Recent studies have shown that women aged 70 years and older have blood testosterone (T) concentrations similar to those of premenopausal women [1,2,], and that circulating T concentrations tend to increase with age from the 8th decade, while concentrations of T’s adrenal precursor dehydroepiandrosterone (DHEA) decline [3,4]. This lead Islam and collaborators to investigate the associations between these sex steroids and cardiovascular disease (CVD) in a large cohort of women aged at least 70 years . They reported on first major adverse (ischaemic) cardiovascular events (MACE) in 5,535 women, mean age 74 years, followed-up for a median of 4.4 years. After adjusting for established risk factors for MACE and study treatment allocation, Islam et al.  reported that women with blood T and /or DHEA concentrations in the lowest quartile had almost a 2-fold greater risk of a first MACE compared with women with higher T/DHEA levels (i.e hazard ratios for the highest quartile versus the lowest quartile were 0.57 [95% CI, 0.36 to 0.91, p=0.02] for T and 0.61 [95% CI, 0.38 to 0.97, p=0.04] for DHEA. No meaningful findings for oestrone or sex hormone binding globulin (SHBG) were seen. Divergence of the cumulative hazard curves for MACE for the lowest quartiles of T and DHEA were statistically significantly different from the other higher quartiles by the third year of follow-up. The authors concluded that higher androgen blood levels are not deleterious to cardiovascular health in older women and that further studies of T and MACE in older women are needed.
Having higher androgen concentrations has been considered to be associated with a greater risk of cardiometabolic disease in postmenopausal women, primarily by extrapolation from studies of women with polycystic ovary syndrome (PCOS). However, available data for postmenopausal women with a prior diagnosis of PCOS do not support these assertions. Meun et al.
 found no association between high T levels and cardiovascular disease (CVD) outcomes or surrogate marker of atherosclerosis in women with and without prior PCOS . More recently, a small longitudinal, observational study reported that by the average age of 81 there were no differences in hypertension (67% vs. 66%), myocardial infarction (10% vs. 13%), angina pectoris (14% vs. 16%), stroke (10% vs. 19%), or transient ischemic attacks (TIAs) (5% vs. 6%) seen between women with past PCOS and a reference group . Furthermore, in postmenopausal women T appears to have favourable effects on vascular function [8,9]. Despite this, studies reporting associations between T and CVD in postmenopausal women have been inconclusive. A major limitation of past studies has been the measurement of T and other sex steroids by immunoassay which lack precision at the levels seen in women. Other past study limitations include sex steroids being measured many years before events occurred, and the use of convenience samples.
The observational study of Islam et al , of which I am a co-author, reported on findings from the ASPirin in Reducing Events in the Elderly (ASPREE) Study. This clinical trial recruited 9,180 Australian women, aged at least 70 years, with no prior CVD events, between 2010 and 2014. Women were randomly allocated to low dose aspirin or placebo and independently adjudicated endpoints were informed by source documents that included clinical notes, hospitalisation records, and imaging studies. It therefore provided a unique opportunity to explore the associations between sex steroids and MACE. Strengths of the study included its size, measurement of sex hormones by liquid chromatography-tandem mass spectrometry LC-MS/MS, exclusion of women known to be taking therapies that influence androgen production and blood levels, and the availability of independently adjudicated MACE. While the findings that both low T and low DHEA were associated with a greater risk of MACE in our study of women aged 70 and older are robust, they do not provide evidence of cause and effect.
There has been a resurgence of interest in T therapy for postmenopausal women beyond its use for low sexual desire. However, available data do not support the use of postmenopausal T therapy for any other indication, including disease prevention . Consistent with this, our study does not justify the use of either T or DHEA replacement in older women for the prevention of CVD. Rather, taken together with the observation that T concentrations measured by LC-MS/MS do not differ between women aged 70-95 years and women aged 18-39 years [1,2], this study signals that androgens may have an important physiological role in older women that needs further exploration.
Susan R. Davis AO, MBBS, FRACP, PhD, FAHMS
Professor of Women’s Health and NHMRC Senior Principal Research Fellow
Director, Women’s Health Research Program
School of Public Health and Preventive Medicine,
Monash University, Melbourne, Australia