Androgens and Heart Failure: New Observations Illuminating an Aging Conundruma

Bu B. Yeap; Girish Dwivedi


J Clin Endocrinol Metab. 2021;106(3):e1476-e1478. 

Epidemiological studies generally associate lower endogenous testosterone concentrations in middle- and older-aged men with higher risks of cardiovascular disease and mortality.[1] Those studies often reported as outcomes the incidence of major atherosclerotic cardiovascular events such as fatal or nonfatal myocardial infarction or stroke, and cardiovascular deaths. However, causality remain unclear: Meta-analyses of randomized clinical trials of testosterone therapy have not shown a clear signal for cardiovascular adverse events,[2] while retrospective case–control studies of men prescribed testosterone therapy which associate treatment with lower mortality are vulnerable to bias and confounding.[3] Of concern, the Cardiovascular Sub-study of the Testosterone Trials reported an increase in coronary atheromatous plaque in men receiving testosterone, the caveat being that the testosterone and placebo groups were unbalanced with the latter having larger plaque volumes at baseline and study end.[4] Therefore, the influence of endogenous and exogenous androgens on cardiovascular risk remains unclear.[1]

Epidemiological studies of testosterone and its associations with cardiovascular disease in aging men have tended not to analyze heart failure (HF) as a distinct outcome. HF is being increasingly recognized as a common contributor to morbidity in middle-aged and older adults, with pathophysiologic, diagnostic, and therapeutic distinctions between HF with preserved ejection fraction (HFpEF, left ventricular ejection fraction ≥50%) or reduced ejection fraction (HFrEF, <50%). HFpEF is thought be a clinical syndrome in which comorbid conditions such as diabetes, obesity, and metabolic syndrome contribute to a sustained low-grade proinflammatory state, exacerbating pre-existing left ventricular and arterial stiffening.[5] It is commoner in women, and its contribution to HF disproportionately increase with age. Patients with HFrEF often have more readily modified risk factors (eg, smoking, hyperlipidemia) and a higher prevalence of atherosclerotic cardiovascular disease. The report by Zhao et al. sheds new light on androgens and cardiovascular risk.[6]

Participants in the Atherosclerosis Risk in Communities (ARIC) study were followed for a median of 19.2 years. Among 4107 men with mean baseline age 63.2 years, there were 873 new cases of hospitalizations or deaths due to HF, determined via annual surveillance telephone calls, review of hospital discharges, and death certificates. In fully adjusted analyses, lower total testosterone concentrations were associated with higher incidence of HF (hazard ratio [HR] = 1.10, 95% confidence interval [CI] = 1.03–1.17, per 1 standard deviation decrease in log-transformed total testosterone). A subgroup analysis was performed in men in whom new HF diagnoses were adjudicated as HFpEF or HFrEF, with diminished statistical power due to fewer outcome events in each analysis. Lower testosterone was associated with increased HFpEF risk in partially adjusted models, but not in the fully adjusted analysis.[6]

Among 4839 postmenopausal women with mean age 62.8 years from the ARIC study, there were 945 new HF cases. There was no association of testosterone concentrations with incidence of HF in women (HR = 1.05, CI = 0.99–1.13), nor with HFpEF or HFrEF in subgroup analyses.[6] However, the testosterone results used were obtained from immunoassay, not mass spectrometry. This is a limitation, particularly for the analysis of postmenopausal women, who in ARIC had mean testosterone of 0.8 nmol/L (compared with 17.7 nmol/L in men).

Of note, lower plasma dehydroepiandrosterone sulfate (DHEAS) concentrations were associated with higher incidence of HF in both men (HR = 1.07, CI = 1.00–1.15) and women (HR = 1.17, CI = 1.09=1.24), and HFpEF in women (HR = 1.12, CI = 1.00–1.25).[6] DHEAS is the sulfated form of dehydroepiandrosterone (DHEA) and is an abundant adrenal steroid, with relatively high circulating concentrations.[7] Conversion of adrenal androgens to testosterone makes a minor contribution to total androgen production in men, but a substantial contribution in women. Therefore, the association of lower DHEAS with higher incidence of HF and HFpEF in postmenopausal women suggests (but does not prove) a possible beneficial influence of androgen on the female heart. An alternative interpretation might be that DHEAS represents a proxy for activity of the hypothalamic–pituitary–adrenal axis. By contrast, plasma sex hormone-binding globulin concentrations were not associated with incidence of HF in either men or postmenopausal women.

Zhao et al., by implicating lower circulating testosterone concentrations with higher incidence of HF hospitalizations and deaths in men, add to existing epidemiological evidence suggesting a beneficial influence of testosterone on the cardiovascular system.[1] However, the effect size was modest, requiring extended follow-up of the ARIC cohort to accumulate large numbers of outcome events to demonstrate it.[6] A similarly modest association of lower DHEAS with higher incidence of HF was seen in men. Lower DHEAS was more prominently associated with incidence of HF in postmenopausal women, and also with HFpEF. Unanswered questions arise: Do plasma testosterone or DHEAS concentrations measured at baseline, provide a robust indicator of hormone exposure over the subsequent 19 years? Could residual confounding be present, even though analyses were adjusted for demographic and lifestyle variables, and for cardiovascular risk factors and disease? Are lower testosterone concentrations in men and DHEAS concentrations in women biomarkers for declining health in ageing? Are there convincing mechanistic explanations for the findings? Zhao et al. discuss potential effects of testosterone on atherogenesis, blood pressure, and cardiac myocytes, all capable of modulating HF risk. To prove causality, interventional studies in men to examine the effects of testosterone on each of these pathways are needed. Corresponding studies of DHEA or DHEAS in postmenopausal women would also be warranted.

A meta-analysis of trials in men with existing HF suggested an improvement in exercise capacity with testosterone treatment, acknowledging the limited numbers of trials and participants involved.[8] However, functional endpoints used in such trials would not necessarily distinguish cardiac or cardiovascular actions of testosterone from whole-body anabolic effects. Studies of DHEA or DHEAS have yielded inconsistent results, and generally have not addressed HF.[7] Thus, clarifying the effects of gonadal and adrenal androgens on the heart may be very relevant to ageing populations across the world, with HF events being of profound clinical importance. The plea by Zhao et al. for the ongoing TRAVERSE trial (NCT03518034) to incorporate adjudication for HF events should be supported. Their work adds to the justification and increases the scope for new, targeted mechanistic and clinical studies to examine the role of androgens to modulate multiple facets of cardiovascular risk.