The Role of Electroencephalography in Children with Acute Altered Mental Status of Unknown Etiology: A Prospective Study

Abstract

Introduction Acute altered mental status (AAMS) is often a challenge for clinicians, since the underlying etiologies cannot always easily be inferred based on the patient's clinical presentation, medical history, or early examinations. The aim of this study is to evaluate the role of electroencephalogram (EEG) as a diagnostic tool in AAMS of unknown etiology in children.

Materials and Methods We conducted a prospective study involving EEG assessments on children presenting with AAMS between May 2017 and October 2019. Inclusion criteria were age 1 month to 18 years and acute (<1 week) and persistent (>5 minutes) altered mental status. Patients with a known etiology of AAMS were excluded. A literature review was also performed.

Results Twenty patients (median age: 7.7 years, range: 0.5–15.4) were enrolled. EEG contributed to the diagnosis in 14/20 cases, and was classified as diagnostic in 9/20 and informative in 5/20. Specifically, EEG was able to identify nonconvulsive status epilepticus (NCSE) in five children and psychogenic events in four. EEG proved to be a poorly informative diagnostic tool at AAMS onset in six children; however, in five of them, it proved useful during follow-up.

Conclusions Limited data exist regarding the role of EEG in children with AAMS of unknown etiology. In our population, EEG proved to be valuable tool, and was especially useful in the prompt identification of NCSE and psychogenic events.

Keywords

acute altered mental status - electroencephalogram - unknown etiology - children - diagnostic assessment

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.

Conclusions

As highlighted by our prospective study in a pediatric population, EEG could be a very useful tool to assess children with AAMS of unknown etiology. EEG's role seems to be very important in the identification and early treatment of patients with NCSE and it plays a pivotal role in the differential diagnosis between “epileptic” and “nonepileptic” events. Indeed, ictal patterns are well recognizable and may rapidly steer treatment.

In other cases, EEG can help suggest a specific underlying etiology and thus guide further testing such as in the case of generalized or focal slowing. In our opinion, this diagnostic tool should be available in pediatric emergency settings, together with prompt EEG interpretation and consultation with a pediatric neurologist. Although obtaining an EEG in an emergency setting is not easy, its role seems to be crucial in the differential diagnosis of AMS, especially in pediatric patients.

Conflict of Interests

None declared.

Acknowledgements

This research did not receive any specific grant from founding agencies in the public, commercial, or not-for-profit sectors.

• References

• 1 Plum F, Posner JB. The diagnosis of stupor and coma [in Japanese]. Brain Nerve 2015; 67 (03) 344-345
• 2 Srinivasan J, Chaves C, Scott B, Small JE. Netter's Neurology. 3rd ed., 2019 2012
• 3 Wong CP, Forsyth RJ, Kelly TP, Eyre JA. Incidence, aetiology, and outcome of non-traumatic coma: a population based study. Arch Dis Child 2001; 84 (03) 193-199
• 4 Kanich W, Brady WJ, Huff JS. et al. Altered mental status: evaluation and etiology in the ED. Am J Emerg Med 2002; 20 (07) 613-617
• 5 Xiao H, Wang Y, Xu T. et al. Evaluation and treatment of altered mental status patients in the emergency department: Life in the fast lane. World J Emerg Med 2012; 3 (04) 270-277
• 6 Button K, Capraro A, Monuteaux M, Mannix R. Etiologies and yield of diagnostic testing in children presenting to the emergency department with altered mental status. J Pediatr 2018; 200: 218-224.e2
• 7 Zehtabchi S, Abdel Baki SG, Grant AC. Electroencephalographic findings in consecutive emergency department patients with altered mental status: a preliminary report. Eur J Emerg Med 2013; 20 (02) 126-129
• 8 Mecarelli O. Clinical Electroencephalography [Internet]. Cham: Springer International Publishing; 2019
• 9 Sejersen T, Wang CH. Acute Pediatric Neurology. Springer; 2014
• 10 Praline J, Grujic J, Corcia P. et al. Emergent EEG in clinical practice. Clin Neurophysiol 2007; 118 (10) 2149-2155
• 11 Ziai WC, Schlattman D, Llinas R. et al. Emergent EEG in the emergency department in patients with altered mental states. Clin Neurophysiol 2012; 123 (05) 910-917
• 12 Duran L, Balci K, Yardan T. et al. The value of electroencephalography in differential diagnosis of altered mental status in emergency departments. J Pak Med Assoc 2014; 64 (08) 923-927
• 13 Newey CR, Kinzy TG, Punia V, Hantus S. Continuous electroencephalography in the critically ill: clinical and continuous electroencephalography markers for targeted monitoring. J Clin Neurophysiol 2018; 35 (04) 325-331
• 14 Kimchi EY, Neelagiri A, Whitt W. et al. Clinical EEG slowing correlates with delirium severity and predicts poor clinical outcomes. Neurology 2019; 93 (13) e1260-e1271
• 15 Egawa S, Hifumi T, Nakamoto H, Kuroda Y, Kubota Y. Diagnostic reliability of headset-type continuous video EEG monitoring for detection of ICU patterns and NCSE in patients with altered mental status with unknown etiology. Neurocrit Care 2020; 32 (01) 217-225
• 16 Rossetti AO, Schindler K, Sutter R. et al. Continuous vs routine electroencephalogram in critically ill adults with altered consciousness and no recent seizure: a multicenter randomized clinical trial. JAMA Neurol 2020; 77 (10) 1225-1232
• 17 Müller M, Rossetti AO, Zimmermann R. et al. Standardized visual EEG features predict outcome in patients with acute consciousness impairment of various etiologies. Crit Care 2020; 24 (01) 680
• 18 Nakae S, Kumon M, Moriya S. et al. Factors associated with prolonged impairment of consciousness in adult patients admitted for seizures: A comprehensive single-center study. Neurol Med Chir (Tokyo) 2021; 61 (10) 570-576
• 19 Alehan FK, Morton LD, Pellock JM. Utility of electroencephalography in the pediatric emergency department. J Child Neurol 2001; 16 (07) 484-487
• 20 Kothare SV, Khurana DS, Valencia I, Melvin JJ, Legido A. Use and value of ordering emergency electroencephalograms and videoelectroencephalographic monitoring after business hours in a children's hospital: 1-year experience. J Child Neurol 2005; 20 (05) 416-419
• 21 Fernández IS, Loddenkemper T, Datta A, Kothare S, Riviello Jr JJ, Rotenberg A. Electroencephalography in the pediatric emergency department: when is it most useful?. J Child Neurol 2014; 29 (04) 475-482
• 22 Yamaguchi H, Nagase H, Nishiyama M. et al. Nonconvulsive seizure detection by reduced-lead electroencephalography in children with altered mental status in the emergency department. J Pediatr 2019; 207: 213-219.e3
• 23 Simma L, Bauder F, Schmitt-Mechelke T. Feasibility and usefulness of rapid 2-channel-EEG-monitoring (point-of-care EEG) for acute CNS disorders in the paediatric emergency department: an observational study. Emerg Med J 2021; 38 (12) 919-922
• 24 Gunawardena S, Chikkannaiah M, Stolfi A, Kumar G. Utility of electroencephalogram in the pediatric emergency department. Am J Emerg Med 2022; 54: 26-29
• 25 Falsaperla R, Striano P, Parisi P. et al. Usefulness of video-EEG in the paediatric emergency department. Expert Rev Neurother 2014; 14 (07) 769-785
• 26 Abend NS, Dlugos DJ. Nonconvulsive status epilepticus in a pediatric intensive care unit. Pediatr Neurol 2007; 37 (03) 165-170
• 27 Abend NS, Topjian AA, Gutierrez-Colina AM, Donnelly M, Clancy RR, Dlugos DJ. Impact of continuous EEG monitoring on clinical management in critically ill children. Neurocrit Care 2011; 15 (01) 70-75
• 28 Abend NS. Electrographic status epilepticus in children with critical illness: epidemiology and outcome. Epilepsy Behav 2015; 49: 223-227
• 29 Schreiber JM, Zelleke T, Gaillard WD, Kaulas H, Dean N, Carpenter JL. Continuous video EEG for patients with acute encephalopathy in a pediatric intensive care unit. Neurocrit Care 2012; 17 (01) 31-38

Publication History

Received: 06 June 2024

Accepted: 04 August 2024

Accepted Manuscript online:
06 August 2024

Article published online:
04 September 2024