Frequency and Predictors of Early Seizures Following First Acute Stroke: Data from a University Hospital in South India
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.325345
Source of Support: None, Conflict of Interest: None
Keywords: Acute stroke, early seizures, lesion location, modified Rankin's scaleKey Message: There is a frequency of 11.8% of early onset post-stroke seizures, which is slightly higher when compared with other studies, with tendency of seizures being the highest in cerebral venous thrombosis (40.9%) followed by ischemic stroke with hemorrhagic conversion (20%), hemorrhagic stroke (8.7%), and ischemic stroke (5.2%). Most of these occur within 24 hours (7.5%), althou
Seizures are well known to occur after acute brain insults such as stroke, cerebral sinus venous thrombosis (CSVT), traumatic brain injury, central nervous system (CNS) infections, including granulomas, febrile seizures, demyelinating disorders, and autoimmune/connective tissue disorders., Acute stroke is known to be associated with seizures and is one of the most common causes of epilepsy in elderly.,,, With increasing life expectancy, stroke patients are expected to increase in number in the coming years. Seizures that occur at the time of brain injury are called early or acute symptomatic seizures. These seizures may suggest the severity of insult and may have an effect on hospital stay, morbidity, and mortality.,, A series of biological changes occur in the long term with variable latency over the injured area, resulting in late unprovoked seizures and epilepsy. At present, there are no effective drugs to prevent epilepsy after brain insult. Conducting well-designed primary prevention trials may help in the development of antiepileptogenic drugs.
Hence, identifying the frequency and predictors of early seizures and epilepsy in acute stroke patients may help in selecting the high risk patients for primary prevention trials. It may help us in appropriately managing these patients and determining the clinical outcome, which may have significant impact on their quality of life. There are not many studies in India that address these aspects, especially early seizures associated with venous stroke [Table 1].
This article addresses the frequency and predictors of early seizures as well as their influence on clinical outcome in patients with first stroke including acute venous stroke.
This is a hospital-based, prospective case-control, observational study conducted among consecutive patients with the diagnosis of first arterial and venous stroke who were admitted in the Nizam's institute of Medical Sciences, Hyderabad, a university health hospital from South India, from January 2015 to November 2015. A total of 1922 patients were admitted in the Neurology department, and out of them, 280 consecutively-admitted eligible patients with diagnosis being confirmed by neurological imaging studies such as computerized tomography (CT)/magnetic resonance imaging brain (MRI)/cerebral angiography were included in this study. Patients and relatives were fully explained regarding the study and a written informed consent was taken. The study protocol was approved by the Institutional Ethics committee. Patients with recurrent stroke, past history of seizures, and presence of metabolic disturbances such as hepatic failure, renal failure, respiratory failure, hypoglycemia, hyperglycemia, electrolyte and acid-base disturbances, drug intoxication, or withdrawal, which can cause seizures, were excluded.
The data was collected at the time of admission and during hospital stay according to the proforma which included demographic data, clinical history, risk factors, clinical examination, and all other relevant investigations. Electroencephalography (EEG) was done only when indicated. Patients were followed up for at least seven days or till discharge, whichever was longer.
Stroke was suspected when there was sudden onset of acute neurologic symptoms with vascular origin lasting for 24 hours or longer, and was further confirmed by neurological imaging studies (CT/MRI brain). CSVT was diagnosed on the basis of clinical features and radiological evidence of thrombosis in any one of the cerebral venous sinuses or cortical veins.
Seizures were diagnosed by reliable history or by medical personnel during the hospital stay. They were divided into three separate categories based on the timing of occurrence, i.e., early onset seizures within 24 hours of stroke, within 7 days, and late onset (>7 days). Seizure semiology was classified as focal seizures, complex partial seizures, focal seizures with secondary generalization, and generalized seizures. EEG was done in most of the patients who developed seizures and was classified as normal, slow, or epileptiform. The antiepileptic drugs (AEDs) used to treat seizures were noted and coded into monotherapy or polytherapy.
For stroke patients, impairment and severity were assessed using the National Institutes of Health Stroke Scale (NIHSS) and Glasgow coma scale (GCS) score, respectively, which were further coded into mild (NIHSS <7; GCS ≥13), moderate (NIHSS = 7–25; GCS = 9–12), and severe (NIHSS >25; GCS ≤8). The modified Rankin's scale (mRS) was used to assess the handicap caused by stroke, which was categorized into mild/absent (mRS ≤2) or moderate/severe (mRS >2). The stroke subtype was categorized as ischemic stroke (IS), IS with hemorrhagic transformation (ISH), hemorrhagic stroke (HS), and CSVT. Lesions on neurological images were classified based on their maximum diameter as large (>3 cm), medium (1–3 cm), small (<1 cm), or lacunar lesion (<1 cm deeply located in the cerebral hemispheres or the brainstem); lesion site was classified as cortical, cortical and subcortical, or subcortical. Lesion location was classified as supratentorial (ST), infratentorial (IT), or both, and the cerebral lobes involved were noted.
In IS patients, Bam ford criteria was used to classify the clinical syndrome and Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria was followed for etiological classification. In CSVT patients, severity was assessed by GCS and coded as mild ≥13, moderate 9–12, and severe ≤8. The thrombosed sinuses and cortical veins were identified and coded as isolated or multiple vein involvement.
A number of recognized risk factors for stroke and CSVT were identified such as arterial hypertension (present treatment with antihypertensive drugs), diabetes (present treatment with insulin or oral hypoglycemic agents), cardiac disorders (any treatment of cardiac disorders or arrhythmia), hyperlipidemia (cholesterol levels >200 mg/dl or triglyceride concentration >140 mg/dl or treatment with statins), hyperhomocysteinemia (serum homocysteine > 15 mg/100 ml in those below 60 years and >20 mg/100 ml in those above 60 years), alcohol intake (current dependency), smoking (daily), previous TIAs (any focal cerebral ischemic event with symptoms lasting <24 hours, current usage of oral contraceptives/hormone replacement therapy, pregnancy, puerperium (till 6 weeks after delivery), hypercoagulable states (both genetic and acquired such as proteins C, S, antithrombin III deficiency, factor V Leiden, and G20210A mutations, APLA syndrome, polycythemia, thrombocytosis), vasculitis (systemic lupus erythematosus, rheumatoid arthritis, not specified), malignancy (solid tumors or hematological), and CNS Infections.
The demographic and clinical data of the patients is presented here as percentage or mean with standard deviation. The number of patients with one or more early onset seizures was calculated to provide the frequency of seizures occurring within 24 hours and seizures occurring within 7 days in percentage. They were compared to patients without seizures in the first 7 days, and the frequency of each seizure predictor was calculated for each group separately by univariate analysis using Students t-test or χ2 test. Those with P values of <0.05 were considered significant and further analyzed by multivariate logistic regression analysis. Statistical significance was chosen at the 5% level. Data was analyzed using the Stastical Package for the Social Sciences version 20 for Windows 8 (SPSS, IBM).
A total of 279 patients were enrolled in the study, with a mean age of 50.04 ± 17.13 years and range of 7–85 years; 67% of the patients were males. The mean GCS was 13.64 ± 2.35; hospital stay was 9.57 ± 4.85 days, and 8 patients expired, with in-hospital mortality rate of 2.9%. IS (62%) was the most common stroke subtype, followed by HS (19.9%), CSVT (15.9%), and ISH (1.8%). Relevant demographic and clinical data of all the stroke subtypes is summarized in [Table 2], [Table 3], [Table 4].
Supratentorial location was the most common and was seen in 72% of all patients. Majority of IS patients had small lesions (36.7%), ISH had large lesions (80%), HS had medium lesions (85.2%), and CSVT patients had no visible parenchymal lesions (46.5%). Cortical lesions were more common in IS (59.2%), ISH (100%), and CSVT (48.9%), whereas subcortical lesions were predominant in HS (96.3%). Post-stroke handicap was more (mRS >2) with IS (69.8%), ISH (80%), and HS (85.2%), however, CSVT (60.5%) patients had mild handicap (mRS ≤2). posterior circular infarct (36%) was the most common clinical syndrome, and most common eetiology was large artery atherosclerotic disease (50%) among patients with IS [Table 3].
Among patients with IS (65.1%), ISH (80%), and HS (85.2%), hypertension was the most prevalent risk factor. In CSVT patients, alcohol (37.2%) followed by hyperhomocysteinemia (34.9%) were the most prevalent risk factors.
Frequency rate of early onset post-stroke seizures was 33 (11.8%), among which 21 (7.5%) occurred within 24 hours of stroke [Table 4]. We performed continuous EEG in patients with unexplained decreased level of unconsciousness, and thus certain patients with seizures were diagnosed retrospectively. In this study, there was only 1 case of complex partial status epilepticus in HS with multilobar bleed; the patient was managed in the intensive care unit (ICU) with phenytoin, levetiracetam, and clobazam. Ten patients had clusters of seizures, of which 7 were managed in the ward and 3 in the ICU with levetriacetam, oxeta, frisium, and eptoin. EEG was performed in 20 patients, out of which majority 9 (45%) were slow; only 5 (25%) had epileptiform abnormalities. We included 7 cases of pediatric strokes. Only 1 pediatric patient had clusters of seizures and was managed with phenytoin, levetracetam, and clobazam in the neurology intensive care unit (NICU).
The significant demographic and clinical features of patients with and without seizures are summarized in [Table 5] and [Table 6]. Among the stroke subtypes, CSVT (P < 0.005) 18 (40.9%) had the highest frequency of early onset seizures followed by ISH (P < 0.005) 1 (20%), HS 5 (8.7%), and IS 9 (5.2%). Patients with early post-stroke seizures were younger (38.88 ± 16.74 years) compared to those without seizures (51.53 ± 16.69 years) (P < 0.005). Among the risk factors, hypertension (P < 0.005) and smoking (P = 0.031) were more prevalent in patients without seizures 157 (63.8%) and 74 (30.04%), respectively when compared to those with seizures 10 (30.3%) and 4 (12.12%), respectively. However, hypercoagulable states (P < 0.005) (6% versus 1.6%) and hyperhomocysteinemia (P = 0.011) (24% versus 13%) were more common in patients with seizures than without seizures. Cortical lesion (P < 0.005) (64% versus 45%), supratentorial lesion (P = 0.001) (76% versus 71%), and large lesion (P < 0.005) (70% versus 24%) were more frequent in patients with seizures compared to those without seizures. Patients with seizures had higher mean duration of hospital stay (P = 0.019) 11.78 ± 6.31 days whereas those without seizures had 9.27 ± 4.56 days. There was no statistically significant difference in disability and mortality between the two groups.
In multivariate analysis, the independent seizure predictors are CSVT patients (OR = 10.2; 95% CI = 4.6–22.5), hypercoagulable states (OR = 5.5; 95% CI = 1.5–20.7), hyperhomocysteinemia (OR = 2.1; 95% CI = 0.9–5.2), large lesion (OR = 7.3; 95% CI = 3.3–16.2), cortical location (OR = 2.1; 95% CI = 1–4.5) and supratentorial location (OR = 1.2; 95% CI = 0.5–2.8). Large lesion (OR = 5; 95% CI = 0.2–12.4), cortical lesion (OR = 1.3; 95% CI = 0.4–4.8), and supratentorial lesion (OR = 1.05; 95% CI = 0.3–3.7) are risk factors in IS. Large lesion (OR = 9.3; 95% CI = 4.2–20.8) and cortical lesion (OR = 1.9; 95% CI = 0.2–17.4) are risk factors in HS. Large lesion (OR = 4.8; 95% CI = 1.8–13.2) is a risk factor in CSVT. Smoking (OR = 0.1; 95% CI = 0.01–0.8), and hypertension (OR = 0.2; 95% CI = 0.1–0.5) are associated with lower risk of seizures. Hospital stay (OR = 1.3; 95% CI = 1–1.6) is increased in patients with early seizures.
The frequency of post-stroke seizures reported different in previous studies range from 1.6 to 33%.,,,, This variability could be due to the differences in the study design, the analysis of retrospective studies, varying definition of acute symptomatic seizures (ranging from 1 to 30 days), and the inclusion of patients with different types of stroke. Several studies that analyzed the predictive factors for early post-stroke seizures reported controversial results. Cortical lesion, large lesion, stroke severity, and hemorrhagic stroke are the common predictive factors reported.,,,,, The prospective design of our study allowed us to perform a systematic search of various risk factors for seizures and assessing the seizures in the hospital, which could have helped us in getting accurate and complete information.
In our study, the incidence of early onset post-stroke seizures was 11.8% and was the highest (7.5%) during the first 24 hours, which tended to decrease thereafter. This was higher in frequency when compared with most of the other prospective studies, which showed cumulative incidence rates of early seizures, ranging from 2.4% to 14% [Table 6], which could be due to the inclusion of venous stroke patients. However, when we compare our individual series such as seizures following IS (5.2%) and venous stroke (40.9%), it is similar to other studies compared in [Table 7]. It has been suggested that early onset post-stroke seizures are caused by transient metabolic and ionic derangements, which may result in the depolarization of the transmembrane potential lowering the seizure threshold. There are not many studies from developing countries regarding the same issue. We found one study from Nizeria; patients with stroke were 28 out of 94 at 29.8%.
Mean age of onset in our study was 50.9 ± 17.27 years, which is less when compared to most other studies.,, Sixty-seven percent of the recruited patients were males in our study, which is similar to another study. In our study, patients with younger age tended to have seizures, which could be due to inclusion of venous stroke.
There are very few studies that have studied the frequency and predictors of seizures after CSVT, which is a common cause of stroke in India. In our study, frequency of early seizures following CSVT were 40.9% and large lesion size was the predictor for early seizures, and those with seizures were found to have longer duration of hospital stay, even though not related to mortality. According to a prospective study conducted by Ferro et al., 39.3% patients presented with seizures and 6.9% had early seizure (<14 days). Supratentorial lesion, cortical vein thrombosis, sagittal sinus thrombosis, and puerperal CSVT were shown to be associated with acute seizures. Maschur et al. reported 44.3% incidence rate of early seizures with motor deficit, with intracranial hemorrhage and cortical vein thrombosis being independent predictors of early seizures. In an Indian study, Kalita et al. reported 50% incidence of presenting seizures, which was independently related to supratentorial lesion. However, in our study, all these factors did not have a significant association.
In our study, the significant predictors of early seizures for all stroke subtypes together were younger age, venous stroke, cortical and supratentorial location, large size, hypercoagulable states, and hyperhomocysteinemia. Duration of hospital stay was increased in patients with seizures. The predictors of early post stroke seizures were compared with other studies in [Table 1].
Our study is in correlation with other studies regarding having a statistical association with supratentorial and cortical location. Cortical location has been consistently found to be associated with post-stroke seizures, which is comparable with other studies. Cortical regions are highly susceptible to seizures following differing brain insults. The neurobiological changes occurring during stroke share some similarities to those induced by status epilepticus or traumatic brain injury.
In our study, it was found that when compared between various stroke subtypes venous stoke and ischemic stroke with hemorrhagic conversion were associated with seizures rather than hemorrhagic stroke, as in the study by Beghi et al. This might be because of the strong association of cortical/supratentorial location and large size with seizures rather than the hemorrhagic component. Among the risk factors for stroke, hypercoagulable states and hyperhomocysteinemia are significantly associated with seizures, which were not seen in other studies.
Duration of hospital stay is more in patients with early seizures, which in turn may increase morbidity to patients. This correlation is strong despite the lack of statistically significant association of early seizures, with stroke severity measured by GCS and in-hospital mortality. There was no association with functional disability in grades of NIHSS and mRS and mortality in our study, which is in contrast with other studies that reported higher disability and mortality in patients with early onset post-stroke seizures., Hence, occurrence of seizures may increase hospital stay, which in turn may increase financial burden to patients.
There is a frequency of 11.8% of early onset post-stroke seizures, which is slightly higher when compared with other studies, with tendency of seizures being the highest in cerebral venous thrombosis (40.9%) followed by ISH (20%), HS (8.7%), and IS (5.2%). Most of these occurred within 24 hours (7.5%) though early seizures can manifest within seven days of stroke onset. Most common predictors of early seizures are venous stroke, large lesion, cortical location, supratentorial location, hypercoagulable states, and hyperhomocysteinemia for all strokes. Large lesion, cortical, and supratentorial location are risk factors in IS. Large lesion and cortical lesion are risk factors for HS. In acute venous stroke, large lesion is a risk factor for early seizures. Duration of hospital stay is more for those with early onset post-stroke seizures, which in turn may increase financial burden to patients. Functional disability and in-hospital mortality are not associated significantly with seizures.
Limitations of the study
The limitation of our study is, that in our study, EEG tracings were performed in 20 out of the 33 patients with early seizures. Our study did not investigate patients by EEG who were not experiencing seizures. Only early seizures are analyzed and predictors of late-onset epilepsy could not be studied.
We could not assess the long-term outcome of patients so that it was useful for assessing epileptogenesis. We excluded recurrent stroke, past history of seizures, and presence of metabolic disturbances such as hepatic failure, renal failure, respiratory failure, hypoglycemia, hyperglycemia, electrolyte and acid-base disturbances, drug intoxication, or withdrawal, which can cause seizures. We could not conduct an age and sex-matched case control study. We have included only few pediatric cases. There is need for a future study which includes only pediatric cases.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7]