Testosterone for secondary prevention in men with ischaemic heart disease?

Atherosclerotic coronary artery disease (CAD) is a leading cause of mortality and morbidity in the western world. It is a chronic progressive condition, and treatment is required indefinitely. Men are more than twice as likely as women to develop CAD:1 this ratio is consistent in all populations and is not related to differences in risk factors. Pre-menopausal females have a lower incidence of CAD, but this rises after the menopause, so that the risk of CAD rapidly approaches that of males after about 10 years. One explanation for these epidemiological observations is that female hormones protect against the development of CAD. Observational studies of hormone replacement therapies in women have suggested some benefit,2 but large randomized controlled trials have failed to show protective effects.3,4 Little attention has been paid to the role of testosterone in the pathogenesis of CAD. Males do not have a menopause equivalent, but sex hormones do fall with advancing age.5,6 The more elderly population, in which the prevalence of CAD is highest, has relatively low testosterone levels. Furthermore, males with CAD have lower testosterone levels than men with normal coronary angiograms males of the same age.7 Moreover, there is evidence that testosterone therapy delays the onset of cardiac ischaemia,8 probably as a consequence of a coronary vasodilator mechanism,8 improving the symptom of angina.8,10 The ability of testosterone replacement therapy to alter disease progression to reduce events or mortality has not been examined. In this article, we review the effects of testosterone therapy on factors known to influence cardiovascular risk, and propose a justification for further studies into its use for males with coronary disease as secondary prevention. Testosterone has well-established functions in the adult male. It induces secondary sexual characteristics, and is important for preserving sexual function, muscle strength, bone mineral density and mood in the post-pubertal male.6 The hormone has a circadian and circannual rhythm, with peak blood levels in the morning and in the spring, respectively. The majority (68%) is tightly bound to sex-hormone-binding globulin, with about 30% loosely bound to albumin and only 2–3% free. The biologically active moiety is the albumin and free testosterone; this is called the ‘bio-available’ portion.11 It is important to bear this in mind when interpreting the studies on testosterone in the literature. Earlier studies only measured total testosterone, which is influenced largely by the level of sex-hormone-binding globulin. As a carrier protein manufactured in the liver, this is in turn influenced by many factors. Hypogonadism, the state of testosterone deficiency, has a prevalence of 7% in the general population, rising to 20% in very elderly males.5 However, males with CAD have lower testosterone levels than unaffected males of the same age. In a small observational study comparing men presenting for scheduled angiography, the bio-available measurement of testosterone was significantly lower in males with CAD compared to men with normal coronaries.7 The mean bio-available testosterone of the CAD group was 2.55 nmol/l; the lower limit of normal range for the bio-available testosterone with this assay is 2.5 nmol/l. This would suggest that a significant proportion of males with CAD may have testosterone levels lower than the normal range, and that the prevalence of frank hypogonadism may be high in the CAD population. Larger studies on the prevalence of hypogonadism in males with CAD have yet to be published. However data have been presented to suggest that the prevalence of bio-chemical hypogonadism in these patients, as defined by a total testosterone < 7.5 nmol/l or a bio-available testosterone < 2.5 nmol/l, may be as high as 23.4%.12 Patients with overt hypogonadism clearly require treatment. However, given that a substantial proportion of men male with CAD appear to have relative androgen deficiency, would correcting this afford any benefit to them? Preliminary studies have suggested objective improvements in symptoms, but would male HRT have beneficial effects on established risk factors, and would this be translated into a clinically relevant reduction in cardiovascular endpoints? Different patterns of lipid levels and apoproteins are associated with CAD. In particular, high total cholesterol, high LDL fraction hypertriglyceridaemia and elevated lipoprotein-A are pro-atherogenic, whereas HDL is protective. The ratio of total cholesterol to HDL has been used to estimate risk in primary prevention in clinical practice.13 In patients with overt vascular disease, virtually all patients benefit from lipid-lowering with statins, whatever their baseline cholesterol level.14 The majority of cross-sectional studies have found a positive correlation of endogenous testosterone with HDL15–19 and a negative correlation with total cholesterol, LDL and triglycerides.20–23 Thus normal men with low testosterone appear to have adverse lipid profiles, and hypogonadal men have a potentially atherogenic dyslipidaemia prior to treatment.24 A few observational/cross-section studies have found no correlation with endogenous androgens and lipid profiles.25–27 It is difficult to account for this variation in the literature other than to comment on the relatively modest numbers of patients in the studies involved, and the use of total testosterone rather than free or bio-available measurements of testosterone.26 There has only been one major longitudinal follow up study of androgens and lipid levels. In the MRFIT study,28 serum was taken from patients at baseline and after an average of 13 years. A small reduction in testosterone level over time was reported and this correlated with a significant increase in triglycerides and HDL. These findings remained positive after controlling for obesity. No significant changes were reported in other lipid fractions but the reduction in testosterone was small. There have been a number of small studies of androgen therapy using several preparations of testosterone in a variety of populations. These studies can be grouped either in terms of their population (hypogonadal men, young men/athletes and healthy elderly men) or the level of androgen therapy (physiological or supra-physiological). Testosterone therapy in most populations causes a reduction in HDL,29–31 but this effect is smaller in elderly or hypogonadal men.32,33 This finding contrasts with data from cross-sectional and longitudinal studies. The relationship of androgens and HDL is complex. In the juvenile male, the rise of endogenous androgens at puberty causes a reduction in HDL.34 Experimental hypogonadism induced with GnRH agonists or withdrawal of maintenance testosterone causes an increase in HDL; this effect is suppressed if testosterone is administered at physiological doses.31,35 There is therefore evidence that androgens suppress HDL, but this is not supported by cross-sectional data. The reason for this disparity in the literature is probably related to obesity and insulin resistance. Testosterone has an inverse relationship with obesity,36,37 and obesity has an inverse relationship with HDL (see later section).16 Thus in cross-sectional analysis testosterone may appear to have a positive relation with HDL. This also explains why the reduction in HDL in young men or athletes is greater than that in old or hypogonadal men. Young men or athletes are lean and insulin-sensitive. Elderly or hypogonadal men have higher body mass index and a degree of insulin resistance—the testosterone has positive effects on body mass index and insulin resistance and this abrogates the reduction of HDL seen with testosterone therapy in the younger populations. There is a more consistent relationship with testosterone and other lipid fractions. Most studies of elderly or hypogonadal populations report reductions in total cholesterol, LDL and apo-protein B,18,32,33,38,39 although there has been at least one conflicting report.40 Testosterone treatment has also been shown to reduce lipoprotein-A level.41 In the hypogonadal and elderly cohorts, testosterone therapy is associated with reported reductions of up to 22% in total cholesterol and 15% LDL.33 This was recently confirmed in a meta-analysis of 19 trials encompassing 272 hypogonadal men receiving intra-muscular testosterone preparations. This study demonstrated significant reductions in total cholesterol and LDL fractions.42 There was also an expected dose-dependent reduction in HDL cholesterol, but this was small and less than that seen in normal men or young athletes. Testosterone replacement in elderly or hypogonadal men would thus appear to improve the dyslipidaemia associated with atherosclerosis. The effects may be sizeable (up to 15% reduction in LDL reported) and approach those seen in the large lipid-lowering trials (for example the 4S study with a 20% reduction in LDL).43 Hypertension is another established risk factor for CAD, and lowering high blood pressure reduces the risk of cardiovascular events. Testosterone is lower in populations of men with hypertension than in normal men,44 and case-controlled studies have demonstrated a relationship between hypotestosteronaemia and high blood pressure.20,22 However, the relationship of testosterone and blood pressure is uncertain. Testosterone has direct vasodilating properties in resistance arterioles by a calcium antagonistic action yet to be fully characterized.45 A study of testosterone therapy in men with heart failure reported an increase in cardiac output, primarily as a result of reducing peripheral vascular resistance, lending support to the action of testosterone as a natural vasodilator.47 In addition, a low testosterone:oestrodiol ratio is linked to activation of the renin-angiotensin system,48 which further increases peripheral vascular resistance. A handful of small trials have recorded blood pressure after testosterone therapy. None have reported increases, but some have reported modest reductions,49,50 of around 4–5 mmHg diastolic. These reductions compare favourably to the effect of anti-hypertensive treatments in major blood pressure trials.51 Myocardial infarction is caused by thrombosis of the coronary artery in 90% of cases. The trigger for the thrombosis occurs when a coronary plaque erodes or ruptures and the endothelium is breached. Thrombosis is a complicated process with its own intrinsic promoters and antagonists which together determine the prevailing coagulation status. Fibrinogen is the precursor of fibrin, which is the final component of the coagulation cascade and the building block of a stable clot. Fibrinogen is also an acute-phase protein and is raised in any disease characterized by inflammation. Some 19 studies have found an association between elevated fibrinogen and atherosclerotic disease, and fibrinogen has been confirmed in meta-analysis to be an independent risk factor for CAD (for review see reference 52). It is not known whether treatments directed at reducing fibrinogen reduce the risk of subsequent vascular disease. However, several of the treatments widely used in cardiac medicine such as fibrates, β-blockers, smoking cessation and physical exercise are associated with fibrinogen reduction,52 and all are associated with improved clinical outcomes and a reduced risk of acute events. Androgens and in particular testosterone have an inverse relationship with fibrinogen such that low endogenous testosterone is associated with elevated fibrinogen.53–55 Moreover, testosterone administration causes a significant fall in fibrinogen,56 which is similar to the reduction with fibrates, for example.57 It is not known how testosterone reduces fibrinogen. Androgens are metabolized extensively in the liver, and androgen therapy may alter coagulation protein production, for example a high factor VII has been associated with low serum testosterone.44,58 An alternative explanation is that testosterone by its intrinsic immune-modulating properties, damps down the inflammation in atherosclerotic plaques causing the reduction of acute-phase proteins in general, including fibrinogen. Testosterone has a number of actions on the haemostatic/fibrinolytic system separate from the effects on fibrinogen. Androgens and testosterone in particular have anti-thrombotic actions, and have been used as therapy in thrombotic diseases.59 Observational studies have consistently found a positive association between testosterone levels and tPA (a major endogenous stimulator of thrombolysis) and a negative association with factor VIIa (a coagulation protein) and PAI-1 (an inhibitor of thrombolysis).48,54,55 Evidence is emerging that PAI-1 in particular is of major importance in coronary disease, since elevated PAI-1 predicts myocardial infarction and progression of atheroma in stable CAD patients60,61 and is associated with a failure of reperfusion with thrombolysis following AMI.62 Hypogonadal men have elevated PAI-1; replacement of testosterone in hypogonadal men and therapy with testosterone or the adrenal androgen dehydroepiandrosterone in normal men leads to a reduction in circulating PAI-1 level.63,64 The fibrinolytic system is of critical importance in the pathogenesis of myocardial infarction. The goal of therapy in this medical emergency is to reopen the thrombosed artery. Therapeutic thrombolysis works by triggering or delivering a large external supply of tPA. Unfortunately, in about 60% of patients it is impossible to dissolve the clot and reopen the artery quickly. One of the reasons for this is thrombolysis resistance, and PAI-1 is one of the blood proteins responsible, because PAI-1 inhibits tPA and stabilizes the clot. During myocardial infarction, testosterone levels fall dramatically and this corresponds with a rise in PAI-1.65 In essence, all men suffering with AMI are temporarily hypogonadal, and this phenomenon leads to a prothrombotic state that may prevent thrombolysis or cause re-infarction. Obesity is a heterogeneous condition, and certain patterns of obesity are more strongly linked to ischaemic heart disease. In particular, abdominal or visceral obesity, as measured by waist:hip ratio, is most strongly linked to cardiac disease and is an independent coronary risk factor.66 Visceral adipose tissue is more highly metabolically active than non-visceral fat.67 Visceral adipocytes produce large amounts of free fatty acids, and their flux of free fatty acids in the portal blood can cause disturbances of hepatic metabolism. In particular, in obese men there is reduced insulin extraction by hepatocytes,68 increased hepatic insulin resistance69 and increased hepatic gluconeogenesis.70 These factors contribute to systemic hyperinsulinaemia and ultimately insulin resistance (see next section). Testosterone blood levels are inversely related to the degree of visceral obesity.37 Visceral fat has a high density of androgen receptors, and these appear to inhibit the action of lipoprotein lipase and fatty acid/triglyceride uptake; the androgen receptors thus limit fat accumulation.71 In the ageing male, the natural diminution of testosterone may allow an increase in fat mass (middle-aged spread), hypogonadal men are also recognized to have increased obesity. Moreover, adipose tissue extensively metabolizes testosterone to oestrodiol, thus obese men also have increased extraction of testosterone because of aromatization.72 Whatever the causal relationship of obesity and testosterone, there is little doubt that testosterone therapy in elderly, hypogonadal, or obese men reduces visceral obesity.50,73,74 Most of these studies also report beneficial effects on other aspects of the metabolic syndrome and an increase in lean mass. The ability of insulin to stimulate glucose uptake is variable. Insulin resistance at peripheral receptors is compensated by an exaggerated release of insulin. Insulin resistance and consequent hyperinsulinaemia are the key components of the metabolic syndrome.75 The other features are type 2 diabetes, hypertension, male pattern (visceral) obesity, impaired fibrinolysis and dyslipidaemia. The dyslipidaemia is characterized by high triglycerides and low HDL. All of these separate conditions are associated with an increased risk of CAD. The metabolic syndrome is complex, and the pathophysiology is incompletely understood. Hyperinsulinaemia is itself a separate risk factor for CAD and mortality,76 but the cardiovascular impact of hyperinsulinaemia is due mainly to the clustering of several potent risk factors in individual patients. The beneficial effects of testosterone on blood pressure, cholesterol, fibrinolysis and obesity have been discussed. Given these earlier reports, testosterone would be expected to have positive effects on the metabolic syndrome. Observational and animal studies have consistently found an inverse correlation between endogenous testosterone and fasting glucose or fasting insulin.20,77,78 Both hyperinsulinaemia and low testosterone separately predict the onset of type 2 diabetes.79 Patients with overt diabetes appear to have lower testosterone levels than normal males.80 A population study of 110 men with type 2 diabetes found that the level of testosterone was inversely related to glycaemic control, and that twice as many men with diabetes had a testosterone in the hypogonadal range than matched controls.81 Intervention studies with testosterone treatment to within the physiological range document improved insulin sensitivity in elderly and obese men.82 These reports concur with animal models.77 Although the risk of the metabolic syndrome is mainly related to the accumulation of cardiovascular risk factors, it seems clear that it is the hyperinsulinaemia that drives the metabolic syndrome. Testosterone inhibits this hyperinsulinaemia, and a general improvement in all associated risk factors is therefore expected; this may include delaying the onset of diabetes mellitus and even improving diabetic control. In men, endogenous levels or testosterone are inversely related to the severity of aortic atheroma and to the progression of aortic atheroma when assessed radiologically.83 Testosterone may have beneficial effects on a number of the risk factors for atherosclerotic heart disease, and these have already been discussed. In addition testosterone may have specific properties that inhibit the progression of atherosclerosis, separate from its effects on acknowledged risk factors. These observations originate from animal models. Castrated male animals rapidly develop atheroma, which is abrogated by testosterone replacement.84 The mechanisms are the subject of extensive investigation, but the key to the whole process may be inflammation. Coronary plaques responsible for angina and the acute coronary syndrome are areas of chronic inflammation. Inflammatory cells are present,85 levels of inflammatory cytokines are high,86 and nuclear mechanisms controlling inflammatory messengers are up-regulated.87 Myocardial infarction is caused when the internal surface of a plaque erodes or ruptures. The plaques most likely to rupture and thus cause AMI are those that are most inflamed.88 The inflammatory process weakens the plaque structure by stimulating matrix metalloproteinase enzymes89 and by triggering apoptosis in vascular smooth muscle cells.90 There are numerous biochemical markers of inflammation, the most recognized and best known is C-reactive protein. A high sensitive assay of CRP is the single best predictor of atherosclerotic disease in otherwise healthy people.91 A high CRP after a coronary event predicts death.92 This reinforces the importance of inflammation, and has led to commentators suggesting the use of hs-CRP to guide stepwise therapy in CAD.93 Testosterone has innate immune modulating properties. There is in vitro and animal evidence that testosterone reduces inflammation and reduces inflammatory Men are less likely to from such as or systemic and hypogonadal or men suffering with chronic inflammatory have been shown to improve following testosterone males with coronary disease have relative androgen deficiency, it is that testosterone therapy would reduce plaque inflammation and therefore limit the progression of The of secondary prevention in patients with overt vascular disease is to reduce the incidence of subsequent events or thus the and of There are well-established therapies for coronary disease that the known cardiovascular risk factors. A evidence these have the influence that endogenous testosterone may have on these risk factors, and we suggest that testosterone treatment in men has potentially beneficial effects on virtually all of the coronary risk factors, as as an independent In addition, testosterone therapy the ischaemic of and in patients with coronary disease. This not be since it is that a secondary prevention treatment a and this may improve However, testosterone replacement therapy can be widely used be is the testosterone reduce the incidence of myocardial infarction, and are the adverse such as the risk of and The evidence endogenous serum androgen levels with risk of is not and are there is no clear evidence that testosterone replacement therapy increases the incidence of although this is on a relatively small number of many would androgen therapy as a risk factor for since tissue is highly Testosterone replacement causes an increase in and in specific levels. Testosterone therapy within the normal range causes increased of cells in and may cause a of in patients. or induced hypogonadism is a of therapy for patients with hypogonadal men may have that only clinically or after testosterone therapy is In these patients, the may be earlier than would be normal because of testosterone therapy in men levels in all patients at baseline and at subsequent and some Thus disease may be earlier in patients on testosterone but it is not known whether these patients have or outcomes with to their Thus testosterone therapy in men is likely to to an increased of but this may be a time and outcomes in terms of mortality may be little The to and of male hormone replacement in men with coronary disease can only be by an