Who Needs Testosterone Replacement Therapy?

Forty five men walked into a clinic in Sydney tired, quiet in the bedroom, flat in mood, and none of them had a diagnosable disease. No testicular failure, no pituitary damage, just middle aged men whose testosterone sat on the low side of normal (00:00). Researchers gave each man a testosterone gel for six weeks and a look-alike placebo gel for six weeks, in random order, with nobody, not even the researchers, knowing which was which. The gel worked exactly as advertised: it pushed levels from about 310 nanograms per deciliter up to about 510, solidly into the range of a healthy young man (39:22, 39:42). And the men felt exactly the same on both.

That result sits at the center of Barbell Medicine physicians Jordan Feigenbaum and Austin Baraki's four part series on testosterone, based on their book Signal. Energy, libido, and mood showed no real difference between drug and placebo (00:43). The trial ran twenty two separate quality of life measures. Exactly one favored testosterone, which the study's authors chalked up to chance rather than a true effect (40:02). The design had more than 90 percent power to catch even a moderate benefit, so this was not a study too small to find something real (40:24). It just did not find it.

"If you take a guy with low normal testosterone levels and you raise them, his symptoms don't really do anything on testosterone that the placebo also doesn't do." (Jordan Feigenbaum, [41:30])

That single sentence explains a strange paradox running through the whole episode: testosterone therapy in the United States is somehow both overused and underused at the same time. On the overuse side, prescriptions have roughly quadrupled over the past twenty years (10:15). A quarter of men who start therapy never had a testosterone level drawn beforehand (10:25). The landmark HIM study of primary care patients found that nearly 40 percent of middle aged men had a total testosterone below 300 ng/dL, but once symptoms were required alongside the low number, the true prevalence of deficiency dropped to somewhere between 2 and 6 percent (10:57, 11:15). Even among men who did get tested, about one in five received a prescription without meeting the actual criteria.

Part of the confusion is a protein called SHBG, sex hormone binding globulin, which binds testosterone in the blood and makes it biologically inactive. Obesity and insulin resistance tend to push SHBG down, which can make a man's total testosterone number look low while the free, usable fraction is perfectly normal (12:05, 12:27). Baraki describes a common patient: mid-40s, BMI above 30, waist over 40 inches, total testosterone of 240, tired, with a normal free testosterone once it is actually checked (12:48). His advice is not to reach for a prescription first, but to negotiate a plan: address sleep apnea, alcohol, and weight before or alongside any hormone therapy, sometimes with the help of a GLP-1 medication (15:10, 16:24).

On the other side of the ledger, real deficiency gets missed constantly. In the same HIM study, only one in ten men with a genuinely low testosterone level was actually being treated (23:06). Chronic opioid users, who suppress their own testosterone production at high rates, are rarely screened at all (23:21). Baraki recalls working with HIV patients whose testosterone came back at literally zero, a level he says he almost never saw checked in that population before (28:34).

A Decade of Fear, Explained by Money and Lawsuits

Part of the reason doctors got skittish traces back to 2013, when a couple of observational studies suggested a heart risk signal and the FDA issued a warning that testosterone might raise the chance of heart attack and stroke (24:17, 24:39). Prescription rates fell by more than half between 2013 and 2016 (24:39). At the same time, more than 25,000 lawsuits piled up against makers of AndroGel, a topical testosterone product, alleging it caused heart attacks and strokes (25:18). AbbVie, the company behind AndroGel, had spent about 75 million dollars marketing the drug in 2012 alone, part of a testosterone market that peaked around 1.6 billion dollars in 2014 before the warnings and lawsuits hit (25:54, 26:05). The FDA had flagged AbbVie's marketing as misleading as far back as 2000, years before the cardiovascular scare (26:05, 26:17).

What does correct treatment actually look like, for the man who truly needs it? Current guidelines aim for a testosterone level in the middle of the normal range, roughly 450 to 600 ng/dL, measured halfway between doses (46:00). Feigenbaum draws a sharp line here.

"A guy sitting at 550 in the middle of his cycle is being treated correctly. But a guy who's sitting at 1,200 isn't being really treated, he's being juiced." (Jordan Feigenbaum, [46:04])

Yet a real world study of more than 9,000 men on testosterone therapy found their average level was around 800 ng/dL, well above the guideline midpoint (47:12). A separate VA analysis found only about 3 percent of men got the full recommended workup before starting treatment, and over half never had follow up labs in the following year (47:23, 47:34). Put those two facts together and a striking number makes more sense: about 70 percent of men who start testosterone therapy quit within a year, usually not because of side effects but because they never felt the improvement they were promised (45:05, 53:16).

Fertility is the risk that gets the least attention and probably deserves the most. Adding testosterone from outside the body tells the brain to stop signaling the testes, which shuts down both testosterone production and sperm production at the same time (57:43, 58:09). In a survey, a quarter of urologists incorrectly believed that testosterone therapy would improve fertility, when in practice it is one of the more common medically caused reasons for male infertility (58:36, 59:02). Recovery after stopping can take six to twelve months, and for some men it never fully returns (59:13). Men who want children can sometimes use HCG or clomiphene instead of, or alongside, testosterone to preserve fertility, a conversation Baraki says needs to happen before the first dose, not after (60:09).

What the Heart and Prostate Data Actually Show

The Traverse trial, published in 2023 in the New England Journal of Medicine, is the best cardiovascular safety data available. Researchers randomized more than 5,000 men, ages 45 to 80, all with testosterone deficiency and either established heart disease or high cardiovascular risk, to topical testosterone or placebo and followed them for two years (62:12, 62:55). Cardiac events landed at about 7 percent in both groups, no meaningful difference (63:06). The catch: the treatment group only reached an average of 375 ng/dL, below the guideline midpoint, using a gel rather than the injections many men actually take, and the trial ran only two years (63:55, 64:00). It is a real reassurance for gently dosed patients on a topical product, not a green light for a 35 year old injecting himself to 900 at a wellness clinic.

The prostate fear traces back to a 1941 study showing castration shrank prostate tumors and testosterone made them grow. The finding was real; the conclusion drawn from it was too simple. The saturation model offers a fix: prostate tissue's androgen receptors seem to saturate at a certain testosterone level, after which PSA, prostate specific antigen, a blood marker used to screen for and monitor prostate cancer, stops climbing even as testosterone keeps rising (67:52, 68:21). A 2026 retrospective study of more than 900 men in active surveillance for prostate cancer found that low baseline testosterone, not high, was linked to a 61 percent higher chance of the cancer progressing to an aggressive, high grade form requiring treatment (71:18, 71:38). And a study of 850 men who had their prostate surgically removed found that those who later went on testosterone therapy had a lower recurrence rate than those who did not (72:12, 72:37).

The one safety issue that is not overblown involves the blood itself. Testosterone tells the kidneys to produce more erythropoietin, the hormone that drives red blood cell production, and up to 20 percent of men on injectable testosterone can develop erythrocytosis, blood that is too thick (74:09, 75:04). A hematocrit above 54 percent is the common threshold doctors use to investigate a dose reduction or look for an underlying cause like untreated sleep apnea (75:24).

The series closes on Mark, a 45 year old architecture firm partner whose focus, energy, and marriage had all been sliding for a year. A wellness clinic drew his blood one afternoon, the wrong time of day for an accurate reading, found a testosterone level of 240 ng/dL, and started him on weekly injections (84:03). By month three he felt worse, with morning headaches and rising blood pressure (84:03). His real problem was severe obstructive sleep apnea, his airway collapsing roughly 60 times an hour, all night, compounded by visceral fat (84:33). A CPAP machine, six weeks of real sleep, a modest training routine, and better food brought his weight and waist down. He eventually came off testosterone entirely, and his own production returned (85:23, 86:01).

"Testosterone is an output. It's a readout of what the rest of your body's doing. And if we can fix the inputs, the output is gonna look a whole lot better." (Jordan Feigenbaum, [83:15])

That is the whole argument in one line. A lab number is a symptom of a system, not a target by itself. For a small number of men, the system really is broken, and replacement is close to a life changing intervention. For a much larger number, the number on the page is just an invitation to look somewhere else first, at sleep, at body fat, at alcohol, before reaching for a syringe.