The TRAVERSE Trial: What It Showed and What It Did Not

Testosterone prescribing has long sat between two legitimate concerns. Untreated pathological hypogonadism can cause distressing symptoms and health effects. At the same time, prescriptions expanded while large randomized evidence about cardiovascular outcomes remained limited.

The TRAVERSE trial was designed to address that safety gap, and it is the largest randomized trial of testosterone replacement with adjudicated cardiovascular outcomes, but its result is often stretched far beyond the question it asked. Reading it well requires attention to the population, treatment target, endpoint, noninferiority design, duration, and subsequent regulatory actions.

Who entered the trial#

TRAVERSE enrolled 5,246 men aged 45 to 80 at 316 sites in the United States. Participants had one or more symptoms consistent with hypogonadism and two fasting testosterone concentrations below 300 ng/dL, measured on separate mornings; they also had preexisting cardiovascular disease or increased cardiovascular risk.

These criteria matter. The trial was not a study of a single low result obtained during acute illness, a random afternoon measurement, bodybuilding doses, or testosterone use in people without symptoms. It did not enroll a general healthy-aging population.

Important exclusions also shape interpretation. Men with certain prostate findings, a history of prostate cancer, severe lower urinary tract symptoms, marked erythrocytosis, or other contraindications were not represented in the same way as the screened participants. Trial monitoring and dose adjustment were more structured than many routine settings.

What treatment was compared#

Participants were randomly assigned to daily transdermal 1.62 percent testosterone gel or matching placebo gel. Testosterone doses were adjusted to maintain concentrations in a prespecified range, while hematocrit and other safety measures were monitored.

The median treatment duration was about 22 months, and median follow-up was about 33 months, which is substantial for an outcomes trial, but short compared with decades of potential use or the long development period of some cancers. Results should not be transferred automatically to injections, pellets, compounded products, supraphysiologic dosing, or use without laboratory and clinical monitoring. Route and dose can affect concentration peaks, erythrocytosis, blood pressure, adherence, and adverse effects.

The primary endpoint and noninferiority question#

The primary cardiovascular safety endpoint was the first occurrence of cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke. Such a composite is commonly called major adverse cardiovascular events.

An event occurred in 182 of 2,596 participants in the testosterone group, 7.0 percent, and 190 of 2,602 in the placebo group, 7.3 percent. The hazard ratio was 0.96, with a 95 percent confidence interval from 0.78 to 1.17. This met the trial's prespecified criterion for noninferiority.

Noninferiority means the data excluded an increase larger than the chosen margin under the prespecified analysis. It does not mean the treatments were identical, that risk was zero, or that testosterone protected the heart. The confidence interval tells you more than the observation that the point estimate sits just below 1.

The margin is also a value judgment built into design: a very generous margin can certify a clinically concerning increase as noninferior; an unrealistically narrow one can make a feasible safety trial impossible. So ask yourself whether the prespecified margin was clinically acceptable, and whether adherence and missing data could bias toward similarity.

Why the result was important#

Earlier observational studies and small trials produced conflicting cardiovascular signals. Confounding was difficult to resolve because men prescribed testosterone differ from men who are not, and low testosterone can be a marker of illness rather than a direct cause of vascular events.

Random allocation in TRAVERSE balanced measured and unmeasured baseline factors more effectively than an observational comparison. Blinding, placebo control, event adjudication, and a large high-risk population made the primary finding unusually informative. The conclusion is still bounded: when used as tested in selected men with confirmed hypogonadism, testosterone did not raise the primary composite beyond the trial's noninferiority margin during follow-up. That is a meaningful reassurance, not a universal clearance certificate.

Adverse-event imbalances deserve attention#

The primary publication reported higher incidences of atrial fibrillation, acute kidney injury, and pulmonary embolism in the testosterone group. These were not the prespecified primary endpoint, and multiple comparisons make any one secondary signal more vulnerable to chance.

The appropriate response is neither to declare each signal proven causal nor to ignore it because the primary endpoint was neutral. A pulmonary embolism signal is clinically relevant if you are weighing this for someone with prior thrombosis or strong thrombotic risk, and atrial fibrillation and kidney events similarly warrant context and surveillance.

Testosterone can also increase hematocrit. Excess red-cell concentration can prompt dose adjustment or discontinuation, depending on the product label and clinical guideline. Blood pressure is another class concern that became central in later FDA action.

The prostate substudy was reassuring within limits#

The prespecified prostate safety study followed 5,204 randomized participants using a structured prostate-monitoring plan. High-grade and any prostate cancer, acute urinary retention, invasive procedures, and new pharmacologic treatment for lower urinary tract symptoms were uncommon, with no statistically significant group differences.

Low event counts mean estimates remain imprecise. The trial excluded men at higher prostate risk, and active screening can identify issues earlier than ordinary care. Median follow-up cannot answer whether decades of therapy alter a cancer process with long latency. The fair conclusion is that the substudy did not demonstrate higher prostate-event incidence in the screened population over the study period. “Testosterone cannot affect prostate cancer” would go beyond the data.

The fracture substudy challenged an assumption#

Testosterone increases bone density in some men, so fewer fractures might appear biologically plausible. The fracture substudy included 5,204 participants and found no fracture reduction. Clinical fractures occurred in 3.50 percent of the testosterone group and 2.46 percent of the placebo group, with a hazard ratio of 1.43 and a 95 percent confidence interval from 1.04 to 1.97.

The investigators did not identify one proven mechanism for the higher incidence. Possible behavioral or fall-related explanations remain hypotheses. What matters for interpretation is that improved bone-density markers should not be substituted for measured fractures. TRAVERSE was not an osteoporosis treatment trial selected for high fracture risk, so its result does not answer every bone-health question. It does show why a favorable surrogate cannot guarantee a favorable clinical endpoint.

Other efficacy substudies#

Nested studies assessed anemia, sexual function, depressive symptoms, and progression from prediabetes. Testosterone corrected anemia more often in men who had anemia and reduced incident anemia in those without it. In men with low libido, it improved sexual activity, desire, and hypogonadal symptoms, while erectile-function effects were less compelling.

These average effects do not establish that every symptom attributed to “low T” will respond. Fatigue, low mood, reduced libido, and erectile difficulty have many causes, including sleep apnea, medication effects, vascular disease, depression, relationship factors, obesity, and endocrine disorders. A diagnostic workup matters because testosterone concentration can fall transiently with acute illness or severe calorie restriction. Repeated morning testing, symptom assessment, and evaluation of the hypothalamic, pituitary, and testicular axis help distinguish primary from secondary hypogonadism and identify reversible causes.

What FDA changed in 2025#

On February 28, 2025, FDA announced class-wide labeling changes after reviewing TRAVERSE and required ambulatory blood-pressure studies. FDA asked manufacturers to remove language from the boxed warning about an increased risk of adverse cardiovascular outcomes; it also required warnings about increased blood pressure for products where appropriate.

Those actions communicate two distinct conclusions. TRAVERSE supported removing the prior cardiovascular warning language, while blood-pressure studies supported a class-wide blood-pressure warning. A neutral major-event trial does not negate a physiological increase in blood pressure. At that time, FDA retained the limitation concerning use for age-related hypogonadism. Product labels remained the controlling source for indications, contraindications, warnings, and monitoring.

What changed in the June 2026 announcement#

On June 18, 2026, HHS and FDA announced that FDA was requesting further updates to testosterone labeling. The requested revisions included removal of limitation-of-use language stating that safety and effectiveness had not been established for age-related or idiopathic hypogonadism, updates to prostate-cancer information, and revisions to benign prostatic hyperplasia warnings and contraindication language.

The grammatical distinction is important: the announcement describes requested label changes. It should not be rewritten as if every marketed product already carries identical revised text. FDA's general Testosterone Information page, checked July 16, 2026, still states that approved products are for low testosterone associated with certain medical conditions and that products are not approved for low testosterone without an associated condition.

That lag is exactly why you check the current product label and the FDA database entry, not the announcement. A policy announcement, a request to manufacturers, a class-wide implementation, and the approved label in a particular package are not interchangeable stages.

What TRAVERSE did not test#

The trial did not test testosterone as a longevity therapy, athletic enhancer, weight-loss drug, or treatment for normal aging without confirmed biochemical deficiency. It did not establish safety at supraphysiologic levels or with unregulated preparations.

It did not settle fertility questions. Exogenous testosterone can suppress gonadotropins and sperm production, so if fertility matters to you, that is a different discussion. It also did not remove the need to assess sleep apnea, edema, prostate risk, hematocrit, blood pressure, medication interactions, and cardiovascular stability.

Finally, it did not prove a class effect across every delivery system and use pattern. The strongest evidence applies to a regimen close to the one tested, in a population close to the one enrolled, with comparable monitoring.

A practical evidence-based reading#

TRAVERSE should make you less likely to repeat an unsupported claim that properly prescribed testosterone necessarily causes excess major cardiovascular events. It should also make you less likely to say the treatment has no meaningful risks.

Diagnosis comes before prescribing. Benefit goals should be explicit and measurable. Monitoring should follow the exact product label and relevant guidelines. If your symptoms do not improve despite an adequate biochemical response, continuing by inertia deserves reconsideration.

The trial's real contribution is narrower and stronger than the slogans around it: in selected men with symptomatic and repeatedly confirmed hypogonadism who had cardiovascular disease or risk, titrated transdermal testosterone was noninferior to placebo for the primary major cardiovascular composite over about three years of follow-up.

References#

1. TRAVERSE cardiovascular safety trial
2. TRAVERSE prostate safety study
3. TRAVERSE fracture study
4. FDA 2025 class-wide testosterone label action
5. HHS and FDA June 2026 requested label updates
6. Current FDA Testosterone Information page

Testosterone diagnosis, treatment, contraindications, and monitoring require individualized clinical care.*