When More Testosterone Made Bones Weaker

LONGEVITY LATEST · DEEP DIVE Issue 11 Companion · 13 May 2026

DEEP DIVE · COMPANION TO ISSUE 11

When More Testosterone Made Bones Weaker

The TRAVERSE Fracture Trial Broke an Endocrinology Textbook

The men in the testosterone arm broke 91 bones across 2,601 participants. The men in the placebo arm broke 64 across 2,603. A 42% relative increase, statistically significant, in the population every endocrinology textbook for forty years has been telling us should benefit most.

That’s the TRAVERSE Fracture sub-trial (Snyder et al., New England Journal of Medicine, January 2024). And the editorial NEJM ran alongside it was titled, with no exaggeration, "Breaking News."

The textbook we were working from

For decades, the case for testosterone and bone has gone like this. Testosterone, at adequate physiological levels, supports osteoblast activity and trabecular bone density. Hypogonadal men — men whose testosterone falls into the diagnostic range — have measurably lower bone density and higher fracture rates than eugonadal peers. Cross-sectional studies, observational cohorts, and small mechanistic trials have lined up behind this for years. Forty-something cardiologists at conferences have been comfortable repeating the line: replace what’s missing, the bone follows.

Cross-sectional and observational data, though, are exactly the kind of evidence that sets up the surprises. They tell you the snapshot. They cannot tell you what happens to a man’s skeleton when you put him on the hormone for three years.

What the trial actually did

TRAVERSE enrolled 5,246 men aged 45 to 80 who had biochemically confirmed hypogonadism plus high cardiovascular risk. They were randomised, double-blind, to topical testosterone or placebo. The primary endpoint was MACE — major adverse cardiac events — which read out as non-inferior in 2023.

The Fracture sub-trial took the same population and asked a different question: does three years of testosterone reduce clinical fractures? The fracture endpoint was pre-specified, adjudicated, and tracked across the full 5,204 participants who reached the analysis. Median follow-up was 3.19 years.

The result is the part that should slow you down. Testosterone did not lower fracture rate. It raised it. And the raised arm wasn’t a small numerical bump — it was 91 fractures versus 64. The hazard ratio favoured placebo, and the confidence interval did not cross unity. This is, in clinical-trial language, a finding the trial actually saw.

The fracture-relevant dimension of bone strength may have been quietly going the wrong way in the men whose surrogate measures were going the right way.

The authors are careful in the paper. They were not designed to test mechanism. They can only speculate. But Snyder and colleagues lean toward a cortical-bone hypothesis: an earlier mechanistic study had shown that testosterone treatment in severely hypogonadal men decreased cortical bone volume fraction and cortical thickness, even as trabecular bone improved. A caveat worth noting: this is the candidate mechanism, not the established one. The paper itself flags the speculation explicitly.

The thing this tells us about "optimisation"

Here’s where the finding lands in the wider hormone-optimisation conversation.

The men in TRAVERSE met a real clinical definition of hypogonadism. Symptoms plus repeated low morning bloods. They were the most defensible TRT population in any trial of this size, anywhere. And in that population, three years of treatment produced a fracture signal nobody saw coming.

Now widen the population. Move from "biochemically defined hypogonadism" to "men in the low-normal band who feel a bit tired and have read a Joe Rogan transcript." That’s the group the £600 pellet clinics from this week’s Hype Check are aimed at. The off-trial population. Lower baseline testosterone, smaller deficit, larger sample of men, and zero RCT data on what three years of replacement does to their bones.

If the indicated population — the group most likely to see net benefit — produced more fractures, the off-indicated population sits on a thinner branch. Not because the mechanism is necessarily worse for them. Because we have no trial data either way, and the closest thing we have is going the wrong direction.

What this doesn’t change

It is worth being precise about what the fracture finding does and doesn’t do.

It does not retract TRAVERSE’s primary cardiovascular result. Non-inferiority on MACE still stands. Replacement for symptomatic, biochemically defined hypogonadism remains supported by the largest RCT in the field — with the asterisks we now know about, but supported.

It does not say testosterone is bad for the skeleton in younger hypogonadal men, where short-term mechanistic data still favour treatment. The TRAVERSE participants were 45–80, and the signal may be specific to older bone biology. We don’t know, because the trial couldn’t tell us.

It does not undermine the broader case for hormone replacement in conditions where the indication is clean and the evidence is dense — most importantly, the transdermal estradiol picture in early-menopause women that Issue 11 covers. Those are different molecules, different routes, different populations, different trial bases. The fracture signal in TRAVERSE belongs to TRT in older men, full stop.

What it does do is collapse the "longevity optimisation" framing for testosterone. The trial that should have validated optimisation TRT — older men, three years on the hormone, hard endpoint — found the opposite. That’s what the £600-a-visit pellet model has not yet absorbed.

Where the field is heading

A few things to watch over the next 18 months.

The cortical-bone hypothesis needs a dedicated trial. A pre-specified imaging sub-study of testosterone’s effect on cortical thickness and porosity in older men, using high-resolution peripheral quantitative CT, would tell us whether the candidate mechanism is the actual one. There is appetite for it. There is no funded protocol yet.

The dose-response question is open. TRAVERSE used standard replacement doses targeting mid-normal range. The clinical bro-science world is using doses that push haematocrit into polycythaemic territory. Whether the fracture signal scales with dose is a question the optimisation community is going to need to answer the hard way unless someone runs the trial first.

The age-stratified question matters. If the signal is older-men-specific, the under-50 cohort is in different territory. If it’s universal, the optimisation thesis is in deeper trouble than it currently looks.

What this means for you

If you are a man being offered TRT outside a clean hypogonadism indication, the fracture data is the part of the conversation you want your prescribing clinician to address. Not "testosterone is dangerous" — that’s not what the trial says. The actual question is: in my population, in my dose range, over my expected treatment window, what does this evidence look like? If the answer is "we don’t know," that is a real answer and you should weigh it.

If you are on TRT for confirmed hypogonadism and the trial result is news to you, the right next step is a conversation with the prescriber about bone-density tracking. Baseline DEXA, repeat at three years. If the cortical-bone hypothesis turns out to be right, that’s the surface where it would show.

If I were designing a protocol from this evidence, it would be boringly simple: replacement-dose testosterone reserved for men who actually meet hypogonadism criteria; baseline DEXA; routine haematocrit and lipid monitoring; a serious sleep, body composition, and resistance-training audit before the script is written. Most of the men currently being offered "optimisation TRT" would walk out of that audit with a lifestyle programme and a follow-up blood test, not a pellet schedule.

The most expensive intervention in that protocol is the gym membership.

This is the Issue 11 Deep Dive. Longevity Latest runs one of these every week alongside the newsletter. Next week’s issue takes apart the Bryan Johnson stack — line by line, with grades.

Sources and further reading

1. Snyder PJ, et al. Testosterone Treatment and Fractures in Men with Hypogonadism. New England Journal of Medicine 2024;390(3):203–211. PMID: 38231621.

2. Lincoff AM, et al. Cardiovascular Safety of Testosterone-Replacement Therapy. New England Journal of Medicine 2023;389(2):107–117. PMID: 37326322.

3. Editorial. Breaking News — Testosterone Treatment and Fractures in Older Men. New England Journal of Medicine 2024;390(3):e6.

4. Snyder PJ, et al. Effect of Testosterone Treatment on Volumetric Bone Density and Strength in Older Men with Low Testosterone. JAMA Internal Medicine 2017;177(4):471–479. (Cortical-bone reduction in earlier mechanistic trial, referenced as candidate mechanism for the 2024 finding.)

5. Hackett G. Cardiovascular safety of testosterone replacement therapy (TRAVERSE trial). Trends in Urology & Men’s Health 2024;15(2):14–18.

6. Androgen Society. Making Sense of the TRAVERSE Trials. Position summary, 2024.

© 2026 FrontWave Media Ltd · Longevity Latest ·