Introduction
Acute altered mental status (AAMS) is a general term used to describe the undifferentiated
presentation of a group of disorders characterized by alterations of consciousness,
with a rapid onset of symptoms. Consciousness itself can be defined by both a quantitative
and a qualitative component. Arousal is the quantitative component and refers to the
overall state of alertness (or wakefulness); it is related to the integrity of ponto-mesodiencephalic
reticular pathways and thalamocortical projections. On the other hand, awareness can
be defined as the qualitative component of consciousness and refers to a subjective
experience, which is the basis for any significant interaction with the environment.[1 ]
[2 ]
[3 ]
A new-onset altered mental status (AMS) is a potentially life-threatening emergency
and, thus, a common reason for admission to the emergency department (ED). It can
represent a diagnostic challenge, especially in children, since it is a nonspecific
diagnosis that encompasses a wide spectrum of diseases with different underlying etioliogies.[4 ] To date, only few studies[3 ]
[4 ]
[5 ] concerning the diagnosis and management of AAMS have been published. The most common
underlying etiologies of AAMS in pediatric patients seem to be neurologic, followed
by toxicological and infectious.[6 ] Several diagnostic tests, such as computed tomography (CT), magnetic resonance imaging
(MRI), and electroencephalography (EEG), are often used to identify the underlying
cause of AAMS.[7 ] Interestingly, EEG findings in children with AAMS will differ based on the specific
underlying etiology. This is because EEG is the graphic representation of the synchronous
activity of a large aggregate of neurons (field potentials). Specifically, it measures
the difference in electrical potential between different areas of the cerebral cortex
where the current flow is generated by synaptic activation of the dendrites of the
pyramidal neurons. Although it is not to date possible to give a completely exhaustive
explanation of the origin of the rhythms we detect in EEGs, the dorsal thalamus is
considered the primary generator of the subcortical EEG rhythm, as it regulates the
synchronization of cortical postsynaptic potentials during sleep and wakefulness.
The thalamus therefore plays a fundamental role in the genesis of the alpha rhythm
(or posterior dominant rhythm), particularly because it is the anatomo-functional
station of ascent of the ascending reticular activation system, closely related to
arousal levels; the origin of slow cerebral rhythms (delta activity) must instead
be sought at least two levels, thalamic and cortical.
Based on these premises, it is easy to understand how alteration of the state of consciousness
causes EEG changes and to give some specific examples. Cerebral inflammations, together
with metabolic and toxic disorders, can cause diffuse, nonspecific EEG slowing; on
the other hand, an abnormal electrical discharge (alterations of nerve cell excitability,
which depend on ionic movements across the membrane through ion channels) originating
within one small group of neurons limited to one hemisphere, causes an initially localized
epileptiform EEG pattern.[8 ]
[9 ]
Published data on the role of EEG in AAMS of unknown etiology are to date lacking,
especially in the pediatric population. To date, no prospective studies focusing exclusively
on the pediatric population are available.
The aim of our study was to evaluate the role of EEG in the management of children
with AAMS of unknown etiology. We conducted a prospective study and enrolled children
with an AAMS diagnosis who were seen in the pediatric ED or who received inpatient
care. To complete our research, a review of the existing literature on the use of
EEG in the diagnosis of AAMS of unknown etiology was also performed.
Materials and Methods
Literature Review
A literature search was performed in the PubMed database up to January 2022. The following
search terms were used: “altered mental status,” “altered state of consciousness,”
“consciousness impairment,” “EEG AND emergency,” and “EEG AND altered mental status.”
The same search terms were input in August 2022 to check for any additional papers.
Studies were included if they were published in English, after January 1990, and if
their main focus was acute and persistent AMS with an unknown etiology—meaning the
etiology could not be inferred from patient history, clinical presentation, or initial
rapid assessments. Studies were excluded if they were case reports, reviews, or if
they were not pertinent to our literature search (e.g., describing an altered state
of consciousness with a specific etiology).
Design and Patients
In this monocentric study, we enrolled children presenting with AAMS to the pediatric
ED or who received inpatient care in the pediatric department from May 2017 to October
2019.
All children were first evaluated by a pediatrician and subsequently by a child neurologist,
who established the diagnosis of AAMS of unknown etiology.
The inclusion criteria were: (1) age between 1 month and 18 years; (2) acute (<1 week)
and persistent (>5 minutes) AMS of unknown etiology defined by at least one of the
following: (i) pediatric Glasgow Coma Scale (pGCS) ≤13, (ii) difficulty or inability
in maintaining vigilance, (iii) confusion and disorientation, (iv) bizarre and inappropriate
behavior, (v) hallucinations.
We believe our criteria allowed us to assess both the quantitative component of consciousness
(through criterion i, the GCS score) and the qualitative component (measured through
criteria ii to v).
Signs/symptoms or a recent history which was suggestive of a specific underlying diagnosis
excluded the patient from the study (e.g., syncope, tonic-clonic seizures, hypoglycemia
or hyperglycemia, recent head trauma, ingestion of sedative drugs, etc.).
Once a diagnosis of AAMS was established, blood tests and toxicological examinations
were performed as part of standard care protocols. Lumbar puncture (LP) and neuroimaging
(CT or MRI) were carried out in selected patients.
Video-EEG was performed on all children on a digital acquisition system (MicroMed)
using caps with 20 permanent electrodes placed according to the 10–20 Standard International
System. Duration was at least 20 to 30 minutes. In patients with nonconvulsive status
epilepticus (NCSE), video-EEG was performed continuously during treatment and until
the complete resolution of symptoms. EEG was acquired using a bipolar montage with
the following filter settings: a low-pass filter at 50 to 70 Hz, a high-pass filter
at 1.0 to 1.6 Hz, and a sensitivity of 100 to 200 Hz. A notch filter was also applied.
The ECG trace was collected for all patients, and polygraphy was acquired as needed
based on clinical manifestations. The recordings were then reviewed by an expert child
neurologist specializing in EEG reading (D.M.C.).
EEGs were subsequently classified as ictal or non-ictal. For non-ictal EEGs, we described:
(1) the organization and symmetry of background activity; (2) interictal discharges;
(3) localized or diffuse slowing; and (4) any specific patterns.
EEGs were classified into three groups based on the role they played in contributing
to the final diagnosis and in assisting further management and treatment:
“Diagnostic,” indicating that the EEG directly contributed to a conclusive diagnosis
(seizure/NCSE/normal EEG despite the altered state of consciousness; in some cases,
neurological signs were also present).
“Informative,” in those cases in which the EEG, together with the patient's clinical
presentation, suggested a presumptive diagnosis, which however required further investigation
to be confirmed (e.g., slowing may suggest encephalitis, but a LP or MRI is essential
to reach a final diagnosis).
“Poorly informative,” indicating all those EEGs that together with the patient's clinical
presentation were not immediately suggestive for a specific etiology. This group included
instances in which repeat EEGs performed during the patient's hospitalization were
of help in guiding the diagnostic process (e.g., a certain degree of slowing, with
subsequent normalization, was suggestive of a migraine episode).
Data and Statistics
Demographic data (age, sex), medical history (developmental milestones, history of
febrile convulsions, medical comorbidities, and treatments), symptoms at the time
of AAMS diagnosis, findings at neurological examination and results from blood tests,
toxicological tests, LP, and neuroimaging were collected.
Continuous variables were presented as median and range, while qualitative data were
described as frequencies and percentages. Due to the small sample size, correlation
analyses for quantitative and categorical variables were not performed.
Ethics Committee Approval
The study was approved by an independent ethics committee (Protocol number 132/2017/O/Oss,
code NPI-EEG-01, May 16, 2017). Parents/legal guardians provided written informed
consent.
Results
Literature Review
Our search strategy identified 53 potentially relevant articles. We excluded 37 studies
that did not meet inclusion criteria. Ultimately, we focused on 16 articles, which
we categorized into two groups, adults and children, as detailed in [Table 1 ].[6 ]
[10 ]
[11 ]
[12 ]
[13 ]
[14 ]
[15 ]
[16 ]
[17 ]
[18 ]
[19 ]
[20 ]
[21 ]
[22 ]
[23 ]
[24 ] Most of the selected studies reported data on EEG monitoring during symptomatic
episodes, although one study did not report EEG results[12 ] and another did not use EEG at all.[6 ]
Table 1
Literature review
Article
(authors, year)
Type of study
[P: prospective
R: retrospective]
Number of patients included
Age
(in years)
Setting
Acute, persistent AMS
[+/− ]
Etiology:
[known/unknown/both]
EEG performed during AMS symptoms:
[+/− ]
Relevant results
Study population consisting of adults and children
1)
Praline et al,
2007[10 ]
P descriptive study
111
(16 children,
95 adults)
Children, mean: 3.4
Adults, mean: 53
ED
+
Both
+
EEG was considered contributory to the diagnosis in 77.5% of cases. It helped confirm
a clinically suspected diagnosis in 36.0% and to rule it out in the remaining 64.0%.
The results of the EEG led to a modification in treatment in 37.8% of cases.
Study population consisting mainly of adults
2)
Ziai et al,
2012[11 ]
Single-center P cohort intervention study
82
Mean: 58.1
(± 2)
ED
+
Both
+
EEG aided the diagnosis in 51.0%, changed ED management in 4.0%, and would be ordered
again if EEG was available in 46.0%. An abbreviated 5-minute full-montage EEG presents
adequate reliability which may improve its use in the ED.
3)
Duran et al,
2014[12 ]
R study
190
Mean: 47.6
(range: 18–87)
ED
+
Unknown
Unspecified
EEG is still an important tool in the differential diagnosis of AMS such as epileptic
seizures, metabolic abnormalities, pseudo-seizures, and syncope.
4)
Newey et al,
2018[13 ]
R study
1,123
Mean: 57.3
(± 18.2)
ICU
+/−
Both
+/−
Continuous EEG monitoring in the critical care setting demonstrates a linear increase
in seizure incidence with declining mental status.
5)
Kimchi et al,
2019[14 ]
P observational cohort study
200
Median: 60
(IQR: 48.5–72)
Medical, surgical,
neurologic floors, and ICUs
+
Unknown
+
Generalized slowing on EEG strongly correlates with delirium and may be a valuable
biomarker for delirium severity; it should trigger elevated concern for the prognosis
of patients with AMS.
6)
Egawa et al,
2020[15 ]
P study with R examination
50
Median: 72 (IQR: 52–80)
Neuro-ICU
+
Unknown
+
HS-cv EEG monitoring demonstrated high reliability for the detection of abnormal EEG
patterns, with moderate reliability for PDs and NCSE, and can facilitate rapid initiation
of cEEG monitoring in patients with AMS of unknown etiology.
7)
Rossetti et al,
2020[16 ]
Multicenter randomized clinical trial
364
Mean: 63
(SD: 15)
ICU and intermediate care unit
+
Both
+
In critically ill adults with AMS and no recent seizures, cEEG leads to increased
seizure detection and modification of ASMs but is not related to improved outcome
compared with repeated rEEG.
8)
Müller et al,
2020[17 ]
P randomized trial retrospectively analyzed
364
Median: 67 (IQR: 55–75)
ICU
+
Unknown
+
EEG background reactivity was the most important EEG feature, both for predicting
survival and a favorable functional outcome. The study supports the fundamental role
of EEG as a prognostic tool
9)
Nakae et al,
2021[18 ]
R study
33
Mean: 67
(range: 20–96)
ED
+
Known
+
Among EEG patterns, PDs were detected only in the delayed recovery group. Diffuse
white matter degeneration around the lateral ventricles contributes to prolonged AMS.
Study population consisting mainly of children
10)
Alehan et al,
2001[19 ]
R descriptive study
56
Mean: 6.6
(range: 2 days–18 years)
ED
+
Both
+
EEG directly contributed to the diagnosis in 84.0% of patients. A prompt EEG should
be considered in children with new-onset seizures and unexplained altered consciousness.
11)
Kothare et al,
2005[20 ]
R descriptive study
32
Mean: 4.5
(range: 0.014–17)
ED
+/−
Both
+/−
Emergency EEGs were useful in decision making in 94.0% of cases. Emergency EEGs and
emergency long-term bedside EEGs during nonbusiness hours have a high yield of offering
useful information that influences treatment decisions.
12)
Fernández et al, 2014[21 ]
R descriptive study
68
Mean: 7.3
(range: 0.096–21)
ED
+/−
Both
+/−
59.1% of individuals were sent home mainly based on the results of the emergency EEG.
These data support the use of an EEG in the pediatric ED, especially for those children
in which ongoing seizures or status epilepticus are suspected.
13)
Button et al,
2018[6 ]
R chart review
336
Mean: 9
(SD: 6)
ED
+
Both
Not performed
Common causes of pediatric AMS include neurologic, toxicological, and infectious,
but often an underlying diagnosis is not found, despite several diagnostic tests.
14)
Yamaguchi et al, 2019[22 ]
R cohort study
206
Median: 3.6 (IQR: 1.8–5.9)
ED
+
Unknown
+
16.9% of patients who underwent EEG had NCS (4.4% of patients with AMS in the ED).
The use of reduced-lead EEG monitoring in the ED might facilitate the recognition
and treatment of NCS.
15)
Simma et al,
2021[23 ]
R observational study of
prospectively collected data
36
Median: 2.8
(range: 0.75–15)
ED
+
Both
+
Physicians found EEG to be “very useful/diagnostic” in 13 cases (36.0%), “useful”
in 21 cases (58.0%), “not useful” in 2 cases (8.0%). The rapid 2-channel-EEG-monitoring
improved recognition of NCSE and facilitated more targeted interventions.
16)
Gunawardena et al,
2022[24 ]
Single-center R study
162
Mean: 7.8
(SD: ± 5.8)
ED
+
Both
+/−
In 87.7% EEG was helpful in confirming or ruling out the suspected diagnosis. When
suspecting new-onset seizures, recurrent seizures, acute AMS, and psychogenic seizures,
EEG was useful in 91.1, 81.3, 81.8, and 100%, respectively. For the 142 patients in
whom routine EEG was diagnostically useful, 59.9% were admitted, compared to 95.0%
of the 20 patients in whom EEG did not help in clarifying the diagnosis (p = 0.002).
Abbreviations: AMS, altered mental status; ASMs, antiseizure medications; cEEG, continuous
EEG; ED, emergency department; HS-cv, headset-type continuous video; ICUs, intensive
care units; NCS, nonconvulsive seizure; NCSE, nonconvulsive status epilepticus; PDs,
periodic discharges; rEEG, routine EEG.
We identified six prospective studies conducted primarily on adults, with the exception
of Praline et al,[10 ] which included 16 children. Additionally, we found 10 retrospective studies, with
3 focusing on adults and 7 on children. These studies were predominantly conducted
in the ED and intensive care unit (ICU) .
In the only prospective study reporting data on both adult and pediatric patients,[10 ] EEG was found to contribute to the diagnosis in 77.5% of cases. It confirmed a clinically
suspected diagnosis in 36.0% and ruled it out in the remaining 64.0%, leading to treatment
modifications in 37.8% of cases.
Among studies involving pediatric patients only, EEG aided the diagnosis and decision
making in 84.0 and 94.0% of cases, respectively[19 ]
[20 ]; in another study, EEG was defined as “very useful/diagnostic” in 36.0%, “useful”
in 58.0%, and “not useful” in 8.0%.[23 ]
Additionally, Fernández et al reported that 59.1% of patients were discharged based
on emergency EEG findings.[21 ] Gunawardena et al[24 ] emphasized the role of EEG in avoiding unnecessary hospitalizations: among patients
for whom routine EEG was diagnostically useful (142 patients, 87.7%), 59.9% were admitted,
compared to 95.0% of the 20 patients in whom EEG did not help clarify the diagnosis
(p = 0.002).
Some authors evaluated the reliability of reduced-lead EEG monitoring to assess whether
it could be a useful option in emergency settings. Yamaguchi et al[22 ] utilized a portable digital EEG system with four channels across bilateral frontal
and occipital regions, identifying NCSE in 4.4% of AMS patients in the ED. Similarly,
Simma et al[23 ] found that rapid two-channel-EEG monitoring improved NCSE recognition and enabled
more targeted interventions. Finally, according to Egawa et al,[15 ] headset-type continuous video (HS-cv) EEG monitoring (eight channels) had high reliability
for detecting abnormal EEG patterns, with moderate reliability for periodic discharges
and NCSE, potentially facilitating the rapid initiation of continuous EEG monitoring
in patients with AMS of unknown etiology. On the other hand, Ziai et al[11 ] compared standard EEG with a 5-minute full-set EEG and found that the reduced-lead
EEG had adequate reliability, potentially improving its utility in the ED.
In conclusion, several authors have explored the topic of determining the cause of
AAMS, and some of them have reported that EEG can be a useful diagnostic aid, thus
demonstrating the relevance of the topic of our article. To date, however, the only
prospective studies performed in pediatric populations have focused exclusively on
patients admitted to the ED. Additionally, there is a lack of studies evaluating how
an EEG performed when the patient is symptomatic can help determine the diagnosis
of AMS of unknown etiology.
The Study
Population
We enrolled 20 patients, 9 males and 11 females. Median age was 7.7 (range: 0.5–15.4)
years.
Most patients did not have a history of developmental delay (n = 14, 70%) or febrile seizures (n = 17, 85%). Eleven (55%) patients showed no comorbidities, whereas 4 (20%) had a
history of seizures (TBC1D24-related epilepsy, hypoxic-ischemic encephalopathy, recent
onset of focal epilepsy of unknown origin, and NCSE in acute lymphoblastic leukemia
[ALL] complicated by posterior reversible encephalopathy syndrome [PRES]), 4 (20%)
had a neurological disorder without epilepsy (dyskinetic encephalopathies, hypoxic-ischemic
encephalopathy with a subsequent ventriculoperitoneal shunt, Prader–Willi syndrome,
and tension headache), and 2 (10%) had an onco-hematological disease (neutropenia
and ALL complicated by PRES).
Fourteen patients (70%) were not prescribed any medications; only four patients (20%)
were on antiseizure medications (ASMs), one of whom was also taking methotrexate (MTX)
and cytarabine for ALL; the other two patients (10%) were treated with other drugs
prescribed for their movement disorder and headache.
Demographic and clinical features of the population are reported in [Table 2 ].
Table 2
Demographic and clinical features
ID, sex
Age
(mo)
Significant past medical history
Signs/symptoms preceding onset of AMS
pGCS
AMS
presentation
New-onset neurological signs
1, M
6 mo
Neutropenia
Hyporexia
pGCS = 10
Inconsolable crying, poor response to stimuli
Tonic deviation of the eyes
2, F
8 mo
None
Paroxysmal events[a ]
Vomit
Hyporexia
pGCS = 11
Unresponsiveness
Tonic deviation of the eyes
3, M
20 mo
None
Vomit
Fever
pGCS = 6
Unresponsiveness
Tonic deviation of the eyes and oro-alimentary automatisms
4, F
23 mo
Developmental encephalopathy
Hyporexia
pGCS = 12
Annoyed,
poor response to stimuli
No
5, F
2 y, 9 mo
PWS
Febrile seizure
Vomit
Fever
pGCS = 10
Drowsy,
poor response to stimuli
No
6, M
3 y, 6 mo
None
Febrile seizures
Vomit, Hyporexia
Fever
pGCS = 15
Difficulty in maintaining vigilance
Upper limb dyskinesia, clonus
7, M
4 y, 1 mo
Hydrocephalous in HIE
None
pGCS = 6
Unresponsiveness
Tonic deviation of the eyes
8, M
4 y, 7 mo
Seizures
Hyporexia
pGCS = 12
Annoyed,
poor response to stimuli
Axial hypotonia
9, F
5 y, 7 mo
TBC1D24 -related epilepsy
Febrile seizures
Hyporexia
Fever
pGCS = 14
Bizarre/inappropriate behavior
Blinking, tongue dyskinesia, hypotonia
10, F
6 y, 2 mo
None
None
pGCS = 12
Drowsy,
poor response to stimuli
No
11, F
9 y
None
None
pGCS = 14
Bizarre/inappropriate behavior
No
12, F
9 y, 8 mo
None
Headache
Vomit
pGCS = 14
Confused, disoriented
No
13, M
10 y, 2 mo
None
Hyporexia
pGCS = 14
Confused, disoriented
No
14, M
10 y, 2 mo
None
Headache
pGCS = 14
Confused, disoriented
Hypotonia, trismus
15, F
10 y, 9 mo
PRES in ALL
None
pGCS = 12
Drowsy,
poor response to stimuli
Facial nerve palsy, dysphonia, asymmetric hypotonia
16, F
11 y, 4 mo
None
None
pGCS = 14
Confused, disoriented
Ascendant upper limb hypoesthesia
17, M
12 y, 7 mo
None
None
pGCS = 3
Unresponsiveness
No
18, M
14 y
HIE, structural epilepsy
Hyporexia
pGCS = 7
Unresponsiveness
No
19, F
14 y, 9 mo
Headache
None
pGCS = 8
Unresponsiveness
No
20, F
15 y, 4 mo
None
Hyporexia
pGCS = 11
Poor response to stimuli
Aphasia
Abbreviations: ALL, acute lymphoblastic leukemia; AMS, altered mental status; HIE,
hypoxic-ischemic encephalopathy; mo, months; pGCS, pediatric Glasgow Coma Scale; PRES,
posterior reversible encephalopathy syndrome; PWS, Prader–Willi Syndrome; y, years.
a Patient 2 presented two episodes of reduced awareness and deviation of the eyes on
the right side.
AMS Features
In the week prior to onset, transient neurological signs (n = 9, 45%) and vomiting and feeding difficulties (n = 2, 10%) were reported. Fever (n = 5, 25%) and hyporexia (n = 7, 35%) were also detected.
All patients met the inclusion criteria as follows: (1) 13 patients had a pGCS ≤13,
(2) 1 patient had difficulty or inability in maintaining vigilance, (3) 4 patients
were confused and disoriented, and (4) 2 patients had bizarre and inappropriate behaviors.
No patients had hallucinations.
Thus, 13 patients (65.0%) were diagnosed with AAMS because of an impairment of the
quantitative component (inclusion criterion 1) while 7 (35%) were diagnosed based
on an impairment of the qualitative component (inclusion criteria 2—4). Three patients
who initially only showed an impairment in the qualitative component of consciousness
subsequently deteriorated and their pGCS score dropped below 13.
The median pGCS for our cohort was 12 (range: 3–14). Eleven patients (55.0%) had a
new onset of neurological signs. Data on AMS features are summarized in [Table 2 ].
Diagnostic Assessments
The roles of EEG and neuroimaging in defining the etiology of AMS are outlined in
[Table 3 ]. EEG was the first diagnostic test to be carried out in 13 patients (65.0%); in
the other patients, neuroimaging was performed before EEG, both CT scan (n = 6, 30.0%) and MRI (n = 1). Additionally, in one patient, LP was also performed before EEG.
Table 3
Diagnostic assessments (EEG and neuroimaging) and their contribution to ascertaining
AMS etiology
ID
Suspected diagnosis
EEG
EEG contribution
Diagnostic assessment
Contribution of diagnostic assessment
Final diagnosis
1
NCSE
Poorly organized background
Informative
CT/MRI
Diagnostic MRI
Oculogyric crisis in VWM
2
NCSE
Rhythmic S/S-W over l-TPO
Diagnostic
MRI
Diagnostic MRI
NCSE in SWS
3
NCSE
Rhythmic theta activity over l-O
Diagnostic
CT/MRI
None
NCSE in SeLEAS
4
Disease progression
High amplitude, biphasic, slow waves with superimposed fast activity
Diagnostic
MRI
None
Epileptic spasms
5
Encephalopathy
Diffuse slowing with posterior slow W
Poorly informative
MRI
None
Nonconvulsive
seizure
6
Onco-hematological disease
Diffuse monomorphic slowing
Informative
LP[a ], CT[a ], MRI
Diagnostic MRI
ADEM
7
Failure of VP shunt
Rhythmic S-W over l-O
Diagnostic
CT, MRI
None
NCSE
8
Epilepsy
relapse
Asymmetric slowing and delta brush
Informative
LP, MRI
Diagnostic LP and MRI
NMDA-r
9
NCSE
Bifrontal S-W
Poorly informative[b ]
MRI
None
Disease progression
10
Migraine
NCSE
Monomorphic delta over the r-posterior hemisphere
Poorly informative[b ]
None
–
Nonconvulsive seizure
11
Encephalitis
Generalized rhythmic poli-S, SW
Diagnostic
CT[a ], MRI
None
NCSE
12
Encephalitis
Migraine
Bifrontal slow W
Informative
None
–
Migraine
13
Migraine
NCSE
l-posterior polymorphic delta
Poorly informative[b ]
CT[a ], MRI
None
Migraine
14
Encephalitis
Normal background
Diagnostic
CT[a ]
None
Psychogenic event
15
Disease progression
Normal background
Poorly informative
MRI[a ]
Diagnostic MRI
MTX encephalopathy
16
CVE
Migraine
NCSE
l-posterior polymorphic delta
Poorly informative[b ]
CT, MRI
None
Migraine
17
Coma
Normal background
Diagnostic
None
–
Psychogenic event
18
NCSE
Asymmetric background (low voltage in l-hemisphere) with asynchronous S,S-W over CPT[c ]
Informative
None
–
CBZ intoxication
19
Coma
Normal background
Diagnostic
CT[a ]
–
Psychogenic event
20
CVE
CNS tumor
Normal background
Diagnostic
CT[a ]
–
Psychogenic event
Abbreviations: ADEM, acute disseminated encephalomyelitis; AMS, altered mental status;
CBZ, carbamazepine; CPT, centro-parieto-temporal; CT, computed tomography; CVE, cerebrovascular
events; l-O, left occipital; LP, lumbar puncture; l-TPO, left temporo-parieto-occipital;
MRI, magnetic resonance imaging; MTX, methotrexate; NCSE, nonconvulsive status epilepticus;
NMDA-r, anti-NMDA receptor encephalitis; r, right; S, spike; SeLEAS, self-limited
epilepsy with autonomic seizures; S-W, spike-and-waves; SWS, Sturge–Weber Syndrome;
VP, ventriculoperitoneal shunt; VWM, vanishing white matter disease; W, waves.
a Diagnostic investigation carried out as a first test.
b EEG during AMS was not contributory to the final diagnosis, but its role was supportive
during follow-up.
c No EEG modification in a patient with a known history of epilepsy.
In the six patients in which a CT scan was used a first diagnostic tool, further testing
was necessary to identify the cause of AAMS, since the information obtained from CT
scanning alone was not sufficient. On the other hand, early use of an MRI successfully
identified the presumptive cause for AMS (MTX leukoencephalopathy—bilateral symmetrical
centrum semiovale areas of restricted diffusion and reduced apparent diffusion coefficient
(ADC) value) as did LP (pleocytosis, suggesting an acute inflammatory involvement
of the central nervous system).
After the first diagnostic assessment, in all patients other procedures were carried
out as follows: MRI was performed in n = 13 (65.0%) and was diagnostic in 5 children (38.5%); CT scan was performed in 10
patients (50.0%) to rule out the possibility of intracranial hemorrhages or major
lesions; LP was carried out in 2 patients and was diagnostic in 1. Toxicological screening
was carried out in all children and revealed a carbamazepine intoxication in one patient
only.
We classified our EEGs as “diagnostic” in 9/20 children, “informative” in 5/20, and
“poorly informative” in 6/20 ([Fig. 1 ]).
Fig. 1 EEG contribution to the final diagnosis: diagnostic EEG 45% of cases, informative
EEG 25% and poorly informative EEG 30%. EEG, electroencephalogram.
Diagnostic EEG
EEG was directly contributory to the diagnosis in nine patients. In 5 patients, it
demonstrated a clear ictal pattern (NCSE in 4/5 and subtle epileptic spasms in 1/5;
[Fig. 2 ]). In 4/9 cases, the procedure ruled out seizures and the patients were subsequently
diagnosed with psychogenic nonepileptic seizures (PNES). In these nine patients, the
EEG not only aided the diagnostic process, but it also enabled a timely and correct
management of their AAMS.
Fig. 2 [sens 150 µV/cm, LF 50 Hz, HF 1.6 Hz, 20 s/p] Rhythmic spikes, polyspikes, and spike-and-waves
were observed over the left occipital and posterior temporal leads, with some theta
rhythmic activity in the contralateral homologous regions. The patient exhibited impaired
awareness and tonic deviation of the eyes toward the right side. The EEG was diagnostic
of NCSE. EEG, electroencephalogram; NCSE, nonconvulsive status epilepticus.
All five patients who had ictal EEG subsequently underwent an MRI, but a specific
etiology was found only in one case (Sturge–Weber syndrome). In another patient the
combination of specific EEG features and a normal MRI permitted the diagnosis of self-limited
epilepsy with autonomic seizures (SeLEAS).
Informative EEG
EEG was considered informative in five patients. In two of them (patient 6 and patient
8), on the basis of their medical history, clinical presentation, and early examinations,
an onco-hematological disease and relapse of known epilepsy were respectively suspected.
However, in patient 6, EEG disclosed diffuse slowing ([Fig. 3 ]), while in patient 8 it showed asymmetric slowing with superimposed delta brushes.
Both findings are typical of encephalitis, which was confirmed by further testing:
an MRI revealed acute disseminated encephalomyelitis (ADEM) in patient 6, while LP
demonstrated anti-N-methyl-D-aspartate (NMDA) receptor-encephalitis in patient 8.
Fig. 3 [sens 100 µV/cm, LF 50 Hz, HF 1.6 Hz, 20 s/p] Diffuse monomorphic slow waves, occasionally
organized into bouffées of high-voltage slow-waves. The patient presented with fever,
vomiting, dyskinesia of the upper limbs, and clonus. The EEG was informative of encephalitis,
and the MRI permitted the diagnosis of ADEM. ADEM, acute disseminated encephalomyelitis;
EEG, electroencephalogram; MRI, magnetic resonance imaging.
In patient 12 a presumptive diagnosis of migraine was made based on the patient's
medical history (recent onset of headache and vomiting) and neurological examination
(no focal neurological signs, improved awareness in time with no specific intervention).
The EEG showed discontinuous anterior slow waves compatible with the clinical suspicion
and which ruled out other differential diagnoses such as encephalitis. No further
testing was deemed necessary.
In the last two cases (patients 1 and 18), NCSE was suspected based on medical history
but was ruled out by EEG. Further diagnostic investigations led to the final diagnosis.
Poorly Informative EEG
In six cases, EEGs were labelled as “poorly informative.” In these patients, the most
common EEG finding was slowing, which was not specific enough to confirm a particular
diagnosis.
In four of these patients, EEG was repeated during hospitalization. Follow-up EEG
together with the clinical picture helped establish the final diagnosis. Specifically,
in two patients, follow-up EEG helped confirm the diagnosis of NCSE: initial features
included persisting slowing, mainly focal, associated with asymmetric spindles, but
the slowing gradually improved and predominantly focal discharges appeared, thus confirming
the presumptive diagnosis.
In the two remaining patients, the clinical suspicion of migraine was supported by
a rapid EEG improvement, especially during sleep. The EEG remained well-organized
throughout the whole hospitalization.
In the last two cases, the EEG was not considered contributory to the diagnosis. One
EEG was normal and only an MRI led to the diagnosis of MTX encephalopathy. The last
patient had previously been diagnosed with TBC1D24 epilepsy, but the EEG at onset
of AAMS showed only a slight increase in the frequency of discharges; however, in
the following days, her symptoms worsened and she developed a NCSE, thus confirming
a disease progression.
Summary of Results
In conclusion, EEGs were considered contributory and helped reach an early diagnosis
in 14 patients (70.0%). Additionally, in five patients (25.0%), follow-up EEGs were
considered contributory to diagnosis, although this was obtained after some days only.
Only in one patient was EEG considered noncontributory throughout the whole hospitalization.
MRI was performed in 13 patients and was diagnostic in 5 (38.5%). Particularly, MRI
established the specific cause of disease in three patients (vanishing white matter
disease, MTX- encephalopathy, ADEM), the etiology of NCSE in one patient (Sturge-Weber
Syndrome) and confirmed the presumptive diagnosis anti-NMDA-r encephalitis in the
remaining case.
[Fig. 4 ] shows the contribution of EEG to diagnosis. In patients aged 9 or less, EEG played
a major role in contributing to the final diagnosis by differentiating between different
neurological disorders; in these children, EEG was also pivotal in prescribing an
adequate treatment. In children over the age of 9, EEG was most helpful in detecting
PNES and thus in excluding other neurological conditions.
Fig. 4 Etiology distribution by age. X -axis: etiologies; Y -axis: number of patients. Epileptic events and encephalitis were seen more frequently
in children younger than 9 years while migraine and PNES in older patients. AMS due
to a nonepileptic neurological condition (disease progression and or a side effect
of medication) was equally distributed in patients younger and older than 9 years
of age. AMS, altered mental status; NCSE, nonconvulsive status epilepticus; PNES,
psychogenic nonepileptic seizures.
Discussion
AMS in children represents a diagnostic challenge, mainly because it encompasses a
wide spectrum of diseases with different underlying etiologies. Among them are complex
migraines, seizures, and encephalitis; this wide spectrum is one of the reasons for
which the literature is unclear on the most useful test to help determine the specific
etiology of AAMS in the pediatric population.[6 ] A recent study of a large cohort of children with AMS revealed that only 114/336
patients received a definite diagnosis in the ED. Interestingly, this study included
all children presenting to the ED with AMS, even those with a medical history suggestive
of a specific etiology (i.e., recent trauma or accidental drug ingestion) and patients
with features suggestive of motor seizures.[6 ] In addition, the same study showed that most of the diagnostic tests conducted did
not assist in identifying the underlying etiology of AMS. Head CT was the most frequently
performed imaging study (n = 80, 36.0%), but its diagnostic yield was low (n = 15, 19.0%), whereas LP (n = 11, 42.0%) and MRI (n = 17, 33.0%), although performed less frequently, often led to a diagnosis.[6 ] Similarly, in our study, MRI was found to contribute to diagnosis in 38.5% of children,
while head CT was mostly helpful in excluding bleeding and major lesions. Unfortunately,
Button et al did not consider the role of EEG in identifying the etiologies of AMS
in pediatric patients presenting to the ED.[6 ]
In our study, we found that EEG has a high diagnostic yield (70.0%) for AMS of unknown
etiology, helping not only to obtain a prompt diagnosis but also to choose the best
course of treatment.
The usefulness of EEG in the ED setting has been demonstrated by several retrospective
studies focusing on adults,[11 ]
[12 ]
[13 ]
[15 ]
[16 ]
[17 ]
[18 ] children,[6 ]
[19 ]
[20 ]
[21 ]
[22 ]
[23 ]
[24 ] or both.[10 ] Specifically, our findings on the high diagnostic yield of EEG are in line with
pediatric findings on retrospective studies: Kothare et al[20 ] reported that 94.0% of EEG were contributory to the management of a cohort of 32
children; Alehan et al[19 ] noted that 84.0% of EEG helped achieve a diagnosis in a population of 56 children;
Gunawardena et al[24 ] found EEG helpful in confirming or ruling out diagnoses in 87.7% of cases.
Similar results were outlined by Falsaperla et al[25 ] in a literature review showing that emergency video-EEG is useful for decision making
in 96.6% of cases. In our prospective cohort, EEG was contributory to the decision-making
process in 70.0% of cases, and helped identify a specific underlying diagnosis (either
seizures or PNES) in 45% of cases. For these patients, EEG not only assisted in the
diagnostic process but also facilitated timely and accurate management of their AAMS,
helping to avoid unnecessary hospitalizations,[21 ] incorrect treatments with ASMs and unnecessary further testing in patients with
PNES, and to monitor the efficacy of ASMs in NCSE.
Thus, our data underline how performing an EEG while the clinical presentation is
still ongoing could lead to a conclusive diagnosis in the case of NCSE and psychogenic
events.[21 ]
Individuals with PNES are often at risk of iatrogenic harm, as they are prone to unnecessary
treatments and hospitalizations. Additionally, undetected PNES places a substantial
burden on individuals, their families, and the health care system.[20 ]
[21 ]
[25 ] In our study, we found a significant percentage of PNES, similar to those in Alehan
et al and Gunawardena et al's studies (10 and 13%, respectively). As described by
Kothare et al,[20 ] 37.0% (12 out of 32) of paroxysmal events or AMSs could be nonepileptic in origin
and their prompt identification could avoid inappropriate use of ASMs and shorten
hospital lengths. Our study, together with the above-mentioned ones, confirms the
usefulness of EEG in identifying PNES and its crucial role in the accurate diagnosis
and effective management of pediatric patients presenting with seizure-like symptoms.
The other diagnosis in which early EEG plays a role is NCSE, and its detection is
extremely important, mostly in the ED and ICU.[26 ]
[27 ]
[28 ]
[29 ] Early detection and treatment of this condition seems to prevent an aggravation
of the clinical picture; however, emergency physicians may underestimate status epilepticus
in patients without motor signs.[25 ]
Across the retrospective pediatric studies on the use of EEG in ED,[19 ]
[20 ]
[21 ]
[22 ] taking into account the slightly different inclusion criteria, the incidence of
NCSE in pediatric AMS ranged from 14.3 to 30%, with our prospective findings falling
within this range. These studies consistently support the use of EEG in the emergency
evaluation of pediatric AMS to detect NCSE, paralleling our results.
Interestingly, not all the studies used continuous (full channel) EEG monitoring.
Particularly, Yamaguchi et al[22 ] focused on the role of reduced-lead EEG in detecting NCSE in children with AMS.
The EEG traces were assessed directly by an ED physician and only in few cases by
a neurologist, who later reviewed all EEGs. The authors reported that reduced-lead
EEG identified not only all those cases of NCSE that were subsequently confirmed by
the neurologist, but also 13 (5.3%) false positives.[22 ]
In our series, EEG was poorly contributive to the diagnosis in 30% of patients. One
of the differential diagnoses where EEG appeared ineffective is the early distinction
between post-ictal states and complicated migraine. The clinical presentations of
these two disorders may overlap, especially in children. Indeed, headache and autonomic
symptoms could suggest both migraine and nonconvulsive seizures, as is the case in
SeLEAS.[25 ] EEG findings are usually similar and include generalized slowing. However, EEG was
useful in the follow-up of some of these patients and helped achieve a final diagnosis
throughout their hospitalization. Misdiagnosing is frequent in patients with SeLEAS[25 ] and a diagnostic error could lead to an aggravation of the clinical picture or an
incorrect treatment.
Access to EEG monitoring in an emergency setting could be not easily attainable, since
it requires a multidisciplinary team (epileptologist, pediatric neurologist, technician)
with high levels of training particularly in an electrically hostile environment.[25 ]
Although traditional EEG seems to be the preferred choice in the evaluation of ongoing
AMS, when it is not available, the use of reduced-lead or abbreviated EEG could also
be considered an option.[11 ]
[15 ]
[22 ]
[23 ]
Strengths and Limits
Importantly, we chose to exclude patients with signs/symptoms or a recent history
suggestive of a specific diagnosis in order to focus on the contribution of EEG in
the determination of a challenging diagnosis. The most important strengths of our
study were its prospective nature, the availability of EEG when the symptoms of AMS
were ongoing for all patients, and its prompt evaluation by a pediatric neurologist.
To date, to our knowledge, there are no prospective studies that have focused specifically
on what we set out to investigate.
The main limitation of our study was undoubtedly the small sample size. The limited
number of patients could make our data less generalizable and more dependent on the
expertise of the pediatric neurologist who reviewed EEG traces. Additionally, it could
be considered a limitation for statistical analysis of the data. However, our sample
was representative of the experience of a regional referral center and was designed
with strict inclusion criteria. To mitigate this issue, it would be very interesting
to collect prospective multicentric data on a national or international basis to confirm
the diagnostic yield of EEG in children with AMS of unknown etiology.
