – the most potent estrogen in humans – being a vasodilator [3], which will potentially lower BP. However, E2 may also induce insulin resistance and thereafter tend to raise BP. High levels of E2 in physiological states, such as puberty, luteal phase of menstrual cycle, and late pregnancy, are associated with insulin resistance [4]. There are epidemiologic studies [21, 22], including MESA [21], showing a strong association between E2 and measures of insulin resistance in postmenopausal women, independent of adiposity. Estrogen may also increase BP through pro-inflammatory activities. Metabolism of estrogens is altered in patients with systemic immune disease such as lupus erythematusus.[5] In postmenopausal women that received hormone replacement therapy, estrogen therapy increased mononuclear cell secretion of tumor necrosis factor alpha (TNF-α).[6] In at least one population-based study, estrone levels were positively associated with inflammatory markers in postmenopausal women [23]. Previous studies on circulating estrogens in relation to BP have mainly used cross-sectional design. One study found higher plasma E2 in 24 hypertensive men and women compared with 24 normotensive controls [9]. However, other studies found no such associations in premenopausal [10] and postmenopausal women [11]. Our study showed that postmenopausal women with higher baseline E2 had increased risk of hypertension and greater BP rise with aging, whereas the association was eliminated by adjustment for BMI. The relation between estrogens and adiposity is likely bidirectional. E2 can be synthesized from aromatization of T in adipocytes. In postmenopausal women, the major source of circulating E2 is derived from the extragonadal conversion of androgen precursors in adipose tissue, with the degree of conversion increasing with the growth of fat mass.[24] On the other hand, some animal studies showed that E2 has direct effects on adipocyte enlargement and weight gain [25]. Because of this bidirectional relation, obesity could be either a confounder or a mediating factor in the link between estrogens and hypertension.
Findings that BP is reduced in male animal models by castration suggest that androgens also control BP, at least in men [26]. Experimental studies show that the most potent androgen – T – can activate both vasoconstriction [7] and vasorelaxation [8]. The balance between the two mechanisms determines the net effect on vascular tone and BP. T may also stimulate cardiac hypertrophy [27], impair insulin sensitivity [28], and increase central adiposity [29]. These effects tilt the cardiovascular system toward a pro-hypertensive state. Previous studies of circulating androgens in relation to BP have also been cross-sectional. Among postmenopausal women, serum T was elevated in hypertensive participants [9, 11, 12], and total T, free T, and DHEA were positively correlated with SBP [11, 12]. Our study extends prior cross-sectional findings to a prospective setting. Similar to E2, the associations for T and DHEA were attenuated by adjustment for BMI, reflecting either a confounding or a mediating effect of obesity.
SHBG, produced by the liver, is the major circulating protein that binds to and transports steroid sex hormones. According to the free-hormone hypothesis, SHBG-bound sex hormone is unavailable to the target tissues. Because the affinity of SHBG to T is higher than to E2, SHBG is conventionally considered an indirect indicator of androgenicity, with its concentrations mainly inversely correlated with free T [14]. Evidence emerged recently for low SHBG as an independent predictor of adverse cardiometabolic outcomes. Compared with the accumulating findings on metabolic syndrome [30] and type 2 diabetes [31], the relation of SHBG with risk of hypertension has not been well studied. Previous cross-sectional studies showed either null [9–11] or inverse [12] association between SHBG and BP among men and women. In the current prospective study of postmenopausal women, we found that SHBG concentration was inversely associated with risk of hypertension and longitudinal rise of BP over time. These associations were strong, monotonic, and independent of known hypertension risk factors including measures of obesity.
The exact biologic mechanisms underlying the association between SHBG and cardiometabolic disorders remain largely unknown. In addition to a direct impact on the bioavailability of E2 and T, SHBG has been postulated as a marker for insulin resistance. In vitro studies showed that insulin inhibits SHBG production from hepatoma cells [32]. In intervention studies, successful weight loss and weight maintenance increased SHBG in men with obesity [33]; suppression of insulin secretion by diazoxide treatment led to increased SHBG levels in both non-diabetic obese men and normal weight men [34]. These findings support a down-regulating effect of insulin on SHBG production. SHBG also inversely correlates with inflammatory markers [35], indicating that a suppressive effect of cytokines on hepatic SHBG synthesis is possible. Consistent with previous observations, our study found inverse relations of SHBG with measures of obesity, insulin resistance, and inflammation. However, adjustment for these factors only slightly attenuated the strong relation between SHBG and hypertension risk, which suggest an independent role of SHBG in the prevention of hypertension.
Several limitations of this study deserve considerations. First, we had only a single baseline measurement of sex hormones, and we lacked information on BP or hypertension status between follow-up examinations. Any random misclassification in these measurements would tend to bias the associations towards null. Second, our study is observational, and thus, although we have adjusted for many covariates, residual confounding cannot be ruled out. Third, since we simultaneously examined several sex hormone variables, false positive findings due to multiple testing are possible. Finally, our study results apply only to postmenopausal women not on hormone therapy. The age range of MESA participants (45–84 years) does not provide us with enough power to assess the association of interest in premenopausal women, which may be different from postmenopausal women.
In conclusion, we found that in a prospective cohort of postmenopausal women without baseline hypertension, higher serum concentrations of endogenous E2, total and bioavailable T, and DHEA and lower concentration of SHBG were each associated with increased risk of hypertension and greater BP rise over time. The associations with E2, T, and DHEA were mostly explained by adiposity, while the association with SHBG was independent of measures of adiposity, insulin resistance, and systemic inflammation. Further studies are needed to identify direct biological effect of SHBG on hypertension and other atherosclerotic disorders.
Supplementary Material
Highlights.
Higher endogenous sex hormones were associated with higher risk of hypertension
Lower sex-hormone binding globulin was associated with higher risk of hypertension
Associations for sex hormones were mostly explained by adiposity
Association for sex-hormone binding globulin was independent of adiposity
Acknowledgments
The authors thank the other investigators, the staff, and the participants of the MESA study for their valuable contributions. A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org. Dr. Lu Wang also thanks the Building Interdisciplinary Research Careers in Women’s Health (BIRCWH) program at Brigham and Women’s Hospital and program director Dr. Ursula Kaiser and program Principal Investigator Dr. Jill M. Goldstein for their generous support to research on health effects of sex hormones.
Funding Sources:
This work was supported by R01 HL074406, R01 HL074338 and contracts N01-HC-95159 through N01-HC-95165 and N01-HC-95169 from the National Heart, Lung, and Blood Institute. Dr. Wang was supported by a Pathway to Independence award K99 HL095649 from the National Heart, Lung, and Blood Institute.
Footnotes
Disclosures of conflict of interest:
None.
Publisher's Disclaimer: 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 citable 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.
Contributor Information
Lu Wang, Email: luwang@rics.bwh.harvard.edu.
Moyses Szklo, Email: mszklo@jhsph.edu.
Aaron R. Folsom, Email: folso001@umn.edu.
Nancy R. Cook, Email: ncook@rics.bwh.harvard.edu.
Susan M. Gapstur, Email: Susan.Gapstur@cancer.org.
Pamela Ouyang, Email: Pouyang@jhmi.edu.
References
- 1.Lopez AD, Mathers CD, Ezzati M, et al. Global and regional burden of disease and risk factors, 2001: systematic analysis of population health data. Lancet. 2006;367:1747–57. doi: 10.1016/S0140-6736(06)68770-9. [DOI] [PubMed] [Google Scholar]
- 2.Reckelhoff JF. Gender differences in the regulation of blood pressure. Hypertension. 2001;37:1199–208. doi: 10.1161/01.hyp.37.5.1199. [DOI] [PubMed] [Google Scholar]
- 3.Caulin-Glaser T, Garcia-Cardena G, Sarrel P, et al. 17 beta-estradiol regulation of human endothelial cell basal nitric oxide release, independent of cytosolic Ca2+ mobilization. Circ Res. 1997;81:885–92. doi: 10.1161/01.res.81.5.885. [DOI] [PubMed] [Google Scholar]
- 4.Livingstone C, Collison M. Sex steroids and insulin resistance. Clin Sci (Lond) 2002;102:151–66. doi: 10.1042/cs1020151. [DOI] [PubMed] [Google Scholar]
- 5.Lahita RG. Sex hormones and systemic lupus erythematosus. Rheum Dis Clin North Am. 2000;26:951–68. doi: 10.1016/s0889-857x(05)70178-2. [DOI] [PubMed] [Google Scholar]
- 6.Brooks-Asplund EM, Tupper CE, Daun JM, et al. Hormonal modulation of interleukin-6, tumor necrosis factor and associated receptor secretion in postmenopausal women. Cytokine. 2002;19:193–200. doi: 10.1006/cyto.2002.1963. [DOI] [PubMed] [Google Scholar]
- 7.Schror K, Morinelli TA, Masuda A, et al. Testosterone treatment enhances thromboxane A2 mimetic induced coronary artery vasoconstriction in guinea pigs. Eur J Clin Invest. 1994;24 (Suppl 1):50–2. doi: 10.1111/j.1365-2362.1994.tb02428.x. [DOI] [PubMed] [Google Scholar]
- 8.Chou TM, Sudhir K, Hutchison SJ, et al. Testosterone induces dilation of canine coronary conductance and resistance arteries in vivo. Circulation. 1996;94:2614–9. doi: 10.1161/01.cir.94.10.2614. [DOI] [PubMed] [Google Scholar]
- 9.Hughes GS, Mathur RS, Margolius HS. Sex steroid hormones are altered in essential hypertension. J Hypertens. 1989;7:181–7. [PubMed] [Google Scholar]
- 10.Sutton-Tyrrell K, Wildman RP, Matthews KA, et al. Sex-hormone-binding globulin and the free androgen index are related to cardiovascular risk factors in multiethnic premenopausal and perimenopausal women enrolled in the Study of Women Across the Nation (SWAN) Circulation. 2005;111:1242–9. doi: 10.1161/01.CIR.0000157697.54255.CE. [DOI] [PubMed] [Google Scholar]
- 11.Phillips GB, Jing TY, Laragh JH. Serum sex hormone levels in postmenopausal women with hypertension. J Hum Hypertens. 1997;11:523–6. doi: 10.1038/sj.jhh.1000481. [DOI] [PubMed] [Google Scholar]
- 12.Haffner SM, Newcomb PA, Marcus PM, et al. Relation of sex hormones and dehydroepiandrosterone sulfate (DHEA-SO4) to cardiovascular risk factors in postmenopausal women. Am J Epidemiol. 1995;142:925–34. doi: 10.1093/oxfordjournals.aje.a117740. [DOI] [PubMed] [Google Scholar]
- 13.Bonithon-Kopp C, Scarabin PY, Bara L, et al. Relationship between sex hormones and haemostatic factors in healthy middle-aged men. Atherosclerosis. 1988;71:71–6. doi: 10.1016/0021-9150(88)90303-6. [DOI] [PubMed] [Google Scholar]
- 14.Anderson DC. Sex-hormone-binding globulin. Clin Endocrinol (Oxf) 1974;3:69–96. doi: 10.1111/j.1365-2265.1974.tb03298.x. [DOI] [PubMed] [Google Scholar]
- 15.Bild DE, Bluemke DA, Burke GL, et al. Multi-ethnic study of atherosclerosis: objectives and design. Am J Epidemiol. 2002;156:871–81. doi: 10.1093/aje/kwf113. [DOI] [PubMed] [Google Scholar]
- 16.Sodergard R, Backstrom T, Shanbhag V, et al. Calculation of free and bound fractions of testosterone and estradiol-17 beta to human plasma proteins at body temperature. J Steroid Biochem. 1982;16:801–10. doi: 10.1016/0022-4731(82)90038-3. [DOI] [PubMed] [Google Scholar]
- 17.Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab. 1999;84:3666–72. doi: 10.1210/jcem.84.10.6079. [DOI] [PubMed] [Google Scholar]
- 18.Ainsworth BE, Irwin ML, Addy CL, et al. Moderate physical activity patterns of minority women: the Cross-Cultural Activity Participation Study. J Womens Health Gend Based Med. 1999;8:805–13. doi: 10.1089/152460999319129. [DOI] [PubMed] [Google Scholar]
- 19.Bertoni AG, Whitt-Glover MC, Chung H, et al. The association between physical activity and subclinical atherosclerosis: the Multi-Ethnic Study of Atherosclerosis. Am J Epidemiol. 2009;169:444–54. doi: 10.1093/aje/kwn350. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Tobin MD, Sheehan NA, Scurrah KJ, et al. Adjusting for treatment effects in studies of quantitative traits: antihypertensive therapy and systolic blood pressure. Stat Med. 2005;24:2911–35. doi: 10.1002/sim.2165. [DOI] [PubMed] [Google Scholar]
- 21.Golden SH, Dobs AS, Vaidya D, et al. Endogenous sex hormones and glucose tolerance status in postmenopausal women. J Clin Endocrinol Metab. 2007;92:1289–95. doi: 10.1210/jc.2006-1895. [DOI] [PubMed] [Google Scholar]
- 22.Kalish GM, Barrett-Connor E, Laughlin GA, et al. Association of endogenous sex hormones and insulin resistance among postmenopausal women: results from the Postmenopausal Estrogen/Progestin Intervention Trial. J Clin Endocrinol Metab. 2003;88:1646–52. doi: 10.1210/jc.2002-021375. [DOI] [PubMed] [Google Scholar]
- 23.Folsom AR, Golden SH, Boland LL, et al. Association of endogenous hormones with C-reactive protein, fibrinogen, and white blood count in post-menopausal women. Eur J Epidemiol. 2005;20:1015–22. doi: 10.1007/s10654-005-3657-0. [DOI] [PubMed] [Google Scholar]
- 24.Burger HG. The endocrinology of the menopause. Maturitas. 1996;23:129–36. doi: 10.1016/0378-5122(95)00969-8. [DOI] [PubMed] [Google Scholar]
- 25.Bryzgalova G, Lundholm L, Portwood N, et al. Mechanisms of antidiabetogenic and body weight-lowering effects of estrogen in high-fat diet-fed mice. Am J Physiol Endocrinol Metab. 2008;295:E904–12. doi: 10.1152/ajpendo.90248.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Dubey RK, Oparil S, Imthurn B, et al. Sex hormones and hypertension. Cardiovasc Res. 2002;53:688–708. doi: 10.1016/s0008-6363(01)00527-2. [DOI] [PubMed] [Google Scholar]
- 27.Marsh JD, Lehmann MH, Ritchie RH, et al. Androgen receptors mediate hypertrophy in cardiac myocytes. Circulation. 1998;98:256–61. doi: 10.1161/01.cir.98.3.256. [DOI] [PubMed] [Google Scholar]
- 28.Polderman KH, Gooren LJ, Asscheman H, et al. Induction of insulin resistance by androgens and estrogens. J Clin Endocrinol Metab. 1994;79:265–71. doi: 10.1210/jcem.79.1.8027240. [DOI] [PubMed] [Google Scholar]
- 29.Lovejoy JC, Bray GA, Bourgeois MO, et al. Exogenous androgens influence body composition and regional body fat distribution in obese postmenopausal women--a clinical research center study. J Clin Endocrinol Metab. 1996;81:2198–203. doi: 10.1210/jcem.81.6.8964851. [DOI] [PubMed] [Google Scholar]
- 30.Kupelian V, Page ST, Araujo AB, et al. Low sex hormone-binding globulin, total testosterone, and symptomatic androgen deficiency are associated with development of the metabolic syndrome in nonobese men. J Clin Endocrinol Metab. 2006;91:843–50. doi: 10.1210/jc.2005-1326. [DOI] [PubMed] [Google Scholar]
- 31.Ding EL, Song Y, Manson JE, et al. Sex hormone-binding globulin and risk of type 2 diabetes in women and men. N Engl J Med. 2009;361:1152–63. doi: 10.1056/NEJMoa0804381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Plymate SR, Matej LA, Jones RE, et al. Inhibition of sex hormone-binding globulin production in the human hepatoma (Hep G2) cell line by insulin and prolactin. J Clin Endocrinol Metab. 1988;67:460–4. doi: 10.1210/jcem-67-3-460. [DOI] [PubMed] [Google Scholar]
- 33.Niskanen L, Laaksonen DE, Punnonen K, et al. Changes in sex hormone-binding globulin and testosterone during weight loss and weight maintenance in abdominally obese men with the metabolic syndrome. Diabetes Obes Metab. 2004;6:208–15. doi: 10.1111/j.1462-8902.2004.00335.x. [DOI] [PubMed] [Google Scholar]
- 34.Pasquali R, Casimirri F, De Iasio R, et al. Insulin regulates testosterone and sex hormone-binding globulin concentrations in adult normal weight and obese men. J Clin Endocrinol Metab. 1995;80:654–8. doi: 10.1210/jcem.80.2.7852532. [DOI] [PubMed] [Google Scholar]
- 35.Laaksonen DE, Niskanen L, Punnonen K, et al. Sex hormones, inflammation and the metabolic syndrome: a population-based study. Eur J Endocrinol. 2003;149:601–8. doi: 10.1530/eje.0.1490601. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.