Utilization of Continuous Electroencephalography in the Intensive Care Unit

TCP1 – Task 2 – Research Paper
Patients are admitted to the intensive care unit (ICU) of a hospital for a myriad of disease and injury conditions. Critically ill patients often present with a specific serious health issue that needs to be addressed, however, can often evolve to include secondary problems. Secondary cardiac issues develop from a long list of unrelated heart conditions, and as such, all ICU patients today are placed on bedside electrocardiogram (EKG) monitoring. Years of research, advancements in technology, and international implementation of the service has led to bedside EKG being the standard of care for all ICU patients, regardless of the primary diagnosis. The heart is not the only organ that can often have secondary issues. For decades physicians have tracked and noted secondary neurologic changes in their critically ill patients, but lacked research-based evidence on what caused the neurologic change onset, and how best to monitor and watch for those changes. Over the past two decades, a lot of research has been conducted looking into this specific issue. Research now suggests that the use of continuous electroencephalography (cEEG) for patient populations in the ICU is an effective, affordable, valuable, and prognostic diagnostic tool to evaluate cerebral function, detection of epileptic activity, and to monitor neurologic changes. Electroencephalography (EEG) is a diagnostic tool used to monitor activity of the brain, and has been in common practice since the 1960s (Koenig, Kaplan, & Thakor, 2006). The early use of EEG was solely evaluation of spontaneous electrical discharge from neurons on the brain, in order to detect for seizures and guide antiepileptic drug (AED) medication for management of seizures. EEG machines were once analog instruments, large and bulky in size. The electrical activity of the neurons were transmitted to the instrument, where finely calibrated ink pens would map out the speed and voltage of the activity, in a trace or pattern on a roll of paper. Advancements in computer technology has transformed the modern day EEG machine into a digital computer, where the patterns are now displaced and recorded on monitors small enough to sit at the bedside of a patient.
Just as the technology advanced, application of the service and utilization for additional monitoring reasons expanded as well. EEG testing began to be utilized to detect cerebral ischemia, assess level of consciousness, monitor sedation levels of patients in induced comas, and localize onset of epileptic activity for surgical resection of diseased cortical tissue (Harris, 2014). However, the utilization of the EEG was in a routine or intermittent length of time, capturing anywhere from thirty minutes to an hour of cortical activity, leaving roughly twenty-three hours of cortical activity unmonitored.
In the mid to late 1990s, neurologists from the University of California Los Angeles (UCLA) began to question the clinical efficacy of utilizing EEG monitoring on a continuous bases versus routine or intermittent length. Patients admitted to the ICU, who showed no clinical signs of seizure, failed to improve neurologically. Additionally, patients placed into medication induced comas following cardiac arrest, or those in unexplained comas, failed to wean successfully from ventilators and properly awaken. The neurologists began to order EEGs lasting longer than one hour, and were surprised to find that patients were suffering from nonconvulsive seizures (NCS) and of nonconvulsive status epilepticus (NCSE) (Vespa, Nenov, & Nuwer, 1999). In both of these situations, the brain is abnormally discharging electrically inducing a seizure, but showing no outward clinical manifestation. Undetected and untreated with AEDs, both situations are extremely dangerous and mortal.
The Vespa, Nenov, and Nuwer (1999) study stated the following:
Never before in the practice of neurology and neurosurgery has an emphasis on timely diagnosis and intervention been as poignant as in the past 2 years. At the same time, development of specialized intensive care units (ICUs) and stroke units has continued to burgeon, and framework of how to deliver care and when to deliver this care is evolving. It is in this context that continuous electroencephalographic monitoring (cEEG) and other forms of neurophysiologic monitoring have begun to develop at a time when timely information may be able to make a significant difference in outcome. (p.2). cEEG would now monitor the electrical activity of the brain for extended periods of time, typically twenty-four to seventy-two hours, providing physicians with a more complete picture of the cortical function of a patient, and the ability to track and manage neurologic care.
EEG testing is labor intensive, and the technical testing has to be carried out by a highly skilled technologist with expertise in the field of neurodiagnostics. The American Board of Registration of Electroencephalographic and Evoked Potential Technologists is the governing body that conducts the examination and testing for registration of an EEG Technologist (R. EEG T.). In the fifty years of credentialing technologists, there are fewer than 5500 R. EEG T. in the country (ABRET, 2015). This labor shortage creates limitations on hospitals to recruit and employ skilled technologists to provide the cEEG studies. In addition, the professional testing is conducted by a neurologist, usually fellowship-trained in epilepsy monitoring, who makes the final read and dictation of the EEG study. In order for a patient to receive a cEEG exam, the technologist applies twenty one specialized electrodes, in predetermined locations according to the international 10/20 system of measuring, performs a calibration of the machine and recording, and then monitors the continued real-time activity of the electrical output of the brain (ACNS, 2006). The neurologists is then responsible for review of the entirety of the study, and provides a professional interpretation of the patterns that are recorded, looking specifically for any abnormalities that may be present. This makes performing a cEEG test not only labor intensive, but time intensive as well.
Sensitive to the labor and time intensive aspects of providing cEEG, neurologists across the World began discussing which patient populations within the ICU would benefit the most from cEEG and have improved outcomes, narrowing the focus onto those patient populations. Dr. Lawrence J. Hirsch authored an article in late 2004 that listed six indications for cEEG monitoring. According to Hirsch (2004), “detection of NCS and NCSE, characterization of spells, assessment level of sedation following trends, management of burst-suppression in anesthetic coma, detection of ischemia, and prognostication” (p. 333), are the most common indications for cEEG monitoring. Expounding upon those six indications, the European Society of Intensive Care Medicine (ESICM), sought out to recommend specific critically ill patients in the ICU that should receive cEEG. The group decided to evaluate every article from the PubMed database that provided quality evidence-based indications where cEEG monitoring in the ICU would or would have increased the positive outcome for a patient. They focused on the underlying etiology that patients first presented to the ICU with, in an attempt to prevent secondary evolution of neurologic changes, and categorized them as to the objective cEEG would supply. The group first focused on Hirsch’s primary indication for utilizing a cEEG, detection of NCS and NCSE, and then systematically worked their way through each category. The result was that they recommended cEEG in the following patient etiology populations: traumatic brain injury, intracerebral hemorrhage, cardiac arrest, encephalitis, comatose patients without primary brain injury, post convulsive status epilepticus, and acute ischemic stroke. For the indication of detecting ischemia, the group recommended cEEG only for patients with a subarachnoid hemorrhage when a clinical examination is unreliable. For indication as a prognostic tool, the recommendation was for all cardiac arrest patients, unexplained comatose patients, and those with toxic and metabolic encephalitis (Claasen, et al, 2013).
The combination of Hirsch’s indications for cEEG and the ESICM group’s recommendation for specific patient populations allowed clinicians to now narrow the scope and utilization of cEEG testing and to prioritize which patients receive the limited personnel resources. Now that patient populations were identified, the next question became how long should the patients be hooked up and monitored to detect the primary indication for NCSs? A study conducted in 2004 retrospectively looked at 110 critically ill patients who had undergone cEEG for the detection of seizures. Of those, approximately 51% were successfully detected in the first hour of recording, 87% within the first 24 hours of recording, and 7% were not detected until more than 48 hours of monitoring. The primary contributing factor to the delay in detection was for patients who were in coma upon initial clinical exam. This study concluded that for patients in a non-coma status, 24 hours of cEEG would be adequate; however, in comatose patients, cEEG would need to be recorded for a minimum of 48 hours, and possibly longer (Kull & Hirsch, 2004). When cEEG it utilized for Hirsch’s other indications, such as assessment of level of sedation and prognostication, those durations are dictated by the management of care. When being used to assess the level of sedation induced by anesthetic drugs, the cEEG monitoring will be utilized for the duration that medication is continued, as directed by anesthesia, pharmacy, nursing, or physician management.
Although the cost of healthcare is an ever changing number, the overall cost of cEEG monitoring is approximately 1% of a patient’s total hospital cost. However, utilization of cEEG is more effectively measured as a cost-avoidance test. When cEEG is utilized, and physicians are able to more timely manage the secondary neurological insult of their patients, length of stay is dramatically decreased. In one study the median length of stay for traumatic brain injury patients with cEEG monitoring declined from 24.3 days to 13.6 days. Decreased patient stays helps to decrease overall hospital cost. The same study showed the average hospital cost declined from $88,690 to $49,578 (Vespa, Nenov, & Nuwer, 1999). Those are declines that are tangible and immediately have an impact on the financial state of a hospital institution. Non-tangible financial impacts though, are that once discharged and sent home, patients are able to rehabilitate and become active members of society again, working, producing, and consuming, all of which add financial gains downstream versus patients who were not monitored with cEEG and either remain in the hospital for longer periods of time, accruing more charges, or those that pass away and never leave.
With the efficacy of cEEG proven in terms of outcomes and finances, recommendations are needed to maximize the labor shortage while providing safe and effective technical monitoring. A staffing model was presented by staff members of Indiana University Health, the nation’s first hospital to provide real-time 24-7-365 coverage of cEEG monitoring. Their neurophysiology department is staffed by a team of technologists that operates on two 12-hour shifts, from 7:00 am to 7:00 pm and 7:00 pm to 7:00 am. The team supplies monitoring support to five facilities that make up their Academic Health Center, which includes an adult level 1 trauma center, the states only pediatric level 1 trauma center, the states only specialized epilepsy center, the nation’s largest adult Neurocritical care unit, and a combined seventeen ICUs. Administration established a protocol for monitoring, in order to maintain high patient safety, in a 4:1 patient to technologist ratio (Thomas, Curtis, Adams, Brown, & Lau, 2014). One technologist is able to real-time monitor the cEEG of up to four patients at any given time.
The American Neurodiagnostic Society (ASET) states “. . . in order to promote and maintain a safe patient environment and best practice in the Epilepsy Monitoring Unit, the patient should never be left unattended” (ASET, 2013). That means there needs to be personnel resources to provide the monitoring in the 4:1 ratio, with separate personnel able to provide the technical hookup and application of the EEG electrodes. A group from Duke University conducted research on determining the quality of electrode lead placement, when done by a nontechnologist compared to a technologist-applied hookup. The nontechnologists who lack the necessary education and skill training to provide a standard measured international 10/20 electrode hookup, instead utilized a special template system in the form of an electrode cap. The electrode cap was available in different sizes depending on patient head circumference, and the overlaying template showed nontechnologists precisely where and which electrodes to place. The nontechnologists were still given a brief period of training on identification of the particular electrodes, techniques for applying and securing the electrodes, and model heads to practice the application process prior to placing on live patients. The technologists applied their hookups according to standards set forth by the American Clinical Neurophysiological Society (ACNS) through a series of technical guidelines (ACNS, 2006). The EEG application was blind to the neurologists providing the professional reads. Analysis of the impedance for both forms of EEG hookup were tracked. Subjective quality assessments from the neurologists were also tracked. The results showed no significant difference in quality between those applied using the template versus the technologist-applied hookup (Kolls, et al, 2012). Skull deformities, ventriculoperitoneal shunts, open wounds, recent craniotomies, and several other scenarios would prevent a patient from being eligible for the template electrode cap application. However, there are many patients who have no cranial anomalies and utilizing ICU staff in the form of nurses or nurse aids, would free up the burden on the R. EEG T. to focus on continued patient monitoring. By allowing nontechnologists to provide the cEEG hookup, delay in initiating cEEG monitoring is also decreased.
There are additional technological challenges to providing cEEG monitoring, many of which are specific to the ICU setting. The ICU is an environment filled with advanced computer instrumentation, medical devices, equipment, and electrical machines. All of these combined make the ICU a high source for electrical artifact. All equipment that is powered through electrical means, leaks electrical current into the surrounding atmosphere. The EEG electrodes placed on the scalp are designed to pick up the electrical activity produced by the neurons of the brain, amplify the signal so it can be read, and transmit the signal back to the digital computer. When ICUs have large sources of electrical leakage, they can be picked up by the sensitive EEG electrodes and become a source of artifact on the EEG recording. There are precautions taken to minimize this electrical interference, such as grounding the recording machine, using specialized coated electrode wires, and applying filters on the digital equipment that cancel out and block electromagnetic artifact that is usually transmitted in 60 Hz noise (Kaplan, 2006). Electrical artifact is only one of many forms of artifact that can be transmitted and recorded on the EEG. Common sources of known artifact include: intravenous (IV) drip artifact, respirator artifact caused by artificial mechanical ventilation, sweat artifact caused by excessive patient perspiration of the scalp, muscle artifact caused by increased temporal tension and activity, ballistocardiographic artifact caused by pulsation of an artery near a recording electrode, movement artifact caused by nurses adjusting and cleaning the patient, and physiological artifact caused by chewing, talking and blinking. Identification and correlation of these artifacts, along with troubleshooting skills to eliminate them are essential to the quality and integrity of cEEG monitoring.
Due to the many volatile artifacts that can obscure or compromise the recording of cEEG, many neurologists advocate for, and support the use of, simultaneous time-locked video recording. Because the patient population in the ICU is often unable to provide an adequate oral clinical neurologic exam, and many of their seizures are from the nonconvulsive category, video cEEG allows the professional reader to correlate EEG activity with patient and environment activity. Some patterns observed while a patient is asleep or has their eyes closed can be considered normal, however, if the same pattern is seen during the waking state, the pattern would be deemed abnormal. Video cEEG allows the professional reader to visualize the patient and in this scenario determine if the patient’s eyes are in fact open or closed, awake or asleep. Many medications, including all barbiturates, cause a dramatic effect on the frequency and activity of the EEG. Video cEEG allows the professional reader to observe when medication was administered to the patient, and correlate with the exact time period that the EEG was altered or affected. When monitoring specifically for epilepsy, not all seizures manifest themselves into complete body jerking and thrusting. Some seizures are very subtle and consist of only small rhythmic movements, such as an index finger moving or the corner of the mouth twitching. It is for these reasons and more that Hirsch recommends video recording be mandatory in all cEEG at his facility, for every recording unless a technical reason makes it unavailable (Kull & Hirsch, 2004).
When patients are admitted to the ICU, one of the primary goals for all treating clinicians is to determine the prognosis in a timely manner. Prognostic data forecasts the likely outcome for the patient and their chance of either recovery or death. For many individuals and families, knowing the prognosis will help direct the course of action for determining lifesaving efforts. For clinicians, it helps determine what hospital resources, effort, and time to apply to patients to promote the best chance for recovery. A patient with a very poor prognosis receives less hospital resources and lifesaving effort compared to a patient with a very high or positive prognosis. Many variables combine together to help determine a patient’s prognosis. They include age, gender, diagnosis, imaging studies, lab results, and EEG. The EEG is a very valuable diagnostic tool that can help clinicians determine the prognosis, by associating certain patterns and grading them as to their influence on prognosis, similarly to how a Glasgow Coma Scale is used (Kramer, Jette, Pillay, Federico, & Zygun, 2012). One of the most highly sought patient populations to determine prognosis, is for patients who are comatose following cardiac arrest. Physicians will utilize the cEEG to look for specific patterns that are associated with poor outcomes. The EEG is classified according to pattern findings and given a grade between I and V, with I being high-grade outcome and V being low-grade outcome. Grade I would resemble a normal adult pattern activity dominant for alpha rhythm and reactive to stimulation. Grade V would be isoelectric, resembling a very flat EEG with no rhythm (Koenig, Kaplan, & Thakor, 2006). Between the spectrum I-V there are many additional patterns that aid in determining a patient’s prognosis and setting the time frame for recovery. Periodic lateralized epileptiform discharges, burst-suppression, triphasic waves, electro cerebral silence, and alpha and theta coma all support poor outcome for prognosis. Furthermore, the EEG can be noted to change dramatically during the course of monitoring and intervention by the clinical team. Burst suppression carries a poor prognosis when seen in a patient who is comatose following cardiac arrest, however, it is the desired pattern when a patient is strategically placed into a drug induced hypothermic state. It is not the single pattern during a lone time frame window, but the overall progression and evolution of the EEG as a whole that is best looked at when determining prognostic value.
The utilization of cEEG in the setting of the ICU is still a relatively new testing procedure, and not widely implemented across the nation. Research over the past two decades has transformed the test from merely a routine hour exam to detect for seizures, to a highly valuable diagnostic tool providing a real-time assessment and look into the function, ischemic, and electric activity of the brain, on a continuous basis. Advancements in computer technology, innovation in delivery and application of electrodes, declines in patient length of stay and total hospital costs, and the prognostic ability of the test, have all catapulted cEEG into becoming a growing standard of care for patients admitted in the ICU. There continues to be a personnel shortage of trained and registered EEG technologists, but staffing models have improved the maximization of technologist numbers while maintaining safe monitoring protocols. With the amount of research that continues to be conducted, the utilization of cEEG and its many valuable applications to clinical management of care will continue to expand.

References
ABRET – American Board of Registration of Electroencephalographic and Evoked Potential Technologists, (2015). Newly Certified Technologists. Retrieved 22 January 2015, from http://abret.org/employers/newly_certified/
ACNS – American Clinical Neurophysiology Society, (2006). Guideline One: Minimum Technical Requirements for Performing Clinical Electroencephalography. Retrieved 22 January 2015, from http://www.acns.org/practice/guidelines
ASET – The Neurodiagnostic Soceity, (2013). LTM Position Statement: long-term monitoring (LTM) for epilepsy. Retrieved 23 January 2015, from http://www.aset.org/i4a/pages/index.cfm?pageid=3992
Claassen, J., Taccone, F. S., Horn, P., Holtkamp, M., Stocchetti, N., Oddo, M., (2013). Recommendations on the use of EEG monitoring in critically ill patients: consensus statement from the neurointensive care section of the ESICM. Intensive Care Medicine, 39, 1337-1351.
Harris, C., (2014). Neuromonitoring Indications and Utility in the Intensive Care Unit. Critical Care Nurse, 34[3], 30-40.
Hirsch, J. L., (2004). Continuous EEG Monitoring in the Intensive Care Unit: An Overview. Journal of Clinical Neurophysiology, 21[5], 332-340.
Kaplan, P. W., (2006). EEG MONITORING IN THE INTENSIVE CARE UNIT. American Journal of Electroneurodiagnostic Technology, 46, 81-97.
Koenig, M. A., Kaplan, P. W., & Thakor, N. V., (2006). Clinical Neurophysiologic Monitoring and Brain Injury from Cardiac Arrest. Neurologic Clinics, 24, 89-106.
Kolls, B. J., Olson, D. M., Gallentine, W. B., Skeen, M. B., Skidmore, C. T., & Sinha, S. R., (2012). Electroencephalography Leads Placed by Nontechnologists Using a Template System Produce Signals Equal in Quality to Technologist-Applied, Collodion Disk Leads. Journal of Clinical Neurophysiology, 29, 42-49.
Kramer, A. H., Jette, N., Pillay, N., Federico, P., & Zygun, D. A., (2012). Epileptiform Activity in Neurocritical Care Patients. Canadian Journal of Neurological Sciences, 39, 328-337.
Kull, L. L., & Hirsch, L. J., (2004). Continuous EEG Monitoring in the Intensive Care Unit. American Journal of Electroneurodiagnostic Technologists, 44, 137-158.
Thomas, J. A., Curtis, C. M., Adams, L. C., Brown, S. L, Lau, R. R., (2014). Indiana University Health Staffing Model for Neurotelemetry and Epilepsy Monitoring Unit Patient Populations: Part 1. Neurodiagnostic Journal, 54, 68-74.
Vespa, P. M., Nenov, V., & Nuwer, M. R., (1999). Continuous EEG Monitoring in the Intensive Care Unit: Early Findings and Clinical Efficacy. Journal of Clinical Neurophysiology, 16, 1-13