Predictive value of the bispectral index for burst suppression on diagnostic electroencephalogram during drug-induced coma.
Study Purpose: To determine correlation and predictive value between data obtained with the bispectral index (BIS) and diagnostic electroencephalogram (EEG) in determining degree of burst suppression during drug-induced coma. This study seeks to answer the question: "To what degree can EEG suppression and burst count as measured by diagnostic EEG during drug-induced coma be predicted from data obtained from the BIS such as BIS value, suppression ratio (SR), and burst count?" Background/Significance: During drug-induced coma, cortical EEG is the gold standard for real-time monitoring and drug titration. Diagnostic EEG is, from setup through data analysis, labor intensive, costly, and difficult to maintain uniform clinician competency. BIS monitoring is less expensive, less labor-intensive, and easier to interpret data and establish/ maintain competency. Validating BIS data versus diagnostic EEG facilitates effective brain monitoring during drug-induced coma at lower cost with similar outcomes. Method: This is a prospective, observational cohort study. Four consecutive patients receiving drug-induced coma/EEG monitoring were enrolled. BIS was initiated after informed consent. Variables recorded per minute included presence or absence of EEG burst suppression, burst count, BIS value over time, burst count, and SR. Pearson's product-moment and Spearman rank coefficient for BIS value and SR versus burst count were performed. Regression analysis was utilized to plot BIS values versus bursts/minute on EEG as well as SR versus burst count on EEG. EEG/BIS data were collected from digital data files and transcribed onto data sheets for corresponding time indices. Results: Four patients yielded 1,972 data sets over 33 hours of EEG/BIS monitoring. Regression coefficient of 0.6673 shows robust predictive value between EEG burst count and BIS SR. Spearman rank coefficient of -0.8727 indicates strong inverse correlation between EEG burst count and BIS SR. Pearson's correlation coefficient between EEG versus BIS burst count was .8256 indicating strong positive correlation. Spearman's rank coefficient of 0.8810 and Pearson's correlation coefficient of .6819 showed strong correlation between BIS value versus EEG burst count. Number of patients (4) limits available statistics and ability to generalize results. Graphs and statistics show strong correlation/predictive value for BIS parameters to EEG suppression. Conclusions: This study is the first to measure correlation and predictive value between BIS monitoring and diagnostic EEG for degree of EEG suppression and burst count in the adult population. Available statistic tests and graphing of variables from BIS and diagnostic EEG show strong correlation and predictive value between both monitoring technologies during drug-induced coma. These support using BIS value, SR, and burst count to predict degree of EEG suppression in real time for titrating metabolic suppression therapy.
Keywords: bispectral index, EEG burst suppression, metabolic suppression therapy, suppression ratio
The purpose of this study is to determine degree of correlation and predictive value between the monitoring parameters obtained with the bispectral index (BIS) and diagnostic electroencephalogram (EEG) in determining level of burst suppression during drug-induced coma. BIS parameters including suppression ratio (SR), burst count, and BIS value will be compared with level of burst suppression (burst count per minute) obtained on diagnostic EEG monitoring. This study seeks to answer the question: "To what degree can EEG suppression and burst suppression be predicted from BIS value, SR, and burst count?" This study fills gaps within the available literature and possibly supports using the BIS to monitor metabolic suppression therapies/burst suppression in select patients. The research hypothesis is that, during drug-induced coma, BIS value, SR, and burst count are predictive for level of burst suppression seen on diagnostic EEG during drug-induced coma.
The gold standard for assessing stability of neurological function remains the clinical neurological evaluation. There are clinical situations where the ability to perform an adequate neurological examination is limited. One situation is after loss of consciousness subsequent to structural or metabolic brain injury. A second situation is neuromuscular blockade therapy where the patient may appear asleep but be potentially awake and in pain. A third situation is drug-induced coma utilizing barbiturates or propofol for metabolic suppression. In these situations, electrophysiologic monitoring of the brain with cortical electroencephalogram (EEG) is an effective means to directly assess stability of the central nervous system (CNS). With cortical EEG being the gold standard of electrophysiologic brain monitoring, it generates a significant amount of data. It may be utilized for determination of seizure focus; audio, visual, or somatosensory evoked potential monitoring; and differentiation of coma states such as alpha coma and as an adjunctive tool in brain death protocols. When diagnostic EEG is utilized for monitoring level of burst suppression in drug-induced coma, far more information may be obtained than is necessary. Using neurophysiologic monitoring more narrowly tailored to display information required can streamline burst-suppression monitoring and be more easily integrated into real-time drug titration. BIS monitoring is one such monitoring device, with display parameters easily interpreted and not requiring additional personnel and expense from setup through data acquisition and display.
Physiologic Basis for the EEG
Approximately 50% of metabolic activity of the brain is directed toward generating and conducting neuronal electrical activity with the balance maintaining cellular structural integrity. Alterations in metabolic state of the brain can be reflected in the cortical EEG. At the most basic level, the cortical EEG reflects electrical activity in the cerebral cortex. Electrical activity measured and recorded at the cerebral cortex is modulated by multiple factors. First, electrical discharges from the thalamus are regulated by relative proportions of excitatory and inhibitory neuronal connections. Second, presynaptic neurotransmitter release facilitates nerve impulse generation and conduction. Multiple energy-consuming steps need to occur to maintain stable brain metabolism and stability of brain metabolism as reflected in the cortical EEG (Arbour, 2003, 2006, 2009; Bader, Arbour, & Sylvan, 2005; Purdon et al., 2013; Wallace, Wagner, Wagner, & McDeavitt, 2001).
Metabolic or structural brain injury as well as variables including multiple drug therapies affects cerebral metabolism and EEG activity. Pharmacological agents affecting brain metabolism and the EEG include benzodiazepines, propofol, and barbiturate therapy (Arbour, 2004, 2009,2013; Friedman, Claassen, & Hirsch, 2009; Honeybul, 2011; Kennedy & Gerard, 2012; Purdon et al., 2013). EEG has long been utilized in monitoring clinical situations including drug-induced coma and effectiveness of anticonvulsant therapy (Arbour, 2013; Friedman et al., 2009; Honeybul, 2011; Kennedy & Gerard, 2012).
Limitations of diagnostic EEG monitoring include size of monitoring technology and electrode application and maintenance as well as difficulty inteipreting the volume of information obtained (Arbour, 2003; Rosow & Manberg, 1998). Dose-dependent changes in EEG activity occurring with administration of sedativehypnotic, opioid, or anesthetic agents include augmentation of beta (high-frequency) activity, correlating with initial stages of sedation such as amnesia or anxiolysis. Further dose increases produce additional decreases in level of consciousness coupled with EEG changes including increased proportion of slow wave (delta and theta) activity. High-dose therapy with agents such as barbiturates or propofol can produce varying degrees of EEG suppression including burst suppression correlating with a state of deep unconsciousness (Arbour, 2003; Bader et al., 2005; Friedman et al., 2009; Kennedy & Gerard, 2012; Purdon et al., 2013; Rhoney & Parker, 2001).
Relationship Between the BIS and EEG
Diagnostic EEG has long been utilized to study drug effects on the brain and monitor drug-induced coma. One option for brain monitoring during drug-induced coma is a derived EEG parameter: the BIS. The BIS was initially studied and utilized in the operating room setting and designed to assess responses to sedation and anesthesia. Over time, it is utilized more frequently in monitoring sedation and metabolic suppression in critically ill adults. BIS determination begins with acquisition of cortical EEG from a frontal-temporal montage, which is then passed through multiple processing steps, producing the specific BIS value. The BIS value is displayed on a linear (0-100) scale and shows correlation with clinical assessment of sedation. A BIS value approaching 100 typically corresponds with an awake state. A BIS value at or near 0 indicates minimal or no brain activity. Additional clinical end points and corresponding BIS values are found on a continuum between these extremes and are well described within the professional literature (Arbour, 2003, 2004, 2009; Bader et al., 2005; Rosow & Manberg, 1998). Correlation between specific EEG state, clinical end points, and corresponding BIS value is found in Figure 1 (available as Supplemental Digital Content I at http://links.lww.com/JNN/A23). BIS is a derived EEG parameter, and under normal conditions, EEG activity and BIS value can change in response to stimulation and altered metabolic states such as cerebral ischemia, brain trauma, metabolic suppression, and hypothermia (Arbour, 2003, 2004, 2009; Bader et al., 2005). Additional parameters monitored with BIS include electromyographic (EMG) activity, SR, and burst count. EMG activity is produced by electrical impulses generated by facial musculature, is typically high frequency in nature, and may produce an elevation of BIS value independent of hypnotic state (Arbour, 2003, 2004, 2009; Bader et al., 2005). The SR represents the percentage of suppressed EEG activity within the previous 63-second interval. For example, if 25% of the EEG activity within the previous 63-second sample is suppressed, the SR is 25. If there is total EEG suppression consequent to catastrophic brain injury or significant overdosing of barbiturate therapy, the SR may approach 100 (Arbour, 2003, 2004, 2009; Bader et al., 2005). The burst count reflects the number of bursts per minute during drug-induced coma. For a comprehensive discussion of BIS derivation, there are multiple excellent references within the professional literature (Arbour, 2003, 2004, 2009; Bader et al., 2005; Rampil, 1998).
EEG-based monitoring using BIS has been studied to evaluate its effectiveness in assessing sedation in critically ill patients. Multiple reports and studies have been published illustrating the potential and limitations of BIS technology in providing feedback on brain function related to therapeutic decisions or clinical changes. A study by Riker, Fraser, and Wilkins (2003) evaluated degree of correlation between parameters monitored with the BIS (SR and BIS value) with level of burst suppression/burst count on EEG. Sixty-two patient-days of data were obtained from 12 patients. The investigators found strong correlation between BIS value, SR, and number of bursts per minute (Riker et al., 2003).
Deogaonkar et al. (2004) studied correlation between BIS values and clinical assessment of sedation in brain-injured patients finding statistically significant correlation between BIS values, clinical sedation assessment, and level of consciousness. The investigators found stronger correlation between BIS and clinical assessment with the newer version of BIS technology, the BIS-XP, and concluded that BIS has the potential as an adjunctive monitoring tool in patients after brain injury (Deogaonkar et al., 2004).
Mondello et al. (2002) also studied correlation between BIS and clinical assessment of sedation. The investigators obtained 980 paired observations and found statistically significant correlations between the BIS and Ramsay Sedation Scale (Mondello et al., 2002).
Courtman, Wardburg, and Petros (2003) found moderate correlation between BIS values and comfort scores in critically ill children. This result is generally consistent with available literature illustrating varying degrees of conelation between BIS values and clinical assessment of sedation (Courtman et al., 2003).
Three case reports illustrate the potential sensitivity of BIS to instability of cerebral physiology and metabolic state. In the first case, BIS was utilized for anesthesia management during a cardiac surgical procedure. Profound hypotension occurred intraoperatively with concurrent dramatic decline in BIS value. BIS values returned to baseline upon restoration of cerebral perfusion. Limitations of BIS for assessing periods of cerebral ischemia include the fact that BIS monitoring reflects EEG activity in a limited area of cerebral cortex. As such, it is nonspecific to location of cerebral instability and reflects global changes in cerebral physiology. It did show sensitivity to compromised cerebral perfusion with resulting cerebral hypometabolic state in this case (Mourisse & Booij, 2003). In the second case, BIS monitoring was utilized after off-pump coronary bypass surgery. In the immediate postoperative period, the patient experienced significant hemorrhage followed by cardiac arrest. Progressive hypotension and pulse-less electrical activity were accompanied by concurrent significant decline in BIS values. BIS values increased with effective cardiac compressions and successful resuscitation (Azim & Wang, 2004). In the third case, BIS was utilized to monitor sedation in an elderly patient receiving sufentanil and midazolam. During one period, BIS values ranged between 40 and 52. Serum glucose level was obtained by point-of-care testing, and the patient was found to be profoundly hypoglycemic. Dextrose 30% was administered by infusion, and rapid elevation of BIS values occurred, ranging between 59 and 82. Occurrence of EEG changes, including slow wave activity and varying degrees of EEG suppression, is well documented consequent to severe hypoglycemia. With BIS values being sensitive to changes in cerebral physiology, it has the potential to facilitate earlier recognition and intervention for cerebral instability (Vivien, Langeron, & Riou, 2002).
BIS was evaluated in monitoring propofol sedation after multisystem trauma during spontaneous awakening trials by Ogilive et al. (2011). The investigators found that the BIS trend waveform/BIS value increased predictably after propofol interruption. Modest correlation between BIS value and Richmond Agitation-Sedation Scale was also documented. The investigators concluded that, during spontaneous awakening trial from propofol sedation, BIS monitoring is a reliable, effective, and continuous adjunct for sedation monitoring in the trauma population (Ogilive et al., 2011).
BIS was studied to determine its potential value in facilitating extubation after cardiothoracic surgery. The investigators compared variables including BIS values over time, medication utilization, and time to extubation between one group (n-25) utilizing BIS monitoring and one group utilizing standard clinical practice (n-25). The investigators found no difference in time to extubation between the two groups. The study highlighted the clinical nurse specialist role as integral in protocol development, developing and reinforcing staff comfort level, and rationale for BIS monitoring (Anderson, Henry, & Hunt, 2010).
BIS monitoring was investigated in monitoring degree of EEG suppression during barbiturate-induced coma (n-8) in a pediatric study (Prins, de Hoog, Blok, Tibboel, & Visser, 2007). BIS monitoring potentially aided monitoring of EEG suppression when used continuously.
EEG and EEG-derived monitoring are well supported by the available evidence for monitoring effectiveness of therapies in managing status epilepticus to determine clinical end points (Claassen, Silbergleit, Weingart, & Smith, 2012). Continuous EEG monitoring, as available, is routinely utilized to guide barbiturate titration to EEG burst suppression in real time for managing intracranial hypertension (Sutter, Stevens, & Kaplan, 2013). EEG-based monitoring can be utilized to detect subclinical seizure activity and guide CNS depressant therapeutics during therapeutic hypothermia protocols (Legriel et al., 2009). EEG-based monitoring may also be integral in determining neurologic prognosis after cardiopulmonary arrest, particularly when interpreted within the context of comprehensive clinical evaluation including brainstem reflexes, body temperature at baseline, and EEG responsiveness to stimulation (Roest et al., 2009).
Literature review reveals several things. First, limited evidence of correlation between BIS and clinical assessment of sedation and drug-induced coma is documented with an older version of BIS technology (BIS A-1050 monitoring system). Second, BIS is sensitive to and reflects changes in brain physiology whether drug induced or consequent to metabolic and/or structural injury. Third, clinical assessment of sedation and neurological stability is and should remain the "gold standard."
Clinical experience and published work utilizing BIS technology clearly support further study and, in limited situations, application during neuromuscular blockade therapy. Further study and potential clinical application related to BIS utilization during drug-induced coma is justified for multiple reasons. One reason is the highly labor-intensive nature of bedside diagnostic EEG monitoring and the potential challenges experienced by clinicians in interpreting the large volume of data obtained. A second reason relates to outcomes related to drug-induced coma including ventilator and intensive care unit (ICU) days. Additional ventilator and ICU days add significantly to financial costs of healthcare in acute care settings (Dasta, McLaughlin, Mody, & Piech, 2005). Closer electrophysiologic monitoring during drug-induced coma facilitates real-time drug titration by bedside clinicians and minimizes increased ICU and ventilator days by avoiding excessive depth of dmg-induced coma, prolonged recovery, and complications such as hospital-acquired infections.
In the literature review and reported cases, one common thread is a degree of correlation between clinical evidence of sedation/CNS depression and data obtained using the BIS. A second common thread is that information obtained from BIS monitoring is actionable in real time providing data not immediately available on clinical assessment. This ease of access to real-time data on brain state may enable the nurse caring for the critically ill neuroscience patient to make both faster assessments and clinical decisions regarding titration of CNS depressants such as propofol or barbiturates in managing drug-induced coma. A third common thread, as described in the cases reported, is that neurophysiologic monitoring using BIS can reveal physiologic instability affecting brain state and enable effective and timely clinical management. This is consistent with the authors' experience in which BIS monitoring during neuromuscular blockade was integral to identifying nonsurvivable brain injury and instrumental in changing direction of care and clinical management (Arbour, 2003). Increased validation and availability of derived EEG parameters such as BIS monitoring can facilitate timely and more streamlined care for the critically ill neuroscience patient.
Patients and Methods
Patients were enrolled consecutively in this study after written, informed consent on their behalf from immediate family member(s). Those requiring drug-induced coma for management of intracranial pressure (ICP) elevation or intractable seizure activity met study inclusion criteria. The investigator then discussed participant recruitment into the study with the patient's neuroscience attending physician. Study goals, drug therapy, clinical end points, and anticipated initiation as well as duration of diagnostic EEG monitoring were discussed. In addition, the point was made that concurrent BIS monitoring would not interfere with diagnostic EEG data collection for monitoring and drug titration for burst suppression as clinically appropriate. When a patient was considered appropriate for study inclusion, the nurse researcher would speak directly with patient family, specifically first-degree relatives giving consent. The informed consent process was reviewed with patient family members in private. All aspects of study participation for the patient were reviewed including study purpose, participant rights/safeguards, right to refuse, and privacy. Time for the informed consent discussion was open-ended, giving ample time for family members to ask questions and provide true informed consent. Upon obtaining informed consent from patient family members, placement of BIS monitoring electrodes was coordinated with EEG technologists. This ensured optimal data collection during current diagnostic EEG and BIS monitoring. Study participants were obtained from medical and surgical critical care areas of a large, tertiary-care referral center. Participant demographics are listed in Table 1.
This was a prospective, observational cohort study conducted between July 11,2007, and March 31,2010. A series of four consecutive patients who received barbiturate- or propofol-induced coma for management of intracranial hypertension or seizure activity constituted the study sample. Bedside diagnostic EEG recording and frontal montage BIS/EEG recording were initiated as quickly as possible after determination that the patient would be undergoing drug-induced coma. Metabolic suppression therapy was titrated to produce a burst-suppression pattern on the diagnostic EEG tracing (goal: 3-6 bursts/minute). Drug dosage (vasopressor and CNS depressant), ICP, cerebral perfusion pressure, and vital signs were recorded daily at the time of EEG recording. Concurrent monitoring with BIS and diagnostic EEG proceeded, and the following variables were recorded for each parameter: EEG, presence or absence of burst suppression; number of bursts per minute; BIS, BIS value/trend over time; burst count; and SR. Variables for each monitoring parameter were compared for corresponding time indices.
Because the EEG and BIS may change in response to stimulation as well as altered drug dosing and physiologic stability, any stimulation of the patient, and change in drug dose, physiologic stability during BIS/EEG monitoring was noted using an event marker, and effects were evaluated on both monitoring parameters. BIS value and SR were correlated with number of bursts per minute and level of EEG suppression during drag-induced coma.
Barbiturate and/or propofol dosing was compared with EEG state and BIS parameters including BIS value, SR, and burst count to determine predictive value of drag dosing for EEG state and level of burst suppression. Statistical analysis included Pearson product-moment correlation and Spearman rank coefficient for BIS value and number of bursts per minute as well as SR and number of bursts per minute. In addition, regression analysis was utilized to plot BIS values against the number of bursts per minute of the diagnostic EEG as well as SR values against the number of bursts per minute on diagnostic EEG.
Data Collection Methods/Procedures Study Location/Administrative Support/Approval and Participant Recruitment
This study took place within the medical and surgical ICUs of a tertiary care referral center (Einstein Medical Center, Philadelphia, PA). After internal approval by nursing department leadership, institutional review board approval was obtained. Support was obtained from all potential stakeholders including neurodiagnostic laboratory personnel, neurologists, and neurosurgeons as well as clinical nurse specialists and nursing management. Education specific to the proposed study was provided by the investigator including research purpose, overview of EEG and BIS technology, and possible clinical applications. Potential research participants were identified, and the investigator was notified by communication with departments of neurology and neurosurgery, neurodiagnostic laboratory, and nursing department colleagues.
Criteria for study participation included determination that the patient in question is a candidate for drag-induced coma. Additional inclusion criteria were intact skin around/near areas of BIS and diagnostic EEG electrode placement as well as obtaining baseline neurological evaluation (if able). Exclusion criteria for the study included significant skin/soft tissue injury/inflammation, which may interfere with BIS and EEG electrode placement, and recent craniotomy/facial trauma disrupting anatomical landmarks and interfering with BIS/EEG electrode placement.
Data Collection Methods
Demographics were collected on each study participant (see Table 1). After application of electrodes for both diagnostic EEG and BIS/EEG data acquisition, data were collected and stored electronically within each monitoring device. In the BIS monitoring system, BIS value/trend over time, SR, burst count, EMG, and signal quality index may be retrieved and reviewed. Diagnostic EEG tracings were stored electronically and downloaded onto a compact disc fonnat for review, manual assessment of burst count/EEG suppression, and comparison with BIS/EEG variables for specific corresponding time indices. The study data for each patient were transcribed onto individualized data collection sheets containing all measured EEG and EEG-derived parameters. All sets of observations, once transcribed, were entered into an Excel spreadsheet. Data sets were de-identified being noted by patient number (1, 2, 3, and 4) only.
Descriptive statistics were performed on demographic variables. Statistical analysis was performed using STATA-10 data analysis and statistical software. Upon completion of data entry into the Excel spreadsheet, all data were imported into STATA-10 for final statistical analysis. Correlation statistics including Pearson's product-moment correlation and Spearman rank coefficient for BIS value and number of bursts per minute as well as SR and number of bursts per minute were performed. Regression analysis was utilized to plot BIS values against the number of bursts per minute on diagnostic EEG and SR values against the number of bursts per minute on diagnostic EEG. BIS, SR, and burst suppression/level of EEG suppression were analyzed specific to identical time indices obtained from each monitoring device.
Four patients yielded 1,972 data sets over 33 hours of continuous and concurrent EEG and BIS monitoring. Regression coefficient of 0.6673 shows robust predictive value between EEG burst count and BIS SR. Spearman rank coefficient of -0.8727 indicates strong inverse correlation between EEG burst count and BIS SR. Pearson's correlation coefficient between burst count obtained from diagnostic EEG versus burst count obtained from BIS monitoring system was .8256 indicating strong positive correlation. Spearman's rank coefficient of 0.8810 showed strong correlation between BIS value versus quantitative burst count obtained from diagnostic EEG recording. Pearson's correlation coefficient of .6819 also indicates positive correlation between BIS value and quantitative burst count on diagnostic EEG. Number of patients (4) limits available statistics and ability to generalize results. Graphic illustration of relationships among parameters and statistical testing show strong correlation and predictive value between BIS parameters (SR, burst count, and BIS value) to quantitative measurement of diagnostic EEG suppression/burst count. Figure 2 (available as Supplemental Digital Content 2 at http://links.lww.com/JNN/A24) illustrates graphic relationship between BIS SR and quantitative burst count obtained from diagnostic EEG. Figure 3 (available as Supplemental Digital Content 3 at http://links.lww.com/JNN/A25) illustrates graphic relationship between burst count obtained from BIS monitoring system and quantitative burst count from diagnostic EEG fracing. Figure 4 (available as Supplemental Digital Content 4 at http://links.lww.com/JNN/A26) illustrates graphic relationship between BIS value and quantitative burst count from diagnostic EEG tracing.
Results related to BIS/EEG monitoring may also be illustrated by review of individual patient trajectory and changes in drug therapy as well as activity/interventions during concurrent monitoring/study participation. Patient 1 was a 64-year-old man with a past medical history of severe alcohol abuse who was found unresponsive at home by a family member. During transport to the emergency department, he was noted to have tonic-clonic movements of the right upper extremity. He received intravenous (IV) lorazepam 4-mg bolus. He was admitted to the medical ICU for management of seizure activity and prophylaxis for severe alcohol withdrawal syndrome. On ICU day 2, he developed status epilepticus. Endotracheal intubation was performed, controlled ventilation was initiated, and IV propofol was administered by continuous infusion for cerebral metabolic suppression and aggressive seizure management. Goal-directed therapy for propofol was to induce burst-suppression EEG pattern (3-6 bursts/ minute), controlling seizure activity pending additional therapeutics. After initiation of bedside continuous EEG monitoring for propofol dose titration, the family was contacted and gave written informed consent for study participation. Figure 5 (available as Supplemental Digital Content 5 at http://links.lww.com/JNN/A27) illustrates BIS/EEG parameters during concurrent monitoring across his trajectory of care.
Patient number 2 was a 57-year-old man who resided at a boarding home and presented originally to the emergency department (ED) with seizures and psychotic symptoms. He was initially managed on a medical floor with neurology consultation after resolution of seizure activity and control of psychotic symptoms in the ED. He experienced status epilepticus after admission to a general medical floor and was emergently intubated, placed on controlled ventilation, and transferred to the ICU for further management. Continuous EEG monitoring was initiated, and pentobarbital sodium by continuous infusion was titrated to control cerebral hypermetabolic state. Goal-directed therapy for pentobarbital sodium titration was a burst-suppression EEG pattern (3-6 bursts/minute) indicating control of cerebral hypermetabolic state and suppression of seizure activity. Upon review of clinical state, medical record, and collaboration with other team members, it was detennined that he met study inclusion criteria. Family meeting was held, and family members gave written informed consent for study participation. Figure 6 (available as Supplemental Digital Content 6 at http://links.lww.com/JNN/A28) illustrates BIS/EEG parameters during concurrent monitoring across his trajectory of care.
Patient number 3 was a 40-year-old woman with a large left middle cerebral artery ischemic stroke who underwent left-sided decompressive craniectomy for declining neurological examination and refractory intracranial hypertension. Postoperatively, intracranial pressures remained elevated, she remained intubated with controlled ventilation, and mechanism-based interventions (osmotic therapy, normocarbia) for ICP control were continued. For refractory ICP elevations, she received propofol by continuous infusion to control cerebral hypermetabolic state and concurrent diagnostic EEG monitoring. Goal-directed therapy for propofol infusion was to attain burst suppression at 3-6 bursts/ minute on EEG. Upon review of clinical state, medical record, and collaboration with other team members, it was determined that she met study inclusion criteria. Family meeting was held, and family members gave written infonned consent for study participation. Figure 7 (available as Supplemental Digital Content 7 at http://links.lww.com/JNN/A29) illustrates BIS/EEG parameters during concurrent monitoring across his trajectory of care.
Patient number 4 was a 54-year-old man with no significant past medical history before admission. He was found unresponsive by family members, and the emergency medical system was activated for transport to ED. Upon ED arrival, the patient was hypothermic and hypotensive and had multiple episodes of tonicclonic seizure activity. He was intubated for airway protection, and controlled ventilation was initiated. Seizure activity was initially halted with IV lorazepam. After stabilization, he was transferred to the medical ICU. Clinical evaluation and computerized tomography scan of the brain revealed that his mental status change was secondary to ischemic stroke affecting the right middle cerebral artery vessel distribution.
The patient had additional refractory seizure activity, progressing to status epilepticus. Sodium pentobarbital infusion was initiated for metabolic suppression/seizure elimination pending treatment of cause and maximizing other therapies. Upon review of clinical state, medical record, and collaboration with other team members, it was determined that he met study inclusion criteria. Family meeting was held, and family members gave written informed consent for study participation. Figure 8 (available as Supplemental Digital Content 8 at http://links.lww.com/JNN/A30) illustrates BIS/EEG parameters during concurrent monitoring across his trajectory of care.
Close association among BIS SR, BIS value, EEG burst count, and BIS burst count indicates predictive value of data obtained from BIS technology for state of cerebral metabolic suppression on diagnostic EEG. Results support use of BIS technology for monitoring of drug-induced coma.
In this study, yielding 1,932 data sets comparing BIS value, SR, burst count (EEG), and burst count (BIS), strong correlation and predictive value were found. This reflects earlier findings by Riker et al. (2003) utilizing an earlier version of BIS technology (A-1050) and analyzed the EEG-derived parameters (burst count, SR, and BIS value) against a single EEG channel obtained from the BIS monitoring device itself. Strong correlation was found between BIS value, SR, and number of bursts per minute as monitored by a single EEG channel displayed on the BIS monitoring device during barbiturate-induced coma (Riker et al., 2003), limiting applicability of study conclusions. This study utilizes the BIS A-2000 XP platform and compares all monitored parameters with the gold standard of diagnostic EEG. This is among the first such studies comparing the newer version of BIS technology with the "gold standard" of diagnostic EEG.
The regression coefficient obtained in this study (0.6673) is comparable with the correlation coefficient obtained by Prins et al. (2007). Those investigators monitored degree of EEG suppression during barbiturate-induced coma in a pediatric study (n-8). The average correlation between BIS SR and quantitative measurement of EEG suppression (on diagnostic EEG tracing) was .68 (moderate correlation). The investigators found, as did the authors of this study, that BIS monitoring potentially aids monitoring of EEG suppression when used continuously and can identify states of overdosing versus underdosing of barbiturate therapy and barbiturate dosing beyond level required for burst suppression (Prins et al., 2007). Graphic illustrations of BIS-derived parameters showed similar correlation as did graphed patient data points in this study. This was among only a very few prior published studies comparing BIS-acquired EEG data with that obtained from diagnostic EEG.
Correlation between BIS-derived parameters was reflected in a recent case report illustrating care and monitoring of a patient with refractory status epilepticus in barbiturate-induced coma (Jaggi, Schwabe, Gill, & Horowitz, 2003). Correlations between BIS, SR, and bursts per minute were strong and reflected results obtained with this study.
Study limitations are multiple. One limitation is sample size. The number of patients (4) limits available statistics and ability to generalize results. A second limitation is that of study subject demographics. Demographics were limited to three male and one female subjects. A third limitation is that of range of brain injuries. Three of the four study subjects underwent metabolic suppression therapy for managing intractable seizure activity. One of the four subjects underwent metabolic suppression therapy for severe ICP elevation consequent to ischemic stroke/decompressive hemicraniectomy. A wider range of brain injuries in study subjects including hemorrhagic stroke, brain trauma, hepatic encephalopathy/brain edema, and ischemic stroke would enhance generalizability of study results and provide evidence for wider practice implications. A fourth limitation is age of the study subjects. Age of study subjects ranged from 40 to 64 years. A subject age range with multiple participants stratified between 21 and 75 years with larger sample size would significantly enhance ability to generalize results and support farther-reaching practice implications.
The "gold standard" for electrophysiologic monitoring of drug-induced coma to objectively determine level of metabolic suppression utilizes bedside diagnostic EEG monitoring. This practice ties up a diagnostic EEG machine, multiple technologists and neurologists for patient setup, data entry, electrode application, and EEG interpretation. In the absence of 24/7 availability of neurology/EEG personnel for continuous interpretation, metabolic suppression therapy and attempting drug titration in real time risk extremes of excessive versus inadequate depth of drug-induced coma. Inadequate dosing of agents utilized for drug-induced coma risks inadequate metabolic suppression therapy and inadequate control of ICR Excessive dosing of agents utilized for drug-induced coma risks excessively deep drug-induced coma causing hemodynamic compromise, decreased brainstem function, prolonged recovery, and larger number of ICU and ventilator days. This circumstance increases human and financial costs of therapy. Even with limited research and evidence validating effectiveness of BIS technology in monitoring drug-induced coma, it is utilized in many clinical settings and incorporated into monitoring protocols. Although BIS technology is utilized in monitoring metabolic suppression, it has limited supporting evidence validating it against the gold standard of diagnostic EEG monitoring. This study is underpowered (n = 4) to support widespread practice changes. The study results do support future research in this area including study duplication utilizing larger sample sizes and more diverse demographics (age, gender, type of injury, and CNS depressant used).
Establishing predictive value with additional research between BIS technology and diagnostic EEG for level of burst suppression in drug-induced coma would objectively test reliability and potentially generate additional evidence supporting BIS for monitoring level of burst suppression. Increased validation and availability of EEG-derived parameters such as BIS adds an additional tool for the nurse caring for the critically ill patient receiving drug-induced coma or after metabolic or structural brain injury.
There was minimal risk associated with study participation. Risk may be associated with skin irritation at the site(s) of BIS electrode placement. This irritation, in the event it occurs, is typically minor and resolves within 1-2 hours after electrode removal. During this study, no instances of skin irritation consequent to electrode placement occurred.
Procedures to Minimize Risks
Procedures to minimize risks included thorough skin assessment and preparation before electrode placement. After electrode placement, frequent skin assessment was done to determine possible onset of tissue irritation. If early indications of tissue irritation were present, removal and reapplication of BIS electrodes in a different location would be done.
Potential study benefits include more streamlined and efficient drug titration to maintain a consistent level of burst suppression during drug-induced coma. This may facilitate more consistent therapy, lower costs, and improved outcomes as measured by fewer ventilator days and shorter ICU length of stay.
Anderson, J., Henry, L., & Hunt, S. (2010). Bispectral index monitoring to facilitate early extubation following cardiovascular surgery. Clinical Nurse Specialist, 24(3), 140-148.
Arbour, R. (2003). Continuous nervous system monitoring: EEG, the bispectral index and neuromuscular transmission. AACN Clinical Issues, 14(2), 185-207.
Arbour, R. (2004). Using bispectral index monitoring to detect potential breakthrough awareness and limit duration of neuromuscular blockade. American Journal of Critical Care, 73(1), 66-73.
Arbour, R. (2006). Impact of bispectral index monitoring on sedation and outcomes in critically ill adults: A case series. Critical Care Nursing Clinics of North America, 18(2), 227-241.
Arbour, R. (2009). Electroencephalographic-derived monitoring. In L. R. Littlejohns & M. K. Bader (Eds.), Monitoring technologies in critically ill neuroscience patients (pp. 175-197). Sudbury, MA: Jones and Bartlett, LLC.
Arbour, R. (2013). Traumatic brain injury: Pathophysiology, monitoring and mechanism-based care. Critical Care Nursing Clinics of North America, 25(2), 297-320.
Azim, N., & Wang, C. Y. (2004). Case report: The use of bispectral index during a cardiopulmonary arrest: A potential indicator of cerebral perfusion. Anesthesia, 59, 610-612.
Bader, M. K., Arbour, R., & Sylvan, P. (2005). Refractory increased intracranial pressure in severe traumatic brain injury: Barbiturate coma and bispectral index monitoring. AACN Clinical Issues, 16(4), 526-541.
Claassen, J., Silbergleit, R., Weingart, S. D., & Smith, W. S. (2012). Emergency neurological life support: Status epilepticus. Neurocritical Care, 17, S73-S78.
Courtman, S. P., Wardburg, A., & Petros, A. J. (2003). Comparison of the bispectral index monitor with the Comfort score in assessing level of sedation of critically ill children. Intensive Care Medicine, 29, 2239-2246.
Dasta, J. F., McLaughlin, T. R, Mody, S. H., & Piech, C. T. (2005). Daily cost of an intensive care unit day: The contribution of mechanical ventilation. Critical Care Medicine, 33(6), 1266-1271.
Deogaonkar, A., Gupta, R., DeGeorgia, M., Sabharwal, V., Gopakumaran, B., Schubert, A., & Provencio, J. J. (2004). Bispectral index monitoring correlates with sedation scales in brain-injured patients. Critical Care Medicine, 32(12), 2403-2406.
Friedman, D., Claassen, J., & Hirsch, L. J. (2009). Continuous electroencephalogram monitoring in the intensive care unit. Anesthesia and Analgesia, 109, 506-523.
Honeybul, S. (2011). An update on the management of traumatic brain injury. Journal of Neurosurgical Sciences, 35(4), 343-355.
Jaggi, P., Schwabe, M. J., Gill, K., & Horowitz, I. N. (2003). Use of an anesthesia cerebral monitor to assess burst-suppression in pentobarbital coma. Pediatric Neurology, 23(4), 219-222.
Kennedy, J. D., & Gerard, E. F. (2012). Continuous EEG monitoring in the intensive care unit. Current Neurology, Neuroscience Reports, 12, 419-428.
Legriel, S., Bruneel, F., Sediri, H., Hilly, J., Abbosh, N., Lagarrigue, M. H., ... Bedos, J. P. (2009). Early EEG monitoring for detecting postanoxic status epilepticus during therapeutic hypothermia: A pilot study. Neurocritical care, 11, 338-344.
Mondello, E., Siliotti, R., Noto, G., Cuzzocrea, E., Scollo, G., Trimarchi, G., & Venuti, F. S. (2002). Bispectral index in ICU: Correlation with Ramsay Score on assessment of sedation level. Journal of Clinical Monitoring and Computing, 17(5), 271-277.
Mourisse, J., & Booij, L. (2003). Bispectral index detects period of cerebral hypoperfusion during cardiopulmonary bypass. Journal of Cardiothoracic and Vascular Anesthesia, 77(1), 76-78.
Ogilive, M. P., Pereira, B. M. T., Ryan, M. L., Gomez-Rodriguez, J. C., Pierre, E. J., Livingstone, A. S., & Proctor, K. G. (2011). Bispectral index to monitor propofol sedation in trauma patients. Journal of Trauma, Injury, Infection and Critical Care, 7/(5), 1415-1521.
Prins, S. A., de Hoog, M., Blok, J. H., Tibboel, D., & Visser, G. H. (2007). Continuous noninvasive monitoring of barbiturate coma in critically ill children using the bispectral index monitor. Critical Care, 77(5), R108. doi:10.1186/cc6138
Purdon, P. L., Pierce, E. T., Mukamel, E. A., Prerau, M. J., Walsh, J. L., Wong, K. F. K., ... Brown, E. N. (2013). Electroencephalogram signatures of loss and recovery of consciousness from propofol. Proceedings of the National Academy of Sciences USA, 7/0(12), El 142-El 151.
Rampil, I. J. (1998). A primer for EEG signal processing in anesthesia. Anesthesiology, 89, 980-1002.
Rhoney, D. H., & Parker, D. (2001). Use of sedative and analgesic agents in neurotrauma patients: Effects on cerebral physiology. Neurological Research, 23, 237-259.
Riker, R. R., Fraser, G. L., & Wilkins, M. L. (2003). Comparing the bispectral index and suppression ratio with burst suppression of the electroencephalogram during pentobarbital infusions in adult intensive care patients. Pharmacotherapy, 23(9), 1087-1093.
Roest, A., van Bets, B., Jorens, P. G., Baar, I., Weyler, J., & Mercelis, R. (2009). The prognostic value of the EEG in postanoxic coma. Neurocritical Care, 10, 318-325.
Rosow, C., & Manberg, P. J. (1998). Bispectral index monitoring. Anesthesiology Clinics of North America, 2, 89-107.
Sutter, R., Stevens, R. D., & Kaplan, P. W. (2013). Continuous electroencephalographic monitoring in critically ill patients: Indications, limitations and strategies. Critical Care Medicine, 41(A), 1124-1132.
Vivien, B., Langeron, O., & Riou, B. (2002). Increase in bispectral index (BIS) while correcting a severe hypoglycemia. Anesthesia and Analgesia, 95, 824-825.
Wallace, B. E., Wagner, A. K., Wagner, E. P., & McDeavitt, J. T. (2001). A history and review of quantitative electroencephalography in traumatic brain injury. Head Trauma Rehabilitation, 16(2), 165-190.
Questions or comments about this article may be directed to Richard B. Arbour, MSN RN CORN CNRN CCNS FAAN, at RichNrs@ aol.com. He is a Neuroscience Clinical Nurse Specialist at Lancaster General Hospital, Lancaster, PA.
Jonathan Dissin, MD, is Medical Director, Neuroscience Unit/Director, Clinical Stroke Service, Einstein Healthcare Network, Philadelphia, PA.
The authors declare no conflicts of interest.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.jnnonline.com).
TABLE 1 Participant Demographics: Age, Ethnicity, Gender, Metabolic Suppression Therapy, and Reason for Therapy Patient Patient Metabolic Number Age, Years Ethnicity Gender Suppression Therapy 1 64 African American Male Propofol 2 57 African American Male Sodium pentobarbital 3 40 Anglo-American Female Propofol 4 54 Middle-eastern Male Sodium pentobarbital descent Patient Reason for Metabolic Number Suppression Therapy 1 Refractory seizure activity 2 Refractory seizure activity 3 Intracranial hypertension secondary to ischemic stroke 4 Refractory seizure activity Note. Descriptive statistics: participants (n) = 4; age (years) = 53.75 (mean), 10.07 (standard deviation); range = 40-64; gender: female = 25%, male = 75%; racial/ethnic background: African American = 2 (50%), Anglo-American = 1 (25%), Middle-eastern = 1 (25%); marital status: married/living with partner = 4 (100%).
|Printer friendly Cite/link Email Feedback|
|Author:||Arbour, Richard B.; Dissin, Jonathan|
|Publication:||Journal of Neuroscience Nursing|
|Date:||Mar 25, 2015|
|Previous Article:||The effect of nocturnal patient care interventions on patient sleep and satisfaction with nursing care in neurosurgery intensive care unit.|
|Next Article:||Certification: value-added care.|