Disease Mechanism PROVEN
Small Vessel Disease (SVD) — Core Mechanism
Arteriosclerosis of perforating arteries (100–400μm). Lipohyalinosis, fibrinoid necrosis. Feeds basal ganglia, deep white matter, thalamus, brainstem. Unlike large vessel disease — strokes are lacunar (small cavity infarcts).
Wardlaw et al. 2013 Lancet Neurol (SVD neuroimaging standards)
White Matter Hyperintensities (WMH) = Leukoaraiosis
T2-FLAIR bright areas. Periventricular + deep white matter. Represent demyelination, axon loss, gliosis from chronic ischaemia. Fazekas scale grades severity. More WMH = more cognitive impairment.
Fazekas et al. 1987 · Wardlaw et al. 2013
Lacunar Infarcts — Basal Ganglia + White Matter
3–15mm cavity infarcts. Putamen, caudate, thalamus, internal capsule. Step-wise cognitive decline with each lacunar event — unlike gradual AD decline. Executive function, processing speed, gait affected first.
Roman et al. 1993 Neurology (NINDS-AIREN criteria)
CADASIL — Genetic Pure SVD
NOTCH3 mutation → EGF-like domain accumulation → vascular smooth muscle degeneration. Pure SVD without traditional vascular risk factors. Onset 30s–40s. Migraines with aura precede dementia by years. 100% penetrance for stroke/dementia.
Joutel et al. 1996 Nature (NOTCH3 discovery) · CADASIL = Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy
VD vs AD Distinction — SORRY: Often Impossible
Mixed dementia (VD + AD) common in 80+. NINDS-AIREN criteria require temporal relationship between stroke and cognitive decline — often unclear. Autopsy shows mixed pathology in majority of "pure VD" cases diagnosed clinically.
SORRY: No pure biomarker separates VD from AD in most elderly patients
Vascular Risk Factors
Hypertension = most important modifiable risk factor. BP control PROVEN to slow WMH progression.
Diabetes mellitus: Advanced glycation damages vessel walls. Doubles VD risk.
Smoking: Direct endothelial damage. Accelerates SVD.
Hyperlipidaemia: Atherosclerosis of feeding vessels. Statin benefit debated for SVD specifically.
Disease-Modifying NONE APPROVED
No VD-Specific DMT Exists
SORRY: Despite decades of trials, no drug has been approved specifically for vascular dementia. The field's biggest gap. ChEIs show modest benefit in some trials — not approved indication. Memantine trials inconclusive for VD.
Cochrane review 2021: insufficient evidence for any pharmacological intervention in VD
Risk Factor Control PROVEN
BP Control — ONLY Proven Intervention
Antihypertensives slow WMH progression. PROGRESS trial: perindopril + indapamide reduced recurrent stroke + cognitive decline. Target BP <130/80 in SVD. The only intervention with Level A evidence.
PROGRESS Collaborative Group 2001 Lancet · Lithell et al. 2003 (SCOPE trial)
Anticoagulation — Cardioembolic Only, NOT SVD
Warfarin/DOACs for AF-related cardioembolic strokes. CONTRAINDICATED in SVD with microbleeds — increases haemorrhage risk. Critical distinction: anticoagulation doesn't help lacunar SVD and may cause harm.
Hart et al. 2014 (microbleed-anticoagulation risk)
Aspirin — Cardioembolic Benefit, Uncertain SVD
Antiplatelet for non-cardioembolic stroke secondary prevention. In pure SVD — benefit unclear, bleeding risk real. CADASIL: aspirin often used empirically.
SORRY: SVD-specific antiplatelet evidence weak
Statins OPEN
Large artery atherosclerosis benefit proven. For pure SVD (lipohyalinosis mechanism): statins may not help the dominant SVD pathway. Pleiotropic anti-inflammatory effects theorised.
HPS 2002 (benefit for large vessel) · SORRY: SVD-specific statin evidence not established
MRI Biomarkers PROVEN
| WMH Volume PROVEN | T2-FLAIR. Fazekas scale 0–3. Best predictor of cognitive decline in SVD. Automated volumetrics now standard. Wardlaw 2013 |
| SVD Score | Composite: WMH + lacunes + microbleeds + enlarged perivascular spaces. 0–4. Predicts dementia risk + mortality. Staals et al. 2014 |
| Lacune count | MRI-visible lacunar infarcts. Each lacune = step-wise decline risk. Basal ganglia/thalamus lacunes = executive function impact. |
| Microbleeds | T2* GRE / susceptibility-weighted. >5 microbleeds = high haemorrhagic risk → avoid anticoagulation. |
| Perivascular spaces | Enlarged PVS (Virchow-Robin spaces) = glymphatic failure, SVD severity marker. |
Blood / CSF Biomarkers
| NfL plasma↑ | Neurofilament light. Non-specific axonal damage marker. Elevated in VD, AD, ALS. Tracks disease burden. SORRY: not VD-specific |
| NOTCH3 mutation | CADASIL genetic confirmation. Blood test. 100% specific for CADASIL when EGFD variant confirmed. GOM deposits on skin biopsy (electron microscopy). PROVEN |
| GFAP↑ | Astrocyte activation. Elevated in WMH progression. Research biomarker. |
Open Sorries
SORRY-1: WMH → cognitive decline causality debated. Many elderly with significant WMH have normal cognition. Threshold unclear.
SORRY-2: VD vs AD distinction often impossible in clinical practice — mixed dementia is the rule, not exception, in 80+.
SORRY-3: CADASIL penetrance is 100% for stroke/dementia but timing is variable. No prediction model for onset.
SORRY-4: No VD-specific biomarker. All current markers are indirect surrogates.
EEG in Vascular Dementia
| Focal slowing | Delta/theta over infarct zones. Reflects focal ischaemia. Localises to lesion territory. Useful in acute setting. SORRY: Not specific for VD type |
| Theta excess | Diffuse frontal theta. Reflects white matter disconnection. Less characteristic than AD theta changes. |
| Alpha coherence ↓ | Reduced interhemispheric coherence. Disconnection signature of white matter damage. |
| EEG non-specificity | SORRY: VD EEG changes are less characteristic than AD, LBD, or FTD. Clinical utility limited. No pathognomonic EEG signature. |
EEG Utility Caveat
VD EEG is the least informative among the major dementias. It reflects the underlying ischaemia but doesn't distinguish VD from other causes of brain injury. Evoked potentials (VEP, SSEP) may show conduction slowing but are rarely used clinically for VD diagnosis. MRI remains primary.
Novel Patterns (7)
PROVENNP-VD-001 · SVD score (0–4) composite predicts dementia risk + mortality better than individual markers
PROVENNP-VD-002 · BP control = ONLY intervention with Level A evidence for slowing VD progression · all others SORRY
PROVENNP-VD-003 · CADASIL NOTCH3 mutation: pure SVD mechanism, genetic, 100% penetrance, no traditional risk factors needed
PROVENNP-VD-004 · Microbleeds >5: anticoagulation contraindicated — haemorrhagic transformation risk outweighs embolic benefit
OPENNP-VD-005 · Enlarged perivascular spaces = glymphatic failure marker in SVD · mechanistic link to WMH accumulation
OPENNP-VD-006 · Step-wise vs gradual decline: VD step-wise, AD gradual — but mixed dementia blurs this in clinical practice
SORRYNP-VD-007 · No VD-specific DMT after decades of trials · WMH causality debated · mixed dementia = norm in 80+ · CADASIL penetrance timing unpredictable