Evidence Explainer · Bone health

Testosterone did not prevent fractures in TRAVERSE

Clinical fractures occurred in 3.50% assigned testosterone and 2.46% assigned placebo over median 3.19 years. An earlier BMD improvement did not establish fracture prevention—and the later clinical outcome moved in the opposite direction.

Reviewed under standing publication authorization

Discovery, counterevidence, claim mapping, identifiers, funding, correction state and prose were reviewed before release.

Evidence at a glance

Clinical fractures—not a bone-density surrogate

TRAVERSE fracture substudy
Study type
Randomized, double-blind clinical-fracture substudy
Studied in
Humans
Participants / sample
5,204 TRAVERSE participants; middle-aged and older men with symptoms, repeated low testosterone and preexisting or high cardiovascular risk
Publication status
Peer reviewed; no linked correction, retraction or expression of concern identified as of verification
Outcome type
Adjudicated clinical fractures
Evidence maturity
Large randomized clinical-outcome evidence over median 3.19 years
Conflicts / funding
Parent TRAVERSE program funded by AbbVie and other testosterone manufacturers
Our assessment
Testosterone did not reduce fractures; the higher observed fracture incidence prevents a fracture-benefit claim

Read this first

Three takeaways

  • Clinical fracture occurred in 91 of 2,601 participants (3.50%) assigned testosterone and 64 of 2,603 (2.46%) assigned placebo; hazard ratio 1.43, 95% CI 1.04–1.97.
  • The result does not show fracture prevention. It also does not, by itself, identify why fractures were higher.
  • A separate 211-person one-year substudy found improved volumetric BMD and estimated strength. Those surrogate measurements cannot replace the later clinical-fracture outcome.

Keep outcomes separate

A surrogate and a patient outcome can point in different directions

Bone density

The earlier substudy measured volumetric BMD and estimated strength after one year.

Clinical fractures

The later, larger substudy measured adjudicated fractures over median 3.19 years.

Mechanism

The trial did not establish a reason for the higher fracture incidence.

Longevity

Neither result demonstrates slower aging, disability prevention or longer life.

01

The bottom line

Testosterone treatment did not reduce clinical fractures in TRAVERSE. Fractures were observed more often in the testosterone group, so the trial cannot support a fracture-prevention claim.

02

What researchers did

The fracture substudy followed 5,204 TRAVERSE participants assigned daily testosterone or placebo gel. At visits, reported fractures were documented from medical records and adjudicated; the primary analysis measured time to first clinical fracture.

03

What they found

Over median 3.19 years, clinical fracture occurred in 3.50% assigned testosterone and 2.46% assigned placebo; hazard ratio 1.43, 95% CI 1.04–1.97. Other fracture endpoints also appeared numerically higher with testosterone.

04

How strong is the evidence?

This is a large randomized clinical-outcome analysis. It directly answers the fracture question more strongly than a smaller, shorter trial measuring only bone density or estimated strength.

05

What this does not show

The analysis does not establish why fractures were higher, identify a causal pathway, or show the same result for every formulation and population. It does not erase appropriate treatment for diagnosed conditions or provide an individual risk estimate.

06

Safety, surrogates and conflicts

A 211-person Testosterone Trials substudy reported a 6.8-percentage-point treatment effect for spine trabecular volumetric BMD after one year. That biomarker did not establish fracture prevention. TRAVERSE was industry funded; the earlier bone substudy was principally publicly supported with manufacturer assistance disclosed.

07

What happens next

Analyses of fracture type, timing, trauma context, falls and longer follow-up may help explain the unexpected pattern. Until then, the mechanism is uncertain and the clinical result should not be softened by the earlier surrogate.

08

Educational boundary

This article explains trial outcomes. It is not personalized medical advice and does not diagnose osteoporosis or hypogonadism, calculate fracture risk, recommend a product or tell a person to start, stop or change testosterone.

Primary sources

Sources, roles and limits

  1. TRAVERSE clinical-fracture substudy
    Identifier
    PMID:38231621 · DOI:10.1056/NEJMoa2308836 · NCT03518034
    Role
    Primary randomized clinical-fracture support
    Limitation
    Unexpected outcome in the selected TRAVERSE population; does not establish mechanism.
  2. Testosterone Trials bone-density substudy
    Identifier
    PMID:28241231 · DOI:10.1001/jamainternmed.2016.9539 · NCT00799617
    Role
    Earlier surrogate contrast
    Limitation
    211 participants, one year, volumetric BMD and estimated strength—not fractures.
  3. TRAVERSE registry record
    Identifier
    NCT03518034
    Role
    Parent-trial identity, population, sponsor and results context
    Limitation
    Registry inclusion or completion is not proof of safety, benefit or scientific validity.
  4. Testosterone Trials registry record
    Identifier
    NCT00799617
    Role
    Bone-substudy parent identity and posted-results context
    Limitation
    The coordinated trials were too short and small for broad fracture or long-term safety conclusions.

Connected topics

Continue through the knowledge system

Health & functionDrugs & supplementsIntervention recordsStudy and trial records

Continue with real coverage

Keep reading

These are learning paths, not rankings or recommendations.

A treatment trial that measured fractures directly

Research AnalysisHealth & function

Zoledronic acid reduced fractures in osteoporosis—not aging itself

HORIZON reduced vertebral and hip fractures in postmenopausal osteoporosis, with outcome-specific duration and safety limits—not evidence of slower aging.

Evidence scopeHuman fracture outcomes · large randomized trial

Another bone-health study that keeps density separate from fractures

Evidence ExplainerEveryday foundations

LIFTMOR improved bone density in a small supervised trial—not fracture prevention

LIFTMOR found BMD and function gains in a small, screened, closely supervised trial—not fracture prevention or proof that unsupervised heavy training is safe.

Evidence scopeHuman BMD and function · small supervised randomized trial

Another randomized prevention trial with a fracture outcome

Claim CheckDrugs & supplements

Vitamin D is essential. VITAL still found no broad prevention benefit

In generally healthy adults, vitamin D3 did not significantly reduce invasive cancer, major cardiovascular events, or fractures; deficiency treatment is a different question.

Evidence scopeHuman clinical outcomes · large randomized prevention trial

Disclosures and history

  • August 3, 2026 — source discovery and independent verification completed before prose drafting.
  • August 3, 2026 — atomic population, formulation, outcome, safety, funding and longevity boundaries reviewed.
  • Corrections: no linked correction, retraction or expression of concern identified for the featured primary papers as of verification; regulator and registry records must be rechecked when this story is updated.