Stroke is one of the most important acquired causes of epilepsy in adults and is a particularly common cause of newly diagnosed epilepsy in older people. Seizures can occur during the acute stroke period or emerge months and even years later after the injured brain has undergone lasting structural and network change.
These are not all the same clinical problem. A seizure occurring soon after a stroke is usually classified differently from a later unprovoked seizure, and the risk of recurrence, treatment decisions and long-term implications differ accordingly.
Post-stroke epilepsy therefore sits at the intersection of stroke neurology, epileptogenesis, vascular disease and long-term rehabilitation.
Stroke can cause both acute symptomatic seizures and later epilepsy
Modern definitions distinguish seizures occurring within seven days of stroke from those occurring later.
A seizure within the first seven days is generally considered an acute symptomatic seizure: a seizure provoked by an acute brain insult.
A seizure occurring more than seven days after the stroke is considered a remote symptomatic or unprovoked seizure. Because recurrence risk after a late post-stroke seizure is high, a single such seizure can be sufficient to meet the practical ILAE definition of epilepsy in some patients.
The seven-day boundary is clinically useful, not biologically absolute
The seven-day threshold is used internationally because early and late seizures have different recurrence risks and mechanisms, but biology does not change abruptly on day eight.
Early seizures are more closely linked to acute ionic disturbance, metabolic stress, excitotoxicity, oedema and blood products. Later seizures reflect longer-term epileptogenesis: gliosis, inflammation, blood–brain barrier dysfunction, synaptic reorganisation and persistent changes in cortical networks.
The distinction is therefore practical rather than a claim that two completely separate processes exist.
How common is post-stroke epilepsy?
Reported rates vary because studies use different definitions, stroke populations and follow-up periods.
A large systematic review and meta-analysis including 128 studies estimated early post-stroke seizures in about 7% of patients and post-stroke epilepsy in about 10%, although individual cohorts vary substantially.
A 2024 review reported that post-stroke epilepsy develops in around 6% of people over several years after stroke. The true risk depends heavily on stroke type, severity and cortical involvement.
Stroke is a major cause of new epilepsy in older adults
Among people aged 60 and over, stroke is one of the leading causes of new unprovoked seizures and epilepsy.
A recent review estimated that stroke accounts for roughly 30–50% of unprovoked seizures in older adults, ahead of several other acquired causes.
This means a first focal seizure in an older person often prompts careful investigation for prior or current cerebrovascular disease.
Ischaemic and haemorrhagic stroke can both lead to epilepsy
Post-stroke epilepsy can follow ischaemic infarction, intracerebral haemorrhage, haemorrhagic transformation of an infarct and other forms of vascular brain injury.
The risk is generally higher when blood or infarcted tissue involves the cortex.
Haemorrhagic stroke tends to carry a higher risk of early seizures than uncomplicated ischaemic stroke.
Cortical involvement is one of the strongest risk factors
Strokes involving the cerebral cortex are substantially more likely to lead to seizures than purely subcortical strokes.
The 2023 meta-analysis found cortical involvement associated with increased risk of both early seizures and later post-stroke epilepsy.
This is biologically plausible because cortical networks generate the abnormal synchronised activity that produces most focal epileptic seizures.
Stroke severity matters
Larger and more severe strokes tend to carry higher seizure and epilepsy risk.
Greater injury means more damaged cortex, inflammation, gliosis and network reorganisation — all factors that can contribute to epileptogenesis.
Stroke severity measured with tools such as the NIH Stroke Scale is consistently associated with later seizure risk.
Haemorrhagic transformation increases risk
An ischaemic stroke can undergo secondary bleeding into the infarcted area. This haemorrhagic transformation is associated with a higher risk of early seizures and post-stroke epilepsy.
Blood products can irritate cortex and contribute to inflammatory and excitotoxic processes.
The increased risk does not mean every person with haemorrhagic transformation will develop epilepsy.
Early seizures predict later epilepsy
Having an acute symptomatic seizure after stroke is itself a strong predictor of later unprovoked seizures.
The large meta-analysis found a markedly increased odds of later post-stroke epilepsy among people who had an early post-stroke seizure.
Even so, an early seizure alone does not automatically mean epilepsy has developed.
Anterior circulation and large cortical infarcts carry greater risk
Anterior circulation strokes frequently involve cortical territories supplied by the middle or anterior cerebral arteries.
Large cortical infarcts are therefore more epileptogenic than small deep lacunar strokes.
This is one reason stroke location matters as much as the fact that a stroke occurred.
Lacunar strokes are less strongly associated with epilepsy
Small-vessel lacunar infarcts usually affect deep subcortical structures rather than cortex.
They can coexist with seizures for other reasons, but a small isolated lacunar infarct is generally less likely to be directly epileptogenic than a cortical infarct.
Clinicians therefore avoid assuming that every vascular lesion seen on MRI explains a seizure.
The seizure may reveal an old stroke
Some people do not know they previously had a stroke until imaging is performed after a first seizure.
Small or clinically subtle cortical infarcts can leave permanent structural injury without having caused an obvious acute neurological deficit.
The later seizure may therefore be the first event that leads to discovery of the old lesion.
Why a stroke can become epileptogenic
A stroke destroys or injures neurons, disrupts normal inhibitory and excitatory circuits and triggers inflammatory responses.
During recovery, surviving networks reorganise. Gliosis develops, synapses remodel, ion channels and receptors change, and the blood–brain barrier can remain altered.
These changes can create a persistent network capable of generating spontaneous recurrent seizures.
Blood–brain barrier dysfunction is part of current research
Stroke can disrupt the blood–brain barrier, allowing proteins and inflammatory mediators to enter brain tissue in abnormal ways.
Experimental and clinical research increasingly links persistent barrier dysfunction with epileptogenesis.
This is one of several mechanisms being investigated as a possible target for future preventive treatment.
Inflammation may help drive epileptogenesis
After stroke, microglia and astrocytes become activated and release inflammatory mediators.
Cytokines such as interleukin-6 and pathways involving HMGB1 and other immune signals have been studied for their role in changing neuronal excitability.
These findings are important scientifically, but they have not yet produced a proven routine treatment that prevents post-stroke epilepsy.
There is currently no proven antiepileptogenic drug after stroke
No antiseizure medicine has been shown convincingly to prevent epilepsy simply by being given prophylactically after stroke.
The 2026 American Heart Association/American Stroke Association acute ischaemic stroke guideline states that evidence is insufficient to support routine prophylactic antiseizure medication to prevent post-stroke seizures or improve functional recovery.
The European Stroke Organisation has likewise historically recommended against routine primary antiseizure prophylaxis in most stroke patients. Its dedicated post-stroke seizure guideline is currently being updated.
Giving levetiracetam to everyone after stroke is not evidence-based
Levetiracetam is widely used when treatment is needed because it has relatively few drug interactions, but routine preventive use after every stroke is not supported by current evidence.
A randomised trial of levetiracetam prophylaxis was stopped early because of poor enrolment and did not establish a preventive benefit.
Treatment should therefore be driven by seizures, risk and individual clinical context rather than automatic prescribing.
When an early seizure occurs
An acute symptomatic seizure may require immediate antiseizure treatment, particularly if prolonged, recurrent or associated with status epilepticus.
The need for continued medication after the acute period is more individualised.
Some patients with a single acute symptomatic seizure may later stop antiseizure medication, while others with repeated seizures, EEG abnormalities or high-risk structural features may require longer treatment.
A late unprovoked seizure changes the picture
A seizure occurring after the acute stroke period carries a higher recurrence risk.
Because that risk can exceed the threshold used in the ILAE practical definition of epilepsy, treatment with long-term antiseizure medication is commonly considered after one remote post-stroke seizure.
The 2026 AHA/ASA guideline states that initiating antiseizure medication after an unprovoked post-stroke seizure is reasonable.
Medication choice should be individualised
There is no single antiseizure medicine proven best for all post-stroke epilepsy.
Selection depends on seizure type, age, cognition, mood, renal and liver function, cardiovascular comorbidity, anticoagulants, antiplatelet therapy and other medications.
Drug–drug interaction burden matters particularly in older stroke survivors who may already take multiple medicines.
Levetiracetam is commonly used, but not automatically ideal
Levetiracetam has few hepatic enzyme interactions and can therefore be convenient in people taking anticoagulants, statins and cardiovascular drugs.
However, it can cause irritability, mood change or behavioural adverse effects in some patients.
Its common use should not be mistaken for universal superiority.
Lamotrigine can be attractive in some older adults
Lamotrigine is often well tolerated cognitively and has relatively few interactions compared with older enzyme-inducing antiseizure medicines.
Its main limitation is the need for slow dose escalation because of the risk of serious rash.
It is therefore not suitable when very rapid therapeutic coverage is required.
Older enzyme-inducing medicines can complicate vascular care
Carbamazepine, phenytoin and phenobarbital can interact with many cardiovascular drugs through hepatic enzyme induction.
These interactions can affect anticoagulants, lipid-lowering treatment and other medicines commonly used after stroke.
They still have clinical roles, but prescribing needs careful review.
Anticoagulation makes interaction checking especially important
Many stroke survivors take anticoagulants for atrial fibrillation or other thromboembolic risks.
Some antiseizure medicines can alter anticoagulant concentrations, particularly certain direct oral anticoagulants and warfarin.
Medication choice should therefore consider both seizure control and stroke-prevention treatment.
EEG can help, but a normal EEG does not exclude post-stroke epilepsy
EEG can identify epileptiform discharges, electrographic seizures and focal slowing.
However, routine scalp EEG samples only a short period of brain activity and may be normal between seizures.
A normal EEG therefore cannot rule out epilepsy when the clinical history is convincing.
Continuous EEG can reveal non-convulsive seizures
In critically ill stroke patients, especially those with unexplained reduced consciousness, continuous EEG can detect seizures that produce little or no obvious motor activity.
The European Stroke Organisation has noted that a substantial proportion of electrographic seizures after stroke can occur without a clear clinical correlate.
Continuous monitoring is most relevant in selected high-risk hospital patients rather than every stroke survivor.
Non-convulsive status epilepticus can mimic stroke deterioration
A patient who becomes less responsive after stroke may have worsening cerebral oedema, recurrent stroke, metabolic disturbance, infection, medication effect or non-convulsive status epilepticus.
EEG can be critical when the neurological state is worse than expected from the structural injury alone.
Failure to consider non-convulsive seizures can delay treatment.
Not every twitch after stroke is epileptic
Post-stroke movement disorders, spasticity, tremor, myoclonus, syncope and functional neurological symptoms can resemble seizures.
Video recordings and EEG may help when events are recurrent and diagnosis is uncertain.
Correct classification matters because unnecessary long-term antiseizure treatment carries its own risks.
Seizure symptoms depend on the damaged network
Because post-stroke epilepsy is usually focal, symptoms vary according to the cortical region involved.
Motor cortex injury can produce focal clonic movements. Temporal or insular involvement may cause impaired awareness, autonomic symptoms or unusual sensory experiences. Occipital involvement can cause visual symptoms.
Focal seizures can spread bilaterally and become tonic–clonic.
Post-stroke seizures can be mistaken for recurrent stroke
After a focal seizure, temporary weakness can occur in the affected limbs — known as Todd’s paresis.
This can resemble a new stroke, especially in someone who already has vascular disease.
Because recurrent stroke is time-critical, new neurological deficits after a seizure still require urgent assessment rather than being assumed to be postictal.
Stroke can also mimic a seizure
Acute stroke can cause involuntary movements, altered awareness and unusual sensory symptoms.
Some transient ischaemic attacks also produce symptoms that are difficult to distinguish from focal seizures.
Diagnosis depends on timing, symptom evolution, vascular territory, imaging and sometimes EEG.
Children can develop post-stroke epilepsy too
Although stroke is most common in older adults, neonatal, childhood and perinatal strokes can also lead to epilepsy.
The 2023 meta-analysis found higher seizure and epilepsy rates after stroke in children than in adults.
Paediatric post-stroke epilepsy intersects with developmental neurology because seizures can affect learning and cognitive development in addition to the original vascular injury.
Perinatal stroke may present with epilepsy much later
Some children with remote perinatal infarction are not recognised as having had a stroke until they later develop seizures, motor asymmetry or developmental differences.
The vascular event may therefore predate the epilepsy diagnosis by years.
This is another example of how structural epilepsy can appear long after the original brain injury.
Cerebral venous thrombosis has a particularly strong seizure association
Cerebral venous sinus thrombosis can cause venous infarction, haemorrhage and cortical irritation.
Seizures are relatively common at presentation compared with many arterial ischaemic strokes.
Long-term epilepsy risk depends on lesion location, haemorrhage and the extent of cortical injury.
Intracerebral haemorrhage carries substantial seizure risk
Blood directly irritates and injures cortex, particularly when the haemorrhage reaches the cortical surface.
Early seizures after intracerebral haemorrhage are more common than after uncomplicated ischaemic infarction.
Remote epilepsy can later develop as the damaged region scars and reorganises.
Subarachnoid haemorrhage is a different vascular context
Aneurysmal subarachnoid haemorrhage can also be associated with seizures, cortical injury and later epilepsy.
Mechanisms can include direct haemorrhagic irritation, infarction from vasospasm, hydrocephalus, surgery and secondary brain injury.
Management therefore depends on the broader neurological context rather than applying a single post-stroke rule.
Prediction tools exist, but they do not decide treatment on their own
Scores such as SeLECT for ischaemic stroke and CAVE for intracerebral haemorrhage have been developed to estimate later seizure risk.
They combine clinical variables such as stroke severity, cortical involvement, early seizure and haemorrhage characteristics.
These tools can support counselling and research but do not replace clinical judgement or automatically justify prophylactic treatment.
The SeLECT score
SeLECT estimates late seizure risk after ischaemic stroke using stroke severity, large-artery atherosclerosis, early seizures, cortical involvement and middle cerebral artery territory.
Higher scores indicate higher predicted risk.
Risk prediction may eventually help select patients for antiepileptogenesis trials, but routine preventive treatment based on the score alone is not currently established.
The CAVE score
CAVE was developed for intracerebral haemorrhage and incorporates cortical involvement, younger age, haematoma volume and early seizure.
Like SeLECT, it provides probabilistic risk rather than certainty.
A low score does not make later epilepsy impossible, and a high score does not mean it will definitely occur.
Risk is highest in the first years but does not disappear
Many late post-stroke seizures occur within the first one to two years, but epilepsy can begin much later.
Persistent cortical scarring means risk can remain elevated for years.
A remote seizure should therefore not be dismissed because the stroke happened long ago.
Rehabilitation and epilepsy interact
Post-stroke recovery often depends on physiotherapy, occupational therapy, speech and language therapy and cognitive rehabilitation.
Seizures, postictal fatigue and antiseizure medication adverse effects can interfere with participation.
Treatment therefore aims not only to prevent further seizures but to protect rehabilitation progress and independence.
Cognition can be affected by both stroke and epilepsy
Stroke itself can cause memory, attention, language and executive difficulties depending on location.
Seizures, interictal epileptiform activity, sleep disruption and medication can add further cognitive burden.
When cognition worsens after treatment begins, clinicians need to distinguish medication effects from seizure activity and the underlying stroke.
Mood symptoms deserve specific attention
Depression and anxiety are common after stroke, and epilepsy can add further psychosocial strain.
Some antiseizure medicines can worsen irritability or mood in susceptible patients.
Mood history is therefore relevant when choosing treatment.
Falls and injury risk can increase
Stroke survivors may already have weakness, balance impairment or visual field loss.
A seizure or postictal confusion can therefore produce a higher risk of falls and injury than in someone without neurological disability.
Safety planning needs to consider both epilepsy and residual stroke deficits.
Driving restrictions follow epilepsy rules, not simply stroke recovery
A seizure after stroke can affect driving eligibility even when motor recovery from the stroke itself is good.
Rules vary by country and depend on seizure timing, recurrence and licence type.
Patients should receive jurisdiction-specific advice.
Secondary stroke prevention still matters after epilepsy develops
Controlling seizures does not reduce the need to prevent another stroke.
Blood-pressure control, anticoagulation where appropriate, antiplatelet therapy, lipid management, diabetes care, smoking cessation and other vascular measures remain essential.
A recurrent stroke can create new neurological disability and potentially add another epileptogenic lesion.
There is no evidence that antiseizure medication prevents recurrent stroke
Antiseizure medicines treat seizures. They do not replace antithrombotic or vascular risk-reduction therapy.
Medication plans therefore often involve two parallel goals: preventing seizures and preventing further cerebrovascular injury.
Surgery is uncommon but not impossible
Most post-stroke epilepsy is managed medically, especially in older adults and people with widespread vascular disease.
However, carefully selected people with focal drug-resistant epilepsy after a well-defined cortical stroke can sometimes undergo presurgical evaluation.
The feasibility depends on lesion location, network organisation, neurological function and overall health.
Neuromodulation may be considered in selected drug-resistant cases
When resection is unsafe or inappropriate, vagus nerve stimulation, deep brain stimulation or responsive neurostimulation may be options in some health systems.
Evidence is much less specific to post-stroke epilepsy than to drug-resistant epilepsy generally.
These are specialist decisions rather than routine post-stroke treatment.
Can post-stroke epilepsy go away?
Some people remain seizure-free for many years on medication and may eventually discuss withdrawal with a specialist.
However, the structural lesion remains, so recurrence risk can persist.
Any withdrawal decision depends on seizure history, EEG, lesion characteristics, medication adverse effects and individual risk tolerance.
Medication withdrawal is highly individualised
There is no universal time point after which antiseizure medicine should automatically be stopped in post-stroke epilepsy.
The 2026 AHA/ASA guideline emphasises individualising treatment duration after an unprovoked post-stroke seizure.
Factors include length of seizure freedom, seizure severity, recurrence risk and medication burden.
Why preventing the stroke is the strongest preventive strategy
There is currently no proven routine medicine that stops epileptogenesis after stroke.
The most reliable way to prevent post-stroke epilepsy is therefore to prevent stroke itself or reduce the amount of permanent cortical injury when stroke occurs.
Blood-pressure control, atrial fibrillation treatment, smoking cessation, diabetes management, vascular prevention and rapid acute stroke treatment all contribute indirectly.
Rapid reperfusion may reduce epileptogenic injury by limiting infarct size
Thrombolysis and mechanical thrombectomy can restore blood flow in eligible ischaemic stroke patients and reduce final infarct volume.
It is biologically plausible that less cortical destruction could reduce later seizure risk, although epilepsy prevention is not the primary reason these treatments are given.
The main goal remains saving threatened brain tissue and reducing disability.
Post-stroke epilepsy is not a sign that rehabilitation has failed
A person can recover well from stroke and still develop epilepsy later.
The seizure reflects how injured networks reorganised, not whether the person worked hard enough in rehabilitation.
This distinction is important because people may wrongly interpret late seizures as evidence that their stroke is actively worsening.
New seizures years later still deserve fresh assessment
An old stroke may explain a seizure, but clinicians should not automatically attribute every new event to the known lesion.
New stroke, tumour, infection, metabolic disturbance, medication toxicity and other causes may need consideration depending on the clinical context.
Structural history should guide diagnosis without closing it prematurely.
What current evidence says overall
Post-stroke epilepsy is common enough to be a major component of acquired epilepsy, especially in older adults. Risk is highest with cortical, severe and haemorrhagic lesions and in people who experience early seizures.
Current evidence supports treating documented seizures appropriately but does not support giving prophylactic antiseizure medication to every stroke patient in an attempt to prevent epilepsy.
Research is increasingly focused on identifying biological markers of epileptogenesis and high-risk patients who might one day benefit from targeted preventive treatment.
What this page does — and does not — explain
This page focuses on epilepsy caused by stroke. Acute symptomatic seizures more broadly are covered on the Acute Symptomatic and Provoked Seizures page. Stroke prevention and rehabilitation are not reproduced here in full. The mechanisms by which an injured brain becomes epileptic are explored separately on Epileptogenesis: How a Brain Can Become Epileptic and Why One Brain Injury Causes Epilepsy and Another Does Not.
Sources
Incidence and risk factors of post-stroke seizures and epilepsy: systematic review and meta-analysis
Post-stroke epilepsy: From clinical predictors to possible mechanisms
From Acute Symptomatic Seizures to Post-Stroke Epilepsy: diagnosis and management
AHA/ASA — 2026 Guideline for the Early Management of Acute Ischemic Stroke
European Stroke Organisation — Post-stroke seizures and epilepsy guideline
European Stroke Organisation — Guideline directory (post-stroke seizure update in development)
Risk factors for post-stroke epilepsy and seizures among stroke survivors
Evidence checked: September 2026.