Androgen therapy in men with testosterone deficiency: can testosterone reduce the risk of cardiovascular disease?


  • Farid Saad

    Corresponding author
    1. Global Medical Affairs Men's Healthcare, Bayer Pharma AG, Berlin, Germany
    2. Research Department, Gulf Medical University, Ajman, United Arab Emirates
    3. Men’s Health Reproduction Study Center, Hang Tuah University, Surabaya, Indonesia
      *Correspondence to: Farid Saad, Bayer Pharma AG, Muellerstr. 178, 13353 Berlin, Germany.
    Search for more papers by this author

*Correspondence to: Farid Saad, Bayer Pharma AG, Muellerstr. 178, 13353 Berlin, Germany.


Obesity, hypertension, insulin resistance (IR), dyslipidaemia, impaired coagulation profile and chronic inflammation characterize cardiovascular risk factors in men. Adipose tissue is an active endocrine organ producing substances that suppress testosterone (T) production and visceral fat plays a key role in this process. Low T leads to further accumulation of fat mass, thus perpetuating a vicious circle. In this review, we discuss reduced levels of T and increased cardiovascular disease (CVD) risk factors by focusing on evidence derived from three different approaches. (i) epidemiological/ observational studies (without intervention); (ii) androgen deprivation therapy (ADT) studies (standard treatment in advanced prostate cancer); and (iii) T replacement therapy (TRT) in men with T deficiency (TD). In epidemiological studies, low T is associated with obesity, inflammation, atherosclerosis and the progression of atherosclerosis. Longitudinal epidemiological studies showed that low T is associated with an increased cardiovascular mortality. ADT brings about unfavourable changes in body composition, IR and dyslipidaemia. Increases in fibrinogen, plasminogen activator inhibitor 1 and C-reactive protein have also been observed. TRT in men with TD has consistently shown a decrease in fat mass and simultaneous increase in lean mass. T is a vasodilator and in long-term studies, it was shown to reduce blood pressure. There is increasing evidence that T treatment improves insulin sensitivity and lipid profiles. T may possess anti-inflammatory and anti-coagulatory properties and therefore TRT contributes to reduction of carotid intima media thickness. We suggest that T may have the potential to decrease CVD risk in men with androgen deficiency.


Relationship between cardiovascular risk biomarkers and androgen deficiency

A host of biochemical and clinical markers are known to constitute increased risk of cardiovascular disease (CVD) [1]. These include high blood pressure (hypertension), insulin resistance (IR), increased body fat, altered lipids and coagulation profiles and increased inflammation, together contributing to the development and progression of atherosclerosis and ultimately heart disease. The role of testosterone (T) in cardiovascular health has been the subject of a number of recent studies, reviews, editorials and controversies [2–48] and the molecular bases of TD and the pathophysiology of vascular disease requires further research.

To address this vital medical and health issue adequately, we felt that data derived from studies using mainly TRT provided some relevant yet limited information, thus inclusion of evidence from other clinical (androgen deprivation therapy – ADT) and epidemiological studies will permit further assessment of the consistency of the information obtained from the various approaches. In this communication, we provide a concise discussion of the potential role of TRT in men with TD in reducing the risk of CVD. This discussion will be based on evidence derived from (i) epidemiological, (ii) ADT and (iii) TRT studies.

Evidence from epidemiological/observational studies

Data from epidemiological studies suggest a positive relationship between physiological T levels and vascular health, albeit indirect [6,7,16,35,38,43,47,49–54]. Low T is associated with hypertension, increased inflammation, increased prevalence of type 2 diabetes mellitus (T2DM) and atherosclerosis progression [1,4,49,51,52,55,56]. These studies also indicated that men with T2DM, obesity or coronary heart disease (CHD) all have reduced T levels. Svartberg et al. [49] investigated the relationship between total and free T levels and hypertension in 1548 men participating in the Tromsø Study. The authors noted a significant difference in total and free T levels in men with hypertension, after adjustments for age and body mass index (BMI). Wu et al. [50] investigated the relationship between T levels, age and BMI in men between the ages of 40 and 79 years in the European Male Aging Study. The data showed a relationship between T levels and BMI with reduced T levels in men with BMI ≥30 kg/m2. In cross-sectional studies, Svartberg et al. [51] demonstrated that reduced T levels are associated with carotid atherosclerosis in men. In longitudinal studies, Hak et al. [52] showed an independent and significant inverse association between bioavailable and total T levels and progression of aortic atherosclerosis in 1032 non-smoking men. Kapoor et al. [55] investigated the relationship between the non-specific inflammatory marker, C-reactive protein (CRP) and total T levels. The authors showed that increased CRP levels are correlated with reduced T levels. A significant number of men with coronary artery disease (CAD) exhibit TD (hypogonadism) with approximately 23% of men with T levels less than 7.5 nm and 53% with T levels less than 12 nm (P J Pugh, K S Channer, Department of Cardiology, Royal Hallamshire Hospital, Sheffield, UK; T H Jones, Academic Unit of Endocrinology, Division of Genomic Medicine, University of Sheffield Medical School, Sheffield, UK, South Yorkshire Study, personal communication). These findings suggest that in men with CAD, reduced T levels may contribute to the progression of the disease. Kaplan et al. [57] examined baseline data from a lipid treatment in 467 men to assess the relationship between T concentrations and high sensitivity C-reactive protein (hsCRP), a cardiovascular risk marker. The authors noted that hsCRP levels were highest in patients in the lower T levels, suggesting an inverse relationship between serum T and hsCRP in aging men. Lin et al. [53] noted an association between reduced T and increased cardiovascular risk in men aged 40 years and older in a study of more than 500 men. In a recent systematic review and meta-analysis study of all-cause mortality [(16,184 subjects); studies of CVD mortality (11,831 subjects)] with mean follow-up time of 9.7 years, Araujo et al. [54] concluded that low T levels are associated with increased risk of all cause and CVD mortality in community-based men.

Evidence from ADT studies

Androgen deprivation therapy has been the mainstay therapy for androgen-dependent prostate cancer (PCa). Thus, surgical or medical castration (GnRH analogues or antagonists) were among the approaches used in ADT. Significant changes in body composition such as decrease in lean mass and increase in fat mass were noted 12 weeks post-ADT [58]. These observations suggest that ADT alters fat and muscle metabolism and produces imbalance in body composition with undesirable consequences. For instance, within the 12 week time period post-ADT, a marked increase in insulin levels (63%) was noted, suggesting increased IR [59]. Data from other studies have suggested that ADT is associated with an earlier onset of myocardial infarction (MI) [60]. ADT was associated with increased all-cause mortality in patients with a history of congestive heart failure or MI, who were previously treated with brachytherapy [61,62]. In an observational study of 37 443 population-based men who were diagnosed with local or regional PCa, Nguyen et al. [61,62] reported that treatment with GnRH agonists was associated with statistically significantly increased risks of incident diabetes, incident CHD, MI, sudden cardiac death and stroke. The data suggested that ADT with GnRH agonists is associated with increased risk of diabetes and CVD [61,62].

Keating et al. [63,64] suggested that GnRH agonist treatment for men with locoregional PCa may be associated with an increased risk of incident diabetes and CVD. ADT with GnRH agonists was associated with an increased risk of T2DM and CVD. In patients undergoing radical prostatectomy for localized PCa, ADT appears to be associated with increased risk of death from CVD causes [65].

Haidar et al. [66] carried out a retrospective analysis of patients with established T2DM, who were then diagnosed with PCa and treated with ADT. The authors noted a progressive and marked deterioration of lipid profiles during the 15 month observation period post-treatment. Interestingly, levels of total cholesterol (TC), low density lipoprotein cholesterol and triglycerides (TGs) were increased and those of high density lipoprotein (HDL) cholesterol levels were markedly decreased with ADT treatment of patients with T2DM. Most significantly, in the same study, a continuous increase in fibrinogen and plasminogen activator inhibitor 1 (PAI-1) levels occurred, suggesting the induction of pro-coagulation factors in these diabetic patients managed with ADT. Further, a continuous increase in the concentrations of the non-specific inflammatory marker CRP was noted. These observations strongly suggest that ADT induces metabolic alterations that contribute to the pathophysiology of CVD via several different mechanisms.

Saylor and Smith [67] have analysed a large database (n = 73 196  66 years or older) in men treated with GnRH agonists and demonstrated a significant association with increased risk of T2DM, CHD, MI and sudden death. The data showed that ADT increased the risk of T2DM by 44%, CHD by 16%, MI by 11% and sudden cardiac death by 16%. These findings were summarized in a report published in Circulation [68] by the American Heart Association, the American Cancer Society, and the American Urological Association (AUA). This scientific article had prompted the Food and Drug Administration (FDA) in the USA to establish the increased cardio-metabolic risk in the label of all drugs used for ADT.

Evidence from studies with TRT

Testosterone replacement therapy is indicated in men with TD defined by blood total T less than 12 nmol/l and a host of clinical signs and symptoms of TD, as discussed previously [69]. Exercise-induced myocardial ischemia in men during stress test, as assessed by time to ST depression, is attenuated by T treatment, suggesting that TRT may mediate vasodilation [70]. After chronic T therapy reduced arterial calcification [52], improvement in lipid profiles [71,72] and reduction in inflammatory cytokines [73–75] were noted. Patients with reduced baseline T levels and ischaemia experienced most benefit of TRT [75]. In men with chronic heart failure, increased cardiac output and reduced systemic vascular resistance were noted subsequent to acute TRT [76]. Men with CHD and TD had increased perfusion in the myocardium supplied by unobstructed arteries and increased left ventricular ejection fraction after oral TRT for 8 weeks [77]. T may exert its beneficial effects on the vasculature through modulation of cardiovascular risk factors such as hypertension [78], hypercholesterolaemia [79–81], T2DM [82] and obesity [72].

A recent review had suggested that TRT shows benefits of glycaemic control, reduced IR, decreased visceral adiposity and reduced TC and low T levels are associated with the presence and progression of atherosclerosis in carotid, coronary and aortic vessels [83].

Insulin resistance correlates positively with the degree of visceral obesity [42] and is a key central biochemical defect in development of the metabolic syndrome (Met S), T2DM and atherosclerosis. The notion that IR promotes atherosclerosis is supported by recent findings that demonstrated that it is an independent predictor of coronary artery calcification, a surrogate marker of subclinical CAD [84].

Testosterone deficiency contributes to increased body fat and adiposity and IR. Young men with hypogonadotrophic hypogonadism who were treated with T demonstrated rapid reduction in insulin sensitivity following acute withdrawal of T within a 2-week period, which is not associated with weight changes during such a short time period [85]. Interestingly, reduced T levels are associated with impaired mitochondrial function including oxidative phosphorylation and VO2max, suggesting multiple mechanisms of T action on insulin sensitivity [86]. TRT improves BMI and glycaemic control and reduces IR, inflammatory factors (such as TNF-α and IL6) IMT, TC and slightly increases Hb [13]. In the following section, we discuss the relationship between TD (hypogonadism) and cardiovascular risk factors and the impact of TRT on the following parameters: (i) body fat, (ii) blood pressure, (iii) IR, (iv) lipids, (v) coagulation, (vi) inflammation and (vii) atherosclerosis [1].

Testosterone therapy and body fat

Visceral obesity is associated with increased atherogenic dyslipidaemia, hypertension, IR, hyperglycaemia, as well as pro-thrombotic and pro-inflammatory states. Adipose tissues secrete a host of adipokines, which may influence metabolism of lipids, glucose homeostasis and contribute to hypertension, as well as thrombotic and inflammatory processes. Several studies have pointed out that reduced sex hormone levels are associated with increased inflammatory markers [55,56,87] with an inverse relationship between plasma T and inflammatory markers noted in elderly men [56]. In young, otherwise healthy, patients acute withdrawal of T caused significant increases in IL-6 and TNF-α within 2 weeks. T treatment for 1 year in patients with newly diagnosed T2DM was associated with an improvement of body composition and increased adiponectin levels [88].

In a placebo-controlled study, Svartberg et al. [89], observed substantial changes in body composition including reductions in fat mass by 5.4 kg and a parallel increase in lean mass by 4.2 kg. The net effect on body weight suggests only a moderate reduction by 1.2 kg. In this study, no controlled diet and exercise intervention other than the usual advice for lifestyle improvement was involved. In the Moscow Study [90], a large placebo-controlled trial in 184 men with Met S, the effects of TRT on body composition were evaluated. The study was double-blind, placebo-controlled during the first 30 weeks; thereafter, all patients were switched to T in an ongoing open label trial. Changes in anthropometric parameters were measured at baseline, after 18 and after 30 weeks. There was a progressive decline in BMI, weight (weight loss of 4.3 kg after 30 weeks) and waist circumference (−6 cm after 30 weeks). Again, all patients received medical advice to modify their lifestyle habits. Interestingly, the placebo group showed the typical U-shaped curve of lifestyle intervention studies with an initial improvement that already faded away during the last 12 weeks of the observation period. In contrast, men with TD treated with T showed significant, progressive reduction in BMI, weight and waist circumference, which was maintained over the study period [90].

In a 2-year long study [31] of which the first year was placebo-controlled with no controlled diet or exercise, a reduction in fat mass by 18.5% after 1 year, which was maintained at 19.2% at the end of the second year, was observed. The placebo group was switched to T undecanoate (TU) in the second year and the same result was achieved with a reduction in fat mass by 17.5%.

In another study [91], a reduction in waist circumference by 8.66 cm after 1 year of T treatment was achieved. Fat mass proportion declined by 18.4% confirming the results of the 2-year study [31], and lean mass increased by 4.22 kg confirming results from the Norwegian study [89]. The question to be answered is whether such striking results are sustainable over time. The longest placebo-controlled studies using T have been conducted over 3 years. However, there are observational data available, which have only recently been evaluated and not yet published (personal communication). In a cohort of 281 men from the Centre of Clinical Andrology at the University of Muenster, 137 have been consistently treated with TU for a period of 4 years, 66 for 5 years and longer. There was a progressive decline of waist circumference over the first 3 years reaching 17.3 cm in the group, which has been treated for a minimum of 5 years. In the same group, a weight reduction by 18.9 kg was observed. This came as a surprise because the short-term studies suggest only moderate weight loss despite a substantial change in body composition with a shift from fat mass to lean mass.

More recently, two more cohorts were evaluated who had undergone treatment for 4 years. In Dr Haidar’s group (personal communication), there was a progressive decline of mean waist circumference up to 8.2 cm and a progressive decline of mean body weight of 13.3 kg confirming the unexpected results from the Muenster group. In Prof. Yassin’s group, the respective results were −10.2 cm waist circumference and −13.5 kg body weight (personal communication). Therefore, a total of 371 men who had received TRT for a minimum of 4 years showed reproducible results with substantial and clinically meaningful reductions in waist circumference and body weight.

Testosterone therapy and blood pressure

In the long-term observation group from the University of Muenster, a reduction in systolic blood pressure by 22 mmHg and in diastolic blood pressure by 19 mmHg was observed. This can, in part, be attributed to the weight loss, but other mechanisms of testosterone include improvements of endothelial function (increase in endothelial nitric oxide synthase, reduction in asymmetric dimethyl arginine, reduction in endothelin-1, suppression of the RhoA-Rho kinase pathway) leading to a vasodilatory effect.

Testosterone therapy and insulin sensitivity

In parallel to the marked improvements in body composition, Aversa et al. [31,91] showed a reduction in IR-measured homeostatic model assessment (HOMA) index. Again, different mechanisms of action are under discussion of which the reduction in fat mass and the increase in muscle mass may be one. A direct effect of T on insulin sensitivity has been demonstrated experimentally. This direct effect may be mediated via mitochondrial function [85].

There is only one pilot study combining a controlled diet and exercise programme with TRT in newly diagnosed patients with T2DM, who all had Met S and TD. This study showed an additive effect of T. While the diet and exercise-only group achieved a reduction in glycosylated haemoglobin (HbA1c) by 0.5% at the end of 1 year, the group that received diet and exercise plus T achieved a reduction by 1.2%. In a meta-analysis [92] of the available studies looking at effects of T on IR, Corona et al. found significant effects on fasting glucose, HOMA and IR [92].

Testosterone therapy and lipids

In the study combining diet and exercise with and without T therapy, the T treatment group had a significantly greater effect on reduction in TG [88]. This positive effect of T on lipids was confirmed in the long-term analysis from Muenster (personal communication).In the study combining diet and exercise with and without T, the T treatment group had a significantly greater effect on increase in HDL cholesterol [88]. Again, this positive effect of T was confirmed in the long-term analysis from Muenster (personal communication).

Testosterone replacement therapy and coagulation factors

In contrast to the pro-coagulatory effect of androgen deprivation demonstrated previously [59–66], TRT has the opposite effect in reducing fibrinogen as well as PAI-1, thus favouring an anti-coagulatory profile [88].

Testosterone therapy and inflammatory markers

Testosterone has been shown to be an effective anti-inflammatory agent [55,56,73]. CRP was reduced to a significantly greater extent in the group receiving T, as compared to the group being only on diet and exercise [88]. The Moscow Study [90] confirmed the anti-inflammatory properties of testosterone in men with Met S: both CRP and TNF-α were significantly reduced in the group on TU [90].

Testosterone therapy and atherosclerosis

The parameter that is perhaps the most clinically relevant measurement with regard to atherosclerosis is the degree of IMT. A reduction of carotid intima media thickness (CIMT) was demonstrated in response to T treatment [31]. The reduction of CIMT appeared to be dose dependent in that men who achieved higher levels of T showed a more pronounced reduction of CIMT [31].

Testosterone deficiency and cardiovascular mortality

Longitudinal epidemiological studies as well other clinical studies showed increased cardiovascular mortality in men with TD [3,6,11,48,53,54]. In the EPIC-Norfolk study, the quartile with the lowest T levels had the shortest survival, the quartile with the highest T levels the longest survival [93]. In a cohort of men with angiographically confirmed CHD, those with normal levels of (bioavailable) testosterone (bioT) had a significantly longer survival than those with lower levels of bioT. The first study that suggested that T treatment could compensate this gap in survival came from Sheffield, UK and was presented at the British Endocrine Societies Meeting in 2011. Men with T2DM who had low T had a shorter survival than men with normal T but with T2DM. The subgroup of men who had received treatment to normalize their low T turned out to have the same survival as those men who had normal T at baseline.

Controversies regarding TRT and cardiovascular safety in men

It is important to point out that a recent study referred to as ‘The Testosterone in Older Men with Mobility Limitations (TOM)’ [33] had cast doubt on the utility of T treatment in elderly men. However, this study is fraught with a number of design and interpretation issues. For instance, in this study, the dose of T administered was not in accordance with the product labelling, which recommended a starting dose of 50 mg T (5 g Testim®) (see 4.2. of the labelling). Thus, administration of high dose (supraphysiological) of T to men with subclinical CVD is expected to cause water and salt retention that would have undesirable effects. Furthermore, the starting dose was 100 mg T (10 g Testim®, Auxilium Pharmaceuticals, Inc., Malvern, PA, USA). The dose of 150 mg T (15 g Testim®) per day is not at all indicated in the product labelling. For this reason, it is not clear why the investigators used this non-recommended high dose of T, knowing clearly that the Endocrine Society Guidelines do not make such recommendations [94]. More perplexing, however, is that the study included patients with high cardiovascular risk profiles and the randomization between the two groups was poor, at best. One may argue that the study design does not comply with guidelines and recommendations on TRT and monitoring for TD (hypogonadism) by the guidelines published by several medical societies. To further question the validity of the conclusions, the study was not sufficiently powered to give significance for cardiovascular events. In addition, the adverse events were not the primary endpoints of this study, but the study conclusion focused only on these events. The adverse events occurred mostly in men with higher doses of T. The cardiovascular events were poorly documented (e.g. self-reported syncope, review of medical records or phone interview). These events are a mixture of different types with no clear evidence that one particular type of event was significant. Based on these obvious shortcomings of the TOM study [33], we do not believe that its conclusions can be generalized to men with TD, who are treated appropriately with TRT and monitored carefully as pointed out in other well executed studies [95]. In four placebo-controlled studies, in which frail, elderly patients were treated with TRT, no safety issues were noted, contrary to the previous study [33]. In one study, Caminiti et al. [96] treated 70 frail men, median age 70 years, with stable chronic heart failure, with TU 3 months and noted no side effect requiring discontinuation. In a large study carried out by Srinivas-Shankar et al. [95], 262 frail and pre-frail men, mean age 73.8 years, were treated with T gel for 6 months. The authors noted that one PCa was found in the placebo group, but no other safety issues were reported. Kenny et al. [97] also treated 131 frail men, mean age 77.1 years (range 57–95), with Tgel up to 24 months. The authors noted 10 deaths (three in the T group, seven in placebo, NS). Similarly, Cornoldi et al. [98] treated 87 men with T2DM and proven CAD, mean age 75 years, with TRT, and no significant side effects were noted.

Summary and conclusions

The relationship between TD and CVD is an evolving area of ongoing investigation and the potential role of TRT in ameliorating many of the CVD risk factors is summarized in Figure 1. Androgen deficiency may be the underlying cause for a number of common clinical conditions such as T2DM, erectile dysfunction, Met S and CVD. The relationship between IR and TD has been discussed recently in a number of current reviews [4,99,100] and a link between IR and TD with vascular disease was postulated. Abnormal lipid profiles, elevation of pro-inflammatory factors, hypertension, IR and endothelial dysfunction are common features in men with TD.

Figure 1.

 Effects of testosterone deficiency (TD) and testosterone replacement therapy (TRT) on various cardiovascular disease (CVD) risk markers

Early identification of patients with TD may help reduce the likelihood that these aforementioned clinical features may progress. Though challenges may lie ahead regarding how data from such clinical trials are to be properly interpreted, and whether long-term safety can be established with T supplementation, findings such as those described herein warrant definite investigation into the beneficial role that T may have in preventing or ameliorating CVD risk in men with TD.

Based on the findings discussed in this review, evidence from epidemiological and observational studies showed that low T is associated with a higher cardiovascular risk in men. In longitudinal studies, low T is associated with increased cardiovascular mortality. Most importantly, in patients with PCa, ADT has profound negative effects on all major cardiovascular risk factors. Normalization of T improves all major cardiovascular risk factors: body fat, IR, dyslipidaemia, coagulation, inflammation and blood pressure. TRT in men with TD (hypogonadal) has the potential to reduce cardiovascular risk.

Conflicts of interest

FS is an employee of Bayer Pharma AG, Berlin, Germany, producers and distributors of testosterone products. The current work has not received any funding other than for travel and accommodation to attend the congress.