Skip to content

Vascular dementia — master index

Last curated: 2026-09-02 · status: draft (Codex build → independent Grok audit → Claude deepening pass → independent Grok re-audit → second Claude deepening pass, 2026-09-02; independent re-audit pending)

The condition in five sentences. Vascular dementia is the major end of vascular cognitive impairment, a spectrum in which cognitive and functional decline is attributed to cerebrovascular injury rather than to one lesion type (Sachdev 2014, PMID 24632990). The label is intrinsically probabilistic because infarcts, small-vessel disease, CAA, and Alzheimer pathology commonly coexist (Schneider 2007, PMID 17568013). MRI standardization through STRIVE-2 improves lesion measurement but does not make imaging alone diagnostic (Duering 2023, PMID 37236211). Prevention evidence is strongest for vascular-risk control: a 2025 cluster-randomized trial in 33,995 adults achieved a 22 mm Hg systolic contrast and cut all-cause dementia by 15%, while cholinesterase-inhibitor effects average roughly 1–2 ADAS-Cog points and may not be clinically important — an effect that six months of aerobic exercise matches or exceeds (He 2025, PMID 40258956; Battle 2021, PMID 33704781; Liu-Ambrose 2016, PMID 27760869). How large those drug effects are is itself contested — raw-scale and standardized-effect syntheses of overlapping trials differ roughly fivefold, and Western and Asia-Pacific evidence bases reach opposite conclusions about which agents work at all (Shi 2022, PMID 35048806; Dang 2024, PMID 39239652). The practical research agenda is to define mechanism-coherent populations, measure mixed pathology, connect screening to action, and use outcomes that matter in daily life.

Start here: overview. Research frontier: OPEN-QUESTIONS.md. Growth history: LOG.md.

Companion conditions: Alzheimer's disease for amyloid/tau biology and therapy; stroke for acute stroke and secondary-prevention infrastructure. The dedicated overlap page prevents a false clean separation.

At a glance

Layer Built content
Canonical wiki pages 19/19 (all draft pending re-audit of the second deepening pass)
Distinct live-resolved PMIDs 416 (21.9 per page)
Mean wiki page length / citations 211 lines / ~31 PMIDs
Landmark notes 5
Guideline/consensus records 25
Statistics sections 34
Patient-voice files 4
Open questions 36 + 27 junctions + 10 closed/downgraded
Live ClinicalTrials.gov records tabulated 34

Reading paths

Reader goal Path
Understand the condition overviewnosologysubtypes
Evaluate a patient phenotype assessmentneuroimagingmixed pathology
Study small-vessel mechanisms SVDpathophysiologybiomarkers
Focus on stroke post-stroke cognitionprevention
Compare interventions treatmentguidelinestrials
Safety and lived impact red flagspatient experience
Find the unresolved arguments overview → thirteen live controversiesOPEN-QUESTIONS.mdSTATISTICS known conflicts

Pages

File Scope Status
overview.md definition, burden, map, index of live controversies draft
nosology-and-diagnostic-criteria.md criteria, prevalence effects, autopsy validation draft
epidemiology-and-burden.md prevalence, incidence, mortality/burden draft
subtypes-and-clinical-syndromes.md post-stroke, subcortical, strategic, hypoperfusion draft
cerebral-small-vessel-disease.md STRIVE markers, progression, genetics draft
cerebral-amyloid-angiopathy.md Boston criteria, cognition, ARIA border draft
monogenic-and-inherited-forms.md CADASIL, HTRA1, COL4A1/2, Fabry draft
pathophysiology.md neurovascular unit, BBB, perfusion, clearance draft
mixed-pathology-and-alzheimer-overlap.md autopsy copathology and repository border draft
cognitive-profile-and-assessment.md profiles, screens, function draft
neuroimaging.md MRI/CT/PET and quantitative markers draft
biomarkers.md MarkVCID qualification, plasma marker dissociation draft
post-stroke-cognitive-impairment.md trajectory, predictors, screening draft
prevention-and-risk-factor-control.md BP, AF, metabolic/lifestyle prevention draft
treatment.md drugs, effect sizes, the two evidence readings draft
guidelines.md recommendation synthesis with grades and conflicts draft
clinical-trials-landscape.md live registry landscape and trial design draft
red-flags-and-safety-concerns.md mimics, bleeding, BP, driving/capacity draft
patient-experience-and-advocacy.md lived experience and priorities draft

Literature layer

Resource Contents Status
BIBLIOGRAPHY.md 418 topic-grouped live-resolved papers and uses; wiki coverage complete draft
notes/ 5 landmark paper notes curated (untouched this pass)
guidelines/REGISTRY.md 25 records, supersession, 22 disagreements, watch list draft
statistics/STATISTICS.md 34 quick-reference sections; 27 conflicts preserved side by side draft
patient-voice/README.md method, ethics, coverage curated (untouched this pass)
patient-voice/themes.md aggregate themes curated (untouched this pass)
patient-voice/sources.md annotated sources curated (untouched this pass)
patient-voice/organizations.md live-fetched organizations curated (untouched this pass; URLs need re-fetch at next audit)

Provenance and audit boundary

Built by Codex from live literature; independently audited by Grok (2026-08-31), which re-fetched every cited PMID and NCT, corrected journal/metadata mismatches, fixed the Filler predictor wording, resolved [unverified] flags with live sources or dated evidence-gap sentences, expanded thin pages, and promoted all 19 pages to curated.

Deepening pass (Claude, 2026-08-31). The audited condition was structurally complete but thin (125 lines and ~15 citations per page). This pass extended all 19 pages with landmark effect sizes and intervals, competing guideline positions with their grades, quantified epidemiology, deeper mechanism, and explicitly two-sided controversies — adding 97 verified records (121 → 218 distinct PMIDs) plus 110 bibliography entries, eleven statistics sections, three guideline records, ten new open questions, and seven new cross-domain junctions. All 218 PMIDs were re-resolved live after writing. The eight live controversies the pass documented are indexed in overview.md; the six questions it closed or downgraded are tabled in OPEN-QUESTIONS.md.

Second deepening pass (Claude, 2026-09-02). The condition was structurally complete but sourced narrowly: 17.6 distinct PubMed records per page across the whole wiki, meaning pages recycled a shared reference pool. This pass widened the evidence base rather than lengthening pages — 335 → 416 distinct PMIDs (21.9 per page) — adding landmark trials with their intervals, competing guideline positions with their grades, quantified epidemiology in place of qualitative statements, deeper mechanism, and five new explicitly two-sided controversies. It also closed a 118-record gap between the wiki and the bibliography, expanded STATISTICS from 23 to 34 sections and the guideline registry from 21 to 25 records, added six open questions and ten junctions, and corrected a stale registry-absence claim (a phase-3 successor to LACI-2 does exist: IMPACT, NCT07252544). All 416 PMIDs were re-resolved live after writing. All 19 pages returned to draft for independent re-audit; the landmark notes and patient-voice layer were deliberately untouched. See LOG.md.

Deepening-pass re-audit (Grok, 2026-08-31). Independent audit of the Claude additions: every wiki PMID and NCT re-fetched live; new numeric claims checked against abstracts; absence searches re-run. Errors fixed (Ding vs Cunningham trial-set identity; Zhang IRVR sign; Cerebrolysin pooling denominator; stale VaD BPSD-absence claim). Unused reference-list padding removed. All 19 pages promoted to curated. See LOG.md.