Borrowing Alzheimer's Drugs for Vascular Dementia: Whose Evidence Actually Applies
A family asks whether the drugs that help Alzheimer's disease might help their father's stepwise, post-stroke decline too. The most specific trial ever run on that exact question already answered it — and the answer depends on which kind of vascular dementia he actually has.
George F., 80, a retired postal supervisor, has had three small strokes over the past four years, each confirmed on imaging and each followed by a visible, stepwise drop in his thinking rather than the slow, gradual decline his daughter has read about with Alzheimer's disease. He managed a twenty-two-person sorting facility for three decades before retiring, well-controlled hypertension and type 2 diabetes his only chronic conditions before the first stroke. Between strokes, his function has plateaued rather than continued to worsen — a pattern his neurologist has pointed to as characteristic of vascular rather than neurodegenerative disease. Brain MRI shows extensive small-vessel ischemic changes and three discrete infarcts; amyloid-PET, obtained specifically to settle the question, was negative.
His daughter, a nurse, asks today whether he should try one of "the Alzheimer's drugs" anyway — she has read that cholinergic deficits show up in vascular dementia too, and reasons that a drug class this well-established for one form of dementia is worth trying for another. The question is more genuinely contested than it might sound: cholinesterase inhibitors have been formally tested in vascular dementia, in trials built specifically to answer this question rather than borrowed from Alzheimer's data — and the most rigorous of those trials found a real answer that depends on exactly the distinction his workup was designed to make: whether his disease is purely vascular, or vascular disease layered on top of underlying Alzheimer's pathology. His negative amyloid-PET puts him squarely in the first group, which is precisely where that trial's own results were least encouraging.
In the stroke follow-up clinic
I think it's reasonable to offer galantamine here. The GAL-INT-6 trial — Erkinjuntti and colleagues, Lancet, 2002 — randomized 592 patients with probable vascular dementia or Alzheimer's disease combined with cerebrovascular disease to galantamine or placebo. Overall, galantamine showed a 2.7-point ADAS-Cog benefit over placebo, with more patients stable or improved on clinician-rated global impression — seventy-four percent versus fifty-nine percent. Cholinergic deficits show up on pathology in vascular dementia too, not just Alzheimer's, so the mechanism has real grounding here.
I want to look at that same trial's own subgroup data before we extend it to George specifically. GAL-INT-6 enrolled both pure vascular dementia and mixed Alzheimer's-with-cerebrovascular-disease patients. The correspondence following its publication noted that galantamine showed no significant benefit over placebo in the pure vascular dementia subgroup — the benefit that drove the overall result came from the mixed-diagnosis patients.
George's amyloid-PET is negative. That's not an untested, ambiguous case — it places him specifically in the subgroup where this exact trial, of this exact drug, already found no significant effect. I don't think we should offer this as though the overall trial result speaks to his situation directly.
Both of those are accurate about what the trial actually showed, including in its own subgroup. I'd note the trial's authors themselves suggested the pure-vascular subgroup may simply have been underpowered to detect a smaller true effect, not proof that none exists — that's a real, if softer, caveat worth naming.
Given that genuine uncertainty, and how little harm this class carries at appropriate doses, I'd offer George a defined eight-week trial with a specific caregiver-rated outcome measure, explicit that we're testing this individually rather than extending a result the drug's own trial couldn't establish for his exact diagnosis. Standard titration, though — 4 mg twice daily to start, and no increase for a full four weeks. The label is deliberate about that because the cholinergic side effects are dose- and speed-dependent, and an eight-week window is not a reason to rush a titration that's meant to take that long.
Agreed: an eight-week trial of galantamine offered, with the family explicitly told this represents an individual trial rather than an extension of GAL-INT-6's positive overall result, given George's own negative amyloid-PET placing him in that trial's less-encouraging subgroup. Vascular risk factor management is reinforced regardless of the outcome.
Not agreed: whether it was right to offer the trial at all once the specific subgroup finding was on the table. The pharmacologist would have preferred not to offer it absent stronger pure-vascular-dementia evidence; the geriatrician's underpowered-subgroup caveat was accepted as a legitimate basis for proceeding, but not as fully resolving the disagreement.