madman
Super Moderator
The Effects of Testosterone Treatment on Cardiovascular Health (2022)
Channa N. Jayasena, MA, Ph.D., MRCP, FRCPatha, Carmen Lok Tung Ho, BSc, Shalender Bhasin, MB, BS
INTRODUCTION
The prescription rates for testosterone products have risen markedly over the last 20 years, due to multiple factors, including heightened awareness about TRT because of direct-to-consumer pharmaceutical marketing, media coverage, and increased off-label use of testosterone for middle-aged and older men with age-related conditions, such as obesity and type 2 diabetes mellitus. However, the cardiovascular safety of long-term TRT remains unknown because of insufficient RCT data and conflicting evidence from epidemiologic, pharmacovigilance, retrospective studies, and small trials. This has affected prescribing behavior among clinicians, leading to disparities in the treatment of men with hypogonadism. This article will provide a critical summary of the available evidence of the cardiovascular effects and safety of TRT. Areas of controversy and gaps in our current evidence are highlighted along with a synthesis of the available evidence.
PHYSIOLOGIC EFFECTS OF TESTOSTERONE ON CARDIOVASCULAR HEALTH
Testosterone exerts several diverse effects on cardiovascular physiology; some of these physiologic effects may increase the risk of cardiovascular events while others may reduce cardiovascular risk. Androgen receptors (ARs) are located in cardiac myocytes, vascular smooth muscle, and vascular endothelial cells.1–4
Testosterone exerts some potentially beneficial effects on the cardiovascular system. Testosterone is a potent vasodilator; it inhibits L-type calcium channels, resulting in coronary vasodilatation and increased coronary blood flow.5,6 DHT is more potent than testosterone in mediating these nongenomic effects on vascular smooth muscle relaxation. Testosterone improves endothelial function, reduces vascular reactivity,7, and shortens QTc interval.8 Furthermore, testosterone administration decreases whole body, subcutaneous, and intraabdominal fat.9,10
In mice, orchiectomy increases sarcoplasmic reticulum (SR) calcium load within ventricular myocytes and the expression of SERCA-2a which is implicated in the preservation of ventricular function after myocardial infarction.11 Testosterone supplementation was associated with left ventricle dysfunction in orchiectomised mice. Testosterone administration by downregulating SERCA-2a expression causes reduced SR calcium accumulation11 thereby attenuating the cardiac inotropic response.12
Several physiologic effects of testosterone could potentially increase the risk of cardiovascular events. As discussed later, testosterone administration reduces plasma HDL cholesterol depending on the administered dose, the route of administration.13,14 Testosterone induces platelet aggregation by stimulating thromboxane A215 and promotes sodium and water retention,16 which can contribute to edema formation and worsen preexisting heart failure. In preclinical models, testosterone promotes smooth muscle proliferation17 and increases the expression of vascular cell adhesion molecule.18 Testosterone increases hematocrit19 by stimulating iron-dependent erythropoiesis by suppressing hepcidin,20,21 increasing erythropoietin,22 and by direct effects on the bone marrow to increase the numbers of erythropoietic progenitors. Older men experience greater increments in hematocrit than younger men.23
Testosterone administration increases the levels of prothrombotic as well as antithrombotic factors. It does not significantly affect myocardial infarct size in preclinical models of myocardial infarction.24 Testosterone has been shown to retard atherosclerosis in some preclinical models25 but not in others,26 and induce myocardial hypertrophy in some mouse strains,4,24 but not in others.
EFFECTS OF TESTOSTERONE ON BLOOD PRESSURE
Blood pressure is higher in men when compared with women.27–29 Testosterone may play a role in the sex differences in BP, which only appears in boys and girls after puberty.30 Orchiectomy or administration of the AR antagonist, flutamide, attenuates the development of salt-induced hypertension has been observed in the male rats.31 However, the 5-alpha-reductase inhibitor, finasteride, does not affect the onset of hypertension, suggesting that testosterone’s conversion to DHT may not be implicated in the development of hypertension in this model.30 Female rats treated with testosterone during the neonatal period, develop higher blood pressures than control animals.31 Furthermore, women with elevated androgens due to virilizing tumors or polycystic ovarian syndrome (PCOS) have higher BP when compared with age-matched women without PCOS.30 AR signaling increases the activity of the renin-angiotensin-aldosterone (RAA) system; male rats have higher plasma renin activity versus females, and castration reduces plasma renin activity in male rats.30 Testosterone exposure also increases renal angiotensinogen mRNA.25 Testosterone administration is associated with transient sodium and water retention in men and women in the first few weeks after starting testosterone treatment and some men, especially older men, with hypogonadism experience edema during TRT.32 However, clinical data suggest that the effects of testosterone on BP are complex; hypogonadal men have been observed to have higher systolic blood pressures than eugonadal men.33 Although most testosterone studies have not reported an increase in BP during testosterone treatment,32,34 recent studies of oral testosterone undecanoate that performed standardized measurements of blood pressure during the clinic visits as well as ambulatory blood pressure measurement found that the BP more than 24 hours was higher following 120 and 180 days of treatment with oral testosterone undecanoate than at baseline; the effects on diastolic BP more than 24 h were less than for the systolic BP.35–37 The US Food and Drug Administration has required a boxed warning on oral testosterone undecanoate labeling stating that the drug can cause blood pressure to rise.38
EFFECTS OF TESTOSTERONE ON SERUM LIPID PARAMETERS
In epidemiologic studies, low testosterone levels are generally associated with a proatherogenic lipid profile and higher HDL cholesterol39–42; some studies have also reported a negative association between circulating testosterone and VLDL cholesterol.43 Testosterone levels are positively associated with smaller or less atherogenic VLDL particles44 and higher testosterone levels have also been associated with a lower apoB to apo A-1 ratio.45
The intervention studies generally have reported modest reductions in total and high-density lipoprotein (HDL) cholesterol during TRT32,34,46,47 in men with hypogonadism; 2 of these studies reported significant reductions in low-density lipoprotein (LDL) cholesterol in hypogonadal men with T2DM.34,46 However, some RCTs have failed to observe any significant changes in total, LDL or HDL cholesterol during treatment with transdermal testosterone.48–52 Nonaromatizable oral androgens suppress HDL cholesterol substantially more than transdermal or injectable testosterone esters. Testosterone-induced suppression of HDL cholesterol is associated with the upregulation of hepatic triacylglycerol lipase, changes in HDL proteome, and suppression of apolipoprotein A1,53 but does not seem to reduce the cholesterol efflux capacity of HDL particles. TRT seems to have no significant effect on serum triglyceride levels. In summary, overall, TRT is generally associated with a modest reduction in total cholesterol and HDL cholesterol without a concomitant reduction in LDL cholesterol.
EFFECTS OF TESTOSTERONE ON VENOUS THROMBOEMBOLISM
Testosterone stimulates erythropoiesis by multiple mechanisms20,21 and increases hematocrit.19 Erythrocytosis is the most frequent adverse event associated with TRT in randomized trials.54 However, there is a paucity of high-quality evidence of an association between testosterone replacement therapy and venous thromboembolism (VTE) in men with hypogonadism.21,55 To date, no RCTs of TRT administration have captured significant numbers of events to accurately determine VTE risk in men with hypogonadism. A case-control study reported an increased risk of VTE in the first 6 months following commencement of testosterone treatment.56 Recently, the IBM MarketScan Commercial Claims and Encounter Database and the Medicare Supplemental Database were used to compare VTE events within 39,622 men during 1, 3, and 6 months before TRT commencement versus 1, 3, and 6 months after starting TRT.57 This study suggested that men with hypogonadism have a twofold increased risk of VTE within the 6 months following TRT commencement. However, the number of confirmed VTE events in RCTs has been exceedingly small. Most of the reported VTE events in published case reports have occurred in men with preexisting hypercoagulable conditions.58 It is prudent to consider VTE risk and counsel men with hypogonadism appropriately when considering TRT.
EFFECTS OF TESTOSTERONE ON GLUCOSE INTOLERANCE AND TYPE 2 DIABETES
In epidemiologic studies, low total testosterone levels are associated with increased visceral fat volume,59 serum glucose concentration,60, and increased risk of type 2 diabetes mellitus (T2DM) both cross-sectionally and longitudinally.61 The association of free testosterone and T2DM has been inconsistent; some studies have reported a weak association62 while others have failed to find any relation.63 The lack of a strong correlation between free testosterone and T2DM suggests that SHBG may be the primary determinant of the observed relation between total testosterone levels and T2DM. Indeed, circulating SHBG level is an independent predictor of incident T2DM even after adjustment for free or total testosterone levels. Polymorphisms of the SHBG gene that are associated with low SHBG levels are associated with an increased risk of T2DM.64
In Mendelian randomization studies, higher genetically determined testosterone levels are associated with the risk of T2DM in a sexually dimorphic manner; in men, higher genetically determined testosterone levels are associated with lower risk of T2DM, but in women, higher genetically determined testosterone levels are associated with increased risk of T2DM.65
The effects of testosterone on insulin sensitivity have been inconsistent across studies. In general studies of men in whom severe testosterone deficiency was induced rapidly by the acute withdrawal of testosterone replacement therapy in men known to have hypogonadism66 develop insulin resistance. Similarly, men receiving with prostate who receive androgen deprivation therapy are at increased risk of developing impaired glucose tolerance, insulin resistance, and T2DM.67–69 The worsening of insulin sensitivity associated with the development of severe testosterone deficiency may be related in part to a loss of skeletal muscle mass, increase in the whole body and visceral fat mass, and to the effects of testosterone on lipid oxidation and mitochondrial function.70 Dhindsa and colleagues performed euglycaemic hyperinsulinemia clamp studies showing that 3 weeks of testosterone administration have no detectable effects on insulin sensitivity or other glucose parameters in men with type 2 diabetes; however, 24 weeks of testosterone administration were associated with significant changes in body composition and improvement in insulin sensitivity.71 Changes in insulin sensitivity was accompanied by reductions in circulating free fatty acids (FFAs)71 and increased adipose expression of insulin signaling markers such as insulin receptor b subunit, insulin reception substrate (IRS) 1, protein kinase B (AKT-2), and glucose transporter types 4 (GLUT-4).
However, well-controlled randomized trials of testosterone treatment that recruited men with mild testosterone deficiency or with low normal testosterone levels have failed to find consistent improvements in insulin sensitivity with testosterone treatment. For example, in the testosterone trials (TTrials),72 testosterone treatment of older men with low testosterone levels and one or more symptoms of testosterone deficiency was associated with a small reduction in insulin but not glucose levels and only a small change in HOMA-IR. In another study, 2 years of testosterone treatment in elderly men with low or low normal testosterone levels did not improve carbohydrate tolerance, insulin secretion, insulin action, glucose effectiveness, hepatic insulin clearance, or the pattern of postprandial glucose metabolism.73 Another placebo-controlled randomized trial also found no significant improvement in insulin sensitivity after 3 years of testosterone treatment relative to placebo in middle-aged and older men with low or low normal testosterone levels.74
The clinical effects of testosterone on diabetes risk and diabetes prevention are covered in detail by Yeap & Wittert elsewhere in this issue. In the T4DM Trial, testosterone treatment administered in combination with a lifestyle intervention for 2 years of men, 50 to 74 years, with impaired glucose tolerance or newly diagnosed type 2 diabetes, but without symptomatic testosterone deficiency, reduced the proportion of randomized men with type 2 diabetes beyond the effects of the lifestyle intervention.75 However, the study participants in this large well-conducted trial were not hypogonadal. In the TIMES2 trial,34 testosterone treatments did not consistently improve hemoglobin A1c in hypogonadal men with T2DM or metabolic syndrome. Thus, in spite of the strong association of low testosterone levels with increased risk of T2DM in epidemiologic studies, randomized intervention trials in hypogonadal men have not provided clear evidence of improvement in glycemic control, prevention of progression from prediabetes to diabetes, or diabetes remission.
THE EFFECTS OF TESTOSTERONE TREATMENT ON THE RISK OF MAJOR ADVERSE CARDIOVASCULAR EVENTS
Epidemiologic Studies
The relation of testosterone levels and coronary artery disease in cross-sectional and prospective cohort studies has been inconsistent. Some cross-sectional studies have shown low levels of testosterone to be associated with increased risk for coronary artery disease,76 while others have shown no association.77 The relationship between serum testosterone levels and the incidence of cardiovascular events also has been inconsistent in prospective epidemiologic studies.
Epidemiologic studies have found a consistent negative association between circulating testosterone concentrations and common carotid artery intima-media thickness, a measure of subclinical atherosclerosis. For example, in the Rotterdam Study, the men in the lowest quartile of testosterone levels had a greater progression of intima-media thickness than men in the highest quartile of testosterone levels.78 However, the same study did not identify a significant difference in the rates of change in coronary artery calcium between the group administrated with testosterone or placebo.78
The relation of testosterone and mortality has been heterogeneous across studies. A meta-analysis by Corona and colleagues reported an association of low testosterone levels with an increased risk of cardiovascular disease. Furthermore, study participants with the lowest testosterone levels seemed to have the highest overall mortality and cardiovascular mortality.79 Another meta-analysis of 11 randomized trials by Araujo and colleagues found that in aggregate, lower testosterone levels were associated with a higher risk of all-cause mortality, especially cardiovascular mortality.80 Epidemiologic studies can only show association but cannot prove causality; reverse causality cannot be excluded. It is possible that testosterone is a marker of health, and those who are at a higher risk of dying have lower testosterone levels. Ruige and colleagues found that higher testosterone levels were associated with a lower risk for cardiovascular events in men more than 70 years of age but not in men who were younger than this age group.77
Pharmacovigilance Studies and Retrospective Analyses of Electronic Medical Records
The pharmacovigilance studies and retrospective analyses of electronic medical records have yielded inconsistent results because of their inherent limitations. In a retrospective analysis of male veterans who underwent angiography and had low testosterone concentrations, Vigen and colleagues observed that TRT was associated with an increased risk of the composite cardiovascular outcome of myocardial infarction, stroke, and death when (hazard ratio, HR 5 1.29)81 relative to no TRT. Finkle used an insurance database and found an increased risk of nonfatal myocardial infarction during the 90 days following initial prescription for TRT when compared with the period before commencing TRT (relative risk (RR) 5 1.36).82 However, another retrospective study of men with low testosterone concluded that TRT was associated with reduced all-cause mortality when compared with no TRT (HR 5 0.61).83 Muraleedharan and colleagues retrospectively concluded that low serum testosterone may predict increased all-cause mortality in 581 men with type 2 diabetes (HR 2.3).84 Furthermore, Boden and colleagues conducted a post hoc analysis of the AIM-HIGH trial of men with metabolic syndrome and low baseline levels of HDL cholesterol. The 643 out of 2118 men with levels of serum testosterone less than 300 ng/dL had a higher risk of the primary composite outcome (coronary heart disease, death, MI, stroke, hospitalization for acute coronary syndrome, or coronary or cerebral revascularization) when compared with the normal testosterone group (HR 1.23).85 Taken collectively, observational studies provide little consensus on TRT safety and have each been used to substantiate claims TRT either reduces or increases the risk of MACE outcomes.
These pharmacovigilance and retrospective studies suffer from many limitations, including heterogeneous study populations and differences in treatment indications, treatment regimens and duration, on-treatment testosterone levels, and other aspects of study design. These studies used variable definitions of cardiovascular outcomes that were often not prespecified, and the ascertainment methods varied across studies. They also suffered from a potential for residual confounding in that the patients assigned to testosterone therapy differed from comparators in baseline cardiovascular risk factors. Due to the inherent limitations and inconsistency of findings, these pharmacovigilance and retrospective analyses do not permit strong inferences about the relation between testosterone therapy and mortality and cardiovascular outcomes.
Randomized Controlled Trials
The testosterone replacement for older men with sarcopenia (TOM) randomized controlled trial was designed to investigate functional mobility following TRT in 209 men with hypogonadism and frailty.32 However, the study was stopped early by its data and safety monitoring board due to an unexpected increase in cardiovascular events within the TRT versus the placebo arm, albeit with a small absolute number of events (23 vs 5, respectively). However, the cardiovascular events were not prespecified nor adjudicated prospectively. The number of major adverse cardiovascular events (MACEs) was small. Subsequent RCTs have often excluded men with the increased baseline cardiovascular risk, so it is unsurprising that the number of MACE has been small in most trials. Several meta-analyses have examined the association between testosterone replacement and cardiovascular events, major cardiovascular events, and death in RCTs, and overall, these meta-analyses have not shown a statistically significant association between testosterone and cardiovascular events, major cardiovascular events, or deaths.23,86–88 These meta-analyses are limited by the heterogeneity of randomized trials included in these analyses with respect to eligibility criteria, testosterone dose, formulation, and intervention durations. The variable quality of adverse event recording in clinical trials has been well documented and was particularly apparent in these trials, which reported a very low frequency of all adverse events as well cardiovascular events. The small size of many trials and the inclusion of pilot studies with very small sample sizes was another constraint. Cardiovascular outcomes were not prespecified, they were often defined post hoc, and were of varying clinical significance. The major cardiovascular events were not adjudicated, not specified prospectively, and the total number of major cardiovascular events was too small to draw strong inferences. None of the trials has been large enough or long enough to determine the effects of testosterone treatment on MACE.
Two randomized trials—the Cardiovascular Trial of the TTrials32 and the Testosterone Effects on Atherosclerosis in Aging Men (The TEAAM Trial)78 determined the effects of testosterone treatment relative to placebo on the rate of atherogenesis progression. The rate of atherosclerosis progressed assessed using the common carotid artery -intima-media thickness or the coronary calcium scores did not differ between testosterone-treated and placebo-treated men in either of the 2 trials. However, in the Cardiovascular Trial of the TTrials,32 testosterone treatment was associated with a greater increase in the volume of noncalcified plaque in the coronary arteries, assessed using computed tomography angiography, compared with placebo.
An extensive review by the FDA concluded that “the studies...have significant limitations that weaken their evidentiary value for confirming a causal relationship between testosterone and adverse cardiovascular outcomes.” Nevertheless, the US Food and Drug Administration (FDA) directed the pharmaceutical companies to include in the label warning about the potential cardiovascular risks of TRT.89 The European Medicines Agency also found no conclusive link between testosterone treatment and cardiovascular risk. Fortunately, 2 ongoing studies are aiming to close the evidence gap. The National Institute for Health Research (NIHR) TestES (Testosterone Effects and Safety) consortium is currently collating individual patient data (IPD) and adverse events from published RCTs to analyze the risks of subtypes of MACE within men with hypogonadism treated with TRT when compared with placebo (https://www.imperial. ac.uk/metabolism-digestion-reproduction/research/diabetes-endocrinologymetabolism/endocrinology-and-investigative-medicine/nihr-testosterone/). Furthermore, the Phase 4, randomized placebo-controlled trial (The TRAVERSE Trial) is recruiting approximately 6000 men aged 45 to 80 years with either preexisting cardiovascular disease or at least 3 cardiovascular risk factors, with the primary objective of comparing the effect of TRT versus placebo on the incidence of MACE.
SUMMARY
Overall, TRT exerts multiple physiologic effects, both positive and negative, on cardiovascular health (Table 1). Finally, there are insufficient RCT data to determine whether TRT increases MACE risk. Studies are underway to clarify this important question. The Endocrine Society’s testosterone treatment guideline recommends avoiding testosterone treatment in hypogonadal men who incurred a MACE in the preceding 6 months and in men with a known hypercoagulable condition, such as a mutation on antithrombin 3, protein C or protein S. Testosterone treatment of hypogonadal men with increased risk of cardiovascular events requires consideration and counseling of the potential risks versus benefits of testosterone replacement therapy.
Channa N. Jayasena, MA, Ph.D., MRCP, FRCPatha, Carmen Lok Tung Ho, BSc, Shalender Bhasin, MB, BS
INTRODUCTION
The prescription rates for testosterone products have risen markedly over the last 20 years, due to multiple factors, including heightened awareness about TRT because of direct-to-consumer pharmaceutical marketing, media coverage, and increased off-label use of testosterone for middle-aged and older men with age-related conditions, such as obesity and type 2 diabetes mellitus. However, the cardiovascular safety of long-term TRT remains unknown because of insufficient RCT data and conflicting evidence from epidemiologic, pharmacovigilance, retrospective studies, and small trials. This has affected prescribing behavior among clinicians, leading to disparities in the treatment of men with hypogonadism. This article will provide a critical summary of the available evidence of the cardiovascular effects and safety of TRT. Areas of controversy and gaps in our current evidence are highlighted along with a synthesis of the available evidence.
PHYSIOLOGIC EFFECTS OF TESTOSTERONE ON CARDIOVASCULAR HEALTH
Testosterone exerts several diverse effects on cardiovascular physiology; some of these physiologic effects may increase the risk of cardiovascular events while others may reduce cardiovascular risk. Androgen receptors (ARs) are located in cardiac myocytes, vascular smooth muscle, and vascular endothelial cells.1–4
Testosterone exerts some potentially beneficial effects on the cardiovascular system. Testosterone is a potent vasodilator; it inhibits L-type calcium channels, resulting in coronary vasodilatation and increased coronary blood flow.5,6 DHT is more potent than testosterone in mediating these nongenomic effects on vascular smooth muscle relaxation. Testosterone improves endothelial function, reduces vascular reactivity,7, and shortens QTc interval.8 Furthermore, testosterone administration decreases whole body, subcutaneous, and intraabdominal fat.9,10
In mice, orchiectomy increases sarcoplasmic reticulum (SR) calcium load within ventricular myocytes and the expression of SERCA-2a which is implicated in the preservation of ventricular function after myocardial infarction.11 Testosterone supplementation was associated with left ventricle dysfunction in orchiectomised mice. Testosterone administration by downregulating SERCA-2a expression causes reduced SR calcium accumulation11 thereby attenuating the cardiac inotropic response.12
Several physiologic effects of testosterone could potentially increase the risk of cardiovascular events. As discussed later, testosterone administration reduces plasma HDL cholesterol depending on the administered dose, the route of administration.13,14 Testosterone induces platelet aggregation by stimulating thromboxane A215 and promotes sodium and water retention,16 which can contribute to edema formation and worsen preexisting heart failure. In preclinical models, testosterone promotes smooth muscle proliferation17 and increases the expression of vascular cell adhesion molecule.18 Testosterone increases hematocrit19 by stimulating iron-dependent erythropoiesis by suppressing hepcidin,20,21 increasing erythropoietin,22 and by direct effects on the bone marrow to increase the numbers of erythropoietic progenitors. Older men experience greater increments in hematocrit than younger men.23
Testosterone administration increases the levels of prothrombotic as well as antithrombotic factors. It does not significantly affect myocardial infarct size in preclinical models of myocardial infarction.24 Testosterone has been shown to retard atherosclerosis in some preclinical models25 but not in others,26 and induce myocardial hypertrophy in some mouse strains,4,24 but not in others.
EFFECTS OF TESTOSTERONE ON BLOOD PRESSURE
Blood pressure is higher in men when compared with women.27–29 Testosterone may play a role in the sex differences in BP, which only appears in boys and girls after puberty.30 Orchiectomy or administration of the AR antagonist, flutamide, attenuates the development of salt-induced hypertension has been observed in the male rats.31 However, the 5-alpha-reductase inhibitor, finasteride, does not affect the onset of hypertension, suggesting that testosterone’s conversion to DHT may not be implicated in the development of hypertension in this model.30 Female rats treated with testosterone during the neonatal period, develop higher blood pressures than control animals.31 Furthermore, women with elevated androgens due to virilizing tumors or polycystic ovarian syndrome (PCOS) have higher BP when compared with age-matched women without PCOS.30 AR signaling increases the activity of the renin-angiotensin-aldosterone (RAA) system; male rats have higher plasma renin activity versus females, and castration reduces plasma renin activity in male rats.30 Testosterone exposure also increases renal angiotensinogen mRNA.25 Testosterone administration is associated with transient sodium and water retention in men and women in the first few weeks after starting testosterone treatment and some men, especially older men, with hypogonadism experience edema during TRT.32 However, clinical data suggest that the effects of testosterone on BP are complex; hypogonadal men have been observed to have higher systolic blood pressures than eugonadal men.33 Although most testosterone studies have not reported an increase in BP during testosterone treatment,32,34 recent studies of oral testosterone undecanoate that performed standardized measurements of blood pressure during the clinic visits as well as ambulatory blood pressure measurement found that the BP more than 24 hours was higher following 120 and 180 days of treatment with oral testosterone undecanoate than at baseline; the effects on diastolic BP more than 24 h were less than for the systolic BP.35–37 The US Food and Drug Administration has required a boxed warning on oral testosterone undecanoate labeling stating that the drug can cause blood pressure to rise.38
EFFECTS OF TESTOSTERONE ON SERUM LIPID PARAMETERS
In epidemiologic studies, low testosterone levels are generally associated with a proatherogenic lipid profile and higher HDL cholesterol39–42; some studies have also reported a negative association between circulating testosterone and VLDL cholesterol.43 Testosterone levels are positively associated with smaller or less atherogenic VLDL particles44 and higher testosterone levels have also been associated with a lower apoB to apo A-1 ratio.45
The intervention studies generally have reported modest reductions in total and high-density lipoprotein (HDL) cholesterol during TRT32,34,46,47 in men with hypogonadism; 2 of these studies reported significant reductions in low-density lipoprotein (LDL) cholesterol in hypogonadal men with T2DM.34,46 However, some RCTs have failed to observe any significant changes in total, LDL or HDL cholesterol during treatment with transdermal testosterone.48–52 Nonaromatizable oral androgens suppress HDL cholesterol substantially more than transdermal or injectable testosterone esters. Testosterone-induced suppression of HDL cholesterol is associated with the upregulation of hepatic triacylglycerol lipase, changes in HDL proteome, and suppression of apolipoprotein A1,53 but does not seem to reduce the cholesterol efflux capacity of HDL particles. TRT seems to have no significant effect on serum triglyceride levels. In summary, overall, TRT is generally associated with a modest reduction in total cholesterol and HDL cholesterol without a concomitant reduction in LDL cholesterol.
EFFECTS OF TESTOSTERONE ON VENOUS THROMBOEMBOLISM
Testosterone stimulates erythropoiesis by multiple mechanisms20,21 and increases hematocrit.19 Erythrocytosis is the most frequent adverse event associated with TRT in randomized trials.54 However, there is a paucity of high-quality evidence of an association between testosterone replacement therapy and venous thromboembolism (VTE) in men with hypogonadism.21,55 To date, no RCTs of TRT administration have captured significant numbers of events to accurately determine VTE risk in men with hypogonadism. A case-control study reported an increased risk of VTE in the first 6 months following commencement of testosterone treatment.56 Recently, the IBM MarketScan Commercial Claims and Encounter Database and the Medicare Supplemental Database were used to compare VTE events within 39,622 men during 1, 3, and 6 months before TRT commencement versus 1, 3, and 6 months after starting TRT.57 This study suggested that men with hypogonadism have a twofold increased risk of VTE within the 6 months following TRT commencement. However, the number of confirmed VTE events in RCTs has been exceedingly small. Most of the reported VTE events in published case reports have occurred in men with preexisting hypercoagulable conditions.58 It is prudent to consider VTE risk and counsel men with hypogonadism appropriately when considering TRT.
EFFECTS OF TESTOSTERONE ON GLUCOSE INTOLERANCE AND TYPE 2 DIABETES
In epidemiologic studies, low total testosterone levels are associated with increased visceral fat volume,59 serum glucose concentration,60, and increased risk of type 2 diabetes mellitus (T2DM) both cross-sectionally and longitudinally.61 The association of free testosterone and T2DM has been inconsistent; some studies have reported a weak association62 while others have failed to find any relation.63 The lack of a strong correlation between free testosterone and T2DM suggests that SHBG may be the primary determinant of the observed relation between total testosterone levels and T2DM. Indeed, circulating SHBG level is an independent predictor of incident T2DM even after adjustment for free or total testosterone levels. Polymorphisms of the SHBG gene that are associated with low SHBG levels are associated with an increased risk of T2DM.64
In Mendelian randomization studies, higher genetically determined testosterone levels are associated with the risk of T2DM in a sexually dimorphic manner; in men, higher genetically determined testosterone levels are associated with lower risk of T2DM, but in women, higher genetically determined testosterone levels are associated with increased risk of T2DM.65
The effects of testosterone on insulin sensitivity have been inconsistent across studies. In general studies of men in whom severe testosterone deficiency was induced rapidly by the acute withdrawal of testosterone replacement therapy in men known to have hypogonadism66 develop insulin resistance. Similarly, men receiving with prostate who receive androgen deprivation therapy are at increased risk of developing impaired glucose tolerance, insulin resistance, and T2DM.67–69 The worsening of insulin sensitivity associated with the development of severe testosterone deficiency may be related in part to a loss of skeletal muscle mass, increase in the whole body and visceral fat mass, and to the effects of testosterone on lipid oxidation and mitochondrial function.70 Dhindsa and colleagues performed euglycaemic hyperinsulinemia clamp studies showing that 3 weeks of testosterone administration have no detectable effects on insulin sensitivity or other glucose parameters in men with type 2 diabetes; however, 24 weeks of testosterone administration were associated with significant changes in body composition and improvement in insulin sensitivity.71 Changes in insulin sensitivity was accompanied by reductions in circulating free fatty acids (FFAs)71 and increased adipose expression of insulin signaling markers such as insulin receptor b subunit, insulin reception substrate (IRS) 1, protein kinase B (AKT-2), and glucose transporter types 4 (GLUT-4).
However, well-controlled randomized trials of testosterone treatment that recruited men with mild testosterone deficiency or with low normal testosterone levels have failed to find consistent improvements in insulin sensitivity with testosterone treatment. For example, in the testosterone trials (TTrials),72 testosterone treatment of older men with low testosterone levels and one or more symptoms of testosterone deficiency was associated with a small reduction in insulin but not glucose levels and only a small change in HOMA-IR. In another study, 2 years of testosterone treatment in elderly men with low or low normal testosterone levels did not improve carbohydrate tolerance, insulin secretion, insulin action, glucose effectiveness, hepatic insulin clearance, or the pattern of postprandial glucose metabolism.73 Another placebo-controlled randomized trial also found no significant improvement in insulin sensitivity after 3 years of testosterone treatment relative to placebo in middle-aged and older men with low or low normal testosterone levels.74
The clinical effects of testosterone on diabetes risk and diabetes prevention are covered in detail by Yeap & Wittert elsewhere in this issue. In the T4DM Trial, testosterone treatment administered in combination with a lifestyle intervention for 2 years of men, 50 to 74 years, with impaired glucose tolerance or newly diagnosed type 2 diabetes, but without symptomatic testosterone deficiency, reduced the proportion of randomized men with type 2 diabetes beyond the effects of the lifestyle intervention.75 However, the study participants in this large well-conducted trial were not hypogonadal. In the TIMES2 trial,34 testosterone treatments did not consistently improve hemoglobin A1c in hypogonadal men with T2DM or metabolic syndrome. Thus, in spite of the strong association of low testosterone levels with increased risk of T2DM in epidemiologic studies, randomized intervention trials in hypogonadal men have not provided clear evidence of improvement in glycemic control, prevention of progression from prediabetes to diabetes, or diabetes remission.
THE EFFECTS OF TESTOSTERONE TREATMENT ON THE RISK OF MAJOR ADVERSE CARDIOVASCULAR EVENTS
Epidemiologic Studies
The relation of testosterone levels and coronary artery disease in cross-sectional and prospective cohort studies has been inconsistent. Some cross-sectional studies have shown low levels of testosterone to be associated with increased risk for coronary artery disease,76 while others have shown no association.77 The relationship between serum testosterone levels and the incidence of cardiovascular events also has been inconsistent in prospective epidemiologic studies.
Epidemiologic studies have found a consistent negative association between circulating testosterone concentrations and common carotid artery intima-media thickness, a measure of subclinical atherosclerosis. For example, in the Rotterdam Study, the men in the lowest quartile of testosterone levels had a greater progression of intima-media thickness than men in the highest quartile of testosterone levels.78 However, the same study did not identify a significant difference in the rates of change in coronary artery calcium between the group administrated with testosterone or placebo.78
The relation of testosterone and mortality has been heterogeneous across studies. A meta-analysis by Corona and colleagues reported an association of low testosterone levels with an increased risk of cardiovascular disease. Furthermore, study participants with the lowest testosterone levels seemed to have the highest overall mortality and cardiovascular mortality.79 Another meta-analysis of 11 randomized trials by Araujo and colleagues found that in aggregate, lower testosterone levels were associated with a higher risk of all-cause mortality, especially cardiovascular mortality.80 Epidemiologic studies can only show association but cannot prove causality; reverse causality cannot be excluded. It is possible that testosterone is a marker of health, and those who are at a higher risk of dying have lower testosterone levels. Ruige and colleagues found that higher testosterone levels were associated with a lower risk for cardiovascular events in men more than 70 years of age but not in men who were younger than this age group.77
Pharmacovigilance Studies and Retrospective Analyses of Electronic Medical Records
The pharmacovigilance studies and retrospective analyses of electronic medical records have yielded inconsistent results because of their inherent limitations. In a retrospective analysis of male veterans who underwent angiography and had low testosterone concentrations, Vigen and colleagues observed that TRT was associated with an increased risk of the composite cardiovascular outcome of myocardial infarction, stroke, and death when (hazard ratio, HR 5 1.29)81 relative to no TRT. Finkle used an insurance database and found an increased risk of nonfatal myocardial infarction during the 90 days following initial prescription for TRT when compared with the period before commencing TRT (relative risk (RR) 5 1.36).82 However, another retrospective study of men with low testosterone concluded that TRT was associated with reduced all-cause mortality when compared with no TRT (HR 5 0.61).83 Muraleedharan and colleagues retrospectively concluded that low serum testosterone may predict increased all-cause mortality in 581 men with type 2 diabetes (HR 2.3).84 Furthermore, Boden and colleagues conducted a post hoc analysis of the AIM-HIGH trial of men with metabolic syndrome and low baseline levels of HDL cholesterol. The 643 out of 2118 men with levels of serum testosterone less than 300 ng/dL had a higher risk of the primary composite outcome (coronary heart disease, death, MI, stroke, hospitalization for acute coronary syndrome, or coronary or cerebral revascularization) when compared with the normal testosterone group (HR 1.23).85 Taken collectively, observational studies provide little consensus on TRT safety and have each been used to substantiate claims TRT either reduces or increases the risk of MACE outcomes.
These pharmacovigilance and retrospective studies suffer from many limitations, including heterogeneous study populations and differences in treatment indications, treatment regimens and duration, on-treatment testosterone levels, and other aspects of study design. These studies used variable definitions of cardiovascular outcomes that were often not prespecified, and the ascertainment methods varied across studies. They also suffered from a potential for residual confounding in that the patients assigned to testosterone therapy differed from comparators in baseline cardiovascular risk factors. Due to the inherent limitations and inconsistency of findings, these pharmacovigilance and retrospective analyses do not permit strong inferences about the relation between testosterone therapy and mortality and cardiovascular outcomes.
Randomized Controlled Trials
The testosterone replacement for older men with sarcopenia (TOM) randomized controlled trial was designed to investigate functional mobility following TRT in 209 men with hypogonadism and frailty.32 However, the study was stopped early by its data and safety monitoring board due to an unexpected increase in cardiovascular events within the TRT versus the placebo arm, albeit with a small absolute number of events (23 vs 5, respectively). However, the cardiovascular events were not prespecified nor adjudicated prospectively. The number of major adverse cardiovascular events (MACEs) was small. Subsequent RCTs have often excluded men with the increased baseline cardiovascular risk, so it is unsurprising that the number of MACE has been small in most trials. Several meta-analyses have examined the association between testosterone replacement and cardiovascular events, major cardiovascular events, and death in RCTs, and overall, these meta-analyses have not shown a statistically significant association between testosterone and cardiovascular events, major cardiovascular events, or deaths.23,86–88 These meta-analyses are limited by the heterogeneity of randomized trials included in these analyses with respect to eligibility criteria, testosterone dose, formulation, and intervention durations. The variable quality of adverse event recording in clinical trials has been well documented and was particularly apparent in these trials, which reported a very low frequency of all adverse events as well cardiovascular events. The small size of many trials and the inclusion of pilot studies with very small sample sizes was another constraint. Cardiovascular outcomes were not prespecified, they were often defined post hoc, and were of varying clinical significance. The major cardiovascular events were not adjudicated, not specified prospectively, and the total number of major cardiovascular events was too small to draw strong inferences. None of the trials has been large enough or long enough to determine the effects of testosterone treatment on MACE.
Two randomized trials—the Cardiovascular Trial of the TTrials32 and the Testosterone Effects on Atherosclerosis in Aging Men (The TEAAM Trial)78 determined the effects of testosterone treatment relative to placebo on the rate of atherogenesis progression. The rate of atherosclerosis progressed assessed using the common carotid artery -intima-media thickness or the coronary calcium scores did not differ between testosterone-treated and placebo-treated men in either of the 2 trials. However, in the Cardiovascular Trial of the TTrials,32 testosterone treatment was associated with a greater increase in the volume of noncalcified plaque in the coronary arteries, assessed using computed tomography angiography, compared with placebo.
An extensive review by the FDA concluded that “the studies...have significant limitations that weaken their evidentiary value for confirming a causal relationship between testosterone and adverse cardiovascular outcomes.” Nevertheless, the US Food and Drug Administration (FDA) directed the pharmaceutical companies to include in the label warning about the potential cardiovascular risks of TRT.89 The European Medicines Agency also found no conclusive link between testosterone treatment and cardiovascular risk. Fortunately, 2 ongoing studies are aiming to close the evidence gap. The National Institute for Health Research (NIHR) TestES (Testosterone Effects and Safety) consortium is currently collating individual patient data (IPD) and adverse events from published RCTs to analyze the risks of subtypes of MACE within men with hypogonadism treated with TRT when compared with placebo (https://www.imperial. ac.uk/metabolism-digestion-reproduction/research/diabetes-endocrinologymetabolism/endocrinology-and-investigative-medicine/nihr-testosterone/). Furthermore, the Phase 4, randomized placebo-controlled trial (The TRAVERSE Trial) is recruiting approximately 6000 men aged 45 to 80 years with either preexisting cardiovascular disease or at least 3 cardiovascular risk factors, with the primary objective of comparing the effect of TRT versus placebo on the incidence of MACE.
SUMMARY
Overall, TRT exerts multiple physiologic effects, both positive and negative, on cardiovascular health (Table 1). Finally, there are insufficient RCT data to determine whether TRT increases MACE risk. Studies are underway to clarify this important question. The Endocrine Society’s testosterone treatment guideline recommends avoiding testosterone treatment in hypogonadal men who incurred a MACE in the preceding 6 months and in men with a known hypercoagulable condition, such as a mutation on antithrombin 3, protein C or protein S. Testosterone treatment of hypogonadal men with increased risk of cardiovascular events requires consideration and counseling of the potential risks versus benefits of testosterone replacement therapy.
