FAQ  /  Men's Hormone Health

Does Testosterone Therapy Cause Prostate Cancer?

What the randomized trials show

---

This is the evidence that speaks most directly to the question, because it compares men who were given testosterone against men who were not, with the assignment made by chance rather than by a physician's judgment.

The largest pooled analysis to date, published in 2026, brought together 41 randomized controlled trials and 11,161 men. It found no increase in prostate cancer events among men treated with testosterone, with an odds ratio of 0.88, and no increase in clinically significant prostate cancer, odds ratio 1.13. The variation between the individual trials was zero for both of those outcomes, meaning the trials agreed with each other rather than pointing in different directions. The confidence intervals are wide, because prostate cancer was an uncommon event in every trial, but the direction of the pooled estimate for any prostate cancer sits slightly below the no-difference mark rather than above it.

The single most informative trial is TRAVERSE. It enrolled 5,246 men aged 45 to 80 across 316 sites in the United States, all of whom had two testosterone measurements below 300 ng/dL along with symptoms of deficiency. They were randomly assigned, with neither the men nor their physicians knowing which they received, to testosterone gel or to a placebo gel. Prostate safety was planned in advance as a formal part of the trial rather than looked for afterward, and every suspected prostate cancer was reviewed by an independent committee.

Here is what it found, in plain numbers. Over about two years of treatment and 14,304 person years of follow up, high grade prostate cancer occurred in 5 of 2,596 men on testosterone, which is 0.19 percent, and in 3 of 2,602 men on placebo, which is 0.12 percent. That is eight cancers in total among more than five thousand men, and the difference between the groups was not statistically significant. Any prostate cancer, acute urinary retention, prostate biopsy, prostate surgery, and new medication for urinary symptoms were all similarly uncommon and none differed between the groups.

Two things about that trial are worth understanding rather than glossing over. Because prostate cancers were so uncommon in it, the trial is much better at showing the absence of a large effect than at settling questions about small differences, and average treatment lasted under two years, which is why longer-term data remain less extensive. And the men who entered had a PSA of 3.0 ng/mL or below, without severe urinary symptoms, with a mean starting PSA of 0.92. That last point is sometimes described as a limitation, but it is closer to a description of how testosterone therapy should be started in the first place. The trial answers the question most men are actually asking, which is what happens when a man is properly evaluated before treatment and then treated.

What happens when you measure testosterone in men before any diagnosis

---

A different and equally useful line of evidence asks whether men who go on to develop prostate cancer had higher testosterone to begin with. The Endogenous Hormones and Prostate Cancer Collaborative Group pooled eighteen prospective studies covering 3,886 men who developed prostate cancer and 6,438 who did not, with hormones measured from blood drawn before anyone had a diagnosis.

They found no association between prostate cancer risk and testosterone, calculated free testosterone, dihydrotestosterone, DHEA sulfate, androstenedione, estradiol, or calculated free estradiol. The studies did not disagree with one another, and adjusting for other risk factors barely moved the estimates.

Put simply, across eighteen separate populations, the men who later developed prostate cancer did not have higher testosterone beforehand than the men who did not.

Why the old belief took hold

---

Understanding where this fear came from makes it easier to see why modern evidence looks so different.

In 1941 Charles Huggins and Clarence Hodges reported that lowering testosterone, by castration or by estrogen treatment, made metastatic prostate cancer regress. That observation was correct, it was important enough to earn a Nobel Prize, and it remains the basis of androgen deprivation therapy for advanced prostate cancer today. In the same work they also reported that giving testosterone made prostate cancer grow. A later review of the original paper established that this second conclusion rested on the results of a single patient.

The reasoning that followed felt obvious at the time. If taking testosterone away treats prostate cancer, then giving testosterone must cause it. But that inference only holds if the relationship runs in a straight line across the whole range of testosterone concentrations, and it does not. A drug that helps a condition when withdrawn does not automatically cause that condition when given, particularly when the tissue in question stops responding above a certain hormone level.

The clinical reports of testosterone worsening prostate cancer came from a specific population. When Fowler and Whitmore reviewed 67 such men in 1981, all of them already had metastatic prostate cancer, and unfavorable responses were most frequent among men whose disease had already relapsed after hormonal treatment. That is a real and relevant finding about advanced, previously treated cancer. It was never a finding about men who do not have prostate cancer and whose testosterone is low. The generalization from one group to the other is where the belief came from, and it is the part that the last two decades of evidence have not supported.

Why raising testosterone behaves differently from lowering it

---

There is a straightforward biological explanation for the pattern in the data, known as the saturation model.

Testosterone and dihydrotestosterone act on prostate cells through the androgen receptor. Each cell has a finite number of receptor binding sites, and clinical data suggest those sites are substantially occupied at relatively low androgen concentrations, with the point of saturation sitting near a total testosterone of about 250 ng/dL. Below that level, changes in testosterone genuinely matter to prostate tissue. Within or above an adequate range, additional testosterone has little further to accomplish through that pathway, because the receptors it would act on are already occupied.

That accounts for what otherwise looks like a contradiction. Taking a man from a normal level down to castrate levels moves him far into the range where prostate tissue is responsive, and the response is dramatic. Taking a man from 300 up to 800 moves him from saturated to more saturated, and there is no further response available to produce.

Two observations in humans fit this. PSA rises with testosterone treatment in men whose starting level was below roughly 250 ng/dL and not in men who started above it, and when PSA was plotted against testosterone in a clinic population of 2,967 men, the curve flattened at approximately the same concentration.

This is a model, not proof of safety, and it should be held as such. A 2020 review questioned how much of the evidence originally assembled to support it had come from experiments designed to test something else, and the model's authors responded that recognizing saturation in data collected for other purposes was the point. Its value is that it offers a coherent reason why the steady, linear increase in prostate cancer risk that clinicians once expected from testosterone has not appeared in the trials. It is an explanation for the findings, not a substitute for them.

What about studies that have linked testosterone to prostate cancer?

---

Some studies have found associations, and they deserve to be described accurately rather than left out. The important thing to understand first is that they ask a different question from the one most men have in mind.

Genetic studies can estimate the effect of the testosterone level a man was set up to have for his entire life, from birth onward, by looking at inherited variation across very large populations. That is not the same question as whether restoring a deficient man's testosterone into a normal range in his fifties or sixties changes his risk.

With that distinction in place, here is what the largest such analysis found. Published in 2022, it used genetic instruments drawn from up to 194,453 men and outcome data on up to 79,148 prostate cancer cases. Genetically predicted higher free testosterone was associated with aggressive prostate cancer, with an odds ratio of 1.23 for each standard deviation higher, and with overall and early onset disease as well. The same publication also measured hormones directly in blood from up to 25 prospective studies, and there it found no association with aggressive disease overall, while total testosterone was associated in the opposite direction, with lower rather than higher risk.

Two points put that in proportion. The genetic estimate is expressed per standard deviation of lifetime exposure and carries no absolute event rate, so it cannot be translated into a personal risk figure for a man considering treatment. And the exposure it describes, a whole lifetime beginning before birth across a population that includes men who sat below the saturation point from childhood, is precisely the exposure the saturation model predicts would behave differently from a few years of replacement in an adult.

What the genetic data reasonably support is that androgen biology is not irrelevant to the prostate. What they do not support is the conclusion that treating a testosterone deficient man raises his risk, which is the question the randomized trials tested directly and did not find.

Low testosterone is not the safer state

---

It is worth knowing that the opposite assumption also fails to hold up. In a series of 937 men who had their prostate removed for cancer, with testosterone measured before surgery, the men with low testosterone had worse disease, not better. Predominant Gleason pattern 4 was found in 41.7 percent of the low testosterone group compared with 29.1 percent of those with normal levels, and 20.1 percent were upgraded from low grade on biopsy to high grade on the surgical specimen compared with 11.6 percent. Low testosterone was an independent predictor of both.

This does not show that testosterone therapy prevents aggressive prostate cancer, and it should not be read that way. What it does show is that the intuition behind the old fear, that less testosterone is the safer place to be, is not supported either.

Your PSA on treatment, and why it is monitored

---

A change in PSA is not the same thing as cancer, and keeping those two ideas separate makes the monitoring far less unnerving.

PSA does tend to rise modestly when testosterone is replaced, and the size of that rise is small. A meta-analysis of 15 randomized trials found an average increase of 0.154 ng/mL more than control, and 0.271 ng/mL with intramuscular dosing. A larger analysis of 28 trials found the same direction with a pooled difference of 0.08 ng/mL that did not reach statistical significance. TRAVERSE found a greater rise on testosterone than placebo, concentrated in the first year of treatment. Consistent with the saturation model, that rise appears in men who started below roughly 250 ng/dL rather than in men who started above it, which means the men most likely to see a measurable change are the ones who were most deficient to begin with, and also the ones with the most to gain from treatment.

It also helps to know how many ordinary things move PSA. In men who do not have cancer, the largest single influence is simply how large the prostate is. Prostatitis raises it, so does a urinary tract infection, recent ejaculation, a catheter or cystoscopy, and a recent biopsy. When the blood was drawn relative to any of those changes what the number means. Prostate enlargement and urinary symptoms are their own question, answered separately in will testosterone make an enlarged prostate worse.

That is the whole case for monitoring, and it is a constructive one. A PSA and a prostate exam before starting, with a repeat during the first year, is what allows a small expected rise to be recognized as exactly that, and allows a change that genuinely warrants attention to be identified early. A rise that is large, that keeps climbing past the first year, or that comes with an abnormal exam is a reason for urologic evaluation. Monitoring is ordinary good medicine, the same reason a blood count gets checked on treatment, and not a sign that the treatment is dangerous.

Careful evaluation is what makes the reassurance meaningful

---

Testosterone therapy is not right for every man, and the honest version of the reassurance on this page depends on that being true. Every man should be properly evaluated before starting, which includes a PSA, a prostate exam, and a discussion of family history, and the evaluation should be repeated on treatment rather than done once and forgotten.

Men with known or suspected prostate cancer are a separate situation that calls for individualized evaluation alongside a urologist. The evidence there is genuinely thinner, and it is observational rather than randomized. A 2026 systematic review gathered seven studies of testosterone therapy in men on active surveillance for prostate cancer, comparing 295 treated men against 6,826 untreated men followed for 2.3 to 6.1 years. No prostate cancer deaths or metastatic events were reported among the treated men, and the one controlled biopsy comparison found similar rates of Gleason upgrading, 10.7 percent against 9.4 percent. PSA stayed stable even as testosterone rose two to threefold. The reviewers were clear that small numbers, retrospective design, and short follow up limit what can be concluded, and that men who were selected for treatment may simply have had more favorable disease. The fair summary is that a history of prostate cancer is no longer treated as an automatic bar to testosterone therapy, and that it remains an individual decision made with a specialist rather than a general recommendation.

Stating it this way is not a hedge. A claim that testosterone has not been shown to increase prostate cancer risk in appropriately evaluated and monitored men is a stronger and more durable claim than an absolute assertion that testosterone can never affect the prostate, because it is the claim the evidence actually supports.

Dr. Darrell Wilcox · @wellnessdoc_4everyoung on Instagram

This content is for educational purposes only and does not constitute medical advice, diagnosis, or treatment. Testosterone is a prescription treatment with its own indications, contraindications, and monitoring requirements, and decisions about prostate cancer screening and about starting or continuing testosterone therapy should be made with your own physician and, where appropriate, a urologist who has your PSA history, exam, and family history in front of them. Individual results differ from trial averages, and the studies described here were conducted in specific populations that may not match your own situation. Dr. Wilcox is licensed to practice in multiple states. See About for current licensure.