Testosterone therapy and the heart: what the evidence actually shows

The alarm was not invented. Between 2010 and 2014 a small trial in frail older men was halted early for cardiovascular events, and two widely covered observational analyses reported more cardiac events in men prescribed testosterone. In 2015 the FDA added cardiovascular risk language to the class labelling and required manufacturers to run a properly powered outcomes trial. What followed is a case study in how long it takes to answer a safety question honestly: the trial the regulator asked for reported eight years later.

Observational data could never settle it. Men prescribed testosterone differ systematically from men who are not — in obesity, diabetes, sleep apnoea and how often they see a doctor — and every one of those differences pushes cardiac risk in the same direction. Only randomisation removes that.

TRAVERSE randomised 5,246 men aged 45 to 80 with two morning testosterone levels below 300 ng/dL, symptoms of hypogonadism, and either established cardiovascular disease or a high risk of it.^[8] Mean age was 63, mean BMI was about 35 kg/m², roughly 70% had type 2 diabetes and about 90% had hypertension and dyslipidaemia. This was not a healthy-volunteer study; it was deliberately stacked toward men who could be harmed.

Participants received daily 1.62% transdermal testosterone gel or placebo, titrated to a target of 350-750 ng/dL, with a mean treatment duration of 21.7 months and mean follow-up of 33 months. The primary endpoint — cardiovascular death, non-fatal myocardial infarction or non-fatal stroke — occurred in 182 men on testosterone (7.0%) and 190 on placebo (7.3%), hazard ratio 0.96 (95% CI 0.78-1.17), meeting the pre-specified non-inferiority margin.^[1]

Read that result precisely. It is evidence of no meaningful increase in heart attack, stroke or cardiovascular death. It is not evidence of benefit: nothing in the trial suggests testosterone protects the heart.

Three secondary events were more frequent on testosterone: atrial fibrillation (3.5% vs 2.4%), acute kidney injury (2.3% vs 1.5%) and pulmonary embolism (0.9% vs 0.5%).^[1] These were secondary endpoints in a trial designed around a different question, so they are hypothesis-generating rather than definitive — but they have not gone away on repeat analysis.

A 2025 meta-analysis of 23 randomised trials and 9,280 men with testosterone deficiency (mean age 64.6) found no increase in all-cause mortality, cardiovascular mortality, myocardial infarction or stroke, and one clear outlier: cardiac arrhythmia, relative risk 1.53 (95% CI 1.20-1.97).^[3] Two independent bodies of evidence pointing at the same rhythm signal is worth more than either alone.

Largely, yes. An individual participant data meta-analysis published in 2022 pooled 35 randomised trials and 5,601 men, with individual data from 17 trials and 3,431 participants. Cardiovascular events occurred in 7.5% on testosterone versus 7.2% on placebo (odds ratio 1.07, 95% CI 0.81-1.42) and deaths in 0.4% versus 0.8% (odds ratio 0.46, 95% CI 0.17-1.24).^[4] The authors were careful about what they had shown: no evidence of short- to medium-term harm, with long-term data still needed. Average treatment exposure across those trials was under a year.

Low testosterone does travel with worse cardiovascular outcomes. An individual participant data meta-analysis of 11 prospective cohorts and 24,109 community-dwelling men, using mass-spectrometry assays, found higher all-cause mortality below roughly 7.4 nmol/L (213 ng/dL) and higher cardiovascular mortality below roughly 5.3 nmol/L (153 ng/dL), after adjustment for age, BMI, lifestyle and cardiovascular risk factors.^[5]

That is an association. Low testosterone is also a faithful marker of obesity, poor metabolic health and chronic illness — and the randomised evidence is the reason to be careful here: correcting the number did not lower cardiac event rates in TRAVERSE.^[1] If low testosterone were driving those events, treating it should have moved them. It did not. Peak Human's overview of cardiovascular risk reduction covers the levers that do have outcome data.

One 2017 trial deserves a direct answer because it circulates widely. In a substudy of the Testosterone Trials, 138 older men had coronary CT angiography before and after 12 months of testosterone gel. Non-calcified plaque volume increased 41 mm³ more in the testosterone group than placebo (P=0.003) and total plaque volume 47 mm³ more (P=0.006), with no difference in coronary artery calcium score.^[6]

It is a real finding and it should not be waved away, but its weight is limited: 138 men, one year, an imaging surrogate rather than clinical events, and a trial explicitly not powered for outcomes. Where a surrogate and a 5,246-man event trial disagree, the event trial wins — while the surrogate remains a reason to keep watching.

On 28 February 2025 the FDA made class-wide labelling changes to all testosterone products. It removed the boxed-warning language about increased risk of adverse cardiovascular outcomes, citing TRAVERSE. In the same action it required a new warning about increased blood pressure, after ambulatory blood pressure monitoring studies confirmed a class-wide rise, and it kept the limitation of use stating these products are not indicated for low testosterone due to ageing alone.^[7]

Both halves matter. The regulator did not conclude testosterone is cardiovascularly neutral; it concluded the old warning overstated one risk while a different one was under-labelled. Blood pressure effects differ by formulation — the oral testosterone undecanoate trade-off is covered in our review of oral testosterone therapy.

The definitive trial answered the question it was built for: in men with genuine hypogonadism, including men with existing heart disease, testosterone therapy did not increase heart attack, stroke or cardiovascular death. It also surfaced a rhythm and clotting signal that survived independent replication, and a blood pressure effect the regulator judged real enough to label.^[1]

That is a narrower claim than either camp usually makes. It supports treating men who are actually hypogonadal, with monitoring that takes atrial fibrillation, haematocrit and blood pressure seriously. It does not support testosterone as a longevity or cardiovascular intervention in men whose levels are normal — a distinction worth keeping when reading marketing that cites TRAVERSE.^[2]

This article is educational and not medical advice. Discuss testosterone therapy with your own physician, particularly if you have heart disease, atrial fibrillation, a clotting history, sleep apnoea or high blood pressure.