Original contribution

Influence of bispectral index monitoring on decisionmaking during cardiac anesthesia

George Vretzakis MD (Assistant Professor)a, Eleni Ferdi MD (Anesthesiologist)a,Helena Argiriadou MD (Anesthesiologist)a, Basilios Papaziogas (Lecturer)b,*,Dimitrios Mikroulis MD (Cardiac Surgeon)c, Miltiadis Lazarides MD (Cardiac Surgeon)c,George Bitzikas MD (Cardiac Surgeon)c, George Bougioukas MDc

aDepartment of Cardiac Anesthesia, University Hospital of Alexandroupolis, 68100 Alexandroupolis, Greeceb2nd Surgical Clinic, Aristotle University of Thessaloniki, 54635 Thessaloniki, GreececClinic of Cardiac and Thoracic Surgery, Democritus University of Thrace, 68100 Alexandroupolis, Greece

Received 5 January 2004; accepted 1 December 2004

0952-8180/$ – see front matter D 2005

doi:10.1016/j.jclinane.2004.12.018

T Corresponding author. 551 33 The

Keywords:Bispectral index

monitoring;

Cardiac anesthesia;

Decision support;

Hypothermia

AbstractStudy Objective: To assess bispectral index (BIS) monitoring on decision making during cardiac

surgery with cardiopulmonary bypass (CPB) by measuring the number of preset standardized comments

with and without knowing the BIS value and by classifying the interventions following the BIS data.

Design: Prospective, randomized study.

Setting: University Hospital.

Patients: One hundred twenty-one patients scheduled for elective cardiac surgery (89 coronary patients,24 valve replacement patients, and 8 valve replacement and coronary surgery).

Interventions: Patients were divided into 3 groups. An observing anesthesiologist recorded on a specialform (bparallelQ anesthesia record) data from the devices of the workstation and the BIS monitor.

Conditions in which BIS monitoring was subjectively considered that might have been useful in

anesthetic decision making were recorded as bevents.Q In group A (36 patients), the responsible

anesthesiologist had continuous access to BIS information. In group B (44 patients), intraoperative

anesthetic management was bblindedQ to BIS values, whereas in group C (41 patients), the

anesthesiologist observing the BIS monitor was free to inform the attending anesthesiologist about

the BIS score. The number of events was considered as negatively reflecting the quality of the clinical

course of a patient. The reduction of events was considered as improvement in decision making. All

patients received the same anesthetic regimen (propofol + remifentanil). Monitoring was equal in all

cases. Mild hypothermic CPB was applied in 73 patients. Statistical analysis used 1-way analysis of

variance, Student 2-tailed t test, and v2 analysis.

Main Results: Patient demographic data, underlying pathology, operation performed, hypothermia

application, times of anesthesia, duration of operation, and CPB were similar in the 3 groups. In group

B, the BIS value was considered by the observer as useful to know in 220 events (5.00 F 1.58 per

patient). In group C, the BIS value was considered by the observer as useful to know in 143 events

Journal of Clinical Anesthesia (2005) 17, 509–516

Elsevier Inc. All rights reserved.

ssaloniki, Greece. Tel.: +30 2310 992562; fax: +30 2310 992563.

G. Vretzakis et al.510

(3.49 F 1.31 per patient, P b 0.001) and, at the same time, the attending anesthesiologist was informed

about BIS. In 112 (78.3%) cases, measures were taken. Titration of anesthetic drugs was done in

79 (70.5%) patients, whereas titration of vasoactive drugs was done in 9 (8.0%) patients, titration of both

in 13 (11.6%) patients, and other diagnostic or corrective actions in 11 (9.8%) patients. Distributions of

BIS values did not differ statistically (39.19 F 10.32, 37.38 F 10.21, and 38.29 F 10.01 in group A,

group B, and group C, respectively). bZenithQ and bnadirQ BIS values after induction also did not differ

statistically. Awakening and extubation times were similar in both groups.

Conclusions: Subjectivity, although reduced as much as possible, can play a confining role in the

value of our results. The usefulness of BIS monitoring is shown by the fact that BIS data resulted in

corrective measures. Attending anesthesiologist’s actions, based on BIS information, reduced the events

in group C.

D 2005 Elsevier Inc. All rights reserved.

1. Introduction

The bispectral index (BIS) offers a simple method for

continuous brain status monitoring throughout the admin-

istration of anesthetic or sedative drugs. Bispectral index

monitoring is considered a reliable tool in assessing the

level of consciousness. It is currently accepted that

induction and maintenance of anesthesia are associated with

a decrease in BIS value, and that increasing concentrations

of either volatile or intravenous anesthetics further decrease

the BIS value. This measurement of the hypnotic effect

proved to be accurate and reliable in nearly all patients and

clinical settings. The use of BIS monitoring has grown

rapidly in the care of patients, as it is safe and inexpensive

and it does not require special training.

During cardiac surgery, one of the aims of anesthetic

management is to achieve a level of adequacy that

eliminates the stress response, without unjustifiably pro-

longing the postoperative mechanical support of breathing.

In these operations, the result of anesthetic drugs may be

affected by various factors. Acute and significant fluctua-

tions in plasma drug concentration as a result of priming,

alterations in the rate of diuresis or in the concentration of

binding proteins, modification of the function of organs

contributing to drug metabolism, and temperature fluctua-

tions are some of the factors that may alter drug kinetics or

affect the pharmacodynamic action of anesthetic or sedative

administration. Furthermore, the relation between the

hypnotic and analgesic depth and the usual clinical criteria

of their estimation is unclear because of the administration

of various drugs such as catecholamines and b-blockers,among others, and because of the manipulations in the

central circulation. All of the above, combined with the

problem of bawarenessQ and the benefits of early extubation,

justify the application of BIS monitoring in assessing the

patient’s hypnotic state. Nevertheless, its utility during

cardiac operations and its role on decision making are

unclear. This study aimed to determine whether BIS

monitoring enhances decision making as measured by

standardized comments and interventions following the

knowledge of BIS value.

2. Materials and methods

2.1. Patient data, anesthesia, and operation

After receiving protocol approval by the Ethical Com-

mittee of the University Hospital of Alexandroupolis/

Greece, we initially enrolled into the study 130 patients

who were scheduled for elective cardiac operations. Those

patients undergoing boff-pumpQ techniques and those with

significantly compromised left ventricular function in the

preoperative evaluation were excluded from the study. In all

patients, ejection fraction (EF) was greater than 45% (mean

EF, 49.5% F 4.6%) as measured by preoperative angioven-

triculography. Patients scheduled for a combined caro-

tid endarterectomy procedure were also excluded from

the study.

Premedication consisted of diazepam 10 mg given orally

1 hour before admission to the operating room (OR). On

arrival in the OR, patients were connected to standard

monitoring (Solar 8000, Marquette Medical Systems,

Milwaukee, Wis) and a peripheral intravenous catheter

and a radial intra-arterial catheter were placed. Intraoper-

ative electrocardioscopy consisted of 5-lead, ST-segment

analysis. Anesthetic induction was achieved with a slow,

single-dose administration of midazolam (1-3 mg) plus

fentanyl (100-250 lg) plus etomidate (0.2 mg/kg) and

neuromuscular blockade with pancuronium (0.12 mg/kg).

After tracheal intubation, all patients were placed on

mechanical ventilation (Julian, Dr7eger, Lqbeck, Germany)

that was targeted to slight hypocapnia (33 mm Hg b Paco2

b37 mm Hg), based on repeated arterial blood gas sampling.

Inspiratory fraction of oxygen in air was between 0.5 and

1.0. For the anesthesia maintenance, patients received a

propofol infusion targeted to a 3-lg/mL plasma concentra-

tion, facilitated by a pump (Fresenius, Vial, Brezins,

France). Infusion rates were manually calculated to ensure

a stable concentration during the next 30 minutes, that is, the

time necessary for further anesthetic and surgical preparation.

A continuous infusion of remifentanil (c20 lg/kg per hour)was also administered as a means of providing analgesia.

The attending anesthesiologist was free to change those

rates during the procedure. The insertion of a triple-lumen

Decision-support by bispectral index during cardiac anesthesia 511

central venous catheter and a pulmonary artery catheter

(Oximetry TD catheter, Edwards Lifesciences, Germany)

was followed by the esophageal placement of a multiplane

echocardiographic probe (Toshiba Medical Systems, Tokyo,

Japan). During field preparation, echocardiographic exami-

nation was performed and the probe was positioned in the

stomach, displaying the two cavities by short axis. All

patients were operated on while on cardiopulmonary bypass

(CPB) and with cardioplegic arrest. Mild hypothermia

(c338C) was applied in 79 (60.7%) patients. In the

remainder of the patients, body temperature was allowed to

decrease naturally. In the period near the completion of the

main surgical procedure, before aortic unclamping and

during the rest of CPB, active rewarming aided by the

heart-lung machine was applied in all patients.

2.2. BIS monitoring, patient groups, and datacollection

Before anesthetic induction and after skin preparation,

the BIS sensor was placed and connected to the monitor

(BIS/XP, Aspect Medical systems, Natick, Mass). We

accepted a BIS value of less than 40 as indicating, with

high probability, that a patient is unconscious. Patients were

randomly allocated to 1 of 3 groups. In group A, the

responsible anesthesiologist had continuous access to BIS

values. In group B and group C, the BIS monitor was

located in a remote position from the bworkstationQ and BIS

values were thus unavailable. An observer anesthesiologist

noted BIS values and collected data from the monitoring

devices and infusion pumps on a special form (a bparallelQanesthesia record). For every case, participating anesthesi-

ologists were appointed randomly from a population of

three. All had at least 1 year of experience in everyday BIS

monitoring. During the anesthetic procedure and in relation

to the BIS value, the observer had the option of writing

down standardized comments as follows:

1. Expected BIS value (ie, from 45 to 35 with the plus

administration of a single dose of 50 mg of propofol

or from 35 to 30 with the initiation of hypothermic

perfusion or from 35 to 40 during rewarming);

2. Expected tendency but unexpected deviation in BIS

value (ie, very low values, less than 20 during the

insertion of the pulmonary artery catheter or rising

with rewarming or falling after the sternotomy

closure but gaps greater than 10 compared with the

previous value);

3. Unexpected BIS value. This option was for values

difficult to interpret for the period of the operation in

relation to the administrated dose of hypnotics (ie, a

value of 50 during left internal mammary artery

preparation, when in usual circumstances this period

is characterized by stability). Abrupt changes in

BIS values, not fitting to the data from the avail-

able monitoring, were included here (ie, from 35 to

45 with the use of an external pacemaker or with other

external interference without artifact recognition);

4. The BIS value would be useful. This option was for

situations where the patient and the BIS trending

were more or less stable but the absolute value

would be useful for the titration of drugs (ie, BIS in

a very low [near 20] or high [near 50] level). It was

also for problematic conditions where the knowl-

edge of patient’s hypnotic state would result in an

integrated or more comprehensive approach (ie,

BIS stable and expected in a patient with compro-

mised hemodynamics);

5. BIS value mandatory to know (ie, BIS values

increasing above 50 for more than 5 minutes or

BIS value above 60 in any case).

This categorization came after analyzing and discussing a

number of pilot cases. It became clear that additional

categorization for more objective clarification of clinical

situations would result in an extended, difficult protocol.

The standardized comments had qualitative character and

numbering was done for ease of data collection and

processing. The observer was free to record a standardized

comment whether it was subjectively estimated that the BIS

value toward the clinical conditions was similar as above.

The participating anesthesiologists considered this catego-

rization adequate for the description and judgment of

clinical events.

In group B, anesthetic support was concluded without

any information about the BIS value (bblindQ to BIS). In

group C, the observer recording a standardized comment

was free to inform the attending anesthesiologist only about

the BIS readings (BIS value, SQI percentage, EMG

percentage). At this moment, the attending anesthesiologist

could also ask to see the electroencephalogram waveform

while BIS trending was covered. The attending anesthesi-

ologist had to evaluate this information and act at his or her

own judgment. Taking into consideration the BIS data and

evaluating its clinical significance (useless, expected,

relationship to the drugs administered or to the condition

of the patient or to the phase of the procedure, unexpected,

useful, etc), the attending anesthesiologist decided whether,

and how, to intervene. However, BIS monitoring is unable

to predict future changes. For this reason, BIS values above

50 for more than 5 minutes (for two bsamplesQ conse-

quently), although until 60 indicating low probability of

explicit recall, were considered a transition from the deep

hypnotic state and the patient was recorded as a standard-

ized comment of 5. In this case (BIS value mandatory to

know), as a protocol obligation to our Ethical Committee,

the anesthesia provider was immediately informed and the

patient was removed to a subgroup (B1 and C1), irrespec-

tive of his or her original group allocation. For patients in

group B, study blindness to BIS had to be violated.

Subgroups were created for further analysis of these

patients. In group A, the attending anesthesiologist had

Table 1 Study patients

Group A

(n = 36)

Group B

(n = 44)

Group C

(n = 41)

Males 22 (61.1) 29 (65.9) 26 (63.4)

Females 14 (38.8) 15 (34.1) 15 (36.6)

Age (y) 66.2 F 8.9 65.8 F 7.9 65.2 F 8.0

Weight (kg) 75.0 F 11.9 76.6 F 13.2 75.2 F 9.7

Height (cm) 162.9 F 10.1 164.5 F 13.2 163.0 F 9.2

NYHA

I 4 (11.1) 6 (13.6) 6 (14.6)

II 29 (80.5) 34 (77.2) 30 (73.1)

III 3 (8.3) 3 (6.8) 3 (7.3)

IV 0 (0.0) 1 (2.2) 2 (4.8)

Type of operation

CABG 26 (72.2) 33 (75.0) 30 (73.1)

VR 8 (22.2) 9 (20.4) 7 (17.0)

VR + CABG 2 (5.5) 2 (4.5) 4 (9.7)

Normothermic

CPB

18 (50.0) 16 (36.3) 14 (34.1)

Hypothermic

CPB

18 (50.0) 28 (63.6) 27 (65.8)

Anesthesia

time (min)

338.8 F 65.5 343.2 F 76.8 348.0 F 72.2

Operation

time (min)

272.0 F 55.2 288.9 F 68.2 298.2 F 65.2

CPB time

(min)

98.8 F 20.5 108.9 F 24.9 110.0 F 25.9

IABP 0 (0.0) 0 (0.0) 1 (2.4)

Time to

extubation

(from ICU)

(min)

290.0 F 108.4 294.1 F 93.0 307.2 F 102.2

Data are expressed as mean F SD or number (%). Group A comprises

patients whose attending anesthesiologist had complete access to BIS

data throughout the procedure; group B, patients whose attending

anesthesiologist was blinded as to BIS data throughout the procedure;

group C, patients whose observing anesthesiologist was free to inform

(or not) the attending anesthesiologist in charge of their procedure of the

BIS data. NYHA indicates New York Heart Association classification;

CABG, coronary artery bypass graft; VR, valve replacement; IABP,

intra-aortic balloon pump.

G. Vretzakis et al.512

no obligation to report when a change was made because of

BIS monitoring, but, for uniformity reasons, a parallel

record was made without checking the standardized

comment made by the observer. The purpose of this group

was to compare only the anesthetic effect. Our hypothesis

was related to the utility of BIS in these operations and a

positive control group was necessary. In case of differences

in BIS values between groups A, B, and C, conclusions

would also extend to quality of anesthesia, as measured by

BIS, with and without its knowledge. The patients were

disconnected and BIS monitoring was stopped before they

were transferred in the intensive care unit (ICU), where

they were treated without any intervention from the study

team. The only parameter that was taken under consider-

ation was the time of extubation. Finally, all patients were

asked about possible intraoperative awareness on the third

postoperative day.

We defined an bevent Q as the clinical situation around a

standardized comment above 1, regardless of whether it

included the transmission of BIS information (group B and

group C, respectively). The only difference between groups

B and C was the awareness or unawareness of BIS readings

by the attending anesthesiologist. In this setting, the

recorded number of events was influenced by the informa-

tion about BIS. We consider the number of events as

negatively reflecting the quality of the clinical course of a

patient. We defined a reduction of events as an improvement

in decision making. Analysis would also include subgroups

B1 and C1.

2.3. Statistical analysis

Continuous data were expressed as mean F SD and

categorical data as absolute and relative frequencies. Groups

were compared in relation to somatometric and demograph-

ic data, underlying pathology, physical status, type of

operation, hypothermia application, duration of extracorpo-

real circulation, need for intra-aortic balloon support, and

duration of postoperative mechanical ventilation. Bispec-

tral index values (15 minutes after induction and every

5 minutes) were analyzed as a continuous variable. The

number of standardized comments for each patient in groups

B and C also was analyzed as a continuous variable.

Normality of continuous data was tested with the Kolmo-

gorof-Smirnof test. Analysis of continuous data was

performed with 1-way analysis of variance, post hoc

Bonferroni test (3-group comparison), and Student 2-tailed

t test (2-group comparison). Comparisons of categorical

data were made with v2 analysis. bZenithQ and bnadir Q BISvalues were also compared between groups. P values less

than 0.05 for 2-group comparisons and adjusted for

Bonferroni correction less than .0167 for 3-group compar-

isons were considered statistically significant. Data were

stored in Excel (Microsoft, Redmond, Wash) and statistical

analysis was performed using SPSS 10.0 for Windows

(SPSS, Chicago, Ill).

3. Results

Nine patients from the original population either devel-

oped major complications, requiring prolonged mechanical

ventilation in the ICU, or else died. In these patients, the

third day evaluation for intraoperative awareness was

impossible and we had to exclude them from the study. In

the remainder of the patients (n = 121), the discussion about

awareness was satisfactory and they were included regard-

less of duration of postoperative mechanical ventilation or

other support measures used. The groups that were formed

were similar and no difference of statistical importance was

shown regarding the analyzed parameters (Table 1). In all

patients, the BIS value was greater than 95 during peripheral

venous catheterization. In group A, the mean BIS value was

39.19 F 10.32; in group B, it was 37.38 F 10.21; and in

Decision-support by bispectral index during cardiac anesthesia 513

group C, it was 38.29 F 10.01, without any difference of

statistical importance between groups (Table 2). No patient

was moved from group B or group C to the B1 or C1

subgroups. In other words, the observer anesthesiologist

never noted a standardized comment of 5. Distribution of

the zenith BIS values did not differ statistically, and there

was no difference in the nadir BIS values.

Our results are shown in Table 2. The difference between

those patients who were operated on during active hypo-

thermia and those who were operated on during normother-

mia is shown. In addition, 4 distinct periods of the procedure

for each subgroup (hypothermic, normothermic) are dis-

played. In all groups, in patients who were operated on

under hypothermia, the BIS values of each period (pre-

sternotomy, pre-CPB, etc) differed statistically (P b 0.01).

In group A patients who were operated on without active

hypothermia, pre-sternotomy values (44.58 F 10.33) were

statistically different from pre-CPB values (39.21 F 9.16,

P b 0.001). Similar statistical differences were noted in

groups B and C. In the same subgroup of patients from

group A, the BIS values did not differ statistically between

the CPB and the rewarming to the end of operation periods,

as was also noted for similar patients in groups B and C.

In group A, in the period from sternotomy to institution of

CPB, BIS values were 38.08 F 8.20 and 39.21 F 9.16 (ns)

with respect to whether hypothermia was to be applied. In

the next operative period, from institution of CPB to the

beginning of rewarming, BIS values were 33.91 F 6.66 and

37.22F 8.98 (P b 0.001), respectively. In group B and with

respect to the above-mentioned periods, BIS values were

Table 2 Results

Group A (n = 36)

BIS valuea 39.19 F 10.32

Zenith BISa 52.08 F 12.22

Nadir BISa 26.73 F 2.82

Standardized comments N1 none

Standardized comment N1 per pt none

Propofol administration (mL/kg per hour) 0.470 F 0.053b

Remifentanil (ml/kg per hour) 0.190 F 0.021

CBP temp (hypothermic pts) (8C) 33.82 F 0.563

CBP temp (normothermic pts) (8C) 5.15 F 0.427

BIS values

Hypothermic pts (n = 18)

Pre-sternotomy perioda 43.90 F 11.82

Pre-bypass period 38.08 F 8.20

CPB 33.91 F 6.66

Rewarming–end of operation 37.54 F 8.02

Normothermic pts (n = 18)

Pre-sternotomy perioda 44.58 F 10.33

Pre-bypass period 39.21 F 9.16

CPB 37.22 F 8.98

Rewarming–end of operation 39.07 F 9.81a Values not including the awake patient until 15 minutes after induction.b Denotes statistical difference (3-group comparison).

35.04 F 7.68 and 37.11 F 8.04 (ns) and 32.47 F 5.78

and 36.09 F 8.67 (P b 0.001). In group C, BIS values were

35.92 F 6.90 and 37.88 F 7.72 (ns) and 33.04 F 5.00 and

37.01 F 7.94 (P b 0.001) in the hypothermic and

normothermic CPB patients, respectively. In the period from

rewarming to the end of the operation, BIS values were

increased significantly for the hypothermic CPB patients

regardless of their group allocation. Finally, as is shown

bhorizontallyQ in Table 2, comparisons of the 3 group

distributions revealed no differences in statistical importance.

In group B, a standardized comment above 1 was noted

in 220 cases (5.00 F 1.58 per patient). In group C, the

observer rated the BIS value qualitatively above 1 (stan-

dardized comment N1) in 143 cases (3.49 F 1.36 per

patient, P b 0.001) and so informed the attending anesthe-

siologist. In 112 (78.3%) of these events in group C, the

attending anesthesiologist proceeded to take corrective

measures. In 79 (70.5%) events, these measures involved

titration of anesthetics; in 9 (8.0%) events, they involved

titration of vasoactive drugs; and in 13 (11.6%) events, titra-

tion of both. Finally, in 11 (9.8%) cases, the attending

anesthesiologist took other diagnostic or corrective actions

or measures. In these cases, actions first written in the

parallel record, when the attending anesthesiologist became

aware of the BIS value of the event, were the following:

1. additional sampling for blood gas analysis (3);

2. zeroing of the pressure signals (3);

3. pump transfer to the heart-lung machine (valve ope-

ration with double venous cannulation) (2);

Group B (n = 44) Group C (n = 41) P

37.38 F 10.21 38.29 F 10.01 ns

53.72 F 12.04 49.82 F 11.90 ns

24.30 F 2.92 25.02 F 2.68 ns

220 143

5.00 F 1.58 3.49 F 1.36 b0.001

0.518 F 0.058 0.512 F 0.056 b0.0167

0.200 F 0.024 0.194 F 0.022 ns

33.74 F 0.478 33.76 F 0.502 ns

35.16 F 0.432 35.16 F 0.587 ns

(n = 28) (n = 27)

42.02 F 11.36 42.38 F 11.11 ns

35.04 F 7.68 35.92 F 6.90 ns

32.47 F 5.78 33.04 F 5.00 ns

36.86 F 7.87 37.49 F 7.11 ns

(n = 16) (n = 14)

42.96 F 12.08 44.09 F 13.62 ns

37.11 F 8.04 37.88 F 7.72 ns

36.09 F 8.67 37.01 F 7.94 ns

38.82 F 8.92 40.66 F 8.61 ns

G. Vretzakis et al.514

4. control and change of the anesthetic administration

line (2);

5. titration of ventilation (1).

These were or were not followed by other actions.

The consumption of propofol, normalized to the patient’s

weight and to duration of the duration, was statistically less

in group A (0.470 F 0.053 mL/kg per hour) than in groups

B or C (0.518 F 0.058 and 0.512 F 0.056 in mL/kg per

hour, respectively; ns). The normalized consumption of

remifentanil showed insignificant differences between the

3 groups (Table 2). Postoperatively, no intergroup differ-

ences were found with respect to time of extubation in the

ICU. Finally, no patient in any group experienced awareness

during anesthesia.

4. Discussion

Bispectral index monitoring has been reported to be

superior in evaluating anesthetic depth compared with other

clinical signs. In noncardiac anesthesia, BIS monitoring

helps to avoid underdosage or overdosage and their

consequences as light anesthesia or unjustified prolongation

of awakening. Furthermore, the usefulness of BIS as an

objective tool for the evaluation of needs and the analysis of

costs has also been documented [1-5]. However, none of

these studies has actually investigated its utility in decision

making. In cardiac anesthesia, some of the advantages of

BIS monitoring are minimized because the intraoperative

anesthetic drug support is usually based on the administra-

tion of large amounts of opioids [6,7]. Although its utility in

the titration of anesthetics or its accuracy is not altered with

the use of high doses of opioids, its ability to facilitate rapid

recovery and spontaneous ventilation is lost. Furthermore,

in these operations, the immediate weaning of ventilatory

support is impossible in most of the cases. Some inves-

tigators therefore wonder about the usefulness of BIS

monitoring in cardiac operations. Our study was designed

to clarify the contribution of intraoperative BIS monitoring

to decision making in the course of cardiac operations under

extracorporeal circulation and hypothermia.

To clarify our hypothesis, the study design excluded

patients undergoing off-pump techniques or patients who

would likely not be available for a third day of evaluation.

The low preoperative EF is a persistent factor affecting the

time of extubation [8,9]. We also excluded patients

programmed for a combined carotid endarterectomy proce-

dure, as the relation of BIS readings to brain ischemia or

carotid clamping (ipsilateral or contralateral) is unclear. In

one report, BIS values were lower during hypoxia [10].

Currently, it seems that during episodes of brain hypoxia,

clinicians follow BIS values with a question mark. Finally,

we excluded 9 patients from the original population, as

they required prolonged mechanical ventilation in the ICU

or had a bad outcome; thus, a third day of evaluation for

intraoperative awareness was impossible. Nevertheless, the

intraoperative BIS values of the excluded subjects showed

no difference compared with all the others. Other character-

istics might affect their outcome, but this evaluation was

considered irrelevant to the hypothesis of our study.

In group B, anesthetic support was performed with no

access to BIS values. In 220 cases (5 times per patient),

checked as standardized comments, the observer considered

that BIS data would be useful, either because a particular act

had an insufficient result or it led to an unnecessarily greater

effect or because interpretation of the patient’s status was

incomplete and the measures taken were inadequate. Free

notes about the BIS value were plenty in the parallel

anesthesia record, concerning the changes in surgical

stimuli, the hemodilution, the alterations of temperature,

and the profound diuresis. In one case, malfunction of the

propofol pump was noticed by the observer before activa-

tion of the alarm.

In group C, the attending anesthesiologist was informed

about the BIS value in 143 cases (approximately 3.5 times

per patient). These cases were qualitatively similar to the

standardized comments in group B, they were checked in

the same way, and they were followed by similar free notes.

Information about BIS value resulted in fewer events. In

other words, the informed provider proceeded to actions that

reduced the number of incidents in which the BIS was

considered of value for management of the patient. The

value of BIS data is reflected in the fact that in most of the

cases (78.3%), measures were taken. Furthermore, this

positive value shows a decreased subjective dependence

by the observer on the standardized comments. The actions

of the attending anesthesiologist, after having the BIS data,

were grouped as per our results above. It seems that

knowledge of the adequacy or inadequacy of anesthesia

completes the hemodynamic image. In 15.3% of the total

events, the attending anesthesiologist addressed his actions

to the direction of hemodynamic handling, in spite of the

adequacy of the available monitoring. In many cases, the

decision to administer b-blockers and to increase or

decrease the inotropic and vasoactive drug administration

was made after taking into consideration the BIS value as

well. Finally, as it is reported, the attending anesthesiologist

proceeded to other diagnostic or corrective actions or

measures. In one case, during the pre-CPB period and after

a satisfactory control of the equipment, in a patient with a

heart rate (HR) of 100 beats per minute (bpm) and a blood

pressure (BP) of 145/88 mm Hg, hypoventilation was

proven when the attending anesthesiologist proceeded

to arterial blood gas analysis, as the BIS value was

47 (standardized comment = 4). It seemed that the BIS-

expressed adequacy of hypnosis turned his attention to other

causes of this hemodynamic picture. In another, a little

before sternotomy, the attending anesthesiologist only

confirmed the accuracy of transducers in a patient with an

HR of 62 bpm and BP of 75/52 mm Hg when the BIS value

was 18 (standardized comment = 2). It seemed that, waiting

Decision-support by bispectral index during cardiac anesthesia 515

for the sternotomy and knowing BIS, the responsible

anesthesiologist decided only to verify the BP level. The

attending anesthesiologist correctly re-zeroed the transducer

systems before transferring the patient to ICU, when in-

fusion of anesthetics was already lowered, HR was 65 bpm,

BIS values were 42 to 45 (standardized comment = 4), and

BP was falsely 178/100 mm Hg. He also proceeded to new

settings in the ventilatory mode in a patient who during left

internal mammary artery preparation was sweating, after

controlling the equipment and considering the BIS value of

45 (standardized comment = 4).

Patients from group A served somewhat as the control

group in this study. In many instances, the observer wrote,

as a free note in the parallel record, that anesthetic

management was targeted to some specific BIS score. At

most of the times, this was higher than what was displayed

on the BIS screen. That was common particularly during the

initiation of hypothermic CPB. Unfortunately, this obser-

vation cannot be further analyzed because of limitations in

our protocol. Although BIS values were highest in group A,

the anesthetic result, as measured by BIS, did not differ

between the 3 groups of patients. Furthermore, BIS values

could be considered low in all groups [11]. We believe that,

in our setting, these observations come from focusing

mainly on propofol’s plasma concentration [12]. Actually,

in group A, continuous BIS monitoring and lesser

uncertainty about adequacy of anesthesia allowed the

provider to significantly reduce propofol’s infusion rate. It

seems that, in the conditions of cardiac anesthesia, BIS

monitoring serves as a tool for significantly less adminis-

tration of anesthetics without increasing the frequency of

moving to clinically undesirable levels of anesthetic

inadequacy. Finally, the regimen that was used in our study

allowed low BIS values without the need for prolonged

postoperative mechanical ventilation, as has been seen

elsewhere [5]. Nevertheless, it could be suggested that

anesthesia guided to values closer to 50 would be more

appropriate for fast-tracking.

Hypothermia is known to cause increased sedation and

loss of consciousness. During cardiac operations, the effect

of hypothermia on BIS values has been extensively

evaluated [13-16], thus allowing decision making to be

based on BIS regardless of the temperature. In our patients,

the application of mild hypothermic CPB resulted in

decreased BIS values. The attending anesthesiologists had

adequate experience in everyday BIS monitoring and they

adjusted, blindly or not, the infusion rate of the anesthetic

drugs. We consider that this experience contributed to the

only small absolute decrease of BIS values during the

hypothermic CPB noted in our patients.

Observer subjectivity could play a limiting role in the

results noted in our study. We tried to reduce the subjective

character in our methods, although we accept that it is

always present to some extent in behavioral observations.

However, given the fact that information resulted in

measures and reduced events in group C and from the rest

of our findings as discussed above, the usefulness of BIS

monitoring during cardiac operations is evident. Actually,

our study directly shows the actions following the intermit-

tent knowledge of BIS value. In conclusion, our results

suggest that BIS data, as part of the intraoperative

assessment in cardiac operations, provide the clinician with

an ability to make more informed decisions about the dosing

and balance of anesthetic agents and other adjuvant

therapies such as analgesics and cardioactive agents.

Acknowledgments

The authors are indebted to several people for their

contribution to this work. We thank the statistician J

Tripsianis for helping us with the statistical analysis. We

also thank M Vergoulidou, D Kyriakidou, D Artemis, and D

Patridas, students of Medicine of the Democritus University

of Thrace, who worked willingly in the OR and ICU for the

collection and processing of the data.

References

[1] Glass PS, Bloom M, Kearse L, Rosow C, Sebel P, Manberg P.

Bispectral analysis measures sedation and memory effects of propofol,

midazolam, isoflurane and alfentanil in healthy volunteers. Anesthe-

siology 1997;86:836-47.

[2] Katoh T, Suzuki A, Ikeda K. Electroencephalographic derivatives as a

tool for predicting the depth of sedation and anesthesia induced by

sevoflurane. Anesthesiology 1998;88:642 -50.

[3] Struys M, Somers AA, Van Den Eynde N, et al. Cost reduction

analysis of propofol versus sevoflurane: maintenance of anaesthesia

for gynaecological surgery using the bispectral index. Eur J

Anaesthesiol 2002;19:727 -34.

[4] Yli-Hankala A, Vakkuri A, Annila P, Korttila K. EEG bispectral

index monitoring in sevoflurane or propofol anaesthesia: analysis of

direct costs and immediate recovery. Acta Anaesthesiol Scand 1999;

43:545-9.

[5] Lehmann A, Karzau J, Boldt J, Thaler E, Lang J, Isgro F. Bispectral

index-guided anesthesia in patients undergoing aortocoronary bypass

grafting. Anesth Analg 2003;96:336-43.

[6] Heck M, Kumle B, Boldt J, Lang J, Lehmann A, Saggau W.

Electroencephalogram bispectral index predicts hemodynamic and

arousal reactions during induction of anesthesia in patients undergoing

cardiac surgery. J Cardiothorac Vasc Anesth 2000;14:693-7.

[7] Driessen JJ, Harbers JB, van Egmond J, Booij LH. Evaluation of the

electroencephalographic bispectral index during fentanyl-midazolam

anaesthesia for cardiac surgery. Does it predict haemodynamic

responses during endotracheal intubation and sternotomy? Eur J

Anaesthesiol 1999;16:622 -7.

[8] Habib RH, Zacharias A, Engoren M. Determinants of prolonged

mechanical ventilation after coronary artery bypass grafting. Ann

Thorac Surg 1996;62:1164-71.

[9] Bezanson JL, Deaton C, Craver J, Jones E, Guyton RA, Weintraub

WS. Predictors and outcomes associated with early extubation in older

adults undergoing coronary artery bypass surgery. Am J Crit Care

2001;10:383-90.

[10] Okawa H, Hirota K, Sakai I, Tsubo T, Ishihara H, Matuki A. Early

detection of hypoxia with BIS monitoring during percutaneous

cardiopulmonary support. Masui 2001;50:429 -32.

G. Vretzakis et al.516

[11] Kelley SD, editor. Monitoring level of consciousness during anesthe-

sia and sedation. A clinician’s guide to the bispectral index. Natick

(Mass)7 Aspect Medical Systems Inc; 2003. p. 32- 47.

[12] White M, Kenny GN. Intravenous anaesthetic agents: delivery

systems. In: Nimmo WA, Rowbotham DJ, Smith G, editors.

Anaesthesia. 2nd ed. Oxford (UK)7 Blackwell Scientific Publications;

1994. p. 106 -8.

[13] Mathew JP, Weatherwax KJ, East CJ, White WD, Reves JG.

Bispectral analysis during cardiopulmonary bypass: the effect of

hypothermia on the hypnotic state. J Clin Anesth 2001;13:301-5.

[14] Hemmerling TM, Fortier JD. Falsely increased bispectral index

values in a series of patients undergoing cardiac surgery using

forced-air-warming therapy of the head. Anesth Analg 2002;95:

322-33.

[15] Doi M, Gajraj RJ, Mantzaridis H, Kenny GN. Effects of cardiopul-

monary bypass and hypothermia on electroencephalographic varia-

bles. Anaesthesia 1999;52:1048-55.

[16] Honan D, Doherty D, Frizelle H. A comparison of the effects of mild

versus moderate hypothermia during cardiopulmonary bypass on the

EEG bispectral index. Eur J Anaesthesiol 2002;19(S24):26.