Anesthesiology, Article

An assessment of the association of bispectral index with 2 clinical sedation scales for monitoring depth of procedural sedation

Original Contribution

An assessment of the association of bispectral index with 2 clinical sedation scales for monitoring depth of procedural sedationB

Christopher S. Weaver MDb,*, William H. Hauter MDa, Cory E. Duncan MDb,

Edward J. Brizendine MSc, William H. Cordell MDb

aDepartment of Anesthesiology, Indiana University School of Medicine, Indianapolis, IN, USA

bDepartment of Emergency Medicine, Indiana University School of Medicine, Indianapolis, IN, USA

cDepartment of Medicine, Division of Biostatistics, Indiana University School of Medicine, Indianapolis, IN, USA

Received 13 October 2006; revised 27 February 2007; accepted 27 February 2007

Abstract

Objective: We conducted a study to assess the correlation of bispectral index (BIS) to 2 clinical sedation scales.

Methods: This was a prospective, observational study. The BIS number was recorded at baseline and every 30 seconds. One investigator separately monitored the patients for depth of sedation using the Observer’s Assessment of Alertness/Sedation and the Continuum of Depth of Sedation scales.

Results: During the 6-month period, 75 patients were enrolled. The spearman correlation between the BIS and the Observer’s Assessment of Alertness/Sedation was 0.59 (95% confidence interval [CI], 0.44-0.74). The Spearman correlation between the BIS and the Continuum of Depth of Sedation was

0.53 (95% CI, 0.36-0.70). The mean minimum BIS for patients without a complication was 70 (SD, 15.9) compared with 68 (SD, 12.9) for patients with a complication (difference = 2; 95% CI, –7-11). Conclusions: Our study demonstrated moderate correlation between BIS and the 2 clinical sedation scales. The correlation is not strong enough to be used reliably in a clinical setting. The mean minimum BIS scores were not significantly different for those with sedation complications vs those without complications.

D 2007

Introduction

The goal of emergency department (ED) procedural sedation is to allow successful completion of common

B The poster presentation for this was made in May 2004 at the Society for Academic Emergency Medicine Annual Meeting in Orlando, Fla.

* Corresponding author. Wishard Memorial Hospital, Indianapolis, IN 46202, USA. Tel.: +1 317 630 7276; fax: +1 317 6564216.

E-mail address: [email protected] (C.S. Weaver).

painful ED procedures while minimizing patient suffering and complications. Accurate assessment of sedation depth is believed to be important in minimizing complications associated with sedation in the ED.

The Observer’s Assessment of Alertness/Sedation (OAA/S) and the Continuum of Depth of Sedation (CDS) are subjective scales used by anesthesiologists to determine the depth of sedation. The OAA/S and CDS require the physician to assess the sedation depth by verbal and tactile

0735-6757/$ – see front matter D 2007 doi:10.1016/j.ajem.2007.02.042

stimuli and monitoring of the airway and cardiovascular and Respiratory functions. Thus, these clinical scales of sedation depth can only intermittently be assessed and require a brief examination that would be particularly difficult for an emergency physician who is also performing a procedure.

Anesthesiologists frequently use bispectral index (BIS) monitoring to numerically assess anesthetic depth. The BIS monitor provides a digital display of the depth of sedation. Bispectral index monitoring has recently been studied during ED procedural sedation [1-9]. The BIS monitor and sensors are Food and Drug Administration-approved devices and have been used in anesthesia and critical care practice for procedural sedation for years [10]. It is a processed electroencephalographic parameter that is used to monitor a patient’s level of sedation [11-19]. The scale is a 0 to 100 numeric readout, with 0 correlating with electroen- cephalographic burst suppression, whereas a score of 95 to 100 is a fully awake level of consciousness.

Although several studies have been conducted in attempts to evaluate the effectiveness of using the BIS monitor during ED procedural sedation [1-9], we are not aware of any studies that have investigated the association between the BIS number and the OAA/S and CDS scales of sedation in an adult population. We therefore conducted a study to assess the correlation of the BIS with clinical scales used to assess sedation depth–the OAA/S and the CDS scales. We also assessed the number of sedation events and complications occurring with different levels of sedation depth as measured by the mean minimum BIS scores.

Methods

Study design

This was a prospective, observational study of the BIS during ED Procedural sedation with propofol. It was a companion study completed in conjunction with a study titled bEmergency department procedural sedation with propofol: is it safe?Q [20] A convenience sample of eligible, consenting patients was enrolled by 1 investigator between August 1, 2002, and January 31, 2003. The convenience sample was based solely on the availability of the study investigator. The investigator, a senior anesthesia resident who is also emergency medicine residency trained and American Board of Emergency Medicine certified, was on a 6-month research sabbatical and available 24 hours a day. The institutional review board governing the research at the 2 study hospitals approved this study.

Study setting and population

This study was conducted in the EDs of 2 urban academic level I trauma centers, with a combined volume of more than 200 000 ED visits annually. All adult (z18 years of age) patients undergoing ED procedural sedation with propofol were eligible for study participation. Patients

were excluded for drug or alcohol intoxication, pregnancy, inability to give consent (Mental status changes, prisoners), significant airway-obstructing disease (tumor, severe Sleep apnea), or obvious cardioVascular abnormalities (recent myocardial infarction or hypotension). The decision to use propofol, and thus include the patient for possible enroll- ment, was exclusively at the discretion of the emergency physician. Informed consent for this study was obtained before and separate from the specific informed consent for all procedural sedation required at both hospitals.

Study protocol and measures

An emergency medicine attending physician or an emergency medicine resident physician under the supervi- sion of the attending physician conducted the procedural sedation with propofol in the usual fashion. Standard monitoring and procedures as required by the procedural sedation protocols at both hospitals were followed. These protocols require an emergency nurse sedation assistant to be in the room to assist the physician, to monitor the vital signs, and to administer the ordered medications. The emergency physician is responsible for the preprocedure assessment, sedation plan, dosing of medication, and standard monitoring of the patient. The frequency and incremental dosing of propofol was at the discretion of the treating physician. By hospital protocol, all patients had an intravenous line inserted, and oxygen was administered 2 L/min via nasal cannula. The heart rate, blood pressure, and oxygen saturation were electronically monitored.

One of the investigators (WHH), who had trained in both emergency medicine and anesthesiology, monitored sepa- rately from the procedure physician the patients in this study. The additional monitoring included the 2 clinical depth of sedation assessments–the OAA/S (Fig. 1) and the CDS (Fig. 2). The American Society of Anesthesiologists has developed the CDS along with standard definitions that has been widely adopted, most notably by the Joint Commission on Accreditation of Healthcare organizations. The investigator also applied a BIS Quatro Sensor (Aspect Medical Systems Inc, Newton, Mass) to the patient’s fore- head. The BIS XP monitor (Aspect Medical Systems Inc) logged the BIS. The BIS XP monitoring system has advanced software that reduces the artifact associated with muscle

Fig. 1 Observer’s Assessment of Alertness/Sedation Scale.

Fig. 2 Continuum of Depth of Sedation Scale.

tension or movement and improves the reliability of the BIS score [21]. The BIS score was recorded automatically with a time stamp on the computer, therefore not requiring direct monitoring of the BIS score by any investigator. An investigator transferred the BIS scores at 30-second intervals to the database the day after each sedation.

The anesthesiologist did not participate in any way with the sedation plan, medication dosing, or standard monitor- ing during the ED procedural sedation. The emergency staff performing the sedation and procedure were blinded to the BIS number and investigator’s clinical assessment data. Blinding was maintained by obscuring the BIS monitor and investigator’s data with drapes.

The blood pressure, heart rate, respiratory rate, oxygen saturation as measured by pulse oximetry (Spo2), and clinical sedation level were recorded at baseline. Blood pressure and heart rate were measured every 5 minutes by the nurse sedation assistant and recorded on the patient’s ED sedation chart. Ventilation, airway reflexes, Spo2, complications, and clinical sedation depth were assessed and recorded every 30 seconds and at nadir by the anesthesiologist.

Sedation events were prospectively defined as snoring/ partial airway obstruction, complete obstruction, clinical hypoventilation, apnea, postprocedure nausea/vomiting, laryngospasm, hypotension, heart rate of less than 50 beats per minute or rhythm disturbance, and rescue maneuver (head repositioning, jaw thrust, use of oral airway, increase in fraction of inspired oxygen, increase in stimulation, unplanned use of reversal agents). Partial obstruction was

defined as snoring or breathing in a way that allowed some limited ventilation. Complete obstruction was defined as respiratory effort without air movement or end-tidal carbon dioxide concentration in the expired air (ETco2) wave- form. Apnea was defined as no ventilation effort, no ETco2 waveform, and no obvious airway obstruction. Hypotension was defined as a systolic blood pressure of less than 90 mm Hg. Bradycardia was defined as a heart rate of less than 50 beats per minute unless the patient had a preexisting bradycardia. Sedation complications were sedation events that were advanced, untreated, or unrecog- nized. These included hypoxemia, aspiration, hypercapnia (ETco2 of N 50 torr on capnography), bag-valve-mask ventilation, intubation, any blood pressure or heart rate interventions, hospital admission, and death. Hypoxemia was defined as an oxygen saturation of less than 90% for longer than 10 seconds.

Data analysis

The study sample size was limited and determined by the total number of patients that the investigator (WHH) could enroll during his 6-month research sabbatical. The strength of the linear association between the BIS values and the level of sedation as measured by the OAA/S and CDS was assessed by estimating the Spearman correlation coefficient. In addition, an analysis of variance model was used to determine if the mean minimum BIS is different among the different levels of sedation within a sedation

Table 1A Mean minimum BIS by OAA/S level of sedation

OAA/S N = 75

(n [%])

1b (painful trapezius squeeze) 31 (41)

2b (mild prodding/shaking) 19 (25)

3b (name called loudly/repeatedly) 25 (33)

Mean (SD)

(95% CI)

60 (14.8) (52-68)

70 (12.4) (59-70)

81 (9.1) (76-83)

Mean difference from

OAA/S level 3 (95% CI)a

21 (13-29)

11 (1-20)

Mean difference from

OAA/S level 2 (95% CI)a

10 (2-19)

a Adjusted using Tukey pairwise comparison procedure.

b See Fig. 1 for complete description.

Fig. 3 A, Distribution of minimum BIS values by OAA/S level of sedation. B, Distribution of minimum BIS values by CDS level of sedation. In the box plots, the line in the box shows the median BIS, whereas the box ends show the 25th (Q1) and 75th (Q3) percentiles. The bwhiskersQ extend to 1.5 times the interquartile range (Q3-Q1). The solid dot (d ) shows the mean minimum BIS.

scale. Tukey procedure was used to adjust pairwise comparisons for multiple comparisons. An appropriate 95% confidence interval (CI) was derived for relevant estimates and differences. If a CI for a difference contains zero, it is an indication that there is not a statistically significant difference in the mean BIS values. Statistical analyses were performed using SAS version 8.2 (SAS Institute, Cary, NC).

Results

During the 6-month study period, 82 patients were enrolled. Bispectral index data were available on 77 patients, with complete BIS data captured on 75. The number of patients with useable BIS data was less than the total number of patients enrolled because of the failure of the BIS sensor to adhere to some of the patients’ foreheads. The investigator was unable to get the sensor to stick to 5 patients. In 2 patients, the sensor was placed on the forehead initially but fell off shortly after beginning the sedation, and adhesion to register reliable BIS data could not be achieved again. The median age was 39 years (range, 18-81 years), and 51% were male. Patients underwent ED procedural sedation most often for incision and drainage (49%), joint reduction (28%), and fracture care (11%).

Table 1A presents the mean minimum BIS by OAA/S level of sedation. Two patients had an OAA/S of 4 (lethargic response to name spoken in normal tone). These 2 subjects were recoded to have an OAA/S of 3 (responds only after name is called loudly and/or repeatedly) for analysis purposes because of the small cell size. The minimum BIS score for these 2 patients fell within the OAA/S 3 patients, and their exclusion would not have changed the results of the analysis. Although there is a significant difference ( P b .001) in the mean minimum BIS among all 3 levels of sedation, there was a large distribution of the BIS values within each sedation level, and the BIS values overlapped each other (Fig. 3A). The Spearman correlation between the BIS score and the OAA/ S was 0.59 (95% CI, 0.44-0.74).

Table 1B presents mean minimum BIS by CDS level of sedation. The mean minimum BIS for those subjects reaching moderate sedation was significantly higher than those subjects who reached deep sedation or who entered general anesthesia. However, there was not a difference in the mean minimum BIS between deep sedation subjects and general anesthesia subjects. Again the distribution of the BIS values within a CDS level of sedation varied greatly (Fig. 3B). The Spearman correlation between the BIS score and the CDS was 0.53 (95% CI, 0.36-0.70).

Six patients experienced a sedation event, and 8 had a sedation complication. Of the 6 patients with a sedation event, 3 experienced hypoventilation, 1 had increase fraction of inspired oxygen, 1 had partial obstruction, and 1 had obstruction. The 8 patients who experienced a sedation complication all had hypoxia. For the purpose of this

Table 1B Mean minimum BIS by CDS level of sedation

CDS

General

Deep Moderate

N = 75

(n [%]) 17 (23%)

28 (37%)

30 (40%)

Mean (SD)

(95% CI)

60 (14.7) (55-65)

65 (15.3) (64-76)

80 (8.9) (77-85)

Mean difference from

moderate sedation (95% CI)a 20 (10-29)

15 (7-23)

Mean difference from deep

sedation (95% CI)a

5 (–5 to 14)

a Adjusted using Tukey pairwise comparison procedure.

Fig. 4 Distribution of minimum BIS values between subjects with and without a sedation event/complication. In the box plots, the line in the box shows the median BIS, whereas the box ends show the 25th (Q1) and 75th (Q3) percentiles. The whiskers extend to 1.5 times the interquartile range (Q3-Q1). The solid dot (d ) shows the mean minimum BIS.

analysis, we combined the sedation events and complications for a total of 14 patients (19%) who experienced a complication. We combined the 2 because of the small numbers, and we wished to evaluate the BIS ability to predict any untoward event regardless of classification. The mean minimum BIS for patients without a complication was 70 (SD, 15.9), whereas the mean minimum BIS for patients with at least 1 complication was 68 (SD, 12.9). The means were not significantly different from each other (difference = 2; 95% CI, –7 to 11). Fig. 4 presents the distribution of the BIS values by event complication occurrence.

Discussion

Our study demonstrated only a moderate correlation between the BIS score and 2 instruments (OAA/S and CDS) used to clinically assess depth of sedation. Other studies have compared the BIS to clinical sedation scores during ED procedural scales [1-9]. Gill et al [6] found a moderate correlation in adults between the BIS and the modified Ramsay Sedation Scale (Spearman = –0.690) that was comparable to our findings. However, they found that BIS monitoring reliably predicted patients who were sedated to the point of general anesthesia from those with a lesser degree of sedation. Bispectral index monitoring did not discriminate as well between patients with mild to moderate sedation and those with moderate to deep sedation in their study. Our results do not demonstrate the same discrimina- tory value of the BIS between deep sedation and the level of general anesthesia. Although it is possible, we did not find a difference because of the small sample size. Fig. 3B displays the considerable overlap of minimum values between various sedation depths.

Agrawal et al [7] found a stronger correlation (Pearson =

–0.78) between the BIS and the modified Ramsay Sedation Scale in 20 children undergoing ED procedural sedation with midazolam/fentanyl and pentobarbital. Similar to the study by Agrawal et al, our study had a single study investigator determine all OAA/S and CDS scores to control for interrater reliability in judging sedation depth. However, our study differed from that of Agrawal et al in several ways. We exclusively enrolled adults, obtained a larger sample size, and recorded sedation depth more frequently (30-second intervals). Three small pilot studies have reported the use of the BIS to monitor the depth of sedation during ED procedural sedation [1,4,5]. Sakles and Keim [1] evaluated the BIS numbers reached in 22 patients sedated with propofol and found that 14 reached BIS scores of less than 60 at some point during the sedation. They did not report a clinical sedation assessment of these patients. Overly et al [4] compared the BIS with the OAA/S in ED sedation of pediatric patients. They enrolled 19 subjects and found that a 25-unit decrease in the BIS predicted a 1-unit decrease in the OAA/S score. Vissers and McHugh [5] enrolled 12 patients undergoing ED procedural sedation with fentanyl and midazolam in a small pilot study. They found a Spearman rank correlation coefficient of 0.66 between the BIS and OAA/S, which was very similar to our result of 0.59.

In 2003, Miner et al [8] investigated the correlation of the BIS with the rate of respiratory depression and patient- derived subjective outcome measures. They concluded that patients with a lowest recorded BIS score between 70 and

85 had the same patient satisfaction scores as did more deeply sedated patients and the same rate of respiratory depression as did less deeply sedated patients. Patients with BIS scores of less than 70 had more respiratory depression than did those with higher scores. There was no significant difference in those with scores of 0 to 85 as compared with those with scores of higher than 85. Miner and his colleagues concluded that BIS scores between 70 and 85 represented the optimal level of sedation. In contrast, we did not categorize BIS scores as described in the article by Miner et al. Instead of trying to determine the optimum level of sedation, our objectives were to assess the relationship of observed BIS scores with established sedation scales and to determine if there is a relationship between a BIS score and a sedation compli- cation or event. Grouping the BIS scores into categories as defined by Miner et al would have decreased our power to detect a significant relationship in both cases. Furthermore, as seen in Figs. 1 and 2, there is a large variation in the BIS scores at each level of each sedation scale, and in Fig. 3, between those patients who experienced a sedation event and those who did not. This large variation in BIS scores would have been clouded had we grouped our data as Miner et al did.

Recently Miner et al [9] published an article to determine whether the knowledge of patients’ BIS scores

would increase the recognition of adequately sedated patients, resulting in less oversedation and respiratory depression. They found a higher rate of respiratory depression in the BIS-blinded group as compared with the group in which the BIS was used. However, they found no difference in the depth of sedation between groups. In contrast, our study was not a blinded comparison of physicians using the BIS for sedation vs those who did not. We evaluated the BIS scores in those patients sustaining complications as compared with those who did not. All physicians performing the sedation were blinded to the BIS score. In our study, the mean minimum BIS scores for those with and without complications are not signifi- cantly different, and the minimum BIS scores for patients in both categories have a great deal of overlap. Our results suggest that the BIS scores themselves do not reliably differentiate which patients have complications.

Limitations and future questions

The results of our study are subject to several limitations. First, we enrolled a relatively small number of patients from a convenience sample. A total of 9 potentially eligible patients sedated with propofol during the study period were not enrolled. Second, because this study was conducted at 2 academic emergency medicine centers, the results may not be generalizable to the community setting. Third, our study was designed to test the correlation of BIS scores to 2 clinical sedation scores. We did not study the predictive value of BIS scores to alert the physician performing a procedure to an impending or overt sedation event or complication (oversedation). Furthermore, we did not study which test was the gold standard in establishing oversedation. Fourth, because propofol was given by Bolus injection in our study, we did not evaluate the value of BIS monitoring in propofol sedation given by other means such as continuous drip, nor did we study other procedural sedation agents such as fentanyl and midazolam. Fifth, we only had 1 investigator performing the clinical sedation scores on all of our patients. Furthermore, this physician was trained in both emergency medicine and anesthesiology. Although the performance of the clinical sedation examinations assured consistency and decreased interrater reliability, it is possible that the poor correlation was secondary to the performance of the sedation scales by this 1 investigator. In addition, this investigator was not blinded to the BIS scores while performing the clinical sedation scores. It is possible that knowledge of BIS values might have subconsciously influenced how the OAA/S and the CDS were scored.

Future studies on this topic should be considered. Studies with larger sample sizes and expanded to multiple sites could add to our knowledge of the use of BIS for ED procedural sedation. Similar studies with other sedatives

would also be of value to evaluate the association of the BIS number with depth of sedation and the ability to predict complications in this population.

Conclusions

In our study, the BIS appears to moderately correlate with the OAA/S and CDS in patients undergoing ED procedural sedation with propofol. The correlation is not strong enough to be used reliably in a clinical setting. In addition, the BIS scores cannot be used to differentiate between the patients who sustained a complication as and those who did not.

References

  1. Sakles J, Keim S. Using the bispectral index (BIS) to monitor the level of sedation during deep procedural sedation in the emergency department with propofol. Acad Emerg Med 2002;9:406.
  2. Miner JR, Heegaard W, Plummer D. End-tidal carbon dioxide monitoring during procedural sedation. Acad Emerg Med 2002;9: 275 – 80.
  3. Miner JR, Kirven K, Biros MH. The utility of the bispectral index in procedural sedation in the emergency department. Acad Emerg Med 2003;10:440.
  4. Overly FL, Fontaine B, Wright RO, et al. Bispectral analysis in the pediatric emergency department during conscious sedation. Acad Emerg Med 2002;9:518.
  5. Vissers RJ, McHugh DF. Bispectral index monitoring as a continuous, non-invasive measure of sedation during procedures in the ED. Acad Emerg Med 2000;7:529.
  6. Gill M, Green SM, Krauss B. A study of the bispectral index monitor during Procedural sedation and analgesia in the emergency depart- ment. Ann Emerg Med 2003;41:234 – 41.
  7. Agrawal D, Feldman HA, Krauss B, et al. Bispectral index monitoring quantifies depth of sedation during emergency depart- ment procedural sedation and analgesia in children. Ann Emerg Med 2004;43:247 – 55.
  8. Miner JR, Biros MH, Heegaard W, et al. Bispectral electroencepha- lographic analysis of patients undergoing procedural sedation in the emergency department. Acad Emerg Med 2003;10:638 – 43.
  9. Miner JR, Biros MH, Seigel T, et al. The utility of the bispectral index in procedural sedation with propofol in the emergency department. Acad Emerg Med 2005;12:190 – 6.
  10. Shah NCS, Chea F, Tayong M, et al. Can bispectral index (BIS) of EEG be useful in assessing sedation in ICU patients? Anesth Analg 1996;82:S400 [abstr].
  11. Avramov MN, White PF. Methods for monitoring the level of sedation. Crit Care Clin 1995;11:803 – 26.
  12. Iselin-Chaves IA, El Moalem HE, Gan TJ, et al. Changes in the auditory evoked potentials and the bispectral index following propofol or propofol and alfentanil. Anesthesiology 2000;92:1300 – 10.
  13. De Deyne C, Struys M, Decruyenaere J, et al. Use of continuous bispectral EEG monitoring to assess depth of sedation in ICU patients. Intensive Care Med 1998;24:1294 – 8.
  14. Kearse Jr LA, Rosow C, Zaslavsky A, et al. Bispectral analysis of the electroencephalogram predicts conscious processing of informa- tion during propofol sedation and hypnosis. Anesthesiology 1998;88: 25 – 34.
  15. Sandler NA, Sparks BS. The use of bispectral analysis in patients undergoing intravenous sedation for third molar extractions. J Oral Maxillofac Surg 2000;58:364 – 8 [discussion 369].
  16. Liu J, Singh H, White PF. Electroencephalographic bispectral index correlates with intraoperative recall and depth of propofol-induced sedation. Anesth Analg 1997;84:185 – 9.
  17. Liu J, Singh H, White PF. Electroencephalogram bispectral analysis predicts the depth of midazolam-induced sedation. Anesthesiology 1996;84:64 – 9.
  18. Glass PS, Bloom M, Kearse L, et al. Bispectral analysis measures sedation and memory effects of propofol, midazolam, isoflurane, and alfentanil in healthy volunteers. Anesthesiology 1997;86:836 – 47.
  19. Leslie K, Sessler DI, Schroeder M, et al. Propofol Blood concentration and the bispectral index predict suppression of learning during propofol/Epidural anesthesia in volunteers. Anesth Analg 1995;81: 1269 – 74.
  20. Weaver CS, Hauter WE, Brizendine EJ. et al. Emergency depart- ment procedural sedation with propofol: is it safe? J Emerg Med (in press).
  21. Bruhn J, Bouillon TW, Shafer SL. Electromyographic activity falsely elevates the bispectral index. Anesthesiology 2000;92:1485 – 7.

Leave a Reply

Your email address will not be published. Required fields are marked *