Abstract
Introduction and objectives
Epileptic seizures (ES) are a recognized complication of stroke, commonly associated with extensive ischemic regions and cortical damage. Despite thorough investigation, reports on the incidence, risk factors, and functional implications of post-stroke ES vary widely in the literature. We aimed to evaluate the predictive factors for post-stroke ES and their effects on the clinical outcome at hospital discharge.
Methods
Patients with acute ischemic stroke (AIS) admitted to our stroke unit from 2015 to 2017 were eligible to this study. A multivariable logistic regression modeling was built to assess associated variables with acute symptomatic seizures (AS). We defined AS as seizures that occurred within the period of hospitalization (stroke unit discharge) without a history of seizures with clinical or electroencephalographic evidence.
Results
Four hundred ninety-two were included in the final analysis. The patients had a mean of 66.7 (±14.4) years; 56% were male. Thirty-eight (7.7%) patients experienced clinical ES in-hospital, with a higher incidence in those with total anterior circulation syndrome. The NIH Stroke Scale score (odds ratio [OR] 1.07, 95% confidence interval [CI], p = .03) and symptomatic hemorrhagic transformation (HT) (OR: 3.53, 95% CI: 1.38–8.99, p = .01) independently predicted ES. We did not find an association between the occurrence of seizures and unfavorable outcomes (Modified Rankin Scale 3–6) at discharge (OR1.26, 95% CI: 0.3–5.32, p = .75). Patients with seizures had a longer hospital stay (18.5 [11–35] vs. 9 [7–14] days).
Conclusions
Patients presenting higher NIH Stroke Scale scores upon admission or experiencing symptomatic HT face an increased risk of in-hospital ES. Nonetheless, acute symptomatic ES following AIS does not independently predict adverse functional outcomes at discharge.
Introduction
Epileptic seizures (ES) are a recognized complication associated with ischemic and intracerebral hemorrhage events, supported by extensive studies.1–4 Prospective research indicates that the incidence of ES following an acute ischemic stroke (AIS) varies between 5.3% and 8.6%, with the majority manifesting within the initial 24 hours post-stroke.1–5 Focal ES emerges as the predominant seizure subtype in this patient cohort.5–8
Large ischemic regions, particularly cortical lesions, have been correlated with an elevated risk of secondary ES. 9 Patients with ES more commonly encounter clinical complications, including fever, pneumonia, urinary tract infection, and pulmonary embolism, as well as neurological issues such as new strokes and hemorrhagic transformations (HT).5, 9–12
Despite extensive investigation, data on the incidence, risk factors, and impact of ES secondary to stroke exhibits variability. This inconsistency may arise from methodological differences, variations in follow-up duration, and heterogeneous baseline clinical characteristics of the studied populations.5–10 Elucidating the clinical predictors and their implications for functional outcomes of acute ES is pivotal for both immediate and long-term patient management.
Our study aims to delineate the prevalence and characteristics of patients who develop clinical ES following AIS during hospitalization and to assess the influence of clinical seizures on functional outcomes at discharge.
Methods
Subjects
Patients admitted to the Stroke Unit at a tertiary public hospital, diagnosed with ischemic stroke through neurologist assessment and neuroimaging, were included in our local dataset. Eligibility depended on admission date and patient consent, obtained through a signed informed consent form. Data collection occurred at least thrice weekly, including weekends.
Exclusions from the final analysis were applied to patients whose ischemic stroke diagnosis was not confirmed, those undergoing intravenous thrombolytic therapy or mechanical thrombectomy, those lacking follow-up neuroimaging within seven days post-admission, and those transferred out within 72 hours of admission. These exclusions were necessitated by the convenience-based data collection schedule, which mandated inclusion of weekend data collection.
Ischemic stroke was characterized as a sudden neurological dysfunction with either persistent symptoms exceeding 24 hours and corroborative neuroimaging, or transient symptoms accompanied by definitive imaging findings of cerebral infarction. Patients with transient symptoms but no confirmed cerebral infarction, diagnosed as transient ischemic attacks, were excluded from the study.
Clinical complications, such as pneumonia, delirium, and urinary tract infections, were assessed exclusively for the duration of the patient’s stay in the Stroke Unit.
Outcomes
Mortality was evaluated throughout the duration of hospitalization, encompassing departments beyond the specialized stroke unit. Functional status at the time of discharge (or transfer from the stroke unit) was ascertained using the Modified Rankin Scale (mRS). The competencies of trained nursing staff were also appraised during the discharge process. An adverse outcome was delineated as a score ranging from 3 to 5 on the mRS. All included patients in the final analysis performed at least one follow-up neuroimaging within seven days after hospital admission. Both computed tomography (CT) and magnetic resonance imaging (MRI) were utilized as a neuroimaging follow-up, according to previous published papers.2, 5
Epileptic Seizures
Acute symptomatic ES were delineated based on established clinical criteria, which encompass single or multiple seizure episodes after a cerebrovascular event, irrespective of the temporal proximity to the ictal onset. Additionally, seizures were identified through electroencephalographic (EEG) evidence obtained via a 30-minute monitoring session. The implementation of EEG monitoring was contingent upon both the availability of the diagnostic service and the presence of clinical indicators suggestive of epileptic activity and could be executed at any juncture during the patient’s hospitalization period.
Hemorrhagic Transformation
HT was evaluated through follow-up neuroimaging conducted within seven days after hospital admission. Each patient incorporated into the final analysis underwent this follow-up neuroimaging. The decision to implement this temporal framework for follow-up neuroimaging was based on the specialized focus of our Stroke Unit, which is exclusively dedicated to the management of acute stroke patients (i.e., those with a time interval between stroke onset and hospital admission not exceeding 24 hours). HT was operationally defined as the presence of any intracranial blood, excluding dural or subarachnoid collections, in patients who had recently experienced an ischemic stroke. 5 Cases of HT identified beyond this specified time frame were omitted from the final analysis. To evaluate this pathological phenomenon, both CT and MRI were utilized. Symptomatic instances of HT were characterized by a neurological deterioration of four or more points on the National Institutes of Health Stroke Scale (NIHSS) within a 48-hour period following the identification of HT. 5
Ethics Rules and Data Collection
After receiving approval from our institution’s Internal Review Board, we gathered clinical, demographic, neuroimaging, and laboratory data from the included patients. This data, spanning from admission to discharge or transfer, was obtained via interviews, medical records, and various forms, including image reports and additional examinations. An 82-item instrument was utilized to ensure consistent data collection and analysis of diverse variables.
Statistical Analysis
In this retrospective study, we analysed variables including age, sex, prior stroke history, ASPECTS, microangiopathy grade, NIH Stroke Scale at admission, admission arterial blood pressure, blood glucose levels, Oxfordshire Community Stroke Project classification, and TOAST etiological classification. Numerical variables were presented as means and standard deviations for normally distributed data, and medians with interquartile ranges for non-normally distributed data. The Kolmogorov–Smirnov test and stem-and-leaf plots evaluated variable distribution. Categorical variables, expressed as percentages and absolute numbers, underwent Fisher’s test for univariate analysis to examine their relationship with seizure incidence. Non-normally distributed quantitative variables, identified by the Kolmogorov–Smirnov test and stem-and-leaf plots, were assessed using the Mann–Whitney test. Automated logistic regression with stepwise backward adjustment, based on literature-reported variables, identified factors independently linked to ES during hospitalization. To determine the association between ES, other variables, and unfavorable outcomes (mRS, scores of 3–5 at discharge), we initially performed univariate analysis, followed by multivariate modeling. Statistical significance was set at a p value ≤ .05, using two-tailed tests, with analyses conducted on IBM SPSS Statistics for Windows, Version 25.
Results
In our study, 492 patients diagnosed with ischemic stroke were incorporated into the final analysis, as delineated in Figure 1. Among these participants, 38 (out 492) (7.7%) experienced acute ES during their hospitalization within our specialized Stroke Unit. The clinical, laboratory, and neuroimaging characteristics of these patients are summarized in Table 1.
Flowchart of Patients Inclusion.
Distribution of Baseline Clinical Characteristics and Complications According to the Presence of Seizures in Patients Hospitalized for Ischemic Stroke.
SD = Standard Deviation; IQR = Interquartile range; NIH = National Institute of Health; ASPECTS = Alberta Stroke Program Early CT score; SBP = Systolic Blood Pressure; DBP = Diastolic Blood Pressure.
*Mann-Whitney Test.
†Fisher Test.
When comparing patients who experienced ES during their hospital stay with those who did not, we observed a statistically equivalent distribution of males and females (57% vs. 55.5%, p = .86). Furthermore, age distribution did not show any significant difference between the two groups (63 [±13.4] vs. 65 [±15.6] years, p = .85).
Patients who manifested seizures presented with higher scores on the NIHSS upon admission (17 [10–21] vs. 13 [6–20], p = .02). Moreover, a greater frequency of Total Anterior Circulation Syndrome (TACS) according to the Oxfordshire classification was observed in this cohort (76.3% vs. 45.8%, p = .01), along with a protracted duration of hospitalization (18.5 [11–35] vs. 9 [7–14] days).
The Alberta Stroke Program Early CT Score (ASPECTS) was assessed in 453 patients, as 39 displayed hypodensity in the posterior circulation territory at the time of admission. No significant difference was noted in the median ASPECTS values between patients with and without clinical seizures. Furthermore, the cardioembolic etiology of ischemic stroke, as defined by the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification, was also comparable between the two groups.
The incidence of any HT was significantly higher among patients who developed acute ES (36% vs. 17% and 23% vs. 7%, p = .01), as was the occurrence of pneumonia (65% vs. 29%, p = .01). The temporal relationship of HT—whether preceding or following seizure identification—was also considered. No difference was observed in the incidence of delirium or urinary tract infections, as indicated in Table 1. We did not assess whether the occurrence of pneumonia, delirium, or urinary tract infections occurred before or after the symptomatic seizure.
A multivariable regression analysis revealed that higher NIHSS scores at admission (Odds Ratio [OR]: 1.07, 95% Confidence Interval [CI]: 1.01–1.13, p = .03) and the presence of symptomatic HT (OR: 3.53, 95% CI: 1.38–8.99, p = .01) were significantly associated with the development of acute symptomatic ES post-ischemic stroke, as shown in Table 2.
Multivariate Logistic Regression for Seizure as Outcome.
With regard to unfavorable outcomes (mRS scores of 3–5) at the time of discharge from the Stroke Unit, several variables were found to be significantly associated in a bivariate analysis: age (OR: 1.02, 95% CI: 1.00–1.04, p = .04), NIHSS scores at admission (OR: 1.26, 95% CI: 1.18–1.34, p = .01), ASPECTS (0.058, 95% CI: 0.71–0.93, p = .002) and pneumonia (OR: 8.35, 95% CI: 2.42–28.78, p = .01). On the other hand, the presence of clinical acute ES was not independently associated with mRS 3–5 (OR: 0.71, 85% CI 0.36–1.43, p = .34). Considering the outcome of mRS 3–6, the presence of clinical acute ES was only significant in the bivariate analysis (OR: 3.77, 95% CI: 1.1–12.5, p = .03) and lost its significance when adjusted for age, ASPECTS, NIHSS scores, and pneumonia, as presented in Table 3.
Logistic Regression Analysis of Factors Associated with Poor Outcome (mRS 3–5) at Discharge.
The prevalence of in-hospital mortality was substantially elevated among patients who experienced seizures, as compared to those without clinical seizures (28.9% vs. 5%, p < .001).
Discussion
In our sample, the prevalence of ES was analogous to that observed in prospective studies and systematic reviews of acute seizures following a stroke.6, 10, 1113–18 Our results corroborate findings from the literature, emphasizing the contribution of HT as a risk factor for symptomatic ES, even in patients not subjected to reperfusion therapies (RT). Generally, the assessment of HT in this patient group is given less attention compared to patients treated with thrombolysis, for instance. Furthermore, our study suggests a causative relationship with poorer functional outcomes, despite the methodological limitations and the lack of measurement of the occurrence times of complications such as pneumonia and urinary tract infection.
Our study revealed no significant differences in gender or age among patients with or without seizures, aligning with previous research.16, 19 While some studies5, 16, 19, 20 identified male gender and younger age as risk factors, our findings suggest that these disparities may stem from baseline characteristics in younger patients, such as ischemic volume and more distinct epileptic features.14, 18 Additionally, clinical baseline characteristics showed no variance between older and younger groups.
In accordance with the methodology employed by select authors in the extant literature, 11 our investigation corroborated a 1.6-fold elevated prevalence of Total Anterior Circulation Strokes (TACS) within the analysed cohort. This lends further credence to the well-documented correlation between large cortical lesions subsequent to a stroke event and the onset of seizures.7, 8, 10, 11, 14–22 However, it is noteworthy that, upon undertaking adjusted analyses, TACS did not exhibit a statistically significant association with seizures.
In concurrence with existing publications, our findings also indicated that a higher score on the NIHSS upon admission was significantly correlated with an elevated risk of acute clinical seizures. Although an array of risk factors has been delineated in prior research, the severity of the stroke, as quantified by the NIHSS score, has consistently emerged as a prognostic indicator for the likelihood of seizure episodes.10, 11, 15– 18
Moreover, our data illuminates an increased seizure risk in association with symptomatic HT. We postulate that this correlation is likely attributable to the expansive and cortical nature of the ischemic regions, compounded by ancillary factors such as aging, cardioembolism, and diabetes mellitus. These factors may engender a proclivity for ES, independent of HT.5, 10, 14, 16, 19, 21 Additionally, the incursion of blood into ischemic cerebral tissue may catalyze an inflammatory cascade, culminating in cerebral edema and disruption of neuronal homeostasis. 5
Contrary to certain studies, our investigation did not reveal a statistically significant prevalence of cardioembolic strokes in patients presenting with ES. We speculate that the observed lack of association may stem from a high degree of collinearity between cardioembolism and HT within our sample.21, 22
The association between post-stroke ES and adverse clinical outcomes remains equivocal.10–12 While several clinical reports have underscored an extension in the duration of hospitalization and elevated morbidity and mortality among patients experiencing seizures,5, 9, 23 our data did not substantiate a correlation between seizures and unfavorable outcomes, including mortality, corroborating some prior research.15, 19
Our study is not without limitations. First, our analysis is retrospective and confined to a single-center dataset, exclusively comprising patients who were not administered RT. Consequently, differences in baseline characteristics between patients who did or did not receive RT could engender divergent outcomes. Second, our dataset lacks temporal specificity concerning the interval between the vascular event and seizure onset, thereby necessitating the categorization of seizures as acute based on the average hospital stay duration of 18 days. Third, the classification of ES was predicated solely upon clinical manifestations corroborated by intermittent EEG analysis; continuous EEG monitoring was not implemented in patients devoid of clinical manifestations. Fourth, the study retrospectively assessed a preselected array of clinical complications, precluding the documentation of additional variables (e.g., pain, pulmonary embolism, hyperglycemia). Lastly, outcomes were ascertained solely during the in-hospital period, limiting the broader applicability of our findings—outcome at three months after stroke onset, for instance, could be more valuable to the practicing clinician to help advise the patient and family.
Our study is not without limitations. First, our analysis is retrospective and confined to a single-center dataset, exclusively comprising patients who were not administered RT. Consequently, differences in baseline characteristics between patients who did or did not receive RT could engender divergent outcomes. Second, our dataset lacks temporal specificity concerning the interval between the vascular event and seizure onset, thereby necessitating the categorization of seizures as acute based on the average hospital stay duration of 18 days. Third, the classification of ES was predicated solely upon clinical manifestations corroborated by intermittent EEG analysis; continuous EEG monitoring was not implemented in patients devoid of clinical manifestations. Fourth, the study retrospectively assessed a preselected array of clinical complications, precluding the documentation of additional variables (e.g., pain, pulmonary embolism, hyperglycemia). Lastly, outcomes were ascertained solely during the in-hospital period, limiting the broader applicability of our findings—outcome at three months after stroke onset, for instance, could be more valuable to the practicing clinician to help advise the patient and family.
Conclusion
Upon admission, a high NIH Stroke Scale score and symptomatic HT are risk factors for acute clinical seizures post-ischemic stroke. Nevertheless, these seizures do not independently predict unfavorable outcomes at hospital discharge.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Ethical Statement
This research has received ethical approval from our Institutional Review Board (IRB), sanctioned under the protocol number (CAAE) 24500719.7.0000.5040.
Funding
Dr Andrade’s visiting scholarship at Columbia University, New York City, is sponsored by the Capes Foundation, Ministry of Education, Brazil.
Informed Consent
All the participating patients have provided their informed consent, acknowledging their understanding and agreement to partake in this study.
MeSH descriptor
[C10.228.140.300] All cerebrovascular disease/Stroke
[E01.789] Prognosis
[C10.597.742] Seizure
