Skip to content

Vascular and metabolic contributions

This is the shared-border page. Alzheimer's disease and vascular dementia are curated as separate conditions in this knowledge base because they have distinct criteria sets, mechanisms and trial pipelines. They are not separate at autopsy. This page describes the vascular and metabolic contribution to Alzheimer's disease; the reciprocal page, mixed-pathology-and-alzheimer-overlap.md, is on that condition's canonical page list but has not yet been written (vascular dementia is seeded, not built). Neither page asserts a clean separation, because the pathology literature does not support one.

TL;DR — Mixed pathology is the common case, not the exception: among community-dwelling people with dementia, Alzheimer pathology plus infarcts (38.0%) was more frequent than pure Alzheimer pathology (30.0%), and more than half had multiple diagnoses (Schneider 2007, PMID 17568013). In 1,079 autopsied participants from two longitudinal cohorts, 94% had at least one neuropathology and 78% had two or more; AD was the commonest (65%) but occurred in isolation in only 9%; more than 230 distinct neuropathologic combinations were observed, each in fewer than 6% of the cohort; and AD accounted for about 50% of cognitive loss on average but 22–100% at the individual level (Boyle 2018, PMID 29244218). Vascular disease acts on AD in at least three distinct ways: it lowers the threshold at which a given tau/amyloid burden produces dementia (Kapasi 2017, PMID 28488154); it appears upstream of amyloid, since midlife — but not late-life — vascular risk factors predicted elevated amyloid PET two decades later (≥2 midlife risk factors OR 2.88, 95% CI 1.46–5.69) (Gottesman 2017, PMID 28399252); and it acts through blood–brain-barrier failure independent of amyloid and tau, most strongly in APOE4 carriers (Nation 2019, PMID 30643288; Montagne 2020, PMID 32376954). The therapeutic implication is unresolved: intensive blood-pressure lowering reduced MCI (HR 0.81, 95% CI 0.69–0.95) but did not significantly reduce probable dementia (HR 0.83, 0.67–1.04) in SPRINT MIND (PMID 30688979).

How much of "Alzheimer's disease" is Alzheimer's disease?

Study Design Result
Schneider 2007 (PMID 17568013) First 141 autopsies, Rush Memory and Aging Project Of 50 with dementia: 38.0% AD + infarcts, 30.0% pure AD, 12% vascular dementia, 12% AD + PD/LBD; >50% with dementia had multiple diagnoses vs >80% with none/one among non-demented; multiple diagnoses raised dementia odds (OR 2.8, 95% CI 1.2–6.7)
Boyle 2018 (PMID 29244218) 1,079 autopsies, two longitudinal cohorts, ≥2 cognitive evaluations (max 22) 94% had ≥1 pathology; 78% ≥2; 58% ≥3; 35% ≥4; AD in 65% but alone in 9%; >230 combinations; AD explained ~50% of cognitive loss on average, 22–100% per person; Lewy bodies and hippocampal sclerosis also potent, with person-specific impact
Kapasi 2017 (PMID 28488154) Clinical-pathological review Concomitant cerebrovascular (macroinfarcts, microinfarcts, atherosclerosis, arteriolosclerosis, CAA) and neurodegenerative (Lewy bodies, TDP-43, hippocampal sclerosis) pathologies lower the threshold for a clinical AD diagnosis
Nelson 2019 (PMID 31039256) LATE consensus report ~25% of community-based autopsy brains have enough LATE neuropathologic change to be associated with discernible cognitive impairment; LATE mimics AD-type amnestic dementia; risk alleles in GRN, TMEM106B, ABCC9, KCNMB2 and APOE
Richardson 2026 (PMID 42184025) 61 Braak-V brains, quantitative p-tau In multivariable analysis only neocortical p-tau burden and microinfarcts independently affected cognitive decline; Aβ, LATE-NC, Lewy bodies and other cerebrovascular measures did not

Boyle's "230 combinations, each in <6%" is the central fact. It means that at the individual level there is no modal disease: the person in front of a clinician has a personal mixture, and a single-target therapy addresses a fraction of it that ranges from 22% to 100%. This is the strongest argument that trial populations selected for biomarker purity are unrepresentative of the treated population — see clinical trials landscape and OQ entries in OPEN-QUESTIONS.md.

Note also that Richardson's result cuts the other way for some vascular markers: with tau quantified properly, microinfarcts remained an independent contributor while other cerebrovascular measures did not (PMID 42184025). "Vascular contribution" is not one variable.

Three distinct mechanisms

1. Vascular pathology lowers the dementia threshold

The threshold model holds that a fixed AD pathological burden produces dementia at a lower level when accompanied by infarcts, arteriolosclerosis, Lewy bodies or TDP-43 — that is, the pathologies are additive or interactive in their effect on the clinical transition rather than on each other (Kapasi 2017, PMID 28488154). Schneider's odds ratio of 2.8 for multiple versus single pathology is the same claim expressed as risk (PMID 17568013).

A population-level consequence is visible in secular trends. In Gothenburg 85-year-olds examined 22 years apart, dementia prevalence fell from 29.8% to 21.7% (OR 0.66, 95% CI 0.50–0.86), the decline was mainly in vascular dementia, and the odds ratio for dementia given stroke fell from 4.3 to 1.8 (interaction P=0.008) — better stroke care changed the threshold, not the amyloid (Skoog 2017, PMID 28733627). Stroke itself remains a strong risk factor: pooled HR 1.69 (95% CI 1.49–1.92) for prevalent stroke across 36 studies (1.9 million participants) and RR 2.18 (1.90–2.50) for incident stroke across 12 studies (1.3 million) (Kuźma 2018, PMID 30177276). The stroke-side literature is curated at stroke.

2. Vascular risk acts upstream of amyloid — but only in midlife

In ARIC-PET, 322 dementia-free participants had vascular risk factors measured at ages 45–64 and florbetapir PET a median 23.5 years later (mean age 76). Elevated SUVR (>1.2) was present in 50.9%. Midlife obesity was associated with elevated amyloid (OR 2.06, 95% CI 1.16–3.65), and the association scaled with the number of midlife risk factors: elevated amyloid in 30.8% with none, 50.4% with one and 61.2% with two or more; adjusted OR for ≥2 versus 0 was 2.88 (1.46–5.69). Crucially, late-life vascular risk factors were not associated with late-life amyloid (OR 1.66, 0.75–3.69) (Gottesman 2017, PMID 28399252).

Two readings compete and this page does not adjudicate: either midlife vascular exposure promotes amyloid accumulation over decades, or midlife risk factors index a lifetime exposure that late-life measurement fails to capture (survivor bias, treatment, reverse causation from weight loss in preclinical disease). Either way, the timing result is a direct challenge to prevention trials that intervene after age 70 — see risk reduction and prevention.

3. Blood–brain-barrier failure, independent of amyloid and tau

Study Population Finding
Nation 2019 (PMID 30643288) Individuals with early cognitive dysfunction Hippocampal BBB breakdown on dynamic contrast-enhanced MRI and elevated CSF soluble PDGFRβ (a pericyte-injury marker) occurred irrespective of Aβ and tau biomarker status
Montagne 2020 (PMID 32376954) APOE ε3/ε4 and ε4/ε4 vs ε3/ε3 BBB breakdown in hippocampus and medial temporal lobe distinguished APOE4 carriers, present even when cognitively unimpaired and worse when impaired, unrelated to CSF or PET Aβ/tau; high baseline CSF sPDGFRβ predicted future cognitive decline in ε4 carriers only, after controlling for Aβ and tau, and correlated with cyclophilin A–MMP9 pathway activity
Shirzadi 2024 (PMID 38452039) 166 cognitively unimpaired, Harvard Aging Brain Study, follow-up 8.5 ± 2.7 y Latent factors of systemic vascular risk, white-matter injury and relative cerebral blood flow each explained PACC decline beyond amyloid and tau; grey-matter atrophy mediated these associations, most strongly for white-matter injury

This is the most consequential recent development for the AD/vascular border: it identifies a mechanism by which the strongest AD risk gene acts vascularly and amyloid-independently. If confirmed, APOE4 is simultaneously an amyloid-clearance gene (amyloid biology), a tau-spreading modifier (tau biology and spread) and a cerebrovascular gene — which would make "is this AD or vascular dementia?" the wrong question for ε4 carriers.

Cerebral amyloid angiopathy: where the two conditions are the same disease

CAA is amyloid deposited in the cerebrovascular wall rather than the parenchyma. It is simultaneously an AD copathology (present in 71% of posterior cortical atrophy autopsies, for instance — see clinical presentation and staging), a cause of lobar haemorrhage that will be curated under vascular dementia's planned cerebral-amyloid-angiopathy.md, and the substrate of the principal harm of anti-amyloid therapy.

The Boston criteria v2.0 define probable CAA in vivo as at least two strictly lobar haemorrhagic lesions (ICH, cerebral microbleeds or cortical superficial siderosis foci) or one strictly lobar haemorrhagic lesion plus one white-matter feature (severe centrum semiovale perivascular spaces or multispot white-matter hyperintensities). Diagnostic accuracy against histopathology: sensitivity 74.8% (95% CI 65.4–82.7) and specificity 84.6% (71.9–93.1) in derivation; 92.5% (79.6–98.4) and 89.5% (66.9–98.7) in temporal validation; 80.2% (70.8–87.6) and 81.5% (61.9–93.7) in geographical validation; and 74.5% (65.4–82.4) sensitivity with 95.0% (83.1–99.4) specificity in the autopsy-standard group, where AUC was 0.848 (0.794–0.901) versus 0.798 (0.741–0.854) for the previous criteria (P=0.0005) (Charidimou 2022, PMID 35841910).

The ARIA connection

Amyloid-related imaging abnormalities share clinical, biological and pathophysiological features with CAA, and baseline microhaemorrhage on MRI increases ARIA risk, as do APOE haplotype and antibody dose (Hampel 2023, PMID 37280110). Autopsy of aducanumab-treated patients showed that regions corresponding to ARIA on MRI contained microinfarcts with haemosiderin, complement activation and CD68-positive vessel walls, originating from Aβ-laden leptomeningeal and penetrating vessels (Boon 2025, PMID 41109234). ARIA is therefore best understood as iatrogenic destabilisation of CAA, which makes CAA burden — a vascular measurement — a determinant of who can safely receive an anti-amyloid drug. Details on anti-amyloid immunotherapy and red flags and safety concerns.

Metabolic contributions

Exposure Evidence Effect
Diabetes mellitus Religious Orders Study, 824 participants followed up to 9 years, 151 incident AD HR 1.65 (95% CI 1.10–2.47) for incident AD adjusted for age, sex, education; lower baseline global cognition, episodic, semantic, working memory and visuospatial ability; 44% greater rate of decline in perceptual speed (P=0.02) but not other domains (Arvanitakis 2004, PMID 15148141)
Midlife obesity ARIC-PET OR 2.06 (1.16–3.65) for elevated late-life amyloid (Gottesman 2017, PMID 28399252)
Brain insulin signalling as a target Intranasal insulin 40 IU daily for 12 months, 289 randomised, primary ITT n=240 No difference in ADAS-cog-12 change at 12 months (0.0258 points, 95% CI −1.771 to 1.822, P=0.98); no differences in secondary clinical or CSF outcomes; no clinically important adverse events; device reliability problems affected the first 49 participants (Craft 2020, PMID 32568367)

The diabetes association is robust and replicated; the mechanism is not established, and the one direct test of the "brain insulin resistance" hypothesis in AD was null. Whether that reflects a wrong hypothesis, an inadequate delivery device or an inadequate dose is unresolved. A second randomised test of a metabolic drug in non-diabetic mild-to-moderate AD was also negative on its primary: ELAD (n=204, liraglutide vs placebo for 52 weeks) found no difference in cerebral glucose metabolic rate (difference −0.17, 95% CI −0.39 to 0.06, P=0.14); an unadjusted secondary ADAS-Exec difference favoured liraglutide (0.15, 0.03–0.28, P=0.01) with no difference on ADCS-ADL or CDR-SB (Edison 2026, PMID 41326666; NCT01843075). Oral semaglutide then failed on CDR-SB at 104 weeks in two phase 3 trials — see clinical trials landscape. Glucose-lowering drug classes and dementia risk remain an active observational literature with substantial confounding by indication; the randomised tests in established AD are not evidence of prevention, and they are not evidence of benefit. Midlife obesity, independently of later comorbidity, carried HR 1.74 (1.34–2.26) for dementia in 10,276 people examined at ages 40–45 (Whitmer 2005, PMID 15863436) — a timing result that belongs with Gottesman rather than with these late-life drug trials.

Does treating the vasculature help?

SPRINT MIND is the best available randomised test. Among 9,361 hypertensive adults ≥50 without diabetes or prior stroke, randomised to systolic targets <120 or <140 mm Hg, over a median 5.11 years of follow-up (median 3.34 years of intervention, trial stopped early for cardiovascular benefit) (PMID 30688979):

Outcome Intensive Standard Effect
Probable dementia 7.2 per 1,000 py (149 cases) 8.6 per 1,000 py (176 cases) HR 0.83 (95% CI 0.67–1.04)
Mild cognitive impairment 14.6 per 1,000 py 18.3 per 1,000 py HR 0.81 (0.69–0.95)
MCI or probable dementia 20.2 per 1,000 py 24.1 per 1,000 py HR 0.85 (0.74–0.97)

The dementia result was underpowered because of early termination and fewer events than expected — an interpretive caution the investigators state explicitly. The imaging substudy adds a complication: at ~4 years, intensive treatment produced a slightly greater hippocampal volume decrease (between-group difference −0.033 cm³, 95% CI −0.062 to −0.003, P=0.03), with no differences in AD-signature regional atrophy, posterior cingulate cerebral blood flow or cingulum fractional anisotropy (Nasrallah 2021, PMID 33683313). Blood-pressure lowering appears to reduce clinical cognitive impairment without moving AD-specific imaging markers — consistent with a threshold effect rather than an effect on AD biology.

Pericyte injury and CAA: human measurements, unresolved direction

CSF soluble PDGFRβ, used as a pericyte-injury marker, was higher in 39 biomarker-defined AD cases than 39 controls and correlated with CSF albumin (r=0.45), total tau (r=0.50) and p-tau (r=0.41), but not Aβ42 (Miners 2019, PMID 31521199). In a separate 158-person cohort, sPDGFRβ peaked at CDR 0.5, correlated with the CSF:serum albumin ratio in CDR 0–0.5 (β=0.314) and predicted annual MMSE change within CDR 0.5 (β=−0.400); those relations were absent or different in later disease (Lv 2023, PMID 36915135). Elevated CSF angiopoietin-2, an endothelial-injury marker, correlated with sPDGFRβ, fibrinogen and the CSF:serum albumin ratio and with tau and neuronal-injury markers in early AD, and was highest at the MCI stage (Van Hulle 2024, PMID 38182581); it is a second barrier-related signal rather than a statistically independent one. All three studies are observational, so barrier failure could be upstream, downstream or a parallel amplifier.

CAA prevalence varies enough that ancestry and ascertainment cannot be ignored. In 848 community autopsies from the US ACT cohort, CAA was present in 38.0%; dementia occurred in 53.7% with CAA versus 40.1% without (age/sex-adjusted OR 1.57, 95% CI 1.18–2.10), but the association disappeared after adjustment for neuritic plaques or Braak stage (Sin 2024, PMID 39481068). In 483 Beijing brain-bank donors, 11% had CAA without AD, 16% AD without CAA and 20% both (Yin 2025, PMID 40201593). The first study suggests mediation by AD neuropathology; the second finds worse cognition with combined pathology. Neither establishes CAA as merely a passenger.

What this page does not claim

  • It does not claim that vascular disease causes Alzheimer's disease. The midlife amyloid association is observational with a 23-year gap (PMID 28399252).
  • It does not claim that AD and vascular dementia are the same condition. They differ in criteria, cognitive profile, imaging and trial pipeline — the companion condition's planned nosology-and-diagnostic-criteria.md will carry that comparison.
  • It does not claim that "pure" AD is rare in absolute terms. It is uncommon in community autopsy series of older people; younger and autosomal-dominant cases are different populations (see genetics).

Open questions

  • If AD explains 22–100% of cognitive loss depending on the person, how should trials measure treatment effect in a population whose remaining pathology is unmeasured (Boyle 2018, PMID 29244218)?
  • Is the midlife-but-not-late-life vascular association with amyloid causal, and if so, what is the mechanism and the latest useful intervention point (Gottesman 2017, PMID 28399252)?
  • Does APOE4-associated BBB breakdown precede, accompany or follow amyloid accumulation, and is sPDGFRβ actionable as a treatment target (Montagne 2020, PMID 32376954; Nation 2019, PMID 30643288)?
  • Why did intensive BP lowering reduce MCI but produce slightly greater hippocampal volume loss — is the volume change ischaemic, haemodynamic, or an artefact of tissue-water change (SPRINT MIND 2019, PMID 30688979; Nasrallah 2021, PMID 33683313)?
  • What proportion of amyloid-positive people with "AD" have a dementia syndrome actually driven by LATE-NC, and can LATE be identified in life (Nelson 2019, PMID 31039256)?
  • Should CAA burden be a formal eligibility variable for anti-amyloid therapy rather than only a monitoring variable, given the ARIA neuropathology (Charidimou 2022, PMID 35841910; Boon 2025, PMID 41109234)?
  • Does diabetes raise AD risk through vascular injury, insulin signalling, or shared upstream determinants — the direct tests of insulin delivery and of GLP-1 agonism in established AD were null on their primaries (Arvanitakis 2004, PMID 15148141; Craft 2020, PMID 32568367; Edison 2026, PMID 41326666)?

In this condition

Companion condition

  • Vascular dementia — master index
  • mixed-pathology-and-alzheimer-overlap.md — the reciprocal page, planned but not yet written.
  • cerebral-amyloid-angiopathy.md — planned but not yet written.
  • cerebral-small-vessel-disease.md — planned but not yet written.
  • Stroke — master index — small-vessel disease and post-stroke outcomes.

References

  1. Schneider JA, et al. Mixed brain pathologies account for most dementia cases in community-dwelling older persons. Neurology. 2007;69:2197-204. PMID 17568013.
  2. Boyle PA, et al. Person-specific contribution of neuropathologies to cognitive loss in old age. Ann Neurol. 2018;83:74-83. PMID 29244218.
  3. Kapasi A, et al. Impact of multiple pathologies on the threshold for clinically overt dementia. Acta Neuropathol. 2017;134:171-186. PMID 28488154.
  4. Nelson PT, et al. Limbic-predominant age-related TDP-43 encephalopathy (LATE): consensus working group report. Brain. 2019;142:1503-1527. PMID 31039256.
  5. Richardson TE, et al. Neocortical tau burden determines the degree of cognitive impairment in individuals with Braak stage V neurofibrillary degeneration. Acta Neuropathol. 2026;151. PMID 42184025.
  6. Charidimou A, et al. The Boston criteria version 2.0 for cerebral amyloid angiopathy: a multicentre, retrospective, MRI-neuropathology diagnostic accuracy study. Lancet Neurol. 2022;21:714-725. PMID 35841910.
  7. Gottesman RF, et al. Association between midlife vascular risk factors and estimated brain amyloid deposition. JAMA. 2017;317:1443-1450. PMID 28399252.
  8. Nation DA, et al. Blood-brain barrier breakdown is an early biomarker of human cognitive dysfunction. Nat Med. 2019;25:270-276. PMID 30643288.
  9. Montagne A, et al. APOE4 leads to blood-brain barrier dysfunction predicting cognitive decline. Nature. 2020;581:71-76. PMID 32376954.
  10. Shirzadi Z, et al. Vascular contributions to cognitive decline: beyond amyloid and tau in the Harvard Aging Brain Study. J Cereb Blood Flow Metab. 2024;44:1319-1328. PMID 38452039.
  11. SPRINT MIND Investigators for the SPRINT Research Group. Effect of intensive vs standard blood pressure control on probable dementia: a randomized clinical trial. JAMA. 2019;321:553-561. PMID 30688979.
  12. Nasrallah IM, et al. Association of intensive vs standard blood pressure control with magnetic resonance imaging biomarkers of Alzheimer disease: secondary analysis of the SPRINT MIND randomized trial. JAMA Neurol. 2021;78:568-577. PMID 33683313.
  13. Hampel H, et al. Amyloid-related imaging abnormalities (ARIA): radiological, biological and clinical characteristics. Brain. 2023;146:4414-4424. PMID 37280110.
  14. Boon BDC, et al. Neuropathological changes and amyloid-related imaging abnormalities in Alzheimer's disease treated with aducanumab versus untreated: a retrospective case-control study. Lancet Neurol. 2025;24:931-944. PMID 41109234.
  15. Kuźma E, et al. Stroke and dementia risk: a systematic review and meta-analysis. Alzheimers Dement. 2018;14:1416-1426. PMID 30177276.
  16. Arvanitakis Z, et al. Diabetes mellitus and risk of Alzheimer disease and decline in cognitive function. Arch Neurol. 2004;61:661-6. PMID 15148141.
  17. Craft S, et al. Safety, efficacy, and feasibility of intranasal insulin for the treatment of mild cognitive impairment and Alzheimer disease dementia: a randomized clinical trial. JAMA Neurol. 2020;77:1099-1109. PMID 32568367.
  18. Skoog I, et al. Decreasing prevalence of dementia in 85-year olds examined 22 years apart. Sci Rep. 2017;7:6136. PMID 28733627.
  19. Edison P, et al. Liraglutide in mild to moderate Alzheimer's disease: a phase 2b clinical trial. Nat Med. 2026;32:353-361. PMID 41326666.
  20. Whitmer RA, et al. Obesity in middle age and future risk of dementia: a 27 year longitudinal population based study. BMJ. 2005;330:1360. PMID 15863436.
  21. Miners JS, et al. CSF evidence of pericyte damage in Alzheimer's disease is associated with markers of blood-brain barrier dysfunction and disease pathology. Alzheimers Res Ther. 2019;11:81. PMID 31521199.
  22. Lv X, et al. Changes in CSF sPDGFRβ level and their association with blood-brain barrier breakdown in Alzheimer's disease with or without small cerebrovascular lesions. Alzheimers Res Ther. 2023;15:51. PMID 36915135.
  23. Van Hulle C, et al. Elevated CSF angiopoietin-2 correlates with blood-brain barrier leakiness and markers of neuronal injury in early Alzheimer's disease. Transl Psychiatry. 2024;14:3. PMID 38182581.
  24. Sin MK, et al. Cerebral amyloid angiopathy, dementia, and Alzheimer neuropathologic changes: findings from the ACT autopsy cohort. Neurology. 2024;103:e210009. PMID 39481068.
  25. Yin XS, et al. Prevalence of cerebral amyloid angiopathy and its correlation with Alzheimer's disease and cognition in an autopsy-confirmed cohort from China. Alzheimers Dement (Amst). 2025;17:e70100. PMID 40201593.