Article In Brief
The use of continuous EEG (cEEG) detected more seizures and more frequent modifications of antiseizure treatment in critically ill patients with impaired consciousness. However, cEEG was not associated with changes in mortality compared with repeated regular EEG.
In hospitals where resources are scarce, continuous EEG (cEEG) may not always be necessary for critically ill adults with impaired consciousness and no recent seizure, according to a July 27 online report in JAMA Neurology.
The study found that cEEG was associated with increased detection of seizures and more frequent modifications of antiseizure treatment but was not associated with changes in mortality compared with repeated regular EEG (rEEG).
For patients similar to those in the study cohort, “more may not always be better,” lead author Andrea Rossetti, MD, of the department of clinical neuroscience at University Hospital in Lausanne, Switzerland, told Neurology Today. “Routine EEG seems appropriate in settings with relatively limited resources. At some institutions, cEEG is probably overused in several patients.”
Dr. Rossetti emphasized, however, that the study excluded patients with known seizures and seizure immediately preceding EEG. “In these patients, cEEG remains indicated, in my view. And these patients represent a considerable proportion of patients in the intensive care unit,” he said.
Dr. Rossetti and colleagues analyzed data on 364 patients in the continuous EEG Randomized Trial in Adults (CERTA) study, conducted at four centers in Switzerland. The study cohort included inpatients older than 18-years-old in intensive or intermediate care units who had impaired consciousness from any cause. Impaired consciousness was defined as a Glasgow Coma Scale (GCS) score of 11 or less or a Full Outline of Responsiveness (FOUR) score of 12 or less.
Patients in palliative care, those at risk for invasive procedures within 48 hours, and those having seizures within 36 hours or determined to be “status epilepticus” (SE) within 96 hours of randomization were excluded.
The researchers randomized the cohort to either continuous EEG lasting 30 to 48 hours, or regular EEG with two 20- to 30-minute recordings over 48 hours. Patients in both arms received standardized reactivity testing with loud sounds and axial nociceptive stimulations at least twice daily.
The primary outcome was mortality at six months. Secondary outcomes—assessed at four and six weeks through semi-structured telephone interviews with patients, relatives, or a treating physician—included detection of seizures and associated features and frequency of change in therapy. Changes in treatment included modification of antiseizure drugs or sedation—either started, stopped, increased, or decreased—triggered by the EEG results.
The researchers found no difference in mortality: at six months, 89 of 182 patients receiving cEEG had died, compared with 88 of 182 patients receiving rEEG. Detection of seizures and associated features was more common with cEEG, as was modification of antiseizure medication or sedatives.
Dr. Rossetti explained that the lack of a change in outcome despite more frequent detection of seizures and associated features is likely because patients with worse prognosis tended to have more changes in medication.
“Underlying brain pathology in some patients may be too strong to be reversed by treatment,” he said. “At least in some subjects, detection of seizures and medication changes are an epiphenomenon among those with more severe conditions, rather than a causal variable influencing the outcome.”
Moreover, detection of seizures and other features may have precipitated decisions to withdraw treatment in some patients, he said.
Among patients who did not have cardiac arrest, there was a small, nonsignificant difference in mortality favoring cEEG. Larger trials are needed to determine whether continuous EEG may be warranted for this and other subgroups, Dr. Rossetti said.
“This study should be the basis for future research to further investigate this issue and identify the categories of patients that really benefit from cEEG,” Dr. Rossetti said. “This would greatly help rationalizing available resources and increasing the benefit in those that really deserve cEEG.”
Experts who reviewed the study for Neurology Today pointed out that there is wide variation in the United States in the application of EEG for acutely ill patients with altered consciousness. In particular, the report may be useful for clinicians working in centers where resources are scarce.
“This study reassures us that intermittent EEG is a reasonable choice for patients with altered levels of consciousness from acute brain injury without evidence of recent clinical seizure activity,” said Holly E. Hinson, MD, MSR, FAAN, associate director of clinical research in the neurocritical care program at Oregon Health & Science University.
“Many community hospitals do not have access to cEEG,” Dr. Hinson said. “Even in tertiary care institutions, practice may vary with the ability, the location, and affiliation of the treating team. A dedicated neuro ICU might have a lower threshold for monitoring than a medical or surgical ICU. When seizures are witnessed clinically, the threshold for cEEG is much lower, but when they are not, choosing to monitor really depends on the setting and treating team.”
Noting that “mortality is a very blunt outcome,” Dr. Hinson said it is possible there were differences in cognitive and functional outcomes in patients not captured by mortality. “We know there were treatment differences, but these data do not fully explore the ramifications of those treatment differences,” she said.
She also agreed that more frequent detection of seizures and associated features might have precipitated withdrawal of treatment. “If cEEG is isoelectric or shows a lack of reactivity to stimulus, that might factor into how a family is counseled about prognosis and limitation of interventions. This is especially pertinent to comatose survivors of cardiac arrest.”
In an editorial accompanying the study, Emily Gilmore, MD, MS, associate professor in the division of neurocritical care and emergency neurology at Yale School of Medicine, wrote that the study should “not support the notion that cEEG should be abandoned or further restricted but rather studied with continued rigor and prioritized for those who need it most.”
“Monitoring isn't what changes outcome; it's what you do with the information that does,” Dr. Gilmore told Neurology Today.
While antiseizure medications were adjusted at a greater rate in the cEEG group, “we have no idea when and for how long seizures occurred or how they were managed,” she said.
Dr. Gilmore said patients with unexplained coma, moderate-to-severe traumatic brain injury, and ischemic and hemorrhagic strokes, as well as aneurysmal subarachnoid hemorrhage, may benefit from cEEG.
“Intermittent routine EEG or rapid EEG systems can be used to help triage which patients may benefit from cEEG within the same institution or need to be transferred to a higher level of care,” she said. “cEEG is not just a tool for detecting seizures or patterns that may be as metabolically demanding as seizures but is the only continuous real-time tool we have for also understanding pathophysiology—such as increased intracranial pressure, delayed cerebral ischemia or hyperemia.”
John Ney, MD, MPH, FAAN, neurologist and a clinical neurophysiologist at the Edith Nourse Rogers Memorial Veterans Hospital in Bedford, MA, agreed and noted that administrative claims studies have shown a mortality benefit to cEEG over rEEG during the hospitalization.
He echoed Dr. Hinson's remarks that mortality as an outcome obscures underlying pathology. “The study did not effectively randomize the underlying structural pathology, which is likely a greater predictor of mortality than whether or not someone was determined to have nonconvulsive seizures,” he said.
“Although this was a randomized study, I would hesitate to say that the cEEG benefit for mortality seen in prior studies should be discounted,” Dr. Ney said. “Instead, it may be that the prognostic value of inpatient usage of cEEG in altering the trajectory of a patient's post-hospital course may be limited where there is more severe structural neurological damage. That is not a reason to ignore seizures, or to not use cEEG when available to reliably detect them in patients with diminished consciousness or engage in active seizure management during the hospitalization.”
So, how can clinicians effectively triage patients when resources are limited? In the JAMA Neurology editorial, Dr. Gilmore wrote, “...risk stratification tools like the 2HELPS2B score, [which uses five EEG findings and one clinical finding (pre-EEG seizure to estimate risk for seizure in hospitalized patients]...can be used to allocate resources to benefit those at highest risk but also ideally with the greatest potential for recovery.”
Dr. Gilmore explained that the 2HELPS2B scale, like the ABCD2 risk assessment for patients with a transient ischemic attack, [which factors in blood pressure, clinical features, duration of TIA, and presence of diabetes] can indicate an inflection point for triage—especially in centers where resources are scarce.
“Patients with acute brain injury who are at a higher risk of secondary brain injury from seizures that will potentially impact their prognosis and recovery trajectory may be prioritized over patients with a lower risk of seizures without acute brain injury or the patient with devastating brain injury, such as global hypoxic ischemic brain injury with impending herniation for whom seizure treatment will have no impact,” she wrote.
Dr. Gilmore echoed the need for further research looking at subgroups and outcomes other than mortality. “There was no comment on effects of seizures on cognition or rates of epilepsy,” she said. “In addition, since we have no idea about how patients were treated, perhaps the next trial also needs to integrate management [into their analysis].”
Drs. Rossetti, Hinson, Gilmore and Ney had no relevant disclosures.