Is Testosterone Replacement Safe for the Heart? The TRAVERSE Trial Answer.

For roughly a decade, the cardiovascular safety of testosterone replacement therapy occupied an uncomfortable middle ground in cardiology: enough signal from observational studies to generate genuine concern, not enough rigorous trial data to settle it. The FDA intervened in 2015. Prescribing became cautious. The conversation stayed messy. In 2023, the TRAVERSE trial produced the randomized evidence that had been missing, answered the primary question clearly, and raised three secondary questions that the wellness industry still has not fully absorbed.

The Mechanism

Testosterone does not act on the cardiovascular system in a single direction. Understanding the full picture requires separating the effects of physiologic replacement from those of supraphysiologic dosing, because the biology differs in ways that matter clinically.

Endothelial and vascular effects. Testosterone stimulates endothelial production of nitric oxide via the enzyme eNOS. Nitric oxide is the primary mediator of vascular relaxation. In hypogonadal men, low testosterone is associated with endothelial dysfunction; testosterone replacement tends to restore NO-mediated vasodilation and improve flow-mediated dilation, the standard non-invasive measure of endothelial health. This effect is physiologic and broadly favorable. At supraphysiologic concentrations, however, testosterone can upregulate endothelin-1, a potent vasoconstrictor, partially opposing the NO effect and raising peripheral resistance. The clinical implication is that the dose range matters: replacement doses behave differently from the supraphysiologic doses used in performance enhancement.

Androgen receptors in myocardium. The cardiac muscle expresses androgen receptors throughout. Testosterone at physiologic levels supports cardiac contractile function and has anti-apoptotic effects on cardiomyocytes in experimental models. At supraphysiologic doses, the story reverses: pathologic left ventricular hypertrophy, impaired diastolic function, and fibrotic remodeling have been documented in anabolic steroid users. This distinction, physiologic replacement versus pharmacologic excess, is the most important conceptual boundary in the testosterone-and-heart literature.

Erythropoiesis and hematocrit. Testosterone increases red blood cell production through stimulation of erythropoietin and direct effects on hematopoietic precursor cells in bone marrow. This is a dose-dependent, reliable effect. Hematocrit typically rises within 3 to 6 months of initiating therapy. In most men the rise is modest. In older men, men with sleep apnea, and men on injectable formulations at higher doses, the rise can be substantial. Above a hematocrit of 54 percent, blood viscosity increases enough to meaningfully elevate thrombotic risk, particularly in the presence of atherosclerosis or irregular cardiac rhythm.

Lipid effects. The lipid impact of testosterone depends on the delivery method and dose. Injectable testosterone tends to reduce HDL more than topical gel formulations. LDL changes are variable. Across the TRAVERSE trial, ApoB and lipid differences between groups were modest but directionally unfavorable. For men starting therapy, a baseline lipid panel including ApoB and a follow-up measurement at 3 to 6 months represents reasonable practice.

Atrial remodeling. The mechanism by which testosterone increases atrial fibrillation risk is not fully established, but two pathways are biologically plausible. First, testosterone promotes cardiac hypertrophy through androgen receptor activation in atrial myocytes, and hypertrophic atrial tissue is the structural substrate for AF. Second, testosterone modulates autonomic nervous system tone in ways that increase sympathetic activity, which increases ectopic atrial firing. The AF signal in TRAVERSE is a pre-specified finding, not a post-hoc statistical artifact, and the mechanistic biology supports it as real.

Platelet and coagulation effects. Testosterone has platelet-activating properties and can modestly increase platelet aggregation at high concentrations. The increase in venous thromboembolic events observed in testosterone trials, including the pulmonary embolism signal in TRAVERSE, likely reflects a combination of hematocrit elevation increasing blood viscosity and direct procoagulant effects on the clotting cascade.

What the Evidence Shows

The story of testosterone and cardiovascular risk follows the typical arc of medical evidence: small signal studies generating alarm, larger observational data adding noise, and eventually a well-designed randomized trial providing clarity at the primary question while opening secondary ones.

The pre-TRAVERSE alarm signals.

In 2010, Basaria and colleagues published the Testosterone in Older Men with Mobility Limitations (TOM) trial in the New England Journal of Medicine. This was a randomized placebo-controlled trial of testosterone gel in 209 men aged 65 and older with mobility limitations and low testosterone. The trial was stopped early because of a higher rate of cardiovascular adverse events in the testosterone group: 23 events in the testosterone arm versus 5 in the placebo arm. This result alarmed the field. Its limitations, however, were important: the population was unusually frail, doses were higher than current clinical practice, and the trial was small enough that the event imbalance may have been partly chance. Still, the TOM trial gave regulators and clinicians a legitimate reason to pause.

In 2013, Vigen and colleagues published a retrospective cohort study in JAMA examining testosterone use in men who had undergone coronary angiography. The study found that testosterone therapy was associated with increased rates of heart attack, stroke, and death compared to no treatment. The paper generated significant media coverage. Its methodological problems were subsequently documented: the analysis included female patients in the testosterone arm due to a coding error, and the cohort selection introduced survivorship bias. These flaws did not eliminate the signal entirely but substantially reduced confidence in the magnitude of the risk estimate.

In 2014, Finkle and colleagues published a large registry study in PLOS ONE examining myocardial infarction rates in men who had filled testosterone prescriptions compared to matched controls and compared to men who had received sildenafil or tadalafil. In the 90 days following an initial testosterone prescription, the rate of nonfatal MI was 1.36 times higher than in the pre-prescription period, with the excess concentrated in men over 65 and men with pre-existing coronary artery disease. The sildenafil/tadalafil comparison group did not show the same elevation, providing a partial control for health-seeking behavior. This study added to the regulatory concern.

The FDA response.

In March 2015, the FDA issued a safety communication requiring changes to testosterone product labeling. The communication stated that the products are approved only for hypogonadism due to a disorder of the testes, pituitary gland, or brain, specifically not for low testosterone due to aging. It also required labeling about potential cardiovascular risks. This communication reshaped prescribing behavior across primary care and urology and increased the regulatory burden on testosterone prescribing.

The TRAVERSE trial.

In 2023, Lincoff and colleagues published the Testosterone Replacement Therapy for Assessment of Long-term Vascular Events and Efficacy Response in Hypogonadal Men trial in the New England Journal of Medicine. This was the cardiovascular safety trial the field needed.

Trial design: 5,246 men aged 45 to 80 with hypogonadism (total testosterone below 300 ng/dL on two morning measurements) and either established cardiovascular disease or elevated cardiovascular risk. Randomized double-blind placebo-controlled, with participants receiving testosterone gel or matching placebo. Median follow-up was approximately 22 months. The primary endpoint was a composite of cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke, assessed under a non-inferiority design.

Primary result: the incidence of major adverse cardiovascular events was 7.0 percent in the testosterone group versus 7.3 percent in the placebo group, hazard ratio 0.96 (95% confidence interval 0.78 to 1.17), non-inferiority p<0.001. 5 / Solid (Lincoff et al. 2023, NEJM) Testosterone replacement at therapeutic doses in properly selected hypogonadal men did not increase the risk of heart attack, stroke, or cardiovascular death. The primary question was resolved.

The secondary findings were not resolved in the same direction.

Atrial fibrillation occurred in 3.5 percent of the testosterone group versus 2.4 percent of the placebo group, a statistically significant increase. 4 / Promising This was a pre-specified secondary endpoint, not a data-mining result.

Pulmonary embolism occurred in 0.9 percent of the testosterone group versus 0.5 percent of the placebo group. 4 / Promising The venous thromboembolic signal is consistent with the known erythropoietic and procoagulant effects of testosterone.

Acute kidney injury occurred in 2.3 percent of the testosterone group versus 1.4 percent of the placebo group. The mechanism is likely related to reduced renal blood flow from elevated hematocrit and viscosity rather than direct nephrotoxicity.

Hematocrit elevation above 54 percent, the threshold where viscosity begins to meaningfully increase thrombotic risk, was significantly more common in the testosterone group. Dose adjustment and therapeutic phlebotomy are the management tools; monitoring is the prerequisite.

The T Trials context. The Testosterone Trials (T Trials), a coordinated set of seven placebo-controlled trials reported primarily by Snyder and colleagues between 2016 and 2018, evaluated testosterone effects on specific symptoms in older hypogonadal men. The sexual function trial showed improvement in sexual desire and erectile function. The physical function trial showed modest improvements in walking distance. The bone trial showed improved bone mineral density. The T Trials were not powered for cardiovascular outcomes but established the symptomatic benefit profile against which cardiovascular risk must be weighed in clinical decision-making.

Testosterone Therapy and Obstructive Sleep Apnea: An Underappreciated Cardiovascular Loop

Obstructive sleep apnea is common in men with hypogonadism and becomes more clinically significant after testosterone replacement therapy begins. This creates a cardiovascular feedback loop that the TRAVERSE secondary findings do not fully capture on their own.

The relationship between testosterone and obstructive sleep apnea is bidirectional. Low testosterone is associated with increased adiposity, particularly visceral fat accumulation, which independently increases upper airway collapsibility. Men with hypogonadism have higher rates of obstructive sleep apnea than age-matched eugonadal men. This might suggest that correcting the hormonal deficiency would improve sleep-disordered breathing. The biology does not support this expectation.

Two mechanisms explain how testosterone worsens sleep apnea. First, testosterone reduces the ventilatory response to rising CO2 during sleep, blunting the respiratory drive that normally triggers arousal when airway obstruction occurs. A sleeping airway that becomes obstructed in a man on testosterone therapy is slower to trigger the corrective arousal response. Second, testosterone has direct effects on upper airway muscle tone that may reduce the neuromuscular defense against collapse of the pharyngeal airway during sleep. Matsumoto and colleagues documented that testosterone administration worsened apnea-hypopnea index in hypogonadal men, with effects that were dose-dependent and appeared within the first weeks of therapy. The FDA testosterone prescribing label includes worsening of obstructive sleep apnea as a labeled risk, reflecting the consistency of this finding across clinical populations. 4 / Promising

The cardiovascular significance of worsening sleep apnea during testosterone therapy is not abstract. Untreated obstructive sleep apnea generates repeated nocturnal hypoxic episodes that activate the sympathetic nervous system acutely, raise circulating catecholamines, produce blood pressure surges during apneic events, and over years produce structural vascular injury and progressive endothelial dysfunction. Yaggi and colleagues, reporting a prospective cohort study in the New England Journal of Medicine in 2005, followed 1,022 patients with and without obstructive sleep apnea and found a hazard ratio of 1.97 for the composite of stroke or death from any cause in the OSA group after adjustment for age, sex, BMI, hypertension, diabetes, and smoking. The dose-response holds: more severe apnea correlates with higher cardiovascular event rates.

For a man starting testosterone therapy who has undiagnosed sleep apnea, the downstream cardiovascular trajectory connects directly to the AF and blood pressure signals TRAVERSE already documented: TRT worsens apnea severity, which increases sympathetic tone and blood pressure variability overnight, which accelerates atrial remodeling, which feeds the AF risk the trial quantified. These pathways are additive, not independent.

The clinical implication is that pre-treatment sleep apnea screening belongs in every pre-testosterone evaluation. A validated questionnaire such as STOP-BANG (Snoring, Tiredness, Observed apnea, blood Pressure, BMI, Age, Neck circumference, Gender) takes under two minutes and identifies men at high risk. A positive screen warrants either a home sleep apnea study before testosterone initiation or a specialist referral. Men with confirmed moderate-to-severe OSA who are not on continuous positive airway pressure therapy should address the apnea before testosterone decisions are finalized.

What to Do This Week

1. Before any testosterone conversation begins, request a morning fasting total testosterone and free testosterone on at least two separate days, along with LH and FSH. These measurements confirm whether true hypogonadism is present and whether the cause is primary (testicular) or secondary (pituitary or hypothalamic). Starting testosterone therapy without this documentation is not evidence-based medicine.

2. Ask your physician specifically about the TRAVERSE atrial fibrillation and pulmonary embolism findings and what they mean for your individual risk. If you have a history of paroxysmal AF, prior blood clot, or untreated sleep apnea, those risk factors require specific discussion before any decision is made.

3. Request baseline labs that a pre-treatment cardiac evaluation should include: complete blood count with hematocrit, ApoB, full lipid panel, basic metabolic panel for kidney function, and blood pressure measured at home over at least seven days on a validated device. A single clinic blood pressure reading is not sufficient.

4. If your physician recommends testosterone therapy, confirm that a monitoring schedule is in place. Hematocrit should be checked at 3 months and 6 months after initiation, then annually. PSA should be checked at baseline and 3 to 6 months. Lipids should be rechecked at 3 to 6 months. A monitoring protocol is not optional; it is what separates appropriate clinical management from dispensing a prescription without follow-through.

5. If you are currently on testosterone therapy prescribed through a telehealth platform and have never had a cardiac evaluation, schedule one now and bring your current dose, delivery method, and any labs you have. The TRAVERSE secondary findings are real and the monitoring they imply belongs in the conversation with a physician who knows your cardiovascular history.

The TRAVERSE trial settled the most urgent question: testosterone replacement at therapeutic doses in documented hypogonadal men does not increase heart attack, stroke, or cardiovascular death risk compared to placebo. That is a meaningful answer that should lift some of the unwarranted fear from a therapy that genuinely helps men whose quality of life has been eroded by inadequate testosterone. What the same trial also documented, in the same data, is that the atrial fibrillation rate was higher, the pulmonary embolism rate was higher, and the acute kidney injury rate was higher in the testosterone group. Those findings are not reasons to avoid a legitimate therapy; they are reasons to have an informed pre-treatment conversation rather than skipping it.

The Signal Check is fifteen questions mapping the male cardiovascular risk pattern, including the physiological domains most commonly missed in standard screenings. It produces a specific starting point for your next clinical conversation.