Clinical Neurophysiology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Clinical Neurophysiology
journal homepage: www.elsevier.com/locate/clinph

Electroencephalography for diagnosis and prognosis of acute

encephalitis q
Raoul Sutter a,b,c,d,⇑, Peter W. Kaplan b, Mackenzie C. Cervenka e, Kiran T. Thakur f, Anthony O. Asemota f,
Arun Venkatesan f, Romergryko G. Geocadin a,f
a
Division of Neurosciences Critical Care, Department of Anesthesiology and Critical Care Medicine, Johns Hopkins University School of Medicine, Baltimore, MD, USA
b
Department of Neurology, Johns Hopkins Bayview Medical Center, Baltimore, MD, USA
c
Department of Neurology University Hospital Basel, Basel, Switzerland
d
Clinic of Intensive Care Medicine, University Hospital Basel, Basel, Switzerland
e
Division of Epilepsy, Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD, USA
f
Johns Hopkins Encephalitis Center, Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD, USA

a r t i c l e i n f o h i g h l i g h t s

Article history:
 EEG provides significant diagnostic and prognostic information in acute encephalitis.
Accepted 3 November 2014
 Periodic discharges and focal slowing were associated with Herpes Simplex encephalitis.
Available online xxxx
 Normal EEG was the strongest association with survival independently from possible confounders.

Keywords:
Encephalitis
Electroencephalography a b s t r a c t
Neurocritical care
EEG patterns
Objectives: To confirm the previously identified EEG characteristics for HSV encephalitis and to
determine the diagnostic and predictive value of electroencephalography (EEG) features for etiology
and outcome of acute encephalitis in adults. In addition, we sought to investigate their independence
from possible clinical confounders.
Methods: This study was performed in the Intensive Care Units of two academic tertiary care centers.
From 1997 to 2011, all consecutive patients with acute encephalitis who received one or more EEGs were
included. Examination of the diagnostic and predictive value of EEG patterns regarding etiology, clinical
conditions, and survival was performed. The main outcome measure was in-hospital death.
Results: Of 103 patients with encephalitis, EEGs were performed in 76 within a median of 1 day (inter
quartile range 0.5–3) after admission. Mortality was 19.7%. Higher proportions of periodic discharges
(PDs) (p = 0.029) and focal slowing (p = 0.017) were detected in Herpes Simplex virus (HSV) encephalitis
as compared to non-HSV encephalitis, while clinical characteristics did not differ. Normal EEG remained
the strongest association with a low relative risk for death in multivariable analyses (RR < 0.001,
p < 0.001) adjusting for confounders as coma, global cerebral edema and mechanical ventilation. None
of the patients with a normal EEG had a GCS of 15.
Conclusions: Normal EEG predicted survival independently from possible confounders, highlighting the
prognostic value of EEG in evaluating patients with encephalitis. EEG revealed higher proportions of
PDs along with focal slowing in HSV encephalitis as compared to other etiologies.
Significance: EEG significantly adds to clinical, diagnostic and prognostic information in patients with
acute encephalitis.
Ó 2014 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights
reserved.

q
1. Introduction
Institutions where the work was performed: The Johns Hopkins Hospital and
the Johns Hopkins Bayview Medical Center.
⇑ Corresponding author at: Department of Neurology and Intensive Care Units, Acute encephalitis is a life-threatening neurologic condition
University Hospital Basel, Petersgraben 4, 4031 Basel, Switzerland. Tel.: +41 61 265 with an incidence of 12/100,000 cases per year, characterized by
25 25. inflammation of the brain (Granerod et al., 2010b). Symptoms
E-mail address: Raoul.Sutter@usb.ch (R. Sutter).

http://dx.doi.org/10.1016/j.clinph.2014.11.006
1388-2457/Ó 2014 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

Please cite this article in press as: Sutter R et al. Electroencephalography for diagnosis and prognosis of acute encephalitis. Clin Neurophysiol (2014), http://
dx.doi.org/10.1016/j.clinph.2014.11.006
2 R. Sutter et al. / Clinical Neurophysiology xxx (2014) xxx–xxx
usually emerge over hours to days and include fever, headache,
altered consciousness, seizures, and focal neurologic deficits
(Davies et al., 2006; Koskiniemi et al., 1981). Approximately one-
third of acute encephalitis cases are thought to be immune-medi-
ated (Granerod et al., 2010a) and etiology frequently remains
undetermined in >50% of cases in some studies (Koskiniemi
et al., 1981). The most frequently identified etiologies are infec-
tions, with viruses representing the predominant pathogens,
although recognition of autoimmune causes is increasing. Mortal-
ity is high, ranging up to 20% in patients with Herpes Simplex virus
(HSV) encephalitis (Whitley and Lakeman, 1995). Although elec-
troencephalography (EEG) is typically recommended in the evalu-
ation of patients with suspected encephalitis (Venkatesan et al.,
2013), the contribution of EEG to diagnosis and prognosis has not
been well characterized. The EEG features mostly described in
association with HSV encephalitis, a type of encephalitis best
examined with EEG, are uni- or bilateral periodic discharges, focal
or generalized slow waves, and electrical seizures (Al-Shekhlee
et al., 2006; Brodtkorb et al., 1982; Ch’ien et al., 1977; Illis and
Taylor, 1972; Lai and Gragasin, 1988; Upton and Gumpert, 1970).
However, investigations regarding the independent diagnostic
and predictive value of EEG patterns are lacking.
We aimed to confirm the previously identified EEG characteris-
tics for HSV encephalitis and to further determine the diagnostic
and predictive value of EEG features for different underlying etiol-
ogies and short-term outcome in the first week of acute encepha-
litis in adults. In addition, we sought to investigate their
independence from possible clinical confounders.
2. Materials and methods
This study was performed at the Johns Hopkins Hospital and the
Johns Hopkins Bayview Medical Center, two academic tertiary care
centers in Baltimore, USA. All neurocritical care units (NCCU), med-
ical ICUs, coronary care units (CCU), and surgical ICUs were
screened. The study was approved by the institutional review
boards; patient consent was waived.
2.1. Patients and data collection
From January 1997 to July 2011, all consecutive patients older
than 16 years of age with the diagnosis of acute encephalitis
requiring critical care for more than 48 h and who had an EEG in
the first week following admission were included. The first EEG
was assessed. Cases were identified using ICD-9 diagnosis codes
corresponding to encephalitis. Diagnoses were confirmed by neu-
rologists’ review of patient charts including physicians’ notes, lab-
oratory results, neuroimaging studies, and other supporting data.
Detailed information regarding clinical data collection was
described previously (Thakur et al., 2013).
2.2. Definition of encephalitis
According to the consensus statement of the international
encephalitis consortium and population-based studies (Granerod
et al., 2010a; Venkatesan et al., 2013), encephalitis was diagnosed
if a patient was encephalopathic (defined by depressed or altered
level of consciousness, lethargy, or personality change lasting at
least 24 h) with at least two of the following characteristics: fever,
seizure, focal neurologic deficit, central spinal fluid (CSF) pleocyto-
sis (white blood cell count >5 cells/mm3), and EEG or neuroimaging
findings consistent with encephalitis. Active malignancy, HIV
infection/AIDS, or use of chronic immunosuppressants defined an
immunocompromised state.
Please cite this article in press as: Sutter R et al. Electroencephalography for diagnosis and prognosis of acute encephalitis. Clin Neurophysiol (2014), http://
dx.doi.org/10.1016/j.clinph.2014.11.006
2.3. Clinical categories
Patients were categorized as having an infectious encephalitis
(i.e., viral, bacterial and fungal), autoimmune encephalitis, or
encephalitis of unknown etiology as previously reported (Thakur
et al., 2013). Infectious encephalitides were defined by serology,
positive polymerase chain reaction, culture, or histopathology.
Autoimmune encephalitis was defined by the presence of antigen-
specific antibodies in the serum and/or CSF or cases with a clinically
recognized autoimmune syndrome and supportive histopathologic
evidence. Cases of acute disseminated encephalomyelitis were cat-
egorized as autoimmune etiology and defined by clinical features
and imaging characteristics of acute disseminated encephalomyeli-
tis or histology-proven cases (Tenembaum et al., 2007).
2.4. Electroencephalography
EEGs were recorded in the first week following admission with
silver–silver chloride disk scalp electrodes placed according to the
International 10–20-System. All patients had one or more EEGs
recorded for at least 30 min. EEGs were analyzed by three neurolo-
gists’ board certified in epilepsy and/or clinical neurophysiology [RS,
PWK, MCC] and blinded to clinical or radiologic information. Inter-
rater agreement was assessed and if divergent, subsequent critical
review was performed to reach agreement. Background activity
was categorized into alpha, alpha/theta, theta, theta/delta, and delta
activity during arousal as defined elsewhere (Sutter et al., 2013a). As
periodic discharges (PDs) have been described in association with
HSV encephalitis in prior case series and smaller studies
(Al-Shekhlee et al., 2006; Brodtkorb et al., 1982; Illis et al., 1972;
Upton et al., 1970), we assessed the occurrence of episodic EEG tran-
sients, such as periodic discharges (PDs) according to the definition
provided by Chatrian et al. (1964), frontal intermittent rhythmic
delta activity (FIRDA), and triphasic waves (TWs) as defined else-
where (Sutter et al., 2013a). In addition, seizures and status epilep-
ticus were noted as defined previously (Ozuna, 2000; Sutter and
Kaplan, 2012). As a standard procedure, patients without arousals
were stimulated by an EEG technician who was trained to first give
verbal commands, open the patients’ eyes, and if still no arousal was
registered, to apply a noxious stimulation. Arousals were defined
according to standard criteria as an abrupt shift in the frequency
range of EEG background activity lasting P3 s that may include
alpha, theta and/or frequencies greater than 16 Hz, but not spindles
(American Sleep Disorders Association, 1992). EEG background
reactivity to stimuli was defined as an abrupt, intermittent shift
from the baseline frequency and/or amplitude ranges towards
higher or lower ranges immediately following arousal, acoustic
(call) or noxious stimulation.
2.5. Outcome
The principal outcome measure was in-hospital death. Second-
ary outcomes were length of hospital and ICU stay, Glasgow Coma
Scale (GCS) on the day of EEG and the modified Rankin Scale (mRS)
at discharge.
2.6. Statistics
Patients were categorized into the following three groups of
acute encephalitis: infectious encephalitis, autoimmune encephali-
tis and encephalitis of unknown etiology. Subsequently, patients
were dichotomized into additional subgroups: patients with
HSV-encephalitis and non-HSV-encephalitis, and into survivors
and non-survivors. Fisher’s exact test was used for comparisons
of proportions. For continuous variables, the Shapiro–Wilk
test was used to distinguish between normal and abnormal
 R. Sutter et al. / Clinical Neurophysiology xxx (2014) xxx–xxx 3

distributions. Normally distributed variables were analyzed by
the Student’s t test and non-normally distributed variables by
the Mann–Whitney U test. Means across all three categories were
compared by ANOVA analyses. In addition, ANOVA analyses were
performed to assess associations between GCS and specific EEG
background frequency ranges. Relative risks of death were esti-
mated by Poisson regression with robust error variance (Zou,
2004), and were adjusted for potential confounders including
coma, global cerebral edema and mechanical ventilation. The Pear-
son and deviance goodness-of-fit tests were performed to assess
the fit of the data to a Poisson distribution in the final regression
models.
Levels for statistical significance were set at two-tailed p-value
of <0.05.
Statistical analysis was performed with STATAÒ 12.0 (Stata
Corp., College Station, TX).

Fig. 1. Study population. EEG = electroencephalography.

Table 1
Comparisons of clinical and EEG characteristics among different types of acute encephalitis (n = 76).

Infectious encephalitis (n = 28) Autoimmune encephalitis (n = 11) Encephalitis, unknown cause (n = 37) p-Value#
Demographics
Gender, male (n, %) 13 46.4 6 54.6 20 54.1 0.816
Age, years (mean, SD) 54.8 ± 12.8 42.4 ± 16.9 49.1 ± 18.3 0.093§
Clinical features
GCS on admission (median, IQR) 10 5–14 13 9–13 10 7–14 0.645§
Comatose on admission, GCS 68 (n, %) 12 42.9 2 18.2 16 43.2 0.328
Charlson comorbidity index (mean, SD) 3.8 ± 2.9 2.3 ± 2.5 1.9 ± 2.7 0.032§
Global cerebral edema (n, %) 3 10.7 2 18.2 5 13.9 0.813
Immunosuppression (n, %) 13 46.4 1 9.1 6 16.2 0.012
Mechanical ventilation (n, %) 17 60.7 7 63.6 23 62.2 1.000
EEG characteristics
Background frequency range (n, %) 0.321
Alpha 15 53.6 7 63.6 10 27
Alpha/theta 2 7.1 0 0 4 10.8
Theta 4 14.3 2 18.2 11 29.7
Theta/delta 6 21.4 1 9.1 7 18.9
Delta 1 3.6 1 9.1 5 13.5
Focal slowing (n, %) 0.645
Frontal 1 3.9 1 9.1 2 5.4
Temporal 3 11.5 1 9.1 1 2.7
Central 1 3.9 0 0 2 5.4
Parietal 1 3.9 0 0 0 0
Occipital 2 7.7 0 0 0 0
Episodic transients (n, %)
FIRDA 0 0 0 0 2 5.4
TWs 0 0 1 9.1 3 8.1
PDs 3 11.5 1 9.1 2 5.4 0.625
Epileptic activity (n, %)
Seizures 1 3.9 1 9.1 2 5.4
Status epilepticus 4 15.4 0 0 2 5.4
Nonreactive EEG background activity (n, %) 17 65.4 2 18.2 4 11.4 0.08

GCS = Glasgow Coma Scale; FIRDA = frontal intermittent rhythmic delta activity; TWs = triphasic waves; PDs = periodic discharges; EEG = electroencephalography;
SD = standard deviation; IQR = inter quartile range. Bold p-values are considered significant.
#
Fisher’s exact test.
§
ANOVA.

Please cite this article in press as: Sutter R et al. Electroencephalography for diagnosis and prognosis of acute encephalitis. Clin Neurophysiol (2014), http://
dx.doi.org/10.1016/j.clinph.2014.11.006
4 R. Sutter et al. / Clinical Neurophysiology xxx (2014) xxx–xxx

3. Results agreed on every EEG characteristic in 68 patients (89%). In the

3.1. Demographics
We identified 103 patients with acute encephalitis of which 76
had interpretable EEGs recorded in the first week after admission
to the ICUs. Mean age was 50.2 years (range 18–80) and mortality
was 19.7%. EEGs were performed because of altered level of con-
sciousness and/or suspected seizures. The 27 patients who did
not receive an EEG within that time interval tended to have a
higher level of consciousness on admission (mean GCSadmission
11.2, SD 4.5 in patients without versus 9.8, SD 4.2 in patients with
EEG, p = 0.111) and a lower mortality (11.1% versus 19.7%,
p = 0.552) as compared to patients with an EEG. Fig. 1 presents
the study population and the patients’ categorization. Demograph-
ics, clinical features, and EEG characteristics are presented in
Table 1. The median Charlson comorbidity index and the propor-
tion of immunosuppressed patients differed significantly among
all three patient groups. Patients with infectious encephalitides
had the highest mean comorbidity index and were most frequently
immunosuppressed, while patients with encephalitis of unknown
cause had the lowest mean comorbidity index and the lowest pro-
portion of immunosuppression.
3.2. EEG among different types of encephalitis
EEGs were performed with a median of 1 day (inter quartile
range [IQR] 0.5–3) after admission. In 79% of patients EEGs were
recorded in the first 72 h after admission. All three EEG readers
remaining 8, subsequent 100% agreement was reached after
additional critical review. Although there were no significant
differences among the EEG characteristics, EEG background
activity was more often nonreactive in patients with
infectious encephalitides as compared to patients with other
encephalitides – a result not reaching statistical significance
(Table 1; p = 0.080). No specific EEG background frequency range
was associated with any GCS on the day of the EEG (p = 0.777),
not even after excluding 6 patients with continuously adminis-
tered anesthetic or sedating drugs prior to or during EEG. While
no clinical characteristics differed significantly between patients
with HSV encephalitis and non-HSV encephalitis, EEG interpreta-
tion revealed significantly higher proportions with PDs
(p = 0.029), and focal slowing in the fronto-temporal and occipital
regions (p = 0.017) in patients with HSV encephalitis, as compared
to patients with encephalitis from other infectious and non-infec-
tious causes (Table 2).
3.3. Outcome predictors
The mean length of ICU stay was 14.0 days (SD 16.5) and the
mean hospital stay 27.8 days (SD 18.3). We performed compari-
sons of clinical and EEG characteristics between survivors and
non-survivors in our cohort. Clinical predictors of mortality
included coma at admission, global cerebral edema and mechani-
cal ventilation (Table 3). The presence of a normal EEG was signif-
icantly associated with survival, as a normal EEG was found in 18
survivors and in none who died. Notably, none of the patients with

Table 2
Comparisons of clinical and EEG characteristics between HSV and non-HSV encephalitis (n = 76).

HSV encephalitis (n = 12) Non-HSV encephalitis (n = 64) p-Value#

Demographics
Gender, male (n, %) 4 33.3 35 54.7 0.217
Age, years (mean, SD) 57.9 ± 12.1 48.8 ± 17 0.079
Clinical features
GCS on admission (median, IQR) 12 9–14 10 6–13 0.071§
Comatose on admission, GCS 68 (n, %) 2 16.7 28 43.8 0.11
Charlson comorbidity index (median, IQR) 3 2–4 1.5 0–5 0.265§
Global cerebral edema (n, %) 2 16.7 8 12.7 0.657
Immunosuppression (n, %) 4 33.3 16 25 0.722
Mechanical ventilation (n, %) 5 41.7 42 65.6 0.194
EEG characteristics
Background frequency ranges (n, %) 0.259
Alpha 8 66.7 24 37.5
Alpha/theta 1 8.3 5 7.8
Theta 2 16.7 15 23.4
Theta/delta 0 0 14 21.9
Delta 1 8.3 6 9.4
Focal slowing (n, %) 0.129
Frontal 1 10 3 4.7
Temporal 2 20 3 4.7
Central 0 0 3 4.7
Parietal 0 0 1 1.6
Occipital 2 20 0 0 0.017
Episodic transients (n, %)
FIRDA 0 0 2 3.1
TWs 0 0 4 6.3
PDs 3 30.0 3 4.7 0.029
Epileptic activities (n, %)
Seizures 1 10 3 4.7
Status epilepticus 1 10 5 7.8
Nonreactive EEG background activity (n, %) 1 10 14 22.6 0.676

GCS = Glasgow Coma Scale; FIRDA = frontal intermittent rhythmic delta activity; TWs = triphasic waves; PDs = periodic discharges; EEG = electroencephalography;
SD = standard deviation; IQR = inter quartile range. Bold p-values are considered significant.
#
Fisher’s exact test.
§
Man–Whitney U-test.

Please cite this article in press as: Sutter R et al. Electroencephalography for diagnosis and prognosis of acute encephalitis. Clin Neurophysiol (2014), http://
dx.doi.org/10.1016/j.clinph.2014.11.006
 R. Sutter et al. / Clinical Neurophysiology xxx (2014) xxx–xxx 5

Table 3
Comparisons of early clinical and EEG characteristics between survivors and non-survivors with acute encephalitis (n = 76).

Survivors (n = 61) Non-survivors (n = 15) p-value#

Demographics
Gender, male (n, %) 29 47.5 10 66.7 0.252
Age, years (mean, SD) 49.1 ± 16.4 54.7 ± 17.3 0.248
Clinical features
GCS on admission (median, IQR) 11 7–14 6 3–11 0.01§
Comatose on admission, GCS 68 (n, %) 20 32.8 10 66.7 0.021
Charlson comorbidity index (median, IQR) 2 0–4 4 0–6 0.103§
Global cerebral edema (n, %) 5 8.3 5 33.3 0.023
Immunosuppression (n, %) 14 23 6 40 0.201
Mechanical ventilation (n, %) 33 54.1 14 93.3 0.006
EEG characteristics
Normal EEG (n, %) 18 29.5 0 0 0.01
Background frequency ranges (n, %) 0.965
Alpha 26 42.6 6 40
Alpha/theta 5 8.2 1 6.7
Theta 14 23 3 20
Theta/delta 11 18 3 20
Delta 5 8.2 2 13.3
Focal slowing (n, %) 0.192
Frontal 4 6.7 0 0
Temporal 5 8.3 0 0
Central 3 5 0 0
Parietal 1 1.7 0 0
Occipital 2 3.3 0 0
Episodic transients (n, %)
FIRDA 2 3.3 0 0
TWs 3 5 1 7.1
PDs 6 10 0 0 0.587
Epileptic activities (n, %)
Seizures 3 5 1 7.1
Status epilepticus 5 8.3 1 7.1
Nonreactive EEG background activity (n, %) 11 18.3 4 33.3 0.258

GCS = Glasgow Coma Scale; PDs = periodic discharges; EEG = electroencephalography; SD = standard deviation; IQR = inter quartile range. Bold p-values are considered
significant.
#
Fisher’s exact test.
§
Man–Whitney U-test.

Table 4
Multivariable Poisson regression analyses of clinical and EEG characteristics between
survivors and non-survivors with acute encephalitis (n = 76).
Death RR 95% CI
Comatose on admission, GCS 68 1.48 0.58–3.72
Global cerebral edema 2.25 1.17–4.33
Mechanical ventilation 4.98 0.67–36.83
Normal EEG <0.001 <0.001 <0.001
GCS = Glasgow Coma Scale; EEG = electroencephalography; RR = relative risks;
CI = confidence interval. Bold p-values are considered significant.
a normal EEG had a normal GCS at admission and on the day of EEG
(mean GCSadmission 11.3, SD 3.91; mean GCSduring EEG 10.0, SD 3.49).
Multivariable Poisson regression analyses revealed an independent
association of normal EEG with a low relative risk for death
(RR < 0.001, p < 0.001) after adjustment for possible confounders
such as coma, global cerebral edema and mechanical ventilation
(Table 4). The Pearson and deviance goodness-of-fit tests revealed
insignificant p-values indicating adequate model-fit (Pearson:
v2 = 197.65, p = 0.940; deviance: v2 = 97.91, p = 1.000). Abnormal
EEG had a 100% sensitivity (15/15) and a low specificity of 29.5%
(18/61) (positive predictive value 25.9% [15/58] and negative pre-
dictive value 100% [18/18]).
There were no EEG patterns associated with duration of hospital
and ICU stay or with mRS at discharge. EEG background reactivity
did not predict outcome – a result persisting after exclusion of 6
patients with continuously administered anesthetic or sedating
drugs prior to or during EEG (data not shown).
4. Discussion
Normal early EEG predicted survival in patients’ independent
p-Value
from important clinical outcome predictors, highlighting the prog-
0.411 nostic value of EEG in evaluating patients with acute encephalitis.
0.015
EEG revealed higher proportions of PDs along with focal slowing in
0.116
<0.001 HSV encephalitis as compared to other etiologies.
Although in many patients EEG patterns did not significantly
contribute to diagnosis and prognosis, all patients with a normal
EEG survived – an unimposing result at first glance. However,
the fact that none of these patients had a normal GCS, neither at
admission nor during EEG, indicates that EEG provides important
prognostic information beyond the patients’ cognitive conditions.
Further, our multivariable model reveals that established clinical
outcome-predictors, such as coma, global cerebral edema and
mechanical ventilation were not as robust as normal EEG (Thakur
et al., 2013; Whitley et al., 1987). This result is supported by a prior
EEG study of 47 patients with acute encephalitis, in which EEG
abnormalities correlated with the severity of illness (Illis et al.,
1972). The calculation of relative risks of survival based on the
presence of normal EEG is challenging in the presence of zero val-
ues and a small-sized population. In our multivariable regression
model the upper limit of the 95% confidence interval for normal
EEG approached infinity. We applied the rule of three to estimate
the upper limit of the 95% confidence interval, as previously rec-
ommended (Hanley and Lippman-Hand, 1983). Given our finding
that 0/18 patients with a normal EEG died, the rule of three indi-
cates that the upper limit of the 95% CI is 0.17, indicating that

Please cite this article in press as: Sutter R et al. Electroencephalography for diagnosis and prognosis of acute encephalitis. Clin Neurophysiol (2014), http://
dx.doi.org/10.1016/j.clinph.2014.11.006
6 R. Sutter et al. / Clinical Neurophysiology xxx (2014) xxx–xxx

Fig. 2. Typical EEG characteristics in patients with Herpes Simplex virus encephalitis. (A) Right-sided periodic discharges (EEG calibration: horizontal unit = 1 s, vertical
unit = 50 uV). (B) Right-sided periodic discharges with EEG background reactivity to stimuli mainly characterized by increased frequency range (EEG calibration: horizontal
unit = 1 s, vertical unit = 70 uV) EEG = electroencephalography.

the maximum risk of dying in the presence of a normal EEG is 17%.
The rule of three overestimates the risk in the setting of a total
sample size which is smaller than 30, so that our estimation is
prudent.
EEG background activity was more often nonreactive in patients
with infectious encephalitis as compared to patients with other
encephalitides – an association that did not reach significance in
our cohort. In addition, unreactive EEG did not differ significantly
between survivors and non-survivors. These results contrast prior
EEG studies in patients with encephalopathies that are caused by
more diffuse or generalized structural abnormalities or neurofunc-
tional alterations from cardiac arrest, systemic metabolic derange-
ments, or intoxication (Rossetti et al., 2010; Sutter and Kaplan,
2013a,b; Sutter et al., 2013a,b), in which EEG background reactivity
predicts outcome. This may be partially explained by the fact that
early encephalitis usually represents an inflammatory or infectious
process restricted to certain cerebral regions possibly sparing cir-
cuits involved in central neuronal reactivity to specific stimuli.
EEG characteristics did not differ significantly across the differ-
ent etiologic categories of encephalitis (i.e., infectious, autoim-
mune and unknown). While there were no significant clinical
differences between patients with HSV and non-HSV encephaliti-
des, EEG interpretation revealed significantly higher proportions
of PDs and focal slowing in the fronto-temporal and occipital
regions in HSV encephalitis. These results underscore the findings
of prior case series and smaller studies (Al-Shekhlee et al., 2006;

Please cite this article in press as: Sutter R et al. Electroencephalography for diagnosis and prognosis of acute encephalitis. Clin Neurophysiol (2014), http://
dx.doi.org/10.1016/j.clinph.2014.11.006
 R. Sutter et al. / Clinical Neurophysiology xxx (2014) xxx–xxx
Brodtkorb et al., 1982; Ch’ien et al., 1977; Illis et al., 1972; Lai et al.,
1988; Upton et al., 1970). The proportion of PDs in our cohort was
lower than in other reports, possibly explained by the fact that in
our cohort, EEGs were performed earlier in the course of disease
as compared to prior studies. Analyses from other series of patients
with HSV encephalitis revealed that PDs were mostly detected
between days 2 and 15 but still before typical structural changes
appeared on neuroimaging (Brodtkorb et al., 1982; Lai et al.,
1988; Upton et al., 1970). Fig. 2A illustrates unilateral PDs and
Fig. 2B unilateral PDs with reactive EEG background activity to
acoustic (call) and noxious stimuli. The EEG in the acute stage of
HSV encephalitis can show a variety of abnormalities, including
uni- or bilateral focal slowing, PDs, or electrical seizures (Lai
et al., 1988). In a previous study of the diagnostic yield of EEG in
HSV encephalitis EEG-patterns did not predict survival (Lai et al.,
1988). This differs from our findings at first glance; however, anal-
yses of the predictive value of normal EEG early in the course of the
disease were not performed and conclusions were drawn mainly
for EEG abnormalities. Further, the authors stated that of five
patients with EEGs that became normal, two had excellent recov-
ery and three survived with persistent disabilities (Lai et al., 1988).
The strength of our study lies in the fact that this is a larger
cohort including patients with different etiologies of acute enceph-
alitis, as compared to most previous studies, case series and case
reports, thereby strengthening the current assumptions and con-
clusions. Not finding additional EEG patterns of diagnostic or prog-
nostic relevance aside from PDs and focal slowing in HSV
encephalitis (except normal EEG) need to be reported, as they
assure EEGers and clinicians about the insignificance of specific
patterns seen in encephalitic patients and thereby allow more
focused analyses of patterns confirmed to have diagnostic and
prognostic impact. Furthermore, our results reveal that in contrast
to EEG patterns, none of the clinical characteristics of our patients
had any diagnostic or prognostic value, underscoring the diagnos-
tic and prognostic yield of EEG in this setting.
5. Limitations
The limited sample size, the retrospective design and lack of fol-
low-up or prolonged EEGs may hamper the generalizability of this
study. However, the latter should not limit the strength of the
results, as we aimed to identify the early predictive value of EEG
characteristics on the first recording. Longer EEG monitoring might
reveal other or more diagnostic findings – a hypothesis that
remains to be proven in future studies in which continuous video
EEG monitoring is used. As during the time period in which the
study was performed (between 1997 and 2011), some of the etio-
logical classification of encephalitis has changed particularly with
the recognition of new forms of autoimmune encephalitis (e.g.,
anti-NMDA receptor encephalitis), it seems likely that some of
the 37 patients with the diagnosis of acute encephalitis of
unknown etiology may have had an autoimmune encephalitis. This
limits our analyses especially in patients with autoimmune
encephalitis, calling for further studies in this context.
6. Conclusions
Normal early EEG predicted survival in patients with acute
encephalitis and altered level of consciousness independent from
clinical outcome predictors highlighting the prognostic value of
EEG in evaluating patients with acute encephalitis. While early
EEG patterns did not differ significantly among patients with
encephalitides from infectious, autoimmune, or unknown causes,
more EEGs were nonreactive to stimuli and revealed higher pro-
portions of PDs along with fronto-temporal and occipital slowing
Please cite this article in press as: Sutter R et al. Electroencephalography for diagnosis and prognosis of acute encephalitis. Clin Neurophysiol (2014), http://
dx.doi.org/10.1016/j.clinph.2014.11.006
7
in patients with HSV encephalitis. Further prospective studies are
needed to characterize the diagnostic and prognostic value of
EEG in acute encephalitis.
Funding
No funding was received for this study. This study was per-
formed and designed without the input or support of any pharma-
ceutical company, or other commercial interest. Dr. R. Sutter was
supported by the Research Fund of the University of Basel, the Sci-
entific Society Basel, and the Gottfried Julia Bangerter-Rhyner
Foundation. Dr. M. Cervenka received funding from the Johns Hop-
kins University School of Medicine Clinician Scientist Award, Nutri-
cia, and NIH (NINDS R01NS075020).
Acknowledgements
The corresponding author had full access to all of the data in the
study and takes responsibility for the integrity of the data and the
accuracy of the data analysis.
We thank Dr. S. Tschudin-Sutter, MD, MSc, University Hospital
Basel, Switzerland, for her statistical support. Dr. R. Sutter has full
access to all of the data in the study and takes responsibility for the
integrity of the data and the accuracy of the data analysis.
Conflict of interest: None of the authors have any conflicts of
interest related to this study.
References
Al-Shekhlee A, Kocharian N, Suarez JJ. Re-evaluating the diagnostic methods in
Herpes Simplex encephalitis. Herpes 2006;13:17–9.
American Sleep Disorders Association. EEG arousals: scoring rules and examples: a
preliminary report from the Sleep Disorders Atlas Task Force of the American
Sleep Disorders Association. Sleep 1992;15:173–84.
Brodtkorb E, Lindqvist M, Jonsson M, Gustafsson A. Diagnosis of Herpes Simplex
encephalitis. A comparison between electroencephalography and computed
tomography findings. Acta Neurol Scand 1982;66:462–71.
Ch’ien LT, Boehm RM, Robinson H, Liu C, Frenkel LD. Characteristic early
electroencephalographic changes in Herpes Simplex encephalitis. Arch Neurol
1977;34:361–4.
Chatrian GE, Shaw CM, Leffman H. The significance of periodic lateralized
epileptiform discharges in EEG: an electrographic, clinical and pathological
study. Electroencephalogr Clin Neurophysiol 1964;17:177–93.
Davies NW, Sharief MK, Howard RS. Infection-associated encephalopathies: their
investigation, diagnosis, and treatment. J Neurol 2006;253:833–45.
Granerod J, Ambrose HE, Davies NW, Clewley JP, Walsh AL, Morgan D, Cunningham
R, Zuckerman M, Mutton KJ, Solomon T, Ward KN, Lunn MP, Irani SR, Vincent A,
Brown DW, Crowcroft NS. Causes of encephalitis and differences in their clinical
presentations in England: a multicentre, population-based prospective study.
Lancet Infect Dis 2010a;10:835–44.
Granerod J, Tam CC, Crowcroft NS, Davies NW, Borchert M, Thomas SL. Challenge of
the unknown. A systematic review of acute encephalitis in non-outbreak
situations. Neurology 2010b;75:924–32.
Hanley JA, Lippman-Hand A. If nothing goes wrong, is everything all right?
Interpreting zero numerators. JAMA 1983;249:1743–5.
Illis LS, Taylor FM. The electroencephalogram in Herpes-Simplex encephalitis. Lancet
1972;1:718–21.
Koskiniemi M, Manninen V, Vaheri A, Sainio K, Eistola P, Karli P. Acute encephalitis.
A survey of epidemiological, clinical and microbiological features covering a
twelve-year period. Acta Med Scand 1981;209:115–20.
Lai CW, Gragasin ME. Electroencephalography in Herpes Simplex encephalitis. J Clin
Neurophysiol 1988;5:87–103.
Ozuna J. Seizure disorders and epilepsy. Lippincotts Prim Care Pract 2000;4:608–18.
Rossetti AO, Oddo M, Logroscino G, Kaplan PW. Prognostication after cardiac arrest
and hypothermia: a prospective study. Ann Neurol 2010;5:161–74.
Sutter R, Kaplan PW. Clinical and electroencephalographic correlates of acute
encephalopathy. J Clin Neurophysiol 2013a;30:443–53.
Sutter R, Kaplan PW. Electroencephalographic criteria for nonconvulsive status
epilepticus: synopsis and comprehensive survey. Epilepsia 2012;53(Suppl
3):1–51.
Sutter R, Kaplan PW. Neuroimaging correlates of acute encephalopathy. J Clin
Neurophysiol 2013b;30:517–25.
Sutter R, Stevens RD, Kaplan PW. Clinical and imaging correlates of EEG patterns in
hospitalized patients with encephalopathy. J Neurol 2013a;260:1087–98.
Sutter R, Stevens RD, Kaplan PW. Significance of triphasic waves in patients with
acute encephalopathy: a nine-year cohort study. Clin Neurophysiol
2013b;124:1952–8.
8 R. Sutter et al. / Clinical Neurophysiology xxx (2014) xxx–xxx

Tenembaum S, Chitnis T, Ness J, Hahn JS. International Pediatric MSSG. Acute
disseminated encephalomyelitis. Neurology 2007;68:S23–36.
Thakur KT, Motta M, Asemota AO, Kirsch HL, Benavides DR, Schneider EB, et al.
Predictors of outcome in acute encephalitis. Neurology 2013;81:793–800.
Upton A, Gumpert J. Electroencephalography in diagnosis of Herpes-Simplex
encephalitis. Lancet 1970;1:650–2.
Venkatesan A, Tunkel AR, Bloch KC, Lauring AS, Sejvar J, Bitnun A, et al. Case
definitions, diagnostic algorithms, and priorities in encephalitis: consensus
statement of the international encephalitis consortium. Clin Infect Dis
2013;57:1114–28.
Whitley RJ, Alford CA, Hirsch MS, Schooley RT, Luby JP, Aoki FY, et al. Factors
indicative of outcome in a comparative trial of acyclovir and vidarabine for
biopsy-proven Herpes Simplex encephalitis. Infection 1987;15(Suppl 1):S3–8.
Whitley RJ, Lakeman F. Herpes Simplex virus infections of the central nervous
system: therapeutic and diagnostic considerations. Clin Infect Dis
1995;20:414–20.
Zou G. A modified Poisson regression approach to prospective studies with binary
data. Am J Epidemiol 2004;159:702–6.

Please cite this article in press as: Sutter R et al. Electroencephalography for diagnosis and prognosis of acute encephalitis. Clin Neurophysiol (2014), http://
dx.doi.org/10.1016/j.clinph.2014.11.006