Bone-Protective Agent Selection During Androgen Deprivation
A man who may remain on androgen deprivation for the rest of his life needs bone protection now. Denosumab has the strongest trial in his exact situation; what happens if he ever has to stop it is the argument against starting it.
H.W., a 70-year-old man, has spent most of his retirement building furniture in a garage workshop, a hobby he picked up after selling his accounting practice and one that keeps him on his feet and lifting lumber several afternoons a week. He was diagnosed with high-risk localized prostate cancer three years ago, treated with radiation and started on leuprolide at that time; his disease has responded well, PSA undetectable, and the current plan is to continue ADT indefinitely given his disease risk category rather than pursue a defined treatment duration. A DEXA scan obtained at the two-year mark of ADT showed a T-score of negative 2.3 at the lumbar spine, down from negative 1.1 at baseline — real, measurable bone loss attributable to the testosterone suppression itself, in a man whose baseline bone density had already been borderline osteopenic before treatment started.
The bone-protective decision is complicated by a fact that would matter less if his ADT had a defined endpoint: he may be on this therapy for the rest of his life. Denosumab's HALT trial, the largest randomized trial specifically in men receiving ADT for nonmetastatic prostate cancer, found a sixty-two percent relative reduction in new vertebral fractures at thirty-six months — three point nine percent with placebo down to one point five percent with denosumab — with no cases of jaw osteonecrosis reported in the trial. But denosumab's antiresorptive effect is fully reversible, and stopping it produces a well-documented rebound in bone turnover. The evidence for that comes from somewhere other than HALT: Cummings and colleagues' post-hoc analysis of the FREEDOM trial and its extension, in which the vertebral fracture rate rose from one point two per one hundred patient-years on treatment to seven point one after stopping. Two features of that analysis matter for H.W. specifically, and they cut in opposite directions. FREEDOM enrolled postmenopausal women with osteoporosis, not men on androgen deprivation, so its rebound estimate is being carried across a population line that HALT's fracture estimate is not. But its own internal finding is that the excess concentrated in participants who already had a prevalent vertebral fracture — and H.W. has never had a fragility fracture of any kind, which places him in the lower-risk half of the very analysis being invoked against him. For a man who might need bone protection for the next fifteen or twenty years, whether he ever stops the drug — by choice, by cost, or because some future circumstance intervenes — is not a hypothetical detail; the question is how much a rebound estimate borrowed from another population should weigh against a fracture estimate generated in his own.
In clinic, planning bone protection for the long run
I'd start denosumab. HALT is the trial built for exactly his situation — nonmetastatic prostate cancer, on ADT, no bone metastases — and it showed a sixty-two percent relative reduction in new vertebral fractures at thirty-six months, with no jaw osteonecrosis reported in the trial itself. Zoledronic acid has real evidence in osteoporosis generally and in men with ADT-related bone loss specifically, but nothing at that scale or that directly matched to his exact clinical scenario.
I'd lean toward zoledronic acid instead, precisely because his ADT has no planned endpoint. Denosumab's antiresorptive effect is fully reversible, and stopping it produces a real rebound — the FREEDOM discontinuation analysis found the vertebral fracture rate jumps from about one per hundred patient-years on treatment to seven per hundred after stopping, which erases the drug's protective effect entirely — it returns to roughly the untreated rate, and what rises above that baseline is not the fracture rate but the share of fracturing patients who sustain several vertebral fractures at once rather than one. Zoledronic acid stays embedded in bone and doesn't produce that same abrupt reversal. If he's going to be on bone-protective therapy for fifteen or twenty years, I'd rather start with the drug whose risk profile doesn't include a cliff on the way down.
This isn't a rejection of HALT's data — it's a question about what happens after the trial's own thirty-six-month window, which the trial itself wasn't designed to answer.
The rebound-fracture risk is real, well-documented, and worth taking seriously — but I don't think it's an argument against starting denosumab now, it's an argument for planning the exit before it's needed. The rebound phenomenon is a known, foreseeable event, not an unpredictable complication: if he ever needs to stop denosumab, transitioning to a bisphosphonate at that point is a well-established mitigation that blunts the rebound significantly, rather than leaving him unprotected. I'd rather start him on the drug with the strongest, most directly applicable randomized fracture data today, with an explicit plan on record that any future discontinuation gets bridged with a bisphosphonate rather than simply stopped.
The counterargument would be stronger if there were real evidence that starting zoledronic acid up front, rather than denosumab, produces comparably strong fracture reduction in exactly his population — and that evidence isn't nearly as direct. The zoledronic acid literature in ADT-associated bone loss is real but thinner and less matched to a fracture endpoint specifically in nonmetastatic prostate cancer than HALT is for denosumab.
Started on denosumab, 60mg subcutaneously every six months, with an explicit plan documented in his chart: if denosumab is ever discontinued for any reason, he transitions to a bisphosphonate rather than simply stopping, to blunt the rebound-fracture risk. Calcium and vitamin D supplementation confirmed adequate.
Not agreed: whether the endocrinologist's preference for starting zoledronic acid up front, specifically to avoid the discontinuation question altogether, should be the default approach in any patient expected to remain on ADT indefinitely. The endocrinologist views the bridging plan as a reasonable mitigation but still would have preferred not needing it in the first place; the urologic oncologist and pharmacologist see HALT's direct evidence as decisive enough to accept a foreseeable, manageable future risk in exchange for the strongest present data.