Periodic Assessment of the Use of Continuous EEG Monitoring

Continuous EEG Monitoring: A Survey of Neurophysiologists and Neurointensivists

Gavvala J, Abend N, LaRoche S, Hahn C, Herman ST, Claassen J, Macken M, Schuele S, Gerard E; Critical Care EEG Monitoring Research Consortium (CCEMRC). Epilepsia 2014;55(11):1864–1871.

OBJECTIVE: Continuous EEG monitoring (cEEG) of critically ill adults is being used with increasing frequency, and practice guidelines on indications for cEEG monitoring have recently been published. However, data describing the current practice of cEEG in critically ill adults is limited. We aimed to describe the current practice of cEEG monitoring in adults in the United States. METHODS: A survey assessing cEEG indications and procedures was sent to one intensivist and one neurophysiologist responsible for intensive care unit (ICU) cEEG at 151 institutions in the United States. At some institutions only one physician could be identified. RESULTS: One hundred thirty-seven physicians from 97 institutions completed the survey. Continuous EEG is utilized by nearly all respondents to detect nonconvulsive seizures (NCS) in patients with altered mental status following clinical seizures, intra cerebral hemorrhage (ICH), traumatic brain injury, and cardiac arrest, as well as to characterize abnormal movements suspected to be seizures. The majority of physicians monitor comatose patients for 24–48 h. In an ideal situation with unlimited resources, 18% of respondents would increase cEEG duration. Eighty-six percent of institutions have an on-call EEG technologist available 24/7 for new patient hookups, but only 26% have technologists available 24/7 in-house. There is substantial variability in who reviews EEG and how frequently it is reviewed as well as use of quantitative EEG. SIGNIFICANCE: Although there is general agreement regarding the indications for ICU cEEG, there is substantial interinstitutional variability in how the procedure is performed.

Commentary

Over time, clinical evidence accumulates in a given area of medicine. When substantial, high-quality evidence is available, it can be summarized in systematic reviews, and from these, guidelines can be written that inform clinical practice. Not uncommonly, the accumulated evidence is insufficient to recommend a specific therapy or practice. To close this gap, expert opinion, or alternatively, consensus statements—an agreement of expert opinion—fills the void. Such is the current state of continuous EEG monitoring (cEEG).

Increased use of cEEG has accompanied the rise of neurocritical care as an important subspecialty of neurology. Some estimates have suggested that the use of cEEG in the intensive care unit (ICU) has quadrupled over the past decade (1). Most neurologists and intensivists are aware of the high rates of subclinical seizures detected when critically ill patients—especially neurologically critically ill patients—undergo cEEG. Although detection and management of seizures in critically ill patients is the most common use of cEEG, the technique has other potential uses, including detecting ischemia in patients at high risk following subarachnoid hemorrhage, monitoring pharmacologic sedation in patients with elevated intracranial pressure or undergoing treatment of refractory status epilepticus, and assessing the severity of brain injury and prognosis after traumatic brain injury or hypoxic ischemic injury. What is the ideal role for cEEG in managing critically ill patients?

In 2012, the Neurocritical Care Society published guidelines for the treatment of status epilepticus (2). These guidelines included some limited recommendations for the use of cEEG in this specific setting. The guidelines recognized broad indications for the use of cEEG to detect possible seizures, including to assess patients with 1) recent seizures who had not returned to baseline; 2) coma, including coma after cardiac arrest; 3) epileptiform or periodic discharges on routine EEG; 4) intracranial hemorrhage of various types; and 5) suspected nonconvulsive seizures in patients with altered mental status. Although based on low quality of evidence, the guidelines made strong recommendations to use cEEG to direct therapy in patients with non-convulsive status; to start cEEG studies promptly, that is, within 1 hour of suspected seizures; to continue cEEG for at least 48 hours in comatose patients; and to have studies interpreted by an experienced and qualified EEG reader.

How closely are these guidelines being applied in practice? Health care systems and physician behavior are notoriously slow to align with new guidelines, even when the level of evidence is strong. How extensively is cEEG being used for other, nonseizure indications?

It is in this context that Gavvala and colleagues, on behalf of the Critical Care EEG Monitoring Research Consortium, surveyed existing practices of 245 neurophysiologists (N-PHYSs) and neurointensivists (N-INTs) at 151 relatively large, experienced U.S. institutions near the end of 2012. In addition to studying the indications for cEEG in practice, they also wished to assess current practices regarding duration of cEEG monitoring, technical resources, and intensity of cEEG review.

The survey response was reasonably good; it included nearly two-thirds of invited institutions and more than half of physicians. Not surprisingly, cEEG was essentially universally used to assess and manage status epilepticus and nonconvulsive seizures, as recommended in the 2012 guidelines. In addition, cEEG was commonly used by N-PHYSs and N-INTs (58–72%) for titration of burst suppression in the management of elevated ICP. However, only a minority of physicians used cEEG to identify vasospasm (15–27%) or to detect elevated intracranial pressure (ICP) (0–22%). N-PHYSs and N-INTs largely used cEEG for similar indications, with only minor exceptions. Respondents who were both an N-PHYS and an N-INT were more likely to monitor patients with traumatic brain injury and altered mental status. No N-INTs surveyed used cEEG to detect elevated ICP; in contrast, its use was limited in N-PHYSs.

In addition to surveying existing practices, the investigators also queried respondents about ideal practices, that is, how they would use cEEG if resources were unlimited. Respondents reported that they would increase use of cEEG in this ideal world for seizure and nonseizure indications. The gap between real and ideal was greatest for use of cEEG to monitor for vasospasm. Currently, approximately 20% of respondents used cEEG for this indication, but 45% would use it under ideal circumstances.

A wide range of cEEG equipment availability and study volume was reported among centers, but overall, 43% of centers reported higher cEEG volume than in the preceding year. EEG technologists were available 24/7 at 86% of centers surveyed, more often on-call than in-house, and 77% of centers reported being able to start urgent studies at any time; nearly all reported that this would be an ideal goal.

Virtually all centers (95%) could perform remote review of studies, but frequency of cEEG review was highly variable. Review of cEEG for nonconvulsive seizures was performed daily (23% of sites), twice daily (35%), three times daily (25%), four times daily (11%), or nearly continuously (5%). The EEG technologist often participated in cEEG review, reviewing the entire record at about one-fourth of centers, but less than half of the record at the majority (56%) of institutions. The N-PHYSs typically reviewed the entire cEEG at 50% of centers and at least half at 80%. Half of surveyed centers used quantitative EEG tools. All sites were able to do emergent interpretation; 80% of the time this was performed by the N-PHYS.

This snapshot of current cEEG practices is timely because it serves as a basis for comparison with the expert consensus statement that was recently published by the Critical Care Continuous EEG Task Force of the American Clinical Neurophysiology Society (ACNS) (3, 4). In publishing the statement, this group recognized that the current state of evidence better supports a consensus statement than formal evidence-based guidelines.

The consensus statement enumerates in detail the recommended indications for use of cEEG to detect nonconvulsive seizures or status (available on the ACNS website at http://ow.ly/M21iO). These indications largely align with current practice at most centers, as reported by the Gavvala et al survey, though cEEG may be more selectively applied at some centers than recommended by the consensus statement. Despite the marked growth in cEEG, rather than overuse, it appears that many centers still fall short of following consensus guidelines in every case. The finite equipment and personnel resources identified in the survey likely explain part of the gap between current practice and the ideal.

The consensus recommendations note the lack of solid evidence linking cEEG use to outcome, but justify its use on the solid logic that poorly controlled seizures may cause neuronal injury, are associated with poor outcome, and require cEEG to guide the treatment that minimizes this risk. Continuous review of ongoing cEEG was thought to be ideal, but the consensus statement recognized the limitations of resources that produced the variable frequency of cEEG review seen in the survey. Given these limitations, tailoring the intensity of review to the frequency of nonconvulsive seizures may be an interim measure to efficiently deploy resources.

In contrast to seizure indications, cEEG for identification of ischemia or to monitor sedation received a weaker “suggested” endorsement as an adjunct to neurologic examination and other studies. The consensus report recognized the limits of cEEG to determine severity of encephalopathy and prognosis after neurologic injury. The relatively weaker evidence for these indications matches the more limited use of cEEG in these patients at surveyed centers.

The iterative process of studying an area of medicine, gathering evidence, writing consensus statements, or, when supported by strong evidence, writing clinical guidelines, helps to direct the efficient and effective use of such resources as cEEG. In many cases, the survey findings of Gavvala et al. appear aligned with current consensus recommendations. Resource limitations and perhaps other factors may explain why not all centers consistently study all patients for the recognized indications. Practices outside of these large, active centers are even less well understood and are likely more variable. Future studies, including those that assess the relationship between cEEG and clinical outcome, will contribute to this iterative process and inform future practice.