Abstract
Background
Conflicting findings exist regarding the associations of sex hormones with subclinical atherosclerosis.
Methods
This is a substudy from MESA of 881 postmenopausal women and 978 men who had both abdominal aortic calcification (AAC) quantified by computed tomography and sex hormone levels assessed [Testosterone (T), estradiol (E2), dehydroepiandrosterone (DHEA), and sex hormone binding globulin (SHBG)]. We examined the association of sex hormones with presence and extent of AAC.
Results
For women, SHBG was inversely associated with both AAC presence [OR=0.62, 95% CI 0.42 to 0.91 for 1 unit greater log(SHBG) level] and extent [0.29 lower log(AAC) for 1 unit greater log(SHBG) level, β= −0.29 (95% CI −0.57 to −0.006)] adjusting for age, race, hypertension, smoking, diabetes, BMI, physical activity, and other sex hormones. After further adjustment for total and HDL-cholesterol, SHBG was not associated with ACC presence or extent. In men, there was no association between SHBG and AAC. In both men and women, neither T, E2, nor DHEA was associated with AAC presence or extent.
Conclusion
After adjustment for non-lipid cardiovascular risk factors, SHBG levels are inversely associated with both the presence and severity of AAC in women but not in men, which may be accounted for by HDL.
Keywords: sex hormone binding globulin, abdominal aortic calcification, sex hormones, subclinical atherosclerosis
Introduction
Differences in the prevalence of cardiovascular disease (CVD) between men and women across age groups suggest that sex hormones may influence the development of subclinical atherosclerosis. However, prior studies evaluating the association of sex hormones with subclinical atherosclerosis report conflicting results. Some studies found that total testosterone (T) levels are inversely associated with carotid intimal medial thickness (cIMT) in both postmenopausal women1 and men2 suggesting a protective effect. However a more recent analysis found both total T and bioavailable testosterone (bioT) were positively associated with cIMT in postmenopausal women after adjustment for risk factors, but no association was found for estradiol (E2) or dehydroepiandrosterone (DHEA).3 Several studies also found an inverse association of sex hormone binding globulin (SHBG) on cIMT.1,3
In women, estrogen replacement has been reported to be associated with less coronary artery calcification (CAC) progression4despite the lack of clinical benefit (and possible increased harm) of hormone therapy (HT) with CVD events.5 Among women with polycystic ovarian syndrome, those with prevalent CAC had higher levels of free T and lower levels of SHBG compared to those without CAC suggesting that increased androgens may promote calcification in women.6
Abdominal aortic calcification (AAC), another measure of subclinical atherosclerosis, is relatively common in older women even in the presence of low levels of coronary calcification. There is less of a gender difference between men and women with AAC than there is with CAC7 but the relationship of sex hormones with AAC is unclear. We sought to characterize the relationship with AAC and to evaluate whether any association differed by gender or race.
Methods
Participants
The Multi-Ethnic Study of Atherosclerosis (MESA) is a prospective cohort study investigating subclinical atherosclerosis in 6814 individuals aged 45–84 years without clinical CVD at baseline. Individuals were excluded if they had clinical CVD, including physician-diagnosed myocardial infarction, angina, stroke, transient ischemic attack, or heart failure, use of nitroglycerine, current atrial fibrillation, or had undergone a procedure related to CVD (coronary artery bypass surgery, angioplasty, valve replacement, pacemaker or defibrillator implantation, any surgery on the heart or arteries).
This report includes a random sample of MESA participants who also participated in the MESA Abdominal Aortic Calcium Study (MESA-AACS). MESA-AACS participants were recruited during follow-up visits between August 2002 and September 2005 from five MESA field centers: Chicago, Illinois, Forsyth County, North Carolina, Los Angeles County, California, Columbia University, and St. Paul, Minnesota. Of 2202 MESA participants recruited, 2172 agreed to participate, and 1990 satisfied eligibility criteria. Ten individuals had missing or incomplete scans for a total of 1980 participants with completed computed tomography scanning of their abdominal aorta.
We further restricted our analyses to postmenopausal women (n=881) and men (n=978) who participated in overlapping ancillary studies evaluating both AAC and serum sex hormone levels. The present analysis was based on data obtained at the baseline visit. Further details about the MESA study design have been published elsewhere8 and are available on the World Wide Web at www.mesa-nhlbi.org.
Risk Factor Assessment
Standardized questionnaires at the baseline examination were used to obtain information about participant demographics, medical history, and medication usage including current HT and cholesterol-lowering medications. Height, weight, and blood pressure were measured at the baseline examination. Body mass index (BMI) was calculated as weight (kg)/ height (m2). Resting blood pressure (BP) was measured three times in the seated position using a Dinamap automated sphygmomanometer, and the average of the 2nd and 3rd readings was used. Blood samples were obtained after a 12-hour fast to measure glucose, total cholesterol, HDL-cholesterol (HDLc), triglycerides, T, bioT, E2, DHEA, and SHBG. LDL-cholesterol (LDLc) was calculated using the Friedewald equation.
Hypertension was defined as systolic BP≥140 mm Hg, and/or diastolic BP≥90 mm Hg, and/or use of anti-hypertensive medications. Diabetes was classified as having a fasting blood glucose≥126 mg/dl and/or the self-reported use of hypoglycemic medications. Smoking was classified by never, former, or current use of cigarettes. Physical activity was calculated as the total minutes of all moderate and vigorous activity multiplied by their respective metabolic equivalents (METs).
Sex Hormone Assessment
Serum hormone concentrations (nmol/L) were measured from stored samples in the Steroid Hormone laboratory at the University of Massachusetts Medical Center in Worcester, MA. Total T and DHEA were measured directly using radioimmunoassay kits, and SHBG was measured by chemiluminescent enzyme immunometric assay using Immulite kits obtained from Diagnostic Products Corporation (Los Angeles, CA). E2 was measured by use of an ultra-sensitive radioimmunoassay kit from Diagnostic System Laboratories (Webster, TX). Percent bioT was calculated according to the method of Södergard et al9, and bioT concentration was calculated as (total T × (percent free T × 0.01)). Assay variability was monitored by including approximately 5% blind quality control samples in each batch of samples analyzed. The quality control serum was obtained from a large pool that was aliquoted into storage vials and labeled identical to MESA participant samples. The overall coefficients of variation for total T, SHBG, DHEA, and E2 were 12.3%, 9.0%, 11.2%, and 10.5%, respectively.
Subclinical Vascular Disease Assessment
Computed tomography of the abdomen was performed a single time for each individual in the MESA-AACS ancillary study. For electron-beam computed tomography (Chicago and Los Angeles; Imatron C-150), scanners were set as follows: scan collimation of 3mm; slice thickness of 6mm; reconstruction using 25 6mm slices with 35cm field of view and normal kernel. For multi-detector computed tomography mode (New York, Forsyth County, and St. Paul field centers; Sensation 64, GE Lightspeed, Siemens S4+ Volume Zoom and Siemens Sensation 16), images were reconstructed in a 35cm field of view with 5mm slice thickness. All scans were brightness adjusted with a standard phantom.
Scans were read centrally by the MESA CT Reading Center, and calcium in an 8cm segment of the distal abdominal aorta ending at the aortic bifurcation was scored. Calcification was identified as a plaque of ≥1mm2 with a density of ≥130 Hounsfield units and quantified using the previously described Agatston scoring method.10 This scoring method has previously shown to be highly reproducible for CAC.11,12
Statistical methods
For all analyses, we evaluated men and women separately because sex hormone levels differ by gender and there is insufficient overlap in levels to test for differences; thus the association of sex hormones on AAC may also differ by gender. Sex hormone levels did not follow a normal distribution, so they were log-transformed to make the distribution more symmetric.
We tabulated the distribution of the clinical characteristics of the study participants by the presence or absence of AAC, and tested differences between groups using chi-squared tests for categorical variables and t-tests for normally distributed variables or log-transformed variables. Using the direct age-standardization method, we also age-adjusted the baseline characteristics by presence or absence of AAC using the whole MESA study population’s age distribution as the standard population.13
We used regression methods to determine the association of each sex hormone with both the presence of AAC and the extent of AAC for those in whom AAC was present. The association of each log(sex hormone) variable with presence and extent of AAC was analyzed singly and then in models including all sex hormones. Because bioT is a calculated value based on total T, it was not included in models including all sex hormones.
For the presence of AAC, we used logistic regression to determine the odds of having any prevalent detectable aortic calcium (Agatston score >0) vs the absence of aortic calcium (Agatston score=0) for each 1 unit greater log(sex hormone) level. This was repeated in multivariate regression after adjusting for covariates of age, race/ethnicity, BMI, former and current smoking, systolic BP, diastolic BP, use of BP medications, diabetes, HT use (women only), and physical activity [Model 1] and after additionally adjusting for total cholesterol, HDLc, and cholesterol-lowering medication use [Model 2].
For the severity of aortic calcium among those that have detectable AAC (>0), non-zero calcium levels were log-transformed, and the association of log(AAC) with log(sex-hormones) was analyzed using linear regression both unadjusted and then adjusted for covariates in Models 1 and 2. We also examined whether the associations between sex hormones and AAC differed by race/ethnicity or HT in models with terms for interaction effects
Scatterplots were created between log(sex hormones) and log(AAC) to verify that the linearity assumption of our models was not violated. Since sex hormones are known to be correlated with each other, variance inflation factors (VIF) and tolerances were tested to ensure that none of the models had VIFs that approached a level of concern (i.e. VIF>10 or tolerances<10%). Stata 9.0 (Stata Corp., College Station, TX) was used for all analyses.
Results
Abdominal aortic calcium was present in 73.1% of men and 73.8% of the women (p=0.74 for the comparison of prevalent AAC between gender). Mean age in women without and with AAC was 58±7 and 66±9 years respectively (p<0.001). Mean age in men without and with AAC was 54±8 and 64± 10 years respectively (p<0.001). Because age is so strongly associated with the prevalence of AAC, age-adjusted clinical characteristics are presented in Table I. After adjusting for age, men and women without AAC were less likely to have hypertension, diabetes, ever-smoked, or used cholesterol-lowering medications, had higher HDLc and lower total cholesterol, LDLc, and triglycerides. There were 301 women on current HT, but HT use did not differ significantly between those without AAC and those with prevalent AAC. Age-adjusted geometric mean sex hormone levels by gender and presence of AAC are shown in Table 2. After adjusting for age only, there was a trend for higher SHBG levels among the women without AAC.
Table 1.
Age-Adjusted Baseline Characteristics by Gender and Presence of AAC, the Multi-Ethnic Study of Atherosclerosis, 2000–2002
| Mean values ± standard error or percent distribution | |||||||
|---|---|---|---|---|---|---|---|
| Women No AAC (N=231) |
Women AAC >0 (N=650) |
p-value | Men No AAC (N=263) |
Men AAC>0 (N=715) |
p-value | ||
| Race/Ethnicity (%) | |||||||
| Caucasian | 26.9 | 42.4 | 23.7 | 46.6 | |||
| Chinese | 12.3 | 13.4 | 0.002 | 8.9 | 14.1 | <0.001 | |
| African-American | 30.4 | 19.6 | 35.7 | 14.6 | |||
| Hispanic | 30.5 | 24.6 | 31.7 | 24.7 | |||
| BMI (kg/m2) | 28.0± 0.4 | 28.6±0.23 | 0.27 | 27.6±0.3 | 28.0±0.2 | 0.31 | |
| Hypertension (%) | 35.6 | 55.1 | <0.001 | 29.3 | 45.8 | 0.002 | |
| Systolic BP (mm Hg) | 125.0 ±1.5 | 129.6 ±0.9 | 0.011 | 122.7 ± 1.3 | 127.6 ± 0.7 | 0.001 | |
| Diastolic BP (mm Hg) | 68.7 ±0.7 | 69.3± 0.4 | 0.49 | 74.3 ± 0.6 | 76.0 ±0.4 | 0.019 | |
| Use of BP meds (%) | 23.5 | 44.5 | <0.001 | 21.7 | 35.9 | 0.009 | |
| Diabetes (%) | 10.5 | 13.4 | 0.03 | 11.9 | 14.5 | 0.43 | |
| Never Smoker (%) | 68.9 | 54.3 | <0.001 | 58.8 | 34.6 | <0.001 | |
| Former Smoker | 24.0 | 31.2 | 34.6 | 48.5 | |||
| Current Smoker | 7.1 | 14.6 | 6.6 | 16.9 | |||
| Total cholesterol (mg/dl) | 193.8±2.3 | 204.6± 1.3 | <0.001 | 186.1±2.2 | 192.3±1.3 | 0.02 | |
| HDLc (mg/dl) | 61.1±1.1 | 55.0±0.6 | <0.001 | 47.3±0.8 | 44.5±0.5 | 0.003 | |
| LDLc (mg/dl) | 110.8±2.1 | 120.9±1.2 | <0.001 | 114.7± 2.0 | 120.2± 1.2 | 0.023 | |
| Triglycerides (mg/dl) | 110.7± 5.4 | 143.7± 3.1 | <0.001 | 122.0±5.4 | 141.3±3.2 | 0.003 | |
| Use of cholesterol medications (%) | 7.9 | 22.2 | <0.001 | 6.6 | 15.1 | 0.03 | |
| Use of HT (%) | 32.7 | 35.4 | 0.38 | N/A | N/A | N/A | |
| Physical activity | 3248.5 | 3422.1 | 0.527 | 4730.9 | 3992.2 | 0.039 | |
| (MET min/week)* | (2834.3–3723.1) | (3163.8–3701.6) | (4137.8–5408.9) | (3694.9–4313.4) | |||
Geometric mean after log transformation, 95% CI
Table 2.
Age-Adjusted Geometric Mean Levels (95% CI) of Sex Hormones by Gender and Presence of AAC, the Multi-Ethnic Study of Atherosclerosis, 2000–2002
| Women No AAC (N=231) |
Women AAC >0 (N=650) |
p-value | Men No AAC (N=263) |
Men AAC>0 (N=715) |
p-value | |
|---|---|---|---|---|---|---|
| Total T (nmol/L) |
0.84 (0.77–0.92) |
0.86 (0.82–0.91) |
0.69 | 14.09 (13.29–14.94) |
13.79 (13.33–14.27) |
0.55 |
| BioT (nmol/L) |
0.19 (0.17–0.21) |
0.20 (0.19–0.22) |
0.21 | 5.22 (4.93–5.54) |
5.10 (4.93–5.27) |
0.50 |
| E2 (nmol/L) |
0.089 (0.078–0.10) |
0.088 (0.082–0.095) |
0.94 | 0.11 (0.10–0.12) |
0.108 0.104–0.110) |
0.50 |
| DHEA (nmol/L) |
10.49 (9.70–11.34) |
9.89 (9.45–10.34) |
0.21 | 12.97 (12.26–13.72) |
12.33 (11.94–12.74) |
0.14 |
| SHBG (nmol/L) |
67.15 (61.57–73.24) |
61.03 (58.05–64.16) |
0.07 | 40.63 (38.73–42.62) |
40.18 (39.08–41.30) |
0.70 |
SHBG levels were correlated with the other sex hormones of total T, E2, and DHEA in women (age-adjusted correlation coefficient [r] of −0.15, 0.37, −0.31, respectively, p<0.0001 for all) and with testosterone in men (r=0.42, p<0.0001). Because sex hormone levels are correlated with each other, but are not collinear, we also examined the independent effects of each sex hormone in models adjusted for all of the other sex hormones in addition to covariates. Scatterplots showed no non-linearity in the association between log(sex hormones) and log(AAC).
For postmenopausal women, Model 1 (adjusted for age, race/ethnicity, BMI, former and current smoking, systolic and diastolic BP, use of hypertension medications, diabetes, physical activity, and HT use) showed a significant inverse association of SHBG with AAC presence (Table 3a) and an inverse trend with AAC extent (Table 4a). There also was a statistically significant inverse association of SHBG with both AAC presence and extent in models additionally adjusted for the other sex hormones [OR 0.62 (95% CI 0.42 to 0.91) for each 1 unit greater log(SHBG) level] [Table 3a] and extent [0.29 lower log(AAC) for every 1 unit greater log(SHBG) level, β= −0.29 (95% CI −0.57 to −0.006)] [Table 4b]. However, after further adjustment for the total cholesterol, HDLc, and lipid medication use (Model 2), SHBG was no longer independently associated with either presence [Table 3a] or extent [Table 4b] of AAC in women. Also, the associations of any sex hormones with AAC were not significantly heterogeneous across race/ethnicity strata, nor was there heterogeneity for the relationship between SHBG and AAC by HT status in women.
Table 3.
Prevalence Odds of AAC by Sex Hormones in Women (3A) and Men (3B): The Multi-Ethnic Study of Atherosclerosis 2000–2002
Per 1 unit greater log(sex hormone) level
| A. | ||||||
|---|---|---|---|---|---|---|
| Women (Model 1) | Women (Model 2) | |||||
| Sex Hormones analyzed separately plus covariates | ||||||
| OR | 95% CI | p-value | OR | 95% CI | p-value | |
| T | 0.96 | 0.73–1.28 | 0.80 | 1.01 | 0.75–1.36 | 0.95 |
| E2 | 0.84 | 0.66–1.07 | 0.15 | 0.83 | 0.65–1.08 | 0.17 |
| DHEA | 0.76 | 0.54–1.08 | 0.13 | 0.76 | 0.53–1.10 | 0.14 |
| SHBG | 0.63 | 0.44–0.92 | 0.016 | 0.91 | 0.61–1.37 | 0.66 |
| Sex Hormones adjusted for each other plus covariates | ||||||
| T | 1.07 | 0.78–1.45 | 0.69 | 1.10 | 0.80–1.52 | 0.55 |
| E2 | 0.92 | 0.71–1.19 | 0.54 | 0.88 | 0.67–1.15 | 0.34 |
| DHEA | 0.72 | 0.49–1.06 | 0.10 | 0.75 | 0.50–1.12 | 0.17 |
| SHBG | 0.62 | 0.42–0.91 | 0.014 | 0.90 | 0.59–1.37 | 0.63 |
| B. | ||||||
|---|---|---|---|---|---|---|
| Men (Model 1) | Men (Model 2) | |||||
| Sex Hormones analyzed separately plus covariates | ||||||
| OR | 95% CI | p-value | OR | 95% CI | p-value | |
| T | 0.93 | 0.59–1.46 | 0.76 | 1.04 | 0.66–1.62 | 0.87 |
| E2 | 0.89 | 0.56–1.43 | 0.64 | 0.86 | 0.53–1.39 | 0.54 |
| DHEA | 0.93 | 0.62–1.41 | 0.75 | 0.97 | 0.64–1.48 | 0.88 |
| SHBG | 1.03 | 0.62–1.70 | 0.91 | 1.33 | 0.79–2.25 | 0.28 |
| Sex Hormones adjusted for each other plus covariates | ||||||
| OR | 95% CI | p-value | OR | 95% CI | p-value | |
| T | 0.95 | 0.56–1.61 | 0.84 | 0.98 | 0.57–1.69 | 0.94 |
| E2 | 0.92 | 0.56–1.52 | 0.75 | 0.86 | 0.51–1.43 | 0.56 |
| DHEA | 0.97 | 0.63–1.48 | 0.88 | 1.04 | 0.67–1.60 | 0.87 |
| SHBG | 1.07 | 0.61–1.89 | 0.80 | 1.39 | 0.77–2.52 | 0.27 |
*Model 1: adjusted for age, race, BMI, systolic BP, diastolic BP, use of BP medications, diabetes, former and current smoking, HT use (women only), and physical activity
*Model 2: adjusted for all Model 1 variables, total cholesterol, HDLc, and use of cholesterol-lowering medications
Table 4.
Severity of AAC by Sex Hormones in Women (4A) and Men (4B): The Multi-Ethnic Study of Atherosclerosis, 2000–2002
β=log-unit difference in AAC per 1 log-unit greater sex hormone level
| A. | ||||||
|---|---|---|---|---|---|---|
| Women (Model 1) | Women (Model 2) | |||||
| Sex Hormones analyzed separately plus covariates | ||||||
| β | 95% CI | p-value | β | 95% CI | p-value | |
| T | −0.101 | −0.31–0.11 | 0.35 | −0.081 | −0.29–0.13 | 0.44 |
| E2 | −0.071 | −0.26–0.11 | 0.45 | −0.091 | −0.27–0.09 | 0.32 |
| DHEA | −0.153 | −0.39–0.085 | 0.21 | −0.167 | −0.40–0.07 | 0.17 |
| SHBG | −0.26 | −0.53–0.015 | 0.064 | −0.155 | −0.44–0.13 | 0.28 |
| Sex Hormones adjusted for each other plus covariates | ||||||
| T | −0.048 | −0.27 – 0.18 | 0.67 | −0.026 | −0.25 – 0.20 | 0.82 |
| E2 | −0.027 | −0.21 – 0.17 | 0.79 | −0.054 | −0.24 – 0.14 | 0.58 |
| DHEA | −0.17 | −0.43 – 0.099 | 0.22 | −0.16 | −0.43 – 0.10 | 0.23 |
| SHBG | −0.29 | −0.57 – −0.006 | 0.045 | −0.18 | −0.46 – 0.11 | 0.23 |
| B. | ||||||
|---|---|---|---|---|---|---|
| Men (Model 1) | Men (Model 2) | |||||
| Sex Hormones analyzed separately plus covariates | ||||||
| β | 95% CI | p-value | β | 95% CI | p-value | |
| T | −0.112 | −0.36–0.14 | 0.38 | −0.064 | −0.32–0.18 | 0.61 |
| E2 | −0.092 | −0.38–0.19 | 0.53 | −0.119 | −0.40–0.16 | 0.41 |
| DHEA | −0.0005 | −0.28–0.27 | 0.997 | 0.0322 | −0.24–0.30 | 0.82 |
| SHBG | −0.061 | −0.39–0.27 | 0.72 | 0.0478 | −0.28–0.38 | 0.78 |
| Sex Hormones adjusted for each other plus covariates | ||||||
| β | 95% CI | p-value | β | 95% CI | p-value | |
| T | −0.095 | −0.38 – 0.18 | 0.50 | −0.06 | −0.34 – 0.22 | 0.67 |
| E2 | −0.043 | −0.35 – 0.26 | 0.78 | −0.096 | −0.40 – 0.21 | 0.53 |
| DHEA | 0.0093 | −0.27 – 0.29 | 0.95 | 0.055 | −0.22 – 0.33 | 0.70 |
| SHBG | −0.021 | −0.37 – 0.33 | 0.91 | 0.081 | −0.27 – 0.44 | 0.65 |
*Model 1: adjusted for age, race, BMI, systolic BP, diastolic BP, use of BP meds, diabetes, former and current smoking, HT use (women only), and physical activity
*Model 2: adjusted for all Model 1 variables, total cholesterol, HDLc, and use of cholesterol-lowering medications
In men, we found no association of SHBG with either presence or severity of AAC in Models 1 or 2. There were no independent associations of T, E2, or DHEA with presence or extent of AAC in either men or women in Models 1 or 2 [Table 3&Table 4]. No significant interactions were found between any of the sex hormones and race/ethnicity with AAC in men either.
Serum SHBG levels were directly associated with HDLc levels (age-adjusted correlation coefficient [r]=0.44 in women, r=0.24 in men, p<0.0001 for both). Greater total cholesterol, lower HDLc, and use of cholesterol medications were all strongly significantly associated with both the presence of AAC and extent of AAC in both women and men in our multivariate model. The lack of association of SHBG with AAC in women after adjusting for lipids seems predominantly mediated by HDLc. After adjustment for total cholesterol only (plus Model 1 variables but not HDLc or lipid medication usage), SHBG was still significantly associated with a reduced presence (OR 0.63, p=0.018) and severity (β= −0.29, p=0.044) of AAC in women. However, after adjusting for HDLc only (plus Model 1 variables but not total cholesterol or cholesterol-medication usage), SHBG was no longer significantly associated with AAC in women, although direction was still the same (OR 0.76, p=0.19 for presence and β= −0.24, p=0.10) for severity. This suggests that HDLc levels may have accounted for the association between SHBG and AAC in women.
Discussion
Unlike the gender differences seen with CAC, our study confirms prior findings7 that there is no significant difference in prevalence of AAC between men and women of similar ages in this multi-ethnic cohort of individuals free of clinical CVD. Abdominal aortic calcification is common in these asymptomatic individuals, and is associated with traditional CVD risk factors such as age, smoking, hypertension, diabetes, and adverse lipid profiles.
We did not find an association of T, E2, or DHEA levels with either presence of extent of AAC in either men or women. However, in postmenopausal women only, there were inverse associations of serum SHBG level with both the presence of AAC and with AAC severity. While the inverse association of SHBG with AAC severity did not quite meet statistical significance in Model 1 when the sex hormones were analyzed separately, the magnitude and direction of the SHBG coefficient, which is the best measure of association strength, was almost identical to the SHBG coefficient in the models adjusted for the other sex hormones. Then, the inverse associations of SHBG with AAC presence and severity were confirmed to be statistically-significantly independent of T and the other sex hormones in fully-adjusted models including diabetes and BMI (which both correlate with SHBG), as well as other major non-lipid CVD risk factors. However, after adjusting for HDLc, there was no significant association of SHBG with either the presence or extent of AAC, suggesting that HDLc levels may explain or confound the protective association of SHBG with AAC.
Sex hormone binding globulin is a glycoprotein synthesized in the liver that binds 17-β-hydroxysteriods such as T and E2 with high affinity.14 SHBG levels are strongly correlated with T levels; however SHBG is negatively correlated with glucose and insulin levels independently of T levels.15,16,17 Low levels of SHBG has been shown to be strongly and consistently related to elevated CVD risk factors of higher insulin, glucose, hemostatic and inflammatory markers, and adverse lipids even after controlling for BMI, 18 but not associated with systolic or diastolic blood pressure.19 Low SHBG levels have been shown to be associated with the metabolic syndrome in postmenopausal women.20 Levels of SHBG in the lowest quartile were also associated with a 2-fold age-adjusted risk of CVD; yet this association was no longer seen after adjustment for obesity, hypertension, diabetes, and lipids, suggesting that the increased CVD risk associated with low SHBG was attributable to the metabolic syndrome.20 Racial differences between Caucasians and African Americans in the association of SHBG with adverse metabolic profile have been reported,21 although our study did not find any racial interactions between SHBG and the presence or severity of AAC.
Many prior cross-sectional studies have found an independent association of SHBG with lipoprotein levels in both men and postmenopausal women with the most consistent finding of a positive association between SHBG and HDLc.22,23,24,25 We also found that SHBG is directly associated with HDLc. Hepatic production of SHBG is regulated by many influences including sex steroid levels, thyroid hormones, cortisol, and growth factors and is inhibited by insulin.14,26,27 Many of these metabolic factors may also influence HDLc levels. For example, hyperinsulinism might negatively regulate both SHBG and HDLc levels. Also, physical activity and HDLc are positively correlated.
SHBG has been postulated to influence atherogenesis through metabolism and/or production of HDLc either directly or indirectly through the estrogen-testosterone balance.28 Increasing levels of HDL are associated with less subclinical atherosclerosis independent of LDL levels,29 and the positive association of SHBG levels with HDLc may be one mechanism that explains the inverse relationship of SHBG with subclinical atherosclerosis. To our knowledge, there are no current prospective studies that document SHBG increases HDL in a causal fashion, but such a pathway is now well established for alcohol which reduces CVD risk through HDLc (in addition to other factors) as a mediator.30 Our study is limited by the cross-sectional design, thus the causality and temporality of the association of SHBG with HDLc and AAC cannot be determined.
In conclusion, SHBG levels are inversely associated with both the presence and severity of AAC in women but not in men after adjustment for non-lipid CVD risk factors. Levels of T, E2, and DHEA, however, do not appear associated with AAC in either gender. The association of SHBG with AAC in women was largely accounted for by HDLc levels. Further prospective studies with baseline measures of SHBG and serial measurements of AAC and/or serum lipids are needed to determine the temporal relationship of SHBG with HDLc and subclinical atherosclerosis.
Acknowledgements
The authors thank the other investigators, the staff, and the participants of the MESA study for their valuable contributions. This research was supported by NHLBI RO1 HL074406, RO1 HL074338, R01 HL72403, and contracts N01-HC-95159 through N01-HC-95165 and N01-HC-95169 from the National Heart, Lung, and Blood Institute. Dr. Michos is also funded by the PJ Schafer Foundation Preventive Cardiology award. A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org.
Footnotes
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.
Disclosures: None
REFERENCES
- 1.Golden SH, Maguire A, Ding J, Crouse JR, Cauley JA, Zacur H, Szklo M. Endogenous postmenopausal hormones and carotid atherosclerosis: a case-control study of the Atherosclerosis Risk in Communities Cohort. Am J Epidemiol. 2002;155:437–445. doi: 10.1093/aje/155.5.437. [DOI] [PubMed] [Google Scholar]
- 2.Svartberg J, von Muhlen D, Mathiesen E, Joakimsen O, Bonaa KH, Stensland-Bugge E. Low testosterone levels are associated with carotid atherosclerosis in men. J Intern Med. 2006;259:576–582. doi: 10.1111/j.1365-2796.2006.01637.x. [DOI] [PubMed] [Google Scholar]
- 3.Ouyang P, Vaidya D, Dobs AS, Golden SH, Szklo M, Liu K, Heckbert SR, Kopp P, Gapstur SM. Endogenous sex hormone levels are associated with intima-medial thickness in postmenopausal women in MESA: Multi-Ethnic Study of Atherosclerosis. Circulation [Abstract] 2006;113:#5. [Google Scholar]
- 4.Budoff MJ, Chen GP, Hunter CJ, Takasu J, Agrawal N, Sorochinsky B, Mao S. Effects of hormone replacement on progression of coronary calcium as measured by electron beam tomography. J Womens Health. 2005;14:410–417. doi: 10.1089/jwh.2005.14.410. [DOI] [PubMed] [Google Scholar]
- 5.Women’s Health Initiative Investigators. Risk and benefits of estrogen plus progestin in healthy postmenopausal women: principle results from the Women’s Health Initiative randomized controlled trial. JAMA. 2002;288:321–333. doi: 10.1001/jama.288.3.321. [DOI] [PubMed] [Google Scholar]
- 6.Christian RC, Dumesic DA, Behrenbeck T, Oberg AL, Sheedy PF, 2nd, Fitzpatrick LA. Prevalence and predictors of coronary artery calcification in women with polycystic ovarian syndrome. J Clin Endocrinol Metab. 2003;88:2562–2568. doi: 10.1210/jc.2003-030334. [DOI] [PubMed] [Google Scholar]
- 7.Post W, Bielak LF, Ryan KA, Chen YC, Shen H, Rumberger JA, Sheedy PF, 2nd, Shuldiner AR, Peyser PA, Mitchell BD. Determinants of coronary artery and aortic calcification in the Old Order Amish. Circulation. 2007;115:717–724. doi: 10.1161/CIRCULATIONAHA.106.637512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bild DE, Bluemke DA, Burke GL, Detrano R, Diex Roux AV, Folsom AR, Greenland P, Jacob DR, Jr, Kronmal R, Liu K, Nelson JC, O’Leary D, Saad MF, Shea S, Szko M, Tracy RP. Multi-ethnic study of atherosclerosis: objectives and design. Am J Epidemiol. 2002;156:871–881. doi: 10.1093/aje/kwf113. [DOI] [PubMed] [Google Scholar]
- 9.Sodergard R, Backstrom T, Shanbhag V, Carstensen H. Calculation of free and bound fractions of testosterone and estradiol-17 beta to human plasma proteins at body temperature. J Steroid Biochem. 1982;16:801–810. doi: 10.1016/0022-4731(82)90038-3. [DOI] [PubMed] [Google Scholar]
- 10.Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M, Jr, Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol. 1990;15:827–832. doi: 10.1016/0735-1097(90)90282-t. [DOI] [PubMed] [Google Scholar]
- 11.Carr JJ, Nelson JC, Wong ND, McNitt-Gray M, Arad Y, Jacobs DR, Jr, Sidney S, Bild DE, Williams OD, Detrano RC. Calcified coronary artery plaque measurement with cardiac CT in population-based studies: standardized protocol of Multi-Ethnic Study of Atherosclerosis (MESA) and Coronary Artery Risk Development in Young Adults (CARDIA) study. Radiology. 2005;234:35–43. doi: 10.1148/radiol.2341040439. [DOI] [PubMed] [Google Scholar]
- 12.Detrano RC, Anderson M, Nelson J, Wong ND, Carr JJ, McNitt-Gray M, Bild DE. Coronary Calcium Measurements: Effect of CT Scanner Type and Calcium Measure on Rescan Reproducibility--MESA Study. Radiology. 2005;236:477–484. doi: 10.1148/radiol.2362040513. [DOI] [PubMed] [Google Scholar]
- 13.Bild DE, Detrano R, Peterson D, Guerci A, Liu K, Shahar E, Ouyang P, Jackson S, Saad MF. Ethnic difference in coronary calcification: the Multi-Ethnic Study of Atherosclerosis (MESA) Circulation. 2005;111:1313–1320. doi: 10.1161/01.CIR.0000157730.94423.4B. [DOI] [PubMed] [Google Scholar]
- 14.Selby C. Sex hormone binding globulin: origin, function, and clinical significance. Ann Clin Biochem. 1990;37:532–541. doi: 10.1177/000456329002700603. [DOI] [PubMed] [Google Scholar]
- 15.Tsai EC, Matsumoto AM, Fujimoto WY, Boyko EJ. Association of bioavailable, free, and total testosterone with insulin resistance: influence of sex hormone-binding globulin and body fat. Diabetes Care. 2004;27:861–868. doi: 10.2337/diacare.27.4.861. [DOI] [PubMed] [Google Scholar]
- 16.Haffner SM, Katz MS, Stern MP, Dunn JF. The relationship of sex hormones to hyperinsulinemia and hyperglycemia. Metabolism. 1988;37:683–688. doi: 10.1016/0026-0495(88)90091-1. [DOI] [PubMed] [Google Scholar]
- 17.Golden SH, Dobs AS, Vaidya D, Szklo M, Gapstur S, Kopp P, Liu K, Ouyang P. Endogenous sex hormones and glucose tolerance status in post-menopausal women. J Clin Endocrinol Metab. 2007;92:1289–1295. doi: 10.1210/jc.2006-1895. [DOI] [PubMed] [Google Scholar]
- 18.Sutton-Tyrrell K, Wildman RP, Matthews KA, Chae C, Lasley BL, Brockwell S, Pasternak RC, Lloyd-Jones D, Sower MF, Torrens JI. SWAN Investigators. 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;111:1242–1249. doi: 10.1161/01.CIR.0000157697.54255.CE. [DOI] [PubMed] [Google Scholar]
- 19.Haffner SM, Katz MS, Stern MP, Dunn JF. Association of decreased sex hormone binding globulin and cardiovascular risk factors. Arteriosclerosis. 1989;9:136–143. doi: 10.1161/01.atv.9.1.136. [DOI] [PubMed] [Google Scholar]
- 20.Weinberg ME, Manson JE, Buring JE, Cook NR, Seely EW, Ridker PM, Rexrode KM. Low sex hormone-binding globulin is associated with the metabolic syndrome in postmenopausal women. Metabolism. 2006;55:1473–1480. doi: 10.1016/j.metabol.2006.06.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Berman DM, Rodrigues LM, Nicklas BJ, Ryan AS, Dennis KE, Goldberg AP. Racial disparities in metabolism, central obesity, and sex hormone-binding globulin in postmenopausal women. J Clin Endocrinol Metab. 2001;86:97–103. doi: 10.1210/jcem.86.1.7147. [DOI] [PubMed] [Google Scholar]
- 22.Bataille V, Perret B, Evans A, Amouyel P, Arveiler D, Ducimetiere P, Bard JM, Ferrieres J. Sex hormone-binding globulin is a major determinant of the lipid profile: the PRIMA study. Atherosclerosis. 2005;179:369–373. doi: 10.1016/j.atherosclerosis.2004.10.029. [DOI] [PubMed] [Google Scholar]
- 23.Haffner SM, Mykkanen L, Valdez RA, Katz MS. Relationship of sex hormones to lipids and lipoproteins in nondiabetic men. J Clin Endocrinol Metab. 1993;77:1610–1615. doi: 10.1210/jcem.77.6.8263149. [DOI] [PubMed] [Google Scholar]
- 24.Haffner SM, Dunn JF, Katz MS. Relationship of sex hormone-binding globulin to lipid, lipoprotein, glucose, and insulin concentrations in postmenopausal women. Metabolism. 1992;41:278–284. doi: 10.1016/0026-0495(92)90271-b. [DOI] [PubMed] [Google Scholar]
- 25.Mudali S, Dobs AS, Ding J, Cauley JA, Szklo M, Golden SH. Endogenous postmenopausal hormones and serum lipids: the Atherosclerosis Risk in Communities Study. J Clin Endocrinol Metab. 2005;90:1202–1209. doi: 10.1210/jc.2004-0744. [DOI] [PubMed] [Google Scholar]
- 26.Loukovaara M, Carson M, Adlercreutz H. Regulation of sex-hormone binding globulin production by endogenous estrogens in vitro. Biochem Biophys Res Commun. 1995;206:895–901. doi: 10.1006/bbrc.1995.1127. [DOI] [PubMed] [Google Scholar]
- 27.Plymate SR, Hoop RC, Jones RE, Matej LA. Regulation of sex hormone-binding globulin production by growth factors. Metabolism. 1990;39:967–970. doi: 10.1016/0026-0495(90)90309-z. [DOI] [PubMed] [Google Scholar]
- 28.Pugeat M, Moulin P, Cousin P, Fimbel S, Nicolas MH, Crave JC, Lejeune H. Interrelations between sex hormone-binding globulin (SHBG), plasma lipoproteins, and cardiovascular risk. J Steroid Biochem Mol Biol. 1995;53:567–572. doi: 10.1016/0960-0760(95)00102-6. [DOI] [PubMed] [Google Scholar]
- 29.Allison MA, Wright CM. A comparison of HDL and LDL cholesterol for prevalent coronary calcification. Int J Cardiol. 2004;95:55–60. doi: 10.1016/j.ijcard.2003.04.013. [DOI] [PubMed] [Google Scholar]
- 30.Hojnacki JL, Cluette-Brown JE, Deschenes RN, Mulligan JJ, Osmolski TV, Rencricca NJ, Barboriak JJ. Alcohol produces dose-dependent antiatherogenic and atherogenic plasma lipoprotein responses. Proc Soc Exp Biol Med. 1992;200:67–77. doi: 10.3181/00379727-200-43395. [DOI] [PubMed] [Google Scholar]