What was the primary endpoint of the TRAVERSE trial?

The primary endpoint was time to first major adverse cardiovascular event (MACE), defined as the composite of cardiovascular death, non-fatal myocardial infarction, or non-fatal stroke. An independent blinded committee adjudicated all events.

How many men were enrolled in TRAVERSE?

A total of 5,246 men were randomized: 2,653 to testosterone 1.62% gel and 2,593 to matching placebo gel. Mean follow-up was 33 months.

What does non-inferiority at a 1.5 margin actually mean?

It means the trial ruled out a 50% or greater increase in MACE with testosterone relative to placebo. The observed upper confidence bound was 1.17, well within the 1.5 threshold. It does not rule out smaller increases in risk.

Were injectable testosterone formulations studied in TRAVERSE?

No. TRAVERSE used transdermal testosterone gel only. Injectable formulations produce different pharmacokinetic profiles and their cardiovascular safety has not been established by a dedicated outcomes trial.

Did TRAVERSE find any safety signals beyond MACE?

Yes. Pulmonary embolism (HR 7.10), atrial fibrillation (HR 1.52), and acute kidney injury (HR 1.77) were more common in the testosterone group. These findings require further investigation in dedicated trials.

Who was eligible for the TRAVERSE trial?

Men aged 45 to 80 with confirmed hypogonadism (two fasting testosterone levels <300 ng/dL), at least one symptom of deficiency, and either established cardiovascular disease or multiple cardiovascular risk factors.

How was blinding maintained despite dose titration?

An unblinded pharmacist managed dose adjustments based on serum testosterone levels. The treating investigator, participant, and endpoint adjudicators remained blinded throughout the trial.

Did the FDA change the testosterone label after TRAVERSE?

The FDA reviewed TRAVERSE data and updated testosterone product labeling in 2023 to reflect the cardiovascular non-inferiority finding while maintaining warnings about venous thromboembolism risk.

How does TRAVERSE compare to the earlier TOM trial?

The TOM trial (2010) was a small (N=209) frailty study in men ≥65 that was stopped early due to excess cardiovascular events with testosterone. TRAVERSE was specifically designed to provide the definitive answer the TOM trial could not, with 25 times the sample size and formal MACE adjudication.

Can TRAVERSE results be applied to younger, healthier men on TRT?

Not directly. The trial enrolled men with established or high-risk cardiovascular disease. Extrapolating to younger, lower-risk men requires caution, as both baseline event rates and risk-benefit ratios differ.

Inside the TRAVERSE Methodology: What Most Summaries Skip

Q: How was blinding maintained despite dose titration?

An unblinded pharmacist managed dose adjustments based on serum testosterone levels. The treating investigator, participant, and endpoint adjudicators remained blinded throughout the trial.

Q: Did the FDA change the testosterone label after TRAVERSE?

The FDA reviewed TRAVERSE data and updated testosterone product labeling in 2023 to reflect the cardiovascular non-inferiority finding while maintaining warnings about venous thromboembolism risk.

Q: How does TRAVERSE compare to the earlier TOM trial?

The TOM trial (2010) was a small (N=209) frailty study in men ≥65 that was stopped early due to excess cardiovascular events with testosterone. TRAVERSE was specifically designed to provide the definitive answer the TOM trial could not, with 25 times the sample size and formal MACE adjudication.

Q: Can TRAVERSE results be applied to younger, healthier men on TRT?

Not directly. The trial enrolled men with established or high-risk cardiovascular disease. Extrapolating to younger, lower-risk men requires caution, as both baseline event rates and risk-benefit ratios differ.

By HealthRX Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

At a glance

| Detail | Value | |---|---| | N | 5,246 (2,653 testosterone; 2,593 placebo) | | Intervention | Transdermal testosterone 1.62% gel (AndroGel), dose-titrated to 350 to 750 ng/dL | | Comparator | Matching placebo gel | | Duration | Mean follow-up 33 months; maximum ~5 years | | Primary endpoint | First occurrence of MACE (CV death, non-fatal MI, non-fatal stroke) | | Key result | HR 0.96 (95% CI 0.78, 1.17); non-inferiority margin 1.5 met |

Why the Design Matters More Than the Headline

Most coverage of TRAVERSE stops at "testosterone didn't increase heart attacks." That framing obscures almost everything a prescriber needs to know. Non-inferiority trials answer a narrower question than superiority trials: "Is the new treatment not unacceptably worse?" Every parameter in that question, the margin, the population, the endpoint, the analysis set, shapes the clinical meaning of the answer.

This page dissects each of those parameters.

Population Enrichment: Who Got In and Who Didn't

TRAVERSE enrolled men aged 45 to 80 with hypogonadism (two fasting morning testosterone levels <300 ng/dL) plus at least one symptom of testosterone deficiency. The critical design feature was the cardiovascular enrichment strategy. Every participant had either pre-existing cardiovascular disease (documented coronary, cerebrovascular, or peripheral arterial disease) or elevated CV risk defined by multiple factors such as hypertension, dyslipidemia, diabetes, smoking, or CKD.

This was intentional. The 2015 FDA advisory mandating cardiovascular outcome data specifically flagged the older, comorbid population where observational signals had emerged. Enrolling lower-risk men would have diluted event rates and required an even larger sample, while also answering the wrong clinical question.

Exclusion criteria removed men with recent acute coronary syndrome (<3 months), uncontrolled heart failure (NYHA class III, IV), hematocrit above 48%, severe untreated sleep apnea, and PSA above 3.0 ng/mL. These cutoffs reflect both safety conservatism and the Endocrine Society clinical practice guidelines for TRT monitoring.

What this means in practice: TRAVERSE results apply most directly to men who resemble TRAVERSE participants. Younger, healthier hypogonadal men were not studied. The trial does not prove testosterone is "safe for everyone."

Randomization and Blinding

Randomization was 1:1 with stratification by age (<65 vs. ≥65) and pre-existing cardiovascular disease (yes vs. no). Stratification by age was important because the 2010 TOM trial that initially raised safety concerns enrolled men ≥65 exclusively. Stratification by existing CVD ensured balanced representation across primary and secondary prevention subgroups.

Blinding relied on identical-appearing gel packets. The dose titration protocol (targeting serum testosterone of 350 to 750 ng/dL based on levels drawn at visits) introduced a practical challenge: clinicians adjusting dose based on testosterone levels could theoretically infer assignment if placebo subjects consistently showed low levels. TRAVERSE addressed this by having an unblinded pharmacist manage dose adjustments while keeping the investigator, participant, and endpoint adjudicators blinded. This pharmacy-based unblinding firewall is standard in titration-based trials but worth noting because it adds one layer of human process that could, in theory, leak.

The Primary Endpoint: Three-Point MACE

The primary composite was time to first occurrence of cardiovascular death, non-fatal myocardial infarction, or non-fatal stroke. This three-point MACE definition matches the FDA guidance for cardiovascular outcome trials originally developed for diabetes drugs and later applied to the testosterone mandate.

An independent clinical events committee adjudicated all potential events while blinded to treatment assignment. Adjudication used standardized definitions aligned with the Standardized Data Collection for Cardiovascular Trials Initiative, which reduces misclassification of ambiguous events like troponin elevations without clear ischemic symptoms.

Choosing three-point over four-point MACE (which adds hospitalization for unstable angina) was conservative in the sense that the softer angina component tends to inflate event counts and dilute signal. The tighter composite keeps the endpoint anchored to hard, less ambiguous outcomes.

The Non-Inferiority Margin: Why 1.5 Matters

The pre-specified non-inferiority margin was a hazard ratio of 1.5. This means the trial was designed to rule out the possibility that testosterone gel increased MACE risk by 50% or more relative to placebo.

Is 1.5 too generous? Cardiovascular outcome trials for diabetes drugs used a margin of 1.3 per FDA guidance. TRAVERSE's wider margin reflects the different regulatory context: testosterone is not treating a disease with established CV-modifying therapies as background. There was no validated active comparator against which to calibrate a tighter margin. The 1.5 boundary was negotiated with the FDA during the Special Protocol Assessment.

In the actual results, the upper bound of the 97.5% one-sided confidence interval was 1.17, clearing the 1.5 margin with room to spare. It also cleared a post-hoc 1.3 boundary. But the trial was not powered for 1.3, so the statistical properties (type I error control, power) only formally hold for the 1.5 threshold.

Statistical Approach and the Estimand Framework

The primary analysis used a Cox proportional-hazards model stratified by the randomization factors (age group, existing CVD). The analysis followed the intention-to-treat principle: all randomized participants were included regardless of adherence.

TRAVERSE also pre-specified a per-protocol analysis excluding major protocol violators, and a sensitivity analysis censoring at treatment discontinuation (an on-treatment estimand). Both supported the primary finding. The on-treatment analysis is particularly relevant because testosterone's biological effects dissipate within weeks of discontinuation. If harms were driven by active drug exposure, the on-treatment analysis would concentrate them. It showed no divergence from the ITT result.

Event-driven design required a minimum of 256 primary endpoint events. The trial accumulated 364 adjudicated MACE events across a mean follow-up of 33.0 months, giving the analysis more statistical information than originally planned.

The proportional-hazards assumption was tested and supported by Schoenfeld residuals. Kaplan-Meier curves for MACE tracked closely throughout follow-up with no late separation, which would have suggested a delayed harm signal missed by the overall hazard ratio.

The Comparator: Why Placebo and Not Another TRT

TRAVERSE compared testosterone gel to placebo, not to injections, patches, or pellets. This was the correct choice for the regulatory question (does testosterone exposure increase CV risk vs. no testosterone?) but limits generalizability across formulations. Transdermal gel produces relatively stable physiological testosterone levels. Injectable testosterone cypionate or enanthate produces supraphysiological peaks followed by troughs. Whether peak-trough kinetics carry different cardiovascular risk remains unanswered by TRAVERSE.

The specific product was 1.62% testosterone gel (commercially, AndroGel). Dose titration targeted mid-normal range. Achieved mean testosterone at 12 months was approximately 358 ng/dL in the active group vs. 196 ng/dL in placebo. The separation was real but modest; some placebo-group men had endogenous levels that overlapped with treated men, which dilutes the treatment contrast and makes the non-inferiority finding more conservative.

Results in Detail

| Outcome | Testosterone (n = 2,653) | Placebo (n = 2,593) | HR (95% CI) | |---|---|---|---| | Primary MACE | 182 events (7.0%) | 182 events (7.3%) | 0.96 (0.78, 1.17) | | CV death | 48 (1.8%) | 46 (1.8%) | 1.00 (0.67, 1.50) | | Non-fatal MI | 100 (3.8%) | 101 (3.9%) | 0.95 (0.72, 1.25) | | Non-fatal stroke | 38 (1.4%) | 40 (1.6%) | 0.91 (0.58, 1.42) | | Pulmonary embolism | 15 (0.6%) | 2 (0.1%) | 7.10 (1.63, 30.94) | | Atrial fibrillation | 57 (2.2%) | 36 (1.4%) | 1.52 (1.00, 2.31) | | AKI | 57 (2.2%) | 31 (1.2%) | 1.77 (1.14, 2.75) |

The MACE result is straightforward non-inferiority. The safety signals in pulmonary embolism, atrial fibrillation, and acute kidney injury warrant separate attention. Venous thromboembolism risk with testosterone is biologically plausible (erythrocytosis increases viscosity) and consistent with the FDA label warning added in 2014. The atrial fibrillation signal was borderline but aligns with emerging mechanistic data on androgen-mediated atrial remodeling.

Limitations the Authors Acknowledged

The published report explicitly listed several limitations worth repeating because they define the boundaries of the conclusion:

Formulation specificity. Results apply to transdermal gel. Injectables and pellets were not tested.
Population specificity. All men had established or high-risk CVD. Lower-risk hypogonadal men may have different baseline event rates and different risk-benefit profiles.
Duration. Mean follow-up was 33 months. Very long-term effects (atherosclerotic plaque progression over decades) remain uncharacterized, though the TRAVERSE coronary CT substudy found no difference in noncalcified plaque volume.
Margin. Non-inferiority at 1.5 does not mean "no risk." It means the data are inconsistent with a ≥50% increase in MACE. A true 20 to 30% increase would not have been detected.
Adherence. Approximately 15% of participants discontinued treatment early. ITT analysis dilutes any on-treatment signal, though the on-treatment sensitivity analysis was consistent.

How This Changes Practice

Before TRAVERSE, clinicians prescribing TRT to men with cardiovascular disease were operating in an evidence vacuum filled by conflicting observational data and the 2015 FDA safety communication. After TRAVERSE, the American Urological Association and prescribers have randomized evidence that testosterone gel does not cause a large increase in MACE in this population.

That is a meaningful clinical advance, but the trial does not say TRT is cardioprotective, does not address injectable formulations, and raises new questions about VTE and atrial fibrillation that require dedicated study.

Frequently asked questions

›

References

Lincoff AM, Bhasin S, Fleg JL, et al. Cardiovascular Safety of Testosterone-Replacement Therapy. N Engl J Med. 2023;389(2):107-117. PubMed
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
Basaria S, Coviello AD, Travison TG, et al. Adverse Events Associated with Testosterone Administration (TOM Trial). N Engl J Med. 2010;363(2):109-122. PubMed
FDA Drug Safety Communication: FDA cautions about using testosterone products for low testosterone due to aging. 2015. FDA.gov
AndroGel (testosterone gel) 1.62% prescribing information. FDA Label
Hicks KA, Tcheng JE, Bozkurt B, et al. 2014 ACC/AHA Key Data Elements and Definitions for Cardiovascular Endpoint Events in Clinical Trials. J Am Coll Cardiol. 2015;66(4):403-469. PubMed