Luis Gaitini MD, Sonia Vaida MD, Geoffrey Collins MB,
Mostafa Somri MD, Edmond Sabo MD
section under general
anaesthesia using EEG
This study examined the relationship between the EEG (spectral
edge frequency 90 - SEFgO) and the occurrence of awareness
defined for the purpose of this study as responsiveness to verbal
commands. Fifty women undergoing general anaesthesia for
elective Caesarean section were examined. Responsiveness to
verbal commands was detected every minute in the period from
the induction of anaesthesia to the delivery of the newborn
using the Tunstall isolated forearm technique and correlated
with the SEF90 value. The patients were assigned by a ran-
domized code to receive either thiopentone (4 mg. kg -t) or
ketamine (1 mg " kg-t) for induction of anaesthesia. Before the
administration of succinylcholine a tourniquet was applied to
the free arm, and inflated to 200 mmHg, to maintain motor
function to one arm. The EEG recordings started five minutes
before induction and were recorded throughout anaesthesia.
The incidence of responsiveness to verbal commands was lower
in the ketamine group (24%) where the average SEFgO was
12.0 -t- 3 Hz, than in the thiopentone group (52%), where the
average SEFgO was 18.09 + 3 Hz (P= 0.01). The results suggest
that SEF values of _<8.6 Hz were sufficient to avoid respon-
siveness to verbal commands.
Cette dtude dvalue la relation entre l'analyse de fr~quence spec-
tra& (spectral edge frequency 90 - SEFgO) sur I'EEG et l~veil
intra-opdratoire d~fini darts cette dtude comme une rdponse ?t
un ordre verbal. L~tude inclut 50 parturientes op~r~es sous
anesth~sie g~n~rale pour une cdsarienne ~lective. Nous avons
ANAESTHESIA: obstetrical, Caesarean section;
From the Department of Anesthesiology, Bnai-Zion Medical
Centre, Haifa, Israel.
Address correspondence to: Dr. Luis Gaitini, Department of
Anesthesiology, Bnai-Zion Medical Center, P.O. Box 4940,
Haifa 31048, Israel.
Accepted for publication 1st January, 1995.
ddceld les rdponses ~un ordre verbal h chaque minute en corre-
lation avec les valeurs du SEFgO pendant l'induction jusqu'd
la naissance, en utilisant sur l'avant-bras la technique de Tuns-
tall du garot isolant. Les patientes ont dtd rdparties au hasard
pour recevoir soit du thiopentone (4 rag" kg -I) soit de la kdta-
mine (1 mg" kg-t) pour l'induction. Avant l'administration in-
traveineuse de succinylcholine, un garot isolant a dtd appliqud
sur l'avant-bras et gonfl~ d 200 mmHg pour conserver l'activitd
motrice de l'avant-bras. L'EEG a dt~ enregistrd 5 rain avant
l~nduction et pendant l'anesth~sie. L~ncidence des r~ponses
des ordres verbaux a dtd inf~rieure dans le groupe k~tamine
(24%) (SEF90 moyenne 12,0 -t- 3 Hz) comparativement au
groupe thiopentone (52%) (SEFgO moyenne 18,09 + 3). Ces
rdsultats suggbrent qu'une valeur de SEF inf~rieure ,~ 8.6 Hz
suffit pour pr~venir la rdponse aux ordres verbaux.
Awareness is the unintentional regaining of consciousness
during presumed general anaesthesia and it has been rec-
ognized increasingly since the routine use of neuromus-
cular relaxants, i It can be a terrifying experience, rep-
resenting a human and a medico-legal problem. 2-4
Although the brain is the major target organ of an-
aesthetic drugs, ironically this organ system has been
largely ignored in routine monitoring during anaesthesia
and major surgery. The complexity of the equipment and
the difficulty of reading an unprocessed EEG tracing has
limited its use in the operating room. 5
During the past decade, as a result of advances in
computer-processed EEG analyzers, routine EEG mon-
itoring in the operation room is becoming more common;
the information obtained from the EEG being converted
into a simplified form for the clinician.
The most widely used technique to process the EEG
is power-spectrum analysis, which uses a computer to
perform a Fourier transformation. In an attempt to sim-
plify the description of the EEG, a number of descriptive
variables has been derived from the power-spectrum anal-
ysis and they have been suggested as indicators of an-
aesthetic depth. 6,7 One of the most commonly used is
CAN J ANAESTH 1995 / 42:5 / pp 377-81
CANADIAN JOURNAL OF ANAESTHESIA
the spectral edge frequency (SEF) which represents the
highest frequency of the EEG. s The SEF90 is the fre-
quency below which 90% of the total brain power is con-
Awareness during Caesarean section is a recognized
problem with general anaesthesia.'
Tunstall's technique for monitoring awareness during
Caesarean section relies on isolation of the forearm from
the effects of the neuromuscular blockade by occlusion
of the circulation with a pneumatic tourniquet inflated
before injection of the muscle relaxant. Movement of the
hand in response to message is then monitored. 3
This study was designed to investigate the correlation
between responsiveness to verbal commands and simul-
taneously recorded SEF90 values.
Fifty women (ASA class I) 19-39 yr of age, undergoing
elective Caesarean section with general anaesthesia were
assigned by a randomized code to receive either thio-
pentone or ketamine for induction of anaesthesia.
The investigation was approved by the Institution
Ethics Committee and all patients gave informed consent.
Preoperatively, all patients received 30 ml, 0.3 M so-
dium citrate po. In the operating room the patients were
positioned on the operating table in a left lateral tilt of
approximately 15 ~ Oxygen, 100%, was administered by
mask for five minutes. Three minutes before anaesthetic
induction d-tubocurarine (3 mg)/v was given.
Anaesthesia was induced according to the drug assign-
ment with either thiopentone (4 mg. kg -I) or ketamine
(1 mg. kg-~). When the patient lost consciousness, suc-
cinylcholine (1.5 mg. kg -I) was given. Cricoid pressure
was maintained until after the trachea has been intubated.
Following intubation the lungs were ventilated with 50%
nitrous oxide and 50% oxygen. End-tidal concentration
of halothane was maintained at 0.5 MAC, measured by
an infrared spectrophotometer (Datex-Ultima).
To assess intraoperative awareness a tourniquet was
applied to the free arm, and was inflated to 200 mmHg
before the administration of the succinylcholine to main-
tain motor function of one arm.
During the induction-to-delivery interval (I-D interval)
the anaesthetist instructed the patient via headphones,
once every minute, to raise her free hand.
After delivery, fentanyl, 2-3 ~g. kg -j /v, was admin-
istered, the FIO2 was reduced to 0.3, halothane was
stopped, and the tourniquet was deflated. Before deflating
the tourniquet, maintenance of motor function in the iso-
lated forearm was confu-rned by train-of-four using a pe-
ripheral nerve stimulator.
On arrival in the recovery room, all patients in the
ketamine group received diazepam, 0.15 mg. kg -j/v, to
prevent postoperative dreams and other unpleasant emer-
The patients were studied using the Cerebrotrac
2500*EEG monitor to monitor and correlate their re-
sponsiveness to verbal commands with the behaviour of
SEF90. The Cerebrotrac receives two channels of real-
time EEG signals from the patient and converts the EEG
waveform from the time domain to the frequency domain
using the Fast Fourier Transform (FFT). The Cerebrotrac
2500 performs FFT using two-second EEG epochs. The
band-width for recording was 0-30 Hz.
Five electrodes were placed in a bilateral frontomastoid
array and electrode impedance was <20000 ohms. The
EEG recordings were started five minutes before induc-
tion, recorded throughout anaesthesia and saved on disk.
The patients were interviewed approximately 24 hr
after surgery to assess recall.
Apgar scores were assigned by a paediatrician.
The time course EEG changes between the groups and
the values preceding and immediately following each in-
cidence of responsiveness were compared with repeated
measurements ANOVA followed by Newman-Keuls mul-
tiple comparisons procedure. Student's t test was used
to compare the averages of the SEF90 values between
the thiopentone and ketamine groups separately for those
patients with responsiveness to verbal commands and for
those without. Fisher's exact test was used to evaluate
the SEF90 value indicative of no awareness. A P value
<0.05 was considered to be statistically significant. All
values are given as the mean + SD.
The groups were similar with respect to age, weight and
induction to delivery interval (Table I).
In the thiopentone group, 13 of 25 patients (52%),
moved their hands in response to the anaesthetist's in-
struction, before delivery. The pre-delivery movements in
response to command occurred at an average SEF90
value of 18.09 + 3.1 Hz. In the ketamine group, five
of 25 patients (24%) moved their hands in response to
command before delivery at an average SEF90 of 12.0
+ 3 Hz. Figure 1 depicts the SEF90 changes.as a function
of time in patients with responsiveness to verbal com-
mand, in both groups (thiopentone and ketamine).
Twelve patients in the thiopentone group did not move
their hands in response to instruction. The average SEF90
in these patients was 13.5 Hz. Twenty patients in the
ketamine group did not move their hands in response
*SRD Shorashim Medical Ltd. Shorashim, DN. Misgav
Gaitini et al.: EEG IN CAESAREAN SECTION
TABLE I Characteristics of the study groups
Induction to delivery interval (rain)
26.2 :t: 5
72.8 4- 7
9.4 :t: l
25.5 4- 5
72.0 -t- 9
9.0 q- l
Values are mean :1: SD.
to the anaesthetist's instruction. The average SEF90 in
these patients was 9.4 + 1 Hz. Figure 2 depicts the SEF90
changes as a function of time in patients without respon-
siveness to verbal command, in both groups (thiopen-
tone and ketamine).
The incidence of awareness was lower in those patients
in whom anaesthesia was induced with ketamine as com-
pared with thiopentone (P = 0.08).
In the thiopentone group there were three spontaneous
movements indicating probable light anaesthesia at an
average SEF90 value of 17.6 + Hz (Figure 3).
Fisher's exact test suggested that SEF90 value of 8.6
Hz was related to unresponsiveness to verbal command
(P = 0.06).
There was a difference between the SEF90 values at
the time of responsiveness to verbal commands (mean
16.5 + 5 Hz) compared with SEF90 values immediately
following each incidence of responsiveness (mean 19.3 -I-
5 Hz; P = 0.003). There was no difference between the
SEF90 values immediately preceding (mean 17.2 ___ 5 Hz)
and at the time of responsiveness to verbal commands
None of the patients recalled anything of the surgery.
With respect to the newborn, no difference could
be shown in the Apgar scores of the two groups
This study used the isolated forearm technique to detect
awareness during elective Caesarean section and exam-
ined the relationship between the EEG (SEF90) and the
occurrence of awareness defined as responsiveness to ver-
bal commands. The results show that if the SEF90 values
recorded during the I-D interval were --<8.6 Hz there
was no response to verbal commands. The SEF90 values
>8.6 Hz at any given moment did not necessarily predict
awareness at that time.
Changes in the SEF have been shown to correlate in
a predictive fashion with serum concentrations of opioids9
and barbiturates. ~~ Rampil and Matteo s demonstrated
that haemodynamic responses to tracheal intubation could
be prevented during a thiopentone-lidocaine-fentanyl-
FIGURE 1 Movements in response to command: The time course
changes of the SEF90.
FIGURE 2 No movements: The time course changes of the SEF90.
FIGURE 3 Thiopentone group: Spontaneous movements.
380 CANADIAN JOURNAL OF ANAESTHESIA
P : 0.003
A - SEFg0 values immediately preceding
responsiveness to verbal command
B - SEFg0 values at the time of responsiveness
to verbal command
C - SEFg0 values immediately following
responsiveness to verbal command
A B C
FIGURE 4 Comparison between the SEF90 values immediately preceding, at and immediately following responsiveness to verbal commands.
TABLE II Apgar scores
(n = 2s)
(n = 25)
Apgar score >7 (n)
- 5 min
droperidol anaesthetic sequence if the SEF was < 14 Hz.
In the Stoeker and Schwilden series it was shown that
median values < 5 Hz reduced the likelihood of intraop-
erative awareness, ii In this study SEF90 values <8.6 Hz
were associated with lack of responsiveness to verbal
commands during the I-D interval and therefore pre-
dictive of non-awareness. Comparing our study with these
studies revealed that there was a correlation between the
computerized EEG descriptors (SEE median power fre-
quency) and awareness. These findings require further
investigation in larger series of comparative studies of
frontalis muscle activity, EEG processed in various ways,
auditory-evoked potentials and isolated forearm tech-
nique in order to resolve some of the questions concerning
Some authors believe that SEF has limitations, being
a poor indicator of anaesthetic depth. 12:3 Arden and Hol-
Icy 12 found that, although the change in SEF mirrors
at an increasing anaesthetic concentration, there is a con-
siderable lag in the recovery of SEF when anaesthetic
levels are falling. White and Boyle ~3 found no consistent
relationship between SEF and haemodynamic responsive-
ness to surgical stimulation in patients undergoing general
anaesthesia with propofol and nitrous oxide.
There are still no guidelines for EEG use during routine
general anaesthesia but the simplicity of this technique
may in the future permit the routine use of intraoperative
This study also confirrns that intraoperative maternal
awareness during Caesarean section is lower after induc-
tion of anaesthesia with ketamine than after thiopen-
tone. 14-16 Ketamine, in doses of I mg .kg -I may be used
instead of thiopentone as an induction agent in Caesarean
section, with no difference in neonatal outcome compared
with thiopentone. 14 However, ketamine can produce un-
pleasant psychological reactions during awakening from
anaesthesia, called emergence phenomena. ~7 The benzo-
diazepines seem to be the most effective drugs in pre-
venting these phenomena. 17 In this study, after arrival
in the recovery room, patients received diazepam which
influenced recall. The standard method of induction of
anaesthesia with thiopentone offers the advantage of less
abnormalities of mental status immediately after anaes-
In conclusion, we found that SEF90 values of --<8.6 Hz
were sufficient to avoid responsiveness to verbal com-
mands. We also found that ketamine induction for
Caesarean section in doses of 1 mg-kg -I had a lower
incidence of intraoperative awareness than with thiopen-
tone induction (4 mg. kg-]). This study demonstrates
continuous monitoring of the EEG during general anaes-
thesia, especially during Caesarean section, using com-
puterized analysis and a graphic display appears to pro-
vide a sensitive method of recognizing awareness.
Gaitini et al.: EEG IN CAESAREAN SECTION 381
1 Bogod DG, Orton JK, Yau HM, Oh TE. Detecting
awareness during general anaesthetic Caesarean section.
An evaluation of two methods. Anaesthesia 1990; 45:
2 Crawford JS. Awareness during operative obstetrics under
general anaesthesia. Br J Anaesth i971; 43: 179-82.
3 Tunstall ME. Detecting wakefulness during general anaes-
thesia for Caesarean section. BMJ 1977; 1: 1321.
4 Wilson J, Turner DJ. Awareness during Caesarean section
under general anaesthesia. BMJ 1969; 1: 281-3.
5 Mori K. The EEG and awareness during anaesthesia (Edi-
torial). Anaesthesia 1987; 42:1153-5.
6 Levy WJ. Power spectrum correlates of changes in con-
sciousness during anesthetic induction with enflurane.
Anesthesiology 1986; 64: 688-93.
7 Long CI4(, Shah NK, Loughlin C, Spydell J, Bedford RF.
A comparison of EEG determinants of near-awakening
from isoflurane and fentanyl anesthesia. Anesth Analg
1989; 69: 169-73.
8 Rampil I J, Matteo RS. Changes in EEG spectral edge fre-
quency correlate with the hemodynamic response to laryn-
goscopy and intubation. Anesthesiology 1987; 67: 139-42.
9 Scott JC, Ponganis KV, Stanski DR. EEG quantitation of
narcotic effect: the comparative pharmacodynarnics of fen-
tanyl and alfentanyL Anesthesiology 1985; 62: 234-41.
10 Hudson R J, Stanski DR, Saidman I_d,, Meathe E. A
model for studying depth of anesthesia and acute tolerance
to thiopental. Anesthesiology 1983; 59: 301-8.
I I Stoekel 1"1, Schwilden 1-1. Median EEG frequency. In:
Rosen M, Lunn JN (Eds). Consciousness, Awareness and
Pain in General Anaesthesia. London: Butterworths 1987:
12 Arden JR, Holley FO,, Stanski DR. Increased sensitivity to
etomidate in the elderly: initial distribution versus altered
brain response. Anesthesiology 1986; 65: 19-27.
13 White PF, Boyle WA. Relationship between bemodynamic
and electroencephalographic changes during general anes-
thesia. Anesth Analg 1988; 68: 177-81.
14 Schultetus RR, Hill CR, Dharamraj CM, Banner TE, Bet-
man LS. Wakefulness during cesarean section after anes-
thetic induction with ketamine, thiopental, or ketamine and
thiopental combined. Anesth Analg 1986; 65: 723-8.
15 Baraka A, Louis E Dalleh R. Maternal awareness and
neonatal outcome after ketarnine induction of anaesthesia
for Caesarean section. Can J Anaesth 1990; 37: 641-4.
16 Baraka A, Louis F, Noueihid R, Diab M, Dabbous A,
Sibai A. Awareness following different techniques of gen-
eral anaesthesia for Caesarean section. Br J Anaesth 1989;
17 Reich DI_o Silvay G. Ketamine: an update on the first
twenty-five years of clinical experience. Can J Anaesth
1989; 36: 186-97.
18 Moretti R J, Hassan SZ, Goodman LI, Meltzer HY.
Comparison of ketamine and thiopenthal in healthy volun-
teers: effects on mental status, mood, and personality.
Anesth Analg 1984; 63: 1087-96.