Post-stroke seizures (PSS) are classified as early (within 7 days) or late (after 7 days). Early seizures, or acute symptomatic seizures (ASS), result from transient neurochemical changes post-stroke and are not typically epileptic. Late seizures, or unprovoked seizures, stem from lasting structural brain changes and signify post-stroke epilepsy (PSE). The 7-day cutoff is widely accepted and aligns with underlying pathophysiology (1). Early seizures occur in 3–6% of stroke patients, more commonly in hemorrhagic (10–16%) than ischemic strokes (2–4%) (2, 3). Stroke causes 73% of acute symptomatic seizures in adults. Late seizures affect 3–5% using the 7-day definition, with incidence up to 12%. According to the International League Against Epilepsy (ILAE), PSE can be diagnosed after a single unprovoked seizure, as it reflects an enduring brain change with a high recurrence risk (>60% over 10 years) (1). Redefining PSE to include single late seizures has raised incidence estimates to 8–12% (1).

Risk factors for PSS include cortical involvement, severe or large strokes, and early seizures (2, 4, 5). Hemorrhagic strokes carry a higher PSE risk (12.4%) than ischemic ones (6.4%). Additional predictors include ICH volume, younger age, hyponatremia, alcohol use, and premorbid disability (4). Stroke treatments, including decompressive craniectomy, craniotomy, intravenous alteplase, or endovascular treatment, are also considered risk factors (3). Routine scalp electrocephalograpm (EEG) has not reliably predicted PSE, but focal epileptiform discharges and lateralized periodic patterns may carry prognostic value (6). Prediction models like the SeLECT score exist but need further validation before widespread use (7).

Studies indicate that PSS is associated with worse functional outcomes and increased disability. Patients with PSS have significantly higher modified Rankin Scale scores and greater odds of poor outcome (3). PSE also contributes to long-term morbidity. There is growing evidence linking PSE with increased dementia risk. A 2.5-fold higher dementia incidence was reported in young stroke survivors with seizures, and pooled analyses confirm that PSS is independently associated with dementia (8). This suggests a feed-forward relationship among stroke, PSS, and neurodegeneration (8).

This minireview discusses antiseizure medication (ASM) therapies for managing PSS, including clinical considerations for initiating and discontinuing treatment.

Routine primary prophylaxis ASMs after stroke is generally not recommended due to the low incidence of PSS or PSE and the significant risk of adverse drug reactions (ADRs), especially in older adults with comorbidities (9, 10). Professional guidelines reflect this: the European Stroke Organization gives a weak recommendation against primary prophylaxis due to very low-quality evidence, and the American Heart Association/American Stroke Association similarly advises against routine ASM use, noting that potential harms outweigh benefits for most survivors (10).

For most patients, harms outweigh the benefits of preventing a first seizure. However, certain high-risk groups may warrant selective primary prophylaxis briefly. Even then, decisions must carefully weigh seizure risk against ASM tolerability and ADRs (11, 12). Tools such as the SeLECT score for ischemic stroke and the CAVE and 2HELPS2B scores for ICH help quantify seizure risk (4, 12). These models include factors like cortical involvement, NIHSS severity, early seizure, and MCA distribution infarcts. Despite this, primary prophylaxis is rarely recommended, even in high-risk patients, as efficacy evidence remains sparse and low quality (9, 11).

Evidence for primary ASM prophylaxis after hemorrhagic stroke is limited. Two randomized trials assessed this: one comparing valproate to placebo in 72 ICH patients showed no significant benefit (13), while the PEACH trial testing levetiracetam yielded mixed results—some reduction in electrographic seizures but no effect on clinical seizures (14). Both were underpowered, with the PEACH trial halted early due to poor recruitment. A Cochrane review incorporating these studies concluded ASMs do not effectively prevent post-stroke seizures, rating the evidence as low quality due to imprecision (9). No trials support long-term prophylaxis for late unprovoked seizures. Two small studies on short-term prophylaxis post-ICH were inconclusive (11, 12).

Observational studies provide important insights. A real-world study in older adults with acute ischemic stroke found higher 30-day mortality among those receiving seizure prophylaxis within 7 days, raising concerns about net benefit (15). Decision models favor secondary over primary prophylaxis. One model showed that starting ASMs after a seizure consistently yields better quality-adjusted life years (QALYs) than primary prophylaxis (11). Another model for ICH suggested short-term prophylaxis (e.g., 7 days) may benefit select high-risk patients, but long-term use generally leads to worse outcomes due to ADRs (12, 16).

In conclusion, current evidence does not support routine primary prophylaxis with ASMs, though select high-risk patients may be considered. When ASMs are indicated, lamotrigine, levetiracetam, lacosamide, and oxcarbazepine are preferred for their safety profiles (15, 17).

Secondary prophylaxis refers to the initiation of ASMs following a seizure in a stroke patient to prevent recurrence. Patients with early seizures carry a relatively low risk of immediate recurrence (10–20%) and a moderate risk of late recurrence (~30% over 10 years) (2, 3). Consequently, long-term ASM therapy is typically not recommended after a single early seizure. However, in cases of acute symptomatic status epilepticus, extended treatment is warranted (14). Short-term ASM use during the acute phase may also be considered to reduce excitotoxicity, with subsequent tapering.

In contrast, patients who experience late seizures are at a substantially higher risk of recurrence—exceeding 70% within 10 years. According to the ILAE, a single unprovoked seizure in this setting qualifies for a diagnosis of epilepsy (1). Discontinuing ASMs in these patients results in relapse rates over 50%, underscoring the need for long-term therapy (18). Therefore, secondary prophylaxis is generally recommended for all patients with post-stroke unprovoked seizures (19). Table 1 summarizes the commonly used ASMs for secondary prophylaxis, along with key clinical considerations to guide individualized treatment decisions based on current evidence.

Current evidence and expert consensus generally advise against routine long-term ASM initiation for primary prophylaxis in patients without post-stroke seizures (9–11, 17). Decision analyses consistently show that starting ASM only after a seizure—secondary prophylaxis—results in better outcomes, measured by quality-adjusted life years (QALYs), compared to primary prophylaxis (9, 11, 16). For example, one comprehensive decision analysis found that long-term primary prophylaxis yielded the lowest QALYs, supporting recommendations against prophylactic ASM use immediately after acute ischemic stroke (11, 16). Clinical trials have not demonstrated that short-term ASM use after stroke prevents epilepsy; ASMs function as antiseizure agents rather than antiepileptogenic therapies (9, 14).

When ASMs are initiated acutely—typically for early seizures within the first 7 days post-stroke—they should be prescribed for a limited duration. Clinical guidelines recommend early withdrawal of ASMs after the acute phase, generally within 1 to 2 weeks (9, 11, 17). In spontaneous ICH (sICH), decision analyses highlight the advantage of short-term (7-day) ASM treatment, urging clinicians to document discontinuation plans in prescriptions, discharge summaries, and patient education (11, 16). Although specific tapering protocols are not well established, consensus stresses early withdrawal due to low recurrence risk and no demonstrated benefit from prolonged primary prophylaxis (9, 11, 17).

Withdrawal of ASMs following secondary prophylaxis depends on whether treatment was initiated for an early or a late seizure consistent with PSE. For early seizures, guidelines recommend limiting ASM therapy to the acute phase—typically 1 to 2 weeks, or 7 days for sICH (9, 17). This is supported by the relatively low seizure recurrence risk (10–20%) and a moderate 10-year risk (~30%) of developing late unprovoked seizures (1, 3). Prolonged treatment after an early seizure does not improve outcomes and may increase adverse effects. Thus, routine ASM discontinuation is advised (9, 17).

In contrast, ASM withdrawal in patients with PSE is more complex and must be individualized. These patients carry a high risk of recurrence after withdrawal—over 50% in some studies—due to the symptomatic, lesional nature of their epilepsy (1, 20). The highest relapse risk occurs within the first 12 months post-withdrawal but may persist for years. Factors influencing withdrawal include age at onset and withdrawal, epilepsy etiology, seizure type, EEG findings, remission duration, and overall burden (11, 17, 22). Warning signs against withdrawal include focal seizures, short seizure-free intervals, abnormal neurological exams, and epileptiform EEG activity. In elderly patients, withdrawal is further complicated by seizure risks and drug interactions, even though relapse rates may be lower in late-onset epilepsy (17, 22). While rationale for withdrawal or continuation is clear, specific tapering protocols remain poorly defined (9, 11, 17).

While ASMs remain the cornerstone of PSS management, non-pharmacologic options also play a key role for patients with drug-resistant epilepsy (DRE). These include resective epilepsy surgery and neuromodulation therapies such as vagus nerve stimulation (VNS) and responsive neurostimulation (RNS) (11, 14, 17, 19).

Surgery is a viable option for select patients with DRE whose seizures arise from a well-localized epileptogenic zone. Surgical candidacy and evaluation follow principles similar to those in other focal epilepsies. Pre-surgical assessment typically involves multimodal imaging—high-resolution 3 T MRI, ictal SPECT, PET—and intracranial EEG monitoring to localize seizure onset (11, 17). In properly selected patients, resective surgery can significantly reduce seizure burden or achieve remission (17).

VNS has become an important adjunctive treatment for DRE. Beyond seizure control, VNS shows promise as a targeted plasticity intervention post-ischemic stroke by modulating brain nuclei linked to neural recovery. Animal studies show that combining VNS with rehabilitative training enhances motor recovery beyond rehabilitation alone (14, 19). Significantly VNS, paired with high-dose occupational therapy, has been shown to be effective in improving upper limb function among patients with ischemic stroke and received regulatory approval from FDA (23, 24). RNS, which delivers electrical stimulation in response to abnormal brain activity, is another option for refractory cases (14, 19). These interventions are generally reserved for patients who have failed adequate ASM trials and meet criteria for DRE, offering hope for improved seizure control and functional outcomes (11, 14, 17, 19).

Advancements in managing post-stroke seizures and epilepsy depend on several key priorities. High-quality randomized controlled trials are urgently needed to assess the efficacy, safety, and tolerability of ASMs for both primary and secondary seizure prevention following ischemic stroke (9, 11, 14, 17). These trials should be double-blind, placebo-controlled, adequately powered, and focused on clinically meaningful outcomes such as seizure freedom and ASM withdrawal rates (9, 11, 17).

Improved risk stratification through validated clinical prediction models, integrated with molecular, imaging, and electrophysiologic biomarkers, is essential (6, 22). Emerging artificial intelligence (AI) techniques offer promise in enhancing EEG and neuroimaging analyses, enabling more personalized assessments and targeted prophylactic strategies (6, 18).

Further research into the underlying pathophysiology—especially excitotoxicity and blood–brain barrier disruption—may identify novel therapeutic targets and biomarkers (6, 14, 18). Advanced imaging methods like GluCEST MRI and permeability imaging are central to these efforts (18). Additionally, exploring the anti-epileptogenic effects of existing drugs such as statins and novel agents like rapamycin could open new preventive avenues (6, 14).

Optimal treatment protocols—including timing, choice of ASM, dosing, duration, and withdrawal—remain incompletely defined and must be individualized based on factors such as comorbidities, medication tolerance, adherence, and risk of seizure recurrence (11, 14, 17). Figure 1 summarizes our proposed algorithm for decision-making regarding primary and secondary ASM prophylaxis in patients presenting with PSS.

Translating these advances into practice demands rigorously designed trials to validate experimental and observational findings. Future research must generate robust evidence, improve predictive accuracy, clarify mechanistic pathways, and personalize therapies to enhance outcomes for stroke survivors vulnerable to epilepsy (6, 11, 14, 17).

PSS and PSE are significant complications of cerebrovascular disease, contributing to morbidity, impaired recovery, cognitive decline, and increased mortality. Management remains limited by a lack of high-quality trials and validated predictive tools. Primary prophylaxis with ASMs is generally discouraged except in high-risk groups (e.g., hemorrhagic stroke, severe cortical injury, prior acute seizures), while secondary prophylaxis after unprovoked seizures is standard.

Newer ASMs like levetiracetam, lamotrigine, lacosamide, and eslicarbazepine are preferred for their safety and minimal interactions. Treatment should be individualized, especially in older patients or those with cardiovascular and cognitive comorbidities, where older EI-ASMs pose greater risks. ASM withdrawal is advised after early seizures, but decisions in established PSE remain complex due to limited biomarker guidance. Non-pharmacologic options, including surgery and neuromodulation, are valuable for drug-resistant cases and may support rehabilitation. Adjunctive therapies such as statins and certain antihypertensives show promise for seizure prevention and need further study.

Ultimately, improving outcomes in PSS and PSE requires well-designed trials, refined prediction models, and integration of emerging biomarkers. As understanding of vascular injury and epileptogenesis evolves, a more personalized, mechanism-driven approach to seizure prevention in stroke survivors is both necessary and achievable.