Do Early Seizures indicate survival of patients with nontraumatic intracerebral hemorrhage?
Toralf Brüning, MD1,2, Samer Awwad1, Mohamed Al-Khaled, MD1
1Department of Neurology, University of Lübeck, Lübeck, Germany
2Department of Neurology, Bundeswehrkrankenhaus, Hamburg, Germany
Corresponding author:
Dr. Toralf Brüning
Department of Neurology
University of Lübeck
Ratzeburger Allee 160
23538 Lübeck
Germany
Tel: +49 (0)451 500-2926
Fax: +49 (0)451 500-2489
E-mail:
Tables: 2
Figures: 1
Key Words: seizure, stroke, nontraumatic spontaneous intracerebral hemorrhage, mortality, outcome
Abstract
Background and Purpose—Early seizures (ESs) in patients with nontraumatic spontaneous intracerebral hemorrhage (sICH) are a frequent complication. The aims of the present study were to determine the frequency of ESs in patients with sICH and to investigate the association of ESs with outcomes in a monocenter study.
Methods—During a 5-year period (2009–2013), 484 consecutive patients (mean age, 72.3 ± 12.6 years; female sex, 51%) with sICH who were admitted to the Department of Neurology at the University of Lübeck, Germany were enrolled and prospectively evaluated.
Results—A total of 52 patients (10.7%; 95% confidence interval [CI], 8-14) experienced ESs during a mean hospitalization of 12 days. Patients with ESs were less affected on the National Institutes of Health Stroke Scale (NIHSS) at admission than those without ESs (7 vs 10; P = 0.02). With the exception of the localization of hemorrhage (P = 0.008), differences in the baseline characteristics between patients with ESs and those without ESs were not found. The logistic regression analysis revealed an increased ES rate in patients with cortical hemispheric sICH (odds ratio [OR], 3.5; 95% CI, 1.8-6.7; P < 0.001). During hospitalization, 109 patients (23%) died, the in-hospital mortality was lower in patients with ESs than those without (9.6% vs 24.0%, respectively; P = 0.02). An association between ESs and good functional outcome on the modified Rankin Scale (mRS) ≤ 2 was not found (P = 0.3).
Conclusion—ESs appear to be correlated with hemorrhage localization and associated with survival of the sICH.
Despite all efforts, nontraumatic spontaneous intracerebral hemorrhage (sICH) remains a devastating disease with poor prognosis.1, 2 Early seizures (ESs) are a frequent symptom and complication in patients suffering from sICH that may be associated with outcome.3, 4 Most available data concerning this topic are derived from cohorts of patients with ischemic stroke.5-10 These data show the frequency of seizures in stroke to be as high as 14%. Some studies have focused on the occurrence of seizures after subarachnoid hemorrhage.11, 12 However, there is a lack of data on patients with nontraumatic sICH, and for this reason, its frequency and predictors are poorly investigated.13 The reported incidences range from 4% to 17%, depending on the diagnostic criteria, study design, and duration of follow-up.4, 12, 14-17 The occurrence of ESs might be a prelude to recurrent seizures but does not necessarily mean that it will develop into epilepsy. Until now, there has been no clear evidence-based guidelines for the treatment of ESs in patients with sICH, which may be related to the fact that data on the preventive use of anticonvulsants are controversial.18-21 The objectives of the present study are to determine the frequency of ESs and to investigate the association of ESs with in-hospital mortality and functional outcome in a monocenter study.
Methods
Study design
During a 5-year period (2009–2013), 484 consecutive patients (mean age, 72.3 ± 12.6 years; female sex, 51%) suffering from sICH who were admitted to the Department of Neurology at the University of Lübeck, Germany were enrolled and prospectively evaluated. Based on International League Against Epilepsy (ILAE) guidelines ESs were defined as seizures occurring within first 7 days of sICH onset.22
We used standardized radiological analysis in characterizing sICH, data from computed tomography (CT) scans and MRI scan were used.
Seizure recording was based on clinical diagnosis. In accordance with the International League Against Epilepsy (ILAE) classification, seizures were defined in this study as a paroxysmal disorder of the central nervous system with or without loss of consciousness or awareness and with or without motor involvement.23 Patients with a history of previous seizures or a preexisting diagnosis of epilepsy were also included in our study. Electroencephalography (EEG) records and findings were analyzed. According to the International Federation of Clinical Neurophysiology glossary of terms most commonly used by clinical electroencephalographers and proposal for the report form for the EEG findings a focus is described as circumscribed area on the surface of head, brain or deeper cortical structures, that shows a selected normal or pathological EEG-pattern; epileptiform patterns (EP) are transients of sharp characteristic, that are distinguished from background activity and can be found interictally in particular but not exclusively in patients with epilepsy; slow activity is characterized by slow, mostly polymorphic waves with a frequency within the delta or theta band, that are not explained by drowsiness, responding to external triggers and exceed dimensions common for the age.24
Baseline and sociodemographic characteristics such as gender, age, comorbid conditions, neurological deficits (quantified by the NIHSS) at admission, incident complications, clinical findings, radiological data were identified from clinical records and the hospital information system (Table 1). Outcome measures were assessed with the modified Rankin Scale (mRS; assessment score for disability and functional status in stroke, ranging from 0 [no symptoms] to 6 points [dead]). Mortality was evaluated at discharge from hospital. All patients included in the study were admitted to the stroke unit or the intensive care unit of the Department of Neurology at the University of Lübeck and were treated by stroke neurologists. The data acquisition was part of the ongoing prospective Stroke Registry at the Department of Neurology. Approval for the Stroke Registry was obtained from the Ethics Committee of the University of Lübeck.
Statistics
We used the SPSS software version 22.0.0.1 (IBM) to analyze the data. The data were described with mean and SD values for continuous variables, median and interquartile range (IQR) values for scores, and absolute numbers and percentages for nominal and categorical variables. We performed a chi-square test to determine the correlation between categorical variables, a t test between continuous variables, and a Mann-Whitney test between scores. The logistic regression was carried out to estimate the odds ratios (ORs). All variables with a P value < 0.1 were entered into the logistic regression model. A P value less than 0.05 was considered significant.
Results
A total of 484 patients (249 female patients) with nontraumatic sICH met the inclusion criteria. Of these, 52 patients (10.7%; 95% CI, 8-14) experienced ESs, 30 patients with simple and complex focal and 17 with generalized and 3 with secondarily generalized seizures. Two patients were not classified due to insufficient data.
The mean duration of hospitalization was 12 days, but patients who experienced ESs stayed longer in hospital than those who did not (14 vs 11 days, respectively, P = 0.001). Patients who suffered from ESs were also less affected as assessed with the NIHSS at admission than those who did not suffer from ESs (7 vs 10, respectively; P = 0.02). Differences in baseline characteristics (age, gender, medical history of arterial hypertension, diabetes mellitus, hyperlipidemia, atrial fibrillation, and prior medication) between patients with ESs and those without ES are shown in Table 1. There were nonsignificant differences involving patients with impaired consciousness at admission: 2 patients in the group of patients who suffered from ESs and 27 in the group of patients who did not suffer from ESs.
The total number of patients with preexisting diagnosis of epilepsy was likewise small (n=21) as the number of patients who were pretreated with AEDs (anti-epileptic drugs) by the time of admission (n=27) and contained of various dosages of Lamotrigine 50-400mg/d; Valproic acid 1.500-2.400mg/d; Carbamazepine 400mg/d; Levetiracetam 250-2.000mg/d; Pregabalin; Clobazam 10mg/d and Phenobarbital 60mg/d alone or in combination.
Six patients were pretreated with AEDs due to other non-anticonvulsant reasons.
The occurrence of ESs in patients with preexisting epilepsy tends to be higher than in those without preexisting epilepsy (9.6% vs. 3.9%; p=0.06).
Three of five patients who were withdrawn from pre-medication with AEDs, died in hospital. 42 of the 52 patients (80.7%) who experienced ESs were treated with AEDs by the time of discharge from hospital. The most frequently prescribed AEDs in these cases were Levetiracetam in a daily dosage range from 1.000 to 2.000mg, followed by Valproic acid, Lamotrigine and Pregabalin. 37 of the 52 ESs patients (71.2%) were initially started on AEDs during hospitalization.
In 180 patients (37.2%) a routine EEG was performed (88% of patients with ESs; n=44; versus 34% of patients without ESs; n=136).
EEG showed in 23% of patients normal findings, in 36% a focal slowing and in 38% slow activity (alone, combined with focal slowing or combined with epileptic potentials). Epileptiform patterns were found among 8% of all the EEG records (table 2).
The localization of the sICH revealed a significant difference (P < 0.001) between the two groups, with much more basal ganglia (n = 144; 35%) and thalamus (n = 45; 11%) bleedings detected in the group of patients without ESs. Other localizations such as those of the pons, cerebellum, ventricle, and hypophysis, as well as multifocal localizations, were similar between the two groups (table 1). The results also showed that 7% of all patients with ventricle extension suffered from ESs. In addition, an association was found between the localization of cortical/lobar hemispheric hemorrhage and the occurrence of ESs (80.4% vs 40.5%, respectively; P = 0.008). The incidence of ESs was 20% in patients with cortical/lobar hemispheric sICH. The multivariate analysis revealed an increase in the rate of ESs in these patients (OR, 3.5; 95% CI, 1.8-6.7; P < 0.001). The incidence rate of ESs in all other cases with non-lobar bleeding localizations was about 4%.
During hospitalization, 109 patients (23%) died, 44% of those within the first 2 days and 69% within the first week (Figure 1). The in-hospital mortality rate was lower in patients who suffered from ESs than in those who did not (9.6% vs 24.0%, respectively; P = 0.02; table 1). An association between ESs and good functional outcome on the mRS (≤2) at the time of discharge of patients from hospital was not found (P = 0.3).
Discussion
We found that ESs occurred in 10.7% of patients with sICH and appeared to be associated with cortical hemispheric localized hemorrhage. In addition, we found that the occurrence of ESs was associated with a lower rate of in-hospital mortality.
In our study, we defined ESs according to the International League Against Epilepsy (ILAE) guidelines as seizures occurring within first 7 days of sICH onset.22 Other study groups have used a 7- or 14- day period after sICH/stroke onset in defining ESs.4, 9 Ischemic stroke and sICH are two different diseases with varying prognoses. Upon closer examination of the in-hospital mortality in our study, we found that approximately two thirds of our patients who did not survive the duration of hospitalization died within the first week. These patients died before the end of the 7-day period stated in the ILAE guidelines for ESs and, in this way, might have dropped out from surveillance before ESs could occur by definition. This high mortality rate within the first week of hospitalization might be an explanation for the small number of patients with ESs at all, particularly among the group of patients who died in hospital. However, it is important to note that the patients who died in hospital were much more severely neurologically impaired at admission to hospital than those who survived.
The incidence rate of ESs in patients with sICH was 10.7%. It is difficult to compare the incidence rates of ESs reported in the literature owing to differences in the methods used, including variables like inclusion criteria, previous seizures, or time window in defining ESs. For example, several studies with a short follow-up period of 7 days reported incidence rates of ESs in patients following sICH that ranged between 1% and 12%.25-27 Other studies with short follow-up periods of 24 hours up to 14 days revealed incidence rates of ESs following sICH of 4% to 8%.6-8, 15, 28 In other prospective studies with longer follow-up periods of 7 days up to 6 months, incidence rates of ESs ranged from 14% to16%.4, 9 Our findings are comparable with those in the literature because they are within the range of previously reported incidence rates.
We found lower rates of in-hospital mortality in patients who suffered from early seizures. One explanation might be that patients with ES were less severely affected neurologically. Another reason is that the occurrence of alteration/loss of consciousness in patients with ES was lower than those without ES. The alteration of consciousness might be linked to severe hemorrhage with non-motor seizures.
These reason may led to survival of the hemorrhage. On the other hand, the functional outcome at discharge did not show a difference between patients with versus without ES.
In our study we used standardized radiological analysis in characterizing sICH. We did not evaluate cortical brain microbleeds, that might serve as a predictor of ischemic brain lesions and may also represent an imaging marker of active vasculopathy29, or other potential prognostic values of neuroradiological biomarkers seen in susceptibility weighted imaging (SWI) sequences in magnetic resonance imaging (MRI). In a recent review studies of the spot sign appeared diverse and therefore complex to interpret due to heterogeneity and potential of bias in the selected studies.30 However, Havsteen et al. found an association between the absence or presence of a spot sign and long-term mortality and functional outcome in sICH patients.31
A recent review described the potential role and limitations of blood biomarkers like GFAP, S100B/RAGE, ApoC-III, ß-Amyloid, MMP-9, BNP and others in the management of nontraumatic ICH but larger prospective studies are needed to validate findings for implementation into clinical routine.32
One difference between our study and previous studies with small cohorts is that we investigated clinical seizures clinically. By contrast, previous studies with small cohorts used continuous EEG monitoring, revealing incidence rates of electric seizures in patients with sICH ranging from 28% to 42%.33, 34
Factors associated with ESs after nontraumatic sICH have been identified in only a few studies. We found in the present the occurrence of ESs seems to be correlated with cortical hemispheric hemorrhage localization. This finding is similar to those of other previous investigations.4, 7, 25 The lobar localization of a hemorrhage has been described as a predictor of ESs following ischemic stroke and sICH owing to potential cortical involvement in comparison to deep sICH.9, 26, 35 Previous research has suggested that direct cortex irritation is the mechanism by which cortical hemispheric sICH provokes seizures.36 In the present study, bleeding localization in the basal ganglia and thalamus was associated with a decrease in the occurrence of ESs. In these cases, direct cortex stimulation is lacking.