The relationship between cerebrovascular disease and epilepsy has long been appreciated since Hughlings Jackson first reported partial seizures in the setting of acute stroke.⁷ Indeed, cerebrovascular disease has been found to be the most common cause of secondary epilepsy. In a population-based study from Rochester, Minnesota, cerebrovascular disease accounted for 11% of cases.³⁷ This chapter will review the limited data on the epidemiology and treatment of poststroke seizures and epilepsy. Because the differentiation between cerebrovascular disorders and seizures is sometimes difficult, the chapter will also aim to make the clinical distinction between epileptic syndromes, transient ischemic attacks (TIAs), and other minor ischemic stroke syndromes. The nosology and natural history of TIAs, the more common TIA syndromes, and some less common clinical syndromes of TIAs and minor strokes that might be confused with paroxysmal epileptic syndromes will be discussed.

Seizures can be a complication of an acute stroke. Traditionally, poststroke seizures have been divided into early and late based on presumed differences in pathophysiology.⁴¹ Early seizures are thought to result from acute biochemical disturbances and may result in part from the damaging effects of the excitatory neurotransmitter glutamate in response to ischemia.^14,17,58 In contrast, late seizures are attributed to gliosis and cortical scarring with resulting selective neuronal loss and hyperexcitability of the surrounding tissue.^39,58 Some authors have argued that early seizures are not reliably related to the strokes themselves because of other concurrent metabolic problems.^38,57 The International League Against Epilepsy (ILAE) defines early poststroke seizures as those that occur before 1 week, but studies have used widely varying definitions, from under 24 hours up to 1 month after stroke.^46,49 Despite varying definitions of what constitutes an early poststroke seizure, several studies have found that the risk of epilepsy is significantly increased among those patients who have seizures within 2 weeks of the onset of a stroke.⁷¹ Late poststroke seizures have also been associated with an increased risk of epilepsy.^10,15,48,71

Data from five prospective studies have shown that the rate of epilepsy after ischemic stroke ranges from about 2% to 4%.^{10,15,48,57,71} However, the reported incidence of epilepsy varies between studies with the definition. Using the ILAE definition of two or more unprovoked seizures more than 1 week after stroke, a study from Rochester, Minnesota, found an overall rate of poststroke epilepsy of 3.3%, and 66% of those that developed an initial late seizure after ischemic stroke went on to develop epilepsy by 4.5 years.⁷¹ In a study from Norway, the overall incidence of poststroke epilepsy was 3.1%, while 11 of 28 subjects with poststroke seizures developed epilepsy (55%).⁵⁷

Limited data are available regarding risk factors for poststroke seizures and more is known about those that occur early than late. Involvement of the cortex is the most well-recognized predictor of early seizure occurrence, having been found to be an independent risk factor in several prospective studies.^{8,10,15,47,48,71} Stroke severity and infarct size have also been found to be independent risk factors.^10,48,57,65 Few studies have examined independent predictors of late poststroke seizures or epilepsy.¹⁷ Early seizures appear to be a risk factor for late seizures and possibly epilepsy, and a few studies have found that infarct size, cortical involvement, and recurrent stroke have been independent predictors.^10,48,71 However, in a large prospective study from Norway, stroke severity was an independent risk factor for epilepsy while cortical location was not.⁵⁷ Thus, further data are needed to clarify the relative importance of these modifying factors.

Treatment of poststroke seizures and epilepsy is complicated by a lack of clinical trial data to support the use of a particular antiepileptic drug on the one hand, and evidence that some commonly used drugs may be detrimental to stroke recovery on the other. For example, treatment with antihypertensive drugs that block the adrenergic system, such as prazosin and clonidine, or those that stimulate γ-aminobutyric acid (GABA) receptors, such as benzodiazepines, have been shown to impair recovery after brain injury in the rat, and the antiepileptic drugs phenytoin and phenobarbital have been implicated as well.³³ Less is known about the effect of these drugs on recovery in humans. One study examining the control group of an acute ischemic stroke treatment trial found that subjects administered any of a group of “detrimental drugs” including benzodiazepines, dopamine receptor antagonists, α-1 blockers, α-2 agonists, phenobarbital, or phenytoin had worse motor recovery and less independence in activities of daily living than those that did not receive any of these drugs.³³ Recovery after subarachnoid hemorrhage may also be impeded by such treatment. In a prospective case series, greater phenytoin exposure was associated with worse functional outcome at 14 days and worse cognitive outcome at 3 months.⁶⁰ However, treatment after experimental ischemia in rodents has shown that many antiepileptics may be neuroprotective if given very early.¹⁷ For example, phenytoin when given 30 minutes after experimental occlusion of the rat middle cerebral artery was neuroprotective, but not if given after 2 hours.²² Separately, diazepam given 30 and 90 minutes after induced forebrain ischemia in the gerbil resulted in protection of hippocampal neurons.⁶⁹ Neuroprotective properties have also been reported for lamotrigine,
topiramate, levetiracetam, and zonisamide.¹⁷ Thus, further data are needed to clarify the importance after stroke of the doses, timing, and length of treatment with antiepileptic drugs that may be helpful or harmful to patients.

The term stroke generally describes a group of vascular disorders diverse in etiology and includes ischemic brain infarction, intracerebral hemorrhage, and subarachnoid hemorrhage.⁴ The term transient ischemic attack has been limited to the description of a brief episode of neurologic dysfunction resulting from ischemia and was introduced in 1957 by C. Miller Fisher. TIA is formally defined as “an abrupt onset of focal loss of brain function lasting less than 24 hours that localizes to a portion of the brain supplied by one vascular system and for which no other cause can be found.”² The arbitrary 24-hour time limit was selected on the basis of prospective studies in the 1970s prior to the widespread use of brain imaging.³ It was thought that focal deficits lasting beyond 24 hours would be expected to result from a focus of ischemic infarction. This definition now requires revision, based upon recent brain imaging evidence of infarction in a substantial proportion of episodes classified by the 24-hour time limit as TIAs.^11,24 The term reversible ischemic neurologic deficit (RIND) was created to define those patients with neurologic dysfunction lasting longer than 24 hours but resolving completely within 1 to 3 weeks.⁴ Neurologic deficits in the minor ischemic stroke syndromes conversely persist beyond these arbitrary time periods. Ischemic stroke may be more appropriately viewed as a continuum that encompasses TIA, RIND, and minor ischemic stroke.¹⁸ As similar risk factors and vascular pathologies underlie these diagnostic subgroups, their continued separate distinction from each other is employed primarily for the purposes of differential diagnosis and prognostic risk strati- fication.²³

Differing methods and ambiguities of definition have resulted in a wide disparity in reported prevalence and incidence rates of TIAs.¹³ Estimated prevalence has varied from 1.1 to 77 per 1,000 persons, but three recent studies have dealt more effectively with the ascertainment bias inherent in capturing transient episodes, and data have shown incidence rates that are more consistent. Population-based data from Rochester, Minnesota, for the period 1985 to 1989 led to a calculated annual age- and sex-adjusted incidence rate of 68 per 100,000 population.¹³ The incidence rate was somewhat higher in the Greater Cincinnati/Northern Kentucky Stroke Study at 83 per 100,000, but 15% of the sample is of black race and may be at higher risk of cerebrovascular disease.⁴² Similar rates were seen in the United Kingdom. In the Oxfordshire study the overall age- and sex-adjusted incidence rate for TIA was 51 per 100,000.⁶⁶ In addition, annual incidence rates for TIA appear to be reasonably stable over time, taking into account lower ascertainment in earlier studies. The incidence rate did not change appreciably in Rochester between the periods 1960 and 1972 and 1985 and 1989, although the rate increased slightly in Oxfordshire, U.K., between 1981 and 2004 (relative incidence 1.27).^13,66 The incidence of TIA was strongly related to age in all three studies.

Transient ischemic attacks precede 11% to 50% of strokes, depending on the specific stroke subtype.³ They have been reported most frequently in association with large artery atherothrombotic disease and less commonly with small vessel disease.^30,59 Approximately 50% to 75% of patients who experience a stroke from extracranial carotid atheromatous disease have a prior TIA.⁵⁹ Perhaps the most important finding in recent studies has been the very high rate of stroke following TIA. Data from a large health maintenance organization in California found that roughly 10% of those presenting with TIA went on to have an ischemic stroke within 90 days and half of these occurred within the first 48 hours after the index event.⁴⁰ Similarly, data from the Greater Cincinnati/Northern Kentucky Stroke Study found a 17% ischemic stroke rate within 6 months after TIA, 65% of which occurred within the first month.⁴² The results from the North American Symptomatic Carotid Endarterectomy Trial (NASCET) helped clarify the outcomes for TIA patients.¹ Patients with TIAs and symptomatic extracranial carotid stenosis exceeding 70% were followed prospectively over a period of 18 months. During this time period, 24% of the medically treated group had a stroke or died, compared with a 7% rate in the endarterectomy-treated group. This striking difference proved that surgery was effective in reducing the risk of stroke after TIA and that TIAs should be regarded as a marker of significantly increased stroke risk. Certain subgroups of TIAs carry a higher stroke risk than do others. Data from NASCET showed that hemispheric TIAs with known high-grade ipsilateral carotid stenosis had a stroke risk exceeding 40% over 2 years.⁷⁵ Early stroke risk is probably greater in those patients with “crescendo” TIAs (multiple and frequent) and in the subgroup with ventricular thrombi.²

Clinical discrimination between TIAs and nonvascular causes of transient neurologic dysfunction is crucial as the occurrence of a TIA is a reliable warning signal of an impending stroke. An episode consistent with a TIA should be rapidly evaluated to determine the cause. Prompt clinical recognition of TIAs and timely institution of appropriate therapy can help prevent stroke. Furthermore, the possibility of occult coronary artery pathology should be considered in the workup of these patients as the occurrence of TIAs may indicate generalized atherosclerosis; ischemic heart disease is the leading cause of death in elderly TIA patients.^2,25

Less than one in ten TIAs are witnessed by a physician; therefore, accurate diagnosis usually depends on a careful interpretation of the patient’s history.²⁶ Health care professionals tend to formulate a preliminary diagnosis within the first few minutes of history taking, which works well when the symptoms follow classic textbook descriptions.⁶⁸ However, many patients cannot offer a clear, concise account of abrupt onset of focal neurologic dysfunction, as required for the diagnosis of TIA. Details are often forgotten or unappreciated regarding time, mode of onset, and subsequent course of symptomatology. Historical reliability becomes more questionable when patients consult a physician weeks or months following the event. Nondominant hemispheric TIAs are particularly susceptible to misclassification because the event may be ignored or misinterpreted by the patient.⁷⁸ The lack of uniform diagnostic criteria for TIA is reflected in significant interobserver disagreement. The Cooperative Group for the study of TIA found that 30% of patients hospitalized with a diagnosis of TIA had been misclassified.¹⁶
This finding has been confirmed in other series.⁴³ Questionnaires on TIA symptomatology have documented a high positive response rate, both in the elderly and in a group of young adults.^56,81 These surveys indicate that episodes of transient central nervous system dysfunction are common. Follow-up of those patients whose transient symptoms were considered too vague to represent TIA has revealed a stroke rate comparable to the TIA group, suggesting that current empiric criteria for TIA are too narrow.⁸² Standardized checklists using nonmedical terminology and computer-based diagnostic algorithms offer useful alternatives to enhance diagnostic consistency.^44,63 There is a need for improved diagnostic guidelines for TIAs with enhanced reliability, sensitivity, and specificity.

Traditionally, TIAs have been classified according to the vascular territory involved: Carotid or vertebrobasilar. This differentiation is important for management and prognosis.⁵⁵ Carotid territory TIAs occur more frequently and may result in the following: Weakness, paralysis, or clumsiness of one side of the body or face; numbness or paresthesias affecting one side of the face or body; loss of vision affecting one eye or, less frequently, one visual field; language disturbance (aphasia); and dysarthria.⁴ Analysis of a large series of carotid TIAs has proved helpful in determining the relative frequencies of various symptoms typical of TIAs (Table 1).³¹