5-Alpha-Reductase Inhibitor Use in the Man on Active Surveillance
A man under active surveillance for low-risk prostate cancer needs relief from worsening BPH symptoms. The drug most likely to help is the one whose fifteen-year-old high-grade cancer signal was never fully closed, only mostly explained.
L.G., a 69-year-old man, has sung in his church choir for over forty years, tenor section, and has started missing rehearsals lately because he can't make it through a two-hour practice without excusing himself twice for the bathroom. He was diagnosed eighteen months ago with low-risk prostate cancer, Gleason 3+3, a single core involved on a twelve-core biopsy, and has been on active surveillance since, with PSA and biopsies stable at his most recent six-month check. His urinary symptoms have worsened over the same period — a weak stream, frequency that now regularly interrupts sleep and choir practice alike, an AUA symptom score that has climbed from 12 to 19 — findings consistent with progressive benign prostatic hyperplasia layered on top of the cancer he's already being monitored for.
The most effective medical option for his obstructive symptoms happens to be the drug carrying the field's longest-running unresolved safety question. Finasteride reduced overall prostate cancer incidence by nearly twenty-five percent in the Prostate Cancer Prevention Trial, but the same trial found a statistically significant increase in high-grade disease among men taking it — Gleason 7 to 10 cancers in thirty-seven percent of the finasteride group's cancers versus twenty-two percent of placebo's. Multiple bias-adjusted reanalyses since have argued this reflects finasteride's own effect on biopsy sensitivity and prostate volume, which together make existing high-grade tumors easier to detect rather than more likely to arise — and REDUCE, the comparable trial for dutasteride, found no difference at all in Gleason 7 to 10 disease. But REDUCE did not come back clean either: it reported more Gleason 8 to 10 cancers on dutasteride than on placebo, one point zero percent versus zero point five percent, and on the strength of both trials together the FDA added a high-grade-prostate-cancer warning to the label of the entire 5-alpha-reductase inhibitor class — finasteride's included, where it remains today. A later modeling analysis of REDUCE's own data attributed that imbalance largely to its biopsy schedule and to prostate size at the time of biopsy, the same explanation offered for PCPT, and could not exclude a real association. But L.G. is not a healthy man deciding whether to take a preventive drug; he already has cancer under surveillance, and whatever residual uncertainty remains in that detection-bias explanation lands most directly on a patient in exactly his position.
In clinic, treating BPH under a cancer's watch
I'd avoid finasteride here and manage his symptoms with an alpha-blocker alone. PCPT's high-grade cancer signal was statistically significant in its primary analysis — Gleason 7 to 10 disease in thirty-seven percent of finasteride-arm cancers versus twenty-two percent on placebo. I understand the detection-bias argument, but he already has a cancer diagnosis under active surveillance. If that explanation turns out to be wrong, or only partially right, the person most exposed to being wrong about it is exactly a patient like him, not a healthy man in a prevention trial.
I'd actually recommend finasteride, and I don't think that's in tension with taking his cancer seriously. The bias-adjusted reanalyses of PCPT are substantial, not a single reassuring footnote — modeling that accounts for finasteride's known effects on prostate volume and biopsy sensitivity found the apparent high-grade excess is consistent with improved detection, not true potentiation. REDUCE, the dutasteride trial, found no difference in Gleason 7 to 10 disease, and the Gleason 8 to 10 imbalance it did report was modeled, using the trial's own data, as largely an artifact of its biopsy schedule and of gland size at biopsy — the same explanation, arrived at independently, in a second drug and a second trial. And there's a practical synergy here that cuts in finasteride's favor specifically for a surveillance patient: a smaller gland means better-yield biopsies going forward, which is exactly the tool his surveillance protocol depends on.
Treating the original PCPT signal as though it stands unexplained ignores how much independent reanalysis has actually gone into understanding it. This isn't a raw, unaddressed red flag — it's a finding that's been investigated from several angles and largely, though perhaps not completely, accounted for.
I'd land between both of you, and I think the honest position is less comfortable than either. The detection-bias explanation is well-supported — multiple independent groups modeling PCPT's own data have converged on it, and REDUCE's investigators reached the same conclusion about their own Gleason 8 to 10 imbalance. But 'well-supported' isn't the same as 'proven' — and I'd note the FDA was not persuaded enough to leave the class labels alone, and I don't think we can tell him the original signal simply isn't real. What I can say is that the explanation has held up under real scrutiny for over fifteen years without a credible alternative mechanism displacing it, which is a meaningfully different epistemic position than an unresolved finding sitting untested.
Given that, I think finasteride is a reasonable choice for him, with informed consent that actually names the history rather than skips past it: PCPT found the signal, it's most likely a detection artifact of finasteride's own biopsy-enhancing effect, REDUCE found a smaller version of the same thing with a related drug and explained it the same way, his label carries an FDA class warning that reflects both trials, and no confirmed mechanism for true potentiation has been found in over a decade of looking — but a small residual uncertainty, however probably explained, exists in a way it wouldn't for a drug with no such history at all.
L.G. chose to start finasteride alongside tamsulosin, after a documented informed-consent conversation covering PCPT's original finding, the bias-adjusted reanalyses, REDUCE's parallel dutasteride finding, the FDA class-wide label warning that applies to the drug he is being handed, and the residual uncertainty the pharmacologist named explicitly. His surveillance biopsy schedule continues unchanged, with the added benefit that a smaller gland should improve future biopsy yield.
Not agreed: whether every active-surveillance patient facing this same BPH-treatment decision should receive the same informed-consent framing, or whether the conversation should differ based on individual risk tolerance and cancer grade. The urologic oncologist remains more comfortable defaulting to an alpha-blocker-only approach for surveillance patients generally; the BPH-focused urologist and pharmacologist think the informed, case-by-case approach used here is the more defensible standard going forward.