Stroke & neurovascular medicine, from anatomy to the bedside

Definition & epidemiology

Acute ischemic stroke (AIS) is an episode of neurological dysfunction caused by focal cerebral, spinal, or retinal infarction. The modern definition is tissue-based rather than time-based: infarction is confirmed by imaging (typically restricted diffusion on MRI) or neuropathology, or inferred from a persistent clinical deficit lasting >24 hours with other causes excluded. A transient ischemic attack (TIA) is by contrast a transient focal deficit without acute infarction on imaging — duration is no longer the discriminator.

Stroke is the second-leading cause of death worldwide and a leading cause of acquired adult disability. Ischemic stroke accounts for roughly 62–87% of all strokes (the remainder being intracerebral and subarachnoid hemorrhage), the proportion varying by population. Non-modifiable risk factors include age, sex, race/ethnicity and heredity; the dominant modifiable factors are hypertension (the single largest attributable risk), atrial fibrillation, diabetes, dyslipidemia, smoking, physical inactivity and obesity.

Pathophysiology: ischemic core vs penumbra

Arterial occlusion produces a spatial gradient of cerebral blood flow (CBF). The ischemic core is tissue with CBF so low that membrane-pump (Na⁺/K⁺-ATPase) failure and cytotoxic edema occur within minutes — irreversibly infarcted. Surrounding it is the ischemic penumbra: hypoperfused tissue that is electrically silent (functionally inactive, hence the clinical deficit) but structurally intact and potentially salvageable if flow is restored before it progresses to infarction. Between the penumbra and normal brain lies benign oligemia, which does not infarct even without reperfusion. The penumbra is the therapeutic target of all reperfusion therapy; its size and survival are governed chiefly by collateral circulation and time.

Time is brain. In supratentorial large-vessel occlusion, an estimated 1.9 million neurons, 14 billion synapses and 12 km of myelinated fibers are lost per minute of untreated ischemia; each hour without reperfusion ages the brain by the equivalent of roughly 3.6 years (Saver, 2006). This quantifies the imperative for rapid recanalization and underlies every door-to-treatment metric.

Etiologic classification (TOAST)

Diagnostic workup

Immediate goals are to exclude hemorrhage, identify treatable large-vessel occlusion (LVO), and estimate core vs penumbra. Non-contrast CT (NCCT) excludes hemorrhage and is scored with ASPECTS; CT angiography localizes occlusion and assesses collaterals; CT perfusion or MRI (DWI/PWI) estimates core and salvageable tissue for extended-window selection. Point-of-care glucose is mandatory before thrombolysis (hypo- and hyperglycemia are mimics/modifiers).

Acute management (framework)

Two pillars: reperfusion — intravenous thrombolysis (tenecteplase or alteplase) and/or endovascular thrombectomy for LVO — layered on supportive neuroprotective physiology (blood pressure, glucose, temperature, oxygenation, aspiration prevention) delivered in an organized stroke unit. These are detailed in the dedicated topics that follow.

Clinical pearls

• Duration no longer defines TIA vs stroke — DWI positivity does. Up to a third of clinically defined "TIAs" show infarction.
• A normal early NCCT does not exclude stroke; it excludes hemorrhage and gross established infarct. DWI is the most sensitive test for acute ischemia.
• The penumbra is a physiological, not anatomical, entity — its extent depends on collaterals, so two patients with identical occlusions can have very different salvageable tissue and time windows.

Key references: AHA/ASA Guideline for the Early Management of Patients With Acute Ischemic Stroke, 2019 & 2019 update, and 2026 update; Sacco et al. AHA/ASA updated definition of stroke/TIA, Stroke 2013; Saver, "Time is Brain — Quantified," Stroke 2006; Adams et al., TOAST classification, Stroke 1993.

Clinical features & recognition tools

Stroke presents as the sudden onset of a focal neurological deficit referable to a vascular territory. Public and prehospital screening uses FAST (Face droop, Arm drift, Speech disturbance, Time to call emergency services). BE-FAST adds Balance (acute gait/truncal ataxia) and Eyes (sudden diplopia, visual loss, or gaze deviation) to capture posterior-circulation events that plain FAST misses — FAST alone fails to detect roughly 1 in 7 strokes, predominantly posterior. Sudden severe "worst-ever" headache, decreased consciousness, or vomiting should prompt consideration of hemorrhage or basilar/cerebellar stroke.

Vascular anatomy & clinical syndromes

Localizing the syndrome guides territory, LVO likelihood and mimic probability.

Cortical vs subcortical vs posterior. Cortical signs (aphasia, neglect, gaze deviation, hemianopia) indicate large-artery/cardioembolic cortical involvement and predict LVO. Classic lacunar (subcortical) syndromes lack cortical signs: pure motor hemiparesis (posterior limb internal capsule/basis pontis), pure sensory stroke (thalamic VPL), ataxic hemiparesis, dysarthria–clumsy hand, and sensorimotor stroke. Posterior-circulation strokes are the most frequently missed and under-scored.

NIHSS overview

The National Institutes of Health Stroke Scale (NIHSS) is a 15-item, 0–42 quantification of deficit severity used for triage, thrombolysis/thrombectomy decisions, and serial monitoring. Domains: level of consciousness (1a–c), best gaze, visual fields, facial palsy, motor arm (L/R), motor leg (L/R), limb ataxia, sensory, best language, dysarthria, and extinction/inattention (neglect).

Caveats: the NIHSS is weighted toward left (dominant) hemisphere function (language items carry ~7 points), so a dominant-hemisphere stroke scores higher than a non-dominant stroke of equal volume. It is relatively insensitive to posterior-circulation and right-hemisphere strokes — a disabling brainstem or cerebellar infarct may score only 1–3. NIHSS ≥6 is a common (not absolute) threshold suggesting LVO warranting vessel imaging.

Differential diagnosis (stroke mimics)

• Seizure / Todd paralysis — postictal focal weakness; a witnessed convulsion or gradual "spreading" onset favors seizure.
• Hypoglycemia / hyperglycemia — can reproduce focal deficits; always check glucose first.
• Migraine with aura — spreading, marching positive symptoms (scintillations, then numbness) over minutes; younger patients.
• Functional (conversion) deficit — non-anatomical, give-way weakness, Hoover sign positive.
• Structural — tumor, subdural hematoma (subacute, fluctuating).
• Toxic-metabolic / infective — hepatic or septic encephalopathy, hyponatremia, Wernicke, HSV encephalitis.
• Peripheral vertigo — vestibular neuritis vs posterior stroke; the bedside HINTS exam (Head-Impulse, Nystagmus, Test-of-Skew) outperforms early MRI in acute vestibular syndrome — a normal head-impulse, direction-changing nystagmus, or skew deviation points to a central (stroke) cause.

Clinical pearls

• Isolated acute vertigo with a normal head-impulse test is a red flag for cerebellar/brainstem stroke, not labyrinthitis.
• Gaze deviation localizes: in hemispheric stroke the eyes look toward the lesion (away from the hemiparesis); in a pontine gaze palsy they look away from the lesion.
• A low NIHSS never excludes LVO — always image the vessels in posterior-circulation presentations.

Key references: AHA/ASA AIS Guideline 2019/2026; Aroor et al. BE-FAST, Stroke 2017; NIH Stroke Scale (Brott et al., 1989); Kattah et al., HINTS, Stroke 2009.

Prehospital care

Emergency dispatch should treat suspected stroke as a time-critical emergency. Field priorities: confirm a prehospital screen (FAST/BE-FAST or Cincinnati/Los Angeles scales), establish last known well (LKW) time, check glucose, obtain a witness/contact and medication history (especially anticoagulants), and pre-notify the receiving center. Prehospital notification measurably shortens door-to-imaging and door-to-needle times. Mobile stroke units (ambulances with onboard CT and telemedicine) enable thrombolysis in the field and, per randomized evidence (BEST-MSU, B_PROUD), improve functional outcomes; the 2026 AHA/ASA guideline gives them a Class 1 recommendation where feasible.

Prehospital LVO triage scales

Beyond detecting stroke, EMS increasingly stratifies severity to route probable LVO patients directly to thrombectomy-capable centers. No scale is perfect; each trades sensitivity against specificity.

In-hospital "code stroke"

A parallel-process pathway compresses evaluation: simultaneous triage, focused history/exam with NIHSS, point-of-care glucose, IV access and labs, and immediate transfer to CT. Advanced imaging (CTA ± CT perfusion) is obtained without delaying IV thrombolysis when the patient is otherwise eligible. Thrombolysis eligibility (time window, BP, contraindications) is determined in parallel so the bolus can be given in the scanner suite.

Time-target (door-to-treatment) metrics

These benchmarks derive from the AHA Target: Stroke quality initiative; each 15-minute reduction in door-to-needle time yields measurable gains in independent survival and reductions in symptomatic hemorrhage and mortality.

Differential diagnosis at triage

Roughly a fifth of "code stroke" activations are mimics (see Stroke Recognition topic). The two exclusions that must be made before thrombolysis are hemorrhage (NCCT) and hypoglycemia (glucose). Reassuringly, treating a mimic with IV thrombolysis carries a low symptomatic-hemorrhage rate, so uncertainty should not cause the treatable stroke patient to miss the window.

Clinical pearls

• "Last known well," not symptom-discovery time, starts the clock — critical for wake-up and aphasic/unwitnessed patients.
• Do not wait for coagulation results to treat unless the patient is on anticoagulants or has a bleeding-risk history; institutional protocols allow empiric thrombolysis when suspicion of coagulopathy is low.
• Routing decisions balance a short delay to reach an EVT center against the benefit of earlier IV thrombolysis at a closer primary center ("drip-and-ship" vs "mothership").

Key references: AHA/ASA AIS Guideline 2019/2026; AHA Target: Stroke Phase III; BEST-MSU (Grotta et al., NEJM 2021) and B_PROUD (Ebinger et al., JAMA 2021); Pérez de la Ossa et al., RACE scale, Stroke 2014.

Airway, breathing & oxygenation

Protect the airway in patients with depressed consciousness (large hemispheric, brainstem, or basilar strokes) or bulbar dysfunction with aspiration risk. Supplemental oxygen is indicated only to maintain SpO₂ >94%; routine supplemental oxygen in non-hypoxic patients confers no benefit and is not recommended.

Glucose

Both extremes worsen outcome. Treat hypoglycemia (<60 mg/dL) promptly. Target a serum glucose of roughly 140–180 mg/dL in the first 24 hours. Intensive insulin protocols targeting near-normoglycemia (e.g., 80–130 mg/dL) do not improve outcomes and increase severe hypoglycemia (SHINE trial); the 2026 guideline recommends against them.

Temperature

Fever worsens ischemic injury and independently predicts poor outcome. Identify and treat the source; give antipyretics (e.g., acetaminophen) for temperature >38 °C. Therapeutic hypothermia is not recommended outside of trials — it has not improved outcomes and adds complications (pneumonia, shivering).

Dysphagia / swallow screening

Perform a validated bedside swallow screen before any oral intake (including oral medications) — dysphagia affects up to half of acute stroke patients and drives aspiration pneumonia. Keep patients NPO until they pass; refer failures for formal speech-language/instrumental (video-fluoroscopy or FEES) evaluation.

Venous thromboembolism (VTE) prophylaxis

Early mobilization & supportive care

Avoid prolonged bed rest, but very early, high-dose out-of-bed mobilization within 24 hours was harmful in AVERT — favor frequent, shorter, individualized mobilization beginning after the first day once stable. Additional bundle elements: aspiration precautions, bowel/bladder care (avoid indwelling catheters — infection and delirium risk), pressure-area care, early nutrition (nasogastric feeding within days if dysphagic; PEG reserved for prolonged dysphagia), glycemic and BP protocols, and depression screening.

Stroke unit care

Admission to an organized, geographically discrete stroke unit with a coordinated multidisciplinary team is one of the most robust interventions in all of stroke medicine: it reduces death and dependency across all severities and ages, independent of thrombolysis, chiefly by preventing and rapidly treating complications. Continuous cardiac monitoring for ≥24 hours detects atrial fibrillation and arrhythmia.

Differential considerations in deterioration

Early neurological worsening should trigger evaluation for cerebral edema, hemorrhagic transformation, seizure, recurrent/extending ischemia, or systemic causes (hypoxia, hypotension, hyper/hypoglycemia, hyponatremia, infection) — several are reversible (see Acute Stroke Complications and Stroke in Evolution topics).

Clinical pearls

• The swallow screen is the single highest-yield nursing intervention for preventing aspiration pneumonia — enforce NPO until passed.
• Do not chase "tight" glucose control; 140–180 mg/dL is the sweet spot.
• Stroke-unit care benefits patients regardless of whether they received reperfusion therapy — it is the default disposition for every stroke.

Key references: AHA/ASA AIS Guideline 2019/2026; SHINE trial (Johnston et al., JAMA 2019); CLOTS-1/2/3 (Lancet 2009–2013); AVERT trial (Lancet 2015); Stroke Unit Trialists' Collaboration (Cochrane).

Principles

Most acute ischemic strokes present with reactive hypertension that spontaneously declines. In the ischemic penumbra, cerebral autoregulation is impaired and perfusion becomes pressure-passive — aggressive BP lowering can extend infarction, whereas extreme hypertension risks hemorrhagic transformation and edema. Management therefore differs sharply by reperfusion status.

Targets by scenario

Randomized trials of intensive post-thrombectomy BP lowering — ENCHANTED2/MT (SBP <120 vs <140–180) and OPTIMAL-BP (SBP <140 vs 140–180) — both showed worse functional outcomes with aggressive control after successful reperfusion, and BEST-II found no benefit to lower targets. Hence the current stance against driving BP down after recanalization.

Preferred agents

Nitroprusside is reserved for refractory hypertension (>220/120) given cyanide toxicity and potential to raise intracranial pressure. Titratable infusions (nicardipine/clevidipine) are generally preferred over repeated boluses when tight control is needed peri-thrombolysis.

Clinical pearls

• Before lytic, the number is 185/110; after lytic, 180/105; without reperfusion, do not treat until 220/120. Memorize these three thresholds.
• After successful thrombectomy the instinct to "protect the brain" by lowering BP is counterproductive — the reperfused but vulnerable tissue tolerates hypotension poorly.
• Always exclude a full bladder, pain, and raised ICP as reversible causes of hypertension before pharmacologic treatment.

Key references: AHA/ASA AIS Guideline 2019/2026; ENCHANTED2/MT (Yang et al., Lancet 2022); OPTIMAL-BP (Nam et al., JAMA 2023); BEST-II (Mistry et al., JAMA 2023).

Cerebral edema & raised intracranial pressure

Cytotoxic then vasogenic edema peaks at 2–5 days. Large hemispheric ("malignant" MCA) and space-occupying cerebellar infarcts cause herniation. Warning signs: declining consciousness, new pupillary asymmetry, worsening NIHSS, Cushing response. Cerebellar infarcts additionally threaten fourth-ventricle compression (obstructive hydrocephalus) and direct brainstem compression. Management: head elevation ~30°, normocapnia, osmotherapy (hypertonic saline or mannitol) as a bridge, avoidance of hypotonic fluids and fever; surgical decompression is the definitive life-saving therapy — hemicraniectomy for malignant MCA infarction and suboccipital craniectomy ± EVD for cerebellar infarction (see dedicated hemicraniectomy topic). Corticosteroids are not effective for cytotoxic ischemic edema.

Hemorrhagic transformation (HT)

HT ranges from asymptomatic petechial staining to a space-occupying parenchymal hematoma. Risk rises with infarct size, cardioembolic mechanism, reperfusion therapy, hyperglycemia, and hypertension. The radiographic ECASS classification is clinically useful:

Symptomatic ICH (sICH) — hematoma plus neurological deterioration (definitions vary: NINDS any decline; ECASS III ≥4 NIHSS points with PH2) — complicates roughly 2–7% of thrombolysis cases. Management of post-lytic sICH: stop the infusion, urgent NCCT, reverse fibrinolysis with cryoprecipitate (to replace fibrinogen) ± tranexamic acid or aminocaproic acid, correct coagulopathy, control BP, and neurosurgical consultation.

Seizures

Early (within 7 days) seizures occur in a minority, more with cortical and hemorrhagic infarcts. Treat clinical or electrographic seizures with standard antiseizure medication; prophylactic antiseizure drugs are not recommended. Consider nonconvulsive status epilepticus (EEG) when consciousness is unexplainedly depressed or fluctuating.

Infection & aspiration

Aspiration pneumonia and urinary tract infection are the commonest medical complications and directly worsen outcome — prevention hinges on swallow screening, oral care, aspiration precautions and avoiding indwelling catheters. Prophylactic antibiotics do not improve outcomes (STROKE-INF, PASS) and are not recommended; treat established infection promptly.

Cardiac complications

Stroke — especially insular cortex involvement — triggers autonomic dysregulation producing arrhythmias (new AF, QT prolongation), troponin elevation, neurogenic (Takotsubo) cardiomyopathy, and ECG changes mimicking ischemia. Continuous telemetry ≥24 h is standard. Distinguish demand ischemia/neurogenic injury from primary acute coronary syndrome, as management (particularly antithrombotics) differs.

Other

• VTE — DVT/PE from immobility (see prophylaxis in General Management).
• Hyponatremia — SIADH or cerebral salt wasting; worsens edema.
• Falls, pressure injury, depression, central/shoulder pain, spasticity — subacute complications requiring rehabilitation-phase attention.

Clinical pearls

• A deteriorating cerebellar stroke is a neurosurgical emergency — decompression/EVD is highly effective and time-sensitive; do not attribute drowsiness to "sedation."
• Reverse post-thrombolysis hemorrhage with cryoprecipitate (fibrinogen replacement), not simply platelets or vitamin K.
• Troponin elevation after stroke is common and usually neurogenic — avoid reflex antithrombotic escalation that could precipitate hemorrhagic transformation.

Key references: AHA/ASA AIS Guideline 2019/2026; ECASS/NINDS HT classifications; STROKE-INF (Kalra et al., Lancet 2015) & PASS (Westendorp et al., Lancet 2015); Wijdicks et al. AHA cerebellar/space-occupying infarct statement, Stroke 2014.

Concept

The penumbra (Astrup, Symon & Siesjö, 1981) is tissue in which CBF has fallen below the threshold for electrical function but remains above the threshold for membrane/metabolic failure — it is electrically silent yet structurally viable, and its fate (recovery vs infarction) depends on depth and duration of hypoperfusion and on collateral support. It is the operational target of every reperfusion therapy; without reperfusion, penumbra progressively recruits into core over hours.

CBF thresholds (approximate, flow in mL/100 g/min)

These thresholds are time-dependent, not fixed: tissue tolerates moderate flow reduction far longer than severe reduction. The relationship between flow depth and time to infarction is the physiological basis for individualized ("tissue-clock") rather than purely time-based selection.

Imaging surrogates & operational definitions

Automated software (e.g., RAPID-type post-processing) generates these volumes and drove patient selection in the extended-window thrombectomy trials DAWN (6–24 h, clinical–core mismatch) and DEFUSE-3 (6–16 h, perfusion–core mismatch), and the thrombolysis trials EXTEND and WAKE-UP.

Collateral circulation

Leptomeningeal (pial) collaterals from adjacent ACA/PCA territories, plus the circle of Willis, sustain the penumbra. Good collaterals slow core growth, enlarge the salvageable region, widen the treatment window and predict better reperfusion outcomes; poor collaterals produce rapid "malignant" core growth ("fast progressors"). Collateral status is assessed on multiphase CTA or from the CTP hypoperfusion-intensity ratio.

Differential / mimics of imaging mismatch

Perfusion "mismatch" is not specific to stroke — seizure (peri-ictal hyperperfusion or post-ictal hypoperfusion), migraine aura, and venous pathology can produce perfusion abnormalities that do not respect an arterial territory. Interpret CTP alongside vessel imaging and the clinical syndrome.

Clinical pearls

• The mismatch concept lets physiology, not just the clock, select patients — a "slow progressor" with excellent collaterals may retain salvageable penumbra well beyond 24 h.
• A large core with a small penumbra ("matched defect") predicts little benefit and higher hemorrhage risk; a small core with large penumbra is the ideal reperfusion candidate.
• Perfusion thresholds are operational conveniences — they approximate, but do not equal, the true pathophysiological flow thresholds and can over- or under-estimate core depending on timing and software.

Key references: Astrup, Siesjö & Symon, "Thresholds in cerebral ischemia — the ischemic penumbra," Stroke 1981; DAWN (Nogueira et al., NEJM 2018); DEFUSE-3 (Albers et al., NEJM 2018); AHA/ASA AIS Guideline 2019/2026.

Overview

Restoring flow is necessary but not by itself sufficient — the return of oxygenated blood to ischemic tissue can itself injure the brain (ischemia–reperfusion injury), microvascular flow may fail to recover despite arterial recanalization (no-reflow), and the blood–brain barrier (BBB) may break down, producing edema and hemorrhagic transformation. These processes constrain the net benefit of recanalization and are active targets of adjunctive neuroprotection research.

Ischemia–reperfusion injury: mechanisms

• Excitotoxicity — energy failure causes glutamate accumulation; NMDA/AMPA over-activation drives Ca²⁺ influx.
• Calcium overload — activates proteases, lipases, endonucleases and nitric oxide synthase.
• Oxidative & nitrosative stress — reperfusion delivers O₂ to dysfunctional mitochondria, generating a burst of reactive oxygen/nitrogen species that overwhelm antioxidant defenses and peroxidize lipids.
• Mitochondrial permeability transition — pore opening triggers apoptotic and necrotic death.
• Inflammation — microglial activation, cytokine release (TNF-α, IL-1β, IL-6), adhesion-molecule up-regulation and leukocyte infiltration extend injury.

No-reflow phenomenon

Despite successful proximal recanalization (e.g., TICI 2b–3 angiographically), tissue-level microcirculatory perfusion can remain impaired — angiographic reperfusion without tissue reperfusion. Proposed mechanisms: pericyte constriction of capillaries persisting after reflow, capillary plugging by leukocytes and platelet–fibrin aggregates, endothelial and astrocyte end-foot swelling narrowing the lumen, and microthrombosis. No-reflow contributes to the "futile recanalization" seen when a technically successful thrombectomy fails to yield clinical improvement.

Blood–brain barrier disruption

Ischemia degrades the neurovascular unit: matrix metalloproteinases (notably MMP-9) and other proteases cleave tight-junction proteins (claudin-5, occludin, ZO-1) and basal lamina, increasing permeability. BBB opening is often biphasic — an early transient opening followed by a later, more severe phase associated with vasogenic edema and hemorrhagic transformation. The consequences are:

Clinical & therapeutic relevance

These mechanisms explain why faster reperfusion yields larger benefit (less accumulated injury before reflow), why hyperglycemia and marked hypertension worsen HT, and why post-reperfusion BP should not be driven too low (impaired autoregulation + no-reflow render tissue perfusion pressure-dependent). Numerous cytoprotectants targeting these pathways (free-radical scavengers, glutamate antagonists, anti-inflammatories) have failed in trials; interest persists in agents given as adjuncts to reperfusion, but no cytoprotectant has definitive proven benefit.

Differential diagnosis (post-EVT hyperdensity)

A frequent bedside problem is the hyperdense lesion on immediate post-thrombectomy CT: contrast staining (blood–brain barrier leak of iodinated contrast — usually resolves within 24 h, Hounsfield units typically lower, respects gyral pattern) versus true hemorrhage. Delayed CT (contrast clears; hemorrhage persists/grows) or dual-energy CT resolves the distinction and prevents unnecessary reversal of antithrombotics.

Clinical pearls

• "Futile recanalization" — an open vessel with no clinical benefit — often reflects no-reflow, already-completed core, or distal embolization, not procedural failure.
• Contrast staining after EVT is common and benign; do not reflexively diagnose hemorrhage — obtain delayed/dual-energy imaging.
• Every minute saved to reperfusion reduces cumulative reperfusion injury and hemorrhage risk — speed matters even once the case is "in the lab."

Key references: Moskowitz, Lo & Iadecola, "The science of stroke: mechanisms in search of treatments," Neuron 2010; del Zoppo & Mabuchi, microvascular no-reflow reviews; Jha et al./Yang & Rosenberg, MMP and BBB disruption reviews; AHA/ASA AIS Guideline 2019/2026.

Definition & epidemiology

"Malignant MCA infarction" denotes a large hemispheric (typically complete MCA ± ACA/PCA) territory infarct that develops space-occupying cytotoxic edema, causing herniation and death in up to ~80% of medically managed patients. Edema peaks at days 2–5. It complicates a minority of MCA strokes but accounts for a disproportionate share of early stroke mortality, often in younger patients with poor collaterals and proximal (ICA-terminus/M1) occlusion.

Clinical & imaging predictors

• High baseline NIHSS (often >15 non-dominant, >20 dominant) with early decreased consciousness, nausea/vomiting.
• Early hypodensity involving >50% of the MCA territory on NCCT; hyperdense MCA sign; ICA-terminus ("T") occlusion; poor collaterals.
• Large early DWI infarct volume (e.g., >~82 mL within 6 h, >~145 mL by 14 h) predicts malignant course.
• The EDEMA score (basal cistern effacement, midline shift, glucose, prior stroke, reperfusion status) and related models help stratify risk of malignant edema, though thresholds are approximate and should not replace serial clinical/imaging assessment.

Medical management (temporizing)

Head elevation ~30°, normocapnia, osmotherapy (hypertonic saline or mannitol), fever and glucose control, avoidance of hypotonic fluids and hypoxia. These buy time but do not prevent the natural history — decompressive surgery is the only intervention shown to reduce mortality. Corticosteroids are ineffective and not recommended.

Decompressive hemicraniectomy: the evidence

Procedure: a large fronto-temporo-parietal bone flap (diameter ≥12 cm, and ≥14 cm preferred) with duraplasty; an inadequately sized craniectomy fails to relieve pressure and risks venous infarction at the bony edge. Timing is ideally early — within 24–48 h and before clinical herniation; waiting for fixed pupillary dilation forfeits much of the benefit.

Patient selection & shared decision-making

Cerebellar (posterior fossa) counterpart

Space-occupying cerebellar infarction is managed analogously but by suboccipital decompressive craniectomy (± external ventricular drain for obstructive hydrocephalus, ± infarct resection). Because the posterior fossa is small, deterioration is rapid; decompression is highly effective and should not be delayed.

Clinical pearls

• Hemicraniectomy is a life-saving, not function-restoring, operation — frame counseling around survival with disability, especially over age 60.
• Operate before the pupil blows: pre-herniation surgery drives the outcome advantage.
• An EVD alone does not relieve brainstem compression from a cerebellar infarct — suboccipital craniectomy is the definitive treatment.

Key references: Vahedi et al. pooled analysis of DECIMAL/DESTINY/HAMLET, Lancet Neurol 2007; Jüttler et al. DESTINY II, NEJM 2014; AHA/ASA scientific statement on cerebral & cerebellar space-occupying infarction (Wijdicks 2014); AHA/ASA AIS Guideline 2019/2026.

Definition & epidemiology

A wake-up stroke is one in which the patient is neurologically normal on going to sleep and has a deficit on awakening, so the exact onset is unknown and "last known well" defaults to bedtime. Wake-up strokes constitute roughly a fifth of ischemic strokes. Historically excluded from reperfusion because the conventional time window could not be established, many are now treatable using imaging-based ("tissue-clock") selection that estimates lesion age or salvageable tissue rather than relying on the wall clock.

DWI–FLAIR mismatch

An acute infarct is visible on diffusion-weighted imaging (DWI) almost immediately but takes several hours to become hyperintense on FLAIR. A lesion that is bright on DWI but not yet visible on FLAIR (DWI–FLAIR mismatch) therefore implies the infarct is relatively young — approximately within 4.5 hours — identifying patients likely to benefit from thrombolysis despite unknown onset.

Evidence

• WAKE-UP (Thomalla, NEJM 2018): unknown-onset stroke with DWI–FLAIR mismatch randomized to alteplase 0.9 mg/kg vs placebo — significantly more patients achieved favorable outcome (mRS 0–1: ~53% vs ~42%), with a non-significant increase in symptomatic hemorrhage. The trial stopped early for funding.
• EXTEND (Ma, NEJM 2019): ischemic stroke 4.5–9 h from onset or wake-up (using time from midpoint of sleep) with automated perfusion mismatch — alteplase improved excellent outcome.
• Individual-patient meta-analysis of wake-up/late thrombolysis trials confirmed net benefit with imaging selection.
• For large-vessel occlusion wake-up strokes, extended-window thrombectomy (DAWN 6–24 h, DEFUSE-3 6–16 h) applies using clinical–core or perfusion–core mismatch.

Current guideline position

Imaging-selected thrombolysis for wake-up / unknown-onset stroke is endorsed: DWI–FLAIR mismatch (MRI) or perfusion mismatch (CTP) identifies candidates. The 2026 AHA/ASA guideline supports IV thrombolysis (tenecteplase or alteplase) as reasonable in the 4.5–9 h window and in wake-up stroke when advanced imaging demonstrates salvageable tissue, and extends both thrombolysis and thrombectomy consideration to selected patients up to 24 h with favorable imaging.

Differential diagnosis

Because onset is unwitnessed, mimics loom larger: nocturnal seizure with Todd paralysis, metabolic disturbance discovered on waking, and functional deficits. A DWI-positive lesion in a vascular territory anchors the diagnosis; a normal DWI in a persisting deficit should prompt reconsideration.

Clinical pearls

• Do not reflexively exclude the wake-up patient — obtain MRI (or CTP) and let the tissue clock decide.
• DWI+/FLAIR− is the practical MRI signature of a treatable young infarct; FLAIR positivity suggests the lesion is likely >4.5 h old.
• For wake-up LVO, pursue thrombectomy evaluation in parallel — the same imaging that selects for lysis also screens for endovascular therapy.

Key references: WAKE-UP (Thomalla et al., NEJM 2018); EXTEND (Ma et al., NEJM 2019); Thomalla et al. individual-patient meta-analysis, Lancet 2020; DAWN & DEFUSE-3 (NEJM 2018); AHA/ASA AIS Guideline 2019/2026.

Overview & why the distinction matters

Neurological worsening after an initial stroke presentation has three importantly different explanations that are frequently conflated: the original infarct is progressing (stroke in evolution), a new infarct has occurred (early recurrence), or an old deficit has transiently re-emerged without new infarction (poststroke recrudescence). Their workups and treatments diverge sharply — one needs urgent reperfusion/secondary-prevention escalation, another needs only correction of a systemic trigger — so distinguishing them at the bedside is a high-yield clinical skill.

Stroke in evolution (progressing stroke / early neurological deterioration)

Defined as objective worsening of the presenting deficit within the first hours to days (commonly ≥2–4-point NIHSS increase). It reflects growth of the same ischemic process. Causes and their remedies:

A recognizable subtype is the capsular warning syndrome — stuttering, recurrent pure-motor lacunar symptoms from a single perforator, with high risk of completed capsular/pontine infarction.

Early recurrence

A genuinely new ischemic event, typically in a different territory, driven by an active embolic or unstable atherosclerotic source. Highest early-recurrence risk accompanies symptomatic large-artery (carotid) stenosis, cardioembolism (especially atrial fibrillation), and intracranial atherosclerosis. Recurrence is confirmed by a new lesion on DWI. Prevention/management centers on rapid etiologic workup and mechanism-specific therapy:

• Short-term dual antiplatelet therapy for minor stroke or high-risk TIA — aspirin + clopidogrel for ~21 days (CHANCE/POINT; started within 24 h, for NIHSS ≤3 or high-risk TIA), or ticagrelor + aspirin (THALES, NIHSS ≤5) — then single-agent, reduces 90-day recurrence.
• Anticoagulation for cardioembolic sources (timing individualized by infarct size/hemorrhage risk).
• Expedited carotid revascularization for symptomatic high-grade stenosis.
• Risk-factor control (statin, BP, glucose).

Poststroke recrudescence

Recrudescence is the transient re-emergence of deficits from a prior, already-established stroke, provoked by a physiological stressor rather than any new ischemia. It is a stroke mimic. Characteristic features:

Recognizing recrudescence avoids unnecessary thrombolysis and instead directs treatment at the underlying cause (e.g., antibiotics, fluids, glucose/sodium correction). Because it can be indistinguishable from recurrence at the bedside, DWI is the arbiter: a new bright lesion means recurrence; its absence in a patient with a known prior stroke and an identified stressor supports recrudescence.

Distinguishing the three

Clinical pearls

• In a patient with a known prior stroke who worsens during an infection or with hypotension, think recrudescence before activating thrombolysis — but confirm with DWI.
• Capsular warning syndrome is fragile: stuttering lacunar symptoms often complete into a fixed deficit — monitor closely and optimize perfusion.
• Early recurrence risk is front-loaded (highest in the first days to weeks) — the value of a fast etiologic workup and early secondary prevention is greatest right after the index event.

Key references: AHA/ASA AIS Guideline 2019/2026; CHANCE (Wang et al., NEJM 2013), POINT (Johnston et al., NEJM 2018) & THALES (Johnston et al., NEJM 2020); Topcuoglu et al., recrudescence of deficits as a stroke mimic; Seners et al., early neurological deterioration reviews.

Acute reperfusion of the ischemic penumbra is the single most powerful modifiable determinant of outcome in acute ischemic stroke (AIS). Two complementary pillars exist: intravenous thrombolysis (IVT) with a fibrinolytic (alteplase or tenecteplase) and endovascular thrombectomy (EVT) for large-vessel occlusion (LVO). The governing principle is that untreated proximal occlusion destroys an estimated ~1.9 million neurons per minute; every 15-minute reduction in onset-to-reperfusion time yields a measurable gain in disability-free survival.

The two pillars and how they combine

• IVT dissolves fibrin systemically and is the default for any eligible patient within the time window regardless of vessel, but recanalization of large proximal clots (ICA terminus, proximal M1, basilar) with lytic alone is poor (~10–30%).
• EVT physically retrieves clot and is the definitive therapy for anterior-circulation LVO (intracranial ICA, M1, proximal M2) and basilar occlusion; recanalization (mTICI 2b–3) exceeds 85% in contemporary practice.
• Bridging therapy = IVT immediately followed by transfer for EVT. IVT is not withheld to wait for EVT; the lytic is started as soon as eligibility is confirmed while the angiography suite is mobilized.

Bridging vs direct (EVT-alone) — the ongoing debate

Six randomized trials tested whether skipping IVT in LVO patients presenting directly to a thrombectomy-capable center is non-inferior to bridging:

Meta-analytic synthesis (including the IRIS collaboration) shows bridging retains a small advantage in functional independence and pre-EVT reperfusion without a definitive safety penalty. Current guidance: give IVT to all eligible patients even when EVT is planned; do not skip the lytic. Direct EVT is reserved for patients with an IVT contraindication.

Workflow metrics and pre-hospital triage

• Door-to-needle (DTN): target <60 min, best-practice <30 min (Target: Stroke Phase III).
• Door-to-groin puncture: target <90 min for direct presenters; door-in–door-out <60 min for transfers.
• Last known well (LKW) anchors all windows, not symptom-discovery time.
• Pre-hospital LVO screens (RACE, LAMS, VAN, C-STAT) route suspected LVO to thrombectomy centers. Mobile stroke units (CT-equipped ambulances; BEST-MSU, B_PROUD) shorten onset-to-thrombolysis and improve outcomes.

Key references: Powers WJ et al. 2019 AHA/ASA Guidelines for the Early Management of AIS (Stroke 2019;50:e344) and the 2026 AHA/ASA update; Turc G et al. ESO-ESMINT thrombectomy guidelines 2019/2023; Fischer U et al. SWIFT-DIRECT (Lancet 2022); Majoie CB et al. IRIS individual-patient meta-analysis (Lancet 2023).

Recombinant tissue plasminogen activator (alteplase) was the first proven reperfusion therapy (NINDS, 1995) and remains a standard agent, now increasingly displaced by tenecteplase. It is a fibrin-selective plasminogen activator with a short plasma half-life (~4–5 min), mandating a continuous infusion.

Indications

• Disabling AIS with treatment feasible ≤4.5 h from LKW (Class I ≤3 h; Class I for 3–4.5 h).
• Age ≥18 (benefit also established ≥80 y within 3 h; IST-3 supported broader age treatment).
• Measurable, disabling deficit — including otherwise-eligible patients with milder but disabling symptoms; NIHSS has no upper cutoff that excludes treatment per se.

Dosing & administration

Worked examples: 70 kg → 63 mg total (6.3 mg bolus, 56.7 mg over 60 min). 85 kg → 76.5 mg (7.65 mg bolus, 68.85 mg infusion). 110 kg → capped 90 mg (9 mg bolus, 81 mg infusion). Overdose from estimated weight is a leading cause of avoidable thrombolysis-related hemorrhage — verify weight.

Practical protocol

• BP <185/110 mmHg before bolus (labetalol 10–20 mg IV, nicardipine 5 mg/h titrated, or clevidipine); maintain <180/105 for 24 h with q15min → q30min → q60min neuro/BP checks.
• Only laboratory result required before treatment in most patients is glucose; do not delay for INR/platelets unless anticoagulant use or bleeding diathesis is suspected.
• No antiplatelet/anticoagulant for 24 h; obtain follow-up CT/MRI at 24 h before starting antithrombotics.
• Avoid arterial punctures, central lines, NG tubes, and bladder catheters during and shortly after infusion where feasible.

Evidence

Complications

Symptomatic ICH (~2–6% depending on definition), orolingual angioedema (~1–5%, higher with ACE inhibitors), systemic bleeding, and rare reperfusion injury. (See dedicated hemorrhagic-transformation topic.)

Clinical pearls

• The 3–4.5 h window historically excluded age >80, NIHSS >25, oral anticoagulant use (any INR), and diabetes-plus-prior-stroke; AHA/ASA now considers many of these patients treatable — individualize rather than reflexively exclude.
• “Rapidly improving” deficits should be treated if the residual deficit is disabling; do not withhold for improvement that still leaves a functionally significant deficit (PRISMS informed but did not resolve the truly-minor question).
• Alteplase does not preclude subsequent EVT — it is bridging.

Key references: NINDS rt-PA Study Group (NEJM 1995); Hacke W et al. ECASS III (NEJM 2008); IST-3 Collaborative Group (Lancet 2012); Emberson J et al. (Lancet 2014); Powers WJ et al. 2019 AHA/ASA Guideline and 2026 update.

Tenecteplase (TNK) is a genetically engineered variant of alteplase with three amino-acid substitutions conferring greater fibrin specificity, resistance to PAI-1, and a longer half-life (~20–24 min) — permitting a single IV bolus. This pharmacology is ideal for stroke workflow, particularly drip-and-ship transfers for EVT. In a landmark change, the FDA approved TNKase (tenecteplase) for AIS on 3 March 2025 — the first new stroke thrombolytic in roughly 30 years.

Dosing & administration

Note: The cardiac TNK dose/reconstitution differs; stroke dosing is 0.25 mg/kg (max 25 mg). Reconstitution and rounding errors are a recognized safety issue — use weight-based charts.

Evidence

Current status & LVO/bridging

The 2026 AHA/ASA guideline places tenecteplase 0.25 mg/kg and alteplase 0.9 mg/kg on equal footing within 4.5 h, noting TNK’s practical advantages (single bolus, simpler workflow). Many comprehensive stroke centers have switched to TNK as the default, especially for LVO/bridging where faster administration and higher early recanalization matter. For the extended and unknown-onset window, TNK data diverge by EVT status (see extended-window topic: TRACE-III positive without EVT; TIMELESS neutral in an EVT-predominant population).

Clinical pearls

• The single-bolus format eliminates infusion-pump interruptions during inter-hospital transfer — a real workflow advantage for bridging.
• Non-inferiority (not proven superiority) underlies the head-to-head data; TNK is favored on pragmatics and reperfusion physiology, not a large efficacy gap.
• Do not confuse the vial and dose with cardiac TNK protocols.

Key references: Campbell BCV et al. EXTEND-IA TNK (NEJM 2018) and part 2 (JAMA 2020); Menon BK et al. AcT (Lancet 2022); Wang Y et al. TRACE-2 (Lancet 2023); Muir KW et al. ATTEST-2 (Lancet Neurol 2024); FDA TNKase AIS approval (Mar 2025); 2026 AHA/ASA Guideline update.

Thrombolysis eligibility balances hemorrhage risk against reperfusion benefit. The 2019 AHA/ASA guideline reclassified many historical “absolute” exclusions as relative or removed them. The list below reflects contemporary practice; when a criterion is relative, weigh individually and document.

Absolute contraindications

Relative contraindications / special situations

Clinical pearls

• Do not delay the bolus for coagulation panels in patients without anticoagulant exposure or bleeding history — only glucose is mandatory.
• DOAC exposure: treat if last dose >48 h with normal renal function, or if a calibrated drug-specific assay (anti-Xa for apixaban/rivaroxaban, dilute thrombin time/ecarin for dabigatran) is below the detection threshold.
• “INR >1.7” applies to warfarin; a mildly elevated INR from other causes still contraindicates.

Key references: Powers WJ et al. 2019 AHA/ASA Guideline (Stroke 2019;50:e344); Fugate JE, Rabinstein AA. Contraindications to IV thrombolysis (Neurohospitalist 2015); Re-examining Acute Eligibility for Thrombolysis (RAET) analyses.

Roughly 20–25% of strokes are “wake-up” or unwitnessed with unknown onset, historically excluded from lysis. Advanced imaging — DWI–FLAIR mismatch (tissue clock: infarct visible on DWI but not yet on FLAIR implies onset <~4.5 h) or perfusion core/penumbra mismatch — identifies patients likely to benefit beyond the conventional window.

Imaging selection paradigms

• DWI–FLAIR mismatch: acute DWI lesion without corresponding parenchymal FLAIR hyperintensity → treat (WAKE-UP paradigm).
• Perfusion mismatch (CTP/MR): small core with salvageable penumbra → treat (EXTEND paradigm). Typical thresholds mirror EVT trials: core (rCBF <30%) modest, mismatch ratio >1.2–1.8, mismatch volume >10–15 mL.

Evidence

Tenecteplase in the late window

Current guidance (2026 update)

The 2026 AHA/ASA update endorses imaging-guided IVT in the extended window: reasonable to give IVT 4.5–9 h after onset or from the midpoint of sleep in wake-up stroke when advanced imaging shows salvageable tissue; and for LVO patients who cannot access EVT, IVT may be considered up to 4.5–24 h with favorable perfusion imaging (informed by TRACE-III). When EVT is available and indicated, EVT remains the priority intervention in the late window.

Clinical pearls

• “Time is brain” becomes “tissue is brain” in the late window — the scan, not the clock, drives eligibility.
• TWIST (no imaging) vs TRACE-III (imaging) crystallizes the lesson: penumbral selection is non-negotiable beyond 4.5 h.
• Do not use late-window IVT as a substitute for EVT in an EVT-eligible LVO patient with access to thrombectomy.

Key references: Thomalla G et al. WAKE-UP (NEJM 2018); Ma H et al. EXTEND (NEJM 2019); Thomalla G et al. pooled analysis (Lancet 2020); Xiong Y et al. TRACE-III (NEJM 2024); Albers GW et al. TIMELESS (NEJM 2024); 2026 AHA/ASA update.

Catheter-directed intra-arterial (IA) fibrinolysis delivers lytic at high local concentration to the clot face. Once a stand-alone strategy, it has been largely superseded by stent-retriever/aspiration thrombectomy but retains defined niche and adjunctive roles.

Historical basis

• PROACT II (1999): IA recombinant pro-urokinase within 6 h for M1/M2 MCA occlusion improved recanalization (66% vs 18%) and mRS 0–2 (40% vs 25%) despite higher early sICH — proof of concept for endovascular reperfusion.
• The device era (2015 onward) demonstrated mechanical retrieval is faster and achieves higher, more durable recanalization than IA lytic alone.

Current niche & adjunctive use

• Distal / medium-vessel occlusions (distal M2/M3, A2/A3, P2) not safely reachable by a retriever — low-dose IA alteplase/TNK or tirofiban may be used as rescue.
• Rescue after incomplete EVT — residual distal branch occlusion or no-reflow despite proximal recanalization.
• Adjunctive IA lytic after successful thrombectomy to improve microvascular (capillary-level) reperfusion — tested by CHOICE.

Evidence for adjunctive IA lytic

CHOICE is promising but underpowered; larger trials (e.g. of adjunctive IA lytic and IA tirofiban) are ongoing. This is not yet standard of care.

Clinical pearls

• IA lytic dosing is empirical and much lower than IV; there is no universally validated regimen — use institutional protocols.
• Consider IA rescue for a symptomatic branch the retriever cannot engage, weighing hemorrhage risk against the deficit.

Key references: Furlan A et al. PROACT II (JAMA 1999); Renu A et al. CHOICE (JAMA 2022); ESO-ESMINT guideline commentary on medium-vessel occlusion.

Endovascular thrombectomy (EVT) with modern devices is the most effective treatment in medicine for anterior-circulation LVO by number-needed-to-treat — roughly NNT 2.6 for reduced disability in the early-window pooled data. Success is defined angiographically as mTICI 2b–3, with a target of 2c/3.

Indications (early window, ≤6 h)

• Age ≥18; pre-stroke mRS 0–1; causative occlusion of ICA or M1; ASPECTS ≥6; NIHSS ≥6; groin puncture ≤6 h from LKW (Class I).
• Reasonable (Class IIa/IIb) for proximal M2, selected A1/A2, and carefully chosen patients outside the strict pivotal criteria (lower ASPECTS, mRS >1, NIHSS <6).

Extended window (6–24 h) — imaging-selected

Evidence (early window, HERMES)

The HERMES individual-patient meta-analysis pooled the five 2015 stent-retriever trials (MR CLEAN, ESCAPE, EXTEND-IA, SWIFT PRIME, REVASCAT; n=1287): mRS shift common OR 2.49; mRS 0–2 46.0% vs 26.5%; benefit across age, sex, NIHSS, and with/without IVT. DAWN and DEFUSE-3 extended benefit to 24 h with dramatic effect sizes (see large-core/summary topics).

Devices & technique

ASTER (2017) and COMPASS (2019) established aspiration-first as a reasonable primary technique; choice is operator/anatomy dependent. First-pass effect (mTICI 2c/3 on a single pass) correlates with best outcomes.

ASPECTS in patient selection

The Alberta Stroke Program Early CT Score (0–10; deduct 1 per involved MCA region) quantifies early ischemic change. Pivotal early-window trials generally required ASPECTS ≥6. Lower scores historically excluded patients — a boundary now overturned by large-core trials (see next topic).

Complications

Vessel perforation/dissection, distal embolization to new territory, symptomatic ICH, groin/access-site hematoma, contrast nephropathy, and reperfusion hemorrhage. Procedural vascular complications were notably higher in some large-core trials.

Clinical pearls

• Reperfusion time is the dominant modifiable variable — every pass and every minute counts; aim for mTICI 2c/3.
• Late-window eligibility hinges on penumbral imaging (CTP/MR perfusion or clinical-core mismatch), not the clock alone.
• Do not withhold IVT before EVT in an IVT-eligible patient to “save time” — bridging remains standard.

Key references: Goyal M et al. HERMES (Lancet 2016); Nogueira RG et al. DAWN (NEJM 2018); Albers GW et al. DEFUSE-3 (NEJM 2018); Lapergue B et al. ASTER (JAMA 2017); Turk AS et al. COMPASS (Lancet 2019); 2019 AHA/ASA Guideline; ESO-ESMINT 2019/2023.

Two frontiers extended EVT beyond the original pivotal criteria: patients with an already-large infarct core (low ASPECTS) and patients with basilar-artery occlusion (BAO), both previously excluded or of uncertain benefit.

Large-core anterior-circulation trials

Synthesis: across trials, EVT for large cores improves functional outcomes and (in LASTE and TENSION) reduces mortality despite a higher absolute rate of symptomatic hemorrhage. Meta-analyses (after TESLA, TENSION, LASTE) confirm benefit even down to ASPECTS 0–2, though absolute good-outcome rates are low and shared decision-making is essential. The 2023 AHA/ASA Science Advisory endorses EVT for selected large-core patients (roughly ASPECTS 3–5, or core up to ~100 mL).

Basilar-artery occlusion

Early BAO RCTs (BEST 2020, BASICS 2021) were individually neutral, undermined by crossover and slow enrollment. Two subsequent Chinese trials were decisively positive:

EVT is now indicated for BAO with a salvageable brainstem/posterior circulation, extending to 24 h in selected patients. Posterior-circulation ASPECTS (pc-ASPECTS) and the pons-midbrain index help gauge established infarct; extensive brainstem infarction portends futility.

Clinical pearls

• A large core is no longer an automatic exclusion — but counsel families that many survivors remain dependent; the benefit is a shift toward less-severe disability and survival, not routine independence.
• For BAO, treat aggressively within 24 h if the brainstem is not already infarcted; outcomes without recanalization are often fatal or locked-in.
• Malignant edema is a major early risk after large-core reperfusion — anticipate decompressive surgery.

Key references: Yoshimura S et al. RESCUE-Japan LIMIT (NEJM 2022); Sarraj A et al. SELECT2 (NEJM 2023); Huo X et al. ANGEL-ASPECT (NEJM 2023); Bendszus M et al. TENSION (Lancet 2023); Costalat V et al. LASTE (NEJM 2024); Yoo AJ et al. TESLA (JAMA 2024); Tao C et al. ATTENTION (NEJM 2022); Jovin TG et al. BAOCHE (NEJM 2022); AHA/ASA Large-Core Science Advisory (Stroke 2023).

Consolidated reference of the landmark EVT randomized controlled trials that define contemporary practice, grouped by era and indication. Effect sizes are for the primary functional endpoint unless noted.

Early-window pivotal (2015) & pooled

Extended window (imaging-selected)

Large core

Basilar / posterior circulation

Bridging vs direct EVT

Key references: Individual trial primary publications (NEJM/Lancet/JAMA 2015–2024) as cited in the respective topics; Goyal M et al. HERMES (Lancet 2016); Majoie CB et al. IRIS bridging meta-analysis (Lancet 2023).

Prior anticoagulation complicates thrombolysis (bleeding risk) far more than thrombectomy (mechanical, no systemic lytic). The general principle: EVT is preferred and generally safe in anticoagulated LVO patients, whereas IVT is constrained by coagulation status.

Warfarin (vitamin-K antagonist)

• IVT permitted if INR ≤1.7 (and PT ≤15 s); INR >1.7 is a contraindication to lysis.
• EVT is appropriate regardless of INR when LVO is present.

Direct oral anticoagulants (DOACs)

Reversal-then-lyse strategy

• Dabigatran + idarucizumab → IVT is the best-established reversal-then-thrombolysis pathway (registry/case-series evidence; generally safe).
• For Xa inhibitors, reversal-then-lyse is not advised because andexanet’s procoagulant profile and its interference with heparin (relevant if EVT follows) create hazard; proceed to EVT instead when LVO is present.

Practical algorithm

• Confirm agent, last dose timing, renal function, and available assays immediately.
• LVO + recent DOAC: proceed to EVT (no lytic exposure needed) rather than delaying for reversal.
• Non-LVO stroke on a DOAC within 48 h: usually forego IVT unless assays confirm safety; treat with best medical care and reassess.
• Antiplatelet therapy (even dual) is not a contraindication to IVT, though it modestly raises sICH risk.

Clinical pearls

• The single most important move in an anticoagulated LVO patient is to route to thrombectomy, not to chase coagulation reversal.
• Do not administer andexanet to enable IVT — it can promote clot propagation and complicate periprocedural heparinization.
• Document assay values and timing; “on a DOAC” alone is insufficient to decide.

Key references: 2019 AHA/ASA Guideline (Stroke 2019;50:e344) § anticoagulation; Kermer P et al. idarucizumab-before-thrombolysis registry (Eur J Neurol / Int J Stroke); Shahjouei S et al. DOAC-thrombolysis meta-analyses; ESO thrombolysis guidance.

Hemorrhagic transformation (HT) is the principal hazard of reperfusion. It spans clinically silent petechiae to space-occupying parenchymal hematoma. Standardized radiographic and clinical definitions are essential for prognosis and trial comparison.

ECASS radiographic classification

PH2 carries the strongest association with neurological deterioration and death.

Definitions of symptomatic ICH (sICH)

Reported sICH rates differ severalfold depending on which definition is used — always specify.

Risk-prediction scores

Management of post-thrombolysis sICH

• Stop the alteplase infusion immediately (if still running).
• Emergent non-contrast CT; send fibrinogen, PT/aPTT, platelets, type & cross.
• Cryoprecipitate 10 units IV (repeat to keep fibrinogen ≥150 mg/dL) — the primary reversal for lytic-associated hemorrhage.
• Tranexamic acid 1000 mg IV over 10 min (or ε-aminocaproic acid) as antifibrinolytic adjunct.
• Platelet transfusion if thrombocytopenic or on antiplatelets (individualized; PATCH cautions against routine platelets in spontaneous ICH but the post-lytic context differs).
• BP control, reverse any concurrent anticoagulant, neurosurgical consult for evacuation/EVD, ICU care.

Other reperfusion complications

• Orolingual angioedema (~1–5%; ACE-inhibitor & insular/frontal cortex association): often contralateral to the ischemic hemisphere. Stop lytic if severe; give antihistamine (H1+H2), corticosteroid, and secure the airway; icatibant/C1-esterase inhibitor for refractory bradykinin-mediated cases; watch for biphasic reactions.
• Malignant cerebral edema after large-territory reperfusion — consider early decompressive hemicraniectomy (DESTINY/DECIMAL/HAMLET; ≤48 h, age-dependent benefit).
• Reperfusion injury / hyperperfusion, contrast-induced encephalopathy or nephropathy (post-EVT), access-site hematoma, and distal embolization.

Clinical pearls

• Acute neuro-worsening during/after lysis is ICH until proven otherwise — scan now, do not wait.
• Cryoprecipitate (fibrinogen replacement) is the cornerstone antidote for lytic hemorrhage; TXA is adjunctive.
• Silent HI1/HI2 does not by itself mandate a change in management, but delays antithrombotic initiation.

Key references: Hacke W et al. ECASS classification; Wahlgren N et al. SITS-MOST (Lancet 2007); Lou M et al. HAT score (Neurology 2008); Strbian D et al. SEDAN (Ann Neurol 2012); 2019 AHA/ASA Guideline; Frontera JA et al. reversal of antithrombotic-associated ICH guideline (Neurocrit Care 2016).

Beyond IVT and EVT, several adjuncts and alternative reperfusion/neuroprotection strategies have been tested. Most are investigational or niche; awareness of the evidence prevents both under- and over-use.

Glycoprotein IIb/IIIa inhibitors & periprocedural antithrombotics

• Tirofiban, eptifibatide, abciximab are not routine in AIS. Intra-arterial or IV GP IIb/IIIa may be used as rescue for procedural re-occlusion, in-stent thrombosis (e.g. after emergent carotid stenting/tandem lesions), or refractory reperfusion.
• RESCUE BT (2022, China): IV tirofiban before EVT did not improve the primary functional outcome overall; possible signal in specific etiologies (e.g. large-artery atherosclerosis). Bleeding risk requires caution.
• Emergent carotid stenting during EVT for tandem occlusion often requires antiplatelet loading — balance against reperfusion-hemorrhage risk.

Intra-arterial adjuncts

• CHOICE (2022): IA alteplase after successful thrombectomy (mTICI 2b–3) improved mRS 0–1 without excess sICH in a small trial — hypothesis-generating, targeting residual microvascular no-reflow. Larger trials pending; not standard of care.
• IA tirofiban after EVT is under study for the same indication.

Sonothrombolysis

• CLOTBUST (2004) phase 2 suggested transcranial-Doppler augmentation of lytic recanalization; the pivotal CLOTBUST-ER (2019) was neutral. Microbubble/ultrasound approaches carried ICH concerns in some studies. Not recommended outside research.

Neuroprotection

No pharmacologic neuroprotectant is approved for AIS. Reperfusion remains the only proven “neuroprotection.”

General supportive measures that matter

• Normoglycemia (avoid >180 mg/dL and hypoglycemia), normothermia, permissive hypertension in non-reperfused tissue, euvolemia, early dysphagia screening, and DVT prophylaxis.

Clinical pearls

• Do not deploy sonothrombolysis or unproven neuroprotectants in place of established reperfusion.
• GP IIb/IIIa agents are a rescue tool for procedural thrombosis, not a routine pre-treatment.
• The strongest “adjunct” remains fast, complete recanalization plus meticulous physiologic support.

Key references: Hill MD et al. ESCAPE-NA1 (Lancet 2020); Chamorro A et al. URICO-ICTUS (Lancet Neurol 2014); Alexandrov AV et al. CLOTBUST-ER (Lancet Neurol 2019); Qiu Z et al. RESCUE BT (JAMA 2022); Renu A et al. CHOICE (JAMA 2022).

Angiographic reperfusion is graded to standardize procedural success and predict outcome. The Thrombolysis in Cerebral Infarction (TICI) scale and its modifications are the lingua franca of thrombectomy reporting.

Original TICI scale

Modified TICI (mTICI) and expanded (eTICI)

Note the 2b threshold difference: original TICI 2b was ≥50% wording but mTICI operationalizes it as ≥50% substantial reperfusion; eTICI subdivides 2b into finer bands (67–89%) and formalizes 2c, improving outcome prediction and inter-rater reliability.

Definitions that drive practice

• Successful reperfusion = mTICI 2b–3. This is the benchmark endpoint of thrombectomy trials.
• Excellent reperfusion = mTICI 2c–3, the contemporary goal — outcomes with 2c/3 clearly exceed 2b.
• First-pass effect (FPE): mTICI 2c/3 achieved with a single device pass and no rescue — the strongest procedural predictor of good functional outcome; “modified FPE” = 2b–3 first pass.

Why it matters

• Functional independence rises stepwise with reperfusion grade; each increment from 2a → 2b → 2c → 3 improves mRS distribution.
• Persisting distal occlusion after proximal recanalization (2b) is a rationale for adjunctive rescue (IA lytic/tirofiban, additional passes) — weighed against hemorrhage and procedural risk.
• Grading should be reported for the target territory at the end of the procedure using the final control run.

Clinical pearls

• Aim for 2c/3, not merely 2b — “good enough” recanalization leaves outcome on the table, but balance additional passes against escalating vessel-injury and hemorrhage risk.
• Document the number of passes and FPE status — they are quality metrics and prognostic markers.
• TICI describes angiographic reperfusion, not tissue-level (microvascular) reperfusion, which no-reflow phenomena can dissociate from — the target of adjunctive strategies like CHOICE.

Key references: Higashida RT, Furlan AJ. TICI recommendations (Stroke 2003); Zaidat OO et al. mTICI/consensus reperfusion grading (Stroke 2013); Liebeskind DS et al. eTICI (J Neurointerv Surg 2019); Zaidat OO et al. first-pass effect (Stroke 2018).

Definition & epidemiology

Cardioembolic stroke is arterial occlusion by thromboembolic material originating in the heart or, by extension, the proximal aorta. It accounts for roughly 20–30% of ischemic strokes and its share rises steeply with age because of the age-dependence of atrial fibrillation (AF). Cardioembolic strokes are, on average, the most severe ischemic subtype (highest admission NIHSS, highest early mortality, highest rate of hemorrhagic transformation) and carry the highest recurrence risk when the underlying source is left untreated, which is why prompt source identification is decisive.

Pathophysiology & vascular anatomy

An embolus travels distally until it impacts at a vessel of matching caliber, typically at a bifurcation. The middle cerebral artery (MCA) territory receives the largest share of emboli owing to flow dynamics, with the superior and inferior divisions and the M2 branches commonly involved; the posterior cerebral artery (PCA) and the basilar apex (“top of the basilar”) are also frequent targets. Hallmarks that suggest an embolic mechanism include involvement of multiple vascular territories (e.g., simultaneous MCA and PCA, or bilateral/anterior-and-posterior circulation infarcts), a wedge-shaped cortical or cortico-subcortical topography, early recanalization producing the “spectacular shrinking deficit,” and hemorrhagic transformation from reperfusion of ischemic tissue.

Sources are stratified by embolic risk. This distinction matters because high-risk sources justify anticoagulation whereas minor/uncertain sources often do not.

Clinical features

Onset is typically abrupt with a maximal deficit at outset (in contrast to the stuttering course of small-vessel or branch atheromatous disease). Suggestive features include depressed level of consciousness, prominent cortical signs (aphasia, neglect, hemianopia, gaze deviation), seizure at onset, and a deficit that improves rapidly as the clot fragments. None of these is pathognomonic; source imaging remains essential.

Diagnostic workup

• Rhythm detection (the central task): admission 12-lead ECG and continuous inpatient telemetry detect only a minority of paroxysmal AF. Escalate to ambulatory monitoring — 24–48 h Holter, then 14–30 day event/patch monitors, and an insertable cardiac monitor (ICM) when suspicion remains high. EMBRACE (30-day event recorder) and CRYSTAL-AF (ICM: AF detection ~8.9% at 6 months rising to ~30% at 3 years) established that longer monitoring markedly increases yield. STROKE-AF and the LOOP study extended this to non-cryptogenic and community populations, though the clinical benefit of treating short, device-detected subclinical AF is being refined by NOAH-AFNET 6 and ARTESIA.
• Cardiac structural imaging: transthoracic echocardiography for ventricular function, wall-motion abnormality and valves; transesophageal echocardiography (TEE) is superior for LAA thrombus, aortic arch atheroma, PFO with bubble study, vegetations and prosthetic valves. Cardiac CT/MRI can characterize thrombus and cardiomyopathy.
• Vascular and parenchymal imaging to demonstrate the embolic pattern and exclude large-artery stenosis; blood cultures and valve imaging if endocarditis is suspected.

Management

Secondary prevention centers on the source. For non-valvular AF, direct oral anticoagulants (DOACs: apixaban, rivaroxaban, edoxaban, dabigatran) are preferred over warfarin for most patients (lower intracranial hemorrhage, no INR monitoring). Warfarin remains indicated for mechanical valves and moderate-to-severe mitral stenosis, where DOACs are contraindicated (RE-ALIGN). Timing of anticoagulation after an acute AF-related infarct has been clarified by ELAN (early initiation, ~day 3–4 for larger strokes, was safe and reasonable), OPTIMAS (early initiation ≤4 days non-inferior to delayed), and the CATALYST individual-patient-data meta-analysis (favoring early DOAC), superseding the older empirical “1–3–6–12 day” rule that keyed timing to infarct size. Left atrial appendage occlusion (e.g., WATCHMAN; PROTECT-AF, PREVAIL) is an option when long-term anticoagulation is contraindicated. Antibiotics (not anticoagulation) treat infective endocarditis; treat the underlying malignancy and use LMWH/anticoagulation for marantic endocarditis.

Prognosis

Cardioembolic infarcts are larger and more disabling on average, but anticoagulation of a high-risk source produces one of the largest relative risk reductions in all of stroke prevention (roughly two-thirds reduction for AF), making accurate diagnosis high-yield.

Clinical pearls

• Multiterritory or bilateral acute infarcts on DWI strongly imply a proximal (cardiac or aortic) embolic source until proven otherwise.
• A single admission ECG effectively never excludes paroxysmal AF; commit to prolonged monitoring in embolic-appearing strokes.
• Hemorrhagic transformation is common and reflects reperfusion—anticipate it when timing anticoagulation.
• Do not anticoagulate infective endocarditis for stroke prevention—it increases intracranial hemorrhage risk.

Key references: Kleindorfer et al., 2021 AHA/ASA Guideline for the Prevention of Stroke in Patients With Stroke/TIA (Stroke). Sanna et al., CRYSTAL-AF (NEJM 2014). Gladstone et al., EMBRACE (NEJM 2014). Fischer et al., ELAN (NEJM 2023). Werring et al., OPTIMAS (Lancet 2024). CATALYST collaboration (Lancet 2025).

Definition & epidemiology

Lacunar (small-vessel) infarcts are small (<15–20 mm) subcortical lesions caused by occlusion of a single deep penetrating arteriole. They constitute roughly 20–25% of ischemic strokes. C. Miller Fisher defined both the pathology and the classic clinical syndromes; the term “lacune” (little lake) describes the small cavity left after the infarct resolves.

Pathophysiology & vascular anatomy

The culprit vessels are small (40–400 µm) end-arteries with little collateral: the lenticulostriate branches of the M1 MCA (supplying putamen, globus pallidus, internal capsule, caudate, corona radiata), the thalamoperforators/thalamogeniculate branches of the PCA and posterior communicating artery (thalamus), the recurrent artery of Heubner (anterior limb of the internal capsule/caudate), and the paramedian pontine perforators of the basilar (basis pontis). Two microangiopathic mechanisms dominate: lipohyalinosis (fibrinoid degeneration of the vessel wall, classically producing the smallest lacunes) and microatheroma at the origin of a single perforator. The principal risk factors are chronic hypertension and diabetes; the mechanism is intrinsic small-vessel disease rather than embolism, so an extensive embolic search is usually low-yield when the clinical and imaging picture is classic. A key clinical rule: lacunar syndromes lack cortical signs (no aphasia, neglect, hemianopia, or depressed consciousness).

Clinical features — the five classic syndromes

Differential diagnosis

Branch atheromatous disease (BAD) produces a larger, often fluctuating/progressive deficit from occlusion at the origin of a perforator; small striatocapsular or cortical embolic infarcts can mimic a lacunar syndrome but usually show a cardioembolic source or additional cortical signs; small deep hemorrhage can present identically and must be excluded with imaging.

Diagnostic workup

DWI-MRI confirms the small deep infarct and its perforator territory (up to a fifth of clinically lacunar strokes are DWI-negative early—re-image if the syndrome is classic). Vascular imaging and cardiac evaluation are warranted to exclude a large-artery or embolic mimic, but yield is lower than in cortical strokes. Look for coexisting markers of cerebral small-vessel disease (confluent white-matter hyperintensities, microbleeds, enlarged perivascular spaces, old lacunes) that reinforce the microangiopathic diagnosis.

Management

Antiplatelet therapy, aggressive risk-factor control (particularly blood pressure), and high-intensity statin. The SPS3 trial informs two points specific to lacunar disease: long-term dual antiplatelet therapy (aspirin + clopidogrel) increased bleeding and mortality without preventing recurrence and should be avoided, and a systolic target of <130 mmHg was reasonable and trended toward fewer hemorrhages. Short-course DAPT still applies to the acute minor-stroke window (see TIA topic).

Prognosis

Early functional recovery is often good, but lacunar disease is not benign: it drives stepwise cognitive decline and vascular dementia, gait disorder, and carries a substantial recurrent-stroke rate reflecting ongoing small-vessel pathology.

Clinical pearls

• A lacunar syndrome plus any cortical sign should prompt reclassification—you are probably dealing with an embolus or BAD.
• Ataxic hemiparesis and dysarthria–clumsy hand overlap and both localize commonly to the pons or internal capsule.
• Avoid long-term dual antiplatelet therapy in established lacunar disease (SPS3).

Key references: Fisher CM, lacunar syndromes (Neurology, classic series). SPS3 Investigators, dual antiplatelet and blood-pressure targets (NEJM 2012; Lancet 2013). Wardlaw et al., STRIVE neuroimaging standards for small-vessel disease (Lancet Neurology 2013).

Definition & epidemiology

Posterior circulation strokes involve the vertebrobasilar system and account for roughly 20–25% of ischemic strokes. They are disproportionately misdiagnosed at first contact because symptoms (dizziness, imbalance, nausea) overlap with benign vestibular disorders and because early CT and even early DWI are less sensitive in the posterior fossa.

Pathophysiology & vascular anatomy

The vertebral arteries (V1–V4) give rise to the posterior inferior cerebellar arteries (PICA) and unite to form the basilar artery, which supplies the pons via paramedian and short/long circumferential perforators, gives off the anterior inferior cerebellar arteries (AICA) and superior cerebellar arteries (SCA), and terminates in the paired PCAs supplying the midbrain, thalamus (via thalamoperforators/thalamogeniculate branches and the artery of Percheron), medial temporal and occipital lobes. Mechanisms include large-artery atherosclerosis (vertebral origin, intracranial vertebral and basilar), artery-to-artery and cardiac embolism (basilar apex a favored target), small-vessel occlusion of pontine perforators, and vertebral artery dissection (a leading cause in the young).

Clinical features — brainstem & cerebellar syndromes

The signature of a brainstem stroke is “crossed” findings (ipsilateral cranial nerve deficit with contralateral long-tract deficit) and the classic “5 D's”—dizziness/vertigo, diplopia, dysarthria, dysphagia, dystaxia—often with nausea, hiccups, and Horner syndrome.

Diagnostic pitfalls & the HINTS examination

Isolated vertigo can be central; features raising alarm include inability to stand/walk unaided, direction-changing or vertical/torsional nystagmus, and any associated brainstem sign. In the acute vestibular syndrome (continuous vertigo with nystagmus, nausea and head-motion intolerance), the bedside HINTS battery outperforms early MRI-DWI for distinguishing central from peripheral causes:

• Head Impulse: a normal (no corrective saccade) horizontal head-impulse test is paradoxically worrying—it points central; an abnormal test (corrective saccade) suggests a peripheral vestibulopathy.
• Nystagmus: direction-changing or vertical/torsional nystagmus is central.
• Test of Skew: any vertical skew deviation on cover-uncover is central.

The mnemonic INFARCT (Impulse Normal, Fast-phase Alternating, Refixation on Cover Test) flags a central lesion. HINTS is validated only for continuous acute vestibular syndrome with nystagmus, applied by trained examiners—not for episodic dizziness. Remember early DWI misses ~15–20% of posterior-fossa infarcts in the first 24–48 h; a negative scan does not exclude the diagnosis, and repeat imaging is warranted when suspicion is high.

Diagnostic workup

MRI-DWI, and CTA/MRA of the vertebrobasilar system to detect stenosis, dissection, or basilar occlusion. Basilar occlusion is a neurological emergency—low threshold for vessel imaging in any patient with fluctuating brainstem signs, decreased consciousness, or a “herald” pattern.

Management

IV thrombolysis by standard criteria. For basilar artery occlusion, endovascular thrombectomy is supported by BAOCHE and ATTENTION (benefit within extended windows), following the earlier neutral BASICS trial; the wide time window reflects the devastating natural history of untreated basilar occlusion. Suboccipital decompression is life-saving in malignant cerebellar infarction with brainstem compression or obstructive hydrocephalus.

Prognosis

Highly variable—from full recovery in a small cerebellar infarct to fatal or locked-in outcomes in basilar occlusion. Early recognition and reperfusion are the main modifiable determinants.

Clinical pearls

• “Crossed” deficits localize to the brainstem; the crossing cranial nerve identifies the level.
• A patient who cannot stand or walk despite seemingly “benign” vertigo has a central lesion until proven otherwise.
• A normal head-impulse test in acute vestibular syndrome is a red flag, not reassurance.
• Sudden bilateral thalamic dysfunction with vertical gaze palsy = think artery of Percheron.

Key references: Kattah et al., HINTS (Stroke 2009). Caplan LR, Posterior Circulation Ischemia. Tao et al., ATTENTION (NEJM 2022); Jovin et al., BAOCHE (NEJM 2022); Langezaal et al., BASICS (NEJM 2021).

Definition & epidemiology

A cryptogenic stroke is an ischemic stroke with no identified cause after standard evaluation; depending on the depth of workup it represents 20–40% of ischemic strokes. ESUS (Hart et al., Cryptogenic Stroke/ESUS International Working Group, 2014) is a more specific construct: a non-lacunar infarct presumed embolic in the absence of a demonstrable high-risk cardioembolic source or significant proximal artery stenosis. The premise was that if these strokes are embolic, anticoagulation might outperform antiplatelets—a hypothesis that has since been tested and largely refuted.

Criteria

The minimum diagnostic workup to assign ESUS includes brain imaging, 12-lead ECG plus ≥24 h cardiac rhythm monitoring, transthoracic echocardiography, and imaging of the extracranial and intracranial arteries.

Diagnostic workup & candidate mechanisms

Because ESUS is a “source-not-yet-found” label, evaluation focuses on the plausible occult sources: covert paroxysmal AF (prolonged monitoring/ICM), atrial cardiopathy even without AF (LA enlargement, elevated NT-proBNP, P-wave terminal force in V1), PFO (especially in younger patients), non-stenotic but complex/ulcerated carotid or aortic arch plaque, occult malignancy with hypercoagulability (marantic endocarditis, elevated D-dimer, multiterritory infarcts), and left ventricular disease.

Secondary prevention evidence

Three landmark randomized trials failed to show benefit of empiric anticoagulation over antiplatelet therapy in unselected ESUS, so antiplatelet therapy remains standard:

• NAVIGATE-ESUS (rivaroxaban 15 mg vs aspirin): stopped early—no reduction in recurrent stroke and more bleeding.
• RESPECT-ESUS (dabigatran vs aspirin): no significant reduction in recurrent stroke.
• ARCADIA (apixaban vs aspirin in ESUS with biomarker-defined atrial cardiopathy): neutral—anticoagulation did not reduce recurrence even in this enriched subgroup, undercutting the atrial-cardiopathy-as-treatable-target hypothesis.

For ESUS attributed to a PFO in a younger patient, percutaneous PFO closure plus antiplatelet therapy reduces recurrence versus medical therapy alone (RESPECT long-term, CLOSE, REDUCE, DEFENSE-PFO); the RoPE score and anatomic features (large shunt, associated atrial septal aneurysm) help identify a probably causal PFO. Closure is generally offered to patients roughly 18–60 years old with a nonlacunar cryptogenic stroke and a high-risk PFO after multidisciplinary evaluation.

Differential diagnosis

Reclassification is common on follow-up: newly detected AF, an emergent malignancy, or a delayed diagnosis of dissection frequently converts an “ESUS” into a defined subtype—hence the value of prolonged monitoring and vigilance for systemic clues.

Management & prognosis

Antiplatelet therapy, aggressive vascular risk-factor modification, high-intensity statin, and prolonged rhythm monitoring to capture treatable AF; PFO closure in selected younger patients. Annual recurrence is meaningful (~4–5%), underscoring that ESUS is a call to keep searching rather than a final etiology.

Clinical pearls

• ESUS is not an indication for empiric anticoagulation—three RCTs (NAVIGATE, RESPECT-ESUS, ARCADIA) were neutral.
• Continue looking: the most valuable “treatment” is often the extended monitor that finds AF.
• Multiterritory ESUS with elevated D-dimer should trigger a malignancy/hypercoagulability search.

Key references: Hart et al., ESUS definition (Lancet Neurology 2014). Hart et al., NAVIGATE-ESUS (NEJM 2018). Diener et al., RESPECT-ESUS (NEJM 2019). Kamel et al., ARCADIA (JAMA 2024). Kent et al., RoPE score; Saver et al., RESPECT long-term (NEJM 2017).

Definition & epidemiology

Stroke in the young is typically defined as ischemic stroke before age 50 (some series use <45). It represents roughly 10–15% of ischemic strokes, and its incidence has been rising. The etiologic spectrum differs markedly from that of older patients—dissection, cardioembolism through a PFO, and uncommon vasculopathies and hematologic disorders replace atherosclerosis as leading causes—so the workup must be correspondingly broad, and a substantial proportion remain cryptogenic.

Pathophysiology & etiologic spectrum

Diagnostic workup

Layer the evaluation on top of the standard stroke workup:

• Vessel imaging with attention to the arterial wall: CTA/MRA head and neck for dissection, vasculopathy, moyamoya; high-resolution vessel-wall MRI to distinguish vasculitis/RCVS/dissection; catheter angiography when non-invasive imaging is inconclusive.
• Cardiac: TTE and TEE with agitated-saline (bubble) study for PFO/shunt; prolonged rhythm monitoring.
• Thrombophilia and immunologic panel: antiphospholipid antibodies (lupus anticoagulant, anticardiolipin, anti-beta2-glycoprotein I—repeat at 12 weeks), inherited thrombophilia when venous/paradoxical mechanism suspected, ANA/ENA, ESR/CRP, homocysteine.
• Targeted testing by phenotype: genetic testing (NOTCH3, mitochondrial mutations), alpha-galactosidase A/leukocyte enzyme for Fabry, lactate, CSF for vasculitis/infection, toxicology, pregnancy testing.

Management & prognosis

Treat the specific cause: antithrombotics (antiplatelet or anticoagulation, comparably effective for extracranial dissection per CADISS/TREAT-CAD), immunosuppression for vasculitis, calcium-channel blockers and trigger removal for RCVS, transfusion/hydroxyurea for sickle cell, enzyme replacement for Fabry, PFO closure in selected patients. Prognosis is generally better than in older cohorts, but long-term risks of recurrence, epilepsy, cognitive and psychosocial morbidity are substantial and warrant sustained follow-up.

Clinical pearls

• Neck pain or headache preceding a young person's stroke suggests dissection until vessel imaging proves otherwise.
• Thunderclap headache with multifocal segmental arterial narrowing that reverses = RCVS, not vasculitis—steroids may worsen it.
• Confirm antiphospholipid antibodies on a repeat sample ≥12 weeks apart before committing to lifelong anticoagulation.
• Consider Fabry, MELAS, and CADASIL when strokes recur without conventional risk factors or with suggestive systemic/imaging clues.

Key references: Ferro et al.; Ekker et al., trends in young stroke (Lancet Neurology). CADISS trial (Lancet Neurology 2015); TREAT-CAD (Lancet Neurology 2021). 2021 AHA/ASA Secondary Prevention Guideline (Stroke).

Definition & epidemiology

Capsular warning syndrome (CWS) is a distinctive TIA pattern: recurrent, stereotyped, crescendo episodes of transient motor (or sensorimotor) deficit—typically involving face, arm and leg without cortical signs—referable to the internal capsule or corona radiata. It is uncommon (a small percentage of TIAs) but clinically important because it carries an exceptionally high short-term risk of progressing to a completed capsular lacunar infarct, on the order of 40–60% within a week, far exceeding the risk predicted by conventional TIA scores.

Pathophysiology & vascular anatomy

The prevailing mechanism is intermittent hemodynamic compromise of a single deep perforator—most often a lenticulostriate branch supplying the internal capsule, or a paramedian pontine perforator in the analogous “pontine warning syndrome.” Small-vessel disease at the perforator origin (microatheroma or lipohyalinosis) produces a critically narrowed end-artery with negligible collateral, so the territory oscillates between ischemia and reperfusion, generating repetitive stereotyped spells before infarction becomes fixed.

Clinical features

Repetitive, near-identical episodes (often three or more within 24–72 h) of pure motor or sensorimotor hemiparesis, characteristically without aphasia, neglect, or visual field loss. Between spells the patient may return to baseline. The stereotyped, escalating pattern is the diagnostic signature and should be treated as an emergency rather than reassuring transient events.

Differential diagnosis

Focal motor seizures (usually positive phenomena, march, post-ictal deficit—capture on EEG when uncertain), hemiplegic migraine, branch atheromatous disease with fluctuating course, and large-artery flow-limiting stenosis producing limb-shaking TIAs (which are usually posture/exertion related and point to carotid disease rather than a perforator).

Diagnostic workup

Urgent MRI-DWI (initially often negative, since infarction may not yet be established), MRA/CTA to exclude a proximal stenotic source, and cardiac evaluation. Consider EEG if seizure is plausible. Early DWI negativity does not exclude CWS and should not lower vigilance.

Management

There is no high-quality randomized evidence; management is pragmatic and aggressive. Approaches include short-course dual antiplatelet therapy (aspirin + clopidogrel), and—because the mechanism is thought to be hemodynamic—permissive hypertension / avoidance of aggressive blood-pressure lowering to maintain perforator perfusion. Case reports describe IV thrombolysis during an acute spell, but efficacy is unproven and must be weighed individually. Some centers use short-term intravenous antithrombotics (e.g., tirofiban or argatroban) in refractory crescendo cases. Admit for monitoring; deficits may fix rapidly.

Prognosis

High early risk of a completed capsular infarct with fixed deficit; once infarction occurs, subsequent recovery follows that of any lacunar stroke. Recognizing CWS is the key opportunity to intervene before the deficit becomes permanent.

Clinical pearls

• Stereotyped, repetitive pure-motor TIAs are a neurologic emergency—do not discharge on the basis of a normal early DWI.
• Resist reflexive blood-pressure lowering; the deficit is thought to be perfusion-dependent.
• The pontine equivalent (“pontine warning syndrome”) behaves similarly and carries the same high conversion risk.

Key references: Donnan et al., capsular warning syndrome (Neurology). Paul et al., outcomes and recurrence in CWS (Stroke). Camps-Renom et al., pontine warning syndrome series.

Definition & epidemiology

Branch atheromatous disease, a concept advanced by Caplan, refers to infarction from occlusion of a large-caliber deep penetrating artery at or near its origin, caused by microatheroma or a junctional plaque of the parent artery blocking the ostium of the branch. It occupies a mechanistic middle ground between classic lipohyalinotic lacunar disease and large-artery atherosclerosis, and it is increasingly recognized as a cause of “lacunar-appearing” strokes that behave badly.

Pathophysiology & vascular anatomy

Two territories are archetypal:

• Lenticulostriate territory: ostial atheroma at the origin of a lateral lenticulostriate branch from the M1 MCA produces a comma-shaped striatocapsular infarct that typically extends across several axial slices—larger than a lipohyalinotic lacune.
• Paramedian pontine territory: atheroma at the ostium of a paramedian perforator from the basilar produces a pontine infarct that characteristically abuts the ventral pontine surface (rather than sitting deep within the pons), a helpful imaging clue.

Because a single larger perforator is occluded at its mouth, the infarct is bigger and the penumbra more vulnerable to progression than in typical small-vessel disease.

Clinical features

Presents like a lacunar syndrome (often pure motor or sensorimotor, without cortical signs) but with two distinguishing behaviors: early neurological deterioration / stepwise or fluctuating progression in a substantial fraction of patients over the first hours to days, and a larger final infarct volume. This progressive course is the clinical hallmark that separates BAD from a stable lacune.

Differential diagnosis

Diagnostic workup

MRI-DWI defines the size and surface relationship (ventral pontine contact); high-resolution vessel-wall MRI can sometimes demonstrate the parent-artery plaque at the perforator origin; CTA/MRA excludes a flow-limiting proximal stenosis or embolic mimic. Monitor closely for progression during the first 72 h.

Management

No dedicated large RCT defines therapy; management extrapolates from minor-stroke antithrombotic data with attention to the high rate of early progression. Options include short-course dual antiplatelet therapy and aggressive risk-factor control; some centers use intravenous antithrombotic agents (argatroban or the GP IIb/IIIa inhibitor tirofiban) for progressive deficits, though evidence is limited and evolving. Avoid excessive blood-pressure lowering acutely given probable perfusion-dependence.

Prognosis

Worse than classic lacunar stroke because of early deterioration and larger infarcts; identifying BAD prospectively allows closer monitoring and earlier escalation.

Clinical pearls

• A “lacunar” deficit that worsens stepwise over hours is BAD until proven otherwise.
• A pontine infarct touching the ventral surface points to a paramedian perforator ostial atheroma (BAD), not a deep lacune.
• Counsel patients and teams to expect possible progression and monitor accordingly.

Key references: Caplan LR, Intracranial branch atheromatous disease (Neurology 1989). Petrone et al., branch atheromatous disease review (Stroke). Nah et al., striatocapsular infarct topography and mechanism.

Definition & epidemiology

TIA is now defined on a tissue basis (AHA/ASA 2009): a transient episode of neurologic dysfunction caused by focal brain, spinal cord, or retinal ischemia without acute infarction on imaging. This replaced the arbitrary time-based definition (deficit <24 h), because up to a third of clinically “transient” events show DWI infarction and are therefore strokes. TIA is a medical emergency: the 90-day stroke risk without treatment approaches 10–17%, with much of that risk concentrated in the first 48 hours—a warning that demands the same urgency as an acute stroke.

Pathophysiology & vascular anatomy

Mechanisms mirror those of ischemic stroke: artery-to-artery embolism from carotid or intracranial plaque, cardioembolism, small-vessel low-flow, and hemodynamic TIAs from critical stenosis (e.g., limb-shaking TIA of severe carotid disease; amaurosis fugax from retinal/ophthalmic artery embolism). The territory of symptoms guides the vascular search (carotid vs vertebrobasilar).

Risk stratification — the ABCD2 score

Total 0–7; higher scores (≥4) predict higher early stroke risk. ABCD2 is a useful triage aid but should not gate imaging or specialist referral—it misses cause-specific high-risk features (carotid stenosis, AF, DWI positivity), which independently mandate urgent evaluation regardless of score.

Differential diagnosis

Distinguish from TIA mimics: focal seizure with post-ictal deficit, migraine aura (positive, spreading, prolonged), transient global amnesia, hypoglycemia, syncope/pre-syncope, peripheral vestibular vertigo, functional episodes, and structural lesions producing intermittent symptoms. “Negative” symptoms of sudden onset without march favor ischemia; spreading positive phenomena favor migraine or seizure.

Diagnostic workup (rapid, ideally same-day)

• Urgent MRI with DWI (defines infarction and often the mechanism) or CT if MRI unavailable.
• Cervical and intracranial vessel imaging (CTA/MRA or carotid ultrasound) to identify symptomatic stenosis amenable to urgent revascularization.
• ECG and cardiac monitoring for AF; echocardiography as indicated; basic labs including glucose.
• Structured urgent pathways (the EXPRESS study and SOS-TIA) demonstrated that immediate assessment and treatment reduce subsequent stroke by roughly 80%, justifying rapid-access TIA clinics or admission.

Management

For high-risk TIA (ABCD2 ≥4) or minor ischemic stroke (NIHSS ≤3–5), start short-course dual antiplatelet therapy promptly:

• CHANCE and POINT established aspirin + clopidogrel started within 24 h reduces recurrent stroke; the benefit is front-loaded and bleeding accrues with time, so DAPT is limited to about 21 days (POINT/CHANCE pooled analysis) before de-escalation to monotherapy.
• INSPIRES extended the treatment window to within 72 h for minor stroke/high-risk TIA of presumed atherosclerotic origin, again favoring clopidogrel–aspirin over aspirin alone (with more, mostly minor, bleeding).
• CHANCE-2 showed that in carriers of CYP2C19 loss-of-function alleles (clopidogrel non-responders), ticagrelor + aspirin was superior to clopidogrel + aspirin—supporting genotype-guided selection where testing is available.

Add high-intensity statin, treat hypertension and diabetes, anticoagulate if AF is found, and pursue urgent carotid revascularization for symptomatic high-grade stenosis (ideally within 2 weeks).

Prognosis

Early treatment transforms outcome: the highest-risk window is the first 48 h, and structured urgent care with DAPT, statin, risk-factor control and revascularization dramatically lowers 90-day stroke risk.

Clinical pearls

• “Transient” symptoms with DWI lesion = stroke, not TIA—image everyone.
• A low ABCD2 does not license delay if there is carotid stenosis, AF, or a DWI lesion.
• Cap DAPT at ~21 days (up to 90 in selected high-risk atherosclerotic cases per INSPIRES) to avoid net harm from bleeding.
• Consider CYP2C19 status: ticagrelor–aspirin for loss-of-function carriers (CHANCE-2).

Key references: Easton et al., tissue-based TIA definition (Stroke 2009). Johnston et al., POINT (NEJM 2018); Wang et al., CHANCE (NEJM 2013); CHANCE-2 (NEJM 2021); INSPIRES (NEJM 2023). Rothwell et al., EXPRESS (Lancet 2007); Lavallee et al., SOS-TIA (Lancet Neurology 2007).

Definition & epidemiology

Transient global amnesia is a striking, self-limited amnestic syndrome: the abrupt onset of profound anterograde amnesia (inability to form new memories) with variable retrograde amnesia, in an otherwise alert patient with preserved personal identity and no other focal neurologic deficit, resolving within 24 hours. It typically affects adults 50–70 years old and often follows an identifiable precipitant (emotional stress, physical exertion, Valsalva, immersion in cold or hot water, sexual intercourse, or a medical procedure). The hallmark at the bedside is repetitive questioning (“Where am I? What are we doing?”) as each answer fails to consolidate.

Criteria (Hodges & Warlow)

Pathophysiology & imaging

The mechanism is debated; leading hypotheses include transient dysfunction of the hippocampal CA1 (Sommer sector, selectively vulnerable to metabolic stress), venous congestion of the mesial temporal lobe (association with internal jugular venous valve incompetence and Valsalva), and a spreading-depression–like phenomenon. Characteristically, MRI shows small punctate DWI hyperintensities in the hippocampus (CA1) that are frequently not visible in the first hours—sensitivity peaks around 24–72 hours, favored by thin-section high-b-value 3T imaging. These transient “dots” support the diagnosis but their absence does not exclude it, and they typically resolve without leaving a permanent lesion.

Differential diagnosis

Diagnostic workup

The diagnosis is clinical. Obtain glucose and basic labs, and MRI-DWI when the presentation is atypical or to document hippocampal changes; consider EEG when recurrence, brevity, or stereotypy suggests transient epileptic amnesia. Extensive vascular or cardiac workup is generally unnecessary for a classic single episode, since TGA is not a marker of increased stroke risk.

Management & prognosis

No specific treatment is required; management is reassurance and observation until the episode resolves. TGA is benign: memory returns to normal (a permanent gap remains only for the event itself), and the recurrence rate is low (roughly 3–6% per year). Frequent, brief, recurrent spells should prompt reconsideration of transient epileptic amnesia.

Clinical pearls

• Preserved identity + isolated anterograde amnesia + repetitive questioning = TGA; loss of identity suggests a psychogenic cause.
• Early DWI may be normal—the hippocampal “dot” is most sensitive at 24–72 h and is confirmatory, not required.
• Short, recurrent, stereotyped amnestic spells (especially on waking) point to transient epileptic amnesia—get an EEG.
• TGA does not confer excess stroke risk and does not warrant an aggressive vascular workup.

Key references: Hodges & Warlow, diagnostic criteria (J Neurol Neurosurg Psychiatry 1990). Bartsch et al., CA1 hippocampal lesions in TGA (Brain). Szabo et al., DWI timing in TGA. Zeman et al., transient epileptic amnesia.

Definition & epidemiology

Stroke mimics are non-vascular conditions producing acute focal neurologic deficits that resemble stroke; they represent roughly 20–30% of suspected strokes at presentation and up to a comparable fraction of “stroke code” activations. Distinguishing mimics matters for avoiding unnecessary treatment—yet, importantly, thrombolysis inadvertently given to mimics carries a low symptomatic hemorrhage risk, so uncertainty should not delay treatment of a plausible ischemic stroke within the window.

Clinical features & distinguishing clues

Diagnostic approach

A focused history (onset tempo, positive vs negative symptoms, precipitants, prior episodes) and examination for non-anatomic or functional signs, plus an immediate glucose, resolve many cases. MRI-DWI is the most powerful discriminator when the diagnosis is uncertain. Prediction aids such as the FABS tool and TeleStroke mimic scores can support triage but do not replace clinical judgment or override the imperative to treat probable stroke within the window.

Management principle

When a plausible acute ischemic stroke is within the treatment window and imaging excludes hemorrhage, do not withhold thrombolysis merely because a mimic is possible—the risk of harm from treating a mimic is low, whereas the cost of missing a true stroke is high. Conversely, recognizing a clear mimic (documented hypoglycemia, obvious seizure with a positive march) avoids needless risk.

Clinical pearls

• Check a fingerstick glucose before every thrombolysis decision—hypoglycemia is an eminently reversible mimic.
• Forehead sparing distinguishes a central (stroke) facial palsy from Bell palsy.
• Spreading, marching, positive symptoms favor migraine or seizure over ischemia.
• Functional deficits show internal inconsistency (Hoover sign, give-way weakness, distractibility)—but remain a diagnosis of caution, not exclusion.

Key references: Libman et al., conditions mimicking stroke. Goyal et al./Zinkstok et al., safety of thrombolysis in mimics. Vilela P, stroke mimics review (Eur J Radiol). Powers et al., 2019 AHA/ASA acute stroke guideline.

Definition & epidemiology

Malignant middle cerebral artery infarction is a large hemispheric infarct (typically ≥50% of the MCA territory, often with additional ACA/PCA involvement) that develops space-occupying cytotoxic edema, producing progressive mass effect, midline shift, and transtentorial (uncal/central) herniation. It complicates roughly 1–10% of supratentorial infarcts and, untreated, carries mortality approaching 70–80%, most deaths occurring between days 2 and 5 as edema peaks. Recognizing it early—ideally before herniation—is the entire clinical point, because decompressive surgery is time-sensitive.

Pathophysiology & vascular anatomy

The substrate is proximal (terminal ICA or M1) occlusion with poor collaterals, giving a complete MCA-territory infarct. Cytotoxic edema accumulates over 2–5 days; because the adult cranium is a fixed volume, tissue swelling raises intracranial pressure and shifts structures, compressing the diencephalon and brainstem and secondarily infarcting the PCA territory and the contralateral hemisphere. Younger patients (with less brain atrophy and therefore less compensatory reserve) are paradoxically at higher risk of fatal herniation.

Clinical features — recognizing deterioration

• Severe deficit at onset: high NIHSS (often >15 in the non-dominant and >20 in the dominant hemisphere), dense hemiplegia, forced eye and head gaze deviation toward the lesion, global aphasia or hemineglect.
• Declining level of consciousness over 24–96 h—the cardinal warning sign—often with early nausea/vomiting and headache.
• Evolving pupillary asymmetry (ipsilateral fixed, dilated pupil from CN III compression), Cushing response, and posturing signal impending or established herniation—late findings that should already have prompted intervention.

Diagnostic workup & radiographic predictors

Admit high-risk patients to a neuro-ICU/stroke unit for close neurologic monitoring and serial imaging; the goal is to detect the trajectory toward malignant edema before consciousness is lost.

Management

Medical measures (osmotherapy with hypertonic saline or mannitol, head-of-bed elevation, normocapnia, avoidance of hypotonic fluids and fever) temporize but do not prevent herniation in a truly malignant course. Definitive treatment is early decompressive hemicraniectomy (see the dedicated surgical topic for full detail). The pooled analysis of DECIMAL, DESTINY, and HAMLET showed that hemicraniectomy within 48 h in patients ≤60 years substantially reduces mortality and improves the odds of favorable outcome; DESTINY II extended a survival benefit to patients >60 years, though with more survivors left moderately-to-severely disabled—making shared decision-making about acceptable outcomes essential.

Prognosis

Without surgery, malignant MCA infarction is usually fatal. With timely hemicraniectomy, mortality falls dramatically, but many survivors—especially older patients—live with significant disability; outcome depends on age, hemispheric dominance, timing, and infarct extent.

Clinical pearls

• A large MCA infarct that is quiet on day 1 can herniate on day 3—anticipate the delayed edema peak and monitor accordingly.
• Falling consciousness, not the pupil, is the signal to act; a blown pupil is already late.
• Early DWI ≥~80 cm³ and >50% CT hypodensity are the most useful early red flags—consult neurosurgery proactively.
• Discuss goals of care early, particularly in patients >60, given the disability profile of survivors.

Key references: Vahedi et al., pooled DECIMAL/DESTINY/HAMLET analysis (Lancet Neurology 2007). Juttler et al., DESTINY II (NEJM 2014). Wijdicks et al., AHA/ASA scientific statement on management of cerebral and cerebellar infarction with swelling (Stroke 2014).

Definition & purpose

The Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification, published in 1993, is the most widely used etiologic scheme for ischemic stroke. It assigns each stroke to one of five mechanistic categories based on clinical features and the results of ancillary testing (brain imaging, vascular imaging, cardiac evaluation, and laboratory workup). It was designed for clinical trials but has become the standard vocabulary for describing stroke etiology in practice and research.

Categories & criteria

Use & limitations

• Strengths: simple, widely adopted, provides a common framework for trials and epidemiology; reasonable inter-rater reliability for the major categories.
• Limitations: the “undetermined” category is heterogeneous and large, lumping together truly cryptogenic strokes, incompletely worked-up strokes, and those with competing causes; it forces a single mechanism even when several coexist; the atherosclerosis threshold (≥50%) ignores non-stenotic but unstable plaque; and category assignment depends heavily on how thoroughly the patient was investigated, creating variability across centers. These shortcomings motivated the more granular successor systems (SSS-TOAST/CCS, ASCOD, CISS).

Key references: Adams et al., Classification of subtype of acute ischemic stroke: TOAST (Stroke 1993). Goldstein et al., inter-rater reliability of TOAST.

Definition & purpose

The Stop Stroke Study TOAST (SSS-TOAST), and its computerized implementation the Causative Classification System (CCS) developed by Ay and colleagues, refine TOAST to reduce the size of the “undetermined” group and improve reproducibility. Two innovations distinguish it: it grades each mechanism by the weight of evidence that it actually caused the stroke, and it is delivered as a web-based algorithm that standardizes the assignment across raters by converting inputs into a category automatically.

Categories & evidentiary weighting

CCS retains the five familiar etiologic groups but subdivides each by causal likelihood, and separately reports a “causative” (most probable single cause) versus a “phenotypic” (all abnormalities present) classification.

Use & limitations

• Strengths: higher inter-rater reliability than original TOAST; the causative-versus-phenotypic distinction preserves information about multiple coexisting mechanisms; the “evident/probable/possible” grading captures diagnostic certainty; the automated web tool reduces subjective drift and improves reproducibility for multicenter research.
• Limitations: more complex and time-consuming than TOAST; still depends on completeness of the diagnostic workup; requires the software/algorithm for consistent application; less familiar to non-researchers.

Key references: Ay et al., A computerized algorithm for etiologic classification of ischemic stroke: the Causative Classification System (Stroke 2007). Ay et al., An evidence-based causative classification system for acute ischemic stroke (Ann Neurol 2005).

Definition & purpose

The Oxfordshire Community Stroke Project (OCSP) classification, described by Bamford and colleagues in 1991, is a purely clinical, bedside scheme that categorizes ischemic stroke by the pattern of maximal deficit—before imaging—into four syndromes. Its enduring value is that it predicts infarct territory, likely etiology, and, notably, prognosis using only the neurologic examination, making it useful for rapid triage and for outcome estimation in resource-limited or hyperacute settings.

Categories & criteria

Use & limitations

• Strengths: requires no investigations—pure bedside classification; strongly prognostic (TACS predicts high mortality/dependency, LACS the best early outcome); useful for triage, communication, and epidemiology.
• Limitations: it classifies the clinical syndrome/territory, not the etiologic mechanism, so it does not direct secondary prevention by itself; clinical–radiological mismatch occurs (a “LACS” may prove cortical on MRI and vice versa); it can misclassify small or resolving deficits; and it does not incorporate modern imaging or vessel data.

Key references: Bamford et al., Classification and natural history of clinically identifiable subtypes of cerebral infarction (Lancet 1991). Pittock et al., validation of the OCSP classification.

Definition & purpose

Both ASCOD and CISS move beyond assigning a single “winner” mechanism. ASCOD (Amarenco et al.), an evolution of the earlier ASCO scheme, is a phenotyping system: rather than choosing one cause, it documents every potential mechanism present in a given patient and grades the causal likelihood of each. CISS (Gao et al.) is an etiology-and-mechanism system developed in China that gives special emphasis to intracranial atherosclerosis (highly prevalent in Asian populations) and to the specific mechanism by which a large-artery plaque causes infarction.

ASCOD — categories & grading

ASCOD records five phenotypic axes for every patient, each graded independently, so a patient can carry, for example, an A1 C3 O0 profile that transparently displays coexisting mechanisms.

Each axis is graded by the strength of the causal relationship: 0 = disease absent; 1 = present and potentially causal; 2 = present, causal link uncertain; 3 = present but unlikely causal (incidental); 9 = incomplete workup. This preserves the full diagnostic picture and is well suited to patients with multiple competing risk factors.

CISS — categories & mechanism

CISS first classifies etiology, then—crucially—the underlying mechanism when large-artery atherosclerosis is responsible:

Use & limitations

• Strengths: ASCOD's phenotyping avoids discarding information when several mechanisms coexist and is ideal for research and for patients with mixed disease; CISS explicitly separates the mechanism of large-artery stroke (penetrator occlusion vs artery-to-artery embolism vs hypoperfusion), which can directly inform management, and it captures intracranial atherosclerosis and branch atheromatous disease that TOAST tends to misclassify as “small-vessel.”
• Limitations: both are more complex and require detailed imaging (including vessel-wall/perforator assessment for CISS); they are less familiar and less used in Western trials; grading still depends on workup completeness; and phenotyping does not by itself designate the single actionable cause a clinician may need for a treatment decision.

Key references: Amarenco et al., The ASCOD phenotyping of ischemic stroke (Cerebrovasc Dis 2013); Amarenco et al., ASCO (Cerebrovasc Dis 2009). Gao et al., Chinese Ischemic Stroke Subclassification (CISS) (Front Neurol 2011).

Definition & purpose

The National Institutes of Health Stroke Scale (NIHSS) is the standard quantitative measure of stroke severity, scored 0–42 across 11 items (level of consciousness, gaze, visual fields, facial and limb strength, ataxia, sensation, language, dysarthria, and extinction/inattention). It is used to communicate severity, guide acute treatment decisions and eligibility, track neurologic change, and predict outcome. Grouping scores into severity bands provides a shared shorthand, and defining “minor” stroke has direct therapeutic consequences (antiplatelet vs thrombolysis decisions, DAPT eligibility).

Categories — NIHSS severity bands

Criteria — defining “minor stroke”

There is no single universal cutoff; commonly used operational definitions include:

• NIHSS ≤3 or ≤5—the most frequent numeric thresholds in trials and practice.
• Non-disabling deficit—a definition based on function rather than a number: the deficit does not impair basic activities of daily living or would not itself preclude independence, regardless of the exact NIHSS. This distinction underlies thrombolysis decisions in mild stroke (e.g., the PRISMS trial addressed alteplase in non-disabling deficits).
• For DAPT eligibility (CHANCE/POINT/INSPIRES), “minor stroke” is typically NIHSS ≤3 (POINT/CHANCE) or ≤5 (INSPIRES), paired with high-risk TIA (ABCD2 ≥4).

Use & limitations

• Strengths: reproducible, quick, validated, strongly predictive of outcome and of large-vessel occlusion (higher scores raise LVO likelihood); a universal language for triage and trials.
• Limitations: the NIHSS is weighted toward the dominant (left) hemisphere—language items contribute heavily, so a disabling non-dominant (right) hemispheric stroke with neglect can score deceptively low; it under-captures posterior circulation and brainstem deficits (a life-threatening basilar stroke may have a modest score); it does not measure disability from a “low but disabling” deficit (isolated hemianopia, aphasia, or hand weakness affecting occupation); and inter-rater reliability, though good with certification, varies for ataxia and dysarthria items. Severity bands therefore inform, but do not replace, individualized clinical and functional assessment.

Key references: Brott et al., original NIHSS (Stroke 1989). Lyden P, Using the NIHSS (Stroke 2017). Khatri et al., PRISMS trial of alteplase in minor non-disabling stroke (JAMA 2018). Powers et al., 2019 AHA/ASA acute ischemic stroke guideline.

Rationale & evidence

Roughly one in four strokes occurs in a person who has already had a cerebrovascular event, and the highest recurrence hazard is concentrated in the first days to weeks. Untreated, the 90-day recurrence risk after minor stroke or high-risk TIA approaches 10–18%; structured secondary prevention lowers relative recurrence by more than 80% when antithrombotic, blood-pressure, lipid and lifestyle interventions are combined. The organising principle of modern secondary prevention is that the mechanism dictates the medicine: a rational regimen cannot be chosen until the stroke is phenotyped. Expedited work-up (vessel imaging of the arch-to-vertex circulation, cardiac rhythm monitoring, echocardiography, and targeted laboratory testing) should therefore begin during the index admission rather than after discharge.

Etiologic classification typically follows the TOAST framework — large-artery atherosclerosis, cardioembolism, small-vessel (lacunar) disease, other determined cause, and undetermined/cryptogenic — refined by the embolic stroke of undetermined source (ESUS) construct and the ASCOD phenotyping system, which grades every co-existing mechanism rather than forcing a single label. Up to 25–40% of ischemic strokes remain cryptogenic after standard evaluation, and a substantial fraction of these harbour occult paroxysmal atrial fibrillation detectable only with prolonged monitoring (CRYSTAL-AF, EMBRACE, STROKE-AF, PER DIEM).

Therapy — matching mechanism to intervention

Targets — the universal bundle

Regardless of subtype, every patient should receive: blood pressure <130/80 mm Hg; LDL-C <70 mg/dL (<1.8 mmol/L) for atherosclerotic mechanisms, with <55 mg/dL (<1.4 mmol/L) an ESC option for very-high-risk patients; HbA1c individualised near <7%; complete tobacco cessation; Mediterranean-style diet; 10 or more MET-hours of physical activity weekly; and treatment of obstructive sleep apnoea and heavy alcohol use. The AHA/ASA 2021 guideline frames these as class I interventions applicable across mechanisms.

Clinical pearls

• Do not discharge a "cryptogenic" stroke without a monitoring plan — the diagnostic label often changes the drug (antiplatelet becomes anticoagulant) once AF is captured.
• Antiplatelet plus anticoagulant combinations should be time-limited and justified (e.g., recent PCI); routine dual pathway therapy after stroke adds bleeding without ischemic benefit for most.
• The steepest recurrence curve is early: statin, antiplatelet and BP therapy should start in hospital, not at the first outpatient visit.
• Adherence is the rate-limiting step; polypharmacy simplification and a single-visit initiation strategy (as in the EXPRESS/SOS-TIA models) sharply cut recurrence.

Key references: Kleindorfer et al. AHA/ASA 2021 Guideline for the Prevention of Stroke in Patients With Stroke and TIA (Stroke 2021); Adams et al. TOAST (Stroke 1993); Hart et al. ESUS construct (Lancet Neurol 2014); ESO Guidelines on secondary prevention.

Rationale & evidence

The INTERSTROKE case-control study (nearly 27,000 participants across 32 countries) established that ten largely modifiable risk factors collectively account for about 90% of the population-attributable risk (PAR) of stroke — 89.7% for ischemic and 88.1% for intracerebral haemorrhage. Because PAR reflects both the strength of association and the prevalence of a factor in the population, common exposures such as hypertension dominate even when their individual odds ratios are moderate. The clinical corollary is that most strokes are preventable, and that population-level risk-factor control yields larger absolute gains than any single pharmacologic advance.

Targets — population-attributable risk (INTERSTROKE, all stroke)

Non-modifiable factors

• Age — the dominant non-modifiable factor; risk roughly doubles each decade after 55.
• Sex — higher lifetime incidence in men at most ages, but women have more events in absolute terms owing to longevity, plus sex-specific exposures (pregnancy/pre-eclampsia, hormonal therapy, migraine with aura).
• Race/ethnicity — higher incidence and earlier onset in Black, Hispanic and East Asian populations; intracranial atherosclerosis more prevalent in Asian, Black and Hispanic groups.
• Genetics/family history — polygenic burden plus monogenic syndromes (CADASIL, Fabry, sickle cell, familial hypercholesterolaemia); a positive family history roughly doubles risk.
• Prior stroke/TIA — the single strongest clinical predictor of a future event.

Clinical pearls

• PAR figures overlap (a patient may carry several factors), so summed percentages exceed the total attributable fraction; the practical message is cumulative, multiplicative risk.
• The relative contribution of factors shifts by region and stroke subtype — hypertension and diet dominate globally, whereas cardioembolic and lipid factors weigh more heavily in high-income settings.
• Targeting a single high-prevalence factor (blood pressure) across a population averts more strokes than perfecting management of a rare high-risk one.

Key references: O'Donnell et al. INTERSTROKE (Lancet 2010; and the expanded 2016 analysis, Lancet 2016); GBD Stroke Collaborators (Lancet Neurol 2021); Feigin et al. global risk-factor burden.

Rationale & evidence

Hypertension is the most important modifiable risk factor for both ischemic and haemorrhagic stroke, with a continuous log-linear relationship between usual blood pressure and stroke risk down to at least 115/75 mm Hg — each 20/10 mm Hg increment roughly doubles stroke mortality. In secondary prevention, PROGRESS (perindopril ± indapamide, >6,100 patients with prior stroke/TIA) reduced recurrent stroke by 28% (and by 43% with combination therapy), with benefit in both hypertensive and "normotensive" participants, establishing that the treatment effect tracks the magnitude of BP reduction rather than a threshold.

SPS3 (lacunar stroke, 3,020 patients) compared a systolic target of <130 vs 130–149 mm Hg: the primary reduction in all stroke did not reach significance (HR 0.81, 95% CI 0.64–1.03), but intracerebral haemorrhage fell significantly (HR 0.37), supporting a lower target particularly for small-vessel disease. SPRINT demonstrated that intensive control (<120 vs <140 mm Hg systolic) cut major cardiovascular events and all-cause mortality in high-risk non-diabetic patients — but it excluded patients with recent stroke, so its result is extrapolated rather than direct. ACCORD-BP in diabetics found no benefit of <120 vs <140 mm Hg for the primary composite, yet a prespecified secondary reduction in stroke (HR 0.59), reinforcing that stroke is the outcome most sensitive to BP lowering.

Targets

Therapy — agent choice

Most patients require two or more agents; a single-pill combination (e.g., ACE-I/ARB + CCB or + thiazide-like diuretic) improves adherence and control. The specific combination matters less than achieving and sustaining the target.

Monitoring & pearls

• Use out-of-office measurement (home or ambulatory) to exclude white-coat effect and detect masked/nocturnal hypertension, which carries excess stroke risk.
• Delay aggressive lowering in the hyperacute phase; permissive hypertension is standard immediately after ischemic stroke, with gradual reduction once stable.
• In symptomatic haemodynamically significant carotid/intracranial stenosis, watch for orthostatic or watershed symptoms with rapid titration.
• Screen for and treat obstructive sleep apnoea, primary aldosteronism (in resistant hypertension), and reinforce sodium restriction (<2 g/day) and potassium-rich diet.

Key references: PROGRESS Collaborative Group (Lancet 2001); SPS3 (Benavente et al., Lancet 2013); SPRINT (NEJM 2015); ACCORD-BP (NEJM 2010); AHA/ASA 2021 Secondary Prevention Guideline; ESO/ESH hypertension guidance.

Rationale & evidence

Diabetes roughly doubles ischemic stroke risk, accelerates large- and small-vessel atherosclerosis, worsens outcomes, and increases recurrence. Yet intensive glycaemic control per se has a modest and inconsistent effect on macrovascular stroke: ACCORD, ADVANCE and VADT showed that driving HbA1c toward normal does not reduce (and in ACCORD increased mortality) — so glucose targets should be individualised, not maximally aggressive. The paradigm has shifted from "glucose-centric" to cardiovascular-outcome-driven agent selection, favouring drug classes with proven event reduction.

For stroke specifically, GLP-1 receptor agonists provide the most direct evidence: SUSTAIN-6 (semaglutide) and REWIND (dulaglutide) reduced non-fatal stroke, and a class meta-analysis shows roughly a 16% relative reduction in stroke. SGLT2 inhibitors (EMPA-REG OUTCOME, CANVAS, DECLARE-TIMI 58) powerfully reduce heart-failure hospitalisation, renal decline and cardiovascular death but have a largely neutral effect on stroke. Pioglitazone occupies a special place in cerebrovascular neurology: the IRIS trial randomised 3,876 non-diabetic insulin-resistant patients with recent ischemic stroke/TIA to pioglitazone 45 mg or placebo and reduced the composite of stroke or MI from 11.8% to 9.0% (HR 0.76, 95% CI 0.62–0.93) over ~4.8 years, and lowered progression to diabetes — at the cost of weight gain, oedema and a higher fracture rate.

Targets

Therapy — preferred agents after stroke

Clinical pearls

• After an ischemic stroke, screen for undiagnosed diabetes/prediabetes (fasting glucose, HbA1c, or OGTT); insulin resistance is highly prevalent even in normoglycaemic stroke patients.
• Choose the glucose-lowering agent for its cardiovascular/renal profile, not merely its HbA1c-lowering potency — a GLP-1 RA is the rational add-on when stroke reduction is the priority.
• Pioglitazone is a consideration in the insulin-resistant, non-diabetic stroke patient without heart failure, weighing fracture and oedema risk; lower doses (15–30 mg) improve tolerability.
• Avoid hypoglycaemia — it triggers arrhythmia, sympathetic surge and may itself precipitate vascular events; relax targets accordingly in the vulnerable.

Key references: IRIS (Kernan et al., NEJM 2016); SUSTAIN-6 (NEJM 2016); REWIND (Lancet 2019); EMPA-REG OUTCOME (NEJM 2015); ACCORD/ADVANCE/VADT; AHA/ASA 2021 guideline; ADA Standards of Care 2024–2025.

Rationale & evidence

Cigarette smoking is an independent, dose-dependent cause of ischemic stroke (roughly doubling risk) and subarachnoid haemorrhage (a two- to fourfold increase), acting through endothelial injury, accelerated atherosclerosis, a prothrombotic and pro-inflammatory milieu, raised fibrinogen and haematocrit, and acute vasoconstriction. Risk rises with pack-years and is synergistic with oral contraceptives and hypertension. Secondhand (passive) smoke exposure raises stroke risk by roughly 20–30%. Crucially, the excess risk is substantially reversible: within 2–5 years of cessation, stroke risk falls toward — though it may not fully reach — that of never-smokers, making cessation among the highest-yield secondary-prevention interventions.

Therapy — pharmacologic cessation aids

Monitoring & behavioural support

• Combine pharmacotherapy with behavioural counselling (quitlines, structured programmes) — the two together roughly double quit rates over either alone.
• Use the "5 A's" (Ask, Advise, Assess, Assist, Arrange) at every encounter; a hospital admission for stroke is a powerful teachable moment with high quit success if support is arranged before discharge.
• Set a quit date, arrange follow-up within the first week, and treat relapse as expected rather than failure.

Clinical pearls

• E-cigarettes may aid cessation for some but are not risk-free and are not endorsed as first-line; dual use (vaping plus smoking) confers no benefit. Long-term cerebrovascular safety data are still maturing.
• Women who smoke and use combined oral contraceptives, especially over age 35, have markedly amplified ischemic stroke risk — a strong prompt for both cessation and contraceptive review.
• Smoking is the dominant modifiable risk factor for aneurysmal SAH and for aneurysm growth/rupture — cessation is central to unruptured-aneurysm management.

Key references: EAGLES trial (Anthenelli et al., Lancet 2016); US Public Health Service Clinical Practice Guideline (Treating Tobacco Use and Dependence); AHA/ASA 2021 Secondary Prevention Guideline; INTERSTROKE (Lancet 2016).

Rationale & evidence

LDL cholesterol is causal in atherosclerotic ischemic stroke, and the benefit of lowering it is proportional to the absolute LDL reduction achieved ("lower is better", and cumulative exposure matters). The landmark SPARCL trial randomised 4,731 patients with recent stroke/TIA and no known coronary disease to atorvastatin 80 mg or placebo: fatal/non-fatal stroke fell from 13.1% to 11.2% over ~5 years (HR 0.84; absolute risk reduction ~2.2%), with a larger reduction in major coronary events. SPARCL also revealed a small excess of haemorrhagic stroke (HR ~1.66; 55 vs 33 events) — a signal that has tempered but not overturned statin use, since the net cerebrovascular and cardiovascular benefit remains clearly favourable.

Treat Stroke to Target (2,860 patients with ischemic stroke/TIA and documented atherosclerosis) directly tested LDL goals: a target <70 mg/dL versus 90–110 mg/dL. Achieved LDL was 65 vs 96 mg/dL, and the primary composite of major cardiovascular events fell from 10.9% to 8.5% (HR 0.78, 95% CI 0.61–0.98) with no significant increase in intracranial haemorrhage (1.3% vs 0.9%). IMPROVE-IT extended the principle beyond statins: adding ezetimibe to simvastatin lowered LDL further and reduced events, including ischemic stroke, confirming that the mechanism of LDL lowering matters less than the level attained.

Targets

Therapy — stepwise intensification

Monitoring & pearls

• Check a lipid panel 4–12 weeks after initiation or dose change to confirm response and adherence; there is no need to withhold statins pending fasting samples in the acute setting.
• Start (or continue) high-intensity statin in hospital — early initiation is associated with better outcomes and abrupt withdrawal worsens them.
• The small haemorrhagic-stroke signal argues for caution (not prohibition) after a lobar/amyloid-pattern ICH; for ischemic stroke and TIA the benefit is unequivocal.
• Non-HDL cholesterol and apoB are useful secondary targets in hypertriglyceridaemia and diabetes where LDL underestimates atherogenic burden.

Key references: SPARCL (Amarenco et al., NEJM 2006); Treat Stroke to Target (Amarenco et al., NEJM 2020); IMPROVE-IT (Cannon et al., NEJM 2015); CTT Collaboration meta-analyses; ESC/EAS 2019 Dyslipidaemia Guidelines; AHA/ASA 2021 Guideline.

Rationale & evidence

Statin-associated muscle symptoms (SAMS) are the leading reason for discontinuation, yet rigorous data show that most reported intolerance is not pharmacologically caused by the statin. In the blinded n-of-1 SAMSON trial, 90% of the symptom burden during statin months was reproduced during placebo months (the nocebo effect), and StatinWISE reached a similar conclusion. True, dose-dependent myopathy — and the rare but serious rhabdomyolysis or immune-mediated necrotising myopathy (anti-HMGCR antibody) — is uncommon. Because untreated intolerance leaves high-risk stroke patients undertreated, the goal is to keep the patient on some effective LDL-lowering therapy rather than abandon it.

Drug interactions (mechanism-based)

Therapy — stepwise approach to reported intolerance

• Confirm and characterise: document symptom pattern, timing, and check CK; symmetrical proximal weakness with very high CK suggests true myopathy, whereas fleeting, migratory aches favour nocebo.
• Exclude reversible contributors: hypothyroidism, vitamin D deficiency, strenuous exercise, and interacting drugs.
• Rechallenge: after a washout, re-trial the same or a different statin — most patients tolerate a second attempt.
• Switch and lower-dose/intermittent: hydrophilic agents (rosuvastatin, pravastatin) or intermittent (alternate-day or twice-weekly) rosuvastatin/atorvastatin, which still yields meaningful LDL reduction.
• Add or substitute non-statins: ezetimibe, bempedoic acid (notably lower muscle complaints — muscle events were similar to placebo in CLEAR trials), PCSK9 inhibitors, or inclisiran to reach LDL goal.

Clinical pearls

• Routine CK monitoring in asymptomatic patients is unnecessary; measure CK when muscle symptoms arise.
• Persistent weakness with markedly elevated CK that does not resolve on withdrawal should prompt testing for anti-HMGCR immune-mediated necrotising myopathy, which requires immunosuppression, not just statin cessation.
• Bempedoic acid is a prodrug activated in the liver (not skeletal muscle), which underlies its favourable muscle-symptom profile — a useful option in genuine SAMS.
• Frame the nocebo phenomenon empathetically; blinded rechallenge or symptom diaries can help patients regain confidence in therapy.

Key references: SAMSON (Wood et al., NEJM 2020); StatinWISE (BMJ 2021); Cheeley et al. NLA statin intolerance statement (2022); SEARCH SLCO1B1 (NEJM 2008); CLEAR Outcomes (Nissen et al., NEJM 2023).

Rationale & evidence

When high-intensity statin plus ezetimibe fails to reach target, three non-statin pathways extend LDL lowering. Monoclonal PCSK9 inhibitors (evolocumab, alirocumab) lower LDL by 50–60% on top of statin. FOURIER (evolocumab, >27,000 patients) reduced the primary composite (HR 0.85) driving LDL to a median ~30 mg/dL, with a significant reduction in ischemic stroke and no excess of haemorrhagic stroke or neurocognitive harm (EBBINGHAUS). ODYSSEY OUTCOMES (alirocumab, post-ACS) reduced MACE and all-cause mortality and likewise lowered ischemic stroke. A prespecified FOURIER analysis in patients with prior stroke confirmed benefit without a haemorrhagic-stroke signal.

Inclisiran is a small interfering RNA that silences hepatic PCSK9 synthesis, lowering LDL by ~50% with twice-yearly dosing after loading (ORION-9/-10/-11). Its cardiovascular-outcomes trials (ORION-4, VICTORION-2 PREVENT) are maturing, with results anticipated around 2026–2027; until then it is used for LDL lowering where adherence or tolerability favours an infrequent injectable. Bempedoic acid, an ATP-citrate lyase inhibitor and liver-specific prodrug, was tested in CLEAR Outcomes (13,970 statin-intolerant patients, 180 mg daily): LDL fell ~21% (~29 mg/dL) and MACE dropped 13% (HR 0.87, 95% CI 0.79–0.96), driven by MI and coronary revascularisation. The stroke-specific reduction was directionally favourable but not independently powered.

Therapy — agents, dosing, monitoring

Monitoring & pearls

• PCSK9 inhibitors and inclisiran require no routine lab monitoring; confirm LDL response and adherence to the injection schedule.
• Inclisiran's biannual, clinician-administered dosing is attractive for patients with poor daily-pill adherence — but reserve firm cardiovascular-outcome claims until ORION-4/VICTORION-2 report.
• Bempedoic acid raises uric acid and can precipitate gout; check baseline and counsel patients with gout history. Its liver-specific activation spares muscle, making it valuable in true statin intolerance.
• These agents are additive: the combination of statin + ezetimibe + PCSK9 inhibitor can bring LDL below 30 mg/dL safely, and very low LDL has not been associated with cognitive harm in randomised data.

Key references: FOURIER (Sabatine et al., NEJM 2017) & EBBINGHAUS; ODYSSEY OUTCOMES (Schwartz et al., NEJM 2018); ORION-9/-10/-11 (Ray et al., NEJM 2020); CLEAR Outcomes (Nissen et al., NEJM 2023); ORION-4/VICTORION-2 PREVENT (pending).

Rationale & evidence

For non-cardioembolic ischemic stroke and TIA, antiplatelet therapy is the antithrombotic foundation. Long-term single antiplatelet therapy is standard; aspirin, clopidogrel, and aspirin–dipyridamole are all acceptable monotherapies (PRoFESS showed aspirin–dipyridamole and clopidogrel to be equivalent). The pivotal modern refinement is short-course dual antiplatelet therapy (DAPT) immediately after minor stroke or high-risk TIA, where the early recurrence risk is highest.

CHANCE (5,170 Chinese patients, minor stroke/high-risk TIA) and POINT (4,881 patients, largely non-Chinese) both showed that aspirin + clopidogrel begun within 12–24 hours reduced 90-day recurrent stroke versus aspirin alone. The benefit accrues almost entirely in the first ~21 days, while major bleeding accumulates with longer exposure — hence guidelines recommend 21 days of DAPT (CHANCE-derived) and no more than ~21–30 days, then monotherapy. THALES (11,016 patients) tested ticagrelor + aspirin for 30 days versus aspirin: it reduced the stroke/death composite (5.5% vs 6.6%) but increased severe/intracranial bleeding, so ticagrelor–aspirin is an alternative rather than default. Ticagrelor monotherapy was not superior to aspirin in SOCRATES.

Therapy — indications, agents, duration

Pharmacogenomics — CYP2C19

Clopidogrel is a prodrug requiring CYP2C19 activation; loss-of-function alleles (*2, *3) impair conversion, blunting platelet inhibition. CHANCE-2 (6,412 CYP2C19 LOF carriers) showed ticagrelor + aspirin superior to clopidogrel + aspirin for 90-day stroke recurrence (6.0% vs 7.6%, HR 0.77) without a significant increase in severe bleeding — supporting genotype-guided selection where testing is available. In LOF carriers, alternatives include ticagrelor-based DAPT or higher clopidogrel dosing.

Monitoring & pearls

• Load promptly — the first 24 hours carry the steepest recurrence risk, and DAPT should begin the day of presentation once haemorrhage is excluded.
• Do not extend DAPT indefinitely: MATCH, CHARISMA and SPS3 showed net harm (bleeding without ischemic benefit) with long-term aspirin + clopidogrel.
• Add a proton-pump inhibitor for GI protection during DAPT in higher-bleeding-risk patients (omeprazole/esomeprazole may modestly reduce clopidogrel activation — pantoprazole is a pragmatic choice).
• Ticagrelor requires twice-daily dosing and causes dyspnoea in ~10–15%; counsel patients and avoid with strong CYP3A inhibitors.

Key references: CHANCE (Wang et al., NEJM 2013); POINT (Johnston et al., NEJM 2018); THALES (Johnston et al., NEJM 2020); CHANCE-2 (Wang et al., NEJM 2021); SAMMPRIS (Chimowitz et al., NEJM 2011); PRoFESS (NEJM 2008); AHA/ASA 2021 Guideline.

Rationale & evidence

"Antiplatelet resistance" describes persistent platelet reactivity despite therapy, but the term conflates several distinct phenomena: true pharmacodynamic non-response, laboratory high on-treatment platelet reactivity (HTPR), and clinical treatment failure (a recurrent event on therapy, which is usually multifactorial rather than a pure drug problem). Assays disagree with one another, and the link between a "resistant" laboratory result and recurrent stroke is inconsistent, which is why routine testing is not recommended for most patients.

Mechanisms

Role of testing

• Platelet-function assays (VerifyNow, light-transmission aggregometry, VASP phosphorylation, PFA-100/200) and CYP2C19 genotyping can identify HTPR/LOF carriers, but randomised evidence that acting on these results improves cerebrovascular outcomes is limited and largely extrapolated from PCI populations (where routine tailoring — GRAVITAS, ARCTIC, ANTARCTIC — did not improve outcomes).
• CYP2C19 genotyping has the strongest cerebrovascular support: CHANCE-2 prospectively randomised confirmed LOF carriers and showed ticagrelor-based DAPT superiority, validating genotype-guided selection more than function-guided titration.
• Testing may be reasonable in the recurrent-event-on-clopidogrel patient to guide a switch, but it should not delay or replace attention to adherence and conventional risk factors.

Clinical pearls

• Before invoking "resistance", confirm adherence and correct dosing — non-compliance is the commonest explanation for apparent aspirin/clopidogrel failure.
• A pragmatic response to clopidogrel treatment failure is to switch drug class (to ticagrelor or aspirin-based therapy) rather than to chase a laboratory target.
• Avoid concurrent ibuprofen with aspirin; if an NSAID is needed, timing separation or an alternative agent reduces the pharmacodynamic interaction.
• Routine platelet-function or genotype testing is not endorsed for unselected secondary prevention; reserve for selected, refractory, or research contexts.

Key references: CHANCE-2 (Wang et al., NEJM 2021); GRAVITAS (JAMA 2011); ARCTIC (NEJM 2012); ANTARCTIC (Lancet 2016); CPIC CYP2C19–clopidogrel guideline; AHA/ASA 2021 Guideline.

Rationale & evidence

Atrial fibrillation increases ischemic stroke risk roughly fivefold, and AF-related strokes are larger, more disabling and more fatal than average. Oral anticoagulation reduces stroke by about two-thirds — far exceeding antiplatelet therapy, which is inadequate and largely abandoned for AF stroke prevention. Direct oral anticoagulants (DOACs) have supplanted warfarin for non-valvular AF: the four pivotal trials (RE-LY dabigatran, ROCKET-AF rivaroxaban, ARISTOTLE apixaban, ENGAGE-AF-TIMI 48 edoxaban), and their meta-analysis, show DOACs are at least as effective as warfarin for stroke/systemic embolism, with a ~50% reduction in intracranial haemorrhage and lower mortality. Warfarin remains mandatory for mechanical heart valves and moderate-to-severe (rheumatic) mitral stenosis, where DOACs are contraindicated (RE-ALIGN halted for harm with dabigatran; INVICTUS confirmed VKA superiority in rheumatic AF).

Risk stratification

Any patient with prior stroke/TIA already scores ≥2 and warrants anticoagulation absent a strong contraindication. HAS-BLED (Hypertension, Abnormal renal/liver function, Stroke, Bleeding history, Labile INR, Elderly, Drugs/alcohol) estimates bleeding risk and flags modifiable factors (BP control, alcohol, concomitant antiplatelets/NSAIDs) — a high score should prompt risk-factor mitigation, not withholding of anticoagulation.

Therapy — DOAC dosing

Left atrial appendage occlusion (cross-reference)

For patients with genuine long-term contraindications to anticoagulation (e.g., recurrent major or intracranial bleeding), percutaneous LAA occlusion (WATCHMAN — PROTECT-AF, PREVAIL; Amulet) offers an alternative, with PRAGUE-17 showing non-inferiority to DOACs in high-risk patients. See the dedicated cardioembolism/LAAO topic for device selection, peri-procedural antithrombotics and DRT (device-related thrombus) surveillance.

Monitoring & pearls

• Do not under-dose reflexively — inappropriate DOAC dose reduction is common and associated with more strokes without less bleeding; apply the label criteria exactly.
• Apixaban and edoxaban have the most favourable GI-bleeding profiles; rivaroxaban and higher-dose dabigatran carry more GI bleeding — individualise by bleeding phenotype and renal function.
• Reversal agents exist: idarucizumab for dabigatran, andexanet alfa for factor Xa inhibitors (PCC as alternative) — relevant when anticoagulated patients present with ICH or need urgent surgery/thrombolysis decisions.
• Reassess renal function at least annually (more often if CrCl <60 or unwell) to keep dosing appropriate.
• Combining a DOAC with antiplatelet therapy should be time-limited (e.g., after PCI, guided by AUGUSTUS/PIONEER-AF), as the combination markedly raises bleeding.

Key references: RE-LY, ROCKET-AF, ARISTOTLE, ENGAGE-AF-TIMI 48; Ruff et al. DOAC meta-analysis (Lancet 2014); INVICTUS (NEJM 2022); PRAGUE-17 (JACC 2022); AHA/ACC/ACCP/HRS 2023 AF Guideline; ESC 2024 AF Guideline; AHA/ASA 2021 Guideline.

Rationale & evidence

Clinicians have long balanced two competing risks after an AF-related ischemic stroke: early recurrent embolism (favouring prompt anticoagulation) versus haemorrhagic transformation of the infarct (favouring delay). The traditional, expert-consensus "1-3-6-12 day rule" (EHRA) staged anticoagulation start by clinical severity — day 1 for TIA, day 3 for mild, day 6 for moderate, and day 12 for large/severe strokes. This heuristic was never validated in randomised trials and, in the DOAC era, has been overtaken by direct evidence favouring earlier initiation.

Four randomised trials now inform practice. TIMING (registry-based, 888 patients) established non-inferiority of early (≤4 days) versus delayed (5–10 days) DOAC start. ELAN (2,032 patients) defined "early" as day 1 (minor/moderate) or day 6 (major) versus later (day 3–4 / 6–7 / 12–14); the 30-day composite was 2.9% early vs 4.1% late, with recurrent ischemic stroke 1.4% vs 2.5% and no increase in symptomatic ICH. OPTIMAS (3,621 analysed), the largest trial, randomised early (≤4 days) versus standard (7–14 days): the 90-day composite was identical at 3.3% in each arm, meeting non-inferiority, with symptomatic ICH 0.6% vs 0.7%. START (response-adaptive, mild/moderate arms days 3–14 and severe arms days 6–21) added further data supporting earlier initiation without an ICH penalty.

The CATALYST individual-participant-data meta-analysis (pooling ELAN, OPTIMAS, TIMING and START, ~5,400 patients) found early anticoagulation (≤4 days) reduced recurrent ischemic stroke with no increase in symptomatic intracranial haemorrhage, supporting early initiation across the severity spectrum studied. Importantly, patients with very large infarcts, haemorrhagic transformation, or those requiring hemicraniectomy were under-represented, so individualised caution persists at the extreme end.

Targets — a pragmatic contemporary approach

Clinical pearls

• The randomised evidence has effectively retired the rigid 1-3-6-12 rule for most patients; early initiation (within ~4 days) is safe and at least as effective for minor-to-moderate strokes.
• Obtain baseline imaging (CT or MRI, including GRE/SWI) to gauge infarct size and detect haemorrhagic transformation before starting; large infarcts remain the group where delay is prudent.
• DOACs — not bridging heparin — are the anticoagulants of choice; parenteral bridging adds bleeding without benefit in this setting.
• Do not defer indefinitely: prolonged delay leaves patients exposed to a high early embolic recurrence risk that outweighs the modest ICH concern in most infarcts.
• Decisions in very large infarcts, mechanical thrombectomy with large cores, or planned surgery should be individualised and imaging-guided.

Key references: ELAN (Fischer et al., NEJM 2023); OPTIMAS (Werring et al., Lancet 2024); TIMING (Oldgren et al., Circulation 2022); START (ISC 2024); CATALYST IPD meta-analysis (Lancet 2025); WSO scientific statement on AF stroke prevention (2025); EHRA practical guide.

Rationale & evidence

Factor XI inhibition aims to uncouple thrombosis from haemostasis. The rationale rests on human genetics and epidemiology: individuals with congenital factor XI deficiency (haemophilia C) are protected from ischemic stroke and venous thromboembolism yet bleed little spontaneously, while elevated factor XI predicts thrombotic events. Because factor XI amplifies thrombin generation through the intrinsic (contact) pathway but is largely dispensable for the initial haemostatic plug, blocking factor XIa promised anticoagulation with less bleeding — a potentially transformative profile for stroke prevention, especially in patients deemed too high-bleeding-risk for DOACs. Agents in development include small molecules (asundexian, milvexian), a monoclonal antibody (abelacimab), and antisense oligonucleotides (fesomersen).

Therapy — the current (evolving) evidence landscape

Monitoring & interpretation

• The OCEANIC-AF failure demonstrated that factor XIa inhibition at the tested asundexian dose does not match the anticoagulant efficacy of apixaban in AF — a reminder that reduced bleeding is meaningless if thromboembolic protection is inadequate.
• The continuing LIBREXIA-AF/STROKE and abelacimab programmes will determine whether a different agent, dose, or higher-risk population can deliver the promised "efficacy with less bleeding" — the AF and secondary-stroke questions remain genuinely open.
• A plausible near-term niche, if efficacy is confirmed, is the anticoagulation-intolerant or high-bleeding-risk patient (e.g., prior ICH, ESUS with atrial cardiopathy) rather than displacement of DOACs in standard AF.

Clinical pearls

• These agents are not yet approved for stroke prevention; do not use outside trials. The field is evolving rapidly and any specific efficacy claim should be regarded as provisional pending 2026 read-outs.
• The class's theoretical appeal (haemostasis-sparing anticoagulation) remains scientifically strong even after OCEANIC-AF — the setback appears drug/dose-specific rather than a refutation of factor XI biology.

Key references: OCEANIC-AF (Piccini et al., NEJM 2024); AZALEA-TIMI 71 (Ruff et al.); PACIFIC-AF phase 2 (Piccini et al., Lancet 2022); LIBREXIA programme rationale (Am Heart J 2024) and BMS/J&J LIBREXIA-ACS discontinuation announcement (Nov 2025).

Rationale & evidence

Residual inflammatory risk persists despite optimal lipid and antithrombotic therapy, and the inflammatory hypothesis of atherosclerosis was validated by CANTOS, in which the IL-1β antibody canakinumab reduced cardiovascular events independent of lipid lowering (but was limited by cost and fatal infection). Colchicine — an inexpensive oral anti-inflammatory that inhibits microtubule assembly, the NLRP3 inflammasome and neutrophil function — became the pragmatic candidate. In coronary disease, LoDoCo2 (low-dose colchicine 0.5 mg daily, 5,522 patients with chronic coronary disease) reduced major cardiovascular events by ~31% (HR 0.69), and COLCOT showed benefit post-MI, leading to cardiology guideline endorsement and FDA approval of colchicine for atherosclerotic cardiovascular disease.

The stroke-specific evidence is more nuanced. CONVINCE (3,144 patients with recent non-cardioembolic stroke/TIA, colchicine 0.5 mg daily, median follow-up ~34 months) missed its primary endpoint in the intention-to-treat analysis: the composite of stroke, MI, cardiac arrest, unstable angina or vascular death was 9.8% vs 11.8% (HR 0.84, 95% CI 0.68–1.05). A prespecified on-treatment analysis was significant (HR 0.80, 95% CI 0.63–0.99), hinting at a real but modest effect diluted by discontinuation. CHANCE-3 (8,343 patients with acute minor stroke/high-risk TIA and elevated hsCRP, colchicine 0.5 mg BID then daily for 90 days) was clearly negative for 90-day recurrent stroke (6.3% vs 6.5%, HR 0.98) — suggesting that short-course colchicine started acutely does not prevent early recurrence. A meta-analysis pooling CONVINCE with the coronary trials nonetheless found a significant reduction in ischemic stroke (RR ~0.73), driven largely by CONVINCE's long-term data.

Therapy

Monitoring & pearls

• Gastrointestinal side effects (diarrhoea, nausea) are the main tolerability issue and drive discontinuation — the likely reason CONVINCE's ITT and on-treatment results diverged.
• Colchicine is renally and hepatically cleared and interacts with CYP3A4/P-gp inhibitors (clarithromycin, ciclosporin, some statins) — cumulative toxicity (myopathy, marrow suppression) can occur; avoid in significant renal/hepatic impairment.
• The current place of colchicine in secondary stroke prevention is adjunctive and selective (atherosclerotic patients with residual inflammatory risk) rather than routine; guidelines have not universally adopted it for stroke, and the acute-phase data are negative.
• hsCRP-guided patient selection is intuitively attractive but was not sufficient to yield benefit in the acute CHANCE-3 population; the durable benefit appears to require prolonged therapy in a stable atherosclerotic phenotype.

Key references: LoDoCo2 (Nidorf et al., NEJM 2020); COLCOT (Tardif et al., NEJM 2019); CONVINCE (Kelly et al., Lancet 2024); CHANCE-3 (Li et al., BMJ 2024); CANTOS (Ridker et al., NEJM 2017); colchicine stroke meta-analysis (eClinicalMedicine 2024).