Secondary Logo

Journal Logo

General Article

Propofol Blood Concentration and the Bispectral Index Predict Suppression of Learning During Propofol/Epidural Anesthesia in Volunteers

Leslie, Kate MBBS, FANZCA; Sessler, Daniel I. MD; Schroeder, Marc BA; Walters, Kristin BA

Author Information
  • Free


Learning is a relatively long-lasting adaptive behavioral change occurring as a result of experience. Learning may be either explicit (accompanied by conscious recall), or implicit (occurring without conscious recall). Both types of learning during anesthesia may have untoward physical and psychologic sequelae [1].

Propofol impairs learning and produces sedation in low doses. It is advocated for sedation during regional anesthesia [2,3]. Although in sufficient concentrations it may produce respiratory and cardiovascular compromise [4] as well as unnecessarily deep hypnosis. Inadequate concentrations may be associated with intraoperative recall. More appropriate propofol blood concentrations can be targeted if the concentration-effect relationship for propofol and suppression of learning is better defined.

Alternatively, electroencephalography (EEG), which offers real-time monitoring of central nervous system activity, might predict learning during anesthesia. This would avoid the problem of predicting actual blood concentrations of a drug from those targeted. Recently, a new EEG indicator, the Bispectral Index, was proposed as a measure of anesthetic effect [5]. This index quantifies the nonlinear relationships between EEG component waves, as well as analyzing their frequency and amplitude. The Bispectral Index can predict movement after painful stimulation during anesthesia [5,6], but it is not known whether it predicts suppression of learning. We therefore tested the hypothesis that propofol blood concentration, the Bispectral Index, and the 95% spectral edge frequency predict suppression of learning during propofol/epidural anesthesia in volunteers.

The EEG usually exhibits a biphasic response to increasing anesthetic concentration; that is, there is activation, followed by slowing of the EEG. This phenomenon has confounded attempts to predict anesthetic effect using the EEG, as one EEG variable value may be associated with two different anesthetic concentrations [7,8]. The Bispectral Index (version 3.0) incorporates features designed to overcome this problem. We therefore tested the hypothesis that the Bispectral Index is linearly related to propofol blood concentration.


With approval from the Committee on Human Research at the University of California, San Francisco, we studied 14 healthy male volunteers. None reported taking prescription medications or illegal drugs. The volunteers' height was 181 +/- 7 cm (mean +/- SD), total body mass 75 +/- 11 kg, lean body mass 60 +/- 7 kg, and age 28 +/- 5 yr.

Volunteers were studied on three randomly ordered days: control (no propofol), a target propofol blood concentration of 1 micro gram/mL, and a target propofol blood concentration of 2 micro gram/mL. Volunteers were blinded to the target concentration. The propofol blood concentration preventing movement after a painful stimulus in 50% of subjects (Cp50) is >8 micro gram/mL [9,10]. After completion of memory testing each day, volunteers participated in a study evaluating the thermoregulatory effects of propofol [11]. Epidural anesthesia was induced, allowing volunteers' core temperatures to be manipulated without altering their sentient skin temperatures.


The volunteers fasted for 8 h before arriving at the laboratory; during studies, they rested supine on a standard operating-room table. The room lights were dimmed and extraneous noise was minimized.

An L2-3 epidural catheter and right internal jugular vein catheter were inserted using standard techniques. An antecubital vein in the nondominant arm was cannulated for fluid and drug administration. Routine anesthesia monitors were attached and were used to monitor standard anesthetic variables, including blood pressure, heart rate, and oxygen saturation.

Induction of epidural anesthesia was preceded by 1 h of cutaneous warming to prevent redistribution hypothermia [12], and we infused 15 mL/kg lactated Ringer's solution. Epidural anesthesia then was induced using 2% 2-chloroprocaine. A test dose of 3 mL with epinephrine 1:100,000 was followed by slow administration of 20 mL 2% 2-chloroprocaine without epinephrine, to produce a dermatomal level of sensory block near T11. Subsequently, the sensory level was maintained with an infusion of 2% 2-chloroprocaine administered at a rate approximately 20 mL/h.

Propofol was infused using a pump (Ohmeda 9000; Ohmeda, Steeton, England) controlled by a computer programmed to target propofol blood concentrations of 1 and 2 micro gram/mL. The pharmacokinetic data were derived from a previous study of propofol pharmacokinetics during mild hypothermia in young healthy volunteers [13]. Combined data from hypothermic and normothermic volunteers were used to program the pump, using the method of Plasma Drug Efflux [14]. Oxygen (4 L/min via nasal cannulae) was administered as necessary to maintain oxygen saturations >95%.

Fifteen minutes after commencement of the infusion, we administered the learning tasks. Headphones were placed over the volunteers' ears, and we confirmed that output volume was adequate. The volunteers were instructed to listen carefully to the tape. A Trivial Pursuit Registered Trademark-type question task, of 3 min duration, was administered at all three test concentrations.

Learning Task

The Trivial Pursuit Registered Trademark-type questions were based on a learning task devised by Chortkoff et al. [15]. They conducted a preliminary test in 100 students to establish a pool of Trivial Pursuit Registered Trademark-type questions of appropriate difficulty and interest (e.g., "How tall was the world's tallest hairdo?"). We selected three sets of 10 questions and presented them to our volunteers in random order on the three study days. During anesthesia, volunteers were presented with answers to six questions, repeated three times. After recovery from anesthesia, the 10 questions were presented in a fiveoption, multiple-choice format.


The EEG was recorded continuously during each study. Data analyzed were collected when the volunteers' eyes were closed. After preparation of the skin with Omniprep Trademark (D.O. Weaver and Co., Aurora, CO), gold cup electrodes were filled with conductive paste (EC2 Trademark; Grass Instrument Co., Quincy, MA) and attached to the skin with collodion adhesive. The electrodes were positioned at Fp1 and Fp2, with the reference electrode at Cz and the ground electrode at the mastoid. A microprocessor-based, four-channel monitor (B500; Aspect Medical Systems, Inc., Natick, MA) was used to collect raw EEG data which were recorded on digital tape for off-line analysis (version 3.0) by Aspect Medical Systems, Inc.

Raw EEG data were inspected for artifact by a blinded observer, and spurious data deleted. Both the power spectrum and the Bispectral Index were calculated continuously. Version 3.0 of the Bispectral Index combines improved artifact detection algorithms to correctly identify awake artifacts (eye blinks and movement, and baseline wander), and improved dynamic range in the light sedation range, and was applied prospectively to our data.

Core temperature was measured at the tympanic membrane (Mon-a-Therm Registered Trademark; Mallinckrodt Anesthesiology Products, Inc., St. Louis, MO). The aural probe was inserted until felt by the volunteer next to the membrane; appropriate placement was confirmed by gently rubbing the wire. The probe then was taped in position and the canal occluded with cotton wool. Only data collected when the volunteers' core temperatures were 35.0-38.0 degrees C were used in the analysis.

Central venous blood was sampled for measurement of propofol blood concentration just prior to memory testing, 5 min later, and at 15-min intervals thereafter. Three-milliliter samples were stored in heparinized tubes at 4 degrees C for up to 10 wk (propofol blood concentrations decrease less than 0.2% per week at 4 degrees C), and later analyzed using a high-pressure liquid chromatography assay, modified from the method of Plummer [16]. This assay is linear to at least 20 micro gram/mL and has a detection limit of 0.025 micro gram/mL and a coefficient of variation of 4.1% at 2 micro gram/mL.

Data Analysis

The percentage of control Trivial Pursuit Registered Trademark-type questions answered correctly over the 3 days was calculated for each volunteer, and then averaged for the group. This quantified the volunteers' ability to guess the correct answer. The number of presented questions correct then was corrected for control: The results were thus negative when the volunteers' answers to presented questions were worse than control. Equation 1 Negative results were set at zero. Nonlinear regression was used to fit a sigmoid Equation tothe data and generate Cp50 values for suppression of learning by propofol (JMP; SAS Institute, Cary, NC). The same analysis was used to compare the Bispectral Index and 95% spectral edge frequency with suppression of learning.

Linear regression, with residual analysis, was used to correlate propofol blood concentration and Bispectral Index or 95% spectral edge frequency. Step-wise linear regression was used to define the influence of propofol concentration and core temperature on the Bispectral Index. Results are presented as mean +/- SD, unless specified otherwise; P < 0.05 was considered statistically significant.


Induction and maintenance of propofol sedation was smooth in all cases; none of the volunteers required airway support. Epidural-induced dermatomal level for loss of temperature discrimination was T11 +/- 1 (range T9-12). One volunteer did not attend his third study day, and another's data was rejected because of poor EEG signal quality and artifact. Variation in propofol concentration over the time of memory testing was 0.47 +/- 0.48 micro gram/mL.

Propofol caused concentration-related impairment of learning, as measured by the Trivial Pursuit Registered Trademark-type question task. Control questions were answered with a frequency a little greater than random guessing (approximate equals 28%). In order to successfully fit the nonlinear regression function, all negative results were set at zero. The Cp50 for propofol-induced suppression of learning was 0.66 +/- 0.1 micro gram/mL Figure 1. The Bispectral Index decreased as learning was suppressed. When the volunteers' responses to 50% of questions were correct, the Bispectral Index was 91 +/- 1 Figure 2. In contrast, the 95% spectral edge frequency correlated poorly with suppression of learning Figure 3. It was not possible to fit a sigmoid curve to these data.

Figure 1:
Propofol blood concentration (micro gram/mL) versus correct responses to Trivial Pursuit Registered Trademark-type questions corrected for control (%). The propofol blood concentration suppressing learning by 50% was 0.66 +/- 0.1 micro gram/mL (nonlinear regression line with 95% confidence limits shown). All questions were answered correctly by all volunteers when no propofol was given.
Figure 2:
Bispectral Index versus correct responses to Trivial Pursuit Registered Trademark-type questions corrected for control (%). The Bispectral Index value when learning was suppressed by 50% was 91 +/- 1 (nonlinear regression line with 95% confidence limits shown).
Figure 3:
95% spectral edge frequency (SEF) (Hz) versus correct responses to Trivial Pursuit Registered Trademark-type questions corrected for control (%). There was no relationship between the 95% spectral edge frequency and suppression of learning.

The Bispectral Index was related to propofol blood concentration by the equation: Bispectral Index = -7.4 centered dot [propofol] + 90 (r2 = 0.47, P = 0.0001, n = 278) Figure 4. In contrast, there was no significant relation between the 95% spectral edge frequency and propofol blood concentration (r2 = 0.094) Figure 5.

Figure 4:
Propofol blood concentration (micro gram/mL) versus Bispectral Index (linear regression line with 95% confidence limits shown). The Bispectral Index was linearly related to propofol blood concentration (Bispectral Index = -7.4 centered dot [propofol] + 90; r2 = 0.47, n = 278).
Figure 5:
Propofol blood concentration (micro gram/mL) versus 95% spectral edge frequency (SEF) (Hz) (linear regression line with 95% confidence limits shown). There was no relationship between propofol blood concentration and 95% spectral edge frequency (r2 = 0.094).

Core temperatures between 35 degrees C and 38 degrees C had an insignificant effect on the Bispectral index (Bispectral Index = 45 centered dot [propofol] + 7.7 centered dot core temperature - 1.4 centered dot ([propofol] centered dot core temperature) - 193).


Propofol produced a concentration-related suppression of learning during epidural anesthesia in volunteers, and therefore may be suitable for suppressing memory of perioperative events. The Bispectral Index correlates with both suppression of learning and propofol blood concentration and thus may be suitable for monitoring anesthetic effect during propofol sedation.

Several other studies have investigated the memory effects of propofol. Veselis et al. [17] administered constant-rate infusions to volunteers, achieving propofol blood concentrations of 0.86 +/- 0.45 micro gram/mL (maximum: 1.10 +/- 0.39 micro gram/mL). Although the volunteers' responses varied substantially, propofol blood concentration predicted impairment of immediate recall and delayed recognition of test material. Other studies, in which no propofol blood concentrations were reported, confirmed a dose-dependent effect on delayed recall, apparently due to impaired storage and consolidation of new material, rather than inadequate retrieval [18,19]. Our study confirms and extends these findings by establishing the concentration-effect relationship for propofol over a clinically relevant range, during steady-state conditions.

Our data support reports that learning is suppressed at propofol blood concentrations insufficient to produce unresponsiveness [17,20]. The Cp50-awake of propofol (blood concentration preventing response to verbal command in 50% of subjects) is approximate equals 3 micro gram/mL when administered alone during steady-state conditions [9,21]. The eyelash reflex is lost at similar propofol concentrations [22]. Therefore, loss of responsiveness, or of the eyelash reflex, will ensure absence of recall, albeit at the cost of greater respiratory depression. However, unparalyzed patients will still be able to move in response to painful stimulation, as the Cp50 (minimum alveolar anesthetic concentration analog) of propofol is >8 micro gram/mL [9,10].

To ensure that propofol blood concentrations equaled effect-site (brain) concentrations, constant blood propofol concentrations were required during administration of the learning tasks. We attempted to minimize variation in drug concentration by including volunteers of similar age and body morphology, and using lean body mass to calculate the dose [23]. Pharmacokinetic data used to program the infusion were derived from a similar group of volunteers [13]. We believe that the constancy of propofol concentration we achieved was adequate to allow definition of the memory effects of propofol. Hysteresis between blood and effect-site concentrations, was minimized by allowing at least 15 min (>three times the half-time for equilibration with the effect compartment*) for equilibration before testing.

**Schuttler J, Schwilden H, Stoekel H. Pharmacokinetic-dynamic modeling of Diprivan [abstract]. Anesthesiology 1986;65:A549.

Pharmacodynamic variation was more difficult to control. Learning may be facilitated by autonomic and endocrine responses [24]. We tried to control ambient light and noise, and minimize distractions in the laboratory, but inevitably some volunteers were more aroused than others. The volunteers' motivation to remember answers likely also was influenced by differences in personality and experience, particularly familiarity with medical environments. As perioperative events are even more stressful, caution should be used when extrapolating our results to the clinical setting.

The Trivial Pursuit Registered Trademark-type question task was easy to use for testing learning during anesthesia. Complete learning of test material in all volunteers when no propofol was given confirms that this material was easy to learn and retain [25]. However, the task was not sufficiently sensitive to allow the definition of propofol effect, as each incorrect answer decreased the percentage of questions answered correctly by 23%. Using more questions, however, may not have improved our results as there is a limit to the amount of information that a volunteer can remember [1].

As an alternative to targeting specific anesthetic concentrations, a monitor of pharmacodynamic effect may be applied. The EEG is suitable as it offers noninvasive, real-time monitoring of central nervous system activity. However, previous attempts to correlate EEG changes during anesthesia with memory have met with mixed success [17,26-28]. Use of drug combinations and inadequate control of drug concentration in some of these studies may have increased the variability observed in the EEG measures.

In this study, we compared the 95% spectral edge frequency, an example of a conventional processed-EEG variable, with the Bispectral Index, and demonstrated that the Bispectral Index was superior to the 95% spectral edge frequency in predicting suppression of learning in our volunteers. We postulate that this is because conventional processed-EEG variables exhibit a biphasic response to increasing propofol concentration [29], whereas the Bispectral Index does not.

We have demonstrated that the Bispectral Index (version 3.0) is monotonically related to propofol blood concentration over the sedative range. In contrast, it is not possible to predict propofol concentration using the 95% spectral edge frequency. The Bispectral Index therefore may be used to measure anesthetic effect during propofol sedation. The reliability of the Bispectral Index as a measure of the anesthetic effect of other drugs awaits further study.

Regression analysis assumed independent data. Repeated testing in each volunteer, although economical, may result in intravolunteer correlation that may violate this assumption. In the absence of a comparable statistic with which to analyze these data, we have accepted, as have others [30], the potential error introduced into our results. Despite these problems, the functions derived appear to adequately describe our data.

In conclusion, propofol caused a concentration-dependent impairment of learning. Propofol is thus suitable for conscious sedation during regional anesthesia. The Bispectral Index was linearly related propofol blood concentration, and predicted suppression of learning.

We would like to acknowledge the advice of Jeff Sigl, PhD. We thank Warren D. Smith, PhD, for assistance in statistical analyses; Ben S. Chortkoff, MD, for advice about memory testing during anesthesia; Andrew R. Bjorksten, PhD, for devising the propofol infusion protocol and performing the propofol assays; and Charles Hackman, FANZCA, (assisted by an Australian and New Zealand College of Anaesthetists Research Project Grant), who programmed the drug-infusion pump. We would like to thank Ohmeda, Inc. for loan of a Modulus Registered Trademark CD-integrated anesthesia machine and infusion pump, and Hermes Systems, Inc, for loan of an IdaCare Trademark automatic record-keeping system.


1. Ghoneim M, Block R. Learning and consciousness during general anesthesia. Anesthesiology 1992;76:279-305.
2. Knell PJW, McKean JF. An investigation of the pharmacokinetic profile of propofol ('Diprivan') after administration for induction and maintenance of anaesthesia by repeat bolus doses in patients having spinal anaesthesic block. Postgrad Med J 1985;61(Suppl 3):60-1.
3. MacKenzie N, Grant I. Propofol for intravenous sedation. Anaesthesia 1987;423-6.
4. Sebel PS, Lowdon JD. Propofol: a new intravenous anesthetic. Anesthesiology 1989;71:260-77.
5. Kearse L, 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.
6. Vernon J, Lang E, Sebel P, et al. Prediction of movement using bispectral electroencephalographic analysis during propofol/alfentanil or isoflurane/alfentanil anesthesia. Anesth Analg 1995;80:780-5.
7. Buhrer M, Maitre P, Hung O, et al. Thiopental pharmacodynamics. 1. Defining the pseudo-steady state serum concentration EEG relationship. Anesthesiology 1992;77:226-36.
8. Hung O, Varvel J, Shafer S, et al. Thiopental pharmacodynamics. 2. Quantitation of clinical and electroencephalographic depth of anesthesia. Anesthesiology 1992;77:237-44.
9. Smith C, McEwan A, Jhavieri R, et al. Interaction of fentanyl on the Cp50 of propofol for loss of consciousness and skin incision. Anesthesiology 1994;81:820-8.
10. Davidson J, Macleod A, Howie J, et al. Effective concentration 50 for propofol with and without 67% nitrous oxide. Acta Anaesthesiol Scand 1993;37:458-64.
11. Leslie K, Sessler D, Bjorksten A, et al. Propofol causes a dose-dependent decrease in the thermoregulatory threshold for vasoconstriction but has little effect on sweating. Anesthesiology 1994;81:353-60.
12. Hynson J, Sessler D, Moayeri A, et al. The effects of preinduction warming on temperature and blood pressure during propofol/nitrous oxide anesthesia. Anesthesiology 1993;79:219-28.
13. Leslie K, Sessler D, Bjorksten A, et al. Mild hypothermia alters propofol pharmacokinetics and increases the duration of action of atracurium. Anesth Analg 1995;80:1007-14.
14. Crankshaw D, Boyd D, Bjorksten AR. Plasma Drug Efflux--a new approach to optimization of drug infusion for constant blood concentration of thiopental and methohexital. Anesthesiology 1987;67:32-41.
15. Chortkoff B, Bennett H, Eger E. Does nitrous oxide antagonize isoflurane-induced suppression of learning? Anesthesiology 1993;79:724-32.
16. Plummer G. Improved method for the determination of propofol in blood by high performance liquid chromatography with fluorescence detection. J Chromatogr 1987;421:171-6.
17. Veselis R, Reinsel R, Wronski M, et al. EEG and memory effects of low-dose infusions of propofol. Br J Anaesth 1992;69:246-54.
18. Zacny JP, Lichtor JL, Coalson DW, et al. Subjective and psychomotor effects of subanaesthetic doses of propofol in healthy volunteers. Anesthesiology 1992;76:696-702.
19. Polster M, Gray P, O'Sullivan G, et al. Comparison of the sedative and amnesic effects of midazolam and propofol. Br J Anaesth 1993;70:612-6.
20. Rupreht J. Awareness with amnesia during total intravenous anaesthesia with propofol [letter]. Anaesthesia 1989;44:1005.
21. Chortkoff B, Eger EI, Crankshaw D, et al. Concentrations of desflurane and propofol that suppress response to command in human volunteers. Anesth Analg 1995;81:737-43.
22. Vuyk J, Engbers FH, Lemmens HJ, et al. Pharmacodynamics of propofol in female patients. Anesthesiology 1992;77:3-9.
23. Leslie K, Crankshaw DP. Lean tissue mass is a useful predictor of induction dose requirements for propofol. Anaesth Intensive Care 1991;19:57-60.
24. Weinberger N, Gold P, Sternberg D. Epinephrine enables pavlovian fear conditionin under anaesthesia. Science 1984;223:605-6.
25. Dwyer R, Bennett H, Eger E, et al. Isoflurane anesthesia prevents unconscious learning. Anesth Analg 1992;75:107-12.
26. Levy W. Power spectrum correlates of changes in consciousness during anesthetic induction with enflurane. Anesthesiology 1986;64:688-93.
27. Veselis R, Reinsel R, Alagesan R, et al. The EEG as a monitor of midazolam amnesia: changes in power and topographu as a function of amnesic state. Anesthesiology 1991;74:866-874.
28. Dwyer R, Rampil I, Eger EI, et al. The electroencephalogram does not predict depth of isoflurane anesthesia. Anesthesiology 1994;81:403-9.
29. Seifert H, Blouin R, Conard P, et al. Sedative doses of propofol increase beta activity of the processed electroencephalogram. Anesth Analg 1993;76:976-8.
30. Chortkoff B, Bennett H, Eger E. Subanesthetic concentrations of isoflurane suppress learning as defined by the category-example task. Anesthesiology 1993;79:16-22.
© 1995 International Anesthesia Research Society