Adverse Symptoms and Physiologic Consequences

Several cross-sectional and longitudinal studies have demonstrated the sexual, psychological, and physical consequences of untreated testosterone deficiency. Low testosterone levels are associated with reduced sexual interest, sexual activity, and libido ; loss of bone mineral density and increased fracture risk ; decreased muscle mass and strength ; increased fat mass ; fatigue and depressed mood ; and a decline in cognitive function.

Increase in Metabolic Syndrome and Other Cardiovascular Risk Factors

Low testosterone levels are also associated with increased risk of the individual components of the metabolic syndrome, including insulin resistance and subsequent type II diabetes mellitus, central adiposity, increased triglycerides, total cholesterol and low-density lipoprotein levels, and decreased high-density lipoprotein levels. Furthermore, several cross-sectional and longitudinal studies have shown that testosterone deficiency is associated with an increased risk of the metabolic syndrome. The Massachusetts Male Aging Study prospectively followed 950 men aged 40 to 70 years for 15 years and found a 2- to 4-fold increase in the risk of metabolic syndrome in men with a body mass index less than 25 who were in the lowest quartile of total testosterone (≤415 ng/dL) compared with those in the highest quartile (>628 ng/dL). Testosterone deficiency has also been shown to be associated with several cardiovascular risk factors, including increased carotid intimal media thickness, peripheral arterial disease, and elevated high-sensitivity C-reactive protein.

Increased All-Cause and Cardiovascular Mortality

Testosterone deficiency is associated with an increase in both all-cause mortality and cardiovascular mortality. A large prospective cohort study of 3014 Swedish men aged 69 to 80 years followed for 4.5 years found that men in the lowest quartile of total testosterone (≤97 ng/dL) had an increased mortality (hazard ratio [HR] 1.65) compared with those in the other 3 quartiles. Another prospective cohort study of 794 men aged 50 to 91 years (mean age 71 years) followed for 11.8 years also showed not only increased mortality (HR 1.40) but also an increased risk of both cardiovascular (HR 1.38) and respiratory disease (HR 2.29) among men with the lowest quartile of total testosterone (<241 ng/dL). A prospective population-based study of almost 2000 men aged 20 to 79 years in Pomerania also found an association between low testosterone levels (<250 ng/dL) and both all-cause mortality (HR 2.32) and cardiovascular mortality (HR 2.56). Several subsequent studies and meta-analyses have similarly shown an association between testosterone deficiency and both all-cause mortality and cardiovascular mortality, including in a subpopulation of men with known coronary disease.

Although some prospective trials of younger men (mean age early 50s) failed to show an association between testosterone deficiency and mortality, they instead showed an increased risk of ischemic heart disease. Other meta-analyses have shown only an increase in cardiovascular risk in elderly men with low testosterone levels, postulating that testosterone deficiency may simply be a marker of poor underlying health status. Although the evidence is conflicting, several studies have found increased cardiovascular morbidity and mortality in men being treated for prostate cancer with androgen deprivation therapy (ADT). Although it is unknown whether this increased risk is limited to those with preexisting coronary artery disease, there are substantial data demonstrating the adverse effects of ADT on traditional cardiovascular risk factors, such as serum lipoproteins, insulin sensitivity, and obesity.

Improvement in Symptoms and Components of the Metabolic Syndrome

Several studies have shown that testosterone replacement therapy in men with clinical androgen deficiency improves sexual function and libido, increases lean body mass and physical function, increases bone mineral density, decreases depression scores, and improves mood and quality of life. Although mostly small studies, several randomized controlled trials of testosterone therapy in men with testosterone deficiency and type II diabetes mellitus or metabolic syndrome have shown improvement in various components of the metabolic syndrome, including atherogenic dyslipidemia, insulin resistance, glycemic control, and inflammatory markers, including c-reactive protein levels. A meta-analysis of 20 cross-sectional, longitudinal, and randomized studies further showed that testosterone replacement therapy is associated with both improved metabolic control and central obesity. Whether the improvement in metabolic syndrome and its individual components is a direct effect of testosterone therapy or the result of the concurrent reduction in abdominal obesity seen in many of these studies is unknown.

Improvement in Cardiac Functional Capacity

Randomized controlled trials of testosterone therapy have shown improvement in functional capacity in both men with coronary disease as well as those with chronic heart failure. A double-blind, randomized, placebo-controlled study of 46 men with stable angina treated with low-dose transdermal testosterone versus placebo showed a delay in time to 1-mm ST-segment depression on treadmill exercise testing in the testosterone arm consistent with reduced exercise-induced myocardial ischemia. Those with lower baseline testosterone levels had a greater magnitude of this response. In those receiving testosterone therapy, testosterone levels increased from 390 ng/dL at baseline to 644 ng/dL at 6 weeks and 535 ng/dL at 14 weeks. Another randomized placebo-controlled study showed similar improvement in myocardial ischemia with testosterone replacement therapy and a sustained effect at 12 months.

Similarly, randomized placebo-controlled trials of testosterone therapy in men with heart failure have demonstrated that low testosterone levels correlate with reduced functional capacity as assessed by both the incremental shuttle walk test (ISWT) and peak oxygen consumption (V o ₂ ). Although guideline recommendations advise against treating men with uncontrolled or poorly controlled heart failure with testosterone therapy, these randomized controlled trials were conducted in men with stable heart failure and a moderately reduced ejection fraction and showed that testosterone therapy improves objective functional capacity based on the ISWT and peak V o ₂ as well as symptomatic functional class without an increase in cardiovascular events, hospital admissions, or mortality.

Progression of Atherosclerotic Risk Factors

A recent randomized placebo-controlled trial of 308 men aged 60 years or older (mean age 67.6 years) with low to low-normal testosterone levels showed no change in the surrogate end points of carotid artery intima-media thickness and coronary artery calcium with the use of testosterone gel versus placebo gel over 3 years. Furthermore, there was no difference in secondary outcomes of sexual function or health-related quality of life. Of note, the mean baseline total testosterone level was 307 ng/dL in both groups, and approximately a third of patients had testosterone levels between 300 and 400 ng/dL. Interestingly, the lower limit of normal reference range for the specific assay used to measure testosterone in this study is 250 ng/dL in men aged 18 to 69 years, which is lower than the 300 to 320 ng/dL cutoff used in most assays. Furthermore, in men aged 70 to 89 years, the lower limit of normal for this assay is 90 ng/dL. This finding suggests that at least a third, if not more, of enrolled patients did not have clinical hypogonadism as defined by the Endocrine Society’s guidelines summarized earlier. The average testosterone level at which symptoms of androgen deficiency occur in most studies corresponds to the lower limit of normal range for young men, which is approximately 300 ng/dL. However, normal ranges vary depending on the assay and laboratory; clinicians are advised to use the lower limit of normal that has been established in their laboratory.

Indeed, the authors agree that some men in the trial had low-normal testosterone levels and that participants were not selected based on a diagnosis of hypogonadism or clinical symptoms of hypogonadism as is often the case in current clinical practice. This trial failed to show not only a reduction in atherosclerotic risk factors with testosterone therapy but also failed to show an improvement in sexual function or quality of life, contrary to most studies of testosterone replacement. This failure is likely explained by the fact that many participants did not have clinical androgen deficiency as defined by the Endocrine Society’s guideline and further supports the recommendation that testosterone therapy should only be administered to those men with appropriately diagnosed hypogonadism.

Signal Toward Increased Cardiovascular Events

Several recent studies suggesting that testosterone replacement therapy may be associated with an increase in cardiovascular events prompted the Endocrine Society to issue a statement regarding this potential risk and reinforcing that testosterone prescription should only be administered to men with appropriately diagnosed clinical androgen deficiency and with adequate safety monitoring.

The Testosterone in Older Men with Mobility Limitations (TOM) trial was a prospective, randomized placebo-controlled study to determine the effects of testosterone gel versus placebo gel on lower extremity strength and physical function in 209 men older than 65 years (mean age 74 years) with low total testosterone (100–350 ng/dL) or free testosterone (<50 pg/mL) levels and limitations in mobility. A higher incidence of cardiovascular signs or symptoms that were not prespecified was observed in the testosterone group, leading to early termination of the trial. It is important to consider several factors when interpreting this study. First, the cardiovascular events found in this study were not predefined and included a wide variety of signs and symptoms, including some of unclear clinical significance, such as self-reported tachycardia with fatigue, peripheral edema, left ventricular strain pattern during exercise testing, and ectopy noted on electrocardiogram. Additionally, the study was not designed to evaluate cardiovascular outcomes but rather the effects of testosterone on muscle function. Furthermore, both the small number of total adverse events and early termination of the trial may have led to an overestimation of treatment differences.

Along with the methodological issues in interpreting these results, there were also multiple confounders that could have played a role in the study results. The testosterone gel dose used in this study was a higher off-label dosage of 100 mg per day, whereas the recommended initial dosage is 50 mg daily. Testosterone levels increased from 243 ng/dL to 574 ±403 ng/dL in the testosterone group, with some subjects achieving testosterone levels well more than the recommended midnormal range. Indeed, there was an increased risk of cardiovascular events in men in the highest quartile of on-treatment testosterone level (512–1957 ng/dL) compared with all other subjects (HR 2.4, P = .05), which may partly be explained by supratherapeutic testosterone levels in these patients. Additionally, the study population was a higher risk older population with a high prevalence of hypertension, hyperlipidemia, diabetes, and obesity at baseline; there was a higher prevalence of patients on antihypertensive and statin therapy in the testosterone arm. Within the testosterone arm, those who experienced cardiovascular events had lower testosterone levels at baseline, suggesting a greater underlying disease burden.

An increased risk of cardiovascular events was also seen in a meta-analysis of 27 small, randomized placebo-controlled trials consisting of 2994 older men treated with at least 12 weeks of testosterone therapy (odds ratio [OR] 1.54, 95% confidence interval [CI] 1.09–2.18). Interestingly, this effect varied with the source of funding but not with baseline testosterone levels. The risk of cardiovascular events was lower in the 13 pharmaceutical industry funded trials compared with those that were not industry sponsored (OR 0.89 vs 2.06). Industry-funded trials had younger subjects and used different definitions of cardiovascular events as well as different reporting of adverse effects, which may have contributed to this observed difference.

Studies Showing No Effect on Cardiovascular Events

Several randomized trials and meta-analyses have shown that testosterone replacement therapy is not associated with an increase in cardiovascular events, although most are small heterogeneous studies of variable duration. A recently published and now the largest meta-analysis of 75 placebo-controlled randomized trials of testosterone therapy including 3016 testosterone-treated patients and 2448 placebo-treated found no association between testosterone therapy and cardiovascular risk and additionally showed a reduction in body fat and increase in lean muscle mass with testosterone replacement therapy.

Reduction in Mortality

A longitudinal cohort study by Muraleedharan and colleagues prospectively followed 581 men with type 2 diabetes for a mean of 41.6 months and found that the 238 men with low total testosterone levels (≤300 ng/dL) had increased mortality rates (17.2% vs 9.0%, HR 2.02, 95% CI 1.2–3.4) compared with the 343 who did not. Those who were treated with testosterone replacement were appropriately diagnosed with clinical androgen deficiency and achieved guideline-recommended on-treatment testosterone levels (mean peak on-treatment level 657 ng/dL). Of the 238 men with low total testosterone levels, the 64 who were treated had reduced mortality (8.4% vs 19.2%, adjusted HR 2.3, 95% CI 1.3–3.9), suggesting that testosterone replacement therapy in these men improves survival. Although this was not a randomized trial, the major strength of this study is that all patients were appropriately diagnosed, treated, and monitored per guideline recommendations.

Potential Increase in Cardiovascular Events and Mortality

There have been a few recent retrospective studies suggesting increased cardiovascular events and mortality with testosterone therapy. A retrospective cohort study of 8709 men who had undergone prior coronary angiography in the Veterans Affairs system and also had low testosterone levels (<300 ng/dL) found that the 1223 men who were prescribed testosterone therapy had increased risk of the composite end point of all-cause mortality, myocardial infarction, and ischemic stroke compared with the 7486 who were not (25.7% vs 19.9%, HR 1.29, 95% CI 1.04–1.58), and this was irrespective of the presence of prior coronary artery disease or revascularization. Although compelling, further inspection reveals that this trial has several limitations, resulting in significant controversy. There is an inherent selection bias in this study given that only 1223 men were prescribed testosterone therapy, despite the fact that all of the men in this cohort had low testosterone levels. The inclusion criteria of this study did not require a clinical diagnosis of hypogonadism; it is possible that more men in the treated group had clinical androgen deficiency, which in itself is associated with higher mortality. Baseline testosterone levels were also lower in the treated group (175.5 vs 206.5 ng/dL), suggesting a greater severity of androgen deficiency and underlying disease burden.

Furthermore, there is no evidence that testosterone was prescribed or monitored as per the guidelines as only 60% of the treated group had repeat testosterone measurements; it is possible that treated patients may have been undertreated or overtreated, both of which can be potentially harmful. Among those patients who did have a repeat measurement, the level was 332.2 ng/dL, which is lower than the recommended goal in the midnormal range, suggesting undertreatment. It is possible that the increased mortality seen in treated patients may reflect inadequate treatment in patients with more severe hypogonadism. Additionally, although the investigators reported a long-term duration of therapy, 18% of subjects had filled only one prescription.

A subsequent cohort study reported increased risk of nonfatal myocardial infarction (relative risk [RR] 1.36) among 55,593 individuals in a large health care database derived from insurance company records in the first 90 days after filling their first prescription of testosterone compared with the prior 1 year. This association was observed in men aged 65 years and older (RR 2.19) and in younger men with preexisting heart disease (RR 2.90). Again, there were no data on follow-up testosterone levels or adequacy of treatment. Given that testosterone deficiency itself is associated with increased cardiovascular mortality, the observed increase in nonfatal myocardial infarction could be due to the underlying androgen deficiency rather than testosterone therapy.

Most recently, a retrospective cohort study comparing the cardiovascular safety of testosterone injections, patches, and gels in the United States and the United Kingdom using insurance claims data in men aged 18 years or older showed that testosterone injection users had a higher rate of cardiovascular events (myocardial infarction, unstable angina, and stroke; HR 1.26, 95% CI 1.18–1.35), hospitalization (HR 1.16, 95% CI 1.13–1.19) and death (HR 1.34, 95% CI 1.15–1.56) when compared with men using testosterone gel. There was no difference in outcomes between men using testosterone gels and patches. The study did not include a comparator group of untreated men. It can be postulated that given the nonphysiologic peaks and troughs of testosterone injection, this formulation may lead to inappropriately high testosterone levels, which can potentially be harmful. However, as in other retrospective analyses, this study also did not assess the validity of testosterone prescription or whether these patients were diagnosed with clinical androgen deficiency. It also lacked data on testosterone levels both at initiation of therapy and during treatment. Furthermore, it is possible that those men prescribed testosterone injection were at a higher baseline cardiovascular risk than those prescribed testosterone gel as evidenced by the fact that the most Medicare patients, who were significantly older, were injection users, whereas most commercially insured US patients were gel users. Given these limitations, it is difficult to draw a conclusion on whether testosterone injections are truly associated with adverse events; further study is needed to evaluate this potential association.

Reduction in Mortality

A retrospective analysis of 1031 men (average age 62 years) with testosterone levels of 250 ng/dL or less from 7 Veterans Affairs medical centers found that the 398 men who were treated with testosterone replacement therapy had lower mortality than those who were untreated (10.3% vs 20.7%, adjusted HR 0.61, 95% CI 0.42–0.88). This study also evaluated the duration of treatment rather than initial prescription only, which has been a limitation of some studies. In a secondary analysis whereby men who stopped treatment were censored at 90 days after their last treatment date also found that testosterone therapy was associated with a decrease in mortality. Furthermore, a correlation was seen between higher mortality and both lower baseline testosterone levels as well as a shorter duration of testosterone treatment, supporting that testosterone deficiency is associated with increased mortality and that testosterone replacement therapy in these patients is associated with improved survival.

Another limitation of recent studies suggesting an increase in cardiovascular events with testosterone therapy is the lack of on-treatment monitoring of testosterone levels to ensure adequate treatment and avoid supraphysiologic testosterone levels. To address this concern, Sharma and colleagues performed a retrospective analysis of 83,010 men in the Veterans Affairs system (median age 66 years) with low testosterone levels and analyzed outcomes based on whether subjects received testosterone replacement prescriptions as well as whether follow-up testosterone levels were in the normal range. The 43,931 men who were treated with any form of testosterone replacement therapy and achieved normal follow-up testosterone levels (mean follow-up 6.2 years) were compared with the 13,378 untreated men (mean follow-up 4.7 years) and were found to have significantly lower rates of all-cause mortality (HR 0.44, 95% CI 0.42–0.46), myocardial infarction (HR 0.76, 95% CI 0.63–0.93) and stroke (HR 0.64, 95% CI, 0.43–0.96). When this group of 43,391 men was compared with the 25,701 men receiving testosterone therapy without achieving normal follow-up testosterone levels (mean follow-up 4.6 years), they again showed significantly lower rates of all-cause mortality, myocardial infarction, and stroke. Compared with untreated men, the group of men who were treated but did not achieve normalization of follow-up testosterone levels showed decreased all-cause mortality (HR 0.84, 95% CI 0.80–0.89) but no difference in rates of myocardial infarction or stroke.

This study includes the largest cohort of patients with hypogonadism and demonstrates the critical role of follow-up testosterone levels on determining the benefits and adverse effects of testosterone therapy. The investigators showed that patients treated with testosterone therapy who achieve normalized testosterone levels indeed have lower rates of mortality as well as myocardial infarction. This study is one of the few studies to analyze outcomes with testosterone replacement therapy based on adequate treatment rather than treatment alone without data on follow-up testosterone levels, which has been a major limitation of previous studies. However, this study does have its own limitations, namely, the likelihood of unmeasured confounders among the 3 groups. Because of the nature of this retrospective analysis, there is no information on why some patients were treated with testosterone replacement, whereas others were not and, among those who were treated, why some did not achieve normalization of levels. Although it is reasonable to label this group as inadequately treated, it can be postulated that the differences observed in this group may be due to differences in baseline characteristics. For example, this group may have been sicker with lower testosterone levels at baseline and more severe hypogonadism; they may have been less motivated and/or noncompliant with testosterone therapy; or they may have received less medical follow-up, all of which can contribute to poorer outcomes. Similarly, it is also unknown why the untreated group was not treated with testosterone therapy and again may have had unmeasured differences in baseline characteristics, such as more severe hypogonadism, greater number of comorbidities precluding treatment with testosterone therapy, and so on, which may have contributed to the reported outcomes. Although the investigators used propensity score-weighted Cox proportional hazard models, the 3 groups were significantly different at baseline and the presence of unmeasured confounders is likely.