The prostate and the heart: what TRAVERSE settled and what it did not
The two questions men ask most, answered from the largest randomized trial ever run on testosterone — including the parts of it that remain open.
Two sentences are commonly offered about testosterone safety, and neither survives contact with the evidence.
“TRT is safe.” The largest trial ever conducted was a non-inferiority trial. It was designed to ask whether testosterone is not unacceptably worse than placebo — not whether it is safe. It ran for a mean of under two years of treatment in men aged 45 to 80, using only a transdermal gel. It found several safety signals that reached statistical significance. It could not address long-term risk because it did not run long enough.
“TRT is dangerous.” The same trial found no increase in heart attacks, strokes or cardiovascular death. The observational studies that triggered a decade of fear were methodologically contested, and one prompted a formal retraction request from a large group of clinicians. The FDA removed the cardiovascular boxed warning from all testosterone products in February 2025.
What follows is organised in four parts: what we know, what TRAVERSE showed, what remains uncertain, and what should be monitored.
Part 1: What we know
The prostate
For most of the last eighty years, the assumption was that testosterone feeds prostate cancer. That assumption traces to work by Huggins and Hodges published in 1941, which established that removing androgens causes metastatic prostate cancer to regress — a finding so clinically important it won a Nobel Prize and remains the basis of androgen deprivation therapy today.
The problem is what got extrapolated from it. That work involved a small number of men with metastatic disease, using serum acid phosphatase as a surrogate marker. It demonstrated that taking androgens away from advanced cancer helps. It never tested whether giving testosterone to a hypogonadal man causes cancer. Those are different questions, and for decades the second was answered by assuming it followed from the first.
The counter-proposal is the saturation model, put forward by Morgentaler and Traish in European Urology in 2009: androgen receptors in prostate tissue become saturated at relatively low testosterone concentrations — in the region of 150 to 250 ng/dL — so raising testosterone above that adds no further stimulus. It explains why castration has dramatic effects while raising testosterone from 250 to 600 appears not to. It rests substantially on retrospective and pooled data, and has been formally challenged on those grounds. Treat it as a useful and plausible framework rather than settled fact.
What the accumulated trial data show: a 2026 meta-analysis of 41 randomized trials including 11,161 men found prostate cancer events at an odds ratio of 0.88 (95% CI 0.52–1.51) and clinically significant prostate cancer at 1.13 (95% CI 0.39–3.26). Neither was statistically significant, and the confidence intervals are wide enough to be honest about: this is evidence of no detected increase, not proof of no increase.
Evidence: 🟡 Promising that testosterone does not cause prostate cancer — with real residual uncertainty about long-term exposure
Cardiovascular risk before TRAVERSE
The fear had specific origins worth knowing, because they explain why the field is still arguing.
The TOM trial (2010). A study of testosterone in older, mobility-limited men — mean age 74 — was stopped early by its safety monitoring board after 23 cardiovascular events in the testosterone group versus 5 on placebo, among 209 men enrolled. Small numbers, a frail population, and an early stop, but alarming.
Vigen et al. (2013). A retrospective analysis of 8,709 men with low testosterone undergoing coronary angiography reported a three-year event rate of 25.7% on testosterone versus 19.9% without, adjusted hazard ratio 1.29 (95% CI 1.04–1.58). This study drew sustained methodological criticism and a formal retraction request from a large group of clinicians, on grounds including how events were counted and the confounding inherent in a non-randomized comparison.
Finkle et al. (2014). A claims analysis comparing 55,593 testosterone prescriptions against 167,279 PDE5-inhibitor prescriptions found the rate of myocardial infarction in the 90 days after starting testosterone was elevated relative to the preceding year: 1.36 (95% CI 1.03–1.81) overall, 2.19 (1.27–3.77) in men 65 and over, 3.43 (1.54–7.56) in men 75 and over, and 1.17 (0.84–1.63) — not significant — in men under 65 without heart disease.
Those studies produced the FDA’s 2015 safety communication, which required labelling to clarify that testosterone is approved for organic hypogonadism rather than age-related decline, and added cardiovascular warning language.
Randomized data pointed elsewhere. A 2022 individual-participant meta-analysis in Lancet Healthy Longevity pooling 17 trials found cardiovascular events in 120 of 1,601 men (7.5%) on testosterone versus 110 of 1,519 (7.2%) on placebo, with deaths in 6 of 1,621 versus 12 of 1,537 — numbers too small to draw mortality conclusions from, but not suggestive of harm.
TRAVERSE was commissioned by the FDA to settle it.
Part 2: What TRAVERSE showed
The design. 5,246 men randomized, aged 45 to 80, all with two fasting testosterone measurements below 300 ng/dL, at least one hypogonadal symptom, and either established cardiovascular disease or at least three cardiovascular risk factors. They received 1.62% transdermal testosterone gel — mean dose 65 mg — or matching placebo, titrated to a target of 350 to 750 ng/dL. Median baseline testosterone was 227 ng/dL. At twelve months, the testosterone group had risen by a median of 148 ng/dL; placebo by 14.
Mean treatment duration was approximately 21.7 months, with mean follow-up around 33 months.
The primary result
Major adverse cardiac events — cardiovascular death, non-fatal myocardial infarction, non-fatal stroke — occurred in 182 men (7.0%) on testosterone and 190 (7.3%) on placebo. Hazard ratio 0.96 (95% CI 0.78–1.17). Non-inferiority was met.
This is the finding that changed the field. In men at elevated cardiovascular risk, treated for roughly two years, testosterone did not increase heart attacks, strokes or cardiovascular death.
Note what it is not. A hazard ratio of 0.96 with a confidence interval crossing 1.0 in a non-inferiority trial does not mean testosterone protects the heart. It means no increase was detected.
The signals that did reach significance
| Event | Testosterone | Placebo |
|---|---|---|
| Pulmonary embolism | 0.9% | 0.5% |
| Atrial fibrillation | 3.5% (91) | 2.4% (63) |
| Acute kidney injury | 2.3% (60) | 1.5% (40) |
These are the findings that make “TRT is safe” untenable. The absolute numbers are small — a rise from 0.5% to 0.9% for pulmonary embolism is roughly four additional events per thousand men — but they are real, they were prespecified, and they are now in the FDA labelling.
Fracture: the counterintuitive one
A separate analysis of 5,204 participants published in the New England Journal of Medicine in 2024 found clinical fractures in 3.50% of the testosterone group versus 2.46% on placebo — hazard ratio 1.43 (95% CI 1.04–1.97) over a median 3.19 years. Most were ribs, wrist and ankle; more than 80% were trauma-associated, largely falls, and the pattern resembled typical osteoporotic fracture.
This runs directly against expectation. Testosterone increases bone mineral density. It was widely assumed that would translate into fewer fractures. In the largest trial ever run, more men on testosterone broke bones.
Nobody has a confident explanation. Increased physical activity leading to more falls is one hypothesis. It has not been demonstrated. What can be said is that testosterone should not be prescribed as fracture prevention, and that men with osteoporosis or a high falls risk warrant particular caution.
Prostate findings
TRAVERSE found a low overall incidence of prostate cancer, and of high-grade prostate cancer, with no significant difference between groups. High-grade prostate cancer — the trial’s primary prostate endpoint — occurred in 5 of 2,596 men on testosterone (0.19%) and 3 of 2,602 on placebo (0.12%), hazard ratio 1.62, 95% CI 0.39 to 6.77, p=0.51. Prostate cancer of any grade: 12 (0.46%) versus 11 (0.42%), hazard ratio 1.07, 95% CI 0.47 to 2.42, p=0.87. PSA rose by an average of 0.15 ng/mL at twelve months (95% CI 0.08 to 0.21) and showed no further between-group difference thereafter. Eighty-five men (1.6%) met protocol criteria for urological referral — 57 on testosterone, 28 on placebo — and of those, only 16 elected biopsy.
Two limitations matter. Digital rectal examination was not mandated, which limits detection sensitivity. And roughly two years of exposure says little about a cancer that typically develops over decades.
One number deserves more attention than it usually gets: invasive surgical prostate procedures for benign enlargement occurred in 23 men on testosterone (0.89%) versus 12 on placebo (0.46%), hazard ratio 1.91, 95% CI 0.95 to 3.84, p=0.07. That is not statistically significant and it is not cancer, but it is the one prostate signal in the trial that points anywhere at all.
Sexual function
In a substudy of 1,161 men, testosterone increased sexual activity — +0.49 acts per day at six months (95% CI 0.19–0.79) and +0.47 at twelve months, sustained to 24 months — with no significant improvement in erectile function.
Desire improved. Erections did not. In a population averaging around 65 with a heavy burden of vascular disease and diabetes, that is what you would predict, and almost no consumer site reports it.
Anemia
In the 814 men with anemia at baseline, testosterone corrected it substantially more often: 41.0% versus 27.5% at six months, and 45.0% versus 33.9% at twelve months. This is one of the clearest benefits the trial produced and gets almost no attention.
Diabetes — a correction worth stating plainly
Testosterone did not improve glycemic outcomes.
Among 1,175 men with prediabetes, progression to diabetes was not significantly different — hazard ratio 0.78 (95% CI 0.58–1.06), omnibus p=0.49. Among 3,880 men with diabetes, remission was not significantly different — hazard ratio 1.15 (95% CI 0.87–1.53), p=0.95.
Earlier reporting of this substudy circulated a figure suggesting testosterone reduced progression to diabetes by around a fifth. The published analysis in JAMA Internal Medicine does not support that. We are noting this explicitly because we nearly repeated it ourselves, and because it is currently repeated on a number of sites.
Adherence
61.4% of the testosterone group and 61.7% of the placebo group discontinued. In a trial of a drug men take because it makes them feel better, roughly three in five stopped. That complicates any interpretation of efficacy, and it is rarely mentioned.
What remains uncertain
Anything beyond about three years. Mean treatment duration was under two years. Prostate cancer takes decades. So does atherosclerosis. TRAVERSE cannot speak to a man who starts at 45 and continues for thirty years — which describes a large share of current patients.
Injectable testosterone. TRAVERSE studied transdermal gel only. The dominant regimen in the United States is intramuscular or subcutaneous injection, which produces markedly different peaks and a substantially higher rate of erythrocytosis. The cardiovascular reassurance does not automatically transfer.
Men under 45. Not studied. Not represented. The fastest-growing group of testosterone users.
Men at low cardiovascular risk. The trial deliberately enrolled high-risk men to accumulate events. Whether the findings generalize downward is an assumption, not a result.
Why fractures increased. Unexplained.
Whether erythrocytosis actually causes thrombosis. See below — this is a larger gap than most clinicians realize.
Mortality. All-cause mortality was not the primary endpoint and the trial was not powered for it.
Supraphysiologic dosing. TRAVERSE targeted 350 to 750 ng/dL. It says nothing about men running levels far above that, which is a substantial share of the “optimization” market.
Undertreatment within the trial. The median rise was 148 ng/dL from a baseline of 227 — reaching roughly 375. A meaningful proportion of treated men never reached a comfortably normal range, which cuts both ways: it may understate both benefit and harm.
Part 4: What should be monitored
Erythrocytosis and hematocrit
Testosterone raises red cell mass, substantially by suppressing hepcidin and thereby increasing iron availability. It is the commonest reason testosterone therapy is modified or stopped.
Thresholds by guideline:
| Body | Position |
|---|---|
| Endocrine Society | Hematocrit >48% (>50% for men living at high altitude) is a contraindication to starting; >54% on treatment is erythrocytosis — withhold, then resume at a lower dose |
| AUA | Withhold if baseline hematocrit >50%; stop or reduce if on-treatment hematocrit reaches 54% |
| EAU | Hematocrit should not exceed 54%; manage by dose reduction, formulation change, or phlebotomy |
Injectable formulations cause erythrocytosis more often than transdermal — this is well established qualitatively, though we could not verify formulation-specific incidence rates and are not going to quote figures.
And here is what almost nobody tells you. There is no controlled outcome data demonstrating that testosterone-induced erythrocytosis causes thrombotic events. Trials exclude men with high hematocrit and stop treatment when it rises, so the question has never been studied properly. In TRAVERSE, the association between hematocrit and both major cardiac events and venous thromboembolism was null.
The management is on equally thin ice. A 2024 review in Endocrine Connections concluded that evidence supporting the efficacy or safety of therapeutic phlebotomy in this setting is lacking. There is even a theoretical concern in the other direction: phlebotomy-induced iron deficiency may itself increase thrombotic risk through hypoxia-inducible factor pathways — the analogy being Chuvash erythrocytosis, where phlebotomy is contraindicated.
So the honest position: hematocrit is monitored because it is measurable and because the guidelines specify thresholds, not because the causal chain from hematocrit to clot has been demonstrated. Dose reduction or formulation change is more defensible than repeated venesection, and routine blood donation as a management strategy has no outcome evidence behind it at all.
Evidence: 🟢 Established that testosterone raises hematocrit · 🟠 Limited that this causes thrombosis · 🟠 Limited that phlebotomy helps
Venous thromboembolism
Beyond TRAVERSE, a case-control study in the BMJ drawing on 19,215 VTE cases and 909,530 controls found an adjusted rate ratio of 1.63 in the first six months of testosterone therapy, returning to approximately 1.00 thereafter.
That pattern — early risk that resolves — is consistent with the pulmonary embolism signal in TRAVERSE and suggests the first six months warrant more attention than the subsequent years.
PSA and the prostate
A baseline PSA before starting is appropriate in men for whom prostate cancer detection would change management. Expect a small rise — around 0.15 ng/mL at twelve months on trial data — then stability.
The Endocrine Society’s 2018 guideline recommends urological consultation for a confirmed PSA rise of more than 1.4 ng/mL above baseline within any twelve-month period, a confirmed PSA above 4.0 ng/mL, an abnormal digital rectal examination, or a substantial worsening of urinary symptoms.
Blood pressure
The FDA added a class-wide blood pressure warning to all testosterone products on 28 February 2025 — the same action that removed the cardiovascular boxed warning. Ambulatory monitoring in one study of 113 men found mean 24-hour systolic pressure rose 2.7 mmHg (95% CI 0.7–4.8) and diastolic 1.1 mmHg (95% CI −0.1–2.3) at sixteen weeks.
Small at an individual level. Meaningful across a population, and worth measuring.
Sleep apnea
Testosterone can worsen obstructive sleep apnea, and the best available characterization is that the effect is mild and time-limited. In a trial of 67 obese men with severe OSA given injectable testosterone undecanoate, the oxygen desaturation index worsened by 10.3 events per hour at week 7 — but showed no difference from placebo by week 18.
Worth knowing about, worth asking about before starting, not a reason for blanket avoidance.
What monitoring depends on
There is no universal schedule, and we decline to publish one. What to check and how often depends on the formulation, baseline hematocrit, prostate risk, age, cardiovascular risk, fertility intentions, and the symptoms being treated. See Monitoring Testosterone Therapy.
The summary a patient can use
Testosterone therapy in appropriately diagnosed men, at replacement doses, over two to three years, does not appear to increase heart attacks, strokes or cardiovascular death — and that is a genuine change from what was believed a decade ago.
It does appear to increase pulmonary embolism, atrial fibrillation, acute kidney injury and clinical fracture, all at low absolute rates.
It corrects anemia. It improves sexual desire. It does not improve erectile function in older men with vascular disease, and it does not improve glycemic control.
Nobody knows what thirty years of it does. Nobody has tested the injectable formulations most American men actually receive. Nobody has studied men under 45.
That is the honest state of the evidence. It supports treating men who need treatment, monitoring them properly, and declining to make promises in either direction.
Questions patients ask
TRT causes prostate cancer.
The evidence does not support it, and the belief rests on an extrapolation that was never tested.
What the evidence showsThe 1941 work establishing androgen deprivation for metastatic prostate cancer never asked whether raising testosterone causes cancer. A 2026 meta-analysis of 41 trials in 11,161 men found prostate cancer at an odds ratio of 0.88 (95% CI 0.52–1.51). TRAVERSE found low incidence with no significant difference.
What remains uncertainTrials run for years; prostate cancer develops over decades. TRAVERSE ran for a median of under two years of exposure, which is short against that timescale.
Bottom lineNo detected increase. Not the same as proven safety over a lifetime.
Moderate
TRT will accelerate a prostate cancer I already have.
Plausible for advanced disease; increasingly questioned for early or treated disease.
What the evidence showsAndrogen deprivation clearly treats metastatic prostate cancer, so androgens clearly matter in that setting. The saturation model proposes receptors are saturated around 150 to 250 ng/dL, which would explain why raising testosterone within the normal range adds little stimulus. The model rests on retrospective data and has been challenged.
What remains uncertainWhether the saturation model holds across disease states.
Bottom lineUntreated or advanced prostate cancer remains a genuine caution. This is a urologist's decision, not a clinic's.
Limited
Men with a history of prostate cancer can never have testosterone.
Not never. The AUA explicitly allows it in defined circumstances.
What the evidence showsThe AUA states testosterone can be considered after radical prostatectomy in men with favourable pathology — negative margins, seminal vesicles and nodes — and an undetectable postoperative PSA. Studies after radiation have not shown a recurrence signal, and limited data in men on active surveillance show no significant PSA rise or new cancer signal.
What remains uncertainLong-term recurrence data are limited and the studies are small.
Bottom linePossible, in selected men, managed by a urologist. Not a routine decision.
Limited
TRT raises your PSA a lot.
A little, early, then it stabilizes.
What the evidence showsIn TRAVERSE, PSA rose by an average of 0.15 ng/mL at twelve months (95% CI 0.08 to 0.21), with no further between-group difference. A modest early rise in a previously deficient man is expected.
What remains uncertainHow to distinguish a treatment-related rise from an early cancer signal in an individual — which is why velocity thresholds exist.
Bottom lineA small early rise is expected. A large or continuing rise is not.
Moderate
Men on TRT need PSA testing.
Baseline and follow-up testing is appropriate for men in whom detecting prostate cancer would change management.
What the evidence showsThe Endocrine Society recommends urological referral for a confirmed rise above 1.4 ng/mL within any twelve-month period, a confirmed PSA above 4.0, an abnormal rectal examination, or substantially worsened urinary symptoms.
What remains uncertainOptimal frequency, and whether men over 70 with limited life expectancy benefit from screening at all.
Bottom lineYes for most men starting treatment — with the same caveats that apply to PSA screening generally.
Moderate
TRT causes heart attacks.
The largest trial found it did not.
What the evidence showsIn TRAVERSE, major adverse cardiac events occurred in 7.0% on testosterone versus 7.3% on placebo — hazard ratio 0.96 (95% CI 0.78–1.17) — in 5,246 men who all had elevated cardiovascular risk. The FDA removed the cardiovascular boxed warning from all testosterone products in February 2025.
What remains uncertainBeyond three years, in injectable formulations, in men under 45, and at supraphysiologic doses — none of which TRAVERSE tested.
Bottom lineThe heart attack fear is not supported for replacement-dose therapy over a few years. That is narrower than "TRT is safe."
Strong within the trial's limits
TRT protects the heart.
No. That is the mirror-image error.
What the evidence showsTRAVERSE was a non-inferiority trial with a hazard ratio of 0.96 and a confidence interval crossing 1.0. That is an absence of detected harm, not a demonstration of benefit.
What remains uncertainWhether any cardiovascular benefit exists in any subgroup. No trial has shown one.
Bottom lineNobody should be taking testosterone for their heart.
Strong (that benefit is unproven)
TRT causes blood clots.
There is a real signal, concentrated early, at low absolute rates.
What the evidence showsTRAVERSE found pulmonary embolism in 0.9% on testosterone versus 0.5% on placebo. A *BMJ* case-control study of 19,215 VTE cases found an adjusted rate ratio of 1.63 in the first six months, returning to about 1.00 afterwards.
What remains uncertainWhich men are susceptible, and whether the risk is mediated by hematocrit — the TRAVERSE data suggest not.
Bottom lineSmall, real, and front-loaded. Unexplained leg swelling or breathlessness in the first months deserves urgent attention.
Moderate
TRT increases hematocrit.
Yes, reliably, and more so with injections.
What the evidence showsTestosterone increases red cell mass substantially by suppressing hepcidin, increasing iron availability for erythropoiesis. It is the commonest reason therapy is modified or stopped.
What remains uncertainFormulation-specific incidence rates — well established in direction, and we could not verify figures.
Bottom lineExpect it, monitor it, and know that dose and formulation drive it.
Strong
A hematocrit of 54% is dangerous.
It's the guideline threshold for acting. Whether it is dangerous has never actually been demonstrated.
What the evidence showsThe Endocrine Society, AUA and EAU all converge on 54% as the point to reduce dose or stop. But no controlled outcome data link testosterone-induced erythrocytosis to thrombotic events — trials exclude and withdraw these men, so the question has never been studied. In TRAVERSE the association between hematocrit and both cardiac events and VTE was null.
What remains uncertainWhether the threshold reflects real risk or measurable convention.
Bottom lineAct on it because the guidelines say so and the downside of acting is small — not because the causal chain is proven.
Strong on the threshold's existence · Limited on the risk it represents
Donating blood solves high hematocrit on TRT.
It lowers the number. Whether that helps you has not been shown.
What the evidence showsA 2024 review in *Endocrine Connections* concluded that evidence supporting the efficacy or safety of therapeutic phlebotomy in this setting is lacking. There is a theoretical concern that phlebotomy-induced iron deficiency could itself increase thrombotic risk through hypoxia-inducible factor pathways.
What remains uncertainWhether repeated venesection helps, does nothing, or causes harm. Nobody has run the trial.
Bottom lineAddressing the dose or the formulation is more defensible than repeatedly removing blood.
Limited
TRT raises blood pressure.
Modestly, consistently enough that the FDA required a class-wide warning.
What the evidence showsThe FDA added a blood pressure warning to all testosterone products on 28 February 2025. Ambulatory monitoring in 113 men found a mean 24-hour systolic rise of 2.7 mmHg (95% CI 0.7–4.8) at sixteen weeks.
What remains uncertainWhether the effect differs meaningfully by formulation, and whether it translates into events.
Bottom lineSmall individually, real at population scale, easily monitored.
Strong
TRT worsens sleep apnea.
It can, and the effect appears mild and temporary.
What the evidence showsIn 67 obese men with severe OSA on injectable testosterone undecanoate, the oxygen desaturation index worsened by 10.3 events per hour at week 7 but showed no difference from placebo by week 18.
What remains uncertainWhether this holds for other formulations, doses, and less severe apnea.
Bottom lineAsk about sleep apnea before starting, not after. Not a blanket contraindication.
Moderate
TRT is safe after a heart attack.
TRAVERSE enrolled men with established cardiovascular disease and found no increase in events — which is the best evidence available, and not a green light for the acute period.
What the evidence showsEvery man in TRAVERSE had either established cardiovascular disease or at least three risk factors, and the primary endpoint showed no increase. The trial did not enrol men in the immediate post-infarction period.
What remains uncertainTiming after an acute event, and management in unstable disease.
Bottom lineReassuring for stable cardiovascular disease. A cardiologist's call for anything recent.
Moderate
TRT causes fractures — surely it prevents them?
This is the trial's most counterintuitive finding, and it went the wrong way.
What the evidence showsIn 5,204 TRAVERSE participants, clinical fractures occurred in 3.50% on testosterone versus 2.46% on placebo — hazard ratio 1.43 (95% CI 1.04–1.97). Most were ribs, wrist and ankle, and over 80% were trauma-associated. Testosterone does increase bone mineral density.
What remains uncertainWhy. Increased activity leading to more falls is a hypothesis, not a finding.
Bottom lineDo not take testosterone to prevent fractures. Extra caution with osteoporosis or falls risk.
Moderate
TRT prevents diabetes.
No — and this claim is circulating with a number attached that the published analysis does not support.
What the evidence showsIn the *JAMA Internal Medicine* analysis, among 1,175 men with prediabetes, progression to diabetes was not significantly different (HR 0.78, 95% CI 0.58–1.06, omnibus p=0.49). Among 3,880 men with diabetes, remission was not significantly different (HR 1.15, 95% CI 0.87–1.53, p=0.95).
What remains uncertainWhether any metabolic benefit exists in a subgroup or over longer treatment.
Bottom lineTestosterone changes body composition. It did not change glycemic outcomes.
Strong (null result)
TRT improves anemia.
Yes, and this is one of the clearest benefits in the trial.
What the evidence showsAmong 814 men with anemia at baseline, anemia corrected in 41.0% on testosterone versus 27.5% on placebo at six months, and 45.0% versus 33.9% at twelve.
What remains uncertainWhether correcting anemia this way improves how men actually function.
Bottom lineA real, replicated benefit that gets far less attention than the risks.
Strong
TRT increases mortality.
No evidence of that, and no trial designed to answer it.
What the evidence showsTRAVERSE showed no increase in cardiovascular death within its primary endpoint. A 2022 individual-participant meta-analysis of 17 trials found deaths in 6 of 1,621 on testosterone versus 12 of 1,537 on placebo — numbers far too small for conclusions.
What remains uncertainAll-cause mortality was never a powered endpoint in any trial.
Bottom lineNo mortality signal. Also no mortality answer.
Limited
TRAVERSE proved TRT is safe.
It proved something narrower and more useful than that.
What the evidence showsTRAVERSE was a non-inferiority trial in men aged 45 to 80 at elevated cardiovascular risk, using transdermal gel only, with a mean treatment duration under two years and 61% discontinuation in both arms. It found no increase in major adverse cardiac events — and increases in pulmonary embolism, atrial fibrillation, acute kidney injury and fracture.
What remains uncertainLong-term risk, injectable formulations, men under 45, low-risk men, and supraphysiologic dosing — none of which it studied.
Bottom lineIt removed the heart attack fear for a defined population over a defined period. It did not certify the drug.
Strong
The old studies showing TRT causes heart attacks were just wrong.
They were methodologically contested and they were also the best evidence at the time.
What the evidence showsThe 2013 Vigen analysis reported an adjusted hazard ratio of 1.29 (95% CI 1.04–1.58) and drew a formal retraction request over its methods. The 2014 Finkle claims study found an elevated 90-day myocardial infarction rate concentrated in older men — 2.19 in men over 65, 3.43 in men over 75 — and not significant in men under 65 without heart disease. Both were non-randomized and subject to confounding by indication.
What remains uncertainWhether the age-concentrated signal reflected something real that TRAVERSE's population and duration could not detect.
Bottom lineObservational data raised a reasonable question. A randomized trial answered it better. That is how it is supposed to work.
Moderate
Where to go next
The trial in detail: What TRAVERSE Actually Showed
Managing a high hematocrit: Erythrocytosis and Testosterone
What to check and when: Monitoring Testosterone Therapy
Understanding a specific lab value: The Lab Library
Ready for testosterone care built on a diagnosis?
The first step is a full endocrine evaluation, not a prescription. We see patients across San Diego County and welcome referrals from other physicians.