The Effect of Testosterone on Cardiovascular Disease

Abstract

Cardiovascular disease is the leading cause of death in the United States. Testosterone is the principal male sex hormone and plays an important role in men’s health and well-being. Historically, testosterone was believed to adversely affect cardiovascular function. However, contemporary literature has refuted this traditional thinking; testosterone has been suggested to have a protective effect on cardiovascular function through its effects on the vascular system. Data from modern research indicate that hypogonadism is closely related to the development of various cardiovascular risk factors, including hyperlipidemia and insulin resistance. Several studies have demonstrated beneficial effects of testosterone supplementation therapy on reversing symptoms of hypogonadism and improving cardiovascular disease risk profiles. In this review, we perform a critical analysis on the association between testosterone and cardiovascular disease.

Cardiovascular disease (CVD) is the leading cause of death in most developed countries, with an estimated 17.3 million deaths worldwide per year (Laslett et al., 2012). Although mortality from CVD has decreased considerably in recent years (Smith et al., 2005; Smolina, Wright, Rayner, & Goldacre, 2012), CVD and its complications remain highly prevalent and are significant burdens to the health system. The American Heart Association projected that CVD care costs would triple from $272.5 billion in 2010 to an estimated $818.1 billion in 2030 (Laslett et al., 2012).

Many of the risk factors for CVD are preventable, including smoking, dyslipidemia, hypertension, diabetes, abdominal obesity, psychosocial factors, inadequate daily consumption of fruits and vegetables, regular alcohol consumption, and lack of regular physical activity (Yusuf et al., 2004). Others risk factors are nonmodifiable and include age (Savji et al., 2013), family history of CVD (Andresdottir, Sigurdsson, Sigvaldason, & Gudnason, 2002; Lloyd-Jones et al., 2004; Sesso et al., 2001), and gender (Kappert et al., 2012).

The lifetime risk of coronary heart disease (CHD) at age 40 is one in two for men and one in three for women (Lloyd-Jones, Larson, Beiser, & Levy, 1999; Lloyd-Jones et al., 2004). Men are at greater risk for CVD than premenopausal women (X. P. Yang & Reckelhoff, 2011). Higher CVD mortality may be partially attributed to behavioral and physical characteristics of males, including increased risk-taking behavior (smoking, drinking), endocrine and metabolic factors like fat distribution, and lower male engagement in preventive care (Tudiver & Talbot, 1999). However, postmenopause, a woman’s risk is similar to a man’s (X. C. Yang, Jing, Resnick, & Phillips, 1993). This finding suggests a possible association of CVD to sex hormones.

Historically, testosterone (T) has been associated with increased CVD risk. However, modern literature shows conflicting data on T and CVD. T may have a protective effect on the cardiovascular (CV) system via its effects on vascular reactivity (Webb, McNeill, Hayward, de Zeigler, & Collins, 1999), immune modulation (Malkin, Pugh, Jones, Jones, & Channer, 2003), arterial wall stiffness (Dockery, Bulpitt, Donaldson, Fernandez, & Rajkumar, 2003), and the endothelium (Pearson, Yandle, Nicholls, & Evans, 2008). Contemporary studies report that hypogonadism is associated with higher CVD incidence as well as insulin resistance, diabetes mellitus, and metabolic syndrome (MetS; Corona, Rastrelli, Vignozzi, Mannucci, & Maggi, 2011). In fact, low serum T has been associated with increased mortality in male veterans (Shores, Matsumoto, Sloan, & Kivlahan, 2006). Contrarily, other studies demonstrated no association between T and incident CVD (Arnlov et al., 2006; Haring et al., 2013).

A critical narrative review of current available literature on the association of T to CVD and MetS was performed. T’s association with CVD as it relates to four of its major diagnostic categories was explored: CHD as contributed by coronary artery disease (CAD) and manifested by myocardial infarction (MI), angina pectoris, heart failure, and coronary death; cerebrovascular disease as manifested by stroke and transient ischemic attack (TIA); peripheral artery disease (PAD) as manifested by intermittent claudication; and aortic atherosclerosis and aneurysms. Details of the studies cited are summarized in Tables A.1 and A.2 in Appendix A.

Method

Literature Search

A PubMed query was performed using the following key phrases: “cardiovascular disease AND testosterone,” “coronary artery disease AND testosterone,” “heart failure AND testosterone,” “TIA AND testosterone,” “stroke AND testosterone,” “cerebrovascular disease AND testosterone,” “aortic aneurysm AND testosterone,” “atherosclerosis AND testosterone,” and “peripheral vascular disease AND testosterone.” Search terms were limited to within the title and abstract. The search included all articles published as of December 16, 2013.

Study Selection

Studies were included if they met the following criteria: (a) research on T, (b) research on CVD, (c) quantifiable outcomes, and (d) published in English. Articles without quantitative outcomes were excluded. Due to the great number of studies on the topic, a narrative review of 129 articles was conducted.

Testosterone and Cardiovascular Disease

Testosterone and Coronary Heart Disease

Beneficial Effect of T Demonstrated in CHD

In this review, CHD consists of CAD, myocardial ischemia or infarction, heart failure, and mortality from CHD. Several studies have demonstrated that patients with CHD and CAD as diagnosed by coronary angiography have lower levels of T compared to control subjects (Cao et al., 2010; Chute et al., 1987; Sieminska et al., 2003), and that hypogonadism is an independent risk factor for developing CHD (Cao et al., 2010). In addition, T has been reported to have an inverse relationship with CAD severity (Hu et al., 2011; L. Li et al., 2012; Phillips, Pinkernell, & Jing, 1994; Rosano et al., 2007) with epidemiologic evidence demonstrating a fivefold decrease in risk of severe atherosclerotic CAD between the lowest and highest quartiles of total T (TT; Chute et al., 1987). This inverse correlation between T and CAD severity is present in both men and postmenopausal women with CAD (Kaczmarek, Reczuch, Majda, Banasiak, & Ponikowski, 2003).

Men with MIs and ischemia have been reported to have lower T and an increased estrogen (E) to T ratio when compared to controls (Lichtenstein et al., 1987; Sewdarsen, Jialal, Vythilingum, & Desai, 1986). It has been postulated that T may have a regulatory effect on lipid metabolism and thus on ischemic heart disease (Lichtenstein et al., 1987; Sewdarsen et al., 1986), with a possible interdependent effect with insulin (Lichtenstein et al., 1987). TT and bioavailable T have been reported to decrease following acute MIs in men, in association with adverse changes in the fibrinolytic profile including an elevation in plasminogen activator inhibitor-1 (PAI-I) activity and reduction of tissue plasminogen activator activity, independent of insulin changes (Pugh, Channer, Parry, Downes, & Jone, 2002).

T deficiency is prevalent in patients with chronic congestive heart failure (CHF), with a prevalence of 39% (Florvaag et al., 2012; Jankowska et al., 2006). Chronic CHF has been associated with significantly lower free testosterone (FT; Kontoleon et al., 2003). New York Heart Association Functional Classification is the most commonly used classification system that divides patients into four categories by limitations in physical exercise; higher New York Heart Association classes and impaired left ventricular function have been associated with lower levels of TT and FT even after adjustment for associated comorbidities (Florvaag et al., 2012; Jankowska et al., 2006; Wehr et al., 2011). Endogenous T has also been linked to increased exercise tolerance in patients with CHF. Bocchi, Carvalho, and Guimaraes (2008) investigated the correlation between T and hemodynamics in CHF patients with erectile dysfunction and reported a direct correlation between T levels and exercise tolerance, as reflected by an increased distance at 6-minute cardiopulmonary walk test; interestingly, T was inversely correlated with diastolic blood pressure and right and left ventricle ejection fraction. Diminished endogenous T levels have also been associated with reduced maximum oxygen uptake (VO2), a measure of exercise capacity, with greater reductions in TT correlating with more severe progression of exercise intolerance (Jankowska et al., 2009).

Several studies have identified an inverse relationship between T levels and mortality in men with CHD including CAD and CHF (Malkin et al., 2010; Ponikowska et al., 2010). In a study of 208 men with CHF, Jankowska et al. (2006) reported that both TT and estimated free T (eFT) were independent prognosticators of mortality. Wehr et al. (2011) reported that patients with FT in the lowest quartile also had increased CHF mortality compared to those in the highest quartile; however, no significant association between TT and CV events or cardiac disease was noted. Others reported that although androgens are commonly decreased in elderly patients with systolic CHF and related to disease severity, after adjustment for confounding factors, neither TT nor eFT independently predicted mortality (Guder et al., 2010; Wu, Wang, Wang, & Li, 2011).

Harmful or Absent Effect of T Demonstrated in CHD

Contrarily, multiple studies have demonstrated a lack of association or an adverse association between T and CAD or identified relationships between other sex hormones. A longitudinal case control study reported no significant difference in T between low-risk men who developed CHD and those who did not (Heller, Miller, Wheeler, & Kind, 1983). Similarly, Contoreggi et al. (1990) reported that plasma levels of T, FT, and dehydroepiandrosterone were not predictive of the risk for developing CAD in men and that cholesterol and systolic blood pressure are more important CAD risk factors. Luria et al. (1982) reported that T levels were similar between patients with acute MIs and proven CAD and healthy individuals; however, although T levels were similar in all groups, estradiol (E2) levels were substantially elevated in male patients with MI and chronic angina pectoris, leading to the hypothesis that increased peripheral conversion of T to E may contribute to MI and CAD. Contrarily, Kajinami et al. (2004) reported that E2 was lower and FT was elevated in patients with proven CAD and controls without but not to a significant degree; however, the ratio of FT to E2 was significantly higher than in the controls, leading to the hypothesis that T may inhibit early atherogenesis by conversion to E2.

In a prospective population-based study of 1,009 older White males, sex hormones measured (T, androstenedione, estrone, E2) did not have a significant association with known CVD at baseline or with subsequent CV or ischemic heart disease mortality (Barrett-Connor & Khaw, 1988). In a longitudinal study of CVD in male twins followed for 20 years, no significant difference in T levels was present in discordant monozygotic twins and similar T, E, and sex hormone–binding globulin were present in both positive and negative concordant twins (Mikulec et al., 2004). In 192 diabetic men, hypogonadism was linked to various CVD risk factors, including body mass index (BMI), waist circumference, triglycerides (TGs), glucose, and insulin but not silent MI or PAD (Hernandez-Mijares et al., 2010). Although Ponholzer et al. (2010) observed an association between low dehydroepiandrosterone levels and CHD, no association was identified between increased vascular risk and impaired T levels.

Testosterone and Cerebrovascular Disease: Stroke and Transient Ischemic Attack

Beneficial Effect of T Demonstrated in Stroke/TIA

Low TT and FT have been reported to be predictive of increased incidence of stroke or TIA, even after adjusting for conventional risk factors for CVD (Yeap et al., 2009). Multiple studies have demonstrated that low T and FT are associated with increased carotid intima-media thickness (IMT), which serves as a measure of cerebrovascular atherosclerosis (De Pergola et al., 2003; Fukui et al., 2003; van den Beld et al., 2003). In a population-based cross-sectional study, Svartberg et al. (2006) reported an inverse association between TT and carotid artery IMT that was present after excluding men with CVD; this relationship, however, was not independent of BMI. Similar results were confirmed thereafter by Debing et al. (2008). Along these same lines, a cross-sectional study looking at a cohort from the Tromso study demonstrated an inverse association between T levels and total carotid plaque area (Vikan, Johnsen, Schirmer, Njolstad, & Svartberg, 2009). Saltiki et al. (2010) reported that subjects with an increased mean carotid IMT had lower levels of FT and bioavailable T, but not TT. Similarly, a recent study of middle-aged Japanese men reported that low serum FT but not TT was associated with carotid artery IMT after multivariate analysis with age, BMI, and other clinically relevant factors (Tsujimura et al., 2012). In a multicity population-based cohort study, Soisson et al. (2012) reported that low bioavailable T and elevated carotid IMT were associated with low-grade inflammation (C-reactive protein [CRP] ≥ 2 mg/L) in elderly men. T’s beneficial effect on carotid IMT is further demonstrated by Aversa et al. (2010), who demonstrated a significant decrease in carotid IMT after a 12-month course of transdermal T.

Harmful or Absent Effect of T Demonstrated in Stroke/TIA

Contrarily, in the Honolulu–Asia Aging Study, although E2 increased the risk of stroke, T had no association to stroke risk (Abbott et al., 2007). In a retrospective longitudinal study, no association was observed for FT or TT with carotid IMT in young to middle-aged women (Calderon-Margalit et al., 2010). Similarly, T was not associated with the progression or level of carotid IMT or adventitial diameter in perimenopausal women (El Khoudary et al., 2012).

Several studies have suggested a male predominance of pediatric stroke (Fullerton, Wu, Sidney, & Johnston, 2007; Golomb et al., 2004; Golomb, Fullerton, Nowak-Gottl, & Deveber, 2009). Normann et al. (2009) reported that elevated T was an independent risk factor and that for each 1 nmol/L increase in T in boys, the odds of cerebral thromboembolism were increased 1.3-fold.

Testosterone, Aortic Atherosclerosis/Aneurysm, and Peripheral Vascular Disease

Beneficial Effect of T Demonstrated in Aortic or Peripheral Vascular Disease

In a large population-based cohort study, Yeap, Hyde, Norman, Chubb, and Golledge (2010) reported that higher FT reduces the odds ratio of abdominal aortic aneurysms in elderly men, suggesting that impaired gonadal function may be involved in abnormal arterial dilation as well as vascular disease in older men. Interestingly, Vaidya et al. (2012) reported that T was negatively associated with baseline waist-to-hip ratio, which has been strongly associated with prevalent atherosclerosis, in men but not in postmenopausal women.

Using ankle brachial index (ABI) as a measure of PAD (ABI < 0.9), Tivesten et al. (2007) demonstrated that elderly men with lower T and higher E had significantly lower ABI and thus lower extremity PAD. Using intermittent claudication as a measure of PAD, Yeap et al. (2013) reported that lower T had an inverse association with symptoms and thus lower-extremity PAD even after multivariate adjustment for age, smoking BMI, and other relevant clinical factors. In a case report of men with complications from diabetic peripheral sensory neuropathy and peripheral vascular disease whose serum T was also subnormal, Kalinchenko, Zemlyanoy, and Gooren (2009) reported that on normalization of serum T, there was an improvement in hyperglycemia, decrease in leukocyte and fibrinogen levels, and increase in antithrombin III and tissue oxygen pressure; most significantly, the wounds showed granulation, which the authors hypothesized may be due to improved vascularization and anti-inflammatory action.

Harmful or Absent Effect of T Demonstrated in Aortic or Peripheral Vascular Disease

Interestingly, Maggio et al. (2012) reported that high T levels are independently associated with the presence of PAD as defined by ABI > 0.90 in older women but not in men. Although Haring et al. (2011) identified an inverse relationship between ABI and TT, the relationship was insignificant in multivariable logistic models. In the Vienna Transdanube Aging study, TT was not associated to vascular disease, including CAD and PAD (Ponholzer et al., 2010). In 192 diabetic men, Hernandez-Mijares et al. (2010) reported that although low T is prevalent in diabetic males, it is not associated with an increase in PAD. Finally, a Cochrane review of three studies with a total of 109 subjects with intermittent claudication or critical leg ischemia reported that T replacement therapy did not improve walking distance and other measures of PAD (venous filling time, muscle blood flow, plethysmography); however, authors indicated that this might reflect limited data available rather than the lack of a real effect (Price & Leng, 2012).

Testosterone and Metabolic Syndrome

MetS encompasses a combination of medical disorders, which increase the risk of developing CVD and diabetes. A meta-analysis of 20 studies reported that even after adjusting for age and BMI, MetS is independently associated with hypogonadism (Corona, Monami, et al., 2011). Contrarily, a more recent epidemiologic study of elderly men and women did not demonstrate an inverse relationship between T and MetS after adjusting for lipids, BMI, inflammation, and insulin resistance (Chrysohoou et al., 2013). T has been inversely associated with all the components of MetS: central obesity (Khaw & Barrett-Connor, 1992; Muller, Grobbee, den Tonkelaar, Lamberts, & van der Schouw, 2005; Vaidya et al., 2012), low high-density lipoprotein (Lichtenstein et al., 1987; Zmuda et al., 1997), high TGs (Lichtenstein et al., 1987; Zmuda et al., 1997), and impaired fasting glucose (Colangelo et al., 2009; Svartberg, Jenssen, Sundsfjord, & Jorde, 2004). T improves fasting insulin sensitivity in men with CHF (Malkin, Jones, & Channer, 2007). Low levels of T have been associated with an unfavorable lipid profile along with low lipoprotein lipase levels (Breier et al., 1985). Contrarily, other studies did not demonstrate an association between lipid profile (high-density lipoprotein, low-density lipoprotein, TG, and total cholesterol) and T in both healthy men and men with CAD (Davoodi et al., 2007; Duell & Bierman, 1990; Wranicz et al., 2005)

In addition to the independent components of MetS, low levels of TT and sex hormone–binding globulin have also been associated with an increased likelihood of having MetS, independent of other known CV risk factors and insulin resistance (Kupelian, Hayes, Link, Rosen, & McKinlay, 2008; C. Li, Ford, Li, Giles, & Liu, 2010; Muller et al., 2005). In fact, a longitudinal study by Laaksonen et al. (2004) reported that the lowest quartile of T levels had a twofold increase in the risk of MetS than the highest quartile independent of important risk factors.

Testosterone Supplementation Therapy

ST Depression/Ischemia

T supplementation therapy has been demonstrated to have a beneficial effect on exercise induced myocardial ischemia in men with CAD (Jaffe, 1977). In a double-blinded, randomized study, men who received 4 to 8 weeks of testosterone cypionate treatment had significant decreases in postexercise ST segment depression on EKG compared to placebo, indicating a decrease in myocardial ischemia; the mechanism of action was unclear and patients did not have a history of CAD (Jaffe, 1977). Similarly, men with chronic stable angina and ischemic heart disease who received daily small dose T supplementation for 3 months experienced prolonged time to ischemia compared to placebo as evidenced by increased time to ST segment depression and improved pain perception and mood; these differences were accentuated in patients with low bioavailable T (English, Steeds, Jones, Diver, & Channer, 2000; Malkin, Pugh, Morris, et al., 2004). Short-term administration of T (30 minutes prior to exercise) has also been demonstrated to decrease chest pain, maximum ST segment depression, and recovery time as well as increase time to ST segment depression and total exercise time; Rosano et al (1999) postulated that T has a direct vasodilatory effect, as evidenced by an increase in heart rate observed at the ST depression and at peak exercise. Long-term administration (12 months) of T appeared to be protective against myocardial ischemia by increasing maximal exercise time, time to 1 mm ST depression, and peak metabolic equivalents but without a difference in quality of life (Mathur et al., 2009). Cornoldi et al. (2010) reported that a 12-week course of low-dose T in diabetic males with CAD decreased the number of anginal attacks, silent ischemic episodes, and total ischemic burden as evidenced in ambulatory EKG compared to placebo.

Contrarily, Thompson et al. (2002) reported that acute administration of T in 32 men with provocable myocardial ischemia did not improve cardiac function, demonstrating no difference on exercise/pharmacologic stress test in peak heart rate, systolic blood pressure, maximal rate pressure product, perfusion imaging scores, or the onset of ST segment depression; they concluded that acute T supplementation does not have a beneficial or adverse effect on the onset and magnitude of stress-induced myocardial ischemia in men with stable CAD.

CHF Exercise Tolerance

As low T independently predicts reduced exercise capacity and poor clinical outcomes in CHF patients, several studies have investigated the effect of T supplementation on exercise tolerance in patients with CHF. T supplementation in elderly men with CHF has been reported to increase exercise capacity as demonstrated by incremental shuffle walk test as well as improve symptoms of CHF by at least one functional class (Malkin et al., 2006). Stout et al. (2012) demonstrated both the feasibility and efficacy of a 12-week program of exercise therapy with T therapy in elderly patients with CHF and hypogonadism, reporting improvements in peak oxygen update, leg strength, and several other quality-of-life domains versus placebo. In a meta-analysis of four randomized and double-blinded trials (Caminiti et al., 2009; Iellamo et al., 2010; Malkin et al., 2006; Pugh, Jones, West, Jones, & Channer, 2004), Toma et al. (2012) concluded that T supplementation produced significant improvements in exercise capacity compared to placebo with no adverse effects as demonstrated by an increase in 6-minute walk test, incremental shuffle walk test, and peak oxygen consumption. In fact, the differences were greater than that seen with other therapies including angiotensin-converting–enzyme inhibitors, beta-blockers, and cardiac resynchronization therapy (Toma et al., 2012). Whereas Malkin et al. (2006) and Pugh et al. (2004) did not find an increase in skeletal muscle bulk or handgrip strength, Iellamo et al. (2010) and Caminiti et al. (2009) observed an improvement. Studies looking at T therapy in men with CVD have been summarized in Table 1.

Table 1.

Summary of Randomized Clinical Trials of TRT in Patients with CVD.

Study	Type of study	Age, years	Sex	No. of intervention-treated subjects	No. of placebo-treated subjects	Inclusion criteria	Treatment	Dosage	Treatment duration	Findings	Values
Jaffe (1977)	R, DB	35-71	Male	25	25	ST segment depression following two-step exercise test	IM testosterone cypionate	200 mg weekly	4 weeks	↓ ST depression	3.9 ± 0.9 a mm(31.7% ± 6.5%)
Jaffe (1977)	R, DB	35-71	Male	25	25	ST segment depression following two-step exercise test	IM testosterone cypionate	200 mg weekly	8 weeks	↓ ST depression	6.0 ± 1.0 a mm(51.2% ± 7.0%)
Rosano et al. (1999)	R, C	58 ± 4	Male	14	14	CAD	IV testosterone	2.5 mg once	30 minutes	↑ time to 1 mm ST depression	108 seconds(579 ± 204 vs. 471 ± 210);p < .01
Rosano et al. (1999)	R, C	58 ± 4	Male	14	14	CAD	IV testosterone	2.5 mg once	30 minutes	↑ exercise time	631 ± 180 vs. 541 ± 204 seconds; p < .01
English et al. (2000)	R, DB, PC	62 ± 2	Male	22	24	Stable angina	Testosterone patches	5 mg qhs	4 weeks, 12 weeks	↑ time to 1 mm ST depression	p = .02
Thompson et al. (2002)	R, C	69 ± 6	Male	34	34	Inducible myocardial ischemia on recent stress perfusion imaging study	IV testosterone	Individualized: to 2× baseline or 6× baseline	Acute	↔ time to ST depression	Placebo: 4.8 ± 2.2 minutes
											2× T: 4.9 ± 2.2 minutes
											6× T: 4.8 ± 2.3 minutes
Thompson et al. (2002)	R, C	69 ± 6	Male	34	34	Inducible myocardial ischemia on recent stress perfusion imaging study	IV testosterone	Individualized: to 2× baseline or 6× baseline	Acute	↔ exercise time	Placebo: 9.3 ± 2 minutes
											2× T: 9.2 ± 2 min
											6× T: 9.3 ± 2 minutes
Malkin et al. (2004)	R, SB, C	60.8 ± 4.6	Male	10	10	Hypogonadism, ischemic heart disease	IM testosterone	100 mg q 2 weeks	4 weeks	↑ time to 1 mm ST depression	74 seconds (35.4, 112.6) p < .001
Pugh et al. (2004)	R, DB, PC	44-81 (median 62)	Male	10	10	CHF	IM testosterone (Sustanon)	100 mg q 2 weeks	12 weeks	↑ shuffle walk test distance	91.0 ± 19.7 a minutes vs. 26.0 ± 15.2 a minutes
											p = .018
											Treatment effect = 65.0 ± 24.9 a minutes
Malkin et al. (2006)	R, DB, PC	64 ± 9.9	Male	37	39	CHF	Testosterone patch (Androderm)	5 mg qhs	12 months	↑ exercise time	25 ± 15 minutes; p = .006
Caminiti et al. (2009)	R, DB, PC	66-76	Male	35	35	CHF	IM testosterone undecanoate (Nebido)	1,000 mg q 6 weeks (×2 total)	12 weeks	↑ peak oxygen consumption, quads maximal voluntary contraction, distance in 6-minute walk test	p < .05
Mathur et al. (2009)	R, DB, PC	64.8 ± 7	Male	7	6	Stable, chronic angina pectoris, proven ST segment depression >1 mm, T < 12 nmol/L	IM testosterone undecanoate (Nebido)	1,000 mg ×4 total	12 months	↑ time to 1 mm ST depression	129 ± 48 vs. 12 ± 18 seconds; p = .02
Aversa et al. (2010)	R, DB, PC	57 ± 8	Male	40	10	Metabolic syndrome, type 2 diabetes	IM testosterone undecanoate (Nebido)	1,000 mg q 12 weeks	24 months	↓ insulin resistance, visceral adiposity
Cornoldi et al. (2010)	R, DB, PC	74 ± 7	Male	43	44	Diabetic with CAD	IM testosterone undecanoate	40 mg tid	12 weeks	↓ episodes per patient, silent episodes, total ischemic burden (seconds/24 hours)
Iellamo et al. (2010)	R, DB, PC	68.2 ± 6.85	Female	20	12	CHF	Testosterone patch (Intrinsa)	300 µg biweekly	24 weeks	↑ peak oxygen consumption, distance in 6-minute walk test	p < .05
Stout et al. (2012)	R, DB, PC	67.2 (51-84)	Male	15	13	CHF (mean T: 10.7 ± 2.6 nmol/L)	IM testosterone (Sustanon)	100 mg q 2 weeks	12 weeks	↑ peak oxygen consumption, leg strength	Peak oxygen uptake: p < .01
											Leg strength: p < .05
											No difference in shuttle walk test, body mass, hand grip strength

Note. TRT = testosterone replacement therapy; CVD = cardiovascular disease; R = randomized; DB = double-blind; IM = intramuscular; C = crossover; IV = intravenous; CAD = coronary artery disease; PC = placebo controlled; qhs = at bedtime; q = every; T = testosterone; SB = single-blind; CHF = congestive heart failure. Ages are represented in ranges or M ± SD, unless otherwise denoted. Continuous values are M ± SD unless otherwise denoted.

a.

Mean ± standard error of mean.

Metabolic Syndrome and Other CV Risk Factors

In a randomized, double-blinded, placebo-controlled study of patients with hypogonadism and MetS, Aversa et al. (2010) reported that 24 months of T supplementation reduced fasting glucose, insulin resistance, waist circumference, and visceral fat mass and improved surrogate markers of endothelial function and atherosclerosis like high-sensitivity CRP and carotid IMT without any significant adverse effects. T supplementation in hypogonadal men also may result in an improvement in the lipid profile, including beneficial reduction of total cholesterol (Cornoldi et al., 2010; Malkin, Pugh, Morris, et al., 2004) and plasma TGs (Cornoldi et al., 2010).

Risks of Testosterone Therapy

Despite the benefits stated above, T supplementation is not without adverse effects. Supratherapeutic doses of anabolic steroids can reduce cardiac function and induce hypertension, ventricular remodeling, MI, and sudden cardiac death (D’Andrea et al., 2007; Karila, Karjalainen, Mantysaari, Viitasalo, & Seppala, 2003; Sullivan, Martinez, & Gallagher, 1998); however, anabolic steroids have different chemical structures and properties from T. T causes salt and water retention, particularly in older men (Johannsson, Gibney, Wolthers, Leung, & Ho, 2005; Quan et al., 2004), which may contribute to edema, hypertension, and CHF.

Erythropoiesis

Of note, elevated hematocrit is the most common side effect of T therapy (Shahani, Braga-Basaria, Maggio, & Basaria, 2009). Hematocrit and hemoglobin have been reported to increase in a linear, dose-dependent fashion with T (Coviello et al., 2008). In a meta-analysis of 19 randomized controlled trials investigating the adverse effects of T therapy, elderly T treated men were 4 times more likely to have increased hematocrit of >50% than placebo-treated men; despite this increase, no significant difference in CV events was noted (Calof et al., 2005). Another meta-analyses of 51 studies reported similar increases in hemoglobin and hematocrit with no significant effect of T therapy on mortality or prostate or cardiac outcomes (Fernandez-Balsells et al., 2010). Moreover, in a randomized clinical trial of 120 men receiving long-acting intramuscular T undecanoate, although elevated, hematocrit and hemoglobin remained in clinically safe limits (Tan et al., 2013). One proposed mechanism by which T induces erythropoiesis is the suppression of hepcidin, an iron-regulatory peptide (Bachman et al., 2010). Several studies have not identified any changes in erythropoietin (EPO) levels (Coviello et al., 2008; Maggio et al., 2013). However, a more recent study demonstrated that T-induced erythropoiesis is associated with both EPO stimulation and reduced ferritin and hepicidin, proposing a possible recalibration of the EPO set point and increased utilization of iron for erythropoiesis in T-treated men (Bachman et al., 2013).

Cardiac Events

The TOM (Testosterone in Older Men With Mobility Limitations) trial was prematurely stopped due to a higher incidence of cardiac, respiratory, and dermatological events in patients receiving T supplementation versus placebo (Basaria et al., 2010). In the TOM trial, elderly community-dwelling men with low serum T and limited mobility were randomly assigned to receive 6 months of daily T 100 mg or placebo transdermal gel to assess the effect of T therapy on exercise tolerance. The trial was stopped after 209 of the planned 252 men were enrolled due to the imbalance in cardiac events between the two groups (23 vs. 5). Despite the alarming increase in cardiac events in men receiving T therapy versus placebo, diversity of the cardiac events that occurred and the lack of structured ascertainment of CV events limit the clinical value of the study; moreover, the population studied and the utilization of higher T dosages may limit the study’s generalizability (Basaria et al., 2010).

Two retrospective cohort Veterans Affairs studies reported differing conclusions regarding the association of T therapy and mortality (Shores, Smith, Forsberg, Anawalt, & Matsumoto, 2012; Vigen et al., 2013). Shores et al. (2012) investigated the association between T therapy and total mortality in men with serum TT levels less than 250 ng/dl and older than 40 years. A total of 398 T-treated men, mean age 60.9 years, were compared with 633 untreated men, mean age 62.8 years. T users were younger, had higher BMIs, and had lower T levels. After adjusting for these factors and other relevant covariates, T-treated men had a 39% reduction in mortality risk (hazard ratio [HR] = 0.61; 95% confidence interval [CI] = 0.42, 0.88; p = .008) versus untreated men. Contrarily Vigen et al. (2013), conducted a retrospective cohort study of men who underwent coronary angiography and had T levels less than 300 ng/dl and demonstrated an increase in adverse effects (HR = 1.29; 95% CI = 1.04, 1.58) in men treated with T. Of the 1,223 T-treated men, mean age 60.6 years, 67 died, 420 had MIs, and 486 had strokes compared to 681 deaths, 420 MIs, and 519 strokes among the 7,486 untreated men, mean age 63.8. The hazard risk remained unchanged even after adjustment for CAD.

Proposed Mechanism of Testosterone’s Action on the CV System

Endothelial Dysfunction

Endothelial dysfunction is considered to be the first step in the formation of atherosclerotic lesions. T has been demonstrated to have a protective effect on endothelial function (Fu, Gao, & Shen, 2008). T has been inversely correlated with vascular cell adhesion molecule-1, which is produced by endothelial cells (ECs) and up-regulated when ECs undergo inflammatory and malignant stimulation (Fu et al., 2008). Foresta, De Toni, Selice, Garolla, and Di Mambro (2010) reported that hypogonadal men had lower levels of endothelial progenitor cells (EPCs), which are important in endothelial regeneration, and higher levels of an osteocalcin-positive subpopulation of EPCs, which are highly correlated with atherosclerosis progression, compared to eugonadal men. Contrarily, Florvaag et al. (2012) reported that despite the frequency of T deficiency in men with CHF, it did not have an effect on EPCs. Incubating ECs and smooth muscle cells (SMCs) with androgens in vitro, Nheu et al. (2011) reported that T can have potentially beneficial or harmful effects, noting that although T induces DNA synthesis and EC growth via activation of MAP (mitogen-activated protein) kinase ERK1/2 activity, T also enhances SMC proliferation, which can contribute to atherosclerosis. Finally, T has been shown to significantly reduce endoplasmic reticulum stress and superoxide generation in human umbilical vein ECs, both of which have been implicated in atherosclerosis; however, when combined with aromatase inhibitors, the protective effect of T was lost, suggesting an E2-mediated mechanism (Haas, Raheja, Jaimungal, Sheikh-Ali, & Mooradian, 2012).

T’s proposed anti-anginal and anti-ischemial effects have been partially attributed to its effects on the vascular system. Although T’s vasodilatory effects are well recognized, the exact mechanism of action is yet to be elucidated. T has been reported to induce endothelium-independent relaxation of numerous vascular beds, including isolated rabbit coronary artery and aorta (Yue, Chatterjee, Beale, Poole-Wilson, & Collins, 1995), human internal mammary arteries (Yildiz et al., 2005), and human radial arteries (Seyrek, Yildiz, Ulusoy, & Yildirim, 2007). This direct relaxation response to T has been attributed to potassium conductance in the form of non–ATP (adenosine triphosphate)-sensitive potassium channel (Yue et al., 1995), ATP-sensitive potassium channel (Seyrek et al., 2007), and large-conductance calcium-activated potassium channel opening action (Yildiz et al., 2005). T has also been reported to induce vasodilation by reducing calcium influx into vascular smooth muscle by acting as a selective and potent inhibitor of L-type calcium channels at physiologic levels and as an inhibitor of T-type channels at supraphysiologic levels (Scragg, Jones, Channer, Jones, & Peers, 2004). Supporting T’s direct vasodilatory effect, despite inducing coronary artery dilation and increasing coronary blood flow in men with CAD, short-term (3-minute) intracoronary administration of T had no effect on acetylcholine-mediated increase in coronary blood flow or vasoreactivity, suggesting a lack of effect on stimulated endothelial nitric oxide (Webb et al., 1999).

Contrarily, other studies have suggested an endothelium-dependent mediated mechanism behind T’s vasodilatory effect (Kang et al., 2002; Ong et al., 2000). Both acute and long-term administration of T in men with CAD resulted in increases in brachial artery flow-mediated reactivity (Kang et al., 2002; Ong et al., 2000), which induces shear stress release of nitric oxide and subsequently leads to vasodilation. This relation has also been demonstrated in postmenopausal women (Montalcini et al., 2007).

Arterial wall stiffness is an independent predictor of CVD risk. Low T levels have been associated with endothelial dysfunction (Laurent, Tropeano, & Boutouyrie, 2006). This inverse relationship has been demonstrated using both pulse pressure (Corona et al., 2009) and pulse wave velocity (Fukui et al., 2007) as reflections of arterial wall stiffness. In fact, the association between T and CVD mortality was lost in male hemodialysis patients after adjusting for pulse wave velocity, suggesting that endothelial dysfunction may be a possible explanation of T’s inverse association with CVD (Kyriazis et al., 2011).

Contrarily, long-term (8-week) administration of T increased myocardial perfusion in unobstructed coronary arteries and decreased radial and aortic augmentation indexes, indicating decreased arterial wall stiffness; however, no effect was observed on global perfusion or on endothelial function (Webb et al., 2008).

Inflammation

Atherosclerosis is mediated by an ongoing inflammatory response, which is induced by cytokines and other inflammatory markers. Cytokines cause cellular and local arterial wall inflammation and may lead to vascular smooth muscle apoptosis, degradation of the fibrin cap, and plaque rupture, thereby leading to platelet adhesion, thrombus formation, and ultimately angina or MI (Malkin et al., 2003). An elevation in inflammatory markers or cytokines has been identified to be predictive of outcomes in patients with CVD (Libby, Ridker, & Maseri, 2002). TT has been negatively associated with inflammatory markers such as macrophage inflammatory protein 1-alpha, 1-beta, and tumor necrosis factor alpha in young men, suggesting a low-grade inflammatory state (Bobjer, Katrinaki, Tsatsanis, Lundberg Giwercman, & Giwercman, 2013). Accordingly, T has been reported to have immune-modulating characteristics, which, due to inflammation’s role in atherosclerosis, may play into the development of CVD.

T may suppress the expression of inflammatory cytokines. T supplementation has been reported to suppress the expression of high-sensitivity CRP and interleukin-6 in patients who underwent coronary artery stent implantation, leading to the hypothesis that T’s anti-inflammatory property could potentially attenuate major CV events after stent placement (Guler et al., 2006). In a randomized, placebo-controlled, crossover study of T supplementation in hypogonadal men, T supplementation reduced levels of proinflammatory cytokines TNF-α (tumor necrosis factor alpha) and IL-1β (interleukin-1 beta) while suppressing levels of cytokine interleukin-10 (Malkin, Pugh, Jones, et al., 2004). Similarly, Corcoran et al. (2010) demonstrated that T suppresses macrophage TNF-α and IL-1β expression. However, the inverse relationship between TNF-α and T was not identified in men with CHF (Pugh et al., 2005). Hernandez-Mijares et al. (2010) did not find a significant difference in IL-6 or TNF-α between diabetic males with hypogonadism compared to eugonadal men.

Clotting

Fibrinogen is a known risk factor for CVD as well as an inflammatory biomarker (Danesh et al., 2005); it increases CVD risk through its effects on atherogenesis, thrombogenesis, and ischemia by increasing plasma and blood viscosity (Kaptoge et al., 2007). T is negatively correlated with fibrinogen (Phillips et al., 1994). A recent study found that patients with prostate cancer on androgen deprivation therapy had elevated levels of fibrinogen compared to healthy controls (Ziaran, Goncalves, & Breza, 2013). In addition to fibrinogen, PAI-1, another clotting mediator and risk factor for ischemic heart disease, has been negatively correlated with T (Phillips et al., 1994; X. C. Yang et al., 1993).

Contrarily, a double-blinded, randomized placebo-controlled trial of T supplementation in men with chronic stable angina demonstrated no changes in fibrinogen or PAI-1, suggesting that T supplementation does not affect blood coagulation status (Smith et al., 2005). Moreover, a study comparing chemically or surgically castrated males to eugonadal controls showed that castrated men had less platelet thromboxane A2 receptors, suggesting that the inhibition of T production may attenuate platelet aggregation responses (Ajayi & Halushka, 2005).

Discussion

Hypogonadism is a growing concern of our aging population; it affects 20% of men older than 60 years, 30% of men older than 70 years, and 50% of men older than 80 years (Harman, Metter, Tobin, Pearson, & Blackman, 2001). Of concern, low T has been linked to increased risk for major adverse cardiac events in hypertensive patients (Vlachopoulos et al., 2013). With increased awareness, there has been great emphasis on studying the association of T to CVD. The current review of literature highlights the important role that T plays in CVD, pointing toward a potential beneficial effect of T on CVD with conflicting studies directed primarily at CAD.

Most of the studies examining T and CVD were observational, cross-sectional, or retrospective studies that cannot demonstrate cause and effect. Whether this inverse association between T and CVD is a result of T or a result of the disease process remains unclear. There also is no consensus on the cutoff T value of hypogonadism in the literature.

T supplementation results in positive short- and long-term physiological and biochemical changes in patients with CVD. Favorable effects have been demonstrated on myocardial ischemia, CHF exercise tolerance, and MetS. T supplementation may reduce the CV risk factors clustered in MetS without any significant hematological or prostate adverse effects (Aversa et al., 2010). Age-dependent decline in T levels in healthy men has also been associated with decreased muscle mass and strength (Izquierdo et al., 2001; van den Beld, de Jong, Grobbee, Pols, & Lamberts, 2000). T therapy, especially when combined with strength training, has been associated with increased skeletal muscle bulk, physical performance, and leg strength (Bhasin et al., 1996; Page et al., 2005; Storer et al., 2003).

Other than the TOM trial (Basaria et al., 2010), clinical trials and meta-analyses investigating the benefits and risks of T therapy have not demonstrated significant adverse CV events in the acute setting. Although the results in Shores et al. (2012) are reassuring, the more recent study by Vigen et al. (2013) raises real concerns regarding the long-term risks of T supplementation therapy. Both studies incorporated multivariate analyses to guard against potential confounders with the latter study utilizing a sophisticated weighted analysis to further improve their analyses. However, as with any retrospective study, unmeasured bias and error may be present. Future randomized control studies are needed to better delineate the risks and benefits of T supplementation in CVD and establish the optimal protocol for T therapy while comparing acute versus chronic adverse effects.

Despite many hypotheses, the specific mechanism by which T protects against CVD is yet to be elucidated. T has been reported to induce vasodilation via both endothelium-derived nitric oxide and ATP-sensitive potassium channels on vascular SMC membranes. This vasodilatory effect may explain its effects on preventing myocardial ischemia and increasing exercise tolerance in patients with CHF. T may also work via decreasing inflammation, protecting endothelial function, decreasing arterial stiffness, and mediating clotting factors to prevent progression of atherosclerosis and improve CV function. The number and variety of mechanisms highlight the important role of androgens in the vascular system.

Conclusion

CVD is the leading cause of death in the United States Review of modern literature suggests a CV -protective effect of T in men. Emerging data have begun to shed light on the mechanism of action of androgens on the CV system. T therapy can reverse symptoms of hypogonadism while potentially reducing CVD risk factors. However, due to conflicting data, future studies should further explore the long-term safety of T therapy. Only by correcting modifiable risk factors and restoring androgen imbalance can optimal management of CVD be achieved.