The TRAVERSE trial

TRAVERSE is the large randomised trial that settled whether testosterone treatment causes heart attacks and strokes in men with genuine deficiency: it does not. But the same trial found more atrial fibrillation, kidney injury, lung clots and fractures on testosterone, no protection against diabetes, no gain in memory or thinking, and no survival benefit.

Read TRAVERSE as a safety result for men with confirmed deficiency, not as a licence to treat normal ageing. If you take testosterone, the trial's side-effect signals — heart rhythm, blood pressure, kidneys, clots and falls — are what to watch.

What the trial tested

TRAVERSE was a double-blind randomised trial that the US Food and Drug Administration (FDA) required after years of worry that testosterone might raise heart risk. It enrolled 5,204 men aged 45–80 with documented low testosterone (two morning levels under 300 ng/dL, about 10.4 nmol/L, plus symptoms) who also had heart disease or a high risk of it. They were randomly assigned to a daily testosterone gel or a placebo gel, treated for an average of 21.7 months and followed for 33 months.^[1]

The trial was built to answer one question: is testosterone no worse than placebo for heart attack, stroke and death from cardiovascular causes? It was designed to rule out a 50% increase in that risk.

The heart result: no difference

Heart attack, stroke or cardiovascular death struck 182 men on testosterone (7.0%) and 190 on placebo (7.3%). The hazard ratio was 0.96, with a 95% confidence interval of 0.78 to 1.17 — the range where the true effect most plausibly lies — comfortably inside the pre-set safety margin of 1.5.^[2] The trial found no mortality benefit either.

On the strength of this result, the FDA in February 2025 ordered the cardiovascular-risk warning off every testosterone label. It kept the statement that safety and effectiveness in "age-related hypogonadism" have not been established, and in the same action extended a blood-pressure warning to every product: systolic pressure rose by about 2 mmHg on gel and about 2–5 mmHg on the oral capsules in required monitoring studies, and in TRAVERSE the gap against placebo was 1.5 mmHg at 36 months.^[3]

Prostate: reassuring, within limits

There was no excess of prostate cancer — 5 high-grade cancers among 2,596 treated men against 3 among 2,602 on placebo — and no worsening of urinary symptoms. Men entering with a prostate-specific antigen (PSA) above 3.0 ng/mL, already-severe urinary symptoms or a prostate nodule were screened out before randomisation, so the reassurance does not extend to them.^[4]

The side-effect signals

On testosterone compared with placebo:

Atrial fibrillation: 3.5% against 2.4%, about one extra case per hundred men. The trial was not designed to detect it and did no routine heart tracings, so atrial fibrillation was only captured when reported as a medical event — a safety signal rather than proof.^[5]
Acute kidney injury: 2.3% against 1.5%. Lung clots (pulmonary embolism): 0.9% against 0.5%. All the men had heart disease or a high risk of it, so read these as an upper bound rather than a figure for the general population.^[6]
Fractures: in a pre-specified fracture substudy, 3.5% against 2.5% over a median 3.2 years — about 43% more (hazard ratio 1.43, 95% CI 1.04–1.97), or roughly one extra fracture per hundred men. Most followed falls rather than occurring at the classic osteoporotic sites, and the trial did not measure bone density.^[7] In a smaller, earlier trial of older men, testosterone raised inner (trabecular) spine density on CT scans by about 7.5% over a year while standard hip scans barely moved; that trial was too small to look at fractures.^[8] A density gain is not the same as fewer fractures. See Bone density.
Blood pressure: the 1.5 mmHg systolic gap above.

Diabetes: no protection

In the trial's pre-specified substudy of men with prediabetes, testosterone did not significantly slow progression to diabetes; the rates tracked closely across four years (for example 10.1% against 14.6% at two years; overall not significant). It did not improve blood sugar or HbA1c, the three-month blood-sugar average, and did not reverse existing diabetes. The authors concluded that testosterone alone should not be used to prevent or treat diabetes.^[9]

One caveat cuts the other way. Doses were adjusted toward 350–750 ng/dL (12.1–26.0 nmol/L), but blood was drawn 24 hours after the gel went on, so recorded levels were daily lows. Treated men rose from 220 ng/dL (7.6 nmol/L) to an average of 409 ng/dL (14.2 nmol/L), the lower half of the target, so the trial cannot fully rule out a blood-sugar effect at higher levels.^[10] The "testosterone prevents diabetes" claim traces to a different trial, in men without true deficiency, discussed in Testosterone therapy.

Mind and mood

TRAVERSE found no effect on cognition, though it did find modest improvements in mood and energy.^[11]

What the trial does not tell you

It is not about healthy men. Every participant had confirmed deficiency and heart disease or high heart risk. It says nothing in favour of raising testosterone in men whose levels are normal; genetic evidence on lifelong higher testosterone points toward a small rise in coronary risk through blood pressure (see Testosterone therapy).
It is not a longevity result. No survival benefit was found.
It tested one form at modest levels. A daily gel, with levels in the lower half of the target range; injections, which cause sharper peaks, were not tested.

What to do

If you have confirmed deficiency, fear of heart attack or stroke alone is no longer a reason to refuse treatment.
Weigh the signals before starting if you have atrial fibrillation, high blood pressure, kidney disease, a past clot, or a high risk of falls and fractures.
Don't take testosterone to prevent diabetes, sharpen memory or live longer — the trial looked for each and found no benefit.
If your PSA is above 3.0 ng/mL, or you have severe urinary symptoms or a prostate nodule, the prostate reassurance does not cover you.