Skip to content

Guidelines

TL;DR — Guidelines agree more strongly on structured assessment, MRI where available, and vascular-risk management than on symptomatic cognitive drugs. VASCOG and VICCCS are diagnostic consensus frameworks; the AHA/ASA statement integrates vascular prevention; ESO addresses covert SVD; the 2024 Canadian update provides a current VCI pathway (Sachdev 2014, PMID 24632990; Skrobot 2018, PMID 29055812; Gorelick 2011, PMID 21778438; Wardlaw 2021, PMID 34414301; Swartz 2025, PMID 39822128). Cholinesterase inhibitors have at most small scale effects and recommendations differ because panels weigh uncertain clinical importance and adverse events differently (Battle 2021, PMID 33704781). CAA and ICH guidance adds safety constraints not captured by generic dementia guidance (Greenberg 2022, PMID 35579034; Cordonnier 2025, PMID 40721902).

Guideline map

Body/framework Year Scope Distinct contribution
NINDS–AIREN 1993 research diagnostic criteria high-specificity VaD definition (PMID 8094895)
ADDTC 1992 ischaemic vascular dementia mixed categories and imaging (PMID 1549205)
NINDS–CSN 2006 VCI harmonization standards 5/30/60-minute cognitive protocols (PMID 16917086)
AHA/ASA 2011 vascular contributions to cognitive impairment prevention and clinical framework; no FDA-approved VCI drug as of 2011 (PMID 21778438)
VASCOG 2014 diagnostic criteria mild/major vascular cognitive disorders (PMID 24632990)
VICCCS 2018 classification/harmonization consensus and NINDS-CSN protocols (PMID 29055812)
ESO 2021 covert SVD risk management with limited direct evidence (PMID 34414301)
ESO–EAN joint 2021 post-stroke cognitive impairment GRADE assessment of prevention, screening, treatment, prognosis; found insufficient evidence for most interventions (PMID 34746430)
Asian Clinical Expert Group on Neurocognitive Disorders 2019 EGb 761 in dementia/MCI with or without cerebrovascular disease regional consensus grading ginkgo at ChEI-equivalent strength (PMID 30648358)
Korean Dementia Association 2025 cholinesterase inhibitors and memantine across dementia types GRADE-graded, VaD-specific conditional recommendation for ChEIs (PMID 39944527)
AHA/ASA 2022 spontaneous ICH haemorrhage/antithrombotic safety (PMID 35579034)
Canadian Stroke Best Practices 2024/2025 publication VCI across care current assessment/management pathway (PMID 39822128)
International CAA Association/WSO 2025 CAA CAA diagnosis and management (PMID 40721902)
VasCog-2-WSO 2025 revised diagnosis contemporary criteria (PMID 40955506)
AHA (Hypertension) 2016 hypertension and cognition concluded insufficient data for evidence-based recommendations while endorsing judicious BP treatment (PMID 27977393)
NICE NG97 2018 dementia assessment, management, support UK national pathway; summarized for practice (PMID 29925626)
WHO 2019 (summary 2021) risk reduction of cognitive decline and dementia first global, multisectoral risk-reduction recommendations (PMID 35082745)
AHA/ASA 2021 secondary stroke prevention the recurrent-stroke prevention base on which delayed-dementia prevention rests (PMID 34024117)
AHA/ASA 2021 primary care after stroke names cognitive screening as a standing primary-care task (PMID 34261351)
AHA/ASA 2023 cognitive impairment after ischaemic and haemorrhagic stroke scoping appraisal; up to one-third develop dementia within 5 years (PMID 37125534)
Lancet standing Commission 2024 dementia prevention, intervention and care the reference point for population-level modifiable-risk policy (PMID 39096926)
European Academy of Neurology 2020 medical management issues in dementia GRADE guideline placing vascular-risk management inside dementia care; anticoagulation in dementia with AF (PMID 32713125)
American Academy of Neurology 2018 (retired) mild cognitive impairment the only guideline giving exercise a graded recommendation over drugs in MCI (PMID 29282327)
American College of Radiology 2024 update imaging appropriateness in dementia scenario-by-scenario imaging selection, including pre/post anti-amyloid therapy (PMID 40409878)
Alzheimer's Association DETeCD-ADRD 2025 diagnostic evaluation, testing, counseling, disclosure structured evaluation aimed at primary care, with a validated-instrument inventory (PMID 39713939)

Areas of agreement

  • VCI spans mild to major impairment.
  • Memory impairment is not required for a vascular syndrome.
  • Diagnosis requires cognition, function, vascular brain evidence, and causal reasoning.
  • MRI is preferred where available.
  • Mixed pathology should be recognized.
  • Vascular-risk treatment is central.
  • Cognitive screening should be linked to fuller assessment and action.
  • Safety, function, mood, gait, and caregivers matter.

These themes recur across VASCOG, VICCCS, AHA/ASA, and Canadian guidance (Sachdev 2014, PMID 24632990; Skrobot 2018, PMID 29055812; Gorelick 2011, PMID 21778438; Swartz 2025, PMID 39822128).

Areas of disagreement or variable emphasis

Topic Position range Evidence driver
diagnostic label VaD vs major vascular NCD/VCI nosology and communication
imaging minimum preferred MRI vs setting-dependent access and sensitivity
cholinesterase inhibitors not routine to considered selectively small cognitive effects (Battle 2021, PMID 33704781)
memantine insufficient/routine-not-supported vs selected mixed cases inconsistent functional evidence
aspirin for covert SVD generally not solely for WMH bleeding and no direct indication
BP intensity broad vascular targets vs individualized caution small cognitive benefit and orthostasis
AD biomarkers recommended selectively vs not routinely accessible diagnostic value, cost, availability

Where guidelines actually contradict each other, with the grades attached

Three documents assess overlapping evidence and reach different operational conclusions. The disagreement is real, and it survives being read carefully.

Question ESO–EAN post-stroke cognitive impairment, 2021 (GRADE) ESO covert SVD, 2021 (GRADE) Korean Dementia Association CPG, 2025 (GRADE) Asian Clinical Expert Group, 2019 (WFSBP grading)
Cholinesterase inhibitors insufficient evidence to make a recommendation (Quinn 2021, PMID 34746430) no evidence for conventional Alzheimer dementia treatments in covert SVD (Wardlaw 2021, PMID 34414301) conditionally recommended for vascular dementia (strongly recommended for Alzheimer disease and for Lewy body dementia) (Kim 2025, PMID 39944527) positioned as one option among several (Kandiah 2019, PMID 30648358)
Memantine / nootropics insufficient evidence to recommend (Quinn 2021, PMID 34746430) not addressed for covert SVD strongly recommended for moderate-to-severe Alzheimer disease; VaD not given a separate strong recommendation (Kim 2025, PMID 39944527) grouped with ChEIs as comparators to EGb 761 (Kandiah 2019, PMID 30648358)
Ginkgo EGb 761 not recommended (no evidence base cited) not addressed not addressed Grade 3 recommendation, Level B evidence — the same strength the WFSBP assigns to cholinesterase inhibitors and memantine, and the only agent with Level B evidence for cognition, behaviour, and ADL in both AD and vascular dementia (Kandiah 2019, PMID 30648358)
Cognitive rehabilitation insufficient evidence to recommend (Quinn 2021, PMID 34746430) — — —
Routine cognitive screening after stroke no evidence found for routine screening; targeted assessment endorsed; no screening test clearly superior (Quinn 2021, PMID 34746430) — — —
Blood pressure lifestyle and vascular risk treatment may have many health benefits but cognitive benefit is not proven (Quinn 2021, PMID 34746430) recommend well-controlled BP in ccSVD with hypertension; lower targets may reduce ccSVD progression (Wardlaw 2021, PMID 34414301) — —
Aspirin / antiplatelets for covert SVD — do not recommend (Wardlaw 2021, PMID 34414301) — —
Lipid lowering — little evidence in ccSVD (Wardlaw 2021, PMID 34414301) — —
Glucose lowering without diabetes — no evidence (Wardlaw 2021, PMID 34414301) — —
Exercise — recommended; may benefit cognition (Wardlaw 2021, PMID 34414301) — —
Imaging prediction acute-imaging associations with PSCI mostly unclear, except that substantial WMH of presumed vascular origin on acute MRI may help predict cognitive outcome (Quinn 2021, PMID 34746430) — — —

The EGb 761 row is the sharpest live disagreement in this field's guideline literature. A regional expert consensus places ginkgo at the same evidence grade as cholinesterase inhibitors and memantine and notes it is uniquely graded Level B across cognition, behaviour, and ADL in both AD and VaD (Kandiah 2019, PMID 30648358); the Cochrane evidence base for the same agent in dementia reports 6-month cognitive and ADL benefits with I² of 91–96% and low certainty (Wieland 2026, PMID 41641880); and the European guideline does not recommend it at all (Quinn 2021, PMID 34746430). The three positions are not reconcilable by reading more papers — they reflect different thresholds for how much heterogeneity a recommendation can tolerate, and different regional trial literatures (see treatment).

The second real contradiction is quieter but more consequential for practice: ESO–EAN concludes that vascular risk treatment has unproven cognitive benefit after stroke, while ESO's covert-SVD guideline recommends blood-pressure control specifically to slow SVD progression (Quinn 2021, PMID 34746430; Wardlaw 2021, PMID 34414301). Both were written under GRADE, in the same year, by overlapping communities. The difference is the endpoint — imaging progression versus cognition — and it is the same imaging-versus-clinical gap that runs through the prevention and biomarkers pages.

Both European guidelines were explicit that their evidence base was thin: ESO–EAN reported "limited randomised controlled trial evidence" across prevention, treatment, and prediction and called for definitive RCTs, and the covert-SVD guideline found "little direct evidence, mostly of low quality," naming randomized trials with clinical endpoints as the field's priority (Quinn 2021, PMID 34746430; Wardlaw 2021, PMID 34414301). Guidance in vascular cognitive impairment is therefore mostly expert consensus wearing GRADE labels, and should be read as such.

Where responsibility for vascular cognition is actually assigned

The dementia-guideline literature and the stroke-guideline literature both claim this territory, and they place the work in different services.

The AHA/ASA scientific statement on post-stroke cognitive impairment sits closest to this condition. Its scoping appraisal states that cognitive impairment is common after stroke, especially in the first year, is reversible in some cases early, and that up to one-third of individuals with stroke develop dementia within 5 years, with pathophysiology framed as an acute stroke precipitating pathological events on a background of pre-existing microvascular and neurodegenerative change. It calls for interdisciplinary management and screening for associated comorbidities, and names the two gaps most relevant here: prospective study of individual PSCI trajectories, and high-quality randomized trials of PSCI management (El Husseini 2023, PMID 37125534). Read next to ESO–EAN's "insufficient evidence to recommend" conclusions, the two documents differ less in what they find than in whether they frame the absence as a reason to withhold recommendations or as a reason to organize services anyway.

The primary-care statement makes the same organizational point operationally: screening for depression, cognitive impairment and fall risk is placed as both a short-term and a long-term component of post-stroke care delivered by primary-care teams, alongside risk-factor management (Kernan 2021, PMID 34261351). The secondary-prevention guideline supplies the recurrent-stroke evidence base on which any claim about preventing delayed dementia depends (Kleindorfer 2021, PMID 34024117); the causal step from "prevents recurrent stroke" to "prevents dementia" is the one no guideline yet certifies — see prevention.

Two population-level documents bound the prevention claim from outside stroke medicine. The AHA statement on hypertension and cognition reviewed the evidence and concluded explicitly that there were insufficient data to make evidence-based recommendations, while judging that treating hypertension is justified on vascular grounds; it separates strong evidence for a deleterious effect of midlife hypertension on late-life cognition from a much less clear effect of late-life hypertension (Iadecola 2016, PMID 27977393). The WHO guidelines on risk reduction of cognitive decline and dementia — the first global recommendations of their kind, developed for a policy audience across lifestyle and behavioural interventions, physical-health conditions, and specific interventions, with implementation and equity considerations — provide the corresponding public-health frame for the majority of people with dementia, who live in low- and middle-income countries (Chowdhary 2021, PMID 35082745). The Lancet standing Commission's 2024 report is the most cited synthesis of modifiable dementia risk and is the document national policy most often references (Livingston 2024, PMID 39096926).

The UK national pathway (NICE NG97) covers dementia assessment, management and support generically rather than by subtype (Pink 2018, PMID 29925626); the absence of vascular-specific content in a major national guideline is itself the implementation gap that the Canadian VCI update and the AHA statement were written to fill.

The two guidelines that recommend exercise, and what that implies

Across the whole table above, exercise is recommended more strongly than any drug. The ESO covert-SVD guideline recommends it and notes it may benefit cognition (Wardlaw 2021, PMID 34414301). The AAN mild-cognitive-impairment guideline goes further and is unusually explicit about the ordering: clinicians should recommend regular exercise (Level B) and may choose not to offer cholinesterase inhibitors (Level B), and if they do offer them they must first discuss the lack of evidence (Level A) — the only Level A recommendation in that document being a disclosure requirement about a drug's weakness. The same guideline supplies the age-stratified MCI prevalence used across this knowledge base (6.7% at 60–64 rising to 25.2% at 80–84) and a 2-year cumulative dementia incidence of 14.9% in MCI over 65 (Petersen 2018, PMID 29282327). It is formally retired and has not been replaced. Two randomized VCI-specific exercise trials report estimates that numerically overlap those of cholinesterase inhibitors, but no head-to-head comparison exists and the trial designs differ (see treatment).

Vascular risk management inside dementia care, not only before it

Most of the guidance above treats vascular risk as prevention. The European Academy of Neurology guideline addresses it as management of established dementia, and its recommendations are graded accordingly: systematic management of vascular risk factors should be performed in patients with mild to moderate dementia, because preventing cerebrovascular pathology may affect dementia progression (Good Practice statement); individuals with dementia and atrial fibrillation but no previous stroke should be anticoagulated (weak recommendation); atypical antipsychotics for agitation or aggression only after non-pharmacological measures have failed or in cases of severe harm (weak recommendation), with discontinuation once symptoms settle; newer anticonvulsants first-line for epilepsy in dementia; and regular preplanned medical follow-up for all patients, minimally through general practice with ready access to specialists (Frederiksen 2020, PMID 32713125). This is the only guideline in the table that assigns anticoagulation a recommendation in the population where microbleed-related bleeding risk is highest, and it does so without conditioning on imaging — a gap that the CAA and ICH documents fill from the other direction.

Imaging selection has a dedicated guideline

Which scan to order is normally left implicit in dementia guidance. The ACR Appropriateness Criteria address it directly and by scenario — initial imaging in suspected mild cognitive impairment, in suspected specific dementia syndromes including vascular dementia and normal-pressure hydrocephalus, in rapidly progressive dementia, and pre- and post-treatment imaging for patients receiving anti-amyloid monoclonal antibodies — using GRADE-adapted evidence assessment supplemented by the RAND/UCLA appropriateness method where the literature is equivocal (ACR 2025, PMID 40409878). The anti-amyloid monitoring scenario is where this document intersects this condition most sharply, because the imaging findings it governs are the CAA markers that also define vascular disease.

The Alzheimer's Association DETeCD-ADRD guideline is the complementary process document, built for primary care from a modified Delphi review of 7,374 publications (133 meeting inclusion criteria), providing a structured evaluation pathway and an inventory of validated instruments for cognition, mood, behaviour and daily function (Atri 2025, PMID 39713939). Its relevance here is negative as well as positive: a US guideline written to be usable by non-specialists for "AD and related dementias" is where most vascular cognitive impairment will actually be evaluated, and vascular-specific content within it is minimal.

Diagnosis pathway synthesis

  1. Identify acquired cognitive and functional change.
  2. Exclude delirium and reversible contributors.
  3. Perform domain-sensitive screening/assessment.
  4. Obtain structural imaging, preferably MRI.
  5. Characterize stroke, SVD, CAA, and atrophy patterns.
  6. Evaluate vascular risks and competing etiologies.
  7. State mild/major severity and attribution confidence.
  8. Communicate uncertainty and reassess longitudinally.

Prevention synthesis

Guidelines support BP treatment and standard stroke/vascular prevention, while cognitive-specific effect estimates remain modest. Meta-analysis reports OR 0.93 (0.88–0.98) for dementia/cognitive impairment with BP lowering (Hughes 2020, PMID 32427305). ESO covert-SVD recommendations appropriately distinguish established cardiovascular indications from treating an incidental MRI marker (Wardlaw 2021, PMID 34414301).

Drug-treatment synthesis

The Cochrane network meta-analysis found statistically significant but probably clinically small benefits for donepezil and galantamine, with dose-related harms; rivastigmine evidence was uncertain (Battle 2021, PMID 33704781). Guideline differences are therefore value judgments under uncertainty, not necessarily contradictory readings of large effects.

Safety synthesis

CAA and ICH guidance changes antithrombotic decisions when lobar haemorrhage, siderosis, or numerous microbleeds are present (Greenberg 2022, PMID 35579034; Cordonnier 2025, PMID 40721902). Driving, capacity, falls, and medication management require functional assessment rather than diagnosis-based prohibition.

Implementation gaps

Gap Consequence
MRI access inequitable etiologic confidence
neuropsychology access overreliance on screens
language norms misclassification
fragmented stroke/dementia services lost follow-up
absent caregiver support recommendations cannot be implemented
inconsistent biomarker access mixed disease hidden

Open questions

  • Why does the strongest graded recommendation for exercise in cognitive impairment sit in a retired AAN guideline with no replacement? (Petersen 2018, PMID 29282327)
  • Should anticoagulation in dementia with atrial fibrillation be conditioned on microbleed burden, as CAA guidance implies but the EAN recommendation does not? (Frederiksen 2020, PMID 32713125; Greenberg 2022, PMID 35579034)
  • Does a primary-care-facing "AD and related dementias" evaluation pathway adequately capture vascular cognitive impairment? (Atri 2025, PMID 39713939)
  • Will VasCog-2-WSO improve real-world reliability? (Sachdev 2025, PMID 40955506)
  • Which guideline implementation package improves outcomes rather than documentation? (Swartz 2025, PMID 39822128)
  • How should guidelines define a worthwhile symptomatic-drug response? (Battle 2021, PMID 33704781)
  • Can MRI-dependent recommendations be adapted without sacrificing integrity? (Wardlaw 2021, PMID 34414301)
  • Why do a European guideline and an Asian expert consensus assign EGb 761 opposite recommendation status from overlapping evidence? (Quinn 2021, PMID 34746430; Kandiah 2019, PMID 30648358)
  • Should a guideline recommend blood-pressure control for an imaging endpoint when a sibling guideline calls the cognitive benefit unproven? (Wardlaw 2021, PMID 34414301; Quinn 2021, PMID 34746430)
  • Is a "conditional recommendation" for cholinesterase inhibitors in vascular dementia defensible when the same evidence yields "insufficient to recommend" elsewhere? (Kim 2025, PMID 39944527; Quinn 2021, PMID 34746430)
  • If no post-stroke cognitive screening test is clearly superior and routine screening is unsupported, what should replace it? (Quinn 2021, PMID 34746430)
  • Why does one body organize services around post-stroke cognition while another declines to recommend the same activities on the same evidence? (El Husseini 2023, PMID 37125534; Quinn 2021, PMID 34746430)
  • Should generic national dementia guidance carry vascular-specific recommendations, or is a separate VCI pathway the better instrument? (Pink 2018, PMID 29925626; Swartz 2025, PMID 39822128)
  • Does the midlife/late-life split in hypertension's cognitive effect mean prevention guidance should be age-stratified rather than target-based? (Iadecola 2016, PMID 27977393)
  • Can WHO's global risk-reduction recommendations be implemented where MRI and neuropsychology are unavailable, and what is lost if they are? (Chowdhary 2021, PMID 35082745)

References

  1. Gorelick PB, et al. Vascular contributions to cognitive impairment and dementia. Stroke. 2011. PMID 21778438
  2. Sachdev PS, et al. VASCOG criteria. Alzheimer Dis Assoc Disord. 2014. PMID 24632990
  3. Skrobot OA, et al. VICCCS guidelines. Alzheimers Dement. 2018. PMID 29055812
  4. Wardlaw JM, et al. ESO guideline on covert cerebral SVD. Eur Stroke J. 2021. PMID 34414301
  5. Greenberg SM, et al. AHA/ASA spontaneous ICH guideline. Stroke. 2022. PMID 35579034
  6. Swartz RH, et al. Canadian Stroke Best Practice Recommendations: Vascular cognitive impairment. Alzheimers Dement. 2025. PMID 39822128
  7. Cordonnier C, et al. Diagnosis and management of CAA. Int J Stroke. 2025. PMID 40721902
  8. VasCog-2-WSO Criteria Consortium. Revised Diagnostic Criteria for Vascular Cognitive Impairment and Dementia—The VasCog-2-WSO Criteria. JAMA Neurol. 2025. PMID 40955506
  9. Battle CE, et al. Cholinesterase inhibitors for VCI. Cochrane Database Syst Rev. 2021. PMID 33704781
  10. Hughes D, et al. BP lowering and cognitive outcomes. JAMA. 2020. PMID 32427305
  11. Román GC, et al. NINDS-AIREN diagnostic criteria. Neurology. 1993;43:250-60. PMID 8094895
  12. Chui HC, et al. ADDTC ischaemic vascular dementia criteria. Neurology. 1992;42:473-80. PMID 1549205
  13. Hachinski V, et al. NINDS-CSN VCI harmonization standards. Stroke. 2006;37:2220-41. PMID 16917086
  14. Quinn TJ, et al. European Stroke Organisation and European Academy of Neurology joint guidelines on post-stroke cognitive impairment. Eur Stroke J. 2021;6:I-XXXVIII. PMID 34746430
  15. Kim Y, et al. Clinical practice guidelines for dementia: recommendations for cholinesterase inhibitors and memantine. Dement Neurocogn Disord. 2025;24:1-23. PMID 39944527
  16. Kandiah N, et al. Treatment of dementia and mild cognitive impairment with or without cerebrovascular disease: expert consensus on the use of Ginkgo biloba extract, EGb 761. CNS Neurosci Ther. 2019;25:288-298. PMID 30648358
  17. Wieland LS, et al. Ginkgo biloba for cognitive impairment and dementia. Cochrane Database Syst Rev. 2026;2:CD013661. PMID 41641880
  18. El Husseini N, et al. Cognitive impairment after ischemic and hemorrhagic stroke: a scientific statement from the American Heart Association/American Stroke Association. Stroke. 2023;54:e272-e291. PMID 37125534
  19. Kernan WN, et al. Primary care of adult patients after stroke: a scientific statement from the American Heart Association/American Stroke Association. Stroke. 2021;52:e558-e571. PMID 34261351
  20. Kleindorfer DO, et al. 2021 guideline for the prevention of stroke in patients with stroke and transient ischemic attack. Stroke. 2021;52:e364-e467. PMID 34024117
  21. Iadecola C, et al. Impact of hypertension on cognitive function: a scientific statement from the American Heart Association. Hypertension. 2016;68:e67-e94. PMID 27977393
  22. Chowdhary N, et al. Reducing the risk of cognitive decline and dementia: WHO recommendations. Front Neurol. 2021;12:765584. PMID 35082745
  23. Livingston G, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. Lancet. 2024;404:572-628. PMID 39096926
  24. Pink J, et al. Dementia: assessment, management and support: summary of updated NICE guidance. BMJ. 2018;361:k2438. PMID 29925626
  25. Frederiksen KS, et al. A European Academy of Neurology guideline on medical management issues in dementia. Eur J Neurol. 2020;27:1805-1820. PMID 32713125
  26. Petersen RC, et al. Practice guideline update summary: mild cognitive impairment. Neurology. 2018;90:126-135. PMID 29282327
  27. Expert Panel on Neurological Imaging. ACR Appropriateness Criteria Dementia: 2024 update. J Am Coll Radiol. 2025;22:S202-S233. PMID 40409878
  28. Atri A, et al. The Alzheimer's Association clinical practice guideline for the diagnostic evaluation, testing, counseling, and disclosure of suspected Alzheimer's disease and related disorders (DETeCD-ADRD): validated clinical assessment instruments. Alzheimers Dement. 2025;21:e14335. PMID 39713939