The reason genetics matters disproportionately in young stroke is a simple denominator effect. In a 70-year-old, atherosclerosis, atrial fibrillation, and hypertension are so prevalent that a monogenic cause is statistically swamped. In a 30-year-old with clean vessels, a normal heart, and no risk factors, those common explanations have been stripped away — so the rarer Mendelian and metabolic causes rise to the surface. This is why the diagnostic question shifts with age: in the young cryptogenic stroke patient, "what single gene or mitochondrial defect could do this?" becomes a leading hypothesis rather than a footnote.

The practical skill is recognizing when to leave the standard stroke algorithm and reach for genetic testing. The triggers cluster into a few patterns:

• The vessels and heart are clean but the brain keeps infarcting. Recurrent strokes — especially small, deep (lacunar) infarcts — in someone without hypertension or diabetes points toward an intrinsic arteriopathy of the small vessels rather than embolism from a proximal source.
• The stroke travels with company. Stroke plus migraine, plus sensorineural hearing loss, plus a rash or angiokeratomas, plus renal disease, plus diabetes — these are not coincidences but the systemic footprint of a syndrome whose vascular component is only one organ's manifestation.
• The imaging behaves wrong. Lesions that cross vascular territory boundaries, disproportionate white matter disease for the patient's age, or microbleeds in a young person all argue that the process is not classical thromboembolism.
• The pedigree lights up. Early stroke or dementia in a parent (autosomal dominant pattern, as in CADASIL) or a string of affected relatives on the maternal line (mitochondrial inheritance, as in MELAS) reframes a "cryptogenic" event as familial.

Making the diagnosis is not academic. It redirects management — antiplatelet versus anticoagulation, enzyme replacement, seizure control, transfusion — and triggers cascade screening of at-risk relatives who can be counseled or treated before their first event.

CADASIL is the most common hereditary stroke disorder in adults, and its mechanism is worth understanding because it explains every facet of management. The disease was traced to NOTCH3 in the landmark study by Joutel et al. 1996, which found strongly stereotyped missense mutations clustered in the epidermal growth factor-like repeat (EGF-r) domain of the receptor's extracellular portion.

Why these mutations are so stereotyped — the cysteine rule. Each EGF-r repeat is folded and stabilized by three disulfide bonds formed between six precisely positioned cysteine residues. Essentially every pathogenic CADASIL variant either removes one of these cysteines or adds a new one, leaving an odd number of cysteines in the repeat. With an unpaired cysteine, the disulfide bonding goes awry, the receptor's extracellular domain misfolds, and the mutant NOTCH3 ectodomain accumulates rather than being cleared. This is the unifying biochemical lesion — and it is why a variant that simply swaps two non-cysteine residues, however rare, is usually not CADASIL.

From misfolded protein to arteriopathy. NOTCH3 is expressed by vascular smooth muscle cells and pericytes. The accumulating ectodomain deposits in the media of small arteries and arterioles, dragging other proteins (including extracellular matrix components) into aggregates. Ultrastructurally these appear as granular osmiophilic material (GOM) — the pathological fingerprint seen on skin biopsy electron microscopy. The smooth muscle cells degenerate, the vessel wall thickens and stiffens, the lumen narrows, and cerebral autoregulation fails. The downstream consequence is chronic hypoperfusion of the deep white matter and small lacunar infarcts in the territory of these end-arteries.

This mechanism dictates therapy. CADASIL is a degenerative arteriopathy, not a thrombotic or embolic disease and not a coagulopathy. There is no clot to dissolve or prevent, which is why no trial has shown anticoagulation to help and why, given the microbleeds of advanced disease, it is avoided where possible (an independent indication such as atrial fibrillation or venous thrombosis still takes precedence). Antiplatelet therapy and aggressive vascular risk-factor control are pragmatic, but no agent reverses the underlying protein aggregation — management is supportive, and the most valuable intervention is often making the diagnosis itself so that relatives can be tested and counseled.

The single most important concept in this section is that a MELAS stroke-like episode (SLE) is not a stroke. The word "stroke-like" is doing real work — it signals that the lesion resembles an infarct on imaging and at the bedside but arises from a completely different mechanism. Confusing the two leads directly to the wrong, and potentially dangerous, treatment.

Metabolic, not vascular. An ischemic stroke is a plumbing problem: a vessel occludes, and the tissue it supplies dies for lack of blood. A MELAS SLE is an energy problem. The genetic defect — most often the m.3243A>G variant in the mitochondrial tRNA-Leucine gene MT-TL1, first linked to MELAS by Goto et al. 1990 — impairs mitochondrial protein synthesis and cripples oxidative phosphorylation. Neurons, which are metabolically ravenous, reach a threshold where they cannot generate enough ATP to maintain ion gradients. They depolarize, become hyperexcitable (hence the prominent seizures), and swell. The injury spreads across cortex by metabolic and electrical contiguity, not along an arterial territory — which is precisely why the lesion crosses vascular boundaries on MRI. A coexisting microvascular endothelial dysfunction and impaired nitric oxide-mediated vasodilation worsen local perfusion, but the primary event is intracellular energy failure.

Why this changes management at the bedside. A confirmed stroke-like episode has no occluding thrombus, so thrombolysis and thrombectomy have no target. That is not a reason to withhold reperfusion when the diagnosis is uncertain: if ischemic stroke cannot be excluded within the treatment window, the standard stroke protocol applies, and m.3243A>G carriers with cardiomyopathy can have genuine embolic strokes. Once the episode is recognized, treatment centers on prompt antiseizure medication whenever seizures are suspected and on avoiding mitochondrial toxins; intravenous corticosteroids may be considered acutely, case by case, on anecdotal evidence. IV L-arginine, a nitric oxide precursor once given to improve microvascular perfusion, is unproven: the Mitochondrial Medicine Society recommended it in 2017, but neither the 2019 international consensus nor the 2026 ERN/MMS consensus (Mancuso et al. 2026) endorses it: the 2026 panel did not support recommending it for routine acute use, given open-label evidence only. It remains a case-by-case option, and practice varies. Two drug-avoidance rules follow directly from the biology — valproate impairs mitochondrial fatty-acid oxidation and has been reported to worsen seizures and stroke-like episodes (it is formally contraindicated in POLG disease), and metformin inhibits complex I and aggravates lactic acidosis.

Inheritance and a testing pitfall. MELAS is maternally inherited and heteroplasmic — the mutant and wild-type mitochondrial genomes coexist in varying proportions across tissues. Blood leukocytes turn over rapidly and tend to purge mutant mtDNA, so a low or even undetectable blood heteroplasmy does not exclude the diagnosis; post-mitotic or high-demand tissues (urinary sediment, buccal cells, or muscle) carry higher mutant loads and are the preferred specimens. For comprehensive coverage of MELAS and mitochondrial disorders, see the Mitochondrial Disorders module.

It helps to split this heterogeneous group by where in the pathophysiology the defect sits, because that determines both the type of stroke and the right intervention.

Thrombophilias — a clot problem, and mostly a venous one. Factor V Leiden and prothrombin G20210A are common (roughly 5% and 2% of Europeans) but each is only a modest risk factor, and their territory is predominantly venous. Factor V Leiden renders activated factor V resistant to cleavage by activated protein C, so the brake on the coagulation cascade fails and clot propagates. The clinical lesson is one of calibration: these are weak alleles that usually require a second hit — estrogen-containing contraceptives, pregnancy, immobility, surgery — to actually produce thrombosis, and in stroke their clearest association is with cerebral venous sinus thrombosis, not arterial infarction. A heterozygote with a single provoked event rarely needs lifelong anticoagulation; removing the provoking factor matters more than the genotype.

MTHFR — the variant to stop testing for. MTHFR C677T is the most over-ordered and over-interpreted result in this whole panel. The variant matters only insofar as it raises plasma homocysteine, and even then the effect is small. If homocysteine is normal, the genotype is clinically inert. The correct workup is therefore to measure homocysteine directly and ignore MTHFR genotyping altogether, as the ACMG 2013 practice guideline advises — a clean example of testing the phenotype, not the gene.

Vasculopathies — a vessel-wall problem. Here the defect is structural rather than thrombotic. COL4A1/COL4A2 encode type IV collagen of vascular basement membranes; mutations weaken small-vessel walls, producing a spectrum from porencephaly and small-vessel disease to frank intracerebral hemorrhage with microbleeds — a hemorrhagic, not ischemic, genetic stroke (COL4A1 also causes the systemic HANAC phenotype with renal and ocular involvement). Fabry disease (X-linked GLA deficiency of alpha-galactosidase A) is the standout because it is treatable: deficient enzyme lets globotriaosylceramide accumulate in vascular endothelium, narrowing small vessels and causing young-adult stroke alongside acroparesthesias, angiokeratomas, corneal verticillata, and renal disease. Recognizing it changes management — enzyme replacement (agalsidase beta; agalsidase alfa outside the US; pegunigalsidase alfa, FDA 2023) or, for patients with an amenable GLA variant, the oral chaperone migalastat (FDA 2018) addresses the underlying storage, making the genetic diagnosis directly therapeutic rather than merely explanatory. Women are frequently affected, and because their enzyme activity can be normal, they are diagnosed by GLA sequencing.

The workup of young stroke is an exercise in disciplined sequencing, not a reflex to send a giant gene panel on day one. The logic is to exclude the common before chasing the rare, then let two pieces of information — the stroke's mechanism and the patient's phenotype — steer the genetic testing.

Settle the mechanism first. Before any gene is ordered, the conventional workup must define what kind of stroke this is: ischemic or hemorrhagic, and if ischemic, large-vessel, small-vessel, or cardioembolic. MRI, vessel imaging, echocardiography, and rhythm monitoring exist to rule out the dissections, patent foramen, and occult atrial fibrillation that explain a large share of young strokes. This step is not a delay to genetic testing — it is the filter that makes genetic testing rational, because each genetic cause maps to a specific mechanism. Small-vessel lacunar disease with white matter change points toward CADASIL; hemorrhage with microbleeds toward COL4A1; lesions crossing territories toward MELAS; venous sinus thrombosis toward thrombophilia.

Then let the phenotype pick the test. This is where targeted testing beats shotgun panels. Acroparesthesias, angiokeratomas, and renal disease → alpha-galactosidase A activity (in males) and GLA for Fabry. Migraine, dominant pedigree, and temporal-pole white matter → NOTCH3. Hearing loss, diabetes, short stature, maternal pedigree, and a lactate peak → mtDNA testing for MELAS (on the right tissue). Only when phenotype-driven testing is unrevealing does a broad stroke gene panel or exome become the cost-effective next step in genuinely cryptogenic cases.

A caveat on thrombophilia panels in the acute setting. Functional assays for protein C, protein S, and antithrombin are unreliable during an acute thrombosis and while a patient is anticoagulated — levels are consumed acutely and altered by warfarin and heparin — so abnormal results must be confirmed off treatment before they are believed.

Prevention follows mechanism, and sometimes precedes the event. Secondary prevention is mechanism-matched: antiplatelet for small-vessel and atherosclerotic disease, anticoagulation for cardioembolic and selected coagulopathic causes, enzyme replacement or migalastat for Fabry, seizure control and avoidance of mitochondrial toxins in MELAS (supplements, including arginine, remain unproven), and discontinuation of estrogen-containing contraceptives in thrombophilia or CADASIL. Two situations are genuinely preventive rather than reactive: enzyme replacement initiated before Fabry strokes accrue, and transcranial Doppler screening in children with sickle cell disease, where a time-averaged mean of the maximum velocity (TAMM) of 200 cm/s or more identifies a child whose first stroke can be averted by transfusion or hydroxyurea — prevention before the brain is ever injured.