Anesthesiologists lack a generally accepted indicator of anesthetic adequacy. Part of the problem is related to the fact that there is no standard against which to assess indicators of anesthetic adequacy. While vital signs are used clinically to monitor patient status during anesthesia, hemodynamic responses alone are not adequate, since many factors contribute to hemodynamic responses  which have poor predictive value for anesthetic depth [2,3].
The movement response to skin incision, minimum alveolar anesthetic concentration (MAC) has classically been used to describe the potency of inhaled anesthetics [4-6]. However, contemporary anesthetic practice often involves the use of neuromuscular blocking drugs which diminish the possibility of using the movement response as an indicator of anesthetic inadequacy. Various electroencephalographic (EEG) variables have been studied as possible indicators of anesthetic adequacy. Power spectral analysis of the EEG permits computation of the spectral edge frequency [7-13] and the median power frequency [14-16]. Neither of these methods has been shown to be sufficiently reliable for general use for assessing anesthetic efficacy during routine procedures [17,18].
Bispectral analysis (BIS) of the EEG  is a signal processing technique that has been proposed as a pharmacodynamic measure of anesthetic effects on the central nervous system. This method decomposes the EEG and quantifies the level of synchronization in the signal, along with the traditional amplitude and frequency variables, thereby providing a more complete description of complex EEG patterns. Previous retrospective studies have demonstrated that an empirically derived bispectral EEG measure (BIS) could be used to predict movement in response to skin incision using isoflurane/oxygen , propofol/nitrous oxide , isoflurane/alfentanil, or propofol/alfentanil anesthesia . These studies suggested that BIS might be useful in predicting movement with several different anesthetics, but it was not clear whether the responses were independent of the anesthetics used.
The present multicenter study was designed to prospectively investigate the use of BIS as a guide to anesthetic dosing under clinically relevant conditions. The hypothesis was that adjustment of anesthetic dose to achieve lower BIS values would decrease the probability of patient movement in response to skin incision. Movement at incision was selected as the primary indicator of inadequate anesthesia because it is a well defined, clinically relevant response which has been widely studied. The present study also intended to evaluate BIS under a wider variety of anesthetic conditions. Pharmacokinetic/dynamic models were then retrospectively used to study the relationship between BIS, changes in vital signs, purposeful movement, and predicted dose estimates (e.g., expired concentrations of volatile anesthetics, plasma, and effectsite concentrations of intravenous anesthetics).
After institutional review board approval at eight study sites, 304 consenting patients (ASA physical status I-III), scheduled for elective noncranial surgical procedures requiring a minimum 1-in. skin incision, were studied. One center enrolled only four evaluable patients and was not included in the analysis, so this report will refer to 300 patients from seven study sites.
At each site, patients were randomly allocated to two groups with a site specific anesthetic technique (Table 1). In the control group, the anesthetic was designed to give an approximately 50% movement response rate to skin incision. BIS was monitored, but no action was taken by the anesthesiologist on the BIS data displayed. In the BIS-guided group, BIS was displayed and the anesthetic concentration altered to decrease BIS to a target of less than 60 (provided the patient was hemodynamically stable).
In the majority of patients, the tracheas were intubated after administration of succinylcholine 1-2 mg/kg intravenously. Two patients underwent awake-tracheal intubation, and a laryngeal mask was used in one.
Prior to skin incision (after return of neuromuscular function as determined by a peripheral nerve stimulator), a supramaximal (50 or 100 Hz) tetanic electrical stimulus was applied to the ulnar nerve for at least 5 s as an initial test stimulation. If the patient showed a purposeful movement, they were classified as a mover for study purposes and anesthesia was deepened. If the patient did not move, anesthesia was maintained at the same level until skin incision. At the time of incision, a determination of any movement in the succeeding 2 min was made. Any movement (except coughing and bucking) was considered a positive response indicative of inadequate anesthesia.
Before induction of anesthesia, gold cup electrodes were secured to the left and right frontal pole (Fp1 and Fp2) regions and referenced to a central vertex electrode (Cz). Impedance was maintained less than 5 K Omega. The raw EEG was continuously recorded using a Model B500 Spectral EEG Monitor (Aspect Medical Systems, Natick, MA). EEG processed variables were calculated in real time and stored on a hard disk. The BIS (Version 1.1) computed from the bilateral frontal channels was displayed on the monitor throughout the procedure. Mean arterial blood pressure (MAP) and heart rate (HR) were measured at 1-min intervals (until 3 min after incision). These data were acquired simultaneously with the EEG variables using a serial interface between the cardiovascular and the EEG monitor. End-tidal anesthetic (if applicable), end-tidal CO2, and oxygen saturation (Spo2) were also continuously recorded.
For pharmacodynamic modeling, individual patient drug profiles were calculated using STANPUMP Software (available at http://pkpd.icon.palo-alto.med.va.gov/) and used to estimate plasma and effect-site anesthetic concentrations from recorded drug doses and patient height, weight, and age. Alfentanil and sufentanil plasma concentrations were converted to fentanyl equivalents in the ratios of 75 to 0.11 to 1, respectively [9,21]. Multivariate nonlinear logistic regression models were then developed using SPSS software (SPSS, Inc., Chicago, IL) to describe the relationships among estimated drug concentrations, BIS, and the probability of movement.
Statistical analysis used analysis of variance, Student's t-test, and logistic regression as appropriate. In regression analysis, single factor (maximum likelihood) logistic models were calculated first, then a second factor was added to the model and compared to the first with a goodness of fit chi squared test to determine whether the fit of the model was improved significantly. Mean values for the BIS (Version 1.1), estimated drug concentrations, and recorded hemodynamic variables immediately prior to incision or tetanic stimulation were used to determine intergroup differences (movers versus nonmovers, control versus BIS-guided treatment groups). Results are expressed as mean values +/- SD. P values < 0.05 were considered statistically significant.
There were no significant differences in demographic variables between the study groups at any of the seven sites (Table 1). Study Sites 1, 2, and 3 used isoflurane-(+/- N2 O) to maintain anesthesia, Sites 4 and 5 used a combination of opioids and isoflurane, and Sites 6 and 7 used continuous infusions of alfentanil and propofol. Figure 1 displays a graph of BIS over time from two sample patients, each receiving an isoflurane anesthetic prior to incision. The first patient belongs to the control group; the second belongs to the BIS-guided group.
Control Versus BIS-Guided Groups
A lower BIS value was found in the BIS-guided group as compared to the control group (Table 2). Among the individual sites, the three sites that maintained anesthesia with isoflurane (Sites 1, 2, and 3), and one site that used isoflurane with opioids (Site 5) had significantly lower BIS values in the BIS-guided group compared with the control group. Two sites that used propofol and alfentanil infusions (Sites 6 and 7), and Site 4 that used isoflurane with opioids had no significant difference in BIS between the BIS-guided anesthesia group when compared with the control group (see Table 2).
Various possible discriminators were examined to attempt to differentiate between the combined (seven sites) control groups compared to the BIS-guided anesthesia groups (Table 2). BIS, HR, and weight were statistically significantly different between the two groups. Prestimulus systolic, mean, and diastolic blood pressure, as well as estimated effect-site opioid concentration were not significantly different between the two treatment groups.
Initial tetanic test stimulation elicited a response in significantly fewer patients in the control group (11%) compared with 3% in the BIS-guided group, with most of the responses occurring at Site 1 (isoflurane alone). The movement response rate to stimulation (tetanus and incision) (Table 3) in the BIS-guided anesthesia group was significantly less than the control group in the combined analysis (43% versus 13%). Incision response rates in the BIS-guided group were also lower at all three sites that used isoflurane (Sites, 1, 2, and 3) compared with the control group. The movement response rates were not different between the two response groups at the two opioid/isoflurane/N2 O sites (Sites 4 and 5), but at Site 5, none of the patients in either group moved in response to tetanic stimulation or surgical skin incision. At the two propofol/alfentanil sites, Site 6 had a greater response rate in the control group as opposed to the BIS-guided anesthesia group, whereas Site 7 showed no statistically significant difference. However, when Sites 6 and 7 were combined, the control group had a significantly higher movement response rate (Table 3).
BIS values were significantly higher in responding patients as opposed to nonresponding patients at the three isoflurane sites combined and the two propofol/alfentanil sites combined. Site 5 (opioid/isoflurane) had no patients who moved in response to skin incision so BIS could not be evaluated as a predictor of movement. When the patients were maintained with an opioid/isoflurane anesthetic, BIS did not predict movement. BIS values at the various sites for movers compared with nonmovers are shown in Table 4.
In Figure 2, the probability of movement response relationship as a function of BIS (calculated using logistic regression) is displayed for six different sites (Sites 1, 2, 3, 4, 6, and 7). The curves for Sites 1, 2, 3, and 7 show a relatively steep dose-response curve between BIS and probability of movement at skin incision. The curves for Site 7 (alfentanil/propofol) and Site 4 (opioid/isoflurane) are flatter, reflecting the lower response rates at high BIS levels found at these centers. It must be recognized, however, that the low number of movers at some sites resulted in less accurate logistic models, due to the unequal distribution of movers/nonmovers. Although these relationships reached significance for each site, considerable variance remained unexplained, so confidence intervals (not displayed) around each curve are broad.
Among the five sites that used isoflurane (Sites 1-5), the end-tidal isoflurane concentrations were lower in the control group compared with the BIS-guided anesthesia group (Table 5). When the five sites were combined, there was still a lower end-tidal isoflurane concentration in the control groups compared with the BIS-guided anesthesia groups, and the difference in the mover groups compared with the nonmover groups also reached statistical significance. Figure 3 shows the corresponding probability of movement response curves as a function of preincision end-tidal isoflurane concentration at the four individual sites which had movers (Site 5 had no movers, so no probability of response relationship could be derived). Regression analysis showed that, in patients who did not receive opioids, higher isoflurane values are associated with no movement and lower isoflurane values are associated with movers. As estimated opioid effect-site concentrations increased, more patients with low preincision isoflurane values (and high BIS levels) did not move in response to skin incision (compare the curves from Sites 1 and 4 in Figure 2 and Figure 3). In order to validate this modeling, estimated effect-site concentration that would reduce isoflurane MAC by 50% was calculated by logistic regression analysis. An estimated fentanyl effect-site concentration of 0.67 ng/mL (0.61-0.73 ng/mL; 95% confidence interval) was associated with a 50% reduction in isoflurane MAC (non-age-adjusted).
Preincision estimated propofol effect-site concentrations at the two alfentanil/propofol sites (Sites 6 and 7) are shown in Table 5. There was no significant difference in predicted propofol effect-site concentration when movers were compared to nonmovers at either site. In the combined database, there were significantly lower estimated effect-site opioid concentrations in the movement response groups (0.7 +/- 1.02 ng/mL, n = 76) as opposed to the nonmovement groups (1.39 +/- 1.27 ng/mL, n = 148). No significant differences were observed between the control and BIS-guided anesthesia groups.
The hemodynamic response to incision was studied in a subset of patients (n = 203) who had not received any preincision vasoactive medications (chronic anti-hypertensives, intraoperative vasoconstrictors, or vasodilators). Using a 20% increase in MAP to define a significant hypertensive response, only 31 patients responded to incision with a significant increase in MAP. Various independent variables were investigated as predictors of a 20% change in MAP with skin incision at all of the sites combined (Table 6). The only variables found to predict this change in MAP were low estimated opioid effect-site concentration and low patient weight.
Logistic Regression Analysis
In order to determine the most significant predictors of patient movement, all available baseline and preincision systolic, diastolic, mean blood pressure, HR, BIS, and anesthetic dosage variables were tested by maximum likelihood logistic regression as potential predictors of patient movement. BIS estimated opioid effect-site concentrations and mean blood pressure were found to be the best independent predictors of response for this population. Combining BIS and estimated opioid effect-site concentration information significantly improved the ability to predict patient response. The modeled relationships between BIS response, probability of movement, and increasing concentrations of isoflurane, propofol, and estimated opioid levels are presented in Figure 4.
Anesthesiologists have no clinically useful monitor of anesthetic effects on brain function [17,22]. Anesthetic dosage is generally adjusted based on a constellation of signs, each of which is sensitive to many factors unrelated to the anesthetic present. An anesthetic adequacy monitor should be able to predict inadequate anesthesia intraoperatively and be able to track recovery from anesthesia. In order to be used on a routine basis, it should perform reliably with a variety of anesthetics.
No prospective multicenter study has been conducted to evaluate the use of a processed EEG variable as a measure of the clinical signs of anesthetic adequacy. Our study design used multiple general anesthetic techniques in an attempt to evaluate the general utility of the BIS in common clinical practice. This study design had the potential for introducing significant confounding variability, especially between study sites, and the possibility of unrecognized site differences influencing our findings certainly exists. Nevertheless, by comparing responses across anesthetic regimens, it is possible to gain valuable insights regarding the components of anesthesia and their relationship to the EEG.
Movement at incision was selected as the primary indicator of inadequate anesthesia based on the use of MAC as an index of inhaled anesthetic potency . The initial anesthetic doses or concentrations were chosen at the various study sites with the anticipation that about 50% of the patients would move in response to surgical skin incision in the control groups. For the combined patient population, a 43% response rate was obtained. However, some of the anesthetic regimens were more effective at preventing patient response than we had predicted, so the response rates in the control group at the propofol/alfentanil sites, and the isoflurane/opioid sites were well below 50%. In contrast, the response rate at two of the isoflurane sites (Sites 1 and 2) were higher than 50%.
In this study, movement response rates at incision were reduced significantly in the BIS-guided group compared to the control group (Table 3). However, the two sites that used opioid/isoflurane maintenance techniques had very low response rates in both treatment groups, even though relatively low concentrations of isoflurane were used and BIS levels remained correspondingly high. The results of the pharmacodynamic modeling suggest that this was due to the higher opioid concentrations present at the time of skin incision in these individuals. It should be recognized that measurement of actual opioid plasma concentrations was not attempted in this study. Since population kinetic variables incorporated in STAN-PUMP were used to estimate opioid concentrations and calculated alfentanil and sufentanil concentrations were expressed as fentanyl "equivalents," there is a possibility of error in the pharmacodynamic modeling. The validity of this model is supported by comparing the estimated fentanyl "equivalents" needed to reduce isoflurane MAC by 50% (0.67 ng/mL, 0.61-0.73 ng/mL, 95% confidence interval [CI]) with published results using measured fentanyl concentrations to achieve the same effect [0.5 ng/mL, 0.0-4.6 ng/mL, 95% CI ; and 1.67 ng/mL, 1.11-2.38 ng/mL, 95% CI ].
Increases in MAP and HR in response to skin incision are usually also considered indicators of inadequate anesthesia, but there is not always good correlation between adequate anesthesia and hemodynamic response [1,25]. In this study, an increase of 20% in mean blood pressure in response to skin incision was selected as a significant hemodynamic response criterion. Interestingly, only 31 individuals exhibited even this moderate degree of blood pressure response, in spite of the fact that a much larger number of patients actually moved at this time. Thus, a lack of hemodynamic (and other autonomic responses) cannot always be relied upon as a good indicator of adequate anesthesia. The only variables found to be associated with an acute hemodynamic response were low estimated opioid concentrations and low weight. While we have no real explanation for the association between low weight and MAP response, our observation that higher estimated levels of opioids prevented both movement and blood pressure response supports the conclusion that adequate analgesia is a critical component of a clinically adequate anesthetic state.
Overall, these results indicate that the BIS is a significant predictor of patient response to incision, but the utility of the BIS depends on the anesthetic technique being used. When hypnotic drugs such as propofol or isoflurane are used as the primary anesthetic, there appears to be a good correlation between changes in BIS and the probability of response to skin incision (Figure 4). When opioid analgesics are used as adjuncts prior to incision, the correlation to patient movement becomes much less significant, so that patients with apparently "light" EEG profiles may not move or otherwise respond to incision. This observation provides evidence of a distinction between the analgesic and hypnotic effects of these anesthetics, which is consistent with recent reports that the neural substrate of the movement response might be separate from the cortical generators of EEG [26-28]. Suppression of movement at incision appears to be mediated by actions on the spinal cord that need not always correlate with anesthetic effects on the EEG or higher cortical functions, such as consciousness and memory . Indeed, several studies [30-35], conducted after the completion of this trial, have reported a strong correlation between BIS, levels of sedation, and impairment of memory function induced by propofol, isoflurane, and midazolam.
In conclusion, this multicenter study demonstrates that dosing anesthetics to lower BIS values results in a lower probability of movement in response to surgical skin incision. However, the usefulness of EEG monitoring with BIS clearly depends on the anesthetic technique and clinical end point used to define anesthetic adequacy. BIS correlated best with the effects of primary hypnotic-based anesthetic techniques. In contrast, opioid-based techniques reduced patient responsiveness at estimated effect-site doses which had little effect on the EEG. This "opioid effect" confounds the use of BIS as a measure of anesthetic adequacy when lack of movement at incision is used to define efficacy.
The authors wish to acknowledge the skilled assistance of Dr. Jin Liu, Dr. Michail Avramov, Dr. Gretchen Hollingsworth, Mr. Steven Behr, and Dr. Dave Martel.
1. Ghouri AF, Monk TG, White PF. Electroencephalogram, spectral edge frequency, lower esophageal contractility and autonomic responsiveness during general anesthesia. J Clin Monit 1993;9:176-85.
2. Hug CC. Does opioid anesthesia exist? Anesthesiology 1990;73:1-4.
3. Dutton RC, Smith WD, Smith NT. The use of EEG to predict movement during anesthesia. In: Rosen M, Lunn JN, eds. Consciousness, awareness and pain in general anaesthesia. London: Butterworth, 1987;72-82.
4. Eger EI II, Saidman LJ, Brandstater B. Minimum alveolar anesthetic concentration: a standard of anesthetic potency. Anesthesiology 1965;26:756-63.
5. Vernon JM, Lang E, Sebel PS, Manberg P. Prediction of movement using bispectral EEG during propofol/alfentanil or isoflurane/alfentanil anesthesia. Anesth Analg 1995;80:780-5.
6. Sebel PS, Bowles SM, Saini V, Chamoun N. EEG bispectrum predicts movement during thiopental isoflurane anesthesia. J Clin Monit 1995;11:83-91.
7. Rampil IJ, Matteo RS. Changes in EEG spectral edge frequency correlate with the hemodynamic response to laryngoscopy and intubation. Anesthesiology 1987;67:139-42.
8. Sidi A, Halimi P, Cotev S. Estimating anesthetic depth by electroencephalography during anesthetic induction and intubation in patients undergoing cardiac surgery. J Clin Anesth 1990;2:101-7.
9. Scott JC, Ponganis KV, Stanski DR. EEG quantitation of narcotic effect: the comparative pharmacodynamics of fentanyl and alfentanil. Anesthesiology 1985;62:234-41.
10. Hudson RJ, Stanski DR, Saidman LJ, Meathe E. A model for studying depth of anesthesia and acute tolerance to thiopental. Anesthesiology 1983;59:301-8.
11. Buhrer M, Maitre PO, Hung OR, et al. Thiopental pharmacodynamics. Defining the pseudosteady-state serum concentration-EEG effect relationship. Anesthesiology 1992;77:226-36.
12. White PF, Boyle WA. Relationship between hemodynamic and electroencephalographic changes during general anesthesia. Anesth Analg 1989;68:177-81.
13. Long CW, Shah NK, Loughlin C, et al. A comparison of EEG determinants of near-awakening from isoflurane and fentanyl anesthesia: spectral edge, median power frequency and delta ratio. Anesth Analg 1989;69:169-73.
14. Schwilden H, Schuttler J, Stoeckel H. Quantitation of the EEG and pharmacodynamic modeling of hypnotic drugs: etomidate as an example. Eur J Anaesth 1985;2:121-30.
15. Schwilden H, Stoeckel H, Schuttler J. Closed-loop feedback control of propofol anaesthesia by quantitative EEG analysis in humans. Br J Anaesth 1989;62:290-6.
16. Schwilden H, Schuttler J, Stoeckl H. Closed loop feedback control of methohexital anesthesia by quantitative EEG analysis in humans. Anesthesiology 1987;67:341-7.
17. Stanski DR. Monitoring depth of anesthesia. In: Miller RD, ed. Anesthesia. New York: Churchill Livingstone, 1986;1001-29.
18. White PF. Clinical uses of intravenous anesthetic and analgesic infusions. Anesth Analg 1989;68:161-71.
19. Sigl JC, Chamoun NC. An introduction to bispectral analysis for the EEG. J Clin Monit 1994;10:392-404.
20. Kearse LA, Manberg P, Chamoun N, et al. Bispectral analysis of the electroencephalogram correlates with patient movement to skin incision during propofol/nitrous oxide anesthesia. Anesthesiology 1994;81:1365-70.
21. Shafer SL, Varvel JR. Pharmacokinetics, pharmacodynamics, and rational opioid selection. Anesthesiology 1991;74:53-63.
22. Sebel PS, Heneghan CP, Ingram DA. Evoked potentials. A neurophysiological indicator of depth of anaesthesia? Br J Anaesth 1985;57:841-2.
23. Westmoreland CL, Sebel PS, Gropper A. Fentanyl or alfentanil decreases the minimum alveolar anesthetic concentration of isoflurane in surgical patients. Anesth Analg 1994;78:23-28.
24. McEwan AI, Smith C, Dyar O, et al. Isoflurane minimum alveolar concentration reduction by fentanyl. Anesthesiology 1993;78:864-9.
25. Moerman N, Bonke B, Oosting J. Awareness and recall during general anesthesia. Anesthesiology 1993;79:454-64.
26. Rampil IJ. Anesthetic potency is not altered after hypothermic spinal cord transection in rats. Anesthesiology 1994;80:606-10.
27. Rampil IJ, Mason P, Singh H. Anesthetic potency (MAC) is independent of forebrain structures in the rat. Anesthesiology 1993;78:707-12.
28. Antognini JF, Schwartz K. Exaggerated anesthetic requirements in the preferentially anesthetized brain. Anesthesiology 1993;79:1244-9.
29. Kandel L, Chortkoff BS, Sonner J, et al. Nonanesthetics can suppress learning. Anesth Analg 1996;82:321-6.
30. Greenwald S, Chiang HH, Devlin P, et al. Bispectral index (BIS 2.0) as a hypnosis measure. Anesthesiology 1994;81:A477.
31. Lui J, Singh H, White PF. Electroencephalogram bispectral analysis predicts the depth of midazolam-induced sedation. Anesthesiology 1996;84:64-9.
32. Leslie K, Sessler DI, Schroeder M, Walters K. Propofol blood concentration and the bispectral index predict suppression of learning during propofol/epidural anesthesia in volunteers. Anesth Analg 1995;81:1269-74.
33. Glass P, Gan TJ, Sebel PS, et al. Comparison of the bispectral index (BIS) and measured drug concentrations for the monitoring effects of propofol, midazolam, alfentanil and isoflurane [abstract]. Anesthesiology 1995;83:A374.
34. Kearse L, Rosow C, Sebel P, et al. The bispectral index correlates with sedation/hypnosis and recall: Comparison using multiple agents. Anesthesiology 1995;83:A507.
35. Glass P, Sebel PS, Greenwald S, Chamoun N. Quantification of the relative effects of anesthetic agents on the EEG and patient responsiveness to incision. Anesthesiology 1994;81:A407.