Figure 39.1. A:Patient who presented with altered mental status and headache with a CT brain showing a left hemisphere subdural hematoma with midline shift (arrows). B:EEG of this patient showing left hemisphere seizure pattern that is slowly evolving in frequency. C:Expansion of the DSA trend analysis showing multiple seizures in the record (arrows).

Intracranial Hemorrhage

In a study by Claassen et al. (6), 102 patients with ICH were recorded on cEEG and 18% were found to be having electrographic seizures and 31% were associated with seizures in some form at presentation (19 had clinical seizures prior to cEEG recording). Seizures in this series were more likely to be recorded with hemorrhages that expanded by 30% or more and with cortical bleeds. A study by Vespa et al. (7) studied 63 patients with ICH and found 18 (28%) with seizures on cEEG monitoring. Seizures were recorded most often in lobar hemorrhages, but were found in 21% of subcortical hemorrhages as well.

Ischemic Stroke

A series of 177 patients with acute ischemic stroke were described by Mecarelli et al. (8) with 15 patients (6.5%) having seizures on EEG monitoring. It was also noted that periodic lateralized epileptiform discharges were recorded in 6% of patients with stroke, and this subset of patients had a 71% incidence of seizures. In the series by Claassen et al. (5), 6 patients of the 56 with ischemic strokes (11%) had seizures recorded on cEEG monitoring. The rate of seizures detected in acute ischemic stroke has been consistently less than ICH and TBI; however, the rates of chronic epilepsy after the acute symptomatic seizures will require future studies.

Traumatic Brain Injury

Acute TBI has been associated with EEG seizure activity that often occurs in the first 14 days after injury. A study by Vespa et al. (9) described 94 patients with TBI who underwent cEEG monitoring, and 21 patients (22%) had electrographic seizures. Another series of 70 patients with TBI found 23 patients (33%) with electrographic seizures recorded on cEEG monitoring. The seizures in these series were often not adequately treated with “prophylactic” doses of phenytoin and required cEEG for diagnosis as well as management of ongoing seizure treatments.

Anoxic Brain Injury

The detection of seizures in the setting of anoxic brain injury has been described in multiple series and often carries a grim prognosis. The use of therapeutic hypothermia has changed the use of cEEG for prognosis (10). A study looking at 101 patients with anoxic injury after cardiac arrest found 12 patients (12%) to have nonconvulsive status epilepticus. The majority of these patients had EEG recorded in the first 8 hours of monitoring (with 12 hours of the cardiac arrest). The outcome in this study was dismal with only one of the patients with EEG seizures survived to a vegetative state. The background reactivity of cEEG recordings to painful stimulation has been reported to have strong predictive value of prognosis after anoxic injury (11). In a study of 34 patients with anoxic brain injury, 7 patients (20%) had EEG seizures recorded. Among the nonsurvivors of the cohort, 75% had nonreactive background to stimulation, 73% had prolonged discontinuous activity (burst suppression), and 47% had seizures. The survivors in the cohort all had reactive backgrounds and did not have EEG seizure activity. The background reactivity changes seen in this study were noted during therapeutic hypothermia. A study of 95 patients with anoxic injury found 26 patients (27%) with EEG seizures. They also described two seizures patterns, one arising from burst suppression and one arising from a cEEG background (12). Seizures were still associated with a dismal prognosis with mortalities around 100% in most studies, but 2 (out of 10) patients who had seizures arising from a cEEG pattern regained consciousness.

CNS Infections/Inflammation

Encephalitis or meningoencephalitis is a risk factor for seizures and has been reported to be associated with EEG seizures in 6% to 29% of patients (5,13). In a study of patients with bacterial meningitis, seizures occurred in 121 of 696 patients (17%). The patients with seizures had worse outcomes (41% mortality) than patients without seizures (16% mortality). Viral encephalitis has also been shown to have risk of EEG seizures with 5% of all seizures being attributed to suspected viral encephalitis in one study (14). Autoimmune encephalitis is becoming increasingly recognized as an etiologic agent in critical care status epilepticus. Some of these patients have a paraneoplastic disorder with antibodies against an occult tumor, but many have autoantibodies without an obvious source. Patients often present with refractory status epilepticus that does not respond to traditional antiseizure medications and requires aggressive immunotherapy. There is also evidence that some patients with infectious viral encephalitis (herpes encephalitis) can develop autoantibodies that then contribute to the generation of seizures and neurologic deterioration (15,16). There has also been a typical EEG pattern often present on the initial EEG of some patients with anti-NMDA receptor encephalitis. This pattern, the so-called delta brush, which is characterized by high-amplitude delta waves and overlying fast activity, may be helpful for identifying these patients and guiding therapy (17).

Subarachnoid Hemorrhage

EEG seizures have been reported in 3% to 26% of comatose patients with SAH. Claassen et al. (18) studied 247 patients with SAH and found EEG seizures in 17 patients (7%). SAHs with associated subdural hemorrhage or cerebral infarction were predictive of patients with seizures. Patients with seizures reported poorer quality of life and experience prolonged recovery times. Another study examined 389 patients with SAH and found 11 patients with nonconvulsive status epilepticus (3%). Patients with poor neurologic grade and associated structural lesions (intracerebral hemorrhage and stroke) were more likely to have seizures (19). In addition to being at risk for EEG seizures, patients with SAH are at risk for delayed cerebral ischemia (DCI). DCI may result in stroke and cause a clinical deficit, but in poor-grade SAH, there is often no neurologic exam to follow (patients are often comatose). Quantitative EEG trend monitoring of cEEG recording has been shown to detect DCI prior to cerebral infarction and in time for intervention. Claassen et al. (20) examined 34 patients with SAH and identified 9 patients with DCI. For patients who experienced DCI, the alpha–delta ratio was significantly decreased (24%) compared to an increase (3%) in the alpha–delta ratio for those patients without DCI. In a study from Montreal, 13 patients were studied with quantitative EEG and 8 patients were determined to have DCI. DCI was predicted by using a quantitative EEG method termed the composite alpha index (CAI). In 3 of the 8 patients, changes in the CAI predicted the changes of DCI 24 hours prior to clinical change (21). cEEG has also been used to develop prognosis in poor-grade SAH. When 116 patients were followed for functional outcome, predictors of poor outcome were periodic epileptiform discharges, electrographic status epilepticus, and the absence of sleep architecture (22).

EEG Monitoring Duration

Determining the length of cEEG monitoring in adult patients admitted to the neuro-ICU with altered mental status is complex and can vary based on etiology, EEG findings, and the degree of altered mental status. If time to first seizure is examined, then a 40% yield of seizures is found at 20 minutes, 50% yield at 1 hour, and 80% to 90% yield at 24 hours for most patients (5). Thus, a minimum recording time of cEEG monitoring for a patient with altered mental status in the critical care setting is 24 hours; however, some situations may require additional monitoring. Patients who have periodic lateralizing epileptiform discharges (PLEDs) have been shown to have a delayed detection of seizure onset that requires 48 hours of monitoring (5). Patients with PLEDs also have a 50% to 70% chance of having seizures recorded sometimes during the record demonstrating the importance of monitoring this population on cEEG monitoring (23,24). Comatose state has also been associated with a prolonged need for monitoring, and 48 hours has been suggested as a minimum time of recording (5). In patients with intracerebral hemorrhage, EEG seizures were detected in 56% in the first hour and 94% by 48 hours, suggesting prolonged monitoring may also be useful in this population (6). Monitoring patients for DCI after SAH typically requires up to 7 to 10 days of monitoring depending on the patient’s clinical course (20,21). Finally, patients with EEG seizures in the record require additional monitoring. In patients who do not present with clinical signs, video EEG is necessary to make the diagnosis, and once the nonconvulsive seizures have been treated, an additional 48 hours of continuous monitoring is common practice, but may vary based on the clinical needs.

EEG Seizures and Outcome

The use of cEEG in the critical care setting has dramatically increased the number of patients diagnosed with seizures in the hospital. Unfortunately, there is not a wealth of outcome data to guide treatment of these seizures and to understand their significance. Outcome studies have shown that patients with an acute neurologic injury (stroke, hemorrhage) with seizures have a worse functional outcome than patients with the same primary neurologic injury without seizures. Such findings may suggest that the seizures may contribute to the extent of injury. Research addressing the effect of treating these acute symptomatic seizures on functional outcome is required. The majority of outcome studies that have been done to date have focused on anoxic brain injury (11,12). From these studies, it is clear that EEG seizures in the setting of anoxia are a negative prognostic sign and are typically associated with a 90% to 100% mortality. Treatment of EEG seizures in anoxic brain injury has not been shown to improve outcome. The use of cEEG to demonstrate reactive background during therapeutic hypothermia has been shown to predict outcome (12). Future studies will need to address the effect of treating nonconvulsive seizures in other etiologies and the factors that predict the development of chronic epilepsy after nonconvulsive seizures are recorded in the setting of an acute neurologic injury.

Current Practice and Guidelines

The use of cEEG has rapidly increased in the past decade. Patients with altered mental status and/or acute neurologic injury are at risk for nonconvulsive seizures and are candidates for monitoring. The number of patients that meet these criteria is often daunting, and a conservation of resources is needed. Focusing on the high-risk etiologies (presented above) and the patients with the most profound altered mental status is a good practice to utilize equipment in the high-yield patients. There is a consensus statement that was published by the neurointensive care section of the European Society of Intensive Care Medicine (25). Their conclusions were to use EEG monitoring in convulsive status epilepticus and to rule out nonconvulsive status epilepticus in brain-injured patients and in comatose ICU patients without primary brain injury who have unexplained and persistent alerted consciousness.

CONTINUOUS EEG MONITORING IN PEDIATRIC PATIENTS

cEEG monitoring in critically ill children is generally performed in multidisciplinary pediatric and neonatal ICUs. There are few designated pediatric or neonatal neuro-ICUs, so care is generally provided by collaborating intensivists, encephalographers, and intensive care neurology consult teams. Despite the paucity of dedicated pediatric or neonatal neuro-ICUs, judicious use of cEEG has seen a growing footprint in the last 5 years. The pediatric literature is rich with available data highlighting the important role played by cEEG in optimizing neurologic health–related outcome in critically ill neonates and children.

Electrographic Seizure Epidemiology

Observational studies of children undergoing clinically indicated cEEG in pediatric intensive care units have reported electrographic seizures in 10% to 40% of children, and about one-third of children with electrographic seizures have a sufficiently high seizure burden to be categorized as electrographic status epilepticus (26–38). Indications for cEEG differ across studies, but most included a primary indication related to “acute ongoing encephalopathy.” The largest study of cEEG in critically ill children retrospectively enrolled 550 children from 11 tertiary care centers in North America. Electrographic seizures occurred in 30% of 550 children, and in 33% of children with electrographic seizures, the seizure burden was classified as electrographic status epilepticus. Many children with electrographic seizures would not be identified without cEEG. In the multicenter study, 35% of children had exclusively EEG-only seizures (37), and these data are consistent with other single-center studies (26,27,29–31,33–36,39). Several studies have demonstrated that EEG-only seizures occur even in children who have not received any or recent paralytics, indicating there is an ongoing dissociation of electrical brain activity and outward mechanical signs (electromechanical uncoupling) (35,36). Identifying children at higher risk of electrographic seizures is complex since electrographic seizures have been reported in both large heterogeneous cohorts (37) and smaller single brain insult etiology cohorts (28,32–34). However, several risk factors have been reported including infants being at higher risk than older children (29,32,34,35,37), convulsive seizures (30,37,39), or status epilepticus (29) prior to cEEG, the presence of acute structural brain injury (28–30,32–34,39), and the presence of interictal epileptiform discharges (29,33,37,39), or periodic epileptiform discharges (26). Although these risk factors are statistically significant, the absolute difference in seizure risk is often only 10% to 20%, so these risk factors may have limited clinical utility in selecting patients to undergo cEEG.

Critically ill neonates are also at high risk for electrographic seizures (40–43), and about one-third of neonates with electrographic seizures have a sufficiently high seizure burden to be categorized as electrographic status epilepticus (41,43,44). Seizures have been identified as common in some specific cohorts of neonates including 33% to 65% of neonates treated with therapeutic hypothermia for hypoxic–ischemic encephalopathy (41–43), as well as many neonates with stroke, undergoing extracorporeal membrane oxygenation, with meningoencephalitis, and in the perioperative period of congenital heart disease surgery (40

Only gold members can continue reading. Log In or Register to continue