Skip to content

Monogenic and inherited forms

TL;DR — Monogenic cerebral arteriopathies are rare but causally informative forms of vascular cognitive impairment. CADASIL from pathogenic NOTCH3 variants is the best characterized, combining migraine with aura, recurrent subcortical stroke, mood symptoms, WMH, and cognitive decline (Chabriat 2009, PMID 19539236; Joutel 1996, PMID 8878478). HTRA1-related disease, COL4A1/2 vasculopathy, and Fabry disease differ in inheritance, systemic clues, haemorrhage risk, and treatment implications (Choi 2015, PMID 25692103; Whittaker 2022, PMID 35699195). Genetic testing should be phenotype- and counseling-led because variants of uncertain significance, a 40% de novo rate in one COL4A1/2 series, and incomplete penetrance can mislead (Meuwissen 2015, PMID 25719457). These disorders are models of vessel injury, not miniature versions of all sporadic SVD.

Core disorders

Disorder Gene / inheritance Vascular phenotype Extra-neurological clues
CADASIL NOTCH3, autosomal dominant WMH, lacunes, migraine, subcortical stroke family history may be obscured (Chabriat 2009, PMID 19539236)
CARASIL HTRA1, usually biallelic early severe SVD alopecia, spondylosis
HTRA1 dominant SVD HTRA1, autosomal dominant later/milder SVD spectrum variable penetrance
COL4A1/2 disorder COL4A1 or COL4A2, dominant porencephaly, ICH, SVD ocular, renal, muscle features
Fabry disease GLA, X-linked stroke and white-matter disease renal, cardiac, neuropathic pain, angiokeratoma
TREX1-related RVCL TREX1, dominant retinal/cerebral small-vessel disease retinal, renal, hepatic involvement
ADA2 deficiency ADA2, recessive inflammatory/vasculitic phenotype systemic inflammation, young stroke

This table is a clinical map, not an exhaustive gene panel. Systematic review of 390 COL4A1, 41 COL4A2, 123 TREX1, 44 homozygous and 82 heterozygous HTRA1, 346 ADA2, and 14 CTSA carriers showed stroke frequencies from 9% (TREX1) to 52% (heterozygous HTRA1) and cognitive features from 0% (ADA2) to 64% (homozygous HTRA1) (Whittaker 2022, PMID 35699195).

CADASIL

NOTCH3 mutations that alter cysteine residues in the extracellular domain cause CADASIL (Joutel 1996, PMID 8878478). MRI often shows early temporal-pole and external-capsule WMH, followed by lacunes and microbleeds; these patterns support but do not replace molecular diagnosis (Chabriat 2009, PMID 19539236).

In 147 consecutive CADASIL patients, lacunar-lesion volume associated independently with global cognitive scores and disability, whereas WMH volume did not; disability also associated with microbleeds, systolic BP, and age (Viswanathan 2007, PMID 17620550). CADASIL is therefore a model in which lacunes, not WMH load alone, track function.

Longitudinal data give that cross-sectional finding a direction. Twenty-five NOTCH3 mutation carriers and 13 controls had standardized neuropsychology and MRI at baseline and after 7 years: carriers declined in global cognition (CAMCOG, P<0.01) and in language, memory and executive function, and the decline — executive dysfunction especially — was associated with increases in lacunar infarcts, microbleeds and ventricular volume, while WMH change and brain atrophy were not associated with cognitive decline at all (Liem 2009, PMID 19139365). In a young, mildly affected population, therefore, the marker most used as a trial endpoint is the marker with no measured relationship to the outcome.

The shape of that decline is not linear, which matters for trial duration. Multivariate latent-process mixed modelling of MMSE and Mattis Dementia Rating Scale trajectories in 185 CADASIL patients followed at the French national referral centre found performance falling with age mainly after 50, unaffected by sex or education (though higher education raised the baseline level), with both scales showing ceiling effects and curvilinearity that make them poor longitudinal instruments (Brice 2020, PMID 32804128). A trial enrolling CADASIL patients in their forties on an MMSE endpoint is measuring the flat part of a curve with a ceiling-limited instrument — which is a second, independent explanation for the null donepezil result above, alongside the mechanistic one.

How common is CADASIL? Clinical versus genetic prevalence differ ~40-fold

Measure Estimate Population/method
Genetically confirmed clinical CADASIL 1.98 per 100,000 adults (95% CI 1.24–3.00) west of Scotland register, 22 individuals from 7 pedigrees, 2002 census denominator (Razvi 2005, PMID 15834040)
Predicted NOTCH3 mutation carriers (clinical + at-risk pedigree members) 4.14 per 100,000 adults (3.04–5.53) same register, adding 37 predicted carriers (Razvi 2005, PMID 15834040)
Clinical prevalence, later Glasgow series 4.6 per 100,000 adults; mutation prevalence 10.7 per 100,000 49 pedigrees, 21 distinct NOTCH3 mutations, 61% in exon 4 (Moreton 2014, PMID 24840674)
Genetic prevalence from population sequencing 1 in 277 individuals (≈361 per 100,000) 129,933 healthy Korean adults screened for NOTCH3 p.Arg544Cys (allele frequency 0.07%), p.Arg640Cys (0.07%), and p.Arg75Pro (0.04%) (Cho 2025, PMID 40982447)
Genetic prevalence, UK exome sequencing 1 in 450 (443 carriers of 67 distinct cysteine-altering variants) 200,632 UK Biobank participants with whole-exome sequencing (Cho 2021, PMID 33712516)

The gap between roughly 2–11 per 100,000 clinically and roughly 361 per 100,000 genetically is the single most important fact about NOTCH3 disease and is not an artifact of ascertainment alone. It means most people carrying a cysteine-altering NOTCH3 variant will never be diagnosed with CADASIL, and it reframes NOTCH3-SVD as a common low-penetrance modifier of sporadic small-vessel disease rather than a rare Mendelian disease — with an obvious corollary that population-scale genomic screening will surface carriers for whom the classical CADASIL prognosis is simply wrong. The Korean estimate is also population-specific by construction (three variants, one ancestry) and should not be transported globally.

The clinical picture has drifted milder as ascertainment has broadened. Median age at first stroke in women was 57 years, and in men rose from 46 years in patients diagnosed before 2006 to 56 years after 2006 (P=0.034); among patients over 58, 13/34 (38%) were living independently and 17/28 (61%) mobile without aids at last review (Moreton 2014, PMID 24840674). Descriptions of near-universal dementia and disability from the fifth decade come from the era of severe-case ascertainment.

What carriers in the general population actually experience

Two UK Biobank analyses convert the prevalence gap into risk estimates, and they do not show nonpenetrance.

Among 200,632 exome-sequenced participants, the 443 cysteine-altering NOTCH3 carriers had adjusted stroke OR 2.33 (P=0.0004) and vascular dementia OR 5.00 (P=0.007), higher WMH volume (standardized difference 0.52, P<0.001) and worse diffusion-MRI white-matter microstructure (0.72, P<0.001). On blinded visual rating of 47 carriers against 148 matched controls, carriers had more lacunes (OR 5.97) and microbleeds (OR 4.38), with the WMH excess concentrated exactly where CADASIL puts it — anterior temporal lobes (OR 7.65) and external capsule (OR 13.32), all P<0.001 (Cho 2021, PMID 33712516). The imaging signature is CADASIL's; only the severity is milder.

Extending to 454,756 participants and all three common monogenic SVD genes gave carrier counts of 973 (NOTCH3), 546 (HTRA1) and 336 (COL4A1/2), with each group at least 66% more likely to have had stroke, and gene-specific risks that track the gene's known biology: vascular dementia OR 5.42 (95% CI 3.11–8.74) for NOTCH3, all-cause dementia OR 2.17 (1.28–3.41) for HTRA1, and intracerebral haemorrhage OR 3.56 (1.34–7.53) for COL4A1/2 (Cho 2022, PMID 36300346). The finding with the most practical consequence is that Framingham cardiovascular risk burden was associated with stroke risk in NOTCH3 and HTRA1 carriers — penetrance is partly modifiable, which turns an untreatable-sounding genetic diagnosis into an argument for aggressive conventional risk-factor control. No trial has yet tested that inference in carriers.

Ancestry matters for which variant is found. Exome-wide association in 75,000 Pakistanis identified NOTCH3 p.Arg1231Cys at an alternate allele frequency of 0.58%, against 0.019% in western Europeans — a 30-fold enrichment — associated with subcortical haemorrhagic stroke (OR 3.39, 95% CI 2.26–5.10, P=3.9×10⁻⁹) and all stroke (OR 2.30, 1.77–3.01, P=7.8×10⁻¹⁰), and with WMH in UK Biobank (1.1 SD, 0.61–1.5). It is attributable for roughly 2.0% of haemorrhagic and 1.1% of all strokes in South Asians (Rodriguez-Flores 2024, PMID 39271666). Panels and prevalence estimates built on European cohorts will under-call this variant.

Staging the spectrum

Because the phenotype now runs from end-stage CADASIL to nonpenetrance, a severity scale was needed and did not exist. The NOTCH3-SVD staging system defines 9 stages/substages from 0 (premanifest) to 4B (end stage), developed in 195 cases and validated in 1,713 from 23 international cohorts across 15 countries plus 101 UK Biobank carriers. 1,789 of 1,908 cases (94%) followed the defined sequence of events, and the stages tracked ischaemic stroke, intracerebral haemorrhage, global cognition, processing speed, brain volume, microstructural damage, serum NfL, and 18-year survival (Gravesteijn 2025, PMID 39610302). This is the first instrument in this knowledge base that stages a vascular cognitive disorder by course rather than by threshold, and it exists because a monogenic cause removes the attribution problem that blocks the same approach in sporadic disease.

The one adequately powered drug trial in a homogeneous vascular population

CADASIL was chosen for a cholinesterase-inhibitor trial precisely because it removes the two confounders that make vascular-dementia trials uninterpretable: mechanism heterogeneity, and Alzheimer copathology in older participants. In 168 patients (mean age 54.8 years) randomized to donepezil 10 mg or placebo for 18 weeks, the primary endpoint — change in V-ADAS-Cog — was flatly null: least-squares mean change −0.85 (SE 0.57) with donepezil versus −0.81 (SE 0.59) with placebo, P=0.956. Three executive secondary endpoints favoured donepezil (Trail Making B P=0.023, Trail Making A P=0.015, EXIT25 P=0.022), and the authors declined to claim clinical relevance for them (Dichgans 2008, PMID 18296124, NCT00103948). This result is load-bearing for treatment: the cleanest available test of the cholinergic hypothesis in pure subcortical vascular disease was negative on its primary endpoint, which is a reason to read the small pooled ADAS-Cog effects in mixed VaD populations sceptically rather than as an underestimate.

Genotype predicts severity: the three-tiered EGFr model

CADASIL severity is not uniform across NOTCH3 cysteine-altering variants, and the variance is now partly predictable from variant position. Using relative variant frequencies between 2,574 CADASIL patients and 1,647 individuals from population databases, each of the 34 EGF-like repeat domains was classified as low-, medium-, or high-risk, then validated in 434 CADASIL patients and 1,003 NOTCH3-cysteine-positive community-dwelling individuals carrying 379 distinct variants (Hack 2023, PMID 36535904).

Risk tier EGFr domains Stroke odds in community-dwelling carriers
High risk 1–6 plus 8, 11, and 26 (9 domains) OR 10.81 (95% CI 5.46–21.37)
Medium risk 10 domains OR 1.81 (95% CI 0.84–3.88)
Low risk 11 domains 1 (reference)

Medium-risk carriers had higher normalized WMH volume (P=0.005) and PSMD (P=0.035) than low-risk carriers; within CADASIL patients, the newly identified high-risk domains 8, 11, and 26 carried stroke risk (P=0.002), disability (P=0.041), WMH volume (P=1.8×10⁻⁸), PSMD (P=2.6×10⁻⁸), and lacune volume (P=0.006) indistinguishable from the classical domains 1–6. Mechanistically, the frequency-derived risk metric tracked vascular NOTCH3 aggregation load (P=0.006) but not NOTCH3 signalling activity (P=0.88) — evidence for a toxic-aggregation model rather than a loss-of-signalling model, and therefore an argument that aggregate clearance, not Notch pathway restoration, is the therapeutic target. The dichotomy it replaced was itself quantitative and remains the reference point for how large the position effect is: comparing 664 CADASIL patients by variant position, EGFr 1–6 carriers had first stroke 12 years earlier, lower survival, and higher WMH volumes than EGFr 7–34 carriers, and while 70% of diagnosed CADASIL patients carry an EGFr 1–6 variant, EGFr 7–34 variants strongly predominate in the general population (Rutten 2019, PMID 30032161). That inversion — the severe variants dominate the clinic, the mild variants dominate the population — is the mechanism generating the 40-fold clinical-versus-genetic prevalence gap above, and it means ascertainment bias in CADASIL is not a nuisance but a structured, quantifiable consequence of variant distribution.

The pre-2023 dichotomy of "EGFr 1–6 versus 7–34" is superseded.

Feature Typical role Caveat
Migraine with aura may precede stroke common in general population
Subcortical ischaemic events central morbidity frequency varies by family
Mood/apathy recognized phenotype nonspecific
Executive slowing tracks accumulated injury Alzheimer copathology still possible
Temporal-pole WMH useful imaging clue not universally present
Skin biopsy can show characteristic deposits molecular testing usually central

Risk management generally extrapolates from sporadic vascular prevention because disease-modifying randomized evidence is limited. Two CADASIL patients treated with intravenous alteplase in a 2023 report had no haemorrhagic complications and good outcomes; a literature review to August 2022 found only case-level data, so efficacy and bleeding risk versus sporadic stroke remain unquantified by trial (Pescini 2023, PMID 36255541). A 2024 multicenter research letter on CADASIL thrombolysis exists (Chen 2024, PMID 39474712) but does not replace randomized evidence. Antithrombotic use must still account for indication and microbleed burden.

Biallelic HTRA1 variants classically cause CARASIL, while heterozygous variants can produce autosomal-dominant SVD with variable age and severity. Whittaker et al. recorded mean age 59 years in heterozygous HTRA1 carriers versus younger biallelic disease, with stroke in 52% of heterozygotes and cognitive features in 64% of homozygotes (Whittaker 2022, PMID 35699195; Choi 2015, PMID 25692103).

Direct comparison with CADASIL is now available and shows the two dominant arteriopathies are separable on imaging, not just on gene. Twenty-one unrelated Taiwanese carriers of 15 heterozygous HTRA1 variants — all functionally validated by in-vitro protease assay — were compared with 406 CADASIL patients (44 high-risk EGFr, 358 moderate-risk, 4 low-risk).

Feature Heterozygous HTRA1 (n=21) NOTCH3 high-risk EGFr (n=44) NOTCH3 moderate-risk EGFr (n=358)
Mean age at symptom onset 54.3 ± 10.7 y 49.2 ± 10.5 y 59.7 ± 9.5 y
Stroke 81.0% — —
Cognitive dysfunction 47.6% — —
Lacunes 90.5% — —
≥10 cerebral microbleeds 66.7% — —
Intracerebral haemorrhage lesions 46.7% — —
Temporal-pole WMH involvement 28.6% 88.6% 32.4%

Temporal-pole WMH severity remained significantly milder in HTRA1 carriers than in high-risk-EGFr CADASIL after adjustment for age and hypertension (P<0.001), and ICH lesions were more frequent in HTRA1 disease. Within HTRA1, genotype tracked bleeding: loss-of-function or protease-domain missense variants gave ≥10 microbleeds in 100% and 83.3% of carriers respectively, against 20% for missense variants outside the protease domain (Lee 2026, PMID 40607620). The practical reading is that the classical temporal-pole sign is a NOTCH3-high-risk sign rather than a monogenic-SVD sign, and that a haemorrhage-predominant dominant arteriopathy should raise HTRA1 before NOTCH3. The cohort is small, single-ancestry, and retrospective.

The TGF-β direction is disputed

CARASIL was mapped to HTRA1 by linkage in five families to a 2.4-Mb region of chromosome 10q, with two nonsense and two missense mutations producing protein with low protease activity; the arteries of affected people showed increased extra-domain-A fibronectin and versican in a thickened intima and increased TGF-β1 in the media. The original interpretation was that HtrA1 represses TGF-β family signalling, so loss of function causes increased signalling (Hara 2009, PMID 19387015). Five years later, work in mouse brain, embryonic fibroblasts and patient skin fibroblasts identified latent TGF-β binding protein 1 (LTBP-1) as an HtrA1 substrate: cleavage occurs at physiological protease concentrations, is prevented by CARASIL mutations, and disrupts LTBP-1 binding to fibronectin and its incorporation into the matrix — making HtrA1 a facilitator of TGF-β signalling, so that loss of function causes attenuated signalling (Beaufort 2014, PMID 25369932).

Hara 2009 (PMID 19387015) Beaufort 2014 (PMID 25369932)
HtrA1's normal role represses TGF-β signalling facilitates TGF-β signalling via LTBP-1 processing
Effect of CARASIL mutation TGF-β signalling increased TGF-β signalling attenuated
Principal evidence TGF-β1 immunostaining in patient cerebral arteries; in-vitro repression assays LTBP-1 identified as substrate; cleavage lost in mutants and in HtrA1-deficient cells

Both are cited as the mechanism of CARASIL and the direction has not been settled by a decisive experiment. It is not a semantic dispute: a TGF-β-lowering and a TGF-β-raising therapeutic strategy are opposite interventions, so no rational drug target follows from the mechanism until the direction is fixed. Detailed features of the recessive phenotype, including the alopecia and spondylosis that distinguish it, are catalogued in Nozaki 2014 (PMID 25116877).

COL4A1 and COL4A2

Type IV collagen variants affect basement membranes. Since 2005 they have been recognized as a monogenic SVD cause spanning antenatal/perinatal haemorrhage, porencephaly, adult ICH, lacunar disease, and microbleeds (Lanfranconi 2010, PMID 20558831). In 52 mutation carriers reviewed to 2010, stroke had occurred in 9 (17.3%; 6 subcortical haemorrhages, 3 lacunar infarcts) at mean onset 36.1 years (range 14–49). Imaging showed leukoaraiosis in 63.5%, microbleeds in 52.9%, lacunar infarction in 13.5%, and intracranial aneurysms in 44.4% of 18 angiogrammed carriers. Haemorrhages were associated with trauma, activity, and anticoagulant therapy (Lanfranconi 2010, PMID 20558831).

The phenotype extends well beyond stroke, and epilepsy rather than cerebrovascular disease is the presenting problem in an appreciable subgroup. Across 44 new and 55 previously reported COL4A1/2 patients, childhood-onset focal seizures — often complicated by status epilepticus and drug resistance — were the commonest phenotype; in 46.4% of new patients with focal seizures the porencephalic cyst colocalized with the focal epileptiform discharges, and a distinct subgroup had epilepsy as the main feature with only nonspecific MRI (periventricular leukoencephalopathy, ventricular asymmetry). Analysis of 15 pedigrees suggested worsening severity in successive generations, particularly with maternal inheritance, and no clear genotype–phenotype correlation emerged — epilepsy-associated variants were spread across COL4A1 (Zagaglia 2018, PMID 30413629). For this knowledge base the consequence is that a COL4A1/2 diagnosis reached through an adult cerebrovascular route describes only one end of the disorder, and that the absence of genotype–phenotype structure here contrasts sharply with the strong position effect in NOTCH3.

HANAC syndrome (hereditary angiopathy, nephropathy, aneurysms, and muscle cramps) is a COL4A1 glycine-mutation phenotype with hematuria, large renal cysts, and both small- and large-artery involvement (Plaisier 2007, PMID 18160688). In a diagnostic series of 183 index tests, 21 COL4A1 and 3 COL4A2 mutations were identified, with a 40% (10/24) de novo rate (Meuwissen 2015, PMID 25719457). Pregnancy-specific risk magnitude is not established by a dedicated trial; the systematic-review association of haemorrhage with trauma and anticoagulation is the quantitative caution available as of 2026-08-31.

Fabry disease

Fabry disease is an X-linked lysosomal glycosphingolipid disorder with cardiac, renal, peripheral-nerve, skin, and cerebrovascular manifestations, including increased stroke incidence attributable in part to endothelial dysfunction (Tuttolomondo 2013, PMID 23448451).

How often it explains a young stroke is now measured rather than asserted. The prospective multinational sifap study enrolled 5,023 patients aged 18–55 with ischaemic stroke (3,396), haemorrhagic stroke (271) or TIA (1,071) across 47 centres in 15 European countries under a standardized protocol: definite Fabry disease in 0.5% (95% CI 0.4–0.8, n=27) and probable Fabry disease in a further 0.4% (Rolfs 2013, PMID 23306324, NCT00414583). Restricting to cryptogenic stroke roughly doubles the yield: pooled prevalence across 16 studies and 7,048 patients was 1.3% (95% CI 0.75–2.32, I²=56.4%), with male sex a significant predictor on meta-regression, and the classical systemic red flags present in a minority even of confirmed cases — hypohidrosis 15.9%, acroparaesthesia 8.9%, pain crises 6.0%, angiokeratoma 3.7%, cornea verticillata 1.9% (Magalhães 2025, PMID 41123702). Waiting for the dermatological signs before testing will therefore miss most cases. Identification matters because enzyme replacement and chaperone strategies exist for selected genotypes/phenotypes. The nearest randomized evidence addresses white-matter lesions rather than cognition, and it is weak. In a phase-4 placebo-controlled analysis of agalsidase beta 1 mg/kg biweekly in Fabry patients with mild-to-moderate renal involvement, 41 patients (mean age 43.9; 25 on ERT, 16 on placebo) had FLAIR MRI at baseline and after a mean 27 months. White-matter lesion burden was present in 63% at baseline and increased over follow-up, correlating with left-ventricular posterior wall thickness but not with renal dysfunction; patients with previous or recent stroke showed lesion increase (P=0.005). Among patients aged ≤50, 44% (8/18) on ERT versus 31% (4/13) on placebo had stable lesion burden (P=0.014), number needed to treat 8 (Fellgiebel 2014, PMID 25502511). The comparison is a subgroup of a small trial, the endpoint is manually measured lesion diameters, and a "stable" proportion of 44% is not disease arrest. It supports early treatment as a hypothesis; it does not establish that enzyme replacement prevents vascular cognitive decline. A live 2026-09-02 PubMed search identified no randomized trial with a cognitive endpoint in Fabry disease; cognitive prevention therefore remains an evidence gap rather than a proven indication.

Retinal vasculopathy with cerebral leukoencephalopathy and systemic manifestations unifies three syndromes once described separately, all caused by C-terminal frameshift TREX1 mutations. In 78 members of 11 unrelated families carrying five different mutations, 64 had vascular retinopathy; neuroimaging showed punctate hyperintense white-matter lesions with or without nodular enhancement in 97%, rim-enhancing mass lesions in 84% — the feature most often mistaken for tumour or demyelination — and white-matter calcification in 52%. Brain involvement was clinical in 90%: focal deficits 68%, migraine 59%, cognitive impairment 56%, psychiatric disturbance 42%, seizures 17%. Systemic features were the rule rather than the exception: liver disease 78%, anaemia 74%, nephropathy 61%, hypertension 60%, Raynaud phenomenon 40%, gastrointestinal bleeding 27%. Mean age at diagnosis was 42.9 ± 8.3 years and at death 53.1 ± 9.6 (Stam 2016, PMID 27604306). The 13 carriers without retinopathy or brain lesions were on average 8 years younger, consistent with high penetrance and age-dependent expression rather than true nonpenetrance — the opposite of the NOTCH3 picture above.

When to suspect an inherited arteriopathy

Trigger Why it raises suspicion
Stroke/SVD at unusually young age low baseline probability of sporadic age-related disease
Multigenerational similar disease dominant inheritance pattern
Consanguinity/siblings affected recessive possibility
Migraine plus characteristic MRI CADASIL pattern
Alopecia/spondylosis CARASIL clue
Ocular/renal/muscle abnormalities COL4A1/2 clue
Renal/cardiac disease plus neuropathic pain Fabry clue
Retinal vasculopathy/systemic organ disease TREX1-related disease

Absence of family history does not exclude a genetic disorder because of de novo variants, small families, early deaths, non-paternity, sex-linked expression, and incomplete penetrance.

Testing workflow

  1. Confirm the imaging phenotype using standardized terminology.
  2. Construct a three-generation vascular, migraine, psychiatric, renal, cardiac, ocular, and movement history.
  3. Examine for systemic clues.
  4. Use targeted testing when phenotype is specific; broader panels when overlap is substantial.
  5. Confirm variant classification and segregation where feasible.
  6. Provide pre- and post-test genetic counseling.
  7. Separate diagnostic testing from predictive testing in unaffected relatives.

What inherited disease teaches—and does not

Translation opportunity Boundary
Known initiating lesion enables natural-history cohorts Sporadic SVD has multiple exposures
Younger cohorts reduce Alzheimer copathology Developmental and lifelong effects differ
Lacune volume tracks cognition in CADASIL WMH burden is not equivalent across diseases
Biomarkers can be measured before symptoms Small populations constrain trials
Targeted molecular therapy is conceivable Benefit may not generalize

Open questions

  • If WMH change is unassociated with cognitive decline in CADASIL over 7 years while lacune, microbleed and ventricular change are, why is WMH volume still the default monogenic-SVD trial endpoint? (Liem 2009, PMID 19139365)
  • Does the post-50 inflection in the CADASIL cognitive trajectory mean trials in younger carriers are underpowered by design rather than by sample size? (Brice 2020, PMID 32804128)
  • Should a haemorrhage-predominant dominant arteriopathy without temporal-pole WMH prompt HTRA1 testing before NOTCH3? (Lee 2026, PMID 40607620)
  • Why does NOTCH3 show one of the strongest genotype–phenotype position effects in neurogenetics while COL4A1/2 shows essentially none? (Rutten 2019, PMID 30032161; Hack 2023, PMID 36535904; Zagaglia 2018, PMID 30413629)
  • Does enzyme replacement in Fabry disease alter cognitive trajectory, or only white-matter lesion counts in a subgroup? (Fellgiebel 2014, PMID 25502511)
  • Which molecular subgroups of HTRA1 disease have actionable progression markers? (Whittaker 2022, PMID 35699195)
  • Can presymptomatic CADASIL trials use lacune volume rather than WMH as a surrogate? (Viswanathan 2007, PMID 17620550)
  • How should variant-specific COL4A1/2 haemorrhage risk alter antithrombotic and pregnancy counseling? (Lanfranconi 2010, PMID 20558831)
  • Does enzyme replacement change cerebrovascular cognitive outcomes in Fabry disease? (Tuttolomondo 2013, PMID 23448451)
  • If genetic NOTCH3-variant prevalence is ~1 in 277 while clinical CADASIL prevalence is ~2–5 per 100,000, what determines penetrance? (Cho 2025, PMID 40982447; Razvi 2005, PMID 15834040)
  • Should the three-tiered EGFr classification replace the EGFr 1–6 versus 7–34 dichotomy in counselling and trial eligibility? (Hack 2023, PMID 36535904)
  • Does the association of risk with NOTCH3 aggregation load but not signalling activity mean aggregate clearance is the correct drug target? (Hack 2023, PMID 36535904)
  • How should incidentally identified NOTCH3 carriers from population sequencing be counselled, given that classical CADASIL prognosis is derived from severe-case ascertainment? (Moreton 2014, PMID 24840674)
  • Does the HtrA1 protease raise or lower TGF-β signalling — and which direction should a CARASIL drug take? (Hara 2009, PMID 19387015; Beaufort 2014, PMID 25369932)
  • Would intensive conventional vascular risk-factor control reduce stroke in NOTCH3 and HTRA1 carriers, as the penetrance-modification finding predicts? (Cho 2022, PMID 36300346)
  • Should NOTCH3 p.Arg1231Cys be included in stroke gene panels used for South Asian populations, where it explains ~2% of haemorrhagic strokes? (Rodriguez-Flores 2024, PMID 39271666)
  • Can the NOTCH3-SVD staging system be adapted to sporadic small-vessel disease, or does attribution uncertainty make course-based staging impossible there? (Gravesteijn 2025, PMID 39610302)
  • Given a null primary endpoint in the cleanest possible subcortical vascular population, is the cholinergic hypothesis in VCI still viable? (Dichgans 2008, PMID 18296124)
  • Should cryptogenic stroke in young men trigger Fabry testing regardless of systemic signs, given that under 16% of confirmed cases had hypohidrosis? (Magalhães 2025, PMID 41123702; Rolfs 2013, PMID 23306324)

References

  1. Chabriat H, et al. CADASIL. Lancet Neurol. 2009;8:643-53. PMID 19539236
  2. Joutel A, et al. Notch3 mutations in CADASIL. Nature. 1996;383:707-10. PMID 8878478
  3. Choi JC. Genetics of cerebral small vessel disease. J Stroke. 2015;17:7-16. PMID 25692103
  4. Whittaker E, et al. Cerebral phenotypes associated with monogenic cerebral small-vessel disease. J Am Heart Assoc. 2022;11:e025629. PMID 35699195
  5. Viswanathan A, et al. Lacunar lesions are independently associated with disability and cognitive impairment in CADASIL. Neurology. 2007;69:172-9. PMID 17620550
  6. Pescini F, et al. Intravenous thrombolysis in CADASIL. Neurol Sci. 2023;44:491-498. PMID 36255541
  7. Chen CH, et al. Safety and effectiveness of intravenous thrombolysis in patients with CADASIL. Stroke. 2024;55:e321-e322. PMID 39474712
  8. Lanfranconi S, Markus HS. COL4A1 mutations as a monogenic cause of cerebral small vessel disease. Stroke. 2010;41:e513-8. PMID 20558831
  9. Plaisier E, et al. COL4A1 mutations and hereditary angiopathy, nephropathy, aneurysms, and muscle cramps. N Engl J Med. 2007;357:2687-95. PMID 18160688
  10. Meuwissen ME, et al. The expanding phenotype of COL4A1 and COL4A2 mutations. Genet Med. 2015;17:843-53. PMID 25719457
  11. Tuttolomondo A, et al. Anderson-Fabry disease: a multiorgan disease. Curr Pharm Des. 2013;19:5974-96. PMID 23448451
  12. Razvi SS, et al. The prevalence of cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoencephalopathy (CADASIL) in the west of Scotland. J Neurol Neurosurg Psychiatry. 2005;76:739-41. PMID 15834040
  13. Moreton FC, et al. Changing clinical patterns and increasing prevalence in CADASIL. Acta Neurol Scand. 2014;130:197-203. PMID 24840674
  14. Cho EH, et al. Genetic epidemiology of moyamoya disease and CADASIL in over 120,000 healthy Korean individuals. PLoS One. 2025;20:e0331174. PMID 40982447
  15. Hack RJ, et al. Three-tiered EGFr domain risk stratification for individualized NOTCH3-small vessel disease prediction. Brain. 2023;146:2913-2927. PMID 36535904
  16. Cho BPH, et al. NOTCH3 variants are more common than expected in the general population and associated with stroke and vascular dementia: an analysis of 200 000 participants. J Neurol Neurosurg Psychiatry. 2021;92:694-701. PMID 33712516
  17. Cho BPH, et al. Association of vascular risk factors and genetic factors with penetrance of variants causing monogenic stroke. JAMA Neurol. 2022;79:1303-1311. PMID 36300346
  18. Rodriguez-Flores JL, et al. NOTCH3 p.Arg1231Cys is markedly enriched in South Asians and associated with stroke. Nat Commun. 2024;15:8029. PMID 39271666
  19. Gravesteijn G, et al. Disease severity staging system for NOTCH3-associated small vessel disease, including CADASIL. JAMA Neurol. 2025;82:49-60. PMID 39610302
  20. Dichgans M, et al. Donepezil in patients with subcortical vascular cognitive impairment: a randomised double-blind trial in CADASIL. Lancet Neurol. 2008;7:310-8. PMID 18296124
  21. Hara K, et al. Association of HTRA1 mutations and familial ischemic cerebral small-vessel disease. N Engl J Med. 2009;360:1729-39. PMID 19387015
  22. Beaufort N, et al. Cerebral small vessel disease-related protease HtrA1 processes latent TGF-β binding protein 1 and facilitates TGF-β signaling. Proc Natl Acad Sci U S A. 2014;111:16496-501. PMID 25369932
  23. Nozaki H, et al. Features of cerebral autosomal recessive arteriopathy with subcortical infarcts and leukoencephalopathy. Stroke. 2014;45:3447-53. PMID 25116877
  24. Rolfs A, et al. Acute cerebrovascular disease in the young: the Stroke in Young Fabry Patients study. Stroke. 2013;44:340-9. PMID 23306324
  25. Magalhães PLM, et al. Prevalence of Fabry disease in cryptogenic stroke: a systematic review and meta-analysis with meta-regression. J Neurol. 2025;272:720. PMID 41123702
  26. Stam AH, et al. Retinal vasculopathy with cerebral leukoencephalopathy and systemic manifestations. Brain. 2016;139:2909-2922. PMID 27604306
  27. Rutten JW, et al. The effect of NOTCH3 pathogenic variant position on CADASIL disease severity. Genet Med. 2019;21:676-682. PMID 30032161
  28. Liem MK, et al. MRI correlates of cognitive decline in CADASIL: a 7-year follow-up study. Neurology. 2009;72:143-8. PMID 19139365
  29. Brice S, et al. Modeling the cognitive trajectory in CADASIL. J Alzheimers Dis. 2020;77:291-300. PMID 32804128
  30. Lee YC, et al. Comparison of clinical and imaging features of cerebral small vessel disease associated with heterozygous HTRA1 and NOTCH3 mutations. Int J Stroke. 2026;21:79-88. PMID 40607620
  31. Zagaglia S, et al. Neurologic phenotypes associated with COL4A1/2 mutations: expanding the spectrum of disease. Neurology. 2018;91:e2078-e2088. PMID 30413629
  32. Fellgiebel A, et al. Enzyme replacement therapy stabilized white matter lesion progression in Fabry disease. Cerebrovasc Dis. 2014;38:448-56. PMID 25502511