Testosterone After Prostate Cancer: The Saturation Model Meets a Cautious Guideline
A single patient, three years past radical prostatectomy with favorable pathology and undetectable PSA, now hypogonadal and asking for treatment. The disagreement isn't about the mechanism, or about whether a trial exists — it's about whether a twelve-week randomized result can carry the weight of an open-ended prescription.
Arthur D., a 71-year-old retired golf course groundskeeper who spends much of his time now helping raise his two grandchildren, is three years past a radical prostatectomy for Gleason 6, organ-confined prostate cancer — undetectable PSA at every check since. What brings him back isn’t his cancer, which by every available measure has stayed gone, but a testosterone of 245 ng/dL on two morning draws and the fatigue and low libido that go with it, symptoms he’s mostly attributed to age until his grandchildren’s energy made the gap harder to ignore. His medical oncologist, when Arthur raised the idea of treatment, pushed back on reflex: testosterone feeds prostate cancer, she told him, the belief that shaped decades of practice before anyone tested whether it actually holds at replacement doses in a man like him.
The saturation model, formalized by Morgentaler and Traish, is the real mechanistic challenge to that belief, and Arthur’s own number sits at its hinge. The model holds that prostate tissue’s androgen receptors saturate at testosterone levels well below the normal physiologic range — roughly 200 to 250 ng/dL — so raising a deficient man’s testosterone back to normal shouldn’t drive additional tumor growth the way the older model, built on castration-level comparisons, implied it would. At 245 ng/dL he is already sitting inside that saturation window rather than below it, which is the specific reason the model predicts treating him adds symptom relief without adding much androgenic signal to whatever prostate tissue remains. The AUA testosterone-deficiency guideline reflects the same shift more cautiously, countenancing testosterone therapy after favorable-pathology prostatectomy while stating plainly that the evidence is inadequate to quantify the risk-benefit ratio for men with a history of prostate cancer. What has changed since that language was written is that a randomized trial in his population now exists: SPIRIT, reported by Bhasin and colleagues in 2026, randomized 136 men with undetectable PSA at least two years after radical prostatectomy for organ-confined, low-grade disease, and found testosterone improved sexual desire, sexual activity, body composition, and aerobic performance, with no biochemical recurrence in either arm. Arthur meets its entry criteria almost exactly. What he does not match is its clock — SPIRIT treated for twelve weeks and followed for twenty-four, and Arthur is asking about the rest of his life.
In the joint urology-oncology visit
Start testosterone. His pathology is favorable, his PSA has been undetectable for three years, and at 245 he’s already inside the saturation range — the model predicts we’d be relieving his symptoms without meaningfully changing the androgen signal his prostate bed sees. SPIRIT randomized men with his pathology and his PSA history and saw no biochemical recurrence in either arm. This is the scenario the AUA guideline says can reasonably be considered, and now there’s a trial in it.
I'd read that guideline language more cautiously than 'reasonably considered' suggests — it's a hedge, saying the evidence is inadequate to quantify the risk-benefit ratio, not that the risk has been shown to be low. And I'd be careful how much weight SPIRIT carries. It treated 136 men for twelve weeks and followed them for twenty-four, powered for sexual function, not for oncologic safety. 'No biochemical recurrence in either arm' sounds reassuring, but recurrence in Gleason 6 organ-confined disease is rare enough that a trial that small and that short would likely have seen none regardless of what testosterone does over years.
Calling SPIRIT 'a randomized trial in his population' is accurate and still doesn't get you where you want to go — a trial can match a patient on entry criteria and miss him entirely on duration, and duration is the whole of what he's asking about.
I don't think this has to be a straight yes-or-no. Start testosterone at a conservative dose, but recheck PSA and testosterone at three months rather than the usual six to twelve — treat this as a monitored trial of therapy rather than an open-ended prescription, giving us a real, early chance to catch anything neither the mechanism nor twelve weeks of SPIRIT would have predicted.
That doesn't resolve the duration question either of you is actually arguing about — it just means we find out on a three-month cycle instead of waiting for a trial nobody has run yet to run long enough.
Agreed: start testosterone cypionate at a conservative dose, with PSA and testosterone rechecked at three months rather than the standard interval, and explicit shared decision-making documentation given how directly the guideline’s own hedge language was discussed with Arthur.
Not agreed: whether a twelve-week randomized result plus hedge-strength guideline language is sufficient basis for therapy Arthur may stay on for a decade. The oncologist would have preferred enrollment in a longer-duration study or registry if one were locally available; none is, and while routine care proceeds, the disagreement about whether SPIRIT’s follow-up reaches Arthur’s actual time horizon is recorded rather than resolved.