In June 2025, the FDA quietly did something it had required since 2015: it told manufacturers of prescription testosterone products they could remove the boxed warning about heart attack and stroke risk from the label. For a drug class that had spent a decade carrying pharmaceutical medicine's most serious safety flag, that's a big deal, and testosterone clinics wasted no time turning it into marketing copy. "FDA confirms testosterone therapy is heart-safe," several low-T websites now claim. That's not quite what happened, and the trial the FDA leaned on to make the change contains findings those same websites tend to leave out of the pitch.
The Trial That Changed the Label
The decision traces back to TRAVERSE — Testosterone Replacement therapy for Assessment of long-term Vascular Events and efficacy ResponsE in hypogonadal men — a randomized, double-blind, placebo-controlled trial published in the New England Journal of Medicine in June 2023. It enrolled 5,246 men aged 45 to 80, all with confirmed low testosterone (under 300 ng/dL) and either existing cardiovascular disease or a high risk profile for developing it. Half received daily transdermal testosterone gel at 1.62%, half received placebo gel, and the trial ran for a mean treatment duration of 21.7 months with about 33 months of total follow-up. Cleveland Clinic coordinated it, and a consortium of testosterone manufacturers funded it — a detail worth knowing given the FDA had specifically mandated this study back in 2015 after earlier retrospective research raised cardiovascular red flags.
The headline result: testosterone was noninferior to placebo on the primary composite endpoint of cardiovascular death, nonfatal heart attack, and nonfatal stroke. A primary event occurred in 182 men on testosterone (7.0%) versus 190 on placebo (7.3%) — a hazard ratio of 0.96, with a 95% confidence interval of 0.78 to 1.17. Study chair Dr. Steven Nissen and his co-investigators framed this as reassurance, and on its own terms, it is. But "noninferior" is a specific statistical claim, not a synonym for "safe," and the difference matters more than the marketing implies.
What "Noninferior" Actually Promises You
A noninferiority trial isn't designed to prove a drug helps — it's designed to prove it doesn't hurt by more than a predefined margin. TRAVERSE set that margin in advance and testosterone came in under it on the primary composite endpoint. That's a real and useful finding for the specific population studied. What it doesn't tell you is anything about men without cardiovascular risk factors, men younger than 45, men on injectable rather than transdermal testosterone, or anyone using testosterone for years rather than the roughly 22-month average treatment window in the trial. The FDA's label change reflects exactly that scope — the boxed warning came off because the specific fear (heart attack and stroke in this population, over this duration) wasn't borne out. It did not come off because testosterone was declared broadly risk-free.
The Findings That Didn't Make the Press Release
Here's the part that gets skipped in the "TRT is finally vindicated" summaries: TRAVERSE and its prespecified substudies turned up three safety signals that had nothing to do with the primary endpoint and everything to do with why a cardiologist might still hesitate.
Atrial fibrillation showed up more often in the testosterone group — 91 cases versus 63 on placebo, a difference reaching statistical significance (p=0.02). Nonfatal arrhythmias overall were also elevated (p=0.001). This one caught researchers somewhat off guard, since low testosterone itself has been linked in prior observational work to higher AFib risk, and some earlier meta-analyses had suggested testosterone therapy might reduce AFib incidence after five or more years of use. TRAVERSE ran shorter than that and found the opposite direction. Acute kidney injury followed the same pattern: 60 cases on testosterone versus 40 on placebo (p=0.04), another signal that hadn't shown up clearly in prior randomized trials of testosterone therapy.
Then, in January 2024, a prespecified TRAVERSE bone-safety substudy published in NEJM delivered the most counterintuitive result of the whole program. Testosterone therapy is known to improve bone mineral density in hypogonadal men — that part isn't in dispute, and it's one of the reasons clinicians have long assumed TRT would lower fracture risk in this population. The substudy found the reverse: men on testosterone had a higher risk of clinical fractures (hazard ratio 1.43, 95% CI 1.04–1.97, p=0.03) and of all fractures combined (hazard ratio 1.52, 95% CI 1.13–2.05, p=0.006) compared with men on placebo.
Nissen, who chaired the trial, called the consistency of that fracture signal "biologically important" despite it running against expectation. One proposed explanation, floated in an accompanying JAMA editorial, is behavioral rather than skeletal: men who feel stronger and more energetic on testosterone may simply do more — more activity, more risk of a fall — and the fractures observed were largely fall-related rather than the kind of spontaneous fragility fracture you'd expect from bone density loss. That's a plausible read. It's also unproven, and until someone tests it directly, "testosterone makes your bones better but somehow also makes you more likely to break one" sits uncomfortably in the same sentence.
Who TRAVERSE Actually Studied — and Who It Didn't
This is the detail that gets lost fastest in gym-forum and low-T-clinic summaries of the trial. TRAVERSE enrolled men who were, on average, older and sicker than the typical guy walking into a testosterone clinic advertising "optimize your levels." Every participant had either diagnosed cardiovascular disease or a risk profile serious enough to qualify as high-risk — this was not a study of healthy 35-year-olds chasing better gym numbers or energy levels. Nissen made this point explicitly in comments accompanying the results: "This study should not be used as a justification for the widespread prescription of testosterone to aging men." If you're a 38-year-old with normal cardiac risk factors considering TRT for low energy, TRAVERSE's cardiovascular reassurance doesn't automatically transfer to you, because you're not who it studied — and neither does its fracture or AFib signal, for that matter, since your baseline risk for both is lower to begin with.
The FDA's 2025 label change followed a European Expert Panel position statement that reviewed TRAVERSE alongside the broader evidence base and concluded the trial represented a genuine advance in understanding testosterone's cardiovascular profile — while also flagging that hematocrit elevation, arrhythmia, and fracture risk deserve continued monitoring rather than a shrug. Regulatory agencies move slowly and conservatively on purpose, and a boxed warning removal after a single large trial — however well-designed — is not the same evidentiary bar as, say, a drug getting approved in the first place.
What This Actually Means If You're Considering TRT
Skip the testosterone clinic that advertises itself with "optimize" and "biohack" language and can prescribe after a single blood draw and a five-minute video call — that model existed before TRAVERSE and TRAVERSE gives you no new reason to trust it. Go through an endocrinologist or a physician who will run two separate morning testosterone measurements before diagnosing hypogonadism, per the Endocrine Society's own diagnostic criteria, rather than treating one low number on a random-time blood draw as sufficient.
Bring three specific questions to that conversation instead of a general "is TRT safe" ask, because the general question gets you a general answer and TRAVERSE doesn't support one. Personal or family history of atrial fibrillation or other arrhythmia belongs first on the list, since that's the signal TRAVERSE flagged most clearly and it's exactly the kind of thing a baseline EKG can catch before you start. Fall risk and existing bone density matter just as much, particularly if you're over 50 or have any history of osteoporosis or prior fracture — Nissen's own comment on this was blunt: men with a history of osteoporosis probably shouldn't be starting testosterone at all. And the third question is about follow-through rather than diagnosis: what monitoring schedule does your prescriber actually plan to run — hematocrit, kidney function, periodic cardiac symptom check-ins — versus a one-time prescription with a refill button and no follow-up?
None of this makes testosterone therapy the wrong choice for a man with genuinely diagnosed hypogonadism and a doctor tracking him properly. It does mean the 2025 label change is a narrower, more conditional piece of good news than the marketing suggests — reassurance about one specific risk, in one specific population, over one specific treatment window, alongside three other risks that trial made harder to ignore, not easier.