© Springer Science+Business Media New York 2015
Mohamad Z. Koubeissi, Amer Alshekhlee and Prachi Mehndiratta (eds.)Seizures in Cerebrovascular Disorders10.1007/978-1-4939-2559-9_11. Epidemiology of Seizures and Epilepsy in Cerebrovascular Disease
(1)
Department of Neurology, Washington University School of Medicine, St. Louis, MO, USA
Keywords
SeizureEpilepsyStrokeEpidemiologyIncidencePrevalenceEpidemiology of Epilepsy
Epilepsy is a very common neurologic disease with significant associated morbidity and mortality. In order to review the occurrence of epilepsy in the population, it is first necessary to clarify some definitions. Most studies are performed using the International League Against Epilepsy (ILAE) guidelines for epidemiological studies [1] which define epilepsy as two or more unprovoked seizures. Acute symptomatic, febrile, and neonatal seizures are excluded. Multiple seizures that occur within a 24-h period are considered to represent a single seizure. Anyone who has had a seizure in the last 5 years is considered to have active epilepsy while someone with a diagnosis of epilepsy who has not had a seizure within the last 5 years, with or without treatment, is considered to have epilepsy in remission.
Incidence is defined as the number of new cases in a specific population in a given period of time. It is typically expressed as the number of cases per 100,000 people per year. Cumulative incidence is defined as the proportion of a population that develops the condition being studied within a certain time. It is typically expressed as the percentage of the population that is affected at any time up to a particular age (e.g., 3 % by age 75 years). Prevalence is defined as the number of patients with a disease at a given point in time. It is typically expressed as the number of cases per 1000 people. For epilepsy prevalence studies, people are considered to have active epilepsy if they have had a seizure and/or used medications to treat seizures in the past 1 or 5 years, depending on the study. As detailed below, the incidence and prevalence of epilepsy vary with age. Therefore, overall population measures of incidence and prevalence depend on the age distribution of the measured population. It is common to correct for the age distribution of the measured population to allow comparisons between studies and age-corrected values are used in this chapter.
The incidence of new-onset seizures is typically measured separately for acute symptomatic (provoked) seizures and unprovoked seizures. Estimates for the rates of provoked seizures range from 20 to 35 per 100,000 [2, 3]. Studies of unprovoked seizures typically combine both individual seizures and new-onset epilepsy (two or more seizures) resulting in incidence rates varying between 40 and 70 per 100,000 in North America and Europe [4, 5]. In studies that specifically assessed new-onset epilepsy, the rates were slightly lower. For example, in the Hauser et al. [3] classic study of the population of Rochester, MN, the rate of new-onset epilepsy was 44 per 100,000 while including single seizures increased the incidence to 61 per 100,000. The same study reported cumulative incidences of 1.2 % by age 24, 3 % by age 75, and 4.4 % by age 85 years.
The incidence of epilepsy varies with age. In general, there is a bimodal distribution with a high incidence in young children that decreases through adolescence to a minimum in young adults. Recently, it has become apparent that the incidence starts to increase again after age 50–60 years [5]. For example, in Iceland the incidence in infants was 130.2 per 100,000 in infants, 110.5 per 100,000 in adults older than 65 years, and 168.5 per 100,000 in adults aged 75–84 years [6]. In Finland, the overall incidence decreased from 71.6 to 52.9 per 100,000 between 1986 and 2002 [7]. Despite this decrease, the incidence in the elderly increased approximately 20 % [7]. It has been suggested that there are some differences between elderly men and women. Specifically, one study reported that the increase in incidence starts in men between ages 60 and 69 years while it starts later in women [2]. Recent data in the US Medicare recipients who were at least 65 years of age suggest some racial differences in the incidence of epilepsy in the elderly. White beneficiaries had incidence rates of 230 per 100,000 compared to an incidence of 410 per 100,000 in black beneficiaries, an incidence of 160 per 100,000 in Asian beneficiaries, and an incidence of 110 per 100,000 in Native American beneficiaries [8]. For all groups, incidence increased with advancing age.
Prevalence studies are typically performed with door-to-door surveys using a standardized and validated screening questionnaire. This is generally considered to be a straightforward and accurate method, although in some cases low prevalence rates have raised concerns about possible concealment. In the developed world, the prevalence of active epilepsy has been estimated to range from 4 to 7 per 1000 [5]. There are also data to suggest that prevalence increases with advanced age. For example, in Finland, the prevalence of epilepsy in those above 85 years was 9.4 per 1000 compared to the overall rate of 4.8 per 1000 [9]. The prevalence of epilepsy in adults older than 75 years increased from 1.9 per 1000 in 1940 to 14.8 per 1000 in 1980 [10]. Recent data in the US Medicare beneficiaries who were at least 65 years of age suggested racial differences in prevalence. White beneficiaries had a prevalence of 10.2 per 1000, while black beneficiaries had a prevalence of 18.7 per 1000, Asian beneficiaries had a prevalence of 5.5 per 1000, and Native American beneficiaries had a prevalence of 7.7 per 1000 [8].
Overall, the most common cause of epilepsy is idiopathic/cryptogenic, and no clear cause is identified in up to two thirds of patients [3, 6]. The most commonly identified cause in adults is cerebrovascular disease [5], a cause that is increasingly important in older adults. For example, it has been estimated that vascular causes explain epilepsy in 32–54 % of elderly patients [11, 12]. Similarly, in the Veterans Administration (VA) Cooperative trial in the elderly #428, 34.1 % of patients had a history of cerebral infarction, and an additional 14.9 % had a history of atherosclerosis [13, 14].
Epidemiology of Stroke
Similar to epilepsy, stroke is a common disease that carries significant societal burden both within the USA and globally. Within the USA, stroke is the fourth leading cause of death and a leading cause of serious, long-term disability, with up to 30 % of survivors being left permanently disabled [15–18]. Worldwide, stroke ranks even higher in mortality as the second leading cause of death while ranking third in disability [19, 20]. In the Framingham Heart Study, a longitudinal community-based cohort study assessing cardiac and cerebral vascular risk factors and events, among ischemic stroke survivors who were older than 65 years of age, several disabilities were observed at 6 months after stroke: 50 % had hemiparesis, 35 % had depressive symptoms, 26 % were institutionalized in a nursing home, 30 % were unable to walk without assistance, 26 % were dependent in activities of daily living, and 19 % had some degree of aphasia [21].
Recent estimates of stroke incidence in the USA suggest approximately 795,000 strokes occur annually [17]. Of these, 75 % are first-time cerebrovascular events with the remaining 25 % being recurrent events. The incidence estimates are derived from several large, stroke epidemiological studies including the Framingham Heart Study [22], the Atherosclerosis Risk in Communities Study [23], the Cardiovascular Health Study [24], and the Greater Cincinnati/Northern Kentucky Stroke Study [25]. Stroke incidence globally is estimated to affect 15 million, leading to death in one third and permanent disability in one third [19]. The US prevalence of stroke (estimate for 2008) was 2.8 % (affecting about seven million Americans older than 20), with slightly greater prevalence in women compared to men (3.0 vs. 2.6 %, respectively) [17]. While men begin having strokes at a younger age than women, women have more strokes than men for the age group greater than 85 years and also carry an overall higher lifetime risk of stroke compared to men [22]. Women have greater disability and likelihood of institutionalization poststroke, in part due to their strokes occurring at older age [21, 22]. Age-adjusted stroke prevalence is more common in Blacks (3.9 %) compared to Whites and Hispanics (2.5 and 2.6 %, respectively) [26]. Much, but not all, of the increased prevalence of stroke in Blacks compared to Whites may be related to their higher prevalence of stroke risk factors [26]. Stroke is relatively uncommon in children, and prevalence increases exponentially with age: 20–39 years, 0.4 %; 40–59 years, 2.0 %; 60–79 years, 7.7 %; and ≥ 80 years, 14.6 % [17, 26–28].
Stroke is often divided into two broad subgroups of ischemic and hemorrhagic stroke, which account for 87 and 13 % of strokes in the USA, respectively [17]. This proportion of ischemic to hemorrhagic strokes varies across the globe, however. For example, eastern countries such as China have a higher proportion of hemorrhagic strokes (33 % of total strokes compared to 12 % in comparative white populations of European origin) [29]. Ischemic strokes are often grouped by etiologic subtype based on suspected underlying cause of thrombosis into five categories with their corresponding proportion of all ischemic strokes: (1) large-artery atherosclerosis (13–17 % of ischemic strokes), (2) small-vessel disease (16–23 %), (3) cardio-embolic (22–29 %), (4) other (uncommon causes, 1–6 %), and (5) cryptogenic/undetermined (35–38 %) [30, 31]. The large percentage of strokes falling into the cryptogenic category is in part due to not having adequate evaluation to determine its etiology in one of the other categories.
While a minority of all strokes is due to primary hemorrhage, hemorrhagic strokes are associated with high morbidity and mortality [20, 32]. Hemorrhagic strokes are often divided into intracerebral hemorrhages (including intraparenchymal and intraventricular hemorrhages accounting for about 10 % of all strokes) and subarachnoid hemorrhages (accounting for about 3 % of all strokes) [17]. Globally, hemorrhagic stroke was found to account for more disability as measured by disability-adjusted life years (DALYs) than ischemic stroke (2.5 vs. 1.6 %) due to the younger age of death in hemorrhagic stroke patients leading to more years of life lost (YLL) which is a factor in DALYs [20]. More than one third of patients with intracerebral hemorrhage die within the first month [32]. Recent data from the Greater Cincinnati/Northern Kentucky Stroke Study comparing stroke rates between 1993–1994 and 1999–2005 indicate declining incidence of ischemic strokes, but not hemorrhagic strokes [33] .
Epidemiology of Seizures and Epilepsy in Stroke
Currently, it is estimated that the incidence of seizures in stroke is approximately 10 % with a very large range of estimates due to a number of factors including differences in study design, definitions, study setting (community vs. hospital), and seizure identification and classification [34]. The Seizures After Stroke Study, a prospective, hospital-based, multicenter study of consecutive stroke patients followed for an average of 9 months identified seizures in 8.9 % of patients [35]. Similarly, 9.8 % of patients admitted to the Stroke Unit of Ghent University Hospital, Belgium had seizures [36]. Community-based studies have also produced similar results. For example, patients followed in the Oxfordshire Community Stroke Project had an 11.5 % risk of having a seizure within the first 5 years following a stroke [37].
Seizures after stroke are often classified as either early or late. Early seizures are typically defined as occurring within 1 week of a stroke while late seizures are defined as occurring more than 1 week after a stroke. This distinction is important because early seizures are considered to be acute symptomatic (provoked) seizures while late seizures are considered to be unprovoked [1]. Of the 535 consecutive, new-onset stroke patients without a prior history of seizures in Rochester, MN, who were followed for 5.5 years, 33 had a seizure within 1 week of a stroke, and 78 % of those occurred within the first 24 h. An additional 27 new-onset stroke patients had a seizure more than 1 week following a stroke [38]. In northern Manhattan, 4.1 % of new-onset stroke patients had a seizure within 7 days of having a stroke [39]. In the Greater Cincinnati/Northern Kentucky Stroke Study, 3.1 % of patients with new-onset strokes had seizures within 24 h [40]. Interestingly, in the Greater Cincinnati/Northern Kentucky Stroke Study, there was no difference in seizure incidence between patients with initial versus recurrent strokes [40].
The risk of seizures varied with stroke subtype. Specifically, many studies reported an increased risk of seizures with hemorrhagic stroke compared to ischemic stroke. In the Seizures After Stroke Study, 10.6 % of patients with a hemorrhagic stroke had seizures, while 8.6 % of patients with an ischemic stroke had seizures [35]. Over 30 % of patients in the Oxfordshire Community Stroke Project with subarachnoid hemorrhage had seizures within 5 years, and over 25 % of patients with intracerebral hemorrhage had seizures within 5 years while about 10 % of patients with an ischemic infarction had seizures within 5 years [37]. In the Greater Cincinnati/Northern Kentucky Stroke Study, 8.4 % of patients with a hemorrhagic stroke had seizures within 24 h compared to 2.4 % of patients with an ischemic stroke [40]. In the Northern Manhattan Study, 14.3 % of patients with lobar hemorrhage, 4 % of patients with a deep hemorrhage, and 8 % of patients with subarachnoid hemorrhage had seizures within 7 days compared to 5.9 % of patients with a lobar ischemic stroke and 0.6 % of patients with a deep ischemic infarct [39].
Stroke is also associated with status epilepticus which can be associated with significant mortality [41, 42]. Status epilepticus was reported in 1.1 % of patients in the northern Manhattan population [39] and 3 % of patients in the Seizures After Stroke Study [35]. Neither study showed a difference between ischemic and hemorrhagic strokes. Other investigators have reported that status epilepticus is associated with more severe strokes [43, 44]. Nonconvulsive status epilepticus has been increasingly recognized in stroke patients. It presents with a wide range of clinical symptoms ranging from confusion to coma [42]. Routine use of continuous electroencephalography (EEG) monitoring has led to the identification of subclinical seizures in stroke patients [45]. Seizures were identified in 11 % of patients with ischemic stroke, 13 % of patients with intracerebral hemorrhage, and 19 % of patients with subarachnoid hemorrhage with the majority being nonconvulsive seizures (9 , 13 , and 18 %, respectively). A significant proportion of the patients were also diagnosed with nonconvulsive status epilepticus (7 , 9 , and 13 %, respectively). Nonconvulsive status epilepticus in subarachnoid hemorrhage has been particularly associated with poor outcome [46].
The risk of developing epilepsy following stroke is less well defined, likely related to the difficulties of the long-term follow-up. It is estimated that the risk of developing poststroke epilepsy is in the range of 2–5 % [47]. In a multivariate model [48], younger age, stroke severity at onset, lesion size, intracerebral hemorrhage, and early seizures were all associated with an increased risk of developing epilepsy. Cortical lesions [49] and lobar hemorrhage [50] increase the risk of developing epilepsy. Late-onset seizures in ischemic stroke have been reported to be a bigger risk factor than early onset seizures for the development of epilepsy [51, 52]. Late-onset seizures in hemorrhagic stroke have also been reported to be a bigger risk factor than early onset seizures [53] although this was not the case in the Seizures After Stroke Study [35].
The effect of seizures on outcomes remains an open question. In a study of 5027 Canadian stroke patients [54], patients with seizures had higher mortality at 30 days and 1 year, longer hospitalization, and worse modified Rankin scores at discharge. Since seizures are associated with more severe strokes, outcomes in patients with seizures are likely to be worse just as a result of having more severe stroke. In the Greater Cincinnati/Northern Kentucky Stroke Study [40], early seizures were associated with an increased risk of mortality at 30 days. However, after correcting for age, stroke type, and stroke severity, early seizures were not a predictor of poor outcome. In the Seizures After Stroke Study [35], seizures were associated with increased mortality at 30 days and 1 year and worse Rankin scores, but only for patients with ischemic stroke. In contrast, in northern Manhattan, early seizures were not associated with increased mortality after accounting for stroke severity [39]. Similarly, in an Italian cohort, early seizures were not associated with worse outcome [55].
Epidemiology of Seizures and Epilepsy in Select Cerebrovascular Diseases
There are several specific neurological syndromes that are clinically dominated by stroke, but also have seizures as a frequent complication. This section is not all inclusive but evaluates several conditions in which strokes and seizures commonly intersect and for which epidemiological literature is available. While the latest literature has been sought on the epidemiology of stroke and seizures related to each condition discussed below, several (such as reversible cerebral vasoconstriction syndrome (RCVS) and amyloid beta-related angiitis (ABRA)) have recently evolved and continue to evolve with respect to their respective clinical phenotypes and underlying pathophysiology.
Primary Central Nervous System Vasculitis
Primary central nervous system (CNS) vasculitis ( PCNSV) is a condition of inflammation of the small- and medium-sized blood vessels involving the brain and spinal cord without any identifiable secondary cause. PCNSV affects men more than women between the ages of 40 and 60 at the time of diagnosis [56, 57]. If not properly diagnosed and left untreated, it may be fatal in up to 95 % of biopsy-confirmed patients [58]. It is an uncommon condition estimated to affect 2.4 people per one million person-years based on a single population-based study from the Mayo Clinic Olmsted County population [59]. This incidence rate was based on 101 cases of PCNSV diagnosed by clinical criteria (less than one third of cases had biopsy confirmation). While PCNSV is primarily a disease of the cerebral vasculature, symptoms due to ischemic or hemorrhagic strokes are not the most common presenting symptoms of the disease. In the Mayo Clinic cohort of 101 patients described above, the most common presenting symptoms were headache (63 %) and altered cognition (50 %) followed by hemiparesis (44 %) [59]. Patients presented with seizures in 16 % of cases. During the disease course of PCNSV, strokes affect approximately 40 % of patients whereas seizures occur in less than 25 % of patients [59, 60]. EEG data are limited to two case series. In the Mayo Clinic case series of 101 patients, EEGs were abnormal 74 % of the time (28 of 38 EEGs) with dysrhythmias in 24 and epileptogenic findings in 4 [59]. In another older case series of 39 histologically confirmed cases with 28 EEGs performed, only two EEGs were normal, and the remainder except one showed generalized slowing, several of which had “superimposed focal findings” [60]. Numerous secondary causes of CNS vasculitis due to autoimmune illness, malignancy, or infection, are commonly associated with seizures at frequencies similar to that seen in PCNSV [61] .
Reversible Cerebral Vasoconstriction Syndrome
RCVS is a syndrome previously called “benign” PCNSV or Call–Fleming syndrome. Recently, greater understanding of the condition and its associations led to its new name along with proposed clinical diagnostic criteria in 2007 [62]. There are no incidence or prevalence data for RCVS, although it appears to be significantly more common than PCNSV given there are three substantial case series published since 2007 totaling 283 patients collected from only four institutions [63]. The condition affects women more than men and affects slightly younger patients than PCNSV with a median age of 42 years. The diagnosis is made by clinical presentation which typically begins with a severe thunderclap headache, and a cerebral angiogram shows focal and segmental narrowing of the cerebral arteries in the anterior or posterior circulation or both. In three case series of 67, 139, and 77 patients, strokes occurred in the form of ischemic strokes, intracerebral hemorrhage, and subarachnoid hemorrhages , leading to abnormal computed tomography (CT) or magnetic resonance imaging (MRI) findings in 12–81 % of cases (these percentages also included abnormal imaging findings related to posterior reversible encephalopathy syndrome). Seizures in these three case series ranged from 1 to 17 % [64–66]. EEG findings have not been widely reported in the literature .
Cerebral Amyloid Angiopathy
Cerebral amyloid angiopathy (CAA) is the most common cause of lobar intracerebral hemorrhage and a major cause of cognitive decline in the elderly. Only definite CAA can be diagnosed via postmortem demonstration of intracerebral hemorrhage in association with severe deposition of beta-amyloid protein in the small cortical and leptomeningeal arteries [67]. Therefore, the incidence of CAA is not fully established. One Japanese autopsy study evaluated 400 brains (only 26 of which had intracerebral hemorrhage) finding that 18 % of men and 28 % of women had evidence of CAA that increased with age such that no one younger than age 50 had CAA rising to greater than 40 % in those aged more than 90 years [68]. Utilizing prospectively collected MRI data, studies have determined recurrent hemorrhage rates in patients suspected of having CAA. In a series of 94 elderly patients with lobar hemorrhage, the total number of hemorrhages at baseline predicted the risk of recurrent hemorrhage such that a 3-year cumulative risk was 14 % with one hemorrhage and more than 50 % with six or more hemorrhages [69]. The clinical presentation of CAA-related hemorrhage is similar to intracerebral hemorrhage due to other etiologies. Symptoms include headache, focal neurological deficits, altered consciousness, and seizures which are in association with the size and location of the hemorrhage. In a recent study by Charidimou et al. [70], 14 % of 172 CAA patients also presented with symptoms termed “transient focal neurological episodes” (TFNE), often described as “recurrent, stereotyped, spreading paresthesias, usually lasting several minutes.” Half of their patients had positive “aura-like” symptoms such as paresthesias, limb-jerking, or visual symptoms such as flashing lights. The other half of their patients had predominately negative “transient ischemic attack (TIA)-like” symptoms including limb weakness or dysphasia. The majority (68 %) of patients had two or more stereotyped episodes with 70 % lasting less than 10 min. It is currently unclear whether these episodes represent seizure activity or may be an aura due to cortical spreading depression .
Amyloid Beta-Related Angiitis
ABRA and CAA-related inflammation (CAA-RI) are recently described syndromes which are diagnosed in patients who demonstrate clinical and histological features of both CAA and PCNSV [71]. The age of presentation falls between that for PCNSV and CAA occurring commonly in the 60s. Although the cohorts of ABRA and CAA-RI cases are small, seizures appear to be more common in ABRA and CAA-RI compared to CAA and PCNSV [71–73]. In a case series of 42 patients with biopsy-proven severe CAA, clinical presentation differed between patients with and without associated inflammation on biopsy [72]. Patients with inflammation were much more likely to present with cognitive decline (43 vs. 3 %) as well as seizures (57 vs. 3 %), while much less likely to present with intracerebral hemorrhage (0 vs. 94 %). A recent study of patients collected at Mayo Clinic between the years of 1983 and 2011 separated patients into four groups: ABRA (n = 28), CAA-RI (n = 10), CAA (n = 40), and PCNSV (n = 118; all biopsy-confirmed except for the PCNSV group which were biopsy-confirmed in one third) [73]. Comparing these four groups, seizures were most common in patients with CAA-RI (90 %) followed by ABRA (39 %), followed by CAA (25 %) and PCNSV (16 %) .
Vascular Malformations
Vascular malformations are associated with a number of neurologic symptoms including seizures and strokes. They are the most common cause of intracerebral hemorrhage in young adults [74]. Vascular malformations occur frequently and are increasingly recognized as imaging quality improves. Common vascular malformations include arteriovenous malformations (AVMs), cavernous malformations (CMs), and venous malformations (VMs). AVMs are congenital abnormalities consisting of multiple arteries and veins that are directly connected without a capillary bed. They most commonly present with hemorrhage, causing about 1 % of strokes. The second most common presentation is seizures, occurring in about 25 % of cases [75]. They can also be identified as incidental findings on MRI at a rate of 1 in 2000 patients [76]. In a prospective study of patients diagnosed with an unruptured AVM for reasons other than seizures, it was estimated that the rate of subsequently having a seizure is about 8 % after 5 years. The rate of developing epilepsy following a seizure with an unruptured AVM was estimated to be 58 % after 5 years. The risk of developing a seizure from an AVM after presenting with a hemorrhage or a focal neurological deficit was 23 % in 5 years and 48 % in patients who had a symptomatic seizure at presentation [77] .
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
