I ran a unit conversion this week that I can’t stop thinking about. Finnish researchers took blood from 2,691 men at age 31, sorted them into four groups by testosterone, then checked on their health at 46. Converted to the units on an American lab report, the lowest group’s median lands at roughly 420 ng/dL. In 2022, the only online testosterone platform a secret shopper found with any treatment cutoff set it at 450. So what happened to the men sitting under that line? Most of them, it turns out, said they were doing well.
The men belong to the Northern Finland Birth Cohort 1966, which has followed people born that year in the country’s north since birth. Each gave a single fasting morning sample at 31, measured by liquid chromatography–mass spectrometry, and the team published the 15-year follow-up in Endocrine Connections. The bottom quarter sat below 17.2 nmol/L, with a median of 14.5. The top quarter sat at or above 25.5, with a median of 29.0.
Among the men who answered the health questionnaire at 46, 63.9% called their health excellent or good, 32.1% average and 4.0% poor or very poor, with no significant difference between the testosterone groups, according to News-Medical’s report on the paper. About 30% of the cohort had no self-rated health data, a gap the authors said was comparable across groups. On the 15D quality-of-life instrument, where 0.9 to 1.0 counts as excellent, every group averaged above 0.9, and the bottom and top quarters differed by about 0.01 points before adjustment; after adjustment the four-way comparison still wasn’t significant, though roughly 39% of men lacked usable 15D data. One signal did hold up in the fully adjusted model: the bottom-quarter men scored higher on the 15D’s depression dimension. They also had a lower prevalence of substance use disorders, and their lower sexual-activity scores stopped being significant once waist, smoking and alcohol entered the model.
| Measure | Share |
|---|---|
| rated their health excellent or good | 63.9% |
Low for Finland, normal on a lab report
That 14.5 nmol/L median works out to about 420 ng/dL, and the 17.2 cutoff to about 500. Now hold those up against a harmonized reference range built, as an Endocrine Society initiative, from healthy nonobese men aged 19 to 39 and calibrated to a CDC reference method. Its limits run from 264 to 916 ng/dL. The typical man in Finland’s “low” group sat comfortably inside it, which is how News-Medical’s write-up describes him too. The paper’s quartiles don’t say how many individual men dipped below 264, so some may have been clinically deficient. The median man was not.
Prescribing has been climbing anyway. Epic Research, drawing on records from 109 million men, found the share of male patients with a testosterone prescription bottomed at 0.64% in 2021 and rose to 0.93% in 2025 and 0.95% in the first half of 2026. By 2025 every age group from 30 up had passed its early-2010s peak, with men aged 50 to 64 leading at 1.57%. Epic’s analysts decline to assign a cause, noting only that telehealth and direct-to-consumer men’s health services have grown.
What those services looked like from the inside was documented in December 2022. In a secret-shopper study in JAMA Internal Medicine, Northwestern urologist Justin Dubin posed as a 34-year-old with low energy, low libido, normal total and free testosterone, and a wish to have children someday. Six of the seven platforms offered him testosterone. Only one set a starting threshold, and the other six had none. About half quoted a treatment target above 1,000 ng/dL, past the top of the healthy range. That one cutoff, 450 ng/dL, sits above the Finnish bottom-quarter median of roughly 420 and far above the 264 lower limit of normal.
The warning that came off the box
On February 28, 2025, the FDA recommended removing the language about increased cardiovascular risk from the boxed warning on every testosterone product. It kept the “Limitation of Use” for age-related hypogonadism and added blood-pressure warnings after ambulatory monitoring showed increases, and both deserve credit. The relabeling rested on TRAVERSE, a trial funded by AbbVie, Acerus Pharmaceuticals, Endo Pharmaceuticals and Upsher-Smith Laboratories, companies that sell the product under review.
TRAVERSE enrolled men aged 45 to 80 with heart disease or high cardiovascular risk, at least one symptom of hypogonadism and two fasting morning readings below 300 ng/dL. Heart attacks, strokes and cardiovascular deaths came out close, at 7.0% on testosterone and 7.3% on placebo. The testosterone arm had more atrial fibrillation (3.5% versus 2.4%), more acute kidney injury (2.3% versus 1.5%) and more pulmonary embolism (0.9% versus 0.5%), per Cleveland Clinic’s summary of the results. Cleveland Clinic cardiologist Steven Nissen, who chaired the trial, was blunt about how far it reaches: “This study should not be used as a justification for the widespread prescription of testosterone to aging men.” On the careful dose monitoring the trial used, he added: “It’s not likely to be happening in low T centers, where some men don’t even get a baseline testosterone measurement.”
| Event | Testosterone | Placebo |
|---|---|---|
| Atrial fibrillation | 3.5% | 2.4% |
| Acute kidney injury | 2.3% | 1.5% |
| Pulmonary embolism | 0.9% | 0.5% |
So a safety result earned in symptomatic men with repeatedly low readings now sits on the label of a drug that, in 2022, six of seven online platforms offered to a man with normal ones. Nothing in the February 2025 change addresses how those platforms decide who qualifies.
Wait, why would a tape measure erase a hormone signal?
Before adjustment, men in the two lower testosterone groups had higher odds of metabolic and gastrointestinal disease than the top group. Once the researchers accounted for waist circumference, those associations were no longer significant. I read that and genuinely stopped, because I had always pictured the hormone as the thing steering the metabolism, and here a belly measurement was soaking up the whole effect.
The biology makes it less strange. Most testosterone in your blood rides on a carrier protein called sex hormone-binding globulin, or SHBG. A review of testosterone in male obesity explains that moderate obesity lowers total testosterone largely through SHBG reductions driven by high insulin, so fewer carriers show up and the total reading drops even when the testes are working. Only more severe obesity starts suppressing the hypothalamic-pituitary signal itself. The authors call it a functional state that can reverse, though it takes substantial weight loss, and they describe a loop running both ways: low testosterone encourages fat to pile on, and the extra fat pushes testosterone down further.
If part of what Finland measured at 31 was the waistline talking, a prescription would change the number on the lab report without touching whatever pushed it down. The study can’t settle that. It measured testosterone once, relied on self-reported diagnoses, followed only Finnish men to 46, and never looked at treatment or diagnosed hypogonadism. Being observational, it cannot prove a low-normal level is harmless, and that depression signal deserves its own follow-up. The researchers still cautioned against unnecessary testosterone replacement, and the burden belongs on whoever sells a potentially long-term hormone to a man reading 420 to show he’ll be better off for it, with evidence from men like him rather than borrowed from a trial of men under 300.
If a man I love got a 420 from an app with a subscribe button underneath, I’d send him for a second morning test and hand him a tape measure before I let one number start a prescription.
Sources
- Endocrine Connections: Tuomisto, Pinola, Pesonen, Morin-Papunen, Niinimäki, testosterone at 31 and health at 46, Northern Finland Birth Cohort 1966 (2026)
- News-Medical: Men with lower testosterone at 31 reported good health 15 years later (2026)
- Cleveland Clinic ConsultQD: TRAVERSE results, adverse events and funding
- FDA: Class-wide labeling changes for testosterone products (February 28, 2025)
- EurekAlert / Northwestern Medicine: secret-shopper study of online testosterone platforms, JAMA Internal Medicine (2022)
- Epic Research: Testosterone prescribing rates rising again (2026)
- Journal of Clinical Endocrinology & Metabolism: Travison et al., harmonized reference ranges for testosterone in men (2017)
- Asian Journal of Andrology: Ng Tang Fui, Dupuis, Grossmann, lowered testosterone in male obesity (2014)