Non-Metastatic CRPC in a Patient With a Seizure History
Three second-generation antiandrogens all extend metastasis-free survival by a similar margin — the actual decision turns entirely on which one his seizure history rules out and which of the remaining two his fall risk favors.
W.T., a 71-year-old man, lives alone since his wife passed two years ago, in the same split-level house they raised their children in — a house with a full flight of stairs he insists he has no intention of leaving. He has a remote seizure disorder from a traumatic brain injury sustained in a car accident at 34, well-controlled for over a decade on levetiracetam with no breakthrough seizures since. His prostate cancer has been followed on ADT for six years after primary radiation, and his PSA has now risen from an undetectable nadir to 2.1ng/mL with a doubling time of 7 months, meeting the definition of castration resistance. Conventional CT and bone scan show no metastatic disease — non-metastatic CRPC, the category where second-generation antiandrogens have shown a metastasis-free survival benefit large enough to have changed practice on their own.
All three approved agents — apalutamide, enzalutamide, darolutamide — showed a similar magnitude of metastasis-free survival benefit in their respective trials, SPARTAN, PROSPER, and ARAMIS, which means the actual choice here is driven almost entirely by toxicity profile rather than efficacy. Enzalutamide crosses the blood-brain barrier more readily than the other two and lowers seizure threshold — patients with a seizure history or on concurrent medications that lower the threshold were excluded from PROSPER for exactly this reason, and post-marketing reports have described breakthrough seizures in patients with prior seizure disorders started on it regardless. His own history, more than a decade controlled, is real but not nothing — the risk enzalutamide would add is not theoretical for him specifically, it is the reason patients like him were kept out of its own pivotal trial. Between the two remaining candidates, their own placebo-controlled comparisons diverge on exactly the endpoint that matters most for a man living alone: ARAMIS reported a fall incidence for darolutamide numerically close to placebo, while both SPARTAN and PROSPER reported meaningfully higher fall and fracture rates for apalutamide and enzalutamide respectively than for their own placebo arms.
Choosing among three drugs with one already ruled out
Enzalutamide is out. PROSPER excluded patients with a seizure history specifically because of the drug's blood-brain-barrier penetration and seizure-threshold effect, and post-marketing reports have described breakthrough seizures in exactly this population. That leaves apalutamide and darolutamide, both of which showed a similar metastasis-free survival benefit in their own trials.
Agreed on enzalutamide. Between the other two, I'd favor darolutamide. SPARTAN reported falls in roughly 16% on apalutamide against 9% on placebo; ARAMIS put darolutamide at about 4% against placebo's 5%. He lives alone and uses a full staircase every day. Darolutamide's lower blood-brain-barrier penetration is the mechanistic reason to expect that separation, and it's the better match for his actual home environment.
I'd get to darolutamide too, but not by that route, and I think the difference matters for how confidently we say it. SPARTAN, PROSPER and ARAMIS were three separate placebo-controlled trials with different enrollment and different fall ascertainment — nobody has ever randomized apalutamide against darolutamide. Reading 16% off one forest plot against 4% off another and calling it a head-to-head is the same move we'd criticize in a drug rep's slide deck.
What actually gets me there is narrower and specific to him: he's on levetiracetam, which is renally cleared and not a CYP substrate, so the interaction argument that sometimes separates these two agents is simply not live in his case. That leaves the fall data as the only discriminator we have — which is a reason to choose darolutamide, but a weaker reason than a within-trial comparison would be, and I'd rather say so out loud than let a cross-trial number harden into a fact in his chart.
Separately, worth putting on record rather than folding silently into "seizure history, therefore avoid enzalutamide": he's been controlled over ten years with no breakthrough events. That doesn't change today's decision, but his degree of control is a real, favorable data point worth documenting rather than treating as identical to a recent or unstable seizure history.
Agreed: darolutamide started, enzalutamide excluded outright given his seizure history, apalutamide passed over specifically for the fall/fracture profile given his living situation. Levetiracetam continued unchanged.
Not fully closed: whether his more-than-a-decade of seizure stability should be documented as a modifying factor for future decisions, or whether a seizure history is treated uniformly regardless of how long it's been controlled.
The neurologist and treating oncologist agreed to note the duration and stability of his control explicitly in the chart, distinct from a bare seizure-history flag, so that any future medication decision — including a hypothetical future need to revisit enzalutamide if his disease progressed on darolutamide — starts from his real risk profile rather than a generic exclusion. Neither treated this as changing today's choice; it was raised as forward-looking documentation, not a live disagreement about the current regimen.