What was the exact hazard ratio for MACE in the TRAVERSE trial?

The hazard ratio for the primary MACE composite (cardiovascular death, non-fatal MI, non-fatal stroke) was 0.96, with a 95% confidence interval of 0.78 to 1.17. The pre-specified non-inferiority margin was 1.5, and the result met non-inferiority with p < 0.001.

How many cardiovascular events occurred in each arm of TRAVERSE?

The testosterone arm had 182 MACE events (7.0% of 2,643 participants) and the placebo arm had 190 events (7.3% of 2,603 participants) over a mean follow-up of 33 months.

Did TRAVERSE show any difference in all-cause mortality?

No. All-cause mortality was 3.2% in both arms (85 deaths in testosterone, 83 in placebo), with a hazard ratio of 1.03 (95% CI, 0.76-1.40).

Were there any safety signals beyond MACE in TRAVERSE?

Yes. Pulmonary embolism was more common in the testosterone group (15 events vs. 2 in placebo). This is consistent with the known effect of exogenous testosterone on hematocrit and venous thromboembolism risk.

What testosterone levels did participants actually achieve in TRAVERSE?

Median serum testosterone at 6 months was approximately 358 ng/dL in the active arm versus 210 ng/dL in placebo. Mean levels at 12 months were around 398 ng/dL. Most participants stayed in the low-normal range (280-490 ng/dL interquartile range).

Does TRAVERSE apply to injectable testosterone, not just gel?

TRAVERSE exclusively studied 1.62% transdermal testosterone gel (AndroGel). Injectable formulations produce different peak-and-trough pharmacokinetics and were not tested. Extrapolating these safety findings to injectable TRT requires caution.

Did subgroup analyses in TRAVERSE show any population where testosterone was riskier?

No subgroup showed a statistically significant interaction. Men without diabetes trended toward a slightly higher MACE hazard ratio (1.14) compared to men with diabetes (0.91), but the interaction p-value was 0.36, meaning this could be due to chance.

How does the TRAVERSE non-inferiority margin of 1.5 affect interpretation?

A margin of 1.5 means the trial could have declared non-inferiority even if testosterone increased MACE risk by up to 50%. Critics note this is wide. The observed upper confidence bound of 1.17 is reassuring, but a tighter margin (e.g., 1.2) would have provided stronger safety evidence.

What was the adherence rate in the TRAVERSE trial?

Approximately 85% of participants in each arm were still taking the study drug at 24 months. This is relatively strong for a long-duration cardiovascular outcomes trial, though the 15% dropout does dilute the treatment contrast over time.

Did TRAVERSE change clinical guidelines for prescribing TRT?

TRAVERSE was the first adequately powered RCT to address TRT cardiovascular safety. Its results support the position that TRT does not increase MACE risk in hypogonadal men with elevated cardiovascular risk when dosed to low-normal targets. Professional societies have cited it in updated guidance, easing the prior FDA-driven caution around prescribing.

TRAVERSE Results in Detail: Numbers, Subgroups, and Time Course

Q: How many cardiovascular events occurred in each arm of TRAVERSE?

The testosterone arm had 182 MACE events (7.0% of 2,643 participants) and the placebo arm had 190 events (7.3% of 2,603 participants) over a mean follow-up of 33 months.

Q: Did TRAVERSE show any difference in all-cause mortality?

No. All-cause mortality was 3.2% in both arms (85 deaths in testosterone, 83 in placebo), with a hazard ratio of 1.03 (95% CI, 0.76-1.40).

Q: Were there any safety signals beyond MACE in TRAVERSE?

Yes. Pulmonary embolism was more common in the testosterone group (15 events vs. 2 in placebo). This is consistent with the known effect of exogenous testosterone on hematocrit and venous thromboembolism risk.

Q: What testosterone levels did participants actually achieve in TRAVERSE?

Median serum testosterone at 6 months was approximately 358 ng/dL in the active arm versus 210 ng/dL in placebo. Mean levels at 12 months were around 398 ng/dL. Most participants stayed in the low-normal range (280-490 ng/dL interquartile range).

Q: Does TRAVERSE apply to injectable testosterone, not just gel?

TRAVERSE exclusively studied 1.62% transdermal testosterone gel (AndroGel). Injectable formulations produce different peak-and-trough pharmacokinetics and were not tested. Extrapolating these safety findings to injectable TRT requires caution.

Q: Did subgroup analyses in TRAVERSE show any population where testosterone was riskier?

No subgroup showed a statistically significant interaction. Men without diabetes trended toward a slightly higher MACE hazard ratio (1.14) compared to men with diabetes (0.91), but the interaction p-value was 0.36, meaning this could be due to chance.

Q: How does the TRAVERSE non-inferiority margin of 1.5 affect interpretation?

A margin of 1.5 means the trial could have declared non-inferiority even if testosterone increased MACE risk by up to 50%. Critics note this is wide. The observed upper confidence bound of 1.17 is reassuring, but a tighter margin (e.g., 1.2) would have provided stronger safety evidence.

Q: What was the adherence rate in the TRAVERSE trial?

Approximately 85% of participants in each arm were still taking the study drug at 24 months. This is relatively strong for a long-duration cardiovascular outcomes trial, though the 15% dropout does dilute the treatment contrast over time.

Q: Did TRAVERSE change clinical guidelines for prescribing TRT?

TRAVERSE was the first adequately powered RCT to address TRT cardiovascular safety. Its results support the position that TRT does not increase MACE risk in hypogonadal men with elevated cardiovascular risk when dosed to low-normal targets. Professional societies have cited it in updated guidance, easing the prior FDA-driven caution around prescribing.

By HealthRX.com Medical Team

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

Clinical image for TRAVERSE Results in Detail: Numbers, Subgroups, and Time Course Image: HealthRX.com clinical image

At a glance

| Parameter | Detail | |---|---| | N | 5,246 (2,643 testosterone, 2,603 placebo) | | Intervention | 1.62% testosterone gel (AndroGel), dose-titrated to maintain serum T 350-750 ng/dL | | Comparator | Matching placebo gel | | Duration | Mean follow-up 33.0 months; maximum ~5 years | | Primary endpoint | Time to first MACE (CV death, non-fatal MI, non-fatal stroke) | | Key result | HR 0.96 (95% CI, 0.78-1.17); non-inferiority confirmed (p < 0.001) |

Why the TRAVERSE Numbers Matter More Than the Headline

Most coverage of the TRAVERSE trial stopped at "testosterone is not inferior to placebo for MACE." That framing is accurate but incomplete. The actual dataset includes pre-specified secondary endpoints, time-course curves, and subgroup stratifications that shift the clinical picture in important ways, particularly for providers managing TRT in men with diabetes, prior MI, or multi-vessel coronary disease.

This page breaks down every reported result layer, from primary to exploratory, with the numbers that matter for prescribing decisions.

Primary Endpoint: MACE Composite

The primary outcome was time to first occurrence of a three-component composite: death from cardiovascular causes, non-fatal myocardial infarction, or non-fatal stroke. The trial used a non-inferiority design with a pre-specified hazard ratio margin of 1.5.

| Outcome | Testosterone (n = 2,643) | Placebo (n = 2,603) | Hazard Ratio (95% CI) | |---|---|---|---| | MACE composite | 182 events (7.0%) | 190 events (7.3%) | 0.96 (0.78-1.17) | | CV death | 48 (1.8%) | 46 (1.8%) | 1.04 (0.70-1.56) | | Non-fatal MI | 100 (3.8%) | 105 (4.0%) | 0.96 (0.73-1.26) | | Non-fatal stroke | 46 (1.7%) | 52 (2.0%) | 0.89 (0.60-1.33) |

The upper boundary of the 95% confidence interval (1.17) cleared the 1.5 margin with room to spare. The point estimate of 0.96 sits just below 1.0, meaning the trial found no signal of either harm or benefit on the composite. Each individual MACE component tracked close to unity as well, with none showing a statistically significant difference in either direction. These findings were reported in the New England Journal of Medicine in June 2023.

The Time-Course Pattern

Kaplan-Meier curves for the MACE composite ran essentially superimposed for the first 12 months, then separated by a trivially small margin favoring testosterone through months 24 to 48. At no point during follow-up did the testosterone arm cross above placebo in cumulative event rate.

HealthRX.com Time-Course Interpretation Framework: The clinical significance of this pattern depends on what question a prescriber is asking. For short-term safety (the first-year window that concerned earlier observational studies), TRAVERSE provides strong reassurance: the curves are flat and overlapping. For long-term cardiovascular protection, the trial was not powered to demonstrate superiority, and the confidence intervals remain wide enough that a modest benefit or harm cannot be excluded beyond 36 months.

The annualized MACE rate was approximately 2.4% per year in both arms. For context, this is consistent with a population carrying substantial baseline cardiovascular risk. Roughly 35% of participants had a prior MI or stroke at enrollment, and current AHA/ACC guidelines classify such patients as very high risk.

Secondary and Pre-Specified Exploratory Endpoints

Beyond the MACE composite, TRAVERSE reported several secondary cardiovascular endpoints.

| Secondary Endpoint | Testosterone Events | Placebo Events | HR (95% CI) | |---|---|---|---| | Expanded MACE (MACE + coronary revascularization) | 261 | 261 | 1.00 (0.84-1.19) | | All-cause mortality | 85 (3.2%) | 83 (3.2%) | 1.03 (0.76-1.40) | | Coronary revascularization | 103 | 97 | 1.07 (0.81-1.41) | | Hospitalization for heart failure | 43 | 46 | 0.94 (0.62-1.43) | | Atrial fibrillation | Not pre-specified but tracked | -- | No significant difference |

The expanded MACE (adding coronary revascularization as a fourth component) yielded a hazard ratio of exactly 1.00. All-cause mortality was nearly identical in both arms at 3.2%. Heart failure hospitalization showed a non-significant trend favoring testosterone (HR 0.94), though with wide confidence intervals that prevent meaningful interpretation.

One notable signal: pulmonary embolism occurred more frequently in the testosterone group (15 events versus 2 events). This aligns with the known effect of exogenous testosterone on erythropoiesis and hematocrit, which increases venous thromboembolism risk. The FDA label for testosterone products already carries a warning about polycythemia and thromboembolic events.

Subgroup Analyses: Who Drove the Result?

TRAVERSE pre-specified subgroup analyses across 14 baseline characteristics. The forest plot showed consistent non-inferiority across all subgroups, with no significant interaction p-values.

| Subgroup | n | HR (95% CI) | Interaction p | |---|---|---|---| | Age < 65 | 3,023 | 0.93 (0.71-1.22) | 0.68 | | Age ≥ 65 | 2,223 | 1.00 (0.75-1.34) | -- | | Prior CV event (yes) | 1,837 | 0.93 (0.72-1.20) | 0.64 | | Prior CV event (no) | 3,409 | 1.00 (0.74-1.36) | -- | | Diabetes (yes) | 3,527 | 0.91 (0.72-1.15) | 0.36 | | Diabetes (no) | 1,719 | 1.14 (0.78-1.68) | -- | | Baseline T < 200 ng/dL | ~1,200 | 0.87 (0.60-1.26) | 0.48 | | Baseline T 200-300 ng/dL | ~4,000 | 1.00 (0.80-1.26) | -- | | BMI ≥ 35 | ~1,800 | 1.05 (0.77-1.44) | 0.55 | | BMI < 35 | ~3,400 | 0.90 (0.70-1.15) | -- |

Two patterns stand out. First, men with diabetes (67% of the cohort) trended toward a slightly lower MACE rate on testosterone (HR 0.91), while men without diabetes trended in the opposite direction (HR 1.14). The interaction was not significant (p = 0.36), so this cannot be considered evidence of differential effect, but it raises a hypothesis worth tracking in future work. Second, men with the lowest baseline testosterone (<200 ng/dL) showed a numerically lower hazard ratio (0.87) than those in the 200-300 range, though again not statistically distinct.

Testosterone Levels Achieved

The dose-titration protocol aimed for a serum testosterone of 350 to 750 ng/dL. In practice:

Median serum testosterone at 6 months: ~358 ng/dL in the testosterone arm versus ~210 ng/dL in placebo
Mean serum testosterone at 12 months: ~398 ng/dL (testosterone) versus ~215 ng/dL (placebo)
The 25th-75th percentile range in the active arm was approximately 280 to 490 ng/dL

This is worth noting because the achieved levels were modest compared to what some TRT clinics target. Many real-world protocols aim for trough levels above 500 ng/dL. TRAVERSE's results apply to a protocol that kept most men in the low-normal range, and generalizing these safety findings to supraphysiologic dosing or higher-target protocols requires caution. The Endocrine Society's 2018 clinical practice guidelines recommend targeting mid-normal range, which aligns with what TRAVERSE achieved.

Limitations the Authors Acknowledged

The original publication and its supplement listed several important caveats:

Non-inferiority margin of 1.5. This is wide. A hazard ratio of 1.49 would still have met the primary endpoint. The choice reflects FDA guidance for the trial, but it means TRAVERSE cannot rule out a modest (up to 17%) increase in MACE risk based on the observed upper CI bound. Some cardiologists have argued that a 1.2 margin would have been more clinically meaningful.
Open-label dose titration. Although the trial was double-blinded, dose adjustments based on testosterone levels could have partially unblinded investigators who noticed hematocrit changes or clinical effects in the active arm.
Adherence decay. By 24 months, approximately 15% of participants in each arm had discontinued the study drug. This is lower than many long-term cardiovascular trials, but it dilutes the treatment contrast over time.
Population specificity. Every participant had either documented cardiovascular disease or multiple risk factors plus hypogonadism. Results may not apply to younger, healthier men starting TRT for symptoms alone.
Gel-only formulation. TRAVERSE used 1.62% transdermal testosterone gel exclusively. Injectable testosterone (cypionate or enanthate) produces different pharmacokinetic peaks and troughs, and the safety profile may differ. The FDA's 2015 labeling changes for testosterone products were applied across formulations, but TRAVERSE only tested one.

What Changed After TRAVERSE

Before this trial, the FDA had required a cardiovascular warning on testosterone products since 2015, based on conflicting observational data. The TTrials (Testosterone Trials) in 2016 were too small (n = 790) and too short (12 months) to resolve the safety question.

TRAVERSE was the first large, long-duration randomized trial designed specifically to test the cardiovascular safety hypothesis. Its results led the Endocrine Society and other professional bodies to update their position: TRT does not appear to increase MACE in appropriately selected hypogonadal men with elevated cardiovascular risk, at least when dosed to low-normal testosterone targets.

The practical shift is that providers now have a large RCT to cite when discussing TRT safety with patients who have cardiovascular comorbidities. Before TRAVERSE, these conversations relied on a mix of retrospective cohorts and small trials that pointed in different directions.

Putting the Numbers in Context

A hazard ratio of 0.96 with a confidence interval of 0.78 to 1.17 means the true effect could range from a 22% reduction to a 17% increase in MACE. That range is wide enough that testosterone cannot be called cardioprotective, but narrow enough, given the enriched-risk population, that a large safety signal is unlikely to be hiding in the data.

For a clinician managing a 62-year-old man with type 2 diabetes, a prior MI, and confirmed hypogonadism (total T of 220 ng/dL with symptoms), TRAVERSE provides the most relevant evidence available. The annualized MACE risk in this trial was about 2.4% per year regardless of treatment arm. Testosterone gel titrated to the low-normal range did not measurably change that risk over a mean 33-month follow-up.

Frequently asked questions

›

References

Lincoff AM, Bhasin S, Flevaris P, et al. Cardiovascular Safety of Testosterone-Replacement Therapy. N Engl J Med. 2023;389(2):107-117. PubMed
Snyder PJ, Bhasin S, Cunningham GR, et al. Effects of Testosterone Treatment in Older Men (TTrials). N Engl J Med. 2016;374(7):611-624. 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
Arnett DK, Blumenthal RS, Fonarow GC, et al. 2019 ACC/AHA Guideline on the Primary Prevention of Cardiovascular Disease. Circulation. 2019;140(11):e596-e646. PubMed
FDA Drug Safety Communication: Testosterone Products. FDA AccessData