If you’d asked me a year ago whether low testosterone at prostate cancer diagnosis was a good or bad thing, I’d probably have said neutral. Maybe even vaguely reassuring. Less fuel for the fire, right?
Two 2025 studies suggest the opposite, and the effect sizes are big enough that I don’t think we can keep ignoring this.
The Data
Finnish cohort (BJUI Compass, 2025). Retrospective, 2,544 prostate cancer patients from a national registry. Low T (under 8.0 nmol/L) at diagnosis was an independent predictor of worse overall survival: multivariable HR 1.58, adjusted for age, PSA, Gleason, stage, Charlson comorbidity index, and diabetes. The effect was consistent whether men were under or over 70. The grey zone (8.0 to 10.4 nmol/L) showed no survival difference against normal.
PSA-screened cohort (Cancer Medicine, 2025). Smaller (n=350) but PSA-screened, so detection bias is somewhat controlled. Low T was associated with more advanced T-stage at diagnosis, higher cancer-specific mortality (adjusted HR 2.70 in men with minimal comorbidity), and higher all-cause mortality (adjusted HR 1.90).
Neither study is an RCT. Both are retrospective. But the Finnish cohort is large, the hazard ratios are consistent across subgroups, and the signal goes in the same direction from two independent datasets.
Why It Might Make Biological Sense
Two plausible mechanisms. The boring one: low T is a marker of metabolic dysfunction and frailty, so these men do worse because they’re sicker at baseline. The interesting one: a tumour that establishes and grows in a low-androgen environment may already be selected for castration-resistant biology. It’s adapted to survive without androgen stimulation before you even start ADT. That would explain both the worse stage at diagnosis and the poorer response to treatment.
Neither mechanism is proven. But the second one is genuinely thought-provoking if you’re trying to understand why some cancers seem to shrug off hormone therapy from the start.
What It Means in Practice
We already measure PSA, grade group, and stage before making treatment decisions. Should testosterone be part of that baseline workup? These data suggest it has prognostic value, at least for risk stratification and counselling. Not to treat the hypogonadism on the spot, but to flag the men who might do worse than their Gleason score alone would predict.
It also makes the “TRT causes prostate cancer” narrative look even more simplistic than it already did. Low T isn’t protective. It may just mean the cancer is already running on a different programme.