- Top of page
- Materials and Methods
Purpose: The incidence of seizures within 24 h of acute stroke has not been studied extensively. We aimed to establish the incidence of acute poststroke seizures in a biracial cohort and to determine whether acute seizure occurrence differs by race/ethnicity, stroke subtype, and/or stroke localization.
Methods: We identified all stroke cases between July 1993 and June 1994 and in 1999 within the population of the Greater Cincinnati metropolitan region. Patients with a prior history of seizures/epilepsy were excluded from analysis.
Results: A total of 6044 strokes without a history of seizure(s) were identified; 190 (3.1%) had seizures within the first 24 h of stroke onset. Of ICH/SAH patients, 8.4% had a seizure within the first 24 h of stroke onset (p ≤ 0.0001 vs. all other stroke subtype). Of the patients with ischemic stroke, we observed higher incidence of seizures in cardioembolic versus small or large vessel ischemic (p = 0.02) strokes. Patients with seizures experienced higher mortality than patients without seizures (p < 0.001) but seizures were not an independent risk factor of mortality at 30 days after stroke. Independent risk factors for seizure development included hemorrhagic stroke, younger age, and prestroke Rankin score of ≥1. Race/ethnicity or localization of the ischemic stroke did not influence the risk for seizure development in the studied population.
Discussion: The overall incidence of acute seizures after stroke was 3.1%, with a higher incidence seen in hemorrhagic stroke, younger patients, and those presenting with higher prestroke Rankin scores. Acute seizures were associated with a higher mortality at 30 days after stroke.
Cerebrovascular disease has long been recognized as a risk factor for the development of epilepsy and it is considered the most common identified antecedent condition that results in symptomatic epilepsy in the elderly (Loiseau et al., 1990; Hauser et al., 1994). The reported incidence of poststroke seizures and epilepsy is dependent on study design, diagnostic criteria, duration of follow up, and population studied (Pohlmann-Eden et al., 1996; Pohlmann-Eden et al., 1997; Silverman et al., 2002; Ferro & Pinto, 2004) Poststroke seizures are often characterized as “early” and “late” with definitions varying considerably (Pohlmann-Eden et al., 1996, 1997; Silverman et al., 2002; Ferro & Pinto, 2004) Early seizures are frequently defined as seizures that occur within the first 7–14 days of stroke symptoms onset (Kotila & Waltimo, 1992; Lamy et al., 2003; Feleppa et al., 2006) but many studies examining the incidence of early poststroke seizures (ES) and epilepsy focus either on poststroke seizures occurring within the first 1–4 weeks after the stroke (Giroud et al., 1994; Arboix et al., 1996, 1997; Reith et al., 1997; Labovitz et al., 2001; Feleppa et al., 2006) or on comparing the characteristics and outcomes of early and late seizures (Gupta et al., 1988; Hornig et al., 1990; Kilpatrick et al., 1990; Sung & Chu, 1990; Kilpatrick et al., 1992; So et al., 1996; Burn et al., 1997; Berges et al., 2000; Bladin et al., 2000; Bentes et al., 2001; Dhanuka et al., 2001; Velioglu et al., 2001; Lossius et al., 2002; Afsar et al., 2003; De Reuck et al., 2005; Misirli et al., 2006). The findings from the largest seizure incidence studies in the immediate poststroke period are similar: approximately 4.2–6.1% of patients develop seizures within the first few weeks after stroke (Kilpatrick et al., 1990; Davalos et al., 1992; Giroud et al., 1994; Reith et al., 1997; Bladin et al., 2000); early seizures may be predictive of epilepsy development (Kilpatrick et al., 1992; So et al., 1996). Only one population-based study focused on seizures occurring in the first 24 h after stroke (6%) but this relatively small study examined incidence of acute seizures in predominantly white and affluent population of Rochester, Minnesota that does not reflect the population characteristics of the United States (So et al., 1996).
The Greater Cincinnati/Northern Kentucky Stroke Study (GCNKSS) is designed to investigate the differences in stroke incidence rates and case mortality in the biracial population of the greater Cincinnati metropolitan area. Our study population generally is representative of the United States with regard to the median age, percentage of black race, median household income, education level, and percent of population below the poverty level (Broderick et al., 1998). Thus, our study should provide an excellent estimate of seizure incidence in the immediate poststroke period for the U.S. population.
The main goal of our population-based study was to establish the incidence of epileptic seizures within the first 24 h of stroke onset. Secondary aims were to determine the subtype and localization of stroke most important to the seizure incidence and to evaluate whether there are any racial differences in the incidence of seizures after stroke. Our hypotheses, based on the available literature, were that patients with hemorrhagic stroke have higher incidence of seizures when compared to patients with ischemic stroke and that blacks have higher incidence of seizures than Caucasian patients.
Materials and Methods
- Top of page
- Materials and Methods
The methodology of the GCNKSS has been previously described (Broderick et al., 1998). The study population for the GCNKSS is defined as all residents of the Cincinnati metropolitan region, which includes two southern Ohio counties and three contiguous Northern Kentucky counties that abut the Ohio River. Included in this area are 19 hospitals (18 in the 1999 data collection). Although residents of surrounding counties seek care at these hospitals, only residents of the five study area counties are included as cases. Previous studies have also documented that residents of the five counties who have a stroke exclusively seek care at these hospitals rather than more distant hospitals in the outlying region (Broderick et al., 1992a, 1992b). This study was approved by the Institutional Review Board at all participating hospitals.
Study nurses reviewed the medical records of all inpatients with primary or secondary stroke-related ICD-9 discharge diagnoses (430–436) from all acute-care hospitals in the study region. The study nurses also reviewed all autopsy cases where stroke was listed as the primary or secondary cause of death. Patients were identified as being from the study area based on their zip code of residence.
This study involved collection of all strokes that occurred in the study population between July 1, 1993 and June 30, 1994 and between January 1, 1999 and December 31, 1999. In addition to ascertaining inpatient strokes using the methodology described above, we determined outpatient strokes by monitoring all visits to all emergency departments, 5 county coroner's offices, 16 public health clinics, 13 hospital-based outpatient clinics and family practice centers, and the neurology and medicine clinics at the VA Hospital. Additional monitoring for outpatient strokes was performed in a random sample of primary care physicians' offices and area nursing homes.
To qualify as an incidence case, a patient must have met the criteria for one of the five stroke categories adapted from the Classification for Cerebrovascular Diseases III and from epidemiological studies of stroke in Rochester, Minnesota: cerebral ischemia, intracerebral hemorrhage (ICH), subarachnoid hemorrhage (SAH), stroke of uncertain cause, or transient ischemic attack (TIA) (Broderick et al., 1998). Cases were excluded if they had: (1) discharge/autopsy diagnosis or neuroimaging consistent with stroke but no clinical history of stroke; or (2) a clinical diagnosis of stroke and died within 24 h of symptom onset but had no focal neurological deficit and no confirmatory neuroimaging or autopsy. All excluded cases and the reasons for exclusion were recorded.
Once cases were identified, the study nurse gathered information from the chart regarding the presence or absence of seizures within the first 24 h after the onset of stroke symptoms. Data regarding the timing or duration of seizures or the clinical characteristics of seizures (focal vs. generalized) were not available in most cases and were not collected. Furthermore, information was collected regarding the stroke symptoms and physical examination findings, medical/surgical history including history of seizures prior to stroke, social history/habits, prehospital evaluation, vital signs and emergency room evaluation, diagnostic test results (including lab testing, electrocardiogram (EKG) and cardiac testing, neuroimaging of any type, etc.), treatments and outcome. Prestroke Rankin scores were determined in all patients. Classification of race/ethnicity was as self-reported in the medical administrative record. The study nurse abstracted all information and then made a determination as to whether a stroke or TIA had occurred. All borderline or possible cases were also fully abstracted for physician review.
Study physicians reviewed all abstracted charts and decided whether a stroke or TIA had occurred. Study physicians also reviewed available neuroimaging studies and characterized imaging findings. As previously, the study physician assigned stroke category and mechanism to each patient based on all available information, using definitions listed above and previously reported (Broderick et al., 1998).
For quality assurance and to assess interrater reliability, three study neurologists and two study nurse abstractors evaluated 18 medical records randomly selected from the database (both, cases and noncases). Each person decided whether the event was a stroke/TIA or not. The kappa (κ) score for these five study personnel was 0.83, indicating excellent agreement.
Data collected from 1993 to 1994 and 1999 were combined for analysis. In order to account for any potential effects of time, these time periods were included as covariates in multivariable analysis. To analyze the significance of observed differences between the patients who developed seizures and those that did not, a bivariate analysis was performed on the demographic characteristics using chi-square for categorical variables and the student's t-test or Wilcoxon rank-sum test for continuous variables, dependent upon the distribution. To determine the independent association of early seizures on mortality and independent factors associated with acute seizure development after stroke, multivariable logistic modeling was used.
- Top of page
- Materials and Methods
Of the 7,432 events that were fully abstracted by study nurses from both data collection periods (out of approximately 25,000 charts), and after physician review determined to meet case criteria, 6,044 met criteria for the current study (Fig. 1). Given our objective of evaluating the acute period after stoke only cases ascertained in the hospital setting were used and thus 976 were eliminated. There were 6,456 cases of hospitalized stroke. Of these, 412 patients were excluded due to previous history of seizures or epilepsy (N = 249), incorrect ICD code (N = 100), age <18 (N = 38), or race other than Caucasian or black (N = 25). Of the 6044 cases, 5,324 (88.1%) patients were determined to have an ischemic stroke or TIA, 715 (11.8%) a hemorrhagic stroke, and 5 (0.1%) a stroke of undetermined subtype.
Our study found the overall incidence of poststroke seizures within 24 h to be 3.1% (190/6044). No differences in seizure incidence was seen when comparing patients with first-ever versus recurrent strokes (51/1494, 3.4%; p = 0.49). Patients with seizures were significantly younger, had lower initial Glasgow Coma Scale (GCS) scores, higher National Institutes of Health (NIH) Stroke Scale (NIHSS) score, higher 30-day mortality and had a higher incidence of hemorrhagic stroke (Table 1) than patients without seizures.
Table 1. Demographic and clinical characteristics of all patients included in the study
|Variable||Seizure (n = 190)||No seizure (n = 5854)||p-value|
|Age (years)a|| 68.0||(15.0)|| 71.7||(13.2)|| 0.001|
|Gender (male)||77||40.5%||2556||43.7% ||0.39|
|Race (black)||33||17.34%|| 999||17.1%||0.91|
|Prestroke Rankinb|| 0||(0–3)|| 0||(0–3)||0.57|
|NIHSSb, c||12|| (5–22)|| 6|| (3–11)|| <0.0001|
|GCSb, d||13|| (7–15)|| 15|| (14–15)|| <0.0001|
|Hemorrhagic stroke||60||31.6%|| 655||11.2%||<0.001 |
|30-day Mortality||61||32.1%|| 776||13.3%|| <0.0001|
The incidence of seizures based on the stroke subtype is shown in Table 2. Patients with ischemic stroke and TIA had an overall incidence of seizures within the first 24 h after acute stroke of 2.4%. In contrast, the incidence of acute seizures in patients with hemorrhagic stroke was 8.4% (p < 0.0001) with 7.9% of ICH and 10.1% of SAH patients having acute seizures, respectively. When examining ischemic stroke mechanism, we noted an increased incidence of seizures in patients with cardioembolic strokes versus those with small or large vessel disease (p = 0.02; Table 3).
Table 2. Incidence of seizures in patients with ischemic and hemorrhagic strokes within the first 24 h of symptom onset (p < 0.0001)
|Type of stroke||Seizure (n = 190)||No seizure (n = 5854)||Rate of seizures|
|Infarct||120||63.2% ||4069||69.5% ||2.9%|
|TIA|| 10||5.3%||1125||19.2% ||0.9%|
|ICH/IVH|| 43||22.6% || 503||8.6%||7.9%|
|SAH|| 17||9.0%|| 152||2.6%||10.1% |
|Unknown|| 0||0.0%|| 5||0.1%||0.0%|
Table 3. Incidence of seizure by subtype of ischemic stroke (p = 0.01)
|Underlying Cause||Ischemic stroke and TIA||Rate of seizure|
|Seizure (n = 130)||No seizure (n = 5194)|
|Small vessel||12|| 9.2%|| 692||13.3%||1.7%|
|Large vessel||14||10.8%|| 805||15.5%||1.7%|
|Other|| 9|| 6.9%|| 137|| 2.6%||6.2%|
Patients who developed seizures within the first 24 h of stroke onset had a higher all-cause mortality rate at 30 days (32.1% vs. 13.3%; p < 0.0001; Table 1). To determine the independent association of seizure with mortality, and also to determine if any patient characteristics were independent predictors of seizure development in the acute setting of stroke, multivariable logistic regression analyses were performed (Table 4). Seizure was associated with a 2-fold increase in risk of 30-day mortality after controlling for age, hemorrhagic stroke, presence of heart disease, prior status as defined by Rankin scale, prior stroke, gender and race. When NIHSS and GCS were included in the multivariable logistic regression analyses, higher scores on the NIHSS (worse stroke) and lower GCS scores (poorer admission status) were associated with seizure occurrence (Table 5) with highest chance of seizures in hemorrhagic stroke patients with highest NIHSS and lowest GCS scores (Table 6). Additional multivariable logistic regression revealed that younger age, hemorrhagic stroke, and prestroke Rankin score of 1 or more were significant and independent predictors of seizure occurrence within the first 24 h after stroke (Table 7). Race/ethnicity was not significantly associated with seizure development or increased mortality.
Table 4. Association of seizure with mortality, showing covariates
|Age (10 years)||1.40||1.30, 1.51||<0.0001|
|Gender (male)||1.00||0.84, 1.18||0.98 |
|Race (black)||0.83||0.66, 1.05||0.12 |
|Prior stroke||0.93||0.77, 1.12||0.46 |
|Heart disease||1.55||1.31, 1.83||<0.0001|
|Prestroke Rankin (1 or more)||2.13||1.78, 2.54||<0.0001|
Table 5. Association of seizure with mortality, showing covariates including NIHSS and NIHSS/GCS
|Variable||Including NIHSS in model (741 missing)||Including NIHSS & GCS in model (1655 missing)|
|Seizure||1.49||0.94, 2.36||0.74||0.42, 1.29|
|Hemorrhage||4.47||3.39, 5.91||4.12||3.00, 5.65|
|Age (10 years)||1.34||1.25, 1.44||1.30||1.19, 1.41|
|Gender (male)||1.27||1.04, 1.55||1.37||1.08, 1.74|
|Race (black)||0.71||0.54, 0.96||0.74||0.53, 1.03|
|Prior stroke||0.75||0.60, 0.94||0.78||0.60, 1.02|
|Heart disease||1.52||1.24, 1.86||1.67||1.32, 2.11|
|Prestroke Rankin (1 or more)||2.06||1.65, 2.56||1.78||1.38, 2.30|
|NIHSS (10–19 vs. <10)||4.00||3.18, 5.02||2.85||2.14, 3.76|
|NIHSS (20+ vs. <10)||22.6||17.6, 29.2||9.01||6.41, 12.7|
|GCS (10–14 vs. 15)|| ||2.24||1.71, 2.94|
|GCS (<10 vs. 15)|| ||6.58||4.47, 9.68|
Table 6. Association of seizure with mortality, separately for Infarct/TIA and hemorrhagic stroke showing covariates
|Variable||Infarct & TIA with NIHSS in model (523 missing)||Hemorrhage with NIHSS & GCS in model (277 missing)|
|Seizure||1.55||0.90, 2.66||0.85||0.33, 2.15|
|Age (10 years)||1.36||1.25, 1.47||1.30||1.08, 1.49|
|Gender (male)||1.31||1.05, 1.64||1.37||1.77, 2.41|
|Race (black)||0.64||0.46, 0.89||1.40||0.70, 2.81|
|Prior stroke||0.68||0.53, 0.88||1.06||0.54, 2.10|
|Heart disease||1.56||1.26, 1.94||1.33||0.68, 2.61|
|Prestroke Rankin (1 or more)||2.33||1.82, 2.97||1.35||0.74, 2.44|
|NIHSS (10–19 vs. <10)|| 4.19||3.26, 5.38||1.71||0.86, 3.42|
|NIHSS (20+ vs. <10)||23.3||17.5, 30.9||5.69||2.69, 12.0|
|GCS (10–14 vs. 15)|| ||2.99||1.53, 5.84|
|GCS (<10 vs. 15)|| ||9.92||4.30, 22.9|
Table 7. Risk factors for seizure development: multivariable logistic regression
|Hemorrhage||3.65||2.62, 5.10||<0.0001 |
|Age (10 years)||0.86||0.77, 0.95||0.004|
|Prestroke Rankin of 1 or more||1.51||1.11, 2.06||0.009|
- Top of page
- Materials and Methods
This study was presented in part at the 125th ANA Annual Meeting, Boston, MA, October 2000 and in part at the 1st North American Regional Epilepsy Congress, San Diego, CA, December 2006. Support for this study was provided by NIH RO1 NS30678.
Conflict of interest: We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. None of the authors has any conflicts of interest to disclose in relation to this paper.