Sex Steroids and Incident Cardiovascular Disease in Postmenopausal Women: New Perspective on an Old Controversy

Loss of ovarian hormones either naturally or surgically increases the aging processes in women including incidence, morbidity and mortality from cardiovascular (CV) disease (1–4). Most observational studies, which evaluated associations among endogenous levels of the sex steroids with CV risk and event incidence in postmenopausal women, indicate that risk and event incidence increase with decreases in endogenous estrogen. However, the association of endogenous testosterone and cardiovascular risk and event incidence in women is less clear. Several factors contribute to this lack of clarity including: associations of singular hormone levels rather than as a collective reflection of the complex interactions among steroid production, transformation, transport, binding, and elimination for each individual. In addition, the ages of women enrolled in those studies vary, and follow-up may not be adequate to identify statistically significant associations in small sample sizes. Also, identification of CV events is often not segregated as to type of event, which reflects the assumption that the hormonal contributions to development of an atherosclerotic lesion, coronary heart disease or heart failure is homogeneous. Differences in population demographics also are difficult to reconcile in regard to regional differences in body mass index, level of physical fitness, and environmental exposures in various geographical regions in which the populations reside [i.e. exposure to sunlight and potential vitamin D deficiencies in persons living in northern latitudes such as Denmark (5) and Germany (6) compared to Southern California (7)].

The paper by Zhao et al., reported in this issue of JACC (8), attends to these factors and provides new insight into relationships of endogenous hormones and incident CV events. Serum levels of estradiol, total testosterone, dehydroepiandrosterone, and sex hormone binding globulin were measured in 2,834 ethnically diverse postmenopausal women participating in the Multi-Ethnic Study of Atherosclerosis (MESA), which represents 6 geographically distinct urban and rural areas within the United States. Importantly, these women were free of cardiovascular disease upon entry into the study, although it is unclear if that assessment included only a coronary calcification score of zero (9). Also, self-reported cardiovascular events were confirmed with review of hospitalization reports and medical records with review and adjudication by two independent physicians. Two important methodologies differentiate this study from others: 1) the ratio of testosterone to estradiol (T:E) was considered in the analysis, and 2) coronary heart disease and heart failure with preserved (HFpEF) and with reduced ejection fraction (HFrEF) were evaluated separately from overall incident CV disease. A higher T:E level associated with increased incidence of overall CV events and coronary heart disease. These results might be expected given that this elevated ratio also associated with current smoking, obesity and C-reactive protein, all conventional risk factors for cardiovascular disease (10). Unexpectedly, the T:E ratio showed a U-shape association with incident heart failure, and with further analysis of subtypes of heart failure, a positive association with HFrEF but not with HFpEF. By addressing a set of defined incident events, this study provides new information needed to develop mechanistic hypotheses of causal relationships of hormones with specific aspects of cardiac function.

Few studies have gone beyond singular associations of hormone levels with CV events (5). It is possible to assess complex pathway interactions in biological systems. Given the complexity of endogenous steroid metabolism, is it time to use such analytic tools to evaluate hormonal profiles to identify women at risk for different types of CV disease, and to identify women who might benefit from hormone treatments for their prevention (11). Individual variability in hormone profiles among women may reflect genetic variation in one more of the enzymes involved in steroid metabolism, and, thus, the bioavailability of testosterone and estradiol (12). For example, multiple copies of a genetic variants for the enzyme SULT1A1, which sulfonates estradiol and estrone for elimination, are associated with an early age of menopause (13). These results suggest that women with multiple copies of this variant might experience reduced bioavailability of estradiol perhaps years before actual decreases in ovarian production of the hormone which puts these women on an accelerated path for aging and CV risks.

Local metabolism of testosterone to estrogen by aromatase in cardiac tissue (14) adds to the complexity of understanding hormonal modulation of cardiac function. Thus, tissue concentrations of the hormones may not reflect those measured in serum. Some studies have evaluated CV risk in women using aromatase inhibitors for breast cancer prevention (15). These studies need to be expanded to include variants in the aromatase gene with types of CV disease (large vessel or microvascular ischemic disease, coronary calcification, and types of heart failure) (16). Also, the gene for the androgen receptor is located on the X chromosome. Mosaic representation of this variant in cardiac tissue of women due to X inactivation further adds to the complexity of understanding how testosterone might influence cardiac function in women. Furthermore, effects of hormones on specific CV outcomes will reflect genetic variants in other systems involved with development of the disease, for example, genetic variants in innate immunity associated with development of carotid intima-media thickness and coronary calcification (17,18).

Experiments are needed to better understand the complex hormonal environment affecting cellular and organ functions involved in development and progression of CV disease in women as they age. Although observational studies such as the report by Zhao et al (8) cannot provide insights into mechanisms by which hormones modulate disease risk, they provide information that directs design of studies to examine mechanisms of atherosclerotic burden compared to coronary arterial calcification or microvascular disease or myocardial function. Defining cardiovascular risk for women should account for individualized profiles of genetic variants in enzymes associated with steroids metabolism, uptake and receptors in conjunction with risk for specific cardiovascular pathologies. This approach is precision medicine.

Biography