British Journal of Anaesthesia 84 (6): 74952 (2000)

Fentanyl and midazolam anaesthesia for coronary bypass

surgery: a clinical study of bispectral electroencephalogram
analysis, drug concentrations and recall
wall1 and J. G. Jakobsson1 2*
G. Barr1, R. E. Anderson1, S. Samuelsson1, A. O
1Department

of Cardiothoracic Anaesthetics and Intensive Care, Karolinska Hospital, Stockholm, Sweden.

2Department of Anaesthetics, Sabbatsbergs Hospital, Stockholm, Sweden

*Corresponding author: Department of Cardiothoracic Anaesthetics and Intensive Care, Karolinska Hospital, S-171 76
Stockholm, Sweden
Bispectral index (BIS) was assessed as a monitor of depth of anaesthesia during fentanyl and
midazolam anaesthesia for coronary bypass surgery. In 10 patients given morphine premedication, anaesthesia was induced with a combination of midazolam and fentanyl and thereafter
maintained with a continuous infusion of a mixture of midazolam and fentanyl 5 and 50 g
kg1 h1, respectively. BIS was recorded continuously but not shown to the attending
anaesthetist. Plasma concentrations of midazolam and fentanyl were measured five times during
the procedure. An auditory stimulus was given during bypass. All patients were interviewed
twice after operation for explicit and implicit recall. No patient had any anaesthetic complications.
BIS decreased during anaesthesia, but varied considerably during surgery (range 3691) with
eight patients having values 60. Midazolam and fentanyl drug concentrations did not correlate
with BIS. No patient reported explicit or implicit recall. During clinically adequate anaesthesia
with midazolam and fentanyl BIS varies considerably. The most likely reason is that BIS is not
an accurate measure of the depth of anaesthesia when using this combination of agents.
Br J Anaesth 2000; 84: 74952
Keywords: analgesics opioid, fentanyl; electroencephalography; monitoring, electroencephalography; surgery, cardiovascular
Accepted for publication: January 24, 2000

Patients about to undergo general anaesthesia may fear that

they will be aware of the operation and be able to remember
it afterwards. Awareness with recall is rare, though, occurring in about 0.2% of patients undergoing anaesthesia for
general surgery.1 A greater incidence, 1%, has been
reported during cardiac surgery.2
The bispectral index (BIS), derived from bispectral analysis of the electroencephalogram, has recently been introduced to monitor the depth of anaesthesia. Ideally such a
monitor should measure the depth of anaesthesia irrespective
of the anaesthetic agent(s) used or the type of surgical
procedure. We investigated the clinical value of BIS during
fentanyl and midazolam anaesthesia for coronary artery
bypass grafting (CABG) using cardiopulmonary bypass
(CPB). A standardized auditory stimulus was presented
at predetermined times during CPB and patients were
interviewed for explicit and implicit recall. Plasma concentrations of fentanyl and midazolam were measured.

Methods
Ten patients with a mean age of 69 yr (range 5083 yr)
and scheduled for elective CABG surgery were studied

after approval from the hospital ethical committee and

informed consent. All patients were seen the day before
surgery and informed about the study and its objectives.
They were told that a sound (not specified in detail) would
be presented during anaesthesia that they would later
be asked to recall. About an hour after intramuscular
premedication with morphine 515 mg, anaesthesia was
induced with midazolam 0.05 (0.030.12) mg kg1 and
fentanyl 10 (712) g1. Pancuronium 0.1 mg1 was given
as a muscle relaxant to facilitate endotracheal intubation.
A continuous infusion of fentanyl and midazolam was
started (5 and 50 g kg1 h1, respectively) and maintained
until completion of surgery; supplementary doses of fentanyl
or midazolam were given at the discretion of the anaesthetist.
Standard procedures were used for CPB. Core temperature was reduced or was allowed to drift to 34C. An
auditory stimulus (100 dB, duration 60 s) was presented
five times during CPB. BIS (Aspect Medical Systems;
version 3.12, Natick, MA, USA) was measured continuously; the attending anaesthesiologist was not shown the
output values.

The Board of Management and Trustees of the British Journal of Anaesthesia 2000

Barr et al.

Fig 1 Bispectral index at various times during coronary bypass surgery

(n10): 1, before induction; 2, sternotomy; 3, heparin; 4, start of bypass;
5, end of bypass; 6, skin closure.

Fig 2 Bispectral index plotted against midazolam plasma concentration

(ng ml1) during bypass surgery (n10).

Table 1 Plasma drug concentrations (median (range)) and bispectral index

(BIS) during bypass surgery (n10)
Time during
procedure

Before induction
Sternotomy
Heparin
Start of bypass
End of bypass
End of surgery

Midazolam
(ng ml1)

0
78
102
90
94
103

(49133)
(52158)
(54113)
(55145)
(87122)

Fentanyl
(ng ml1)

0
5.0
6.4
4.8
4.9
5.2

(3.77.8)
(3.411.9)
(2.87.8)
(3.06.6)
(3.210.2)

BIS (no. of patients)

60

60

75

0
6
7
5
7
6

10
4
3
5
3
4

10
0
1
2
1
1

Blood samples were drawn from the radial artery catheter

before induction, at sternotomy, when heparin was given,
at the start, middle and end of CPB, and at skin closure.
Samples were stored in a freezer for blinded, batch analysis.
Patients were interviewed on each of the first 2 days
after the operation using a fixed predefined questionnaire.
At the first interview they were asked to describe their
experience of anaesthesia. At the second meeting the patients
were also asked specifically about recall of any special
sound. They were given a number of associative proposals.
If there was still no recall, the actual sound was presented
and the patients were asked again for recollections from
the surgical procedure.

Results
The 10 patients studied had mean (SD) body weight of 78
(13) kg. All patients had an uncomplicated surgical course,
with a mean surgical time 158 (19) min and mean anaesthesia
time 198 (35) min. Haemodynamic measurements were
stable.
All patients were awake and fully oriented at arrival in
the operating theatre and at induction BIS was between 96
and 99. BIS decreased in all patients with induction, but
during surgery it varied between 36 and 91. Almost 40%
of BIS values at the time of blood sampling were 60
(Table 1, Fig. 3).

Fig 3 Bispectral index plotted against fentanyl plasma concentration (ng

ml1) during bypass surgery (n10).

Only two of the 10 patients studied did not have BIS

values or 60 at some point during surgery. The BIS values
in each subject during anaesthesia varied, with a range that
was as small as 15 or as large as 50 (median range was 30).
No drug accumulation occurred. The median plasma
concentrations of fentanyl and midazolam did not change
during anaesthesia (Table 1). Plasma drug concentrations
varied widely between individuals. BIS did not correlate
with fentanyl (r20.13) or midazolam plasma concentration
(r20.034; Figs 2 and 3). In individual patients, there was
no correlation between drug plasma concentration changes
and BIS.
No patients reported explicit recall of intraoperative
events at interview on the first or second day after operation.
No patient could recall the acoustic stimulus presented
repeatedly during CPB, either after given strong associations
or when actually hearing it.

Discussion
General anaesthesia is a state of pharmacologically induced
depression of the central nervous system causing uncon-

750

Cardiac surgery and BIS

sciousness. Anaesthesia is not an onoff phenomenon but a

continuum, a gradual depression of cognitive and autonomic
functions. Most anaesthetics cause a dose-dependent
increase in anaesthetic depth. Many hypnotics and anaesthetics also cause varying degrees of amnesia.3 Recollection
of a stimulus requires an initial activation of the sensory
system, central processing of the information and, finally,
retrieval of information. The memory process can be separated into explicit or conscious learning and implicit or
unconscious retrieval of information.
Awareness with explicit recall of intraoperative events is
a feared complication of general anaesthesia with potentially
serious adverse psychological sequelae.4 The incidence of
explicit awareness with recall after general anaesthesia is,
however, small: about two per thousand patients.1 Implicit
processing has been shown to be less affected by general
anaesthesia than explicit memory processing and may
continue subconsciously during clinically adequate anaesthesia.5 Implicit intraoperative awareness and postoperative
recall of auditory stimuli have been reported in up to half
of patients undergoing cardiac surgery with opiate and
benzodiazepine anaesthesia.6 7
Most studies of recall suggest that hearing is one of the
most receptive sensory channels for perception during
general anaesthesia.8 9 Although explicit recall, especially
of pain and paralysis, is a serious complication, adjusting
anaesthetic depth to preserve implicit processing of auditory
information could be useful. Several studies suggest new
therapeutic possibilities with implicit processing10 that
would need carefully titrated depth of anaesthesia however
this places additional demands on monitoring.
The ideal monitor of depth of anaesthesia should be able
to measure accurately different degrees of anaesthetic depth,
such as the transition from wakefulness to unconsciouness,
loss of explicit as well as implicit memory processing, loss
of motor response to surgical stimulation and loss of
autonomic response. Furthermore, all common anaesthetics
should be comparable on the same scale. Bispectral analysis
of the electroencephalogram can estimate the hypnotic
effects of propofol and halogenated agents. BIS has been
proposed to be a useful monitor during cardiac surgery with
CPB11 when the usual haemodynamic responses and signs
such as sweating or tear production are less dependable
markers of anaesthetic depth. In addition, the requirement
for early extubation in some patients would make such a
monitor useful to avoid excessive anaesthesia. BIS may
also be useful in determining the degree of sedation in
intensive care patients.12
Anaesthesia for coronary surgery is frequently based on
a combination of a large dose of an opioid with a hypnotic
or a volatile agent. In this study a combination of fentanyl
and midazolam was used and anaesthesia was adjusted
according to clinical needs at the discretion of the anaesthesiologist. Midazolam and fentanyl have a synergic effect on
cognitive depression. In studies of the interactions between
fentanyl and sevoflurane, and fentanyl and propofol, a

maximum effect for fentanyl occurs at a plasma level of

about 34 ng ml1.1315 The plasma concentrations in the
present study were greater than this, but with large differences between patients, and considerably less than those
found in the study by van der Maaten and colleagues of
patients having cardiac surgery.16 The large variation in
plasma concentrations we found supports the observations
of van der Maaten and colleagues who also found a two to
threefold variation when using midazolam and an opioid.16
The midazolam and fentanyl concentrations throughout
the procedure did not differ from their initial values,
demonstrating that the continuous drug infusion technique
provides unchanged drug concentrations during CPB and
that no drug accumulation occurs during the procedure.
The considerable variation in BIS observed here is, in
some respects, similar to that found during sedation of
intensive care patients.12 In our study BIS not only varied
considerably but was, in general, surprisingly large. Such
values are associated with a great likelihood of awareness
and the capability to follow command.17 18 A BIS value of
60 is often regarded as the criterion for adequate anaesthesia, while a value of 70 is frequently seen during
awakening.18 Several factors may have led to the observed
variation in BIS. Hypothermia may have affected the BIS
measurements during bypass, as discussed by Doi and
colleagues.19 Cardiac surgery with profound hypothermia
affects electroencephalographic recordings because of
changes in conductance,19 but our patients were operated
on at 34C and gave no problems with recording apart from
during diathermy. Furthermore, the variability was present
both before and after the bypass period. The variability
may, of course, be explained by inadequate depth of
anaesthesia. Although we detected no apparent awareness
during surgery, and although haemodynamic measurments
were stable, muscle relaxants, beta-blocking drugs and
blood volume changes could have masked many of these
clincial indicators of adequate anaesthesia. No patient
reported recall when actively interviewed on two occasions
after operation, even though auditory stimuli are resistant
to suppresion by fentanyl and midazolam anaesthesia. No
consistent decrease in BIS was found with increasing
plasma concentrations of either drug. The most plausible
explanation for the high BIS values seems to be that BIS
reflects anaesthetic depth less accurately when fentanyl and
midazolam are used as anaesthetics. We have shown in a
previous study that increasing concentrations of nitrous
oxide have no effect on BIS,20 and addition of fentanyl to
propofol anaesthesia causes large BIS values.21 22 The
weak correlation between decreasing BIS and an increasing
fentanyl concentration supports results obtained in
volunteers.17
As discussed above, no patient recalled events during the
operation. The timing of, and technique for, evaluating
awareness and recall have considerable impact on the
results.1 The lack of any explicit recall in this study with
so few patients is not surprising, as explicit recall is

751

Barr et al.

infrequent.1 The lack of implicit recall of the auditory

stimulation during surgery, which should be detectable even
in the small number of patients in this study, confirms the
adequacy of the anaesthetic depth. Technique and time of
stimulation are important for implicit memory testing.23
In conclusion, this study has shown that although anaesthetic depth was adequate and no recall was reported,
BIS varied widely and values that are usually related to
excessively light anaesthesia or wakefulness were occasionally observed. BIS did not correlate with midazolam and
fentanyl plasma concentrations. A measurement such as
BIS should, ideally, be suitable for determining the depth
of anaesthesia for all anaesthetics on a common scale. For
coronary surgery with fentanyl and midazolam anaesthesia,
the BIS algorithm too often underestimates the depth
of anaesthesia and does not accurately reflect plasma
concentrations. Administering more anaesthetic to try to
decrease large BIS values would result in an excessive
depth of anaesthesia with delayed awakening and, perhaps,
haemodynamic instability. BIS may be well suited to
predominantly hypnotic drugs, but it is not useful for
the midazolamfentanyl combination commonly used for
cardiac surgery.

References
1 Nordstro m O, Engstro m AM, Persson S, Sandin R. Incidence of
awareness in total i.v. anaesthesia based on propofol, alfentanil
and neuromuscular blockade. Acta Anaesthiol Scand 1997; 41:
97884
2 Phillips AA, McLean RF, Devitt JH, Harrington EM. Recall of
intraoperative events after general anaesthesia and
cardiopulmonary bypass. Can J Anaesth 1993; 40: 9226
3 Polster MR, Gray PA, OSullivan G, McCarthy RA, Park GR.
Comparison of the sedative and amnestic effects of midazolam
and propofol. Br J Anaesth 1993; 70: 61216
4 Bailey AR, Jones JG. Patients memories of events during general
anaesthesia. Anesthesia 1997; 52: 46076
5 Bethune DW, Bhosh S, Gray B, Kerr L, Walker IA, Doolan LA,
Harwood RJ, Sharples LD. Learning during general anaesthesia:
implicit recall after methohexitone or propofol infusion. Br J
Anaesth 1992; 69: 1979
6 Goldman L, Shah MV, Hebden MW. Memory and cardiac
anaesthesia. Anesthesiology 1987; 42: 596603
7 Schwender D, Kaiser A, Klasing S, Peter K, Po ppel E. Midlatency
auditory evoked potentials and explicit and implicit memory in
patients undergoing cardiac surgery. Anesthesiology 1994; 80:
493501

8 Hilgenberg JC. Intraoperative awareness during high-dose

fentanyloxygen anesthesia. Anesthesiology 1981; 54: 3413
9 Jones JG. Hearing and memory in anaesthetised patients. BMJ
1986; 292: 12913
10 Dreher H. Mindbody interventions for surgery: evidence and
exigency. Adv MindBody Med 1998; 14: 20722
11 Sebel PS. Central nervous system monitoring during open heart
surgery: an update. J Cardiothorac Vasc Anesth 1998; 12: 38
12 DeDeyne C, Struys M, Decruyenere J, Creuplandt J, Hoste E,
Colardyn F. Use of continuous bispectral EEG monitoring to
assess depth of sedation in IDU patients. Intens Care Med 1998;
24: 12948
13 Thomson AR, Henderson BR, Singh K, Hudson RJ. Concentration
response relationships for fentanyl and sufentanil in patients
undergoing coronary artery bypass grafting. Anesthesiology 1998;
89: 85261
14 Katoh T, Ikeda K. The effects of fentanyl on sevoflurane
requirements for loss of consciousness and skin incision.
Anesthesiology 1998; 88: 1824
15 Kazama T, Ikeda K, Morita K. Reduction by fentanyl of the Cp50
values of propofol and hemodynamic responses to various noxious
stimuli. Anesthesiology 1997; 87: 21327
16 Van der Maaten JMAA, Epema AH, Gallandat Huet RCG, Hennis
PJ. The effect of midazolam at two plasma concentrations on
hemodynamics and sufentanil requirement in coronary artery
surgery. J Cardiothor Vasc Anesth 1996; 10: 35663
17 Glass P, Bloom M, Kearse L, Rosow C, Sebel P, Manberg P.
Bispectral analysis measures sedation and memory effects of
propofol, midazolam, isoflurane, and alfentanil. Anaesthesiology
1997; 86: 83647
18 Katoh T, Suzuki A, Ikeda K. Electroencephalographic derivatives
as a tool for predicting depth of sedation and anaesthesia induced
by sevoflurane. Anaesthesiology 1998; 88: 64250
19 Doi M, Gajraj RJ, Mantzaridis H, Kenny GN. Effects of
cardiopulmonary bypass and hypothermia on electroencephalographic variables. Anaesthesia 1997; 52: 104855
wall A, Anderson RE. Nitrous oxide does
20 Barr G, Jakobsson JG, O
not alter bispectral index: a study with nitrous oxide as sole agent
and as adjunct to intravenous anaesthesia. Br J Anaesth 1999; 82:
82730
21 Iselin-Chaves I, Flaishon R, Sebel P, Howell S, Gan TJ, Sigl J,
Ginsberg B, Glass P. The effect of the interaction of propofol and
alfentanil on recall, loss of consciousness, and the bispectral index.
Anesth Analg 1998; 87: 94955
22 Mi WD, Sakai T, Singh H, Kudo T, Kudo M, Matsuki A. Hypnotic
endpoints vs the bispectral index, 95% spectral edge frequency
and median frequency during propofol infusion with or without
fentanyl. Eur J Anaesth 1999; 16: 4752
23 Munte S, Kobbe I, Demertzis A, Lullwitz E, Munte TF, Piepenbrock
S, Leuwer M. Increased reading speed for stories presented during
general anesthesia. Anesthesiology 1999; 90: 6629

752