Resuscitation (2006) 69, 207—212

CLINICAL PAPER

An observational study of bispectral indexmonitoring for out of hospital cardiac arrest�

Daniel M Fatovich ∗, Ian G Jacobs,Antonio Celenza, Michael J Paech

University of Western Australia, Department of Emergency Medicine,Royal Perth Hospital, Box X2213 GPO, Perth, WA 6847, Australia

KEYWORDSAdvanced life support;

Summary Cerebral resuscitation is the most important goal of advanced life sup-port. Currently, there are no objective monitoring methods available to gauge the

I

Copodit

o1

f

(

0d

Anaesthesia;Brain ischaemia;Cardiac arrest;External chestcompression;Resuscitation

effectiveness of advanced life support on cerebral resuscitation. We assessed theutility of bispectral index (BIS) monitoring during cardiopulmonary resuscitation asa marker of cerebral resuscitation. Twenty one patients with out of hospital cardiacarrest had a BIS monitor applied during the resuscitation, in addition to standardadvanced life support. The BIS monitor was also applied to a cadaver to assess therole of artefact. Illustrative data are presented, outlining the process of evaluationundertaken. A major component of the BIS tracing during external chest compres-sions appears to be due to movement artefact. Our pilot data indicate that withcurrent technology, BIS monitoring is not a clinically reliable marker of the efficacyof external chest compressions.© 2005 Elsevier Ireland Ltd. All rights reserved.

ntroduction

erebral resuscitation is the most important goalf advanced cardiac life support (ALS); so that theatient achieves their best possible neurologicalutcome. During ALS, clinical staff monitor car-iorespiratory function using end tidal carbon diox-de (CO2) and electrocardiogram (ECG) monitoringo gauge the effectiveness of resuscitation efforts.

� A Spanish translated version of the summary and keywordsf this article appears as Appendix in the online version at0.1016/j.resuscitation.2005.07.022.∗ Corresponding author. Tel.: +61 8 92242244;

ax: +61 8 92247045.E-mail address: daniel.fatovich@health.wa.gov.au

D.M. Fatovich).

Currently, there are no objective monitoringmethods available to gauge the effectiveness ofALS on cerebral resuscitation. Such a monitor wouldneed to be portable, easy to use and provide realtime data indicative of the adequacy of cerebralperfusion.

Bispectral index (BIS) monitoring is a new tech-nology that provides a bispectral analysis of theelectroencephalogram (EEG). It was developed tomonitor the depth of anaesthesia and may preventawareness in the anaesthetised patient.1 The mon-itor uses an easily applied, four-electrode foreheadsensor that calculates and provides a real-timenumber with graphical presentation of trends.

The BIS index is a numerical processed, clinicallyvalidated EEG parameter. It is derived utilisinga composite of multiple advanced EEG signal

300-9572/$ — see front matter © 2005 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.resuscitation.2005.07.022

208 D.M. Fatovich et al.

processing techniques—–including bispectral anal-ysis, power spectral analysis and time domainanalysis. These components were combined tooptimize the correlation between the EEG and theclinical effects of anaesthesia.2

During development of the BIS Index, EEG fea-tures or descriptors were identified by analysing adatabase of EEGs from over 5000 subjects who hadreceived one or more of the most commonly usedhypnotic agents and who had been evaluated withsimultaneous sedation assessment.2 The BIS valueis integrated from various EEG descriptors into asingle dimensionless variable scaled from 100 to 0,where 100 represents an awake EEG and zero com-plete electrical silence. BIS values of 65—80 havebeen recommended for sedation, whereas valuesof 40—60 are said to reflect an adequate depth ofgeneral anaesthesia.1,3

Case reports suggest that BIS monitoring could behelpful in determining the adequacy of resuscita-tion efforts and could guide prolonged efforts.4 Ananaphylactic arrest on induction of anaesthesia pro-vided some information regarding the maintenanceof BIS values above 40 during cardiopulmonaryresuscitation (CPR).5 In a patient with BIS monitor-

Australia. The combined annual census is some94,000, with an admission rate of 43%. The twohospitals are the largest University hospitals inWestern Australia and serve a population of 800,000residents, with 11.3% aged 65 years or older.11

Patients were brought to the ED by the emer-gency medical service (who do not use vasopres-sors) either in, or following, out of hospital cardiacarrest. They received standard ALS in the ED. Theonly additional intervention was the application ofa BIS monitor (BIS Monitor A-2000XP, Aspect Med-ical Systems) sensor to the patient’s forehead byan investigator. The BIS monitor requires 3 × 2.2 sepochs of pure, artefact-free EEG to calculate avalue, so the fastest interval between EEG record-ing and the appearance of the BIS value is about7 s.

BIS monitor tracings were recorded for the dura-tion of the resuscitation efforts and significantevents such as interruption of ECCs were high-lighted on the tracing. The investigators had noclinical role during the resuscitation. Clinical stafftreating the patient were blinded to the BIS data,by concealment of the monitor screen.

Approval for this study was granted by the hos-puc

R

Tupts

C

Fw

ing during hypovolaemic cardiac arrest, BIS valuescorrelated with haemodynamic deterioration andimprovement.6 Another case described fluctuationsin BIS values during external chest compression(ECC) followed by internal cardiac massage, thatappeared to correlate with arterial pressure.7 Theauthors suggest a BIS value of 30—40 during resusci-tation may indicate adequate cerebral perfusion.7,8

The relationship between BIS, the level of seda-tion and the plasma concentrations of propofol hasbeen established.1,9 There is also evidence thatBIS values correlate with cerebral function andblood flow. The EEG is tightly linked to cerebralmetabolism.10 We hypothesised that the BIS trac-ing and value might reflect the adequacy of cerebralperfusion during CPR, and therefore act as a markerof cerebral resuscitation.

This paper presents data from our observationalpilot work, describing the data obtained from BISmonitoring during cardiac arrest. To our knowledge,these data represent the first such series for out ofhospital cardiac arrest.

Methods

This was a convenience sample of patients with outof hospital cardiac arrest who attended two Emer-gency Departments (ED) of the North MetropolitanArea Health Service (Royal Perth Hospital andSir Charles Gairdner Hospital) in Perth, Western

itals’ Ethics Committees. As these patients werenconscious and critically ill, a waiver of informedonsent was granted.

esults

wenty one patients and one cadaver were eval-ated. These data are reported in Table 1. Allatients, except case 6 and the cadaver, died inhe ED. We present illustrative sample data fromix of these.

ase 1

igure 1 shows the BIS data from a 72 year femaleith a witnessed out of hospital cardiac arrest due

Figure 1

An observational study of bispectral index monitoring for out of hospital cardiac arrest 209

Table 1 Summary data of patients enrolled, in chronological order

Age Sex Duration of EDresuscitation (min)

Rhythm onarrival

Initial BISvalue

Comments (cases 1—6 refer to text description)

91 F 0 SR 97 PEA cardiac arrest, remarkable recovery; case 691 M 6 PEA 54 Lung cancer, BIS value peak 7155 F 8 AS 38 Motor neurone disease, respiratory failure72 F 22 AS 29 Case 159 M 8 AS 23 Case 256 M 48 AS 26 Myeloma with septicaemia, arrested in ED50 M 4 AS 17 MI, movement artefact noted36 F 13 AS 32 Following heroin overdose80 M 16 AS 38 MI, rising tracing noted with deeper ECC50 M 15 AS 42 MI, head movement caused sinusoidal BIS tracing58 M N/A N/A 0 BIS monitor applied to cadaver, case 346 M 10 AS 36 First patient to have electrode strapped onto forehead77 F 8 PEA 35 Rocking the head after death produced a rising BIS40 M 15 AS 47 MI57 M 36 PEA 39 Arrested in ED; massive PE; case 438 M 26 PEA 37 Following blunt trauma; case 544 M 30 PEA 36 Massive PE64 M 20 AS 42 MI68 M 10 AS 38 Artefact visible with body movement62 F 23 PEA 39 PE; BIS maintained >35 for 17 min with CPR63 M 15 AS 32 BIS > 30 for 12 min; see Figure 454 M 10 AS 27 Declining BIS with ECC

AS, asystole; ECC, external chest compression; MI, myocardial infarction; N/A, not applicable; PEA, pulseless electrical activity;PE, pulmonary embolus; SR, sinus rhythm.

to ventricular fibrillation. Despite nine defibrilla-tions, she arrived at the ED in asystole, 48 min afterthe cardiac arrest.

The first arrow (starting from the left, indicatinga fall in BIS value) reflects interruption of CPR tocheck the position of the tracheal tube. Followingthis, ECCs recommence and the BIS rises. The nextthree arrows also indicate falls in BIS values withinterruption of CPR. However, with continued ECCsthe BIS value peaks at 80, coinciding with the clin-ical observation of small pupil size. The final arrowindicates cessation of CPR, after which the BIS trac-ing becomes isoelectric within 2 min. The final smallpeak represents artefact resulting from moving thepatient.

Case 2

In Figure 2, the tracing comes from a 59-year-oldman with asystole following out of hospital cardiacarrest. It demonstrates a declining trace despiteECCs. This was believed to be due to insufficientdepth of compression. The first arrow shows thetime that the operator was advised to performdvw

Case 3

At this time in the study, we investigated the sig-nificance of artefact. After obtaining appropriateapproval, we applied the BIS monitor to a cadaver.Death had occurred 48 h previously. The first declin-ing tracing (Figure 3, arrow A) followed applicationof the monitor. Arrow B resulted from performingECCs, thus representing movement artefact. ArrowC resulted from simply rocking the cadaver’s headfrom side to side, mimicking the head movementsthat occur during CPR. However, by simply ban-daging the electrode to the forehead with a crepebandage to prevent movement artefact, the BIStracing remained isoelectric (arrow D).

eeper ECCs. This results in a rising BIS, but thealue remained low and the second arrow indicateshen CPR efforts were terminated.

Figure 2

210 D.M. Fatovich et al.

Figure 3

As a result of this case, further applicationsof the BIS monitor to patients with out of hospi-tal cardiac arrest were made with the electrodebandaged. BIS tracings were still obtained, how-ever, their interpretation was more difficult, e.g.changes with the interruption of CPR were less obvi-ous. Also, we found it remained possible to inducemovement artefact, simply by rocking the headfrom side to side, mimicking the head movementsthat occur during resuscitation. ECCs were associ-ated with a ‘sine wave’ pattern on the BIS monitor(Figure 4), which appeared to result in a spuriousBIS tracing and score.

Case 4

Figure 5 demonstrates the recordings obtainedfrom a 58-year-old man with a massive pulmonaryembolus (confirmed at post-mortem). This is thefirst tracing we obtained that recorded an isoelec-tric BIS during CPR (arrow A). No disconnectionof the BIS monitor electrode had occurred. Thissuggests that artefact is a component of the trac-ing, but not completely so. Hence, we speculatethat the isoelectric period represents a complete

Figure 5

obstruction of pulmonary blood flow. The subse-quent brief period of ‘recovery’ of the BIS tracingrepresents the effects of the resuscitation efforts,which included thrombolysis, thus resulting insome cerebral oxygenation.

Case 5

Figure 6 demonstrates recordings from a 38-year-old man with an out of hospital cardiac arrestfollowing blunt trauma. He had an emergencythoracotomy in the ED. The BIS tracing rose withthe performance of internal cardiac massage(arrow A). There was no general body movement,such as occurs during standard CPR, and which webelieve causes movement artefact. Internal cardiacmassage was associated with a transient return ofspontaneous circulation, reflected by a prominentupward spike in the BIS tracing (arrow B).

Case 6

A 91-year-old woman with multiple medical prob-lems had an out of hospital cardiac arrest due topulseless electrical activity. She had 8 min of CPRbcww

lack of cerebral oxygenation, consequent to total

Figure 4

y paramedics, followed by a return of spontaneousirculation. She arrived at the ED intubated, andith a Glasgow Coma Score of 5. Her prognosisould normally be considered to be poor but her BIS

Figure 6

An observational study of bispectral index monitoring for out of hospital cardiac arrest 211

value at this time was 97. She subsequently madea full recovery to her pre-arrest status.

Discussion

We believe this is the first report of BIS monitor-ing for out of hospital cardiac arrest. The litera-ture on the use of BIS monitoring during cardiacarrest is limited to four case reports of periopera-tive patients.5—8

We hypothesised that because the EEG is tightlylinked to cerebral metabolism, BIS values andtrends would reflect the adequacy of cerebral per-fusion during CPR, and therefore be markers ofcerebral resuscitation. Unfortunately, movementartefact that occurs during ECC has a major impacton the BIS tracing. It is noteworthy that during thecase involving internal cardiac massage, this prob-lem was not apparent.

BIS monitoring systems utilise a variety of signalanalysis methods to detect and reduce extraneousartefacts that contaminate the EEG.2 Despite thesetechnological improvements, artefacts produced bysome non-EEG signals can interfere with the abilityofidtgctha

maathi

daiaadfofaHddo

Recent literature has emphasised the impor-tance of ECCs in the return of spontaneous circula-tion (ROSC) following cardiac arrest. Interruption ofECC significantly decreases the likelihood of ROSC.Currently, resuscitation authorities are consider-ing changes to resuscitation algorithms based onthis literature.18—23 Ideally, a reliable and sensitiveform of cerebral monitoring would provide objec-tive information on the effects of interruption ofECCs and on ROSC.

It must be stressed that the BIS monitor was notdesigned, nor has it been validated, for detectingischaemic brain injury.13 It provides a very limitedpicture of the total brain EEG activity.24 Unfortu-nately, our experience indicates that it does noteliminate the mechanical artefact induced by headmovement, such as occurs during ECCs, which has asymmetric frequency of about 2 Hz. Our pilot dataindicate that with current technology, BIS monitor-ing is not a clinically reliable marker of the efficacyof ECC. If it becomes possible to filter out the lowfrequency movement artefact, further research ofBIS monitoring in cardiac arrest may be worthwhile.

A

WA(I

R

f the BIS system to render an accurate value. Theltering system used in the BIS-XP platform usesual-channel EEG processing, making it resistanto the contamination effects of the electromyo-ram, which is the most frequent source of EEGontamination.2 The BIS system concentrates on fil-ering high frequency contamination. During CPRowever, the artefact is a much lower frequency,round 2 Hz.

Artefact has been previously reported with BISonitoring. Examples include pacmaker-induced

rtefact,12 and artefact due to the patient’s ECG orrterial pulsation being detected by the BIS moni-or but not recognised as artefact.13 Vibration ofead wires caused by air circulation has also beenmplicated.14,15

There are few descriptions of EEG monitoringuring cardiac arrest. Moss and Rockoff described62-year-old woman who had 27 s of asystole dur-

ng a carotid endarterectomy.16 Cerebral electricalctivity was lost within 15 s of asystole and returnedlmost instantly on resuscitation. Losasso et al.escribed a 72-year-old man who developed dif-use EEG slowing and suppression within 10—20 sf asystole during carotid endarterectomy.17 High-requency activity began to return to the EEGpproximately 15—20 s after CPR was initiated.owever, in the setting of out of hospital car-iac arrest, recovery of EEG activity is likely to beelayed, due to the greater magnitude and durationf the ischaemic episode.

cknowledgements

e are grateful for the support of Steve Cramp,spect Medical Systems, for loan of a BIS monitorA-2000 XP), and for the assistance of the Medicalllustrations Department, Royal Perth Hospital.

eferences

1. Myles PS, Leslie K, McNeil J, Forbes A, Chan MT. Bispec-tral index monitoring to prevent awareness during anaes-thesia: the B-Aware randomised controlled trial. Lancet2004;363:1757—63.

2. Kelley S. Monitoring level of consciousness during anesthe-sia and sedation. A clinician’s guide to the Bispectral Index.Aspect Medical Systems, Inc.; 2003.

3. Rosow C, Manberg PJ. Bispectral index monitoring. Anesthe-siol Clin North America 2001;19:947—66, xi.

4. Hayes MM, Berg RA, Otto CW. Monitoring during cardiacarrest: are we there yet? Curr Opin Crit Care 2003;9:211—7.

5. Kluger MT. The bispectral index during an anaphylactic cir-culatory arrest. Anaesth Intensive Care 2001;29:544—7.

6. England MR. The changes in bispectral index during a hypo-volemic cardiac arrest. Anesthesiology 1999;91:1947—9.

7. Szekely B, Saint-Marc T, Degremont AC, Castelain MH, Fis-chler M. Value of bispectral index monitoring during car-diopulmonary resuscitation. Br J Anaesth 2002;88:443—4.

8. Chakravarthy M, Patil T, Jayaprakash K, Shivananda N, JawaliV. Bispectral index is an indicator of adequate cerebral per-fusion during cardiopulmonary resuscitation. J CardiothoracVasc Anesth 2003;17:506—8.

9. Johansen JW, Sebel PS. Development and clinical applicationof electroencephalographic bispectrum monitoring. Anes-thesiology 2000;93:1336—44.

212 D.M. Fatovich et al.

10. Jordan KG. Continuous EEG monitoring in the neuroscienceintensive care unit and emergency department. J Clin Neu-rophysiol 1999;16:14—39.

11. Australian Bureau of Statistics. 2003 Census. Canberra: Aus-tralian Bureau of Statistics; 2003.

12. Gallagher JD. Pacer-induced artifact in the bispectral indexduring cardiac surgery. Anesthesiology 1999;90:636.

13. Myles PS, Cairo S. Artefact in the bispectral index in apatient with severe ischemic brain injury. Anesth Analg2004;98:706—7.

14. Guignard B, Chauvin M. Bispectral index increases anddecreases are not always signs of inadequate anesthesia.Anesthesiology 2000;92:903.

15. Hemmerling TM, Fortier JD. Falsely increased bispectralindex values in a series of patients undergoing cardiacsurgery using forced-air-warming therapy of the head.Anesth Analg 2002;95:322—3.

16. Moss J, Rockoff M. EEG monitoring during cardiac arrest andresuscitation. JAMA 1980;244:2750—1.

17. Losasso TJ, Muzzi DA, Meyer FB, Sharbrough FW. Electroen-cephalographic monitoring of cerebral function during asys-

tole and successful cardiopulmonary resuscitation. AnesthAnalg 1992;75:1021—4.

18. Ewy GA. A new approach for out-of-hospital CPR: a bold stepforward. Resuscitation 2003;58:271—2.

19. Frenneaux M. Cardiopulmonary resuscitation-some physio-logical considerations. Resuscitation 2003;58:259—65.

20. Kern KB. Limiting interruptions of chest compressionsduring cardiopulmonary resuscitation. Resuscitation2003;58:273—4.

21. Koster RW. Limiting ’hands-off’ periods during resuscitation.Resuscitation 2003;58:275—6.

22. Steen S, Liao Q, Pierre L, Paskevicius A, Sjoberg T. Thecritical importance of minimal delay between chest com-pressions and subsequent defibrillation: a haemodynamicexplanation. Resuscitation 2003;58:249—58.

23. Wik L. Rediscovering the importance of chest compressionsto improve the outcome from cardiac arrest. Resuscitation2003;58:267—9.

24. Mychaskiw G, Heath BJ, Eichhorn JH. Falsely elevated bis-pectral index during deep hypothermic circulatory arrest. BrJ Anaesth 2000;85:798—800.