Ictal Interictal Continuum in the pediatric intensive care unit

The ictal interictal continuum (IIC) is a common EEG pattern identi�ed in critically ill adults. The IIC has not been systematically studied in critically ill children. We report the incidence of the IIC in the pediatric intensive care unit (PICU). We then compare IIC patterns to periodic and rhythmic patterns (PRP) not meeting IIC criteria looking for associations with cerebral injury, electrographic seizures (ES), and in-hospital mortality.


Introduction
The ictal interictal continuum (IIC) is an electrographic pattern commonly seen in adults with critical neurologic injury 1,2 .It consists of periodic and rhythmic patterns (PRP) that do not qualify as an electrographic seizure (ES) or status epilepticus (ESE) but that may be contributing to impaired alertness, causing other clinical symptoms, and/or contributing to neuronal injury' 3 .As such the ictal nature, signi cance, and approach to treatment remain unclear [3][4][5][6] .Recently, the American Clinical Neurophysiology Society (ACNS) published criteria de ning the IIC, facilitating formal evaluation and study of these patterns 3,6 .Several studies in the adult neurocritical care setting have shown associations of the IIC with ES and cerebral injury 1,[7][8][9][10][11][12] .Some of these studies, performed prior to well-de ned criteria for the IIC, suggest that PRP thought to be on the IIC are associated with secondary brain injury due to increased focal cerebral metabolism and reduction in brain tissue oxygenation, even in the absence of seizures.Despite these ndings, studies of these patterns in the pediatric neurocritical care setting have been limited.
In this prospective study, we describe the incidence of the IIC in a large cohort of pediatric neurocritical care patients, using the 2021 ACNS criteria.We correlate neuroimaging ndings with the IIC and evaluate associations of the IIC and PRP not meeting IIC criteria, with mortality, cerebral injury, and ES.

Methods
The Children's National Hospital Institutional Review Board for human subject studies approved this prospective study of continuous EEG (cEEG) in critically ill pediatric patients in the pediatric intensive care unit (PICU) from July 1st, 2021through January 2023 (Pro00015473).Patients with both cEEG and neuroimaging were included.Patients with epilepsy and/or a history of a prior cerebral injury were excluded.

Patient characteristics
At our institution, cEEG is used in accordance with the ACNS consensus statement on cEEG in pediatrics and adults, with standardized protocols for cEEG in patients with traumatic brain injury, cardiac and respiratory arrest, clinical seizures with continued altered mental status or need for pharmacologic paralysis, extracorporeal membrane oxygenation (ECMO), and altered mental status of either unclear etiology or with a known or suspected supratentorial etiology,(i.e., encephalitis, intracranial hemorrhage, etc.) 13,14 .Patients on cEEG are monitored for a minimum of 24 hours with prolonged recordings for patients with high-risk cEEG backgrounds (i.e., those with epileptiform discharges) or higher propensity to seize after 24 hours (i.e., patients on ECMO) 13 .Data extracted from the medical record included age, sex, past medical and neurologic history, admitting diagnosis, duration of cEEG, presence or absence of new cerebral injury, and in-hospital mortality.

Electroencephalography
EEG was performed using a standard 10-20 electrode montage according to ACNS guidelines 14 .All patients admitted to the PICU with cEEG were screened for sporadic epileptiform discharges (SED) and PRP by a board-certi ed epileptologist (AJS) blinded to the clinical data.The background EEG was classi ed as normal, slow/disorganized, discontinuous, burst suppression/attenuation, or attenuated/featureless 15 .SED and PRP were described in accordance with ACNS Critical Care EEG nomenclature 3,6 .We documented if SEDs were isolated or present in combination with PRPs.Patients with PRPs were classi ed as either having periodic discharges alone, rhythmic patterns alone, or a combination of PRP.Patients were then further classi ed according to location(s) of the PRP (generalized, lateralized, bilateral independent, unilateral independent, and/or multifocal).
When PRP were identi ed, the ACNS critical care EEG nomenclature for the IIC was applied.These criteria include 1) Any periodic discharge (PD) or spike wave (SW) pattern that averages > 1.0 Hz and ≤ 2.5 Hz over 10 seconds (s) or 2) Any PD or SW pattern ≥ 0.5 Hz and ≤ 1.0 Hz over 10s with uctuation or a plus modi er or 3) Lateralized rhythmic delta activity averaging > 1 Hz for 10s with uctuation or a plus modi er, and 4) Does not meet criteria for ES 3 .
ES were de ned as either epileptiform discharges averaging > 2.5 Hz for ≥ 10s or any pattern with de nite evolution lasting greater than 10s 3 .The locations of ES onset were documented and categorized as generalized, focal, or multifocal.
Reports for both CT-and MR-based imaging obtained during the patient's hospitalization were screened for acute ndings.We classi ed the ndings as focal if abnormalities were con ned to the left or right cortical and/or subcortical region (i.e., left middle cerebral artery stroke) or bilateral if the ndings were present in the left and right cortical and/or subcortical regions (i.e., global anoxic injury).

Data analysis
Study data were collected and managed using REDCap electronic data capture tools hosted at Children's National Hospital 16 .REDCap (Research Electronic Data Capture), a secure, web-based application designed to support data capture for research studies, funded at Children's National Hospital by UL1TR001876.
Descriptive statistics for nonparametric data included median (interquartile range, IQR) (i.e., age, duration of EEG) and proportions as percentages (i.e., sex, percentage of periodic and rhythmic patterns).
We assessed associations of the IIC with ES, cerebral injury, and mortality and compared the IIC to periodic and/or rhythmic patterns alone using the same outcome measures.These associations are presented using odds ratios (OR), with 95% con dence interval (CI).P values of < 0.05 were considered signi cant.Statistical analyses were performed using MedCalc statistical software version 22.007.

Demographics and clinical features
201 patients were monitored on EEG, exclusive of those with epilepsy or known prior cerebral injury.The median age was 5.8 years (IQR 0.7-12.2),and 40% (81/201) were female.The most common reason for admission was rst time seizure (24%(49/201)), followed by cardiac arrest (23%(47/201)).Seventy-two percent had a slow disorganized background, 53% had an acute imaging abnormality and 13% died in the hospital.In patients with rhythmic patterns alone, acute brain injury on neuroimaging was seen in 75% (9/12) and ES in 25% (3/12).Acute neuroimaging ndings were seen in 71% (5/7) of patients with GRDA.One patient with GRDA had ES, and this patient also had SED.
When IIC patterns were identi ed, a single criterion was met in 71% (16/24) with the remainder meeting two individual criteria during the recording.In patients meeting a single criterion, there was relatively equal distribution between the three speci c criteria, with criterion 3 being most common (38%, (6/16)).In patients meeting multiple criteria, a combination of criteria 1 and 2 was most common (50% (4/8)).All patients with an IIC pattern except one had evidence of acute brain injury on neuroimaging.The most common abnormality was global anoxic injury (33% (8/24)) after trauma or cardiac arrest.CNS infection was also common, seen in 21% (5/24) of patients.The single patient without acute neuroimaging ndings had new onset epilepsy (Febrile Infection-Related Epilepsy Syndrome).We identi ed ES in 55% (23/42) of patients with a PRP and in 83% (20/24) of those with an IIC pattern.Table 3 shows associated neuroimaging with IIC patterns.

Discussion
In this prospective single center study of the IIC in pediatrics we found that the IIC is common in pediatric neurocritical care patients.IIC patterns have a strong association with cerebral injury of varying etiologies, and when compared to PRP not meeting IIC criteria, the IIC is more likely to be associated with electrographic seizure.
Information on PRP and particularly the IIC has been limited in the pediatric neurocritical care setting.
In our cohort, the incidence of both PRP and IIC are higher than prior reports in other pediatric studies.
This may be due to differences in cEEG monitoring protocols at the different pediatric centers.At our institution, we use cEEG routinely for the evaluation of acute encephalopathy due to suspected or con rmed cerebral injury despite not having concern for clinical seizure.Examples of this include patients with CNS infections, traumatic brain injury and ECMO.This may inherently increase the incidence of PRP as compared to other pediatric centers that may focus on concern for clinical seizure prior to cEEG initiation.
Importantly, our study is the rst in the pediatric neurocritical care setting to show a clear association of PRP, and speci cally the IIC, with cerebral injury.In the adult Neurocritical care setting, the association of periodic patterns and the IIC with acute brain injury is well established.In a study of 152 adults with moderate to severe traumatic brain injury, 14% had IIC patterns and these had a direct correlation with the injury severity score 19 .Similar ndings were reported in a cohort of patients with intraventricular (IVH) or intraparenchymal hemorrhage (IPH).Among 11 patients with IVH, 64% had PRP compared to 38% of 32 patients with IPH 20 .Small studies in children have shown associations of LPDs, BIPDs, and GPDs with acute structural abnormalities (i.e.neonatal stroke) and intracranial infection [21][22][23][24][25][26] .LPDs in neonates and children were initially described in 1979 by PeBenito and colleagues.They described seven infants and children with LPDs in association with metabolic and/or brain structural abnormalities 22 .A later study of 15 pediatric patients with LPDs showed that most patients (12/15) had an associated structural radiographic correlate that included diffuse anoxic injury, focal infarction, subdural hemorrhage, and/or white matter lesions 21 .In our study, the IIC was associated mostly with global anoxic injury (30%,7/23) and CNS infections (17%,4/23).
Further studies are necessary to assess the independent impact that IIC patterns may have on cerebral injury in pediatrics.In adults, there are clear associations with increased metabolic demand and decreased regional oxygenation when the IIC is identi ed.In a prospective study of 90 comatose adults with poor grade subarachnoid hemorrhage, periodic discharges identi ed at a frequency ≥ 2 Hz were associated with a 25% decrease in PbtO 2 5 to 10 minutes after their onset 27 .Another adult study of nine patients with LPDs showed that seven of them had an acute brain injury.In this cohort, changes in glucose metabolism using FDG-PET scans were seen in patients with LPDs with a frequency ≥ 1 Hz 28 .These studies suggest that IIC patterns may contribute to secondary brain injury and, in theory, treatment leading to resolution of the IIC could have some impact on degree of cerebral injury and subsequent outcome.However, the extent to which IIC patterns independently impact outcome, particularly in patients with severe brain injury, remains unclear.In an adult study of 68 patients with subarachnoid hemorrhage, 79% (54/68) of whom had poor-grade subarachnoid hemorrhage, 33 patients had periodic patterns and 38 had rhythmic delta and there was no correlation with mortality or disability at discharge for either pattern 29 .It is possible that PRP and IIC may have greater impact in patients with a mild to moderate degree of injury or in those with etiologies that have not yet led to cerebral injury such as early-stage CNS infections.
We also add to the limited body of literature that highlights the association of IIC patterns with ES (OR 121 (33-451), p < 0.0001).Interestingly, PRP alone not meeting IIC criteria did not have an association with ES.This may offer insight to our approach to these patterns and the importance of assessing PRP for IIC features in clinical practice.In theory, treating highly epileptiform patterns such as high frequency SEDs, some PRP, and the IIC may limit the development of seizures and/or lower seizure burden 17,18 .
While the impact of treatment of ES on outcome remains unclear, several studies highlight associations of worse long-term outcome in children with frequent seizures and status epilepticus as compared to those without 30,31 .The associations of ES on short term outcomes are variable with some suggesting an association of higher seizure burden with worsening discharge PCPC and others showing no association of ES on other short term metrics, particularly mortality 32,33 .
The question of how to approach the treatment of patients with IIC patterns in pediatrics remains unclear.
In the adult Neurocritical care setting, treatment approaches target patterns of faster frequencies (i.e.>2Hz) with plus modi ers with anti-seizure medications (ASM) 4,5 .Some have speculated that decisions about whether to treat an IIC pattern with ASMs might be informed by metabolic studies or measures of cerebral blood ow through multimodal monitoring 34,35 .
Our study is limited by the single center design and inherent selection bias as our use of cEEG may differ when compared to other tertiary care pediatric neurocritical care units.In addition, our study focuses on the PICU alone, which may impact the frequency of patterns seen if other intensive care units such as the cardiac and neonatal ICU were included.Finally, we have excluded patients with epilepsy and chronic injury which, if included, may have increased the overall incidence of the IIC.

Conclusion
In conclusion, the IIC is a common EEG pattern in pediatric neurocritical care and has an association with cerebral injury.Further study is needed to evaluate the independent impact of these patterns on cerebral injury in children.

Declarations
This manuscript complies with the instructions to the authors set forth in the author instructions.Author requirements have been met and the nal manuscript presented was edited and approved by all authors.This manuscript has not been submitted or published elsewhere and adheres to all ethical guidelines.This is an IRB approved study at our institution, and it has been assigned the following identi cation number (Pro00015473).The authors have no disclaimers or con icts of interest relevant to this manuscript and have not received funding for this work.Manuscript submission has been approved by each author.This study is not funded.This is a prospective cohort study, and as such abides by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement guideline for reporting observational studies.
The STROBE checklist has been included in this submission.
Author contributions: Dr. Sansevere participated in conception, study design, data collection, data analysis, manuscript preparation, editing and submission.As the primary author Dr. Sansevere takes full responsibility for all aspects of this work.
Ms. Keenan participated in conception, study design ,data collection, data analysis ,manuscript preparation, editing, and approval.
Dr. Pickup participated in study design ,data collection, manuscript preparation editing, and approval.
Dr. Conley participated in study design ,data collection, manuscript preparation, editing, and approval.
Ms. Staso participated in study design ,data collection, manuscript preparation editing, and approval.
Dr. Harrar participated in conception, study design ,data collection, manuscript preparation, editing, and approval.

Table 1 :
Demographics and clinical features Fishers Exact Test and Mann Whitney U test used to assess associations between demographic and clinical features IIC-Ictal interictal continuum, IQR-interquartile range, PRES-Posterior Reversible EncephalopathySyndrome, PRP: periodic and/or rhythmic patterns

Table 2
highlights details of SEDs, PRP and IIC showing proportions of each with cerebral injury, ES, and mortality.