Did the TRAVERSE trial prove that testosterone causes fractures?

Not definitively. The fracture endpoint was secondary, and the confidence interval was wide (HR 1.43 to 95% CI 1.04-1.97). The finding is statistically significant but needs replication in a trial designed specifically to measure fractures as the primary outcome.

How can testosterone increase fractures if it improves bone density?

BMD measured by DXA reflects mainly trabecular bone quantity. Fracture resistance depends also on cortical thickness, bone microarchitecture, and fall mechanics. Testosterone may improve the surrogate (BMD) without improving the structural properties that prevent fractures under real-world loading conditions.

Does this mean all men on TRT should stop treatment?

No. TRAVERSE enrolled older men (mean age 63) with significant cardiovascular comorbidities. The fracture risk must be weighed against symptom relief and other outcomes. However, fracture risk should now be part of every TRT informed consent discussion.

Were the fractures caused by more falls or weaker bones?

TRAVERSE did not systematically collect falls data, so this question remains unanswered. Both mechanisms are plausible. Testosterone can increase hematocrit and body weight, potentially affecting balance and impact force.

Should men on TRT get a DXA scan?

Baseline DXA is reasonable for men starting TRT who are over 50 or who have other fracture risk factors. The TRAVERSE data add urgency to this recommendation, though no guideline has formally mandated it yet.

Did the Endocrine Society change its TRT guidelines after TRAVERSE?

Not yet. The current Endocrine Society guideline dates to 2018 and predates TRAVERSE. An update is anticipated but has not been scheduled. The AUA issued a 2024 panel update acknowledging the bone findings.

Does this fracture signal apply to testosterone injections, not just gel?

TRAVERSE used transdermal testosterone gel. No large fracture-endpoint trial has been conducted with injectable testosterone. The mechanism, if systemic (related to circulating testosterone levels rather than topical effects), would plausibly apply across formulations, but this has not been tested.

What was the absolute fracture risk increase?

Approximately 1 additional fracture per 100 men treated over 3.2 years (3.50% in the testosterone group vs. 2.46% in placebo). The relative increase was 43%, but the absolute numbers are modest.

Are younger hypogonadal men at the same fracture risk from TRT?

Unknown. TRAVERSE enrolled men 45 to 80 with high comorbidity burden. Younger, healthier men with hypogonadism were not represented. Extrapolating the fracture signal to men under 45 without cardiovascular disease is not supported by these data.

Has the FDA updated testosterone labels to include fracture risk?

As of mid-2025, FDA testosterone labeling includes cardiovascular warnings from the parent TRAVERSE trial but does not yet include specific fracture warnings from the bone substudy. This may change as the agency reviews the secondary endpoint data.

What TRAVERSE Bone Fracture Substudy Actually Changes in Clinical Practice

By HealthRX Medical Team

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

At a glance

| Parameter | Detail | |---|---| | N | 5,204 (fracture analysis population from 5,246 randomized) | | Intervention | 1.62% testosterone gel, dose-titrated to maintain serum T 350-750 ng/dL | | Comparator | Matching placebo gel | | Duration | Median 3.19 years | | Primary endpoint | First clinical fracture (confirmed by imaging or medical records) | | Key result | HR 1.43 (95% CI 1.04-1.97); 91 fractures in testosterone arm vs 64 in placebo |

Why this result caught everyone off guard

For over two decades, TRT advocates pointed to small, short studies showing that testosterone increases bone mineral density (BMD) at the spine and hip. The Testosterone Trials (TTrials), published in JAMA Internal Medicine in 2017, showed a measurable BMD gain at 12 months in older men with low testosterone. That finding was real. BMD did go up.

But BMD is a surrogate. Fractures are what matter. And TRAVERSE is the first trial powered to count actual fractures in men on TRT. The disconnect between BMD improvement and fracture increase is the central clinical problem this substudy creates.

The parent TRAVERSE trial enrolled 5,246 men aged 45 to 80 with hypogonadism (two fasting testosterone levels <300 ng/dL) and either preexisting cardiovascular disease or elevated CV risk. The primary outcome of the parent trial was major adverse cardiovascular events (MACE). The bone fracture substudy was prespecified as a secondary analysis.

Methodology details that matter for interpretation

Several design elements separate this from prior testosterone-bone studies and determine how much weight the findings carry.

Population selection. Every participant had documented hypogonadism plus cardiovascular risk factors. Mean age was 63. Mean BMI was 33. Over 70% had diabetes or prediabetes. This is not a general TRT population; it is a high-comorbidity cohort. That distinction matters when extrapolating.

Fracture ascertainment. Fractures were identified through participant self-report at scheduled visits, then confirmed by central adjudication reviewing imaging and medical records. Pathologic fractures and high-trauma fractures were excluded. The adjudication process caught fractures that might have been missed in less rigorous designs, but also means the true rate of vertebral fractures (often silent) was likely undercounted in both arms.

Dose titration. Testosterone gel was adjusted to keep serum levels between 350 and 750 ng/dL, checked at months 1, 3, 6, and every 6 months after. Compliance was high. This means the fracture signal cannot be dismissed as a supraphysiologic dosing artifact.

No DXA requirement at baseline. Participants were not selected based on bone density. The trial did not require osteoporosis or osteopenia for enrollment. This is critical: the fracture increase occurred in a population not specifically at skeletal risk, making it harder to confine the signal to a narrow subgroup.

The HealthRX Practice-Impact Framework

We score every major trial on three axes to determine how much it should actually change what clinicians do on Monday morning.

| Axis | Score | Rationale | |---|---|---| | Signal strength | Moderate | HR 1.43 with CI just clearing 1.0 (1.04-1.97). Statistically significant, but the lower bound sits close to null. A larger trial or longer follow-up could narrow or widen this. | | Applicability to real-world TRT patients | Moderate-High | Mean age 63, obese, high CV burden. Overlaps heavily with the typical TRT-prescribing demographic in primary care and men's health clinics. Less applicable to younger, leaner hypogonadal men. | | Guideline responsiveness | Early but moving | The Endocrine Society has not yet formally revised its 2018 guideline, but the AUA's 2024 panel discussion cited TRAVERSE bone data as requiring updated risk-benefit counseling. FDA label changes remain pending. |

Overall practice-change grade: Act now on counseling. Wait for guideline revision before changing prescribing algorithms.

Fracture results in detail

| Outcome | Testosterone | Placebo | Hazard Ratio (95% CI) | |---|---|---|---| | All clinical fractures | 91/2,601 (3.50%) | 64/2,603 (2.46%) | 1.43 (1.04-1.97) | | Upper extremity | 28 | 16 | 1.75 (0.95-3.24) | | Lower extremity | 22 | 15 | 1.47 (0.76-2.82) | | Vertebral (clinical) | 18 | 13 | 1.38 (0.67-2.85) | | Other sites | 23 | 20 | 1.15 (0.63-2.10) |

The overall signal is driven by a consistent trend across skeletal sites rather than a spike at one location. No single site reached statistical significance alone, which is expected given event counts. The consistency across sites argues against the fracture increase being an artifact of one outlier fracture type.

Falls data were not systematically collected in TRAVERSE. This is the study's most important missing variable. Testosterone can increase hematocrit, alter blood pressure, and affect balance through changes in body composition. Whether the fracture increase reflects more falls, weaker bone microarchitecture despite higher BMD, or some combination remains unanswered.

What changed between the TTrials BMD data and TRAVERSE fractures

The apparent contradiction, BMD goes up but fractures also go up, has several candidate explanations. None are proven.

Cortical vs. trabecular effects. Testosterone may preferentially increase trabecular BMD (measured well by DXA) while failing to strengthen cortical bone, which bears most of the mechanical load in long bones. The TTrials bone substudy used both DXA and quantitative CT, but peripheral cortical endpoints were not powered for fracture prediction.

Body composition shifts increasing fall force. Testosterone increases lean mass and can increase body weight. Greater mass means greater impact force during a fall. If fall frequency stays constant or rises slightly (perhaps due to polycythemia-related dizziness or fluid shifts), the mechanical load at impact increases.

Bone turnover suppression. Testosterone suppresses bone resorption markers. Reduced remodeling can allow microdamage to accumulate, a mechanism proposed for bisphosphonate-associated atypical fractures. Whether TRT produces enough turnover suppression to trigger this pathway over 3 years is speculative but biologically plausible.

Which guidelines have responded

Endocrine Society (2018 guideline, last updated). The current guideline predates TRAVERSE entirely. It recommends testosterone for men with "unequivocally low serum testosterone" and symptoms, noting bone density improvement as a secondary benefit. That bone-benefit language will need revision. An update is expected but no timeline has been announced.

AUA (2018 guideline, 2024 panel update). The American Urological Association's 2024 panel review acknowledged TRAVERSE cardiovascular and bone findings, recommending that clinicians "discuss the potential for fracture risk" during informed consent for TRT. This stops short of a contraindication but represents a meaningful shift in counseling language.

FDA. The testosterone product labeling was updated in 2023 to add cardiovascular warnings based on the parent TRAVERSE trial. Bone fracture warnings have not yet been added. Given the secondary-endpoint status of the fracture data, an FDA label change may require additional confirmatory evidence.

What should change in practice today

Stop listing bone protection as a TRT benefit. Before TRAVERSE, it was common for prescribers and direct-to-consumer TRT clinics to include "protects bone density" in the list of testosterone benefits. That claim is no longer supportable without qualification. BMD may improve, but the hard endpoint, fractures, moved in the wrong direction.

Add fracture risk to informed consent. Any man starting TRT, particularly those over 55 with cardiovascular risk factors, should be told about the TRAVERSE fracture signal. The absolute risk difference is modest (approximately 1 extra fracture per 100 men treated over 3 years), but informed consent requires disclosure.

Consider baseline DXA in men starting TRT. If a man already has osteopenia or osteoporosis, the risk-benefit calculation shifts. Testosterone should not be used as a bone-protective strategy in these patients. Standard osteoporosis therapies (bisphosphonates, denosumab) remain the evidence-based options for men with documented low BMD.

Do not extrapolate to younger men without comorbidities. The TRAVERSE population was older, obese, and cardiovascularly compromised. A 35-year-old with hypogonadism from a pituitary adenoma does not share this risk profile. The fracture signal may or may not apply outside the trial population, but there is no evidence to support applying it broadly to all TRT patients.

Monitor for falls. Until the mechanism is clarified, clinicians should ask about falls, dizziness, and balance changes at TRT follow-up visits. Hematocrit monitoring (already standard) takes on added importance if polycythemia contributes to fall risk.

Limitations the authors acknowledged

The investigators noted several important caveats in the published analysis. The fracture endpoint was secondary, not primary. The trial was not designed or powered specifically for fracture detection. Falls were not prospectively tracked. Morphometric vertebral fractures (asymptomatic compression fractures visible only on imaging) were not captured because routine spine imaging was not performed. Baseline DXA was not required, so the bone-density distribution of participants is incompletely characterized.

The confidence interval for the hazard ratio (1.04-1.97) is wide. The lower bound barely excludes 1.0. A single misadjudicated fracture could theoretically shift the result to nonsignificance. This does not invalidate the finding, but it means replication is essential before the signal can be treated as definitive.

What comes next

The TRAVERSE investigators have called for a dedicated fracture-endpoint trial in hypogonadal men. No such trial is currently registered on ClinicalTrials.gov. Until one is completed, the TRAVERSE bone data will remain the best available evidence on testosterone and hard skeletal endpoints.

Ongoing analysis of TRAVERSE subgroups (by age, baseline testosterone level, BMI, and diabetes status) may clarify whether the fracture signal concentrates in identifiable patient subsets. These secondary analyses are expected over the next 12 to 18 months.

Frequently asked questions

›

References

Snyder PJ, Bauer DC, Engel SS, et al. Testosterone treatment and fractures in men with hypogonadism. N Engl J Med. 2024;390(3):203-211. PubMed
Snyder PJ, Kopperdahl DL, Stephens-Shields AJ, et al. Effect of testosterone treatment on volumetric bone density and strength in older men with low testosterone. JAMA Intern Med. 2017;177(4):471-479. 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
Mulhall JP, Trost LW, Brannigan RE, et al. Evaluation and management of testosterone deficiency: AUA guideline (2024 panel update). J Urol. 2024;211(1):15-21. PubMed
FDA. AndroGel (testosterone gel) prescribing information. Revised 2023. FDA Label