Non convulsive status epilepticus after
J Neurol
DOI 10.1007/s00415-014-7471-z
ORIGINAL COMMUNICATION
Non-convulsive status epilepticus after ischemic stroke: a hospitalbased stroke cohort study
Vincenzo Belcastro • Simone Vidale •
Gaetano Gorgone • Laura Rosa Pisani •
Luigi Sironi • Marco Arnaboldi • Francesco Pisani
Received: 10 May 2014 / Revised: 19 July 2014 / Accepted: 11 August 2014
Ó Springer-Verlag Berlin Heidelberg 2014
Abstract To evaluate in the setting of a stroke unit ward
the usefulness of a prolonged ([6 h) video-EEG recording
(PVEEG) in identifying non-convulsive status epilepticus
(NCSE) in patients with an acute ischemic stroke. Predictors of NCSE were also evaluated. Patients with an acute
ischemic stroke, referred to our unit, were included in this
prospective observational study. A PVEEG recording was
implemented after stroke in all patients during the first
week: (a) promptly in those exhibiting a clear or suspected
epileptic manifestation; (b) at any time during the routine
activity in the remaining patients. After the first week, a
standard EEG/PVEEG recording was hooked up only in
presence of an evident or suspected epileptic manifestation
or as control of a previous epileptic episode. NCSE was
identified in 32 of the 889 patients (3.6 %) included in the
study. It occurred early (within the first week) in 20/32
(62.5 %) patients and late in the remaining 12. Diagnosis
was made on the basis of a specific clinical suspect
(n = 19, 59.4 %) or without any suspect (n = 13, 40.6 %).
In a multivariate analysis, a significant association of
NCSE was observed with NIHSS score, infarct size and
large atherothrombotic etiology. NCSE is not a rare event
after an acute ischemic stroke and a delayed diagnosis
V. Belcastro (&) S. Vidale L. Sironi M. Arnaboldi
Neurology Unit, S. Anna Hospital, Como, Italy
e-mail: vincenzobelcastro@libero.it
G. Gorgone
Neurology Unit, Treviglio-Caravaggio Hospital, Treviglio, Italy
L. R. Pisani
IRCCS Centro Neurolesi ‘‘Bonino-Pulejo’’, Messina, Italy
F. Pisani
Department of Neurosciences, University of Messina, Messina,
Italy
could worsen patient prognosis. Since NCSE can be difficult to be diagnosed only on clinical grounds, implementation of a prompt PVEEG should be kept available in a
stroke unit whenever a patient develop signs, although
subtle, consistent with NCSE.
Keywords Non-convulsive status epilepticus
Post-stroke seizures Prolonged video-EEG recording
Introduction
Stroke is the most common cause of symptomatic epilepsy
in older adults and status epilepticus (SE) can be a presenting clinical feature of an acute stroke [1–6]. Values
concerning the incidence of post-stroke seizures and
development of epilepsy after stroke show a great variability in the literature. This variability is due to different
reasons: small sample sizes, different study designs, heterogeneous terminology, different periods of follow-up, no
specification on the ischemic/hemorrhagic nature of the
lesion. Incidence values of ‘‘early’’ (i.e., within the first
7 days) seizures range from 3 to 15 % [1, 7–11], but values
[30 % have been also reported in the literature [5]. ‘‘Late’’
seizures, i.e., those after the first week, have been reported
to occur in a percentage ranging from 3 to 9 % with a
gradual increase by increasing the time period after stroke,
the higher values being observed at 10 years [1, 4, 5].
Similarly, post-stroke epilepsy has been seen to develop in
6.4 % of patients as a mean value, ranging from 1.5 % at
3 months to 12.4 % at 10 years after stroke [2]. Concerning the incidence of SE, it has been estimated to range from
0.2 to 1.4 % [9–14] with very small differences observed
between ischemic and hemorrhagic stroke (0.2 vs. 0.3 %)
[14]. Additionally, most of these data regard convulsive
123
J Neurol
generalized SE (CGSE) and only few observations have
been specifically reported on non-convulsive SE (NCSE)
after stroke [6, 15]. In particular, in comatose patients with
NCSE, 20 % of patients had an ischemic stroke [15].
NCSE has been defined as a state of ongoing (or nonrecovery between) seizures without convulsions, usually
for more than 30 min [16]. It is a clinical entity very difficult to be detected in acute and critically ill patients,
including those with stroke, because of its subtle, nonconvulsive manifestations associated with transient consciousness disturbances [16–20].
Using a hospital-based stroke cohort, the first aim of the
present study was to evaluate in the setting of a stroke unit
ward the usefulness of a prolonged, namely lasting for at
least 6 h, video-EEG recording (PVEEG) in identifying
episodes of NCSE after an acute ischemic stroke. Predictors of NCSE were also evaluated.
Patients and methods
The Institutional Review Board of the S. Anna Hospital,
Como, Italy, approved the study and granted a waiver of
informed consent. The present investigation is a prospective observational study.
Patient population
Patients, referred to the Stroke Unit of S. Anna Hospital of
Como between January 2010 and September 2013, were
included in this study according to the following inclusion
criteria: (1) acute ischemic stroke defined by the occurrence of acute neurologic signs and symptoms of vascular
origin lasting 24 h or longer and documented through brain
CT/MRI; (2) no other concomitant causes potentially
responsible of acute seizures (for example: alcohol withdrawal, psychotropic drugs, electrolyte disturbance, previous brain lesions). A direct interview was required, except
for aphasic or uncooperative patients, for whom the interview was released by his relatives.
Patients with: (1) a previous history of seizures; (2)
recurrent stroke; (3) i.v. or p.o. treatment with antiepileptic
(AEDs) or sedative drugs, including barbiturates, benzodiazepines and propofol, prior to video-EEG execution; (4)
hemorrhagic transformation of the ischemic area were
excluded from analysis.
EEG recording and epilepsy characterization
Given that [50 % of epileptic seizures/SE occur in the
first days after stroke [4, 5, 7–10, 13], a PVEEG
recording, including ECG tracing, was implemented in all
patients admitted to our Stroke Unit within the first week
123
after stroke onset: it was hooked up as soon as an epileptic activity was clear or suspected or at any time
during the routine activity in the remaining patients.
Additional conventional (i.e., of *30 min) EEG recordings, or even PVEEG when clinically judged more
appropriate, were made after the first week only in presence of clear or suspected epileptic seizures/SE or as
control of a previous epileptic episode. EEG was recorded digitally using caps with 21 fixed gel electrodes
placed according to the International 10–20 system. EEGs
were
immediately
evaluated
by
board-certified
electroencephalographers.
Seizures were divided into convulsive and non-convulsive. Convulsive seizures (CSz) were described as
‘‘tonic–clonic,’’ ‘‘clonic’’ or ‘‘tonic’’ (including also synonyms like ‘‘jerking’’ or ‘‘twitching’’). SE was reported
as convulsive (CSE) for CSz lasting longer than 5 min or
if 2 or more fits occurred without a return to baseline in
between. Non-convulsive seizures (NCSz) were considered those characterized by subtle movements like slow
facial twitching and eye deviation. NCSE was defined as
30 min of continuous seizure activity without major
motor phenomena and according to the criteria of Young
et al. [16]. At least one of the following three primary
EEG criteria and at least one of the four secondary EEG
criteria had to be fulfilled. Primary criteria were (1)
repetitive focal or generalized spikes, sharp waves, spikeand-wave or sharp-and-slow wave complexes at a frequency of [3/s; (2) repetitive focal or generalized spikes,
sharp waves, spike-and-wave or sharp-and-slow wave
complexes at a frequency of [3/s and secondary significant improvement in clinical state or baseline EEG after
administration of AEDs; or (3) sequential rhythmic waves
and secondary criteria (1), (2), (3) with or without significant improvement in clinical state or baseline EEG
after administration of AEDs. Secondary EEG criteria
were the following discharge patterns with a duration of
[10 s: (1) incrementing onset with increase in voltage
and/or increase or slowing of frequency; (2) decrementing
offset with decrease in voltage or frequency; (3) postdischarge slowing or voltage attenuation; or (4) significant improvement in clinical state or baseline EEG after
administration of AEDs.
Patients with EEG showing only bilateral periodic discharges or stimulus-induced rhythmic, periodic or ictal
discharges (SIRPIDs) were excluded. The presence of
periodic lateralized epileptiform discharges (PLEDs), an
EEG pattern usually observed in patients with acute stroke,
was not the reason for exclusion.
NCSE was coded into two categories: ‘‘early’’ referred
to that occurring within a week after stroke and considered
to be provoked by the acute brain insult; ‘‘late’’ that
developing after the first week [10]. Video recordings were
J Neurol
screened to determine clinical correlates for NCSE and
episodes of electroencephalographic seizures. Level of
consciousness, evaluated before pharmacological treatment
and upon EEG initiation, was categorized on clinical
grounds as alert/somnolent/stuporous/comatose. Each
patient was monitored through video-EEG recording for at
least 6 h following SE resolution, and a further conventional EEG recording was made after 24 h.
Antiepileptic drug treatment
When a true diagnosis of SE was made, pharmacological
treatment was promptly started and its effect evaluated
through both EEG and clinical monitoring. As a standard
therapeutic approach, i.v. administered drugs were: lorazepam (0.1 mg/kg within 1–2 min), diazepam (10 mg,
5 mg/min) or phenytoin (PHT) (18 mg/kg, 5 mg/kg/min).
According to the recent literature suggestions [21], levetiracetam (LEV) 25 mg/kg over 15 min or lacosamide
(LCM) as initial bolus dose of 400 mg over 30 min was
used in patients with concomitant medical conditions in
whom the use of traditional AEDs was judged potentially
unsafe because of their adverse effects (hypotension, QT
prolongation, arrhythmias, respiratory complications).
During i.v. treatment, patients were monitored through
video-EEG, ECG, heart rate and blood pressure measurements. SE was considered fully controlled only after
both clinical and electrographic resolution.
Stroke characterization
Based on history and the results of diagnostic studies
including brain imaging, Doppler sonography, echocardiograms, and electrocardiograms, patients were classified
according to stroke syndrome, etiology and severity.
Clinical syndrome classification was assessed using the
Oxfordshire Community Stroke Project (OCSP) criteria
[22]. Stroke etiology was classified using the TOAST
(Trial of ORG 10172 in Acute Stroke Treatment) criteria
[23]. Stroke severity was assessed at admission using the
National Institutes of Health (NIH) stroke scale (NIHSS)
score [24]. Neuroimaging data (CT or MRI) were evaluated
according to the arterial territory circulation involved
(anterior or posterior) and lesion size (large, medium,
small, or lacunar). A large lesion was required to have a
diameter [3 cm on neuroimages; a small lesion a diameter
\1 cm; a lacunar lesion was a CT hypodense area or T2
MRI hyperintense area \1 cm in maximum diameter deeply located in the cerebral hemispheres or in the brainstem.
Finally, functional disability was obtained using the modified Rankin scale (mRS) [25] and mortality was assessed
during hospitalization.
Statistical analysis
Distribution of gender, vascular risk factors, functional
disability, OCSP and TOAST criteria were summarized as
frequencies and percentages and comparisons were made
using the Pearson Chi-square test or Fisher exact test, as
appropriate. Age and NIHSS score were indicated as
median and interquartile range, and were investigated using
the Mann–Whitney U test. Univariate and multivariate
logistic regression models were used to evaluate the following potential determinants of NCSE: stroke syndrome,
stroke etiology, size of infarction, NIHSS score, age,
gender, vascular risk factors. In the analysis, the size of
infarction was expressed as a three-level ordinal variable,
according to the classification above mentioned. Multivariate models were constructed using all variables resulted
significant in the univariate analysis. Measures of association were odds ratios (ORs) with 95 % confidence intervals. Each covariate was tested independently and with the
main interaction terms. Finally, the ROC analysis was
performed to check the goodness of fit of the implemented
logistic regression final model. Statistical significance was
chosen at the 5 % level. All statistics were implemented
using STATA version 12 (StataCorp, USA).
Results
Cohort characteristics
A total of 1,362 patients were admitted to our Stroke Unit
between January 2010 and September 2013: 1,034 (76 %)
have an ischemic stroke; 227 (16.6 %) have an intracerebral hemorrhage (ICH); 101 (7.4 %) have a subarachnoid hemorrhage. Of the 1,034 patients with ischemic
stroke, 72 (7 %) were excluded from the analysis because
of a hemorrhagic transformation of the ischemic area.
Moreover, 73 patients were also excluded from the analysis
because of a previous history of seizures, sedative medication prior to PVEEG recording, a history of recurrent
stroke or EEG recordings showing only bilateral periodic
discharges. The final cohort consisted of 889 patients. The
characteristics of the sample are summarized in Table 1.
Seizures and status epilepticus
Overall 59/889 (6.6 %) patients exhibited epileptic manifestations. Early CSz and CSE occurred in 41/889 (4.6 %)
patients: seizures were tonic–clonic in 14, clonic without
impairment of consciousness in 13, and clonic with
automatisms in 6; focal motor SE occurred in four patients
and CSE in the remaining four. Sixteen out of these 41
patients also developed late CSzs (n = 12) or focal motor
123
J Neurol
Table 1 Demographic and
clinical characteristics of the
screened population
Variable
NCSE (n = 32)
No NCSE (n = 857)
Overall (n = 889)
66.5 (61.0–78)
16 (50)
65.0 (58.0–77.0)
550 (64.1)
65.0 (58.0–77.0)
566 (63.6)
Demographic data
Age (years)
Male
Stroke severity indicators
NIHSS
mRS [2
13.0 (9.0–15.0)*
3.0 (2.0–7.0)
4.0 (2.0–8.0)
12 (37.5)
343 (40.0)
355 (39.5)
Large size infarct
B1 cm
2 (6.3)*
447 (52.2)
449 (50.5)
[1 cm
9 (28.1)*
284 (33.1)
293 (33.0)
[3 cm
21 (65.7)*
126 (14.7)
147 (16.5)
Vascular risk factors
Discrete variables (sex, large
size infarct, vascular risk
factors, OCSP and TOAST
ischemic stroke subtypes) are
indicated as number of cases
(percentage). Age is indicated
as median (interquartile range)
and NIHSS score as geometric
mean (standard error)
mRS Modified Rankin Scale,
NCSE non-convulsive status
epilepticus, NIHSS NIH Stroke
Scale, PACI partial anterior
circulation infarct, POCI
posterior circulation infarct,
TACI total anterior circulation
infarct, TIA transient ischemic
attack
* p \ 0.0001, p = 0.0001 and
à
p = 0.001 in comparison to
No NCSE patients
Hypertension
Diabetes
430 (50.1)
446 (50.1)
9 (28.1)
186 (21.7)
195 (21.7)
Cardiac arrhythmia
5 (15.6)
188 (15.6)
192 (21.4)
Smoking
2 (6.2)
310 (36.2)
312 (35.0)
Hyperlipidemia
8 (25)
276 (32.2)
284 (31.5)
TIA
5 (15.6)
120 (14.0)
125 (13.9)
15 (46.8)à
123 (14.3)
138 (15.5)
11 (34.3)
126 (14.7)
137 (15.4)
385 (41.5)
385 (45.0)
268 (31.2)
274 (30.8)
OCSP classification
TACI
PACI
LACI
POCI
0
6 (18.7)
TOAST classification
Large artery atherothrombosis
13 (40.6)
69 (8.0)
82 (9.2)
Cardioembolic
10 (31.2)
227 (26.5)
237 (26.6)
Lacunar
315 (36.7)
315 (35.4)
Cryptogenetic
7 (21.9)
214 (25.0)
221 (24.8)
Other
2 (6.2)
58 (6.8)
60 (6.8)
SE (n = 4) during the period of hospitalization. Two
patients exhibited for the first time tonic–clonic Sz during
the third week after stroke. Complex partial seizures,
although suspected in eight additional patients on the basis
of apparently not oriented motor manifestations and/or
level of consciousness oscillations, were not firmly demonstrated through EEG recording. Similarly, electrical
seizures not associated to clinical manifestations were not
intercepted in our population of patients during hospitalization. In our cohort, NCSE was identified in 32/889
(3.6 %) patients: it occurred early in 20/32 (i.e., 62.5 %)
and late in the remaining 12 patients. Early NCSE developed as continuation of a CSE in five patients and during
the post-ictal phase of a tonic–clonic seizure in two
patients. Late NCSE was diagnosed in 12 patients: during
the second week in eight and during the third week in the
remaining four. Seven of these 12 patients had already had
early tonic–clonic seizures and two early focal motor SE;
the remaining three patients developed NCSE without any
previously identified epileptic activity. Impairment of
123
16 (50.0)
0
consciousness and agitation were the main manifestations
in 22/32 (68.8 %) patients and aphasia in the remaining 10
(31.2 %). Noteworthy, NCSE was diagnosed on the basis
of a specific clinical suspect in 19 patients (59.4 %) and
without any clinical suspect in the remaining 13 patients
(40.6 %). In 6 out of these 13 patients, it was intercepted in
the first week during PVEEG recording implemented as
part of the routine activity, and in the other seven patients
during PVEEG made to monitor the time course of a tonic–
clonic seizure (1 patient in the first week), a CSE (3
patients in the first week), or made just as a control in 3
patients (1 in the second week and 2 in the third week) who
had had previous CSzs.
EEG features
PVEEG was recorded for a mean duration of 9 h 22 min,
range 6 h 15 min–13 h 27 min.
PVEEG recording was implemented within the first
3 days in the majority of patients, namely 605, and
J Neurol
between the forth and the seventh day in the remaining 284
patients.
Ictal EEG features showed: (a) high-voltage, theta
rhythmic activity with intermingled spikes over the temporo-occipital regions in eight patients; (b) high-voltage
theta activity intermingled with sharp waves over frontal
regions in seven patients; (c) bilateral, over frontal regions,
continuous spike- and slow-wave discharges in six patients.
Continuous focal sharp waves with change in amplitude,
frequency and spatial distribution were observed in 11
patients.
Treatment and epilepsy outcome
Initial treatment of NCSE was efficacious in 21 patients in
a time ranging from 30 to 60 min after i.v. administration:
lorazepam (n = 4), diazepam (n = 9), LCM (n = 5) and
PHT (n = 3). Eight patients did not show any response to
initial diazepam treatment and were treated with PHT
(n = 6) or LEV (n = 2) introduced as second-line treatment. None of these patients relapse. In three patients,
NCSE was refractory to diazepam followed by PHT and
LCM and a transfer to the intensive care unit was needed.
A chronic treatment with AEDs was started in all SE
patients. The following drugs were given to patients with
NCSE: LEV in 24 patients, oxcarbazepine in six and
phenobarbital in the remaining two patients. The present
follow-up for 28 of these patients ranges from 2 to
43 months. In the control visits, usually planned every
6–8 weeks, a global neurological evaluation, seizure frequency recordings through ad hoc calendars, therapy
adjustments and standard EEG execution have been performed. Of these patients, 25 developed post-stroke epilepsy with seizure frequency of 3–8 fits per month (18
patients) and \1 per month (7 patients). Three patients are
seizure free (present follow-up 5, 13 and 16 months).
Detailed data concerning long-term outcome of both epilepsy and stroke in our population will be object of a
forthcoming article. No information has been obtained on
the remaining four patients after discharge from hospital
and until the preparation of the present manuscript.
Stroke features and outcome
Of the 32 patients with NCSE, a total anterior circulation
infarct (TACI) was diagnosed in 15, a partial anterior circulation infarct (PACI) in 11 and a posterior circulation
infarct (POCI) in 6 patients. Three patients with early
NCSE received intravenous recombinant tissue plasminogen activator (t-PA). None of the patients developing
NCSE died during the period of hospitalization, and all
were discharged to rehabilitation or to home within
14–30 days after stroke onset. No differences in outcome
Table 2 Independent predictors of NCSE in patients with ischemic
stroke
SE
z
p
OR
95 % CI
TACI
0.24
-1.46
0.1
0.34
0.12
1.03
Large artery
atherothrombosis
1.57
3.07
0.002
3.7
1.60
6.49
Infarct size
NIHSS score
1.56
0.05
2.92
2.72
0.003
0.007
3.6
1.14
1.52
1.04
8.44
1.24
Variables not in equations: gender, age, hypertension, diabetes mellitus, cardiac arrhythmia, smoking, hyperlipidemia, TIA, PACI,
POCI, other TOAST classification aetiologies
CI confidence interval, OR odds ratio, SE standard error, NIHSS NIH
Stroke Scale
of stroke between NCSE patients and the other patients
were detected through the NIHH score and other functional
assessment scales used (Barthel index, modified Rankin
Scale). As above mentioned, detailed data on the long-term
outcome of our population will be illustrated in a forthcoming study.
Statistical significance
Univariate analysis showed significant differences between
patients with NCSE and those without NCSE with regard
to etiology, clinical presentation and size of ischemic area
(Table 1). Regarding stroke etiology, large artery atherothrombosis showed a different distribution between NCSE
and non-NCSE groups (Table 1). In the multiple logistic
regression models, we considered as predictors of the
NCSE those variables resulting significant from the univariate analysis (Table 1). A significant association of
NCSE was observed with NIHSS score, infarct size and
large atherothrombotic etiology (Table 2). No significant
differences between the expected and the observed
regression matrix were observed (area under ROC
curve = 0.87).
Discussion
The results of the present study indicate that NCSE is not a
rare event after an acute ischemic stroke and occurs in a
percentage of 3–4 % of patients. In our cohort, significant
predictors of NCSE were a large size infarct, large artery
atherothrombosis and NIHSS score at admission.
With regard to infarct area, it has been proposed that the
ischemic penumbra of a stroke can contain electrically
irritable tissue that provides a focus for seizure activity
[26]. Enhanced release of glutamate, ionic imbalances
associated to breakdown of membrane phospholipids,
release of free fatty acids, reduced GABAergic function
123
J Neurol
with functional and structural impairment of GABAergic
interneurons have been described to occur in the ischemic
area after an acute insult [27, 28]. Cortical location and
large size infarct have been constantly found to be associated with post-stroke seizures [2, 9, 11], a finding that has
emerged also from the present study. On the basis of the
cascade of synaptic and intracellular events shared by
seizure and vascular brain injuries, it can be hypothesized
that post-stroke NCSE could be triggered by acute cellular
biochemical disturbances occurring with acute ischemic
stroke.
In the present study, large artery atherothrombosis was
another significant predictor of NCSE.
Large artery atherothrombosis is more common in cortical than lacunar infarcts, implying that an embolic etiology would be more common in TACIs and PACIs than in
LACIs [29]. Notably, in our NCSE group, TACI was
diagnosed in 15/32 (46.8 %) and PACI in 11/32 (34.3 %)
patients while lacunar infarct was completely absent. These
observations reinforce the crucial role of cortical involvement in seizure genesis.
A clinically relevant aspect is the possible impact of
NCSE on outcome of patients with ischemic stroke. Previous studies have shown that NCSE is a risk factor for the
development of refractory SE, poor outcome, and increased
mortality independently of the underlying pathology [16,
30]. In one of these studies, for example, the mortality rate
in an intensive care unit was high: more than half of the
patients with NCSE, 13/23 (57 %) died [16]. Specific data
on NCSE after an ischemic stroke are difficult to be
extrapolated from published studies because of a number of
reasons, including no reported distinction between CSE
and NCSE in the analysis, combination of both patients
with ischemic stroke and ICH, focus mainly on EEG
activity, generic report on patients with acute insults as
seen in medical intensive care units [8, 16–18, 20, 31]. In
our cohort, cases of death were not observed during hospitalization and until the present follow-up. Additionally,
NCSE patients did not show, as compared to patients
without NCSE, differences concerning outcome and complications until the present follow-up.
The NIHSS score was a significant predictor of NCSE.
Interpretation of NIHSS score is difficult since the impairment of consciousness, which is a hallmark of NCSE, could
have increased itself NIHSS score at admission. In fact,
early NCSE occurred in 62.5 % of our 32 NCSE patients,
this value being in agreement with that reported in critically
ill patients, in whom seizures activity was observed in the
first 24 h after the insult in a large majority of patients
(88 %) [20]. Additionally, as above mentioned, SE has been
seen to be a possible presenting clinical feature of an acute
stroke [1–3]. On this basis, it can be hypothesized that a
percentage of our patients might have at admittance a
123
condition of NCSE which has not been detected. In patients
with ischemic stroke, values concerning the incidence of SE
range between 1 and 10 % [20] and NCSE has been
reported to occur in 30 % of patients developing post-stroke
SE [13]. In our cohort, 5 out of 20 patients developed early
NCSE as continuation of CSE and in two patients it
occurred in the post-ictal phase of a tonic–clonic seizure.
These values are very close to those reported in other
studies [17, 32] and emphasize the importance to have the
possibility in a stroke unit to implement a VEEG recording
longer than a conventional one, which lasts about
20–30 min on average. In a population of 110 patients, the
first seizures were detected in slightly more than half of the
cases within the first hour of recording [33].
The second relevant finding deriving from the present
study, concerning EEG, is that the diagnosis of NCSE has
been made during a routine EEG recording and without any
clinical suspect or any previous epileptic manifestation in a
considerable proportion of patients. This result, in line with
a large body of literature observations [18, 31, 32], provides further support in favor of the view that the incidence
of NCSE after an ischemic stroke is underappreciated and
that this pathology is more frequent than thought in the
past. As above mentioned, in fact, NCSE is a clinical entity
which goes easily unnoticed especially in elderly patients
and in patients with acute brain disorders in whom
impairment of consciousness is frequent as a consequence
of various pathologies [32–36]. In particular, in patients
with an acute ischemic stroke diagnosis of NCSE might be
made difficult because of various factors: (1) NCSE occurs
more frequently in the first days after stroke onset when
impairment of consciousness is a common expression of
stroke itself, (2) NCSE can occur just after CSz/CSE and
might be easily interpreted as a post-ictal state, (3) manifestations of NCSE can be so subtle as no to raise the
suspect of an epileptic activity.
Our study has different limitations: (1) the single-institution approach makes the study vulnerable to bias; (2) the
relatively small sample size of NCSE group does not
necessarily represent all the characteristics of patients
suffering from acute stroke and developing NCSE; (3)
cases of NCSE before and after PVEEG implementation
could have been easily missed. The latter limit, in particular, has been also emphasized by various authors [36, 37].
Conclusions
This population-based study shows that implementation of
a PVEGG should be permanently available in a stroke
unit. In this setting, a conventional EEG recording is of
limited diagnostic value because of the intermittent nature
of subtle, difficult-to-detect seizures and PVEEG
J Neurol
recording increases the possibility to diagnose NCSE. An
increased awareness that NCSE is a possible, not rare
event after an ischemic stroke and that it can most frequently occur with subtle signs can contribute to speed a
diagnosis of NCSE. A quick diagnosis of this condition is
crucial to implement a prompt treatment and hence to
avoid possible clinical complications in patients with
ischemic stroke [33, 37].
17.
18.
19.
Acknowledgments We would like to acknowledge the contribution
of Neurology EEG technologists of S. Anna Hospital in connecting
these patients to the PVEEG at all odd hours.
20.
Conflicts of interest
21.
None.
22.
References
23.
1. So EL, Annegers JF, Hauser WA, O’Brien PC, Whisnant JP
(1996) Population-based study of seizure disorders after cerebral
infarction. Neurology 46:350–355
2. Graham NS, Crichton S, Koutroumanidis M, Wolfe CD, Rudd
AG (2013) Incidence and associations of post-stroke epilepsy: the
prospective South London stroke register. Stroke 44:605–611
3. Ryvlin P, Montavont A, Nighoghossian N (2006) Optimizing
therapy of seizures in stroke patients. Neurology 67:S3–S9
4. Lossius MI, Ronning OM, Slapo GD, Mowinckel P, Gjerstad L
(2005) Post-stroke epilepsy: occurrence and predictors: a longterm prospective controlled study (Akershus Stroke Study).
Epilepsia 46:1246–1251
5. Camilo O, Goldstein LB (2004) Seizures and epilepsy after
ischemic stroke. Stroke 35:1769–1775
6. Afsar N, Kaya D, Aktan S, Aykut-Bingol C (2003) Stroke and
status epilepticus: stroke type, type of status epilepticus, and
prognosis. Seizure 12:23–27
7. Beghi E, D’Alessandro R, Beretta S et al (2011) Incidence and
predictors of acute symptomatic seizures after stroke. Neurology
77:1785–1793
8. Pezzini A, Grassi M, Del Zotto E et al (2013) Complications of
acute stroke and the occurrence of early seizures. Cerebrovasc
Dis 35:444–450
9. Labovitz DL, Hauser WA, Sacco RL (2001) Prevalence and
predictors of early seizure and status epilepticus after first stroke.
Neurology 57:200–206
¨ zmenoglu M, Boz C, Alioglu Z (2001) Status
10. Velioglu SK, O
epilepticus after stroke. Stroke 32:1169–1172
11. Procaccianti G, Zaniboni A, Rondelli F, Crisci M, Sacquengna T
(2012) Seizures in acute stroke: incidence, risk factors and
prognosis. Neuroepidemiology 39:45–50
12. Kilpatrick CJ, Davis SM, Tress B et al (1990) Epileptic seizures
in acute stroke. Arch Neurol 47:157–160
13. Rumbach L, Sablot D, Berger E, Tatu L, Vuillier F, Moulin T
(2000) Status epilepticus in stroke: report on a hospital-based
stroke cohort. Neurology 54:350–354
14. Bateman BT, Claassen J, Willey JZ et al (2007) Convulsive status
epilepticus after ischemic stroke and intracerebral haemorrhage:
frequency, predictors, and impact on outcome in a large administrative dataset. Neurocrit Care 7:187–193
15. Towne AR, Waterhouse EJ, Boggs JG et al (2000) Prevalence of
nonconvulsive status epilepticus in comatose patients. Neurology
25:340–345
16. Young GB, Jordan KG, Doig GS (1996) An assessment of nonconvulsive seizures in the intensive care unit using continuous
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
EEG monitoring: an investigation of variables associated with
mortality. Neurology 47:83–89
De Lorenzo RJ, Waterhouse EJ, Towne AR et al (1998) Persistent
nonconvulsive status epilepticus after the control of convulsive
status epilepticus. Epilepsia 39:833–840
Bottaro FJ, Martinez OA, Pardal MM, Bruetman JE, Reisin RC
(2007) Nonconvulsive status epilepticus in the elderly: a case–
control study. Epilepsia 48:966–972
Ney JP, van der Goes DN, Nuwer MR, Nelson L, Eccher MA
(2013) Continuous and routine EEG in intensive care: utilization
and outcomes, United States 2005–2009. Neurology
81:2002–2008
Sutter R, Stevens RD, Kaplan PW (2013) Continuous electroencephalographic monitoring in critically ill patients: indications,
limitations, and strategies. Crit Care Med 41:1124–1132
Shorvon S (2011) The treatment of status epilepticus. Curr Opin
Neurol 24:165–170
Burn J, Dennis M, Bamford J, Sandercock P, Wade D, Warlow C
(1994) Long-term risk of recurrent stroke after a first-ever stroke.
The Oxfordshire Community Stroke Project. Stroke 25:333–337
Goldstein LB, Jones MR, Matchar D et al (2001) Improving the
reliability of stroke subgroup classification using the Trial of
ORG 10172 in Acute Stroke Treatment (TOAST) criteria. Stroke
32:1091–1098
Brott T, Adams HP Jr, Olinger CP et al (1989) Measurements of
acute cerebral infarction: a clinical examination scale. Stroke
20:864–870
van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn
J (1988) Interobserver agreement for the assessment of handicap
in stroke patients. Stroke 19:604–607
Heiss WD, Huber M, Fink GR et al (1992) Progressive
derangement of periinfarct viable tissue in ischemic stroke.
J Cereb Blood Flow Metab 12:193–203
Luhmann HJ, Mudrick-Donnon LA, Mittmann T, Heinemann U
(1995) Ischemia-induced long-term hyperexcitability in rat neocortex. Eur J Neurosci 7:180–191
Belcastro V, Pierguidi L, Tambasco N (2011) Levetiracetam in
brain ischemia: clinical implications in neuroprotection and prevention of post-stroke epilepsy. Brain Dev 33:289–293
Mead GE, Lewis SC, Wardlaw JM, Dennis MS, Warlow CP
(2000) How well does the Oxfordshire community stroke project
classification predict the site and size of the infarct on brain
imaging? J Neurol Neurosurg Psychiatry 68:558–562
Mayer SA, Claassen J, Lokin J, Mendelsohn F, Dennis LJ,
Fitzsimmons BF (2002) Refractory status epilepticus: frequency,
risk factors, and impact on outcome. Arch Neurol 59:205–210
Claassen J, Mayer SA, Kowalski RG, Emerson RG, Hirsch LJ
(2004) Detection of electrographic seizures with continuous EEG
monitoring in critically ill patients. Neurology 62:1743–1748
Meierkord H, Holtkamp M (2007) Non-convulsive status epilepticus in adults: clinical forms and treatment. Lancet Neurol
6:329–339
Claassen J, Lokin JK, Fitzsimmons BF, Mendelsohn FA, Mayer
SA (2002) Predictors of functional disability and mortality after
status epilepticus. Neurology 58:139–142
Jordan KG (2004) Emergency EEG and continuous EEG monitoring in acute ischemic stroke. J Clin Neurophysiol 21:341–352
Carrera E, Michel P, Despland PA et al (2006) Continuous
assessment of electrical epileptic activity in acute stroke. Neurology 67:99–104
Sutter R, Fuhr P, Grize L, Marsch S, Ru¨egg S (2011) Continuous
video-EEG monitoring increases detection rate of nonconvulsive
status epilepticus in the ICU. Epilepsia 52:453–457
Huang CW, Saposnik G, Fang J, Steven DA, Burneo JG (2014)
Influence of seizures on stroke outcomes: a large multicenter
study. Neurology 82:768–776
123
DOI 10.1007/s00415-014-7471-z
ORIGINAL COMMUNICATION
Non-convulsive status epilepticus after ischemic stroke: a hospitalbased stroke cohort study
Vincenzo Belcastro • Simone Vidale •
Gaetano Gorgone • Laura Rosa Pisani •
Luigi Sironi • Marco Arnaboldi • Francesco Pisani
Received: 10 May 2014 / Revised: 19 July 2014 / Accepted: 11 August 2014
Ó Springer-Verlag Berlin Heidelberg 2014
Abstract To evaluate in the setting of a stroke unit ward
the usefulness of a prolonged ([6 h) video-EEG recording
(PVEEG) in identifying non-convulsive status epilepticus
(NCSE) in patients with an acute ischemic stroke. Predictors of NCSE were also evaluated. Patients with an acute
ischemic stroke, referred to our unit, were included in this
prospective observational study. A PVEEG recording was
implemented after stroke in all patients during the first
week: (a) promptly in those exhibiting a clear or suspected
epileptic manifestation; (b) at any time during the routine
activity in the remaining patients. After the first week, a
standard EEG/PVEEG recording was hooked up only in
presence of an evident or suspected epileptic manifestation
or as control of a previous epileptic episode. NCSE was
identified in 32 of the 889 patients (3.6 %) included in the
study. It occurred early (within the first week) in 20/32
(62.5 %) patients and late in the remaining 12. Diagnosis
was made on the basis of a specific clinical suspect
(n = 19, 59.4 %) or without any suspect (n = 13, 40.6 %).
In a multivariate analysis, a significant association of
NCSE was observed with NIHSS score, infarct size and
large atherothrombotic etiology. NCSE is not a rare event
after an acute ischemic stroke and a delayed diagnosis
V. Belcastro (&) S. Vidale L. Sironi M. Arnaboldi
Neurology Unit, S. Anna Hospital, Como, Italy
e-mail: vincenzobelcastro@libero.it
G. Gorgone
Neurology Unit, Treviglio-Caravaggio Hospital, Treviglio, Italy
L. R. Pisani
IRCCS Centro Neurolesi ‘‘Bonino-Pulejo’’, Messina, Italy
F. Pisani
Department of Neurosciences, University of Messina, Messina,
Italy
could worsen patient prognosis. Since NCSE can be difficult to be diagnosed only on clinical grounds, implementation of a prompt PVEEG should be kept available in a
stroke unit whenever a patient develop signs, although
subtle, consistent with NCSE.
Keywords Non-convulsive status epilepticus
Post-stroke seizures Prolonged video-EEG recording
Introduction
Stroke is the most common cause of symptomatic epilepsy
in older adults and status epilepticus (SE) can be a presenting clinical feature of an acute stroke [1–6]. Values
concerning the incidence of post-stroke seizures and
development of epilepsy after stroke show a great variability in the literature. This variability is due to different
reasons: small sample sizes, different study designs, heterogeneous terminology, different periods of follow-up, no
specification on the ischemic/hemorrhagic nature of the
lesion. Incidence values of ‘‘early’’ (i.e., within the first
7 days) seizures range from 3 to 15 % [1, 7–11], but values
[30 % have been also reported in the literature [5]. ‘‘Late’’
seizures, i.e., those after the first week, have been reported
to occur in a percentage ranging from 3 to 9 % with a
gradual increase by increasing the time period after stroke,
the higher values being observed at 10 years [1, 4, 5].
Similarly, post-stroke epilepsy has been seen to develop in
6.4 % of patients as a mean value, ranging from 1.5 % at
3 months to 12.4 % at 10 years after stroke [2]. Concerning the incidence of SE, it has been estimated to range from
0.2 to 1.4 % [9–14] with very small differences observed
between ischemic and hemorrhagic stroke (0.2 vs. 0.3 %)
[14]. Additionally, most of these data regard convulsive
123
J Neurol
generalized SE (CGSE) and only few observations have
been specifically reported on non-convulsive SE (NCSE)
after stroke [6, 15]. In particular, in comatose patients with
NCSE, 20 % of patients had an ischemic stroke [15].
NCSE has been defined as a state of ongoing (or nonrecovery between) seizures without convulsions, usually
for more than 30 min [16]. It is a clinical entity very difficult to be detected in acute and critically ill patients,
including those with stroke, because of its subtle, nonconvulsive manifestations associated with transient consciousness disturbances [16–20].
Using a hospital-based stroke cohort, the first aim of the
present study was to evaluate in the setting of a stroke unit
ward the usefulness of a prolonged, namely lasting for at
least 6 h, video-EEG recording (PVEEG) in identifying
episodes of NCSE after an acute ischemic stroke. Predictors of NCSE were also evaluated.
Patients and methods
The Institutional Review Board of the S. Anna Hospital,
Como, Italy, approved the study and granted a waiver of
informed consent. The present investigation is a prospective observational study.
Patient population
Patients, referred to the Stroke Unit of S. Anna Hospital of
Como between January 2010 and September 2013, were
included in this study according to the following inclusion
criteria: (1) acute ischemic stroke defined by the occurrence of acute neurologic signs and symptoms of vascular
origin lasting 24 h or longer and documented through brain
CT/MRI; (2) no other concomitant causes potentially
responsible of acute seizures (for example: alcohol withdrawal, psychotropic drugs, electrolyte disturbance, previous brain lesions). A direct interview was required, except
for aphasic or uncooperative patients, for whom the interview was released by his relatives.
Patients with: (1) a previous history of seizures; (2)
recurrent stroke; (3) i.v. or p.o. treatment with antiepileptic
(AEDs) or sedative drugs, including barbiturates, benzodiazepines and propofol, prior to video-EEG execution; (4)
hemorrhagic transformation of the ischemic area were
excluded from analysis.
EEG recording and epilepsy characterization
Given that [50 % of epileptic seizures/SE occur in the
first days after stroke [4, 5, 7–10, 13], a PVEEG
recording, including ECG tracing, was implemented in all
patients admitted to our Stroke Unit within the first week
123
after stroke onset: it was hooked up as soon as an epileptic activity was clear or suspected or at any time
during the routine activity in the remaining patients.
Additional conventional (i.e., of *30 min) EEG recordings, or even PVEEG when clinically judged more
appropriate, were made after the first week only in presence of clear or suspected epileptic seizures/SE or as
control of a previous epileptic episode. EEG was recorded digitally using caps with 21 fixed gel electrodes
placed according to the International 10–20 system. EEGs
were
immediately
evaluated
by
board-certified
electroencephalographers.
Seizures were divided into convulsive and non-convulsive. Convulsive seizures (CSz) were described as
‘‘tonic–clonic,’’ ‘‘clonic’’ or ‘‘tonic’’ (including also synonyms like ‘‘jerking’’ or ‘‘twitching’’). SE was reported
as convulsive (CSE) for CSz lasting longer than 5 min or
if 2 or more fits occurred without a return to baseline in
between. Non-convulsive seizures (NCSz) were considered those characterized by subtle movements like slow
facial twitching and eye deviation. NCSE was defined as
30 min of continuous seizure activity without major
motor phenomena and according to the criteria of Young
et al. [16]. At least one of the following three primary
EEG criteria and at least one of the four secondary EEG
criteria had to be fulfilled. Primary criteria were (1)
repetitive focal or generalized spikes, sharp waves, spikeand-wave or sharp-and-slow wave complexes at a frequency of [3/s; (2) repetitive focal or generalized spikes,
sharp waves, spike-and-wave or sharp-and-slow wave
complexes at a frequency of [3/s and secondary significant improvement in clinical state or baseline EEG after
administration of AEDs; or (3) sequential rhythmic waves
and secondary criteria (1), (2), (3) with or without significant improvement in clinical state or baseline EEG
after administration of AEDs. Secondary EEG criteria
were the following discharge patterns with a duration of
[10 s: (1) incrementing onset with increase in voltage
and/or increase or slowing of frequency; (2) decrementing
offset with decrease in voltage or frequency; (3) postdischarge slowing or voltage attenuation; or (4) significant improvement in clinical state or baseline EEG after
administration of AEDs.
Patients with EEG showing only bilateral periodic discharges or stimulus-induced rhythmic, periodic or ictal
discharges (SIRPIDs) were excluded. The presence of
periodic lateralized epileptiform discharges (PLEDs), an
EEG pattern usually observed in patients with acute stroke,
was not the reason for exclusion.
NCSE was coded into two categories: ‘‘early’’ referred
to that occurring within a week after stroke and considered
to be provoked by the acute brain insult; ‘‘late’’ that
developing after the first week [10]. Video recordings were
J Neurol
screened to determine clinical correlates for NCSE and
episodes of electroencephalographic seizures. Level of
consciousness, evaluated before pharmacological treatment
and upon EEG initiation, was categorized on clinical
grounds as alert/somnolent/stuporous/comatose. Each
patient was monitored through video-EEG recording for at
least 6 h following SE resolution, and a further conventional EEG recording was made after 24 h.
Antiepileptic drug treatment
When a true diagnosis of SE was made, pharmacological
treatment was promptly started and its effect evaluated
through both EEG and clinical monitoring. As a standard
therapeutic approach, i.v. administered drugs were: lorazepam (0.1 mg/kg within 1–2 min), diazepam (10 mg,
5 mg/min) or phenytoin (PHT) (18 mg/kg, 5 mg/kg/min).
According to the recent literature suggestions [21], levetiracetam (LEV) 25 mg/kg over 15 min or lacosamide
(LCM) as initial bolus dose of 400 mg over 30 min was
used in patients with concomitant medical conditions in
whom the use of traditional AEDs was judged potentially
unsafe because of their adverse effects (hypotension, QT
prolongation, arrhythmias, respiratory complications).
During i.v. treatment, patients were monitored through
video-EEG, ECG, heart rate and blood pressure measurements. SE was considered fully controlled only after
both clinical and electrographic resolution.
Stroke characterization
Based on history and the results of diagnostic studies
including brain imaging, Doppler sonography, echocardiograms, and electrocardiograms, patients were classified
according to stroke syndrome, etiology and severity.
Clinical syndrome classification was assessed using the
Oxfordshire Community Stroke Project (OCSP) criteria
[22]. Stroke etiology was classified using the TOAST
(Trial of ORG 10172 in Acute Stroke Treatment) criteria
[23]. Stroke severity was assessed at admission using the
National Institutes of Health (NIH) stroke scale (NIHSS)
score [24]. Neuroimaging data (CT or MRI) were evaluated
according to the arterial territory circulation involved
(anterior or posterior) and lesion size (large, medium,
small, or lacunar). A large lesion was required to have a
diameter [3 cm on neuroimages; a small lesion a diameter
\1 cm; a lacunar lesion was a CT hypodense area or T2
MRI hyperintense area \1 cm in maximum diameter deeply located in the cerebral hemispheres or in the brainstem.
Finally, functional disability was obtained using the modified Rankin scale (mRS) [25] and mortality was assessed
during hospitalization.
Statistical analysis
Distribution of gender, vascular risk factors, functional
disability, OCSP and TOAST criteria were summarized as
frequencies and percentages and comparisons were made
using the Pearson Chi-square test or Fisher exact test, as
appropriate. Age and NIHSS score were indicated as
median and interquartile range, and were investigated using
the Mann–Whitney U test. Univariate and multivariate
logistic regression models were used to evaluate the following potential determinants of NCSE: stroke syndrome,
stroke etiology, size of infarction, NIHSS score, age,
gender, vascular risk factors. In the analysis, the size of
infarction was expressed as a three-level ordinal variable,
according to the classification above mentioned. Multivariate models were constructed using all variables resulted
significant in the univariate analysis. Measures of association were odds ratios (ORs) with 95 % confidence intervals. Each covariate was tested independently and with the
main interaction terms. Finally, the ROC analysis was
performed to check the goodness of fit of the implemented
logistic regression final model. Statistical significance was
chosen at the 5 % level. All statistics were implemented
using STATA version 12 (StataCorp, USA).
Results
Cohort characteristics
A total of 1,362 patients were admitted to our Stroke Unit
between January 2010 and September 2013: 1,034 (76 %)
have an ischemic stroke; 227 (16.6 %) have an intracerebral hemorrhage (ICH); 101 (7.4 %) have a subarachnoid hemorrhage. Of the 1,034 patients with ischemic
stroke, 72 (7 %) were excluded from the analysis because
of a hemorrhagic transformation of the ischemic area.
Moreover, 73 patients were also excluded from the analysis
because of a previous history of seizures, sedative medication prior to PVEEG recording, a history of recurrent
stroke or EEG recordings showing only bilateral periodic
discharges. The final cohort consisted of 889 patients. The
characteristics of the sample are summarized in Table 1.
Seizures and status epilepticus
Overall 59/889 (6.6 %) patients exhibited epileptic manifestations. Early CSz and CSE occurred in 41/889 (4.6 %)
patients: seizures were tonic–clonic in 14, clonic without
impairment of consciousness in 13, and clonic with
automatisms in 6; focal motor SE occurred in four patients
and CSE in the remaining four. Sixteen out of these 41
patients also developed late CSzs (n = 12) or focal motor
123
J Neurol
Table 1 Demographic and
clinical characteristics of the
screened population
Variable
NCSE (n = 32)
No NCSE (n = 857)
Overall (n = 889)
66.5 (61.0–78)
16 (50)
65.0 (58.0–77.0)
550 (64.1)
65.0 (58.0–77.0)
566 (63.6)
Demographic data
Age (years)
Male
Stroke severity indicators
NIHSS
mRS [2
13.0 (9.0–15.0)*
3.0 (2.0–7.0)
4.0 (2.0–8.0)
12 (37.5)
343 (40.0)
355 (39.5)
Large size infarct
B1 cm
2 (6.3)*
447 (52.2)
449 (50.5)
[1 cm
9 (28.1)*
284 (33.1)
293 (33.0)
[3 cm
21 (65.7)*
126 (14.7)
147 (16.5)
Vascular risk factors
Discrete variables (sex, large
size infarct, vascular risk
factors, OCSP and TOAST
ischemic stroke subtypes) are
indicated as number of cases
(percentage). Age is indicated
as median (interquartile range)
and NIHSS score as geometric
mean (standard error)
mRS Modified Rankin Scale,
NCSE non-convulsive status
epilepticus, NIHSS NIH Stroke
Scale, PACI partial anterior
circulation infarct, POCI
posterior circulation infarct,
TACI total anterior circulation
infarct, TIA transient ischemic
attack
* p \ 0.0001, p = 0.0001 and
à
p = 0.001 in comparison to
No NCSE patients
Hypertension
Diabetes
430 (50.1)
446 (50.1)
9 (28.1)
186 (21.7)
195 (21.7)
Cardiac arrhythmia
5 (15.6)
188 (15.6)
192 (21.4)
Smoking
2 (6.2)
310 (36.2)
312 (35.0)
Hyperlipidemia
8 (25)
276 (32.2)
284 (31.5)
TIA
5 (15.6)
120 (14.0)
125 (13.9)
15 (46.8)à
123 (14.3)
138 (15.5)
11 (34.3)
126 (14.7)
137 (15.4)
385 (41.5)
385 (45.0)
268 (31.2)
274 (30.8)
OCSP classification
TACI
PACI
LACI
POCI
0
6 (18.7)
TOAST classification
Large artery atherothrombosis
13 (40.6)
69 (8.0)
82 (9.2)
Cardioembolic
10 (31.2)
227 (26.5)
237 (26.6)
Lacunar
315 (36.7)
315 (35.4)
Cryptogenetic
7 (21.9)
214 (25.0)
221 (24.8)
Other
2 (6.2)
58 (6.8)
60 (6.8)
SE (n = 4) during the period of hospitalization. Two
patients exhibited for the first time tonic–clonic Sz during
the third week after stroke. Complex partial seizures,
although suspected in eight additional patients on the basis
of apparently not oriented motor manifestations and/or
level of consciousness oscillations, were not firmly demonstrated through EEG recording. Similarly, electrical
seizures not associated to clinical manifestations were not
intercepted in our population of patients during hospitalization. In our cohort, NCSE was identified in 32/889
(3.6 %) patients: it occurred early in 20/32 (i.e., 62.5 %)
and late in the remaining 12 patients. Early NCSE developed as continuation of a CSE in five patients and during
the post-ictal phase of a tonic–clonic seizure in two
patients. Late NCSE was diagnosed in 12 patients: during
the second week in eight and during the third week in the
remaining four. Seven of these 12 patients had already had
early tonic–clonic seizures and two early focal motor SE;
the remaining three patients developed NCSE without any
previously identified epileptic activity. Impairment of
123
16 (50.0)
0
consciousness and agitation were the main manifestations
in 22/32 (68.8 %) patients and aphasia in the remaining 10
(31.2 %). Noteworthy, NCSE was diagnosed on the basis
of a specific clinical suspect in 19 patients (59.4 %) and
without any clinical suspect in the remaining 13 patients
(40.6 %). In 6 out of these 13 patients, it was intercepted in
the first week during PVEEG recording implemented as
part of the routine activity, and in the other seven patients
during PVEEG made to monitor the time course of a tonic–
clonic seizure (1 patient in the first week), a CSE (3
patients in the first week), or made just as a control in 3
patients (1 in the second week and 2 in the third week) who
had had previous CSzs.
EEG features
PVEEG was recorded for a mean duration of 9 h 22 min,
range 6 h 15 min–13 h 27 min.
PVEEG recording was implemented within the first
3 days in the majority of patients, namely 605, and
J Neurol
between the forth and the seventh day in the remaining 284
patients.
Ictal EEG features showed: (a) high-voltage, theta
rhythmic activity with intermingled spikes over the temporo-occipital regions in eight patients; (b) high-voltage
theta activity intermingled with sharp waves over frontal
regions in seven patients; (c) bilateral, over frontal regions,
continuous spike- and slow-wave discharges in six patients.
Continuous focal sharp waves with change in amplitude,
frequency and spatial distribution were observed in 11
patients.
Treatment and epilepsy outcome
Initial treatment of NCSE was efficacious in 21 patients in
a time ranging from 30 to 60 min after i.v. administration:
lorazepam (n = 4), diazepam (n = 9), LCM (n = 5) and
PHT (n = 3). Eight patients did not show any response to
initial diazepam treatment and were treated with PHT
(n = 6) or LEV (n = 2) introduced as second-line treatment. None of these patients relapse. In three patients,
NCSE was refractory to diazepam followed by PHT and
LCM and a transfer to the intensive care unit was needed.
A chronic treatment with AEDs was started in all SE
patients. The following drugs were given to patients with
NCSE: LEV in 24 patients, oxcarbazepine in six and
phenobarbital in the remaining two patients. The present
follow-up for 28 of these patients ranges from 2 to
43 months. In the control visits, usually planned every
6–8 weeks, a global neurological evaluation, seizure frequency recordings through ad hoc calendars, therapy
adjustments and standard EEG execution have been performed. Of these patients, 25 developed post-stroke epilepsy with seizure frequency of 3–8 fits per month (18
patients) and \1 per month (7 patients). Three patients are
seizure free (present follow-up 5, 13 and 16 months).
Detailed data concerning long-term outcome of both epilepsy and stroke in our population will be object of a
forthcoming article. No information has been obtained on
the remaining four patients after discharge from hospital
and until the preparation of the present manuscript.
Stroke features and outcome
Of the 32 patients with NCSE, a total anterior circulation
infarct (TACI) was diagnosed in 15, a partial anterior circulation infarct (PACI) in 11 and a posterior circulation
infarct (POCI) in 6 patients. Three patients with early
NCSE received intravenous recombinant tissue plasminogen activator (t-PA). None of the patients developing
NCSE died during the period of hospitalization, and all
were discharged to rehabilitation or to home within
14–30 days after stroke onset. No differences in outcome
Table 2 Independent predictors of NCSE in patients with ischemic
stroke
SE
z
p
OR
95 % CI
TACI
0.24
-1.46
0.1
0.34
0.12
1.03
Large artery
atherothrombosis
1.57
3.07
0.002
3.7
1.60
6.49
Infarct size
NIHSS score
1.56
0.05
2.92
2.72
0.003
0.007
3.6
1.14
1.52
1.04
8.44
1.24
Variables not in equations: gender, age, hypertension, diabetes mellitus, cardiac arrhythmia, smoking, hyperlipidemia, TIA, PACI,
POCI, other TOAST classification aetiologies
CI confidence interval, OR odds ratio, SE standard error, NIHSS NIH
Stroke Scale
of stroke between NCSE patients and the other patients
were detected through the NIHH score and other functional
assessment scales used (Barthel index, modified Rankin
Scale). As above mentioned, detailed data on the long-term
outcome of our population will be illustrated in a forthcoming study.
Statistical significance
Univariate analysis showed significant differences between
patients with NCSE and those without NCSE with regard
to etiology, clinical presentation and size of ischemic area
(Table 1). Regarding stroke etiology, large artery atherothrombosis showed a different distribution between NCSE
and non-NCSE groups (Table 1). In the multiple logistic
regression models, we considered as predictors of the
NCSE those variables resulting significant from the univariate analysis (Table 1). A significant association of
NCSE was observed with NIHSS score, infarct size and
large atherothrombotic etiology (Table 2). No significant
differences between the expected and the observed
regression matrix were observed (area under ROC
curve = 0.87).
Discussion
The results of the present study indicate that NCSE is not a
rare event after an acute ischemic stroke and occurs in a
percentage of 3–4 % of patients. In our cohort, significant
predictors of NCSE were a large size infarct, large artery
atherothrombosis and NIHSS score at admission.
With regard to infarct area, it has been proposed that the
ischemic penumbra of a stroke can contain electrically
irritable tissue that provides a focus for seizure activity
[26]. Enhanced release of glutamate, ionic imbalances
associated to breakdown of membrane phospholipids,
release of free fatty acids, reduced GABAergic function
123
J Neurol
with functional and structural impairment of GABAergic
interneurons have been described to occur in the ischemic
area after an acute insult [27, 28]. Cortical location and
large size infarct have been constantly found to be associated with post-stroke seizures [2, 9, 11], a finding that has
emerged also from the present study. On the basis of the
cascade of synaptic and intracellular events shared by
seizure and vascular brain injuries, it can be hypothesized
that post-stroke NCSE could be triggered by acute cellular
biochemical disturbances occurring with acute ischemic
stroke.
In the present study, large artery atherothrombosis was
another significant predictor of NCSE.
Large artery atherothrombosis is more common in cortical than lacunar infarcts, implying that an embolic etiology would be more common in TACIs and PACIs than in
LACIs [29]. Notably, in our NCSE group, TACI was
diagnosed in 15/32 (46.8 %) and PACI in 11/32 (34.3 %)
patients while lacunar infarct was completely absent. These
observations reinforce the crucial role of cortical involvement in seizure genesis.
A clinically relevant aspect is the possible impact of
NCSE on outcome of patients with ischemic stroke. Previous studies have shown that NCSE is a risk factor for the
development of refractory SE, poor outcome, and increased
mortality independently of the underlying pathology [16,
30]. In one of these studies, for example, the mortality rate
in an intensive care unit was high: more than half of the
patients with NCSE, 13/23 (57 %) died [16]. Specific data
on NCSE after an ischemic stroke are difficult to be
extrapolated from published studies because of a number of
reasons, including no reported distinction between CSE
and NCSE in the analysis, combination of both patients
with ischemic stroke and ICH, focus mainly on EEG
activity, generic report on patients with acute insults as
seen in medical intensive care units [8, 16–18, 20, 31]. In
our cohort, cases of death were not observed during hospitalization and until the present follow-up. Additionally,
NCSE patients did not show, as compared to patients
without NCSE, differences concerning outcome and complications until the present follow-up.
The NIHSS score was a significant predictor of NCSE.
Interpretation of NIHSS score is difficult since the impairment of consciousness, which is a hallmark of NCSE, could
have increased itself NIHSS score at admission. In fact,
early NCSE occurred in 62.5 % of our 32 NCSE patients,
this value being in agreement with that reported in critically
ill patients, in whom seizures activity was observed in the
first 24 h after the insult in a large majority of patients
(88 %) [20]. Additionally, as above mentioned, SE has been
seen to be a possible presenting clinical feature of an acute
stroke [1–3]. On this basis, it can be hypothesized that a
percentage of our patients might have at admittance a
123
condition of NCSE which has not been detected. In patients
with ischemic stroke, values concerning the incidence of SE
range between 1 and 10 % [20] and NCSE has been
reported to occur in 30 % of patients developing post-stroke
SE [13]. In our cohort, 5 out of 20 patients developed early
NCSE as continuation of CSE and in two patients it
occurred in the post-ictal phase of a tonic–clonic seizure.
These values are very close to those reported in other
studies [17, 32] and emphasize the importance to have the
possibility in a stroke unit to implement a VEEG recording
longer than a conventional one, which lasts about
20–30 min on average. In a population of 110 patients, the
first seizures were detected in slightly more than half of the
cases within the first hour of recording [33].
The second relevant finding deriving from the present
study, concerning EEG, is that the diagnosis of NCSE has
been made during a routine EEG recording and without any
clinical suspect or any previous epileptic manifestation in a
considerable proportion of patients. This result, in line with
a large body of literature observations [18, 31, 32], provides further support in favor of the view that the incidence
of NCSE after an ischemic stroke is underappreciated and
that this pathology is more frequent than thought in the
past. As above mentioned, in fact, NCSE is a clinical entity
which goes easily unnoticed especially in elderly patients
and in patients with acute brain disorders in whom
impairment of consciousness is frequent as a consequence
of various pathologies [32–36]. In particular, in patients
with an acute ischemic stroke diagnosis of NCSE might be
made difficult because of various factors: (1) NCSE occurs
more frequently in the first days after stroke onset when
impairment of consciousness is a common expression of
stroke itself, (2) NCSE can occur just after CSz/CSE and
might be easily interpreted as a post-ictal state, (3) manifestations of NCSE can be so subtle as no to raise the
suspect of an epileptic activity.
Our study has different limitations: (1) the single-institution approach makes the study vulnerable to bias; (2) the
relatively small sample size of NCSE group does not
necessarily represent all the characteristics of patients
suffering from acute stroke and developing NCSE; (3)
cases of NCSE before and after PVEEG implementation
could have been easily missed. The latter limit, in particular, has been also emphasized by various authors [36, 37].
Conclusions
This population-based study shows that implementation of
a PVEGG should be permanently available in a stroke
unit. In this setting, a conventional EEG recording is of
limited diagnostic value because of the intermittent nature
of subtle, difficult-to-detect seizures and PVEEG
J Neurol
recording increases the possibility to diagnose NCSE. An
increased awareness that NCSE is a possible, not rare
event after an ischemic stroke and that it can most frequently occur with subtle signs can contribute to speed a
diagnosis of NCSE. A quick diagnosis of this condition is
crucial to implement a prompt treatment and hence to
avoid possible clinical complications in patients with
ischemic stroke [33, 37].
17.
18.
19.
Acknowledgments We would like to acknowledge the contribution
of Neurology EEG technologists of S. Anna Hospital in connecting
these patients to the PVEEG at all odd hours.
20.
Conflicts of interest
21.
None.
22.
References
23.
1. So EL, Annegers JF, Hauser WA, O’Brien PC, Whisnant JP
(1996) Population-based study of seizure disorders after cerebral
infarction. Neurology 46:350–355
2. Graham NS, Crichton S, Koutroumanidis M, Wolfe CD, Rudd
AG (2013) Incidence and associations of post-stroke epilepsy: the
prospective South London stroke register. Stroke 44:605–611
3. Ryvlin P, Montavont A, Nighoghossian N (2006) Optimizing
therapy of seizures in stroke patients. Neurology 67:S3–S9
4. Lossius MI, Ronning OM, Slapo GD, Mowinckel P, Gjerstad L
(2005) Post-stroke epilepsy: occurrence and predictors: a longterm prospective controlled study (Akershus Stroke Study).
Epilepsia 46:1246–1251
5. Camilo O, Goldstein LB (2004) Seizures and epilepsy after
ischemic stroke. Stroke 35:1769–1775
6. Afsar N, Kaya D, Aktan S, Aykut-Bingol C (2003) Stroke and
status epilepticus: stroke type, type of status epilepticus, and
prognosis. Seizure 12:23–27
7. Beghi E, D’Alessandro R, Beretta S et al (2011) Incidence and
predictors of acute symptomatic seizures after stroke. Neurology
77:1785–1793
8. Pezzini A, Grassi M, Del Zotto E et al (2013) Complications of
acute stroke and the occurrence of early seizures. Cerebrovasc
Dis 35:444–450
9. Labovitz DL, Hauser WA, Sacco RL (2001) Prevalence and
predictors of early seizure and status epilepticus after first stroke.
Neurology 57:200–206
¨ zmenoglu M, Boz C, Alioglu Z (2001) Status
10. Velioglu SK, O
epilepticus after stroke. Stroke 32:1169–1172
11. Procaccianti G, Zaniboni A, Rondelli F, Crisci M, Sacquengna T
(2012) Seizures in acute stroke: incidence, risk factors and
prognosis. Neuroepidemiology 39:45–50
12. Kilpatrick CJ, Davis SM, Tress B et al (1990) Epileptic seizures
in acute stroke. Arch Neurol 47:157–160
13. Rumbach L, Sablot D, Berger E, Tatu L, Vuillier F, Moulin T
(2000) Status epilepticus in stroke: report on a hospital-based
stroke cohort. Neurology 54:350–354
14. Bateman BT, Claassen J, Willey JZ et al (2007) Convulsive status
epilepticus after ischemic stroke and intracerebral haemorrhage:
frequency, predictors, and impact on outcome in a large administrative dataset. Neurocrit Care 7:187–193
15. Towne AR, Waterhouse EJ, Boggs JG et al (2000) Prevalence of
nonconvulsive status epilepticus in comatose patients. Neurology
25:340–345
16. Young GB, Jordan KG, Doig GS (1996) An assessment of nonconvulsive seizures in the intensive care unit using continuous
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
EEG monitoring: an investigation of variables associated with
mortality. Neurology 47:83–89
De Lorenzo RJ, Waterhouse EJ, Towne AR et al (1998) Persistent
nonconvulsive status epilepticus after the control of convulsive
status epilepticus. Epilepsia 39:833–840
Bottaro FJ, Martinez OA, Pardal MM, Bruetman JE, Reisin RC
(2007) Nonconvulsive status epilepticus in the elderly: a case–
control study. Epilepsia 48:966–972
Ney JP, van der Goes DN, Nuwer MR, Nelson L, Eccher MA
(2013) Continuous and routine EEG in intensive care: utilization
and outcomes, United States 2005–2009. Neurology
81:2002–2008
Sutter R, Stevens RD, Kaplan PW (2013) Continuous electroencephalographic monitoring in critically ill patients: indications,
limitations, and strategies. Crit Care Med 41:1124–1132
Shorvon S (2011) The treatment of status epilepticus. Curr Opin
Neurol 24:165–170
Burn J, Dennis M, Bamford J, Sandercock P, Wade D, Warlow C
(1994) Long-term risk of recurrent stroke after a first-ever stroke.
The Oxfordshire Community Stroke Project. Stroke 25:333–337
Goldstein LB, Jones MR, Matchar D et al (2001) Improving the
reliability of stroke subgroup classification using the Trial of
ORG 10172 in Acute Stroke Treatment (TOAST) criteria. Stroke
32:1091–1098
Brott T, Adams HP Jr, Olinger CP et al (1989) Measurements of
acute cerebral infarction: a clinical examination scale. Stroke
20:864–870
van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn
J (1988) Interobserver agreement for the assessment of handicap
in stroke patients. Stroke 19:604–607
Heiss WD, Huber M, Fink GR et al (1992) Progressive
derangement of periinfarct viable tissue in ischemic stroke.
J Cereb Blood Flow Metab 12:193–203
Luhmann HJ, Mudrick-Donnon LA, Mittmann T, Heinemann U
(1995) Ischemia-induced long-term hyperexcitability in rat neocortex. Eur J Neurosci 7:180–191
Belcastro V, Pierguidi L, Tambasco N (2011) Levetiracetam in
brain ischemia: clinical implications in neuroprotection and prevention of post-stroke epilepsy. Brain Dev 33:289–293
Mead GE, Lewis SC, Wardlaw JM, Dennis MS, Warlow CP
(2000) How well does the Oxfordshire community stroke project
classification predict the site and size of the infarct on brain
imaging? J Neurol Neurosurg Psychiatry 68:558–562
Mayer SA, Claassen J, Lokin J, Mendelsohn F, Dennis LJ,
Fitzsimmons BF (2002) Refractory status epilepticus: frequency,
risk factors, and impact on outcome. Arch Neurol 59:205–210
Claassen J, Mayer SA, Kowalski RG, Emerson RG, Hirsch LJ
(2004) Detection of electrographic seizures with continuous EEG
monitoring in critically ill patients. Neurology 62:1743–1748
Meierkord H, Holtkamp M (2007) Non-convulsive status epilepticus in adults: clinical forms and treatment. Lancet Neurol
6:329–339
Claassen J, Lokin JK, Fitzsimmons BF, Mendelsohn FA, Mayer
SA (2002) Predictors of functional disability and mortality after
status epilepticus. Neurology 58:139–142
Jordan KG (2004) Emergency EEG and continuous EEG monitoring in acute ischemic stroke. J Clin Neurophysiol 21:341–352
Carrera E, Michel P, Despland PA et al (2006) Continuous
assessment of electrical epileptic activity in acute stroke. Neurology 67:99–104
Sutter R, Fuhr P, Grize L, Marsch S, Ru¨egg S (2011) Continuous
video-EEG monitoring increases detection rate of nonconvulsive
status epilepticus in the ICU. Epilepsia 52:453–457
Huang CW, Saposnik G, Fang J, Steven DA, Burneo JG (2014)
Influence of seizures on stroke outcomes: a large multicenter
study. Neurology 82:768–776
123