Monogenic cerebral small-vessel diseases: diagnosis and therapy. Consensus recommendations of the European Academy of Neurology

Phase I: pre-meeting

Endorsed by the Neurogenetics and Stroke Panels of the European Academy of Neurology, two facilitators (M.M. and H.S.M.) invited experts from established centres of excellence in cSVD within Europe. All experts are opinion leaders in the fields of neurogenetics, neuroimaging and cSVD. The experts decided to focus on the following selected monogenic cSVDs and working groups were created: CADASIL (led by S.L.O., with A.Be., A.Fe., H.C. and H.S.M.), cerebral autosomal recessive arteriopathy with subcortical infarcts and leukoencephalopathy (CARASIL) (led by A.Be., with A.Fe., H.S.M. and E.T.-L.), cathepsin-A-related arteriopathy with strokes and leukoencephalopathy (CARASAL) (J.F.), pontine autosomal dominant microangiopathy and leukoencephalopathy (PADMAL) (led by H.C., with A.Be. and E.T.-L.), FD (led by A.Bu., with J.F. and M.A.), mitochondrial encephalopathy, lactic acidosis and stroke-like episodes (MELAS) (led by M.M., with J.F.), type IV collagen (COL4)A1/2 (led by H.C., with A.Be. and E.T.-L.) and retinal vasculopathy with cerebral leukoencephalopathy and systemic manifestations (RVCL-S) (D.H.).

Each working group leader executed a comprehensive search strategy to November 2018 in the MEDLINE database. Only English language studies were included. Each leader screened citations, reviewed full-text papers for inclusion, appraised study quality and abstracted the data. Consensus was achieved for inclusion of papers by discussion within working groups (Fig. 1). Finally, each group provided a list of queries to be voted on, as well as the references used to address the specific disease (File S1), which was distributed to all experts.

The two facilitators created a survey (File S2) to gauge the level of consensus among experts; responses and votes were collected anonymously through the Google forms platform (https://docs.google.com/forms/u/0/) and analysed prior to the face-to-face meeting.

Participants voted using a five-point Likert scale 1 to indicate their level of agreement on each statement (1, absolutely disagree; 2, disagree; 3, no judgement; 4, more than agree; 5, absolutely agree). A ‘strong consensus’ was defined if >70% of scores were ≥4 and the mean score was ≥4. If only one of these two parameters was met, the consensus was considered a ‘good consensus’. If both parameters were not met, the statement was considered to lack consensus agreement.

Phase II: Delphi panel

The 14 experts met in a face-to-face meeting in Vienna from 28 February to 1 March 2019. A delegate from the Lily Foundation participated in the meeting as patient advocate, without voting.

M.M. introduced the workshop aims. H.S.M. gave a brief overview on proposed red flags suggesting a diagnosis of monogenic cSVD. M.M. presented results from Phase I. Statements from Phase I without consensus were selected for discussion. The working group leaders presented individual diseases, discussing results of the first round of the survey and presenting the available data supporting the queries. After discussion, when required, statements from the first round were modified or deleted and participants voted again on the revised statements. All participants were involved in all Delphi phases.

Statements were divided into three main areas: (i) identification of monogenic cSVD; (ii) clinical and features and diagnosis; and (iii) management.

Clues for diagnosis of monogenic cerebral small-vessel disease

Experts were asked to identify ‘red-flag’ features that suggested to them that a diagnosis of monogenic disease should be considered in a patient with cSVD. Strong consensus was reached on the following: family history, young age at onset, consanguinity, suggestive extracerebral/systemic features and clinical and neuroimaging characteristics of a specific monogenic disease. There was strong consensus that, even in patients with conventional vascular risk factors, a diagnosis of monogenic cSVD should not be excluded. Negative workup for other causes of cSVD reached good consensus.

A number of specific causes of cSVD were individually considered. These, with the associated key genetic, clinical and neuroimaging features are shown in Table 2.

Management recommended for all monogenic cerebral small-vessel diseases

General principles for all monogenic cSVDs, discussed and elaborated during the face-to-face meeting, are listed in Table 1.

Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy

CADASIL, caused by NOTCH3 gene variants 2, is the most prevalent monogenic cSVD, with a minimal estimated prevalence of 2–4:100 000 3, although the true prevalence is likely to be much higher 4. NOTCH3 is a single-pass receptor protein that plays an important role in vascular smooth muscle cell differentiation and maturation, and mutant NOTCH3 protein deposits are found in small arteries and capillaries throughout the body. However, despite this widespread vascular pathology, clinical manifestations are only related to the central nervous system, in the form of migraine with aura, acute encephalopathy, recurrent lacunar strokes, cognitive decline, gait and mood disorders 5, 6. Most strokes are ischaemic, but intracerebral haemorrhage can rarely occur, especially in patients treated with anticoagulants 6, 7.

Stroke onset was classically described to be between the ages of 40 and 60 years, with a severe progression to a mute and bedridden state after 10–20 years. However, individuals with milder disease are increasingly reported 4. NOTCH3 variants leading to a cysteine alteration in one of the NOTCH3 epidermal growth factor-like repeat (EGFr) domains 1–6 have been recently reported to predispose to an earlier onset of stroke than variants located in one of the EGFr domains 7–34 4. Classical vascular risk factors, such as smoking and hypertension, have been associated with earlier age of stroke onset 8.

The magnetic resonance imaging (MRI) features are central to diagnosis and features include symmetrical and progressive subcortical white matter hyperintensities (WMHs), lacunes, microbleeds and enlarged perivascular spaces (Fig. 2a) 9. WMHs are widespread. Anterior temporal lobe WMHs are highly suggestive and often considered a radiological hallmark for CADASIL, but can be lacking in some cases and are also seen in CARASIL and High Temperature Requirement A Serine Peptidase 1 (HTRA1)-autosomal dominant disease.

Pathological hallmarks of the disease are small-vessel wall abnormalities, which typically include an increased and granular pattern of NOTCH3 staining using an antibody with an epitope in the extracellular domain and pathognomonic granular osmiophilic material (GOM) adjacent to mural cells using electron microscopy 10. There is usually a family history, but this may not be apparent due to mild disease course or misdiagnosis in other family members or occasionally due to a de-novo variant.

CADASIL is confirmed by genetic testing, revealing a NOTCH3 variant altering the number of cysteines in one of the 34 EGFr domains of the NOTCH3 protein encoded by exons 2–24, thereby disrupting normal disulphide bridge formation and resulting in a pro-aggregatory mutant NOTCH3 protein that sequesters extracellular matrix proteins.

There are few data on management of CADASIL and a paucity of randomized controlled trials. A multicentre trial of the cholinesterase inhibitor donepezil showed no improvement in the primary endpoint of the V-ADAS-cog score in CADASIL with cognitive impairment 11.

Consensus statement results

Cerebral autosomal recessive arteriopathy with subcortical infarcts and leukoencephalopathy

CARASIL is a rare autosomal recessive disorder caused by biallelic variants in the HTRA1 gene on chromosome 10q26 13, 14, mainly reported in Japan but also in Chinese and Caucasian cases 15, 16. HTRA1 encodes a homotrimeric serine protease that probably has a role in maintaining cerebral small-vessel integrity by regulating transforming growth factor beta signalling.

Recurrent lacunar stroke with onset from 20 to 44 years, early vascular dementia (usually by age 30–40 years), gait disturbance, depression, premature head alopecia in 90% of patients during adolescence and severe back pain with lumbar disc herniation are the main clinical features. Additional manifestations include seizures, psychiatric disturbances, pseudobulbar palsy and spondylosis deformans. The disease is rapidly progressive with stepwise deterioration of motor and cognitive function, and the average illness duration is 20–30 years, although most patients became bedridden within 10 years after onset 16. T2-weighted cerebral MRI by age 20 years usually shows diffuse and symmetrical WMHs and lacunar infarcts (Fig. 3) 17. WMHs sometimes involve the temporal lobes as in CADASIL.

In contrast to CADASIL, histological changes are not disease specific and are characterized by arteriolosclerosis and loss of vascular smooth muscle cells with subsequent hyalinization, intima fibrous thickening and internal elastic lamina disruption. GOM deposits have not been found.

No effective treatment of CARASIL is available.

Consensus statement results

HTRA1-autosomal dominant disease

In addition to biallelic HTRA1 variants that cause CARASIL, heterozygous HTRA1 variants have been shown to cause an autosomal dominant cSVD18. These dominant negative variants lead to decreased residual activity of the HTRA1 trimer through different mechanisms 19. Clinical features differ from CARASIL by a later age of stroke onset (around 60 years) and less frequent alopecia and spondylosis deformans (Fig. 4). Cognitive impairment and encephalopathy also occur. Histological findings demonstrate diffuse and focally intensive myelin pallor in the white matter sparing U-fibres, resembling CARASIL 19.

Not all rare missense HTRA1 variants are pathogenic. In order to establish pathogenicity of a HTRA1 variant, the activity of the mutated enzyme and its ability to form trimers should be evaluated, as well as whether it segregates with disease within a family.

Collagen 4A1 and collagen 4A2 microangiopathy

Type IV collagen is a component of basement membranes and expressed ubiquitously in blood vessels and organs throughout the body. Missense or null variants in COL4A1 and COL4A2 cause an autosomal dominant cSVD of the haemorrhagic type that can be associated with extracerebral clinical manifestations 20. These two genes encode the α1 and α2 chains of COL4 and share a common locus on chromosome 13 (13q34). Most pathogenic variants of COL4A1/2 lead to either substitution of a glycine with a different amino acid or haploinsufficiency.

The clinical spectrum associated with pathogenic variants of monogenic COL4A1/2 is highly variable 21, 22. Mean age at stroke onset is 30–40 years but intracerebral haemorrhage can occur in fetuses, children and adults, and has been associated with physical activity, trauma and anticoagulant therapy. Therefore, it is recommended that a carrier should avoid hard physical activity and anticoagulants. Due to risk of haemorrhage, a fetus carrying a pathogenic COL4A1/2 variant should be delivered by Caesarean section. Haemorrhagic strokes are reported twice as often as ischaemic strokes. COL4A1/2 present in the paediatric and neonatal period with porencephaly or schizencephaly. Other manifestations include: intracerebral aneurysms, proteinuria, renal insufficiency, renal and hepatic cysts, tortuosities of retinal arteries, focal retinal haemorrhages, and early cataract or abnormal development of the anterior segment of the eye 23, 24.

The clinical phenotype can vary widely between different affected family members.

Brain imaging may show intracerebral haemorrhages or cerebral microbleeds, WMHs and lacunes that are usually subcortical (Figs 5 and 6) 25.

Missense variants of COL4A1 that disrupt the NC1 domain seem to be associated with antenatal cerebral haemorrhages and porencephaly, but not with kidney disease. Missense variants of COL4A1 that encode the CB3 (IV) segment of the triple helical domain (exons 24 and 25) have been associated with hereditary angiopathy, nephropathy, aneurysms and cramps (HANAC) syndrome 24.

In contrast with COL4A1/A2 microangiopathies associated with either glycine or stop codon variants, PADMAL is an autosomal dominant ischaemic cSVD caused by variants in the 3’ untranslated region of COL4A1. These variants affect a binding site of a microRNA, miR29, and lead to upregulated expression of COL4A1. PADMAL clinical features are characterized by adult early-onset pons infarcts and early death 26. No clinical manifestations have been reported so far in children.

Consensus statement results

Cathepsin-A-related arteriopathy with strokes and leukoencephalopathy

CARASAL is an adult, ultra-rare, hereditary cSVD reported in only 19 patients 27-29, caused by a dominant variant in the CTSA gene, located on chromosome 20q13.12 27, with the variant c.973C> T detected in 14/19 patients 27, 28. CTSA encodes the serine carboxypeptidase cathepsin-A, predominantly located in lysosomes where it stabilizes a multi-enzyme complex built up of beta-galactosidase and neuraminidase, and degrades endothelin-1 27.

CARASAL manifests clinically as ischaemic or haemorrhagic stroke, cognitive impairment, dementia, headache, migraine, vertigo, movement disorders including, rarely, dystonia, gait disturbance or depression. Some patients present with extra-central nervous system manifestations, such as dysphagia, dry mouth, dry eyes, facial pain, muscle cramps, arterial hypertension or diabetes 27-29. Cerebral imaging shows leukoencephalopathy involving the brainstem (pyramidal tracts, tegmental tracts, middle and superior cerebellar peduncles) and the subcortical white matter sparing the U-fibres 27. Grey matter, including the thalamus, basal ganglia and dentate nuclei can be also affected 28. Clinical manifestations can be absent or mild also in cases with extensive leukoencephalopathy.

There is no causal treatment available and no evidence that antithrombotic medication, anticoagulation or thrombolysis is indicated.

Consensus statement results

Retinal vasculopathy with cerebral leukoencephalopathy and systemic manifestations

RVCL-S is a genetic multi-organ small-vessel disease. The brain and retina are almost always affected, but microangiopathic disease can develop in other organs, most notably the kidneys and liver. The terms cerebroretinal vasculopathy, hereditary vascular retinopathy and hereditary endotheliopathy, retinopathy, nephropathy and stroke (HERNS) were initially used to describe affected families. However, the identification of a common genetic basis in 2007 suggested a single microvascular disorder, subsequently termed RVCL-S 30. RVCL-S is an autosomal dominant disease caused by heterozygous C-terminal frameshift variants in the TREX1 gene 30. Although the biological functions of TREX1 are increasingly well described, the exact molecular mechanism whereby RVCL-S variants cause disease is unknown. TREX1 is a negative regulator of innate immune pathways such as the type I interferon response, but prominent activation of this pathway is not typically observed in RVCL-S 31.

RVCL-S is a very rare disease and our understanding of clinical phenotype is largely based on details from an international study of 11 unrelated families with 78 variant carriers 32. The clinical picture is usually dominated by microangiopathy of the retina and brain. Disease onset is typically in middle age 32. Ophthalmological examination shows progressive retinal vasculopathy in almost all symptomatic patients. Clinical manifestations of the cerebral microangiopathy often develop in the fourth to fifth decade. Although the majority of patients develop cognitive impairment, brain disease can manifest as a large enhancing lesion, which can mimic tumours with pronounced mass effect or tumefactive inflammation 32. Focal deficits can be associated with the development of contrast-enhancing lesions on MRI. These can become large and confluent, sometimes with perilesional oedema (Fig. 7). Calcifications in the white matter detected on computed tomography are common and were identified in 52% of mutation carriers 32. Lacunar stroke is not a common manifestation and systematic screenings for TREX1 in 1000 young-onset lacunar stroke patients detected no TREX1 pathogenic variants 33.

Microvascular disease affecting other organs is commonly seen and can be severe, mainly in kidney and liver, and can cause organ failure. Raynaud’s disease, gastrointestinal bleeding, bone disease and thyroid disease are also seen.

Consensus statement results

Fabry disease

Fabry disease is an inherited metabolic disease that results from a deficiency or absence of the enzyme α-galactosidase A due to a variant in the GLA gene. FD is inherited as an X-linked trait. Male patients can develop a classic severe phenotype of the disease (type 1), whereas heterozygous female patients present with phenotypes ranging from asymptomatic to severe involvement of vital organs. The classic forms exhibit early manifestations, whereas later-onset forms (type 2) can present clinical signs later in life. About 900 variants, most of them private, have been reported 34. FD is a lysosomal storage disease, due to accumulation of the sphingolipid globotriaosylceramide (GL-3 or Gb3) and its deacetylated derivative lyso-globotriaosylceramide (lyso-GL-3 or lyso-Gb3).

Many different cell types are involved in FD, including vascular cells (endothelial cells and vascular smooth muscle cells), cardiomyocytes, renal cells, enteric neurons and nerve cells 34. Therefore, heart, kidney, brain, peripheral nervous system and the gastrointestinal system are heavily involved in the disease 34, 35.

There is a causal relationship between FD and ischaemic stroke, but the underlying mechanisms are still unclear. Ischaemic stroke is the most common subtype, but intracerebral haemorrhages and cerebral venous thrombosis have been reported 36. The posterior circulation seems to be predominantly affected. The most common age of stroke onset is 20–50 years 36. WMHs seen on T2-weighted MRI are a common finding in patients with FD, present in up to 46% (Fig. 8) 37, and progression of WMHs occurs in a significant proportion of them. Other neuroradiological features are basilar artery dolichoectasia and the pulvinar sign 36.

Fabry disease explains less than 1% of cryptogenic ischaemic strokes in young adults. An early study 38 suggested that FD might account for 4.9% of ischaemic stroke in young male patients but this has not been confirmed in replication studies 39, 40.

Cognitive function has been less studied than other neurological complications in FD.

Besides symptomatic drugs, specific therapies are available. The latter include enzymatic replacement therapy (ERT) (agalsidase beta or alfa) and chaperone therapy for amenable variants. Recent recommendations regarding diagnosis, monitoring and treatment have been published 35.

Consensus statement results

Mitochondrial encephalopathy, lactic acidosis and stroke-like episodes

MELAS is one of the most severe syndromes in mitochondrial disease. Although the majority of patients (80%) have the common MELAS pathogenic variant m.3243A>G in the mitochondrial DNA tRNA-leucine (MT-TL1) 43, the number of genotypes causing the MELAS phenotype continues to expand and may include nuclear genes, mainly POLG 44.

One of the clinical hallmarks of MELAS, the stroke-like episode (SLE), is characterized by headache, nausea and vomiting, encephalopathy, focal-onset seizures and focal neurological deficits. On conventional MRI, the affected areas do not correspond to vascular territories and involve the cortex and juxtacortical white matter (Fig. 9). Lesions may spread over days, weeks and even months or go to the contralateral side after the acute episode. Some lesions reverse completely with time; however, most severe lesions develop into cortical laminar necrosis, gliosis and atrophy.

Although SLEs were originally described in individuals<40 years of age, late-onset presentation is recognized. Although SLEs cannot be strictly classified as cSVD, they have to be considered in the differential diagnosis of cSVD and therefore MELAS was included in the consensus statement.

Additional common clinical and laboratory findings in MELAS are eyelid ptosis, cardio(myo)pathy, muscle weakness, diabetes and lactic acidosis.

Compared with strokes of vascular origin, there is higher incidence of bilateral clinical symptoms such as cortical blindness or auditory agnosia.

The underlying pathophysiological mechanism of SLEs remains unclear 45. Limitations in the understanding of the disease mechanism are reflected in the controversies about the acute management of SLEs.

Consensus statement results