During general anesthesia including a neuromuscular blocking drug (NMBD) the incidence of explicit recall is more frequent than during general anesthesia without a NMBD, and in patients receiving a NMBD the posttraumatic consequences of explicit recall seem to be more pronounced (1). Bispectral index (BIS), a processed derivative of the human electroencephalogram (EEG), is approved for monitoring the level of hypnosis during anesthesia (2). BIS has been associated with reduced incidence of explicit recall by 80% during general surgery (3) and during high-risk procedures, such as rigid bronchoscopy, cardiac surgery, and cesarean delivery (4). Nevertheless, the BIS technology is based on an empirically derived algorithm, and sources of bias have been identified (5). Among those ambiguities is whether the reliability of BIS as a measure of hypnosis is adversely affected by variations in frontal electromyogram (EMG) activity. In this context, there are conflicting data in the literature concerning the influence of NMBDs and the resultant changes of the EMG signal on the BIS (6–12). It has been suggested that NMBDs may affect the level of hypnosis by functional deafferentation, i.e., by suppressing muscle afferents to the spinal cord and thereby reduce arousing signaling to the brain (13). The afferentation theory is supported by reduced halothane requirement in humans when pancuronium was administered (14) and by studies in which central neuraxial blockades (i.e., epidural and spinal blockades), in the absence of noxious stimulation, have been shown to reduce the need for hypnotics (15,16). Notably, the conflicting results in human studies assessing the effect from neuromuscular block (NMB) on anesthetic depth have been done in the absence of noxious stimuli (6–12).
We hypothesize that a NMB affects BIS both at rest and during noxious stimulation and that such effects are not related to NMB-induced changes in the EMG. To confirm this, BIS was analyzed in parallel with directly measured EMG (EMGdirect), (Nervus®, Cephalon A/S, Nörresundby, Denmark) and another hypnosis monitoring technology, not primarily based on EEG. Thus, midlatency auditory-evoked potential (MLAEP) index (AAI™) known not to interfere with BIS recording, was simultaneously obtained in all patients (17–19).
The objectives of this study were to investigate the following: 1) the effect of a NMB on BIS, 2) if an interference with BIS could be attributed to changes in frontal EMG activity, 3) whether such effects are seen also during noxious stimulation.
Twenty-five patients (18–65 yr, classified as American Society of Anesthesiologists physical status 1–2), scheduled for elective surgery during general anesthesia, were included in this prospective randomized study after giving written informed consent. The protocol was approved by the local ethical committee at Linköping University. Patients with a body mass index >30, known neurological or neuromuscular disorders, impaired hearing during normal conversation or using hearing aids, or with medications known to interfere with neuromuscular function or the central nervous system, were not included. No premedication was given.
Anesthesia Induction and Maintenance
Anesthesia was induced by inhalation of 8% sevoflurane for 2 min using a fresh gas flow of >8 L/min followed by 4% for 3 min. Tracheal intubation was performed without a NMBD 2 min after an IV bolus dose of remifentanil 1 μg/kg. Anesthesia was maintained with sevoflurane in oxygen and air (fraction of inspired oxygen was 50%). All patients were mechanically ventilated at respiratory rates of 9–14 breaths/min and tidal volumes were adjusted to maintain end-tidal concentration of CO2 (ETco2) = 4.3–5.4 kPa.
Monitoring and Data Acquisition
Standard monitoring including electrocardiogram, oxygen saturation by pulse oximetry, noninvasive arterial blood pressure, and end-tidal gas concentrations [ETo2, ETco2, and ETsevoflurane (ETSEV)] were obtained using a Datex Ohmeda ADU 98/S5 monitor (Datex Ohmeda, Helsinki, Finland). All data, as well as BIS and AAI parameters, were downloaded into laptop computers for subsequent analysis. BIS, AAI, and standard parameters were downloaded every 5 s. Noninvasive arterial blood pressure was measured every 5 min.
The NMB was assessed by isometric mechanomyography of the adductor pollicis muscle using a Biometer Myograph 2000® neuromuscular transmission analyzer (Organon Teknika, Turnhout, Belgium) (20).
The evoked train-of-four (TOF) response (2 Hz for 1.5 s every 11.5 s) was continuously recorded after supramaximal ulnar nerve stimulation (300 μs2 impulses) by a Myotest nerve stimulator (Biometer, Odense, Denmark) connected to standard surface electrodes at the wrist. After waiting for a stable twitch response during 15–20 min of 1 Hz ulnar nerve stimulation, a TOF stimulus pattern was initiated and the recorder calibrated. A continuous preload of 200–350 g was applied to the thumb. Skin temperature above the contracting adductor pollicis muscle was monitored using a surface probe (Datex, Helsinki, Finland) and was maintained above 32°C using a warming blanket (20).
After control recordings, rocuronium [30 mg Esmeron® (Organon) dissolved in 50 mL normal saline; i.e., 0.6 mg/mL] was administered as a continuous IV infusion, using a motor syringe, IVAC P4000 (Cardinal Health, Sollentuna, Sweden), and in a randomly determined order adjusted until 50% or 95% depression of the first twitch (T1) in the TOF. Each degree of NMB was kept at steady state for 10 min and T1 in the TOF response was compared with baseline (T1 = 100%). All patients received 2.5 mg neostigmine and 0.5 mg glycopyrrolate (1 mL Robinul-neostigmine®, Meda AB, Solna, Sweden), IV for reversal of the NMB.
The EEG was recorded with an Aspect A-2000® XP monitor (BIS version 4.0; Aspect Medical Systems, Newton, MA). Surface electrodes (BIS-sensor XP) were applied on the forehead as recommended by the manufacturer. Electrode impedances were accepted if <7.5 kΩ (default settings). The averaging interval of the BIS monitor was set to 15 s. BIS ranges from 100 (awake) to 0 (isoelectric EEG) and the recommended value for surgical anesthesia is 40–60. The spectral power in the 70–110 Hz frequency band is considered to represent EMG and is displayed in decibels (dB relative to 0.0001 μV2, interval 30–80 dB).
The MLAEP were recorded using the auditory-evoked potential (AEP) monitor/2™ (version 1.6, Danmeter A/S, Odense, Denmark). Three surface electrodes (A-line™ electrodes; Danmeter A/S) were positioned at midforehead (+), left forehead (reference), and left mastoid (−). Electrode impedances were accepted if <5 kΩ (default settings). The Alaris AEP/2 index (AAI) is a composite index calculated from MLAEP or EEG changes or both changes depending on signal-to-noise ratio of the extracted MLAEP. If the signal-to-noise ratio <1.45, the AAI is not calculated from MLAEP but instead is based on EEG β-ratio and burst suppression. AAI ranges from 99 (awake) to 0 (deep hypnosis) and recommended range for surgical anesthesia is 15–25. The spectral power in the 65–85 Hz band is considered by the Alaris AEP/2 monitor as EMG (expressed 0–100 logarithmic).
EMG activity was continuously recorded using a Nervus C16 amplifier (Cephalon A/S, Denmark) connected to pairs of subdermal needle electrodes (12 mm, 25 gauge, Oxford Instruments, Surrey, UK) placed 1 cm apart in the frontal (close to the BIS-sensor, electrode one) and temporal (in-between BIS-sensor electrodes 3 and 4) muscles and secured with tape. The recordings were made between the two electrodes in each pair. The signals were filtered through a band width of 0.5–100 Hz. Visual offline analysis was done using Nervus software (version 3.3). The adequacy and functional integrity of the EMG recordings was checked with voluntary facial muscle activity before and after the experiment. This directly measured EMG (EMGdirect) is what we consider as the “true” EMG to which we compared the BIS and AAI monitors estimation of EMG as specified above.
Noxious stimulation was applied to the arm not used for NMB monitoring. Cutaneous tetanic electrical stimulation (TET) of 60 mA at 100 Hz for 3 s was delivered to the ulnar nerve via surface electrodes using a nerve stimulator (CEFAR Tempo®, CEFAR Medical AB, Lund, Sweden).
Noxious stimulation was delivered to 20 patients at four separate occasions in each patient: the first occasion was used as baseline control and was performed after equilibration for 30 min at an ETSEV of 1.2%, or slightly higher in order to maintain BIS at 50, the second and third TET at either 50% or 95% T1 twitch depression, and the fourth TET after administration of reversal drugs and return of the TOF-ratio T4/T1 to >0.90. The resultant changes in BIS, AAI, and the EMG derived from the BIS and AAI (EMGBIS, EMGAAI) as well as the direct EMG recordings (Nervus) from frontal and temporal muscles (EMGdirect) were recorded from 2 min before until 2 min after each noxious stimuli and were analyzed offline (Fig. 2). The remaining five patients served as controls to determine changes in BIS, AAI, and EMG over time. Noxious stimulation was not applied in the control patients who received sevoflurane only to an ETSEV of 1.2% or slightly higher if BIS exceeded 50 at the start of the observation period. If the patient showed BIS values that persisted >60 for more than 1 min, or prolonged/severe coughing after TET, rescue drug was given (increasing ETSEV). After rescue drug had been given, the anesthesia was returned to the previous level of ETSEV and allowed to equilibrate for another 10–15 min before subsequent TETs. The patients were questioned for explicit recall according to the method of Brice et al. (21), in the postanesthesia care unit and a second time 1–3 mo postoperatively. The experimental protocol was completed before the start of surgery.
Data Analysis and Statistics
Mean values of the monitored parameters over a 2-min period before (pre) and after (post) stimulation were first calculated for each patient at the four occasions and are presented as median (quartile range) in Table 2. Differences between post and pre values for each patient are presented individually (Fig. 1). To show the time course of the responses to noxious stimulation, the index values were normalized to 100 at noxious stimulation (time zero) (Fig. 2). Using data from one of our previous studies (22) we estimated a sample size of 20 (paired t-test) giving a power of 80% (α = 0.05). Data were not normally or log-normally distributed and hence nonparametric tests were used (Wilcoxon's matched pairs test). The coupling between monitor EMG and corresponding index values was analyzed using ordinary linear regression. A commercial statistical software was used (Statistica; Statsoft®, Tulsa, OK) and P values <0.05 were considered statistically significant.
Demographic data are displayed in Table 1. One patient at baseline was excluded because of nerve stimulator electrode failure. Four patients were excluded after neostigmine reversal: 2 patients because of spontaneous coughing before TET, 1 patient because of high BIS values before TET, and 1 because of shortage of time. The ETSEV at baseline (after 30 min equilibration) was 1.26% ± 0.11% (mean ± sd) (control group 1.22% ± 0.04%) corresponding to a median (quartile range) of 44 (39–50) for BIS and 15 (14–16) for AAI (Table 2). The maximal variation in prestimulus BIS at the four different experimental situations in each patient was 4.6 (2.3–20.1) (median and range). The corresponding data for AAI were 2.9 (1.2–7.5). Rescue anesthesia was required by 7 (6 coughers) of the patients at baseline, 8 (7 coughers) patients at T1 50% depression, 1 (1 cougher) patient at T1 95% depression, and 7 (6 coughers) patients after neostigmine reversal. No patient reported explicit recall.
Absence of Noxious Stimulation
In the absence of noxious stimulation, BIS, AAI, EMGBIS, and EMGAAI did not change at either 50% or 95% twitch depression or after neostigmine reversal with a TOF ratio >0.90 when compared with baseline values (Table 2). EMGdirect was also unchanged by different degrees of NMB and after neostigmine reversal.
After Noxious Stimulation
During profound NMB (T1 95% depression) noxious stimulation altered the BIS and AAI responses significantly when compared with during partial NMB (T1 50% depression) (P = 0.01 and P < 0.01, respectively), or after neostigmine reversal (P < 0.01 and P = 0.01, respectively) (Fig. 1 and Table 2). The AAI response at T1 95% depression was also significantly different from baseline, whereas the BIS response did not reach statistical significance (P = 0.02 and, P = 0.08 respectively) (Table 2). The temporal aspects of the changes in BIS and AAI at 2 min after noxious stimulation at different degrees of NMB are shown in Figure 2. The responses in EMGBIS and EMGAAI after noxious stimulation were similar to the alterations in BIS and AAI (Table 2). Regression analysis yielded significant correlations between EMGBIS, EMGAAI, and their corresponding index changes after noxious stimulation (Fig. 3).
After noxious stimulation minimal EMGdirect activity in the frontalis and temporalis muscles close to the electrode tips was evident in 10 of the 20 patients, but the EMGdirect signal was too small for accurate quantification.
In the control group (n = 5) sevoflurane was administered over the same experimental time period (60 min) as in the study group (n = 20). No NMBD or noxious stimuli were applied. During the 60 min control period BIS, AAI, EMGBIS, EMGAAI, and EMGdirect did not change from baseline.
We demonstrate that profound NMB blunts BIS and AAI responses to a standardized noxious stimulation during sevoflurane anesthesia when compared with a partial NMB or after neostigmine-induced reversal. This effect was probably not related to a depression of the EMG component of the BIS or AAI, because directly measured EMG (EMGdirect) in frontal and temporal muscles was absent or negligible.
The results indicate that a profound NMB exerts an effect on the level of hypnosis as measured by BIS and AAI under noxious stimulation. Because nondepolarizing NMBDs do not cross the intact blood–brain barrier and, thus, have not been assigned any direct central effect (23,24) our findings support the afferentation theory [i.e., decreased arousing afferent input from muscles (13)]. Our findings are further strengthened by the fact that two independent hypnosis monitoring principles, BIS and AAI, yielded similar results. Still, our data do not completely exclude the possibility that profound NMB directly suppresses electrical activity in the muscles. Thus, the mechanism for the observed interaction between the degree of NMB and BIS/AAI response to noxious stimulation still remains unclear.
Studies on the effect of NMB on the level of hypnosis have produced conflicting results (6–12). One reason for this may be that EMG activity has contaminated the EEG signals to various degrees in different studies. EMG is generally considered to constitute an increasing part of the electrical power spectrum above the frequency of about 30 Hz. Thus, EMG has a potential to distort the index calculating algorithms if higher frequencies are considered. The BIS algorithm quantifies the spectral power between 70 and 110 Hz as an estimate of EMG activity, and the corresponding range for AAI is 65–85 Hz. When using surface electrodes to record the EEG and monitor the level of hypnosis, it cannot be determined whether the electrical activity detected is generated in the brain or in the frontal muscle. In this study, the EMG was measured directly in temporal and frontal muscles and was found to be negligible both with and without NMB. The integrity of the EMG recording system was confirmed by voluntary action in the awake state, before, and after the experiments. Negligible EMG activity was also confirmed in a study by Sleigh et al. (9). By measuring submental EMG they quantified the contamination of the EEG by frontal EMG and found that the EMG signal was, on average, 10 times less in magnitude than the EEG in the low γ range (i.e., 30–47 Hz). Hence, the EMG activity during light sevoflurane anesthesia probably did not influence the EEG recordings in this study.
Results in previous studies may also be related to whether an EEG response to different degrees of NMB was studied during noxious stimulation or not (6–12). No effect of NMB on BIS could be shown in a study in which propofol and mivacurium, without noxious stimulation, were administered to volunteers (8), nor could any effect be seen in another study using atracurium and propofol/remifentanil in nonstimulated patients (11). This agrees with our finding that NMB does not affect BIS or AAI during resting conditions. However, after noxious stimulation the NMB changed the reaction pattern of the displayed parameters from the BIS and AAI monitors (Figs. 1 and 2 and Table 2). A similar finding was also demonstrated by Lanier et al. (13) in dogs. Using pancuronium and halothane, they found a weaker cerebral response to noxious stimulation when the dogs were subjected to a NMB. Lanier et al. (13) put forward the afferentation theory, which states that proprioceptive afferent input from muscle stimulation will produce cerebral arousal. However, the afferent theory was not confirmed by Fahey et al. evaluating different NMBDs and their effect on minimum alveolar concentration (MACskin incision) for halothane in humans (25).
Interestingly, although EMGdirect, which we consider as the true representation of EMG, was almost absent, significant electrical activity considered as EMG by the BIS and AAI monitors was evident after noxious stimulation. EMG from the BIS and AAI monitors was highly correlated with their respective index values, even at profound NMB in patients who did not cough or move (Fig. 3). This leads us to postulate that the NMB-related alteration in EMG content, as estimated by BIS and AAI monitors may, in fact, constitute EEG activity in the high frequency band (i.e., the γ-band), an assumption that should be further investigated by the use of a multiple lead EEG. In addition, if coherence between multiple EEG leads is affected by the degree of NMB, our hypothesis, that the blunted response to noxious stimulation is because of a true EEG effect, would be further strengthened. If this hypothesis is further supported, our findings may constitute an argument for using a deep NMB in hemodynamically unstable patients because of less need for circulating more compromising general anesthetics. This would require that current or future hypnosis monitoring technologies are considered sufficiently accurate to avoid awareness. At present our findings do not justify any change in clinical practice.
Critiques of Methods
The interpretation of our results should take some limitations into consideration. The major drawback in this study is that only the responses to noxious stimulation at baseline as measured by AAI, but not BIS, followed the expected pattern—namely, the same response at baseline as after neostigmine reversal. We assume that remifentanil, used to facilitate intubation, had a lingering effect, even though we used 30 min to equilibrate baseline levels. Another plausible explanation could be the windup phenomenon, where repeated noxious stimulation gives rise to a stronger reaction to subsequent stimulation (26). The need for rescue anesthesia in a considerable number of patients at no, or partial, NMB may also have served to reduce the difference from the recordings at baseline. Theoretically, the degree of NMB could moderate the effect from TET by reducing noxious muscular contractions. However, we think this is of minor importance, because the TET gives rise to pronounced afferent sensory signaling via pain neurons, which is the same whether or not the patient is neuromuscular blocked. Therefore, we believe that the intensity of the noxious stimulus is equivalent regardless of the degree of NMB.
We aimed to conduct this investigation under light anesthesia without causing awareness with explicit recall. Therefore, an ETSEV of 1.2% (approximately 2 × MACawake) was used during the experiment, anticipated to give a BIS of approximately 50 (22,27,28). In six of the patients, ETSEV had to be increased slightly (at most to 1.6% in one patient) to prevent BIS from exceeding 50 at baseline. The resulting ETSEV at baseline was 1.26% ± 0.11%, corresponding to BIS 44 (39–50). Thus, our patients may not have been investigated in exactly the same hypnotic state. However, there was no correlation between the BIS value immediately before noxious stimulation and the magnitude of the subsequent BIS response (Fig. 1). In addition the individual ETSEV at baseline was kept constant in all patients during the three following episodes with noxious stimulation. Therefore, we regard the slight variation among patients in ETSEV as a minor drawback.
Another limitation in our study is that we used only two degrees of NMB. However, our main purpose was to determine if there was any difference between partial and profound NMB, and therefore we consider these two degrees sufficient. Spontaneous alterations in BIS, AAI, or their corresponding estimates of EMG activity over time that could have affected our results were found negligible in control patients.
In conclusion, BIS and AAI responses to noxious TET are affected by the degree of NMB during sevoflurane anesthesia, whereas there is no effect of NMB on BIS and AAI in the absence of noxious stimulation.
Bodil Alexanderson, RNA, for recording and collecting the data and engineer Mats Johansson for computer support.
1. Sandin RH, Enlund G, Samuelsson P, Lennmarken C. Awareness during anaesthesia: a prospective case study. Lancet 2000;355:707–11
2. Rampil IJ. A primer for EEG signal processing in anesthesia. Anesthesiology 1998;89:980–1002
3. Ekman A, Lindholm ML, Lennmarken C, Sandin R. Reduction in the incidence of awareness using BIS monitoring. Acta Anaesthesiol Scand 2004;48:20–6
4. Myles PS, Leslie K, McNeil J, Forbes A, Chan MT. Bispectral index monitoring to prevent awareness during anaesthesia: the B-aware randomized controlled trial. Lancet 2004;363:1757–63
5. Dahaba AA. Different conditions that could result in the bispectral index indicating an incorrect hypnotic state. Anesth Analg 2005;101:765–73
6. Messner M, Beese U, Romstöck J, Dinkel M, Tschaikowsky K. The bispectral index declines during neuromuscular block in fully awake persons. Anesth Analg 2003;97:488–91
7. Vivien B, Di Maria S, Ouattara A, Langeron O, Coriat P, Riou B. Overestimation of bispectral index in sedated intensive care unit patients revealed by administration of muscle relaxant. Anesthesiology 2003;99:9–17
8. Greif R, Greenwald S, Ekkehard S, Laciny S, Rajek A, Caldwell JE, Sessler DI. Muscle relaxation does not alter hypnotic level during propofol anesthesia. Anesth Analg 2002;94:604–8
9. Sleigh JW, Steyn-Ross DA, Steyn-Ross ML, Williams ML, Smith P. Comparison of changes in electroencephalographic measures during induction of general anaesthesia: influence of the gamma band and the electromyogram signal. Br J Anaesth 2001;86:50–8
10. Bruhn J, Bouillon TW, Shafer SL. Electromyographic activity falsely elevates the bispectral index. Anesthesiology 2000;92:1485–7
11. Vasella FC, Frascarolo D, Spahn DR, Magnusson L. Antagonism of neuromuscular blockade but not muscle relaxation affects depth of anesthesia. Br J Anaesth 2005;94:742–7
12. Liu N, Chazot T, Huybrechts I, Law-Koune JD, Barvais L, Fischler M. The influence of a muscle relaxant bolus on bispectral and Datex-Ohmeda entropy values during propofol-remifentanil induced loss of consciousness. Anesth Analg 2005;101:1713–8
13. Lanier WL, Iaizzo PA, Milde JH, Sharbrough FW. The cerebral and systemic effects of movement in response to a noxious stimulus in lightly anesthetized dogs. Possible modulation of cerebral function by muscle afferents. Anesthesiology 1994;80:392–401
14. Forbes AR, Cohen NH, Eger EI. Pancuronium reduces halothane requirement in man. Anesth Analg 1979;58:497–9
15. Hodgson PS, Liu SS. Epidural lidocaine decreases sevoflurane requirement for adequate depth of anesthesia as measured by the bispectral index monitor. Anesthesiology 2001;94:799–803
16. Tverskoy M, Shagal M, Finger J, Kissin I. Subarachnoid bupivacaine blockade decreases midazolam and thiopental hypnotic requirements. J Clin Anesth 1994;6:487–90
17. Litvan H, Jensen EW, Galan J, Lund J, Rodriguez BE, Henneberg SW, Caminal P, Villar Landiera JM. Comparison of conventional averaged and rapid averaged, autoregressive-based extracted auditory evoked potentials for monitoring the hypnotic level during propofol induction. Anesthesiology 2002;97:351–8
18. Absalom AR, Sutcliffe N, Kenny GNC. Effects of the auditory stimuli of an evoked potential system on levels of consciousness, and on the bispectral index. Br J Anaesth 2001;87:778–80
19. Struys MMRF, Jensen EW, Smith W, Smith TN, Rampil I, Dumortier FJE, Mestach C, Mortier EP. Performance of the ARX-derived auditory evoked potential index as an indicator of anesthetic depth. Anesthesiology 2002;96:803–16
20. Eriksson LI, Sundman E, Olsson R, Nilsson L, Witt H, Ekberg O, Kuylenstierna R. Functional assessment of the pharynx at rest and during swallowing in partially paralyzed humans: simultaneous videomanometry and mechanomyography of awake human volunteers. Anesthesiology 1997;87:1035–43
21. Brice DD, Hetherington RR, Utting JE. A simple study of awareness and dreaming during general anaesthesia. Br J Anaesth 1970;42:535–42
22. Ekman A, Brudin L, Sandin R. A comparison of bispectral index and rapidly extracted auditory evoked potentials index responses to noxious stimulation during sevoflurane anesthesia. Anesth Analg 2004;99:1141–6
23. Matteo RS, Pua EK, Khambatta HJ, Spector S. Cerebrospinal fluid levels after d-tubocurarine in man. Anesthesiology 1977;46:396–9
24. Werba A, Gilly H, Weidlmayr-Goettel M, Spiss CK, Steinbereithner K, Czech T, Agoston S. Porcine model for studying the passage of non-depolarizing neuromuscular blockers through the blood brain barrier. Br J Anaesth 1992;69:382–6
25. Fahey MR, Sessler DI, Cannon JE, Brady K, Stoen R, Miller RD. Atracurium, vecuronium, and pancuronium do not alter the minimum alveolar concentrations of halothane in humans. Anesthesiology 1989;71:53–6
26. Herrero JF, Laird JMA, Lopez-Garcia JA. Wind-up of spinal cord neurons and pain sensation: much ado about something. Prog Neurobiol 2000;61:169–203
27. Katoh T, Suzuki A, Ikeda K. Electroencephalographic derivates as a tool for predicting the depth of sedation and anesthesia induced by sevoflurane. Anesthesiology 1998; 88:642–50
© 2007 International Anesthesia Research Society
28. Olofsen E, Dahan A. The dynamic relationship between end-tidal sevoflurane and isoflurane concentrations and bispectral index and spectral edge frequency of the electroencephalogram (EEG). Anesthesiology 1999;91:1687–92