Full text of DOI:10.1007/BF03017323

GENERAL ANESTHESIA
361
Midlatency auditory evoked potentials do not
allow the prediction of recovery from general
anesthesia with isoflurane
[Les potentiels évoqués auditifs de milatence ne permettent pas de prédire la
récupération après une anesthésie générale à l’isoflurane]
Ingrid Rundshagen MD,* Kai Schnabel,† Jochen Schulte am Esch MD‡
Purpose: To investigate midlatency auditory evoked potentials
(MLAEP) waveforms during recovery from anesthesia. The hypoth-
esis was that MLAEP are sensitive variables to discriminate between
states of consciousness and unconsciousness during emergence
from anesthesia.
Objectif : Examiner les tracés des potentiels évoqués auditifs de
milantence (PEAML) pendant la récupération de l’anesthésie.
L’hypothèse était que les PEAML sont des variables sensibles qui per-
mettent de distinguer les états de conscience et d’inconscience pen-
dant le réveil postanesthésie.
Methods: MLAEP were recorded in the awake state and during the
wake-up phase from isoflurane anesthesia in 22 female patients
undergoing ophthalmologic surgery. During emergence from anes-
thesia the changes in latency and amplitude of MLAEP components
Na, Pa and Nb were compared with the awake level. The next day
the patients were asked for explicit memory for the recovery period.
Méthode : Les PEAML ont été enregistrés à l’état vigil et pendant le
réveil postanesthésie à l’isoflurane chez 22 femmes en chirurgie oph-
talmologique. Pendant le retour à la conscience, les modifications du
temps de latence et de l’amplitude des composantes Na, Pa et Nb
des PEAML ont été comparées avec celles de l’état vigil. Le jour sui-
vant, on a interrogé les patientes sur le souvenir qu’elles avaient de la
période de récupération.
Results: In 72% of the patients the MLAEP waveforms were com-
pletely suppressed during isoflurane anesthesia. When the patients
responded and opened their eyes spontaneously 38 ± 12 min after
anesthesia, the latencies of Na (18.3 ± 1.2 vs 17.6 ± 1.3; P =
0.013) and Nb (47.4 vs 7.1 vs 44.7 ± 7.8; P = 0.048) remained
prolonged compared with awake values. In contrast, the ampli-
tudes NaPa and PaNb had regained baseline level. Nine patients
had explicit memory for the immediate recovery period. However,
there was no difference for any MLAEP component between
patients with and without memory at any time.
Résultats : Chez 72 % des patientes, les tracés des PEAML ont été
complètement supprimés pendant l’anesthésie à l’isoflurane. Quand
les patientes ont repris conscience et ont ouvert les yeux spontané-
ment, 38 ± 12 min après l’anesthésie, les temps de latence de Na
(18,3 ± 1,2 vs 17,6 ± 1,3 ; P = 0,013) et de Nb (47,4 vs 7,1 vs
44,7 ± 7,8 ; P = 0,048) sont demeurés longs comparés aux valeurs
vigiles. Par ailleurs, les amplitudes NaPa et PaNb ont repris les valeurs
de base. Neuf patientes ont eu un souvenir explicite de la période de
récupération immédiate. Cependant, il n’y a pas eu de différence
interpatientes concernant l’une ou l’autre composante des PEAML,
sans égard à la présence ou non de souvenir.
Conclusions: The persistent changes of MLAEP latency compo-
nents Na and Nb indicated impaired auditory signal processing 38
min after isoflurane anesthesia. There was a marked intra- and inter-
individual variability during reversal of the anesthetic induced MLAEP
changes. This limits the prediction of recovery of consciousness in the
individual patient during emergence from anesthesia.
Conclusion : Les changements persistants des composantes Na et
Nb des temps de latence des PEAML ont indiqué une modification du
traitement du signal auditif 38 min après l’anesthésie à l’isoflurane. Il
y a eu une variabilité intra-individuelle et interindividuelle marquée
pendant le renversement des changements des PEAML induits par
l’anesthésie. Ce résultat limite les possibilités de prédire le retour à la
conscience pendant le réveil postanesthésie.
From the Department of Anesthesiology,* University Hospital Charité, Humboldt University of Berlin, Campus Charité Mitte, Berlin,
Germany; the Department of Psychology,† University of Michigan, Ann Arbor, USA; and the Department of Anesthesiology,‡ University-
Hospital Eppendorf, Hamburg, Germany.
Address correspondence to: Dr. Ingrid Rundshagen, Department of Anesthesiology, University Hospital Charité, Campus Charité Mitte,
Schumannstr. 20/21, D-10117 Berlin, Germany. Phone: 49-30-450 531 012; Fax: 49-30-450 531 911;
E-mail: ingrid.rundshagen@charite.de
Accepted for publication July 3, 2001.
Revision accepted January 16, 2002.
CAN J ANESTH 2002 / 49: 4 / pp 361–368

362
CANADIAN JOURNAL OF ANESTHESIA
OST of the anesthetics used during
general anesthesia alter the functional
state of the central nervous system
(CNS). However, in routine clinical
Methods
After obtaining approval from the local Ethics
Committee and written informed consent, 22
patients, scheduled for elective ophthalmological
surgery, were included in this prospective study.
Patients with neurological diseases or hearing abnor-
malities were excluded. The patients did not take any
centrally acting drugs.
M
practice the CNS is not monitored. The anesthesiolo-
gist usually relies on autonomic variables such as
blood pressure, heart rate, pupillary reactivity, lacrima-
tion and/or muscle responses to estimate depth and
quality of anesthesia. For clinical routine and for sci-
entific questions there is a need to evaluate variables,
which might be useful to quantify and to objectify dif-
ferent functional states of the CNS during general
anesthesia. Quantification of the pharmacodynamic
characteristics of anesthetic agents is necessary to
avoid over- or undermedication and potential risks to
the patients.1,2
Several studies have been carried out to define the
anesthetic action of drugs on the CNS using different
variables of the spontaneous or evoked electroen-
cephalogram (EEG).3,4 While the EEG provides infor-
mation on the actual functional state of the brain, the
evoked responses indicate stimulus induced changes of
the neural pathway from the receptor at the peripher-
al site all the way to the specific projection areas in the
brain. It has been shown that the different compo-
nents of the evoked responses are generated in specif-
ic areas of the brain. However, the exact location of
some components is still debated.4 Several recent
studies indicate that the midlatency auditory evoked
potentials (MLAEP) are a promising tool to quantify
the hypnotic component during anesthesia.5
In 1989, Thornton and colleagues could demon-
strate in seven patients, anesthetized with thiopentone
and nitrous oxide after premedication with morphine,
that the midlatency AEP component Nb could differ-
entiate, with a 100% specificity, whether patients would
follow a command using the isolated forearm tech-
nique.6 They suggested that a threshold of 44.5 msec
for the Nb latency could be used as a possible indicator
for intra-operative awareness. In addition to some other
investigations supporting this finding, Davies et al.
could demonstrate in patients, sedated with propofol
during regional anesthesia, that the changes of the Nb
latency were the most sensitive variable to differentiate
between loss and return of consciousness.7
The present study investigated MLAEP during
emergence from isoflurane anesthesia in comparison
to the awake state. We tested the hypothesis that the
MLAEP latencies correlate with a decreasing level of
sedation and discriminate states of consciousness and
unconsciousness during recovery from anesthesia.
Anesthesia and recovery
All patients received midazolam 7.5 mg orally as a pre-
medication 45 min prior to anesthesia. Anesthesia was
induced with propofol 2 mg·kg–1, sufentanil 0.4
µg·kg–1 and vecuronium 0.1 mg·kg–1. Endotracheal
intubation was performed and anesthesia maintained
with isoflurane 0.6% (end-tidal) in nitrous oxide and
oxygen (30%). An additional bolus of sufentanil 0.25
µg·kg–1 was administered if necessary. Approximately
20 min before the end of surgery nitrous oxide was
stopped and anesthesia maintained with 1.2% isoflu-
rane. After the end of surgery and after the MLAEP
had been recorded under anesthesia, the administra-
tion of isoflurane was stopped, 100% oxygen was
delivered (fresh gas flow 3 L·min–1), and the patients
allowed to recover. After extubation patients received
oxygen via a face mask.
AEP-data acquisition
The day before surgery patients were familiarized with
the procedure of MLAEP recording in order to obtain
awake baseline values. MLAEP recording was per-
formed in a standardized manner with an Evomatic
4000® system (Dantec, Copenhagen, Denmark). The
individual hearing threshold was defined. The stimulus
intensity was about 70 dB above the individual thresh-
old and kept constant throughout the whole study.
Baseline recording was replicated twice. An 8 Hz stim-
ulus frequency was applied using binaural randomized
click stimulation via headphone. The MLAEP wave-
forms were recorded on two amplifier channels using
zinc/lead cup electrodes placed over Cz (international
10–20 system) and bilateral mastoid. Electrode imped-
ances were kept below 5 k. A poststimulus period of 90
msec was analyzed. During anesthesia and recovery a
period of 180 msec was recorded. The brainstem com-
ponent V served as a reference that the auditory stimu-
lus was delivered correctly. Moreover, the device
included an automatic artefact rejection mode. Using a
bandpass 0.02 to 2 k Hz for each response 2000 stim-
uli were averaged and stored on disk for later analysis.
Latencies and peak-to-peak amplitudes were obtained
using a software package (EvoPC®, Müller, Hamburg,
Germany). The following peak latencies were obtained:

Rundshagen et al.: MLAEP DURING RECOVERY FROM ANESTHESIA
363
brainstem component V, two negative components
(Na, Nb) and one positive peak in between (Pa). The
peak-to-peak amplitudes NaPa and PaNb were mea-
sured. When patients presented a postauricular reflex
(PAR), the latency and amplitude were determined
(LatPAR, AmpPAR).
thesia, only measurements at AWAKE, Pre-EXT, Post-
EXT and RECOVERY were included in the overall
multivariate analysis resulting in a full two-factor-with-
in design (3 latencies × 4 time points). For descriptive
purposes, a posteriori paired t tests including all mea-
surements were performed comparing each single
MLAEP component at the different measurement
points. When normal distribution was violated, non-
parametric tests were used (Friedman test).
Correlation coefficients were tested for the MLAEP
components, the vital and clinical parameters. P <
0.05 was adopted to determine significance.
Prediction probability PK according to the method of
Smith was calculated.9
Measurements
MLAEP recording started on the day before surgery
in order to obtain baseline data in the awake state
(AWAKE). Then, the MLAEP recording was per-
formed on the day of surgery during steady state
isoflurane anesthesia after the end of surgery
(ANESTH), and every five to ten minutes during
recovery from anesthesia. In accordance with the clin-
ical findings, Pre-EXT was defined as the last MLAEP
recording before extubation and Post-EXT as the first
recording following extubation, when the patients
opened their eyes on command. When the patients
opened their eyes spontaneously, they were asked to
verbally define a shown object (a small red booklet
[about 12 × 15 cm in size], which was opened and
closed in front of the patient). The final MLAEP
recording was performed (RECOVERY), when they
correctly named the object shown.
Results
AEP-parameter
The mean stimulation intensity was 105
5 dB.
During the awake state the brainstem component V
was identified in all patients. While 13 patients had a
small or no postauricular reflex (PAR < 2 µV), in seven
patients it was larger. Figure 1 depicts the original
MLAEP recordings of patients at AWAKE indicating
the typical midlatency components. Following the
postauricular reflex, two negative and one positive
deflections of MLAEP components were clearly iden-
tified. However, there was one patient who did not
present the component Nb. An example of the
Clinical variables
Heart rate (beats·min–1), non-invasive mean arterial
blood pressure (mmHg), percutaneous oxygen satura-
tion (%) and body temperature were recorded in parallel
to the MLAEP recording. During and after anesthesia
end tidal carbon dioxide (petCO2) and isoflurane
(petISO) concentrations were measured (via the endotra-
cheal tube; after extubation, via a tightly held facial
mask). In addition, the total duration of anesthesia, the
time of the extubation and of RECOVERY were record-
ed. The day after the operation the patients were asked
(structured interview), whether they recalled anything
during anesthesia or the wake-up phase.
Statistics
Latencies and peak-to-peak amplitudes NaPa and
PaNb were calculated. The sample distributions of
data were checked with Kolmogorov-Smirnov tests.
The MLAEP data were analyzed using multivariate
analysis of variance (Hotellings T-square; repeated
measurement design) provided the data were normal-
ly distributed. This statistical approach is most appro-
priate for studies involving repeated measurements
with the same individuals. However it is applicable
only to a data set in which all measurements have been
made at all time points.8 Since in some patients the
MLAEP waves completely disappeared during anes-
FIGURE 1 The original auditory evoked potential (AEP) trac-
ings of three different patients in the awake state are shown. The
postauricular reflex, the brain stem component V and the three
midlatency components are indicated. (a) Patient SK, 25 yr, male;
(b) patient FH, 42 yr, male; (c) patient SU, 33 yr, male. Vertical
lines indicate the Nb latencies to show the inter-individual variabil-
ity. Montage: Cz vs linked mastoid, binaural click stimulation.

364
CANADIAN JOURNAL OF ANESTHESIA
FIGURE 3 The individual responses of the midlatency auditory
evoked potential (MLAEP) latency component Nb are shown dur-
ing the different measurements. AWAKE = the day before anesthe-
sia; ANESTH = after the end of surgery at 1 MAC isoflurane;
PRE-EXT = before extubation; POST-EXT = after extubation,
when the patients opened their eyes at command; RECOVERY =
when the patients opened their eyes spontaneously.
FIGURE 2 The original auditory evoked potential (AEP) trac-
ings of patient FH, 42 yr, male, are shown during the different
measurements. AWAKE = the day before anesthesia; ANESTH =
after the end of surgery at 1 MAC isoflurane; PRE-EXT = before
extubation; POST-EXT = after extubation, when the patients
opened their eyes at command; RECOVERY = when the patients
opened their eyes spontaneously. Montage: Cz vs linked mastoid,
binaural click stimulation.
command, indicated MLAEP deflections in 20 patients.
The latencies were still significantly prolonged (P < 0.03)
and the amplitudes NaPa (P < 0.001) reduced in com-
parison to AWAKE. The PAR recovered in 19 patients,
but was smaller than 1 µV except in one patient.
Immediately after extubation (Post-EXT), when the
patients opened their eyes on command, the decrease in
latencies and the increase in amplitudes was progressive
in comparison to ANESTH. The latency Na was still
significantly prolonged (P = 0.025), while the compo-
nents Pa and Nb did not differ in latency from the base-
line values. The amplitudes NaPa and PaNb reached
baseline level. In six patients the postauricular reflex was
larger than 2 µV (maximum 8.7 µV).
During the final measurement RECOVERY, when
the patients opened their eyes spontaneously, the laten-
cies Na and Nb were still prolonged (P < 0.05). Eight
patients presented a large postauricular reflex (> 3.5 µV,
maximum 8.9 µV). However, only five of these patients
had shown a PAR larger than 1.5 µV at AWAKE. Figure
3 demonstrates the change of the Nb latencies for all
patients during the different measurements.
MLAEP changes during the various stages of mea-
surements is given in Figure 2.
The Kolmogorov-Smirnoff test supported the
assumption of normal distribution for all MLAEP
latencies. In contrast, the amplitude NaPa was not
normally distributed, therefore the amplitudes were
analyzed by non parametric tests (Friedman test,
Wilcoxon test). The multivariate analysis revealed an
overall significant difference for the latency compo-
nents comparing the different measurements (P =
0.005) The absolute values of the MLAEP latencies
are given in Table I. The Friedman test revealed sig-
nificant changes for the amplitudes NaPa (P < 0.001).
The absolute values are documented in Table II.
However, the data at ANESTH were excluded from
testing, because, in 16 patients, the MLAEP compo-
nents were suppressed completely.
During anesthesia with 1.2% isoflurane the brain-
stem component V was identifiable in all patients, the
PAR was absent except in one patient. The MLAEP
components were completely suppressed in 16
patients. In the other six patients the latencies Na and
Nb were significantly prolonged in comparison to
AWAKE (P # 0.03).
PK values of prediction probability were as follows for
the latency components
Na – 0.68, Pa – 0.64, Nb – 0.71; for the amplitudes
respectively: NaPa – 0.71, PaNb – 0.57.9 Therefore,
with respect to the decreasing anesthetic depth during
recovery, the latency Nb and the amplitude PaNa were
the best predictors.
During emergence the anesthetic-induced MLAEP
changes were in part reversed. The last measurement
before extubation (Pre-EXT), when the patients still tol-
erated the endotracheal tube and did not respond to

Rundshagen et al.: MLAEP DURING RECOVERY FROM ANESTHESIA
365
TABLE I MLAEP-latency components
AWAKE
ANESTH
PRE-EXT
POST-EXT
RECOVERY
Na
Pa
Nb
17.6 1.3
29.8 4.8
44.7 7.8
28.5 5.6*
37.8 7.6
63.5 14.3*
19.0 1.4*
32.9 2.6*
54.9 5.5*
18.5 1.8*
30.3 4.2
48.5 8.5
18.3 1.2*
30.2 4.3
47.4 7.1*
Means SD of midlatency auditory evoked potential (MLAEP) waveforms Na, Pa and Nb in the awake state, during 1 MAC isoflurane
anesthesia and during emergence from anesthesia (pre-extubation, postextubation and 38 12 min after stopping isoflurane). n = 22
patients. *P < 0.05 vs AWAKE by paired t test (see text for more details).
TABLE II MLAEP amplitude components
AWAKE
ANESTH
PRE-EXT
POST-EXT
RECOVERY
NaPa
PaNb
2.4 (11.7)
1.0 (2.5)
0.7 (2.6)
0.6 (0.7)
0.8 (1.5)*
0.9 (1.1)
2.3 (7.7)
1.1 (3.9)
2.7 (10.7)
1.3 (4.6)
Midlatency auditory evoked potential (MLAEP) amplitudes (median and ranges) during 1 MAC isoflurane anesthesia and during emer-
gence from anesthesia (pre-extubation, postextubation and recovery [38 12 min after stopping isoflurane]). n = 22 patients. *P < 0.05 vs
AWAKE by Wilcoxon’s test (see text for more details).
TABLE III Clinical variables
AWAKE
67
ANESTH
71
PRE-EXT
POST-EXT
RECOVERY
Heart rate (beats·min–1)
MAP (mmHg)
SaO2 (%)
PetCO2 (mmHg)
Temperature (rectal; °C)
Pet isoflurane (%)
9
9
65 13
97 11
79 12*
100 10
81 15*
101 10
98
78 9*
101
99
43
9
1
99
36
1
2
99
39
1
3
99
45
1
7
1
4
36.3 0.5
36.1 0.5
1.1 0.02
36.2 0.5
0.3 0.08
36.2 0.5
0.2 0.05
36.2 0.5
0.1 0.04
Means SD of the clinical variables at the awake state, during 1 MAC isoflurane anesthesia and during emergence from anesthesia (pre-
extubation, postextubation and 38 12 min after stopping isoflurane). n = 22 patients; *P < 0.05 vs AWAKE by paired t test; MAP =
mean arterial pressure.
MLAEP and the postauricular reflex
Vital parameter and clinical findings
Data of seven female and 15 male patients (age 38
Correlation was significant between PAR amplitude
and MLAEP amplitude NaPa (P < 0.001) and
between PAR latency and MLAEP latency Na (P #
0.042) at all measurements. While MLAEP latency
Nb was not correlated with PAR latency at any time,
the latency Pa showed a significant correlation at
RECOVERY only (P = 0.003). In order to identify a
difference between MLAEP waves of patients with a
small or a large PAR the patients were divided into
two groups: Group I included patients with a PAR
amplitude < 2 µV (at AWAKE); Group II patients with
a PAR amplitude > 2 µV. With respect to the MLAEP
latencies the multivariate analysis revealed an overall
significant difference between the different measure-
ments for Group I only (P = 0.009, Group II: P =
0.150) The Friedman test revealed significant changes
comparing the different measurements for the ampli-
tudes NaPa for both groups (Group I: P = 0.002;
Group II: P = 0.01).
9
yr; height 1.77 0.10 m; weight 75 13 kg, ASA
I–II) were collected. The mean duration of anesthesia
was 100 28 min, while the surgical procedure took
52 26 min. For induction of anesthesia the patients
received 31 7 µg of sufentanil. Additionally 5 4 µg
sufentanil were administered 71 25 min before the
end of anesthesia. The mean time for extubation was
24 12 min after the end of anesthesia. The average
time between interruption of isoflurane and sponta-
neous eye opening was 38 12 min. Patients did not
require any additional medications either for stabiliz-
ing hemodynamics or against pain or nausea during
the study period.
Table III summarises the significant changes for the
vital signs. The changes in MLAEP variables were sig-
nificantly correlated with the isoflurane concentration
(P # 0.004). The differences in isoflurane concentra-
tion reached statistical significance among all mea-
surements (paired t test: P # 0.004). There was no

366
CANADIAN JOURNAL OF ANESTHESIA
correlation between the sufentanil dose per weight
and total time of anesthesia and any of the MLAEP
variables. The changes in MLAEP latencies were cor-
related with the changes of mean arterial pressure (P <
0.001), while Nb only correlated with heart rate
changes (P = 0.004).
During the interview at the first postoperative day
13 patients did not remember anything of the recov-
ery period, while nine patients indicated memory (two
patients remembered the auditory cliques, seven
patients the shown object). There was no difference in
any of the MLAEP parameters at any time between
patients with and without explicit memory.
large variation might be explained by the different
techniques used for recording the MLAEP. Often the
recording modalities are determined by the equipment
used. However, there are no strict guidelines for record-
ing of MLAEP. Inter-individual variation was also
reported in subsequent studies, when the same record-
ing techniques were used. Therefore, substantial inter-
individual differences have to be taken into account.
With respect to the changes of MLAEP waveforms
induced during isoflurane anesthesia, our findings cor-
roborate results of previous studies, which document-
ed that isoflurane decreases MLAEP amplitudes and
increases latencies.13,14 In 73% of our patients the
MLAEP components disappeared completely during
anesthesia. Similar findings have been documented for
enflurane (1.2%) and sevoflurane (1.5%). However,
for isoflurane it had not been documented precisely
before.15,16 For patients undergoing cardiac surgery
with different anesthetic regimes, Schwender et al.
showed that, in patients with implicit memory, the
waveform pattern of MLAEPs continued during anes-
thesia.17 In all patients with implicit memory, the
latency increase of Pa was less than 12 msec during
anesthesia when auditory information was presented.
In contrast, when the latency increase was greater than
12 msec during anesthesia, no case of implicit memo-
ry was found. They concluded that if MLAEP are pre-
served during general anesthesia, auditory
information may be processed and indicated, when
implicit memory tests are used postoperatively.
However, in the present population we could not
document a correlation between explicit memory per-
formance and any of the MLAEP components during
recovery from anesthesia.
To determine threshold values of MLAEP for pre-
diction of movement during light anesthesia
Schwender et al. investigated patients anesthetized
with isoflurane or propofol in combination with
epidural anesthesia.18 A threshold of 60 msec of the
Nb latency proved to be most predictive of movement
during anesthesia and during recovery from anesthe-
sia. Comparisons with the present study are difficult
since our study was not designed to predict move-
ment. Schwender et al. documented Nb latencies of
about 52 msec when spontaneous movement
occurred. This fits our data almost perfectly, where the
mean Nb latency was about 55 msec before extuba-
tion, when the patients tolerated the endotracheal
tube and did not move, and 49 msec, immediately
after extubation at a time when patients opened their
eyes on command. However, in light of our findings,
it remains questionable whether defining a general
threshold value for individual prediction is a sensible
Discussion
In comparison to the awake state we could demonstrate
profound changes of MLAEP components, which
were, in part, reversed during recovery from anesthesia.
In most patients the MLAEP components were com-
pletely suppressed. In all other patients the latencies
were significantly prolonged. During emergence, the
MLAEP components reappeared. The latencies gradu-
ally decreased and the amplitudes increased in relation
to the decreasing anesthetic isoflurane concentration.
However, when the patients were responsive again, the
latencies Na and Nb were still prolonged, indicating
impairment in processing of the auditory signal.
While the group effect was strongly significant, the
individual MLAEP responses differed markedly,
expressing a large inter- and intra-individual variation.
We could document that at AWAKE and during
RECOVERY the MLAEP variable Na was correlated
with the PAR. However, patients with a large PAR
response during AWAKE did not always present a large
PAR during RECOVERY. Therefore there does not
seem to be a linear recovery profile of the PAR in all
patients. When interpreting our results for the patients
with and without a large PAR, it must be taken into
account that the statistical power was low due to the
small number of patients in the two groups. In order to
elucidate the relationship between PAR and MLAEP
components during light stages of anesthesia a larger
population should be investigated, combining MLAEP
monitoring with electromyography. Different stimula-
tion and recording modalities like lower stimulation
intensity or using only one mastoid electrode could
help to decrease the PAR response during clinical mon-
itoring.10 However, during surgical anesthesia with 1.2
volume % isoflurane the PAR was completely sup-
pressed in all patients except in one.
Data in AWAKE reported here are in line with prior
findings, stating mean values of the Nb latencies rang-
ing from 36 to 50 msec (SD 1–7 msec).11,12 In part, this

Rundshagen et al.: MLAEP DURING RECOVERY FROM ANESTHESIA
367
endeavour, given the substantial inter-individual vari-
ance observed. The low PK values in the present study
indicate that the prediction of recovery in the individ-
ual case is limited by MLAEP components.
References
1 Kress JP, Pohlman AS, O’Connor MF, Hall JB. Daily
interruption of sedative infusions in critically ill patients
undergoing mechanical ventilation. N Engl J Med
2000; 342: 1471–7.
Newton et al. studied the relation between MLAEP
and simple tests of awareness at subanesthetic isoflu-
rane concentrations (0.0, 0.1, 0.2 and 0.4 MAC) in
eight anesthesiologist volunteers.19 They documented
graded changes in MLAEP latencies and amplitudes,
which were significantly correlated to the anesthetic
concentration. The MLAEP variables were related to
the level of response more closely than the end-expi-
ratory gas concentrations. The absolute values of
MLAEP data are comparable in our study, yet the
changes in MLAEP amplitudes we observed were less
sensitive in comparison with those of latencies.
Plourde et al. investigated the auditory steady state
response (ASSR), a sustained electrical response of the
brain to auditory stimuli delivered at fast rates, in ten
volunteers.20 The ASSR was attenuated in a dose-
dependent manner (end-tidal isoflurane concentra-
tion: 0%, 0.26%, 0. 38%, 0.50%). However, two
volunteers had to be excluded in their study due to
myogenic potentials, probably from the temporal
muscle, contaminating the ASSR.
With respect to the question: “which method is most
promising in measuring depth of anesthesia?”, MLAEP
have been compared to the bispectral index (BIS), a
computationally derived parameter of the spontaneous
EEG. Gajraj et al. documented that the AEP index was
best at distinguishing the transition from unconscious-
ness to consciousness during sedation with propofol.21
In contrast, in a volunteer study the BIS correlated bet-
ter with sedation scores than the MLAEP.22
The present study documented that anesthetic
induced changes of MLAEP were, in part, reversed
during emergence from isoflurane anesthesia, indicat-
ing impairment of auditory signal processing in the
CNS. In contrast to the distinct group effects, predic-
tions in individual patients were limited and it was not
possible to predict arousal based on population
derived values. New methods for analyzing MLAEP
during anesthesia have been suggested recently. The
AEP-index, derived from the raw AEP signal, or on-
line fast extraction of AEP waveforms (markedly
reducing the number of sweeps for averaging), have
been advocated to be superior to conventional visual
analysis.23,24 Further studies are needed to investigate
whether these new methods may overcome the limita-
tions in MLAEP interpretation demonstrated in the
present study.
2 Sandin RH, Enlund G, Samuelsson P, Lennmarken C.
Awareness during anaesthesia: a prospective case study.
Lancet 2000; 355: 707–11.
3 Schneider G, Sebel PS. Monitoring depth of anaesthesia.
Eur J Anaesthesiol 1997; 14(Suppl. 15): 21–8.
4 Thornton C, Sharpe RM. Evoked responses in anaesthe-
sia. Br J Anaesth 1998; 81: 771–81.
5 Pockett S. Anesthesia and the electrophysiology of audi-
tory consciousness. Conscious Cogn 1999; 8: 45–61.
6 Thornton C, Barrowcliffe MP, Konieczko KM, et al. The
auditory evoked response as an indicator of awareness.
Br J Anaesth 1989; 63: 113–5.
7 Davies FW, Mantzaridis H, Kenny GNC, Fisher AC.
Middle latency auditory evoked potentials during
repeated transitions from consciousness to uncon-
sciousness. Anaesthesia 1996; 51: 107–13.
8 Winer BJ. Multifactor experiments having repeated
measures on the same elements. In: Winer BJ (Ed.).
Statistical Principles in Experimental Design, 2nd ed.
New York: McGraw Hill Book Company, 1971:
514–603.
9 Smith WD, Dutton RC, Smith NT. Measuring the per-
formance of anesthetic depth indicators.
Anesthesiology 1996; 84: 38–51.
10 Jones JG, Aggarwal S. Monitoring the depth of anaes-
thesia. In: Ghoneim MM (Ed.). Awareness During
Anesthesia. Oxford: Butterworth Heinemann, 2001:
69–91.
11 Özdamar Ö, Kraus N. Auditory middle-latency
responses in humans. Audiology 1983; 22: 34–49.
12 Picton TW, Hillyard SA, Krausz HI, Galambos R.
Human auditory evoked potentials. I: Evaluation of
components. Electroencephalogr Clin Neurophysiol
1974; 36: 179–90.
13 Newton DEF, Thornton C, Creagh-Barry P, Dore CJ.
Early cortical auditory evoked response in anaesthesia:
comparison of the effects of nitrous oxide and isoflu-
rane. Br J Anaesth 1989; 62: 61–5.
14 Heneghan CPH, Thornton C, Navaratnarajah M, Jones
JG. Effect of isoflurane on the auditory evoked
response in man. Br J Anaesth 1987; 59: 277–82.
15 Thornton C, Catley DM, Jordan C, Lehane JR, Royston
D, Jones JG. Enflurane anaesthesia causes graded
changes in the brainstem and early cortical auditory
evoked response in man. Br J Anaesth 1983; 55:
479–6.
16 Tatsumi K, Hirai K, Furuya H, Okuda T. Effects of
sevoflurane on the middle latency auditory evoked

368
CANADIAN JOURNAL OF ANESTHESIA
response and the electroencephalographic power spec-
trum. Anesth Analg 1995; 80: 940–3.
17 Schwender D, Kaiser A, Klasing S, Peter K, Pöppel E.
Midlatency auditory evoked potentials and explicit and
implicit memory in patients undergoing cardiac
surgery. Anesthesiology 1994; 80: 493–501.
18 Schwender D, Daunderer M, Mulzer S, Klasing S,
Finsterer U, Peter K. Midlatency auditory evoked
potentials predict movements during anesthesia with
isoflurane or propofol. Anesth Analg 1997; 85:
164–73.
19 Newton DEF, Thornton C, Konieczko KM, et al.
Auditory evoked response and awareness: a study in
volunteers at sub-MAC concentrations of isoflurane. Br
J Anaesth 1992; 69: 122–9.
20 Plourde G, Villemure C, Fiset P, Bonhomme V, Backman
SB. Effect of isoflurane on the auditory steady- state
response and on consciousness in human volunteers.
Anesthesiology 1998; 89: 844–51.
21 Gajraj RJ, Doi M, Mantzaridis H, Kenny GNC.
Analysis of the EEG bispectrum, auditory evoked
potentials and the EEG power spectrum during repeat-
ed transitions from consciousness to unconsciousness.
Br J Anaesth 1998; 80: 46–52.
22 Iselin-Chaves IA, El Moalem HE, Gan TJ, Jinsberg B,
Glass PSA. Changes in the auditory evoked potentials
and the bispectral index following propofol or propofol
and alfentanil. Anesthesiology 2000; 92: 1300–10.
23 Mantzaridis H, Kenny GNC. Auditory evoked poten-
tial index: a quantitative measure of changes in audito-
ry evoked potentials during general anaesthesia.
Anaesthesia 1997; 52: 1030–6.
24 Urhonen E, Jensen EW, Lund J. Changes in rapidly
extracted auditory evoked potentials during tracheal
intubation. Acta Anaesthesiol Scand 2000; 44: 743–8.