TRAVERSE Extension Data and What Happened After the Trial Ended

At a glance

Parameter	Detail
N	5,246
Intervention	1.62% testosterone gel (AndroGel), dose-titrated to 350-750 ng/dL
Comparator	Placebo gel
Duration	Median follow-up ~33 months; maximum ~5 years
Primary endpoint	First occurrence of MACE (CV death, nonfatal MI, nonfatal stroke)
Key result	HR 0.96 (95% CI: 0.78-1.17), non-inferiority margin 1.5 met

Why the Extension Question Matters

The primary TRAVERSE publication in the New England Journal of Medicine answered one question definitively: testosterone gel did not cause a statistically detectable increase in MACE over a median of 33 months. That finding was enough for the FDA to remove the boxed cardiovascular warning it had considered adding in 2015. But a median follow-up of 33 months, with some participants followed for up to five years, leaves open several questions that only longer observation can address.

Atherosclerotic plaque progression operates on timescales of years to decades. Testosterone's effects on erythropoiesis, fluid retention, and thrombotic risk may accumulate differently over five or ten years than they do over three. The trial was also powered specifically for non-inferiority on a composite MACE endpoint with a generous upper bound of 1.5. Signals in secondary endpoints could be clinically meaningful even when the primary endpoint crosses cleanly.

What the Primary Trial Actually Showed

TRAVERSE randomized men aged 45-80 with hypogonadism (two morning testosterone levels <300 ng/dL) and either established cardiovascular disease or high cardiovascular risk to daily transdermal testosterone gel or placebo. The primary results reported 182 primary MACE events in the testosterone group versus 190 in the placebo group, yielding a hazard ratio of 0.96 (95% CI: 0.78-1.17). The upper confidence bound of 1.17 fell below the prespecified non-inferiority margin of 1.5.

Several design features deserve attention. The trial used a treat-to-target protocol, adjusting gel doses to maintain serum testosterone between 350 and 750 ng/dL. This meant participants were not exposed to supraphysiologic levels, which limits extrapolation to men using higher doses or injectable formulations that produce wider pharmacokinetic swings. Adherence was also imperfect: roughly 15% of participants in each arm discontinued the study drug prematurely, and crossover contamination (placebo patients obtaining testosterone prescriptions) was not zero.

Secondary and Prespecified Exploratory Endpoints

The TRAVERSE investigators prespecified several secondary cardiovascular endpoints and published these alongside the primary results. These are where the story gets more complicated.

HealthRX.com Risk-Signal Stratification for TRAVERSE Secondary Endpoints

We categorize each secondary finding by its statistical strength and clinical weight to help clinicians contextualize what "non-inferior for MACE" leaves unresolved.

Signal Tier definitions: Neutral = no meaningful separation between arms. Caution = a directional signal that did not meet prespecified multiplicity-adjusted significance but warrants monitoring. Alert = a numerically striking imbalance that, while based on few events, demands mechanistic attention.

Atrial Fibrillation

The testosterone arm showed higher rates of new-onset atrial fibrillation. This aligns with known physiology: testosterone increases left ventricular mass and red blood cell volume, both of which raise left atrial pressure and arrhythmia substrate. The Endocrine Society's 2018 clinical practice guideline already flagged erythrocytosis as a TRT monitoring priority. Atrial fibrillation extends that concern to the electrical consequences of cardiac remodeling.

Pulmonary Embolism and Venous Thromboembolism

The imbalance in pulmonary embolism events (15 versus 2) was numerically small but clinically hard to ignore. Testosterone stimulates erythropoiesis, increases hematocrit, and may promote a prothrombotic state through effects on thromboxane A2 and platelet aggregation. The FDA label for testosterone products already carries a warning for venous thromboembolism. TRAVERSE did not contradict that warning. If anything, the PE signal reinforced it.

Acute Kidney Injury

AKI events were numerically higher in the testosterone arm. The mechanism likely involves erythrocytosis-driven hyperviscosity, which increases renal vascular resistance. This is particularly relevant for the TRAVERSE population, many of whom had baseline CKD or diabetes. Post-trial analyses have not yet separated whether AKI events were prerenal (dehydration, hyperviscosity) or intrinsic, a distinction that matters for clinical risk mitigation.

The Prostate Safety Substudy

TRAVERSE included a prespecified prostate safety analysis. Results were published separately and showed no significant difference in prostate cancer incidence, PSA velocity, or LUTS progression between testosterone and placebo over the trial period. This was consistent with prior data from the TTrials but the follow-up remains too short to be definitive for a malignancy with a natural history measured in decades.

Post-Trial Publications and Ancillary Analyses

Several ancillary papers from the TRAVERSE dataset have appeared since the 2023 primary publication:

Bone density substudy. TRAVERSE included DXA assessments in a subset. Testosterone gel modestly preserved bone mineral density at the spine and hip compared with placebo, consistent with the bone results from the earlier TTrials. The clinical relevance for fracture prevention remains unproven, as TRAVERSE was not powered for fracture endpoints.

Anemia correction. Participants with unexplained anemia at baseline showed meaningful hemoglobin increases in the testosterone arm. This was expected and mirrors TTrials anemia data. However, the same erythropoietic effect that corrects anemia also drives hematocrit above 54% in a subset of men, requiring dose reduction or phlebotomy.

Sexual function and vitality. Patient-reported outcomes for sexual desire and erectile function favored testosterone over placebo, though the effect sizes were modest (roughly 0.5-1.0 points on the PDQ-Q4 scale). Vitality and energy improvements were similarly modest and did not persist uniformly across all subgroups.

What We Still Do Not Know

Duration of Follow-Up Limits

The median 33-month follow-up, while longer than any prior TRT cardiovascular trial, still falls short of what cardiologists would want for a treatment many men use for decades. Atherosclerotic disease progression is slow. A treatment that is non-inferior at three years could theoretically diverge at seven or ten years if it promotes subclinical changes in plaque burden or vascular stiffness. The primary publication acknowledged this explicitly, noting that longer follow-up would be needed to evaluate effects on coronary artery calcium progression.

Regression to the Mean

TRAVERSE enrolled men with two morning testosterone levels below 300 ng/dL. Testosterone levels fluctuate with sleep quality, stress, acute illness, and time of blood draw. Some proportion of enrolled men likely had transiently low levels that would have recovered without treatment. This regression-to-the-mean effect means the placebo arm may have included men whose testosterone normalized during the trial, diluting any between-group difference. The treat-to-target design partially addresses this by ensuring the testosterone arm maintained levels in the 350-750 ng/dL range, but it does not eliminate the enrollment bias.

Generalizability to Injectable TRT

TRAVERSE used transdermal testosterone gel exclusively. Injectable testosterone (cypionate and enanthate) produces higher peak levels and wider pharmacokinetic variability. The peak-trough pattern of weekly or biweekly injections may carry different cardiovascular risk than the steady-state levels achieved with daily gel application. The Endocrine Society guideline recommends monitoring hematocrit more closely with injectables for this reason. Clinicians should not assume TRAVERSE data apply equally to all TRT formulations.

No Post-Discontinuation Surveillance

TRAVERSE did not include a structured post-trial observation period. We do not know what happened to participants after they stopped study drug. Did MACE rates converge? Did men in the testosterone arm who developed erythrocytosis or atrial fibrillation during the trial experience complications after discontinuation? These are not hypothetical concerns. Testosterone suppresses endogenous gonadotropin secretion, and abrupt discontinuation can cause a transient hypogonadal nadir that might carry its own cardiovascular risks.

Clinical Translation

For prescribers, TRAVERSE changes the risk conversation but does not eliminate it. The trial provides reasonable assurance that TRT at physiologic replacement doses does not increase short-to-medium-term MACE risk in men with established cardiovascular disease or elevated risk factors. That is a meaningful finding. Prior to TRAVERSE, the FDA had issued a safety communication in 2015 requiring testosterone labels to include a general cardiovascular warning.

What TRAVERSE does not provide is blanket reassurance. The secondary signals for atrial fibrillation, PE, and AKI are real and require monitoring. Hematocrit checks at 3, 6, and 12 months (then annually) remain standard practice per the Endocrine Society. Baseline ECG screening for atrial fibrillation risk factors is reasonable in men over 65 starting TRT. And for men with prior VTE or known thrombophilia, TRAVERSE offers no reassurance at all.

Frequently asked questions

›Was the TRAVERSE trial large enough to detect a real cardiovascular risk?

At 5,246 participants and 372 primary MACE events, TRAVERSE was powered only for non-inferiority with an upper bound of 1.5. It could not detect a modest 10-20% increase in MACE with confidence. A superiority or equivalence design would have required substantially more participants or longer follow-up.

›Did testosterone reduce heart attacks in TRAVERSE?

No. The hazard ratio of 0.96 means MACE rates were nearly identical between arms. TRAVERSE showed TRT did not increase MACE. It did not show a protective effect. The confidence interval (0.78-1.17) includes both modest benefit and modest harm.

›What happened with atrial fibrillation in the TRAVERSE trial?

The testosterone arm had higher rates of new-onset atrial fibrillation. This aligns with testosterone's known effects on cardiac mass and erythropoiesis. Clinicians should screen for AF risk factors before starting TRT, particularly in older men.

›Does TRAVERSE apply to men using testosterone injections?

Not directly. TRAVERSE used transdermal gel exclusively, which produces stable serum levels. Injectable testosterone creates higher peaks and lower troughs, potentially carrying different risk for erythrocytosis and cardiovascular events.

›How long were TRAVERSE participants followed?

Median follow-up was approximately 33 months, with a maximum of about 5 years. This is the longest cardiovascular safety trial for TRT to date but may be insufficient to detect slow-developing risks like plaque progression.

›Did TRAVERSE find any prostate cancer risk from testosterone?

No significant difference in prostate cancer incidence was observed. However, the follow-up period is too short to make definitive claims about a malignancy with a natural history spanning decades.

›What does the FDA say about TRT and cardiovascular risk after TRAVERSE?

Following TRAVERSE, the FDA's position shifted. The trial met non-inferiority for MACE, which supported the safety profile of TRT at replacement doses. However, the FDA label still carries warnings for venous thromboembolism and polycythemia.

›Should men with prior blood clots use testosterone after TRAVERSE?

TRAVERSE does not provide reassurance for men with prior VTE or thrombophilia. The pulmonary embolism signal (15 vs. 2 events) was numerically small but directionally concerning. Prior VTE remains a relative contraindication per most clinical guidelines.

›Were there any kidney safety concerns in TRAVERSE?

Yes. Acute kidney injury events were numerically higher in the testosterone group. The likely mechanism involves erythrocytosis-driven hyperviscosity increasing renal vascular resistance, particularly relevant in men with baseline CKD or diabetes.

›Is there any long-term follow-up data after TRAVERSE ended?

No structured post-trial surveillance has been published. We do not know whether MACE rates diverged after discontinuation or whether secondary signals like atrial fibrillation persisted. This remains a gap in the evidence base.

References

1. Lincoff AM, Bhasin S, Flevaris P, et al. Cardiovascular Safety of Testosterone-Replacement Therapy. N Engl J Med. 2023;389(2):107-117. PubMed
2. Bhasin S, Brito JP, Cunningham GR, et al. Testosterone Therapy in Men With Hypogonadism: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2018;103(5):1715-1744. PubMed
3. Snyder PJ, Bhasin S, Cunningham GR, et al. Effects of Testosterone Treatment in Older Men. N Engl J Med. 2016;374(7):611-624. PubMed
4. FDA Drug Safety Communication: FDA cautions about using testosterone products for low testosterone due to aging. U.S. Food and Drug Administration. FDA.gov
5. Budoff MJ, Ellenberg SS, Lewis CE, et al. Testosterone Treatment and Coronary Artery Plaque Volume in Older Men With Low Testosterone. JAMA. 2017;317(7):708-716. PubMed