You are on page 1of 7

TECHNOLOGY, COMPUTING

,
SECTION EDITOR JEFFREY M. FELDMAN

AND

SIMULATION

SOCIETY

FOR

TECHNOLOGY

IN

ANESTHESIA

Spectral Entropy and Bispectral Index as Measures of the Electroencephalographic Effects of Propofol
Richard Klaus Ellerkmann, MD*, Martin Soehle, MD*, Thorsten Michael Alves*, Vidal-Markus Liermann*, Ingobert Wenningmann, MD*, Heiko Roepcke, MD*, ¨ rgen Bruhn, MD* Sascha Kreuer, MD†, Andreas Hoeft, MD, PhD*, and Jo
*Department of Anesthesiology and Intensive Care Medicine, University of Bonn; and †Department of Anesthesiology and Intensive Care Medicine, University of Saarland, Homburg/Saar, Germany

Recently, Datex-Ohmeda introduced the Entropy ModuleTM for measuring depth of anesthesia. Based on the Shannon entropy of the electroencephalogram, state entropy (SE) and response entropy (RE) are computed. We investigated the dose-response relationship of SE and RE during propofol anesthesia in comparison with the Bispectral IndexTM (BIS). Twenty patients were studied without surgical stimulus. Anesthesia was induced by a constant propofol infusion of 2000 mg/h (451 Ϯ 77 ␮g ⅐ minϪ1 ⅐ kgϪ1) via a large forearm vein. Propofol was infused until substantial burst suppression occurred (more than 50%) or mean arterial blood pressure decreased to Ͻ60 mm Hg. Hereafter, infusions were stopped until recovery of BIS values up to 60 was reached. Subsequently, the constant propofol infusion of 2000 mg/h was restarted to increase depth

of anesthesia and again decreased (infusion was stopped) within the BIS value range of 40 – 60. The coefficient of determination (R2) and the prediction probability (PK) were calculated to evaluate the performance of SE, RE, and BIS to predict changing propofol effectsite concentrations. R2 values for SE, RE, and BIS of 0.88 Ϯ 0.08, 0.89 Ϯ 0.07, and 0.92 Ϯ 0.06, respectively, were similar. The calculated PK values, however, revealed a significant difference between SE and RE compared with BIS, with PK ϭ 0.77 Ϯ 0.09, 0.76 Ϯ 0.10, and 0.84 Ϯ 0.06, respectively. BIS seems to show slight advantages in predicting propofol effect-site concentrations compared with SE and RE, as measured by PK but not as measured by R2. (Anesth Analg 2006;102:1456 –62)

M

onitors intraoperatively analyzing the electroencephalogram (EEG) are used to measure anesthetic drug effect on the central nervous system. Because analyzing the raw EEG signal during anesthesia at real-time is difficult, several EEG monitors have been developed to extract and process EEG information and to present the content in a continuous index from 0 to 100. Zero represents the deepest level of anesthesia (isoelectric EEG line) and 100 the awake state of a patient. The use of such EEG monitors can decrease drug consumption during anesthesia (1,2) and lead to a faster recovery from anesthesia (1,3). The use of the Bispectral IndexTM (BIS) monitor (Aspect Medical Systems, Newton, MA) may decrease the incidence of intraoperative awareness (4,5).

Accepted for publication December 21, 2005. Address correspondence and reprint requests to Richard Klaus Ellerkmann, MD, Sigmund Freud St. 25, 53105 Bonn, Germany. Address e-mail to richard.ellerkmann@ukb.uni-bonn.de. DOI: 10.1213/01.ane.0000204289.47792.56

In this study, we investigated the dose-response relationship of the new Entropy ModuleTM (DatexOhmeda, Helsinki, Finland) during propofol anesthesia in comparison with the BIS monitor. Whereas the BIS monitor uses different algorithms to calculate the BIS during the different stages of anesthesia, e.g., burst suppression (BS) (6) and frequency power calculation (7), as well as bispectral analysis (8), the Entropy ModuleTM measures depth of anesthesia with a single algorithm, i.e., calculating the Shannon Entropy (9) of the power spectrum called the Spectral Entropy. The Entropy ModuleTM calculates two different Spectral Entropy indicators: the state entropy (SE), computed over the frequency range from 0.8 to 32 Hz, reflecting the EEG-dominant part of the spectrum, in addition to the response entropy (RE), computed over the frequency range of 0.8 to 47 Hz, including both the EEG and electromyographic (EMG) dominant part of the recorded spectrum (10). The aim of our study was to investigate the ability of the EEG monitors to differentiate between different effect-site concentrations of propofol calculated by the prediction probability. In addition the correlation
©2006 by the International Anesthesia Research Society 0003-2999/06

1456

Anesth Analg 2006;102:1456–62

11 men and 9 women) scheduled for minor surgery with general anesthesia. and the BIS (BIS-XP sensor. Ceff the effect-site concentration. No fluids were given before the induction of anesthesia. The smoothing time period for the BISTM monitor was set to 15 s. as shown exemplary for a patient in Figure 1. C50 no BS is the propofol concentration associated with a 50% decrease from E0 to Eplateau. After arrival in the induction room. respectively. (11) using the parameter set published by Marsh et al. All participants were ASA physical state I or II. C50 BS is the propofol concentration associated with 50% decrease from Eplateau to Emax. A ke0 value of 0. Cplateau is the propofol concentration at Eplateau.ANESTH ANALG 2006. Methods After IRB approval. and ␭no BS is the steepness of that relationship. WA). and standard monitors were applied. 23– 64 yr. an IV catheter was inserted into a larger forearm vein. as recommended by the manufacturers. Entropy values (SE and RE) were recorded with the Datex-Ohmeda software S/5TM Collect (version 4. The bi-sigmoidal . To verify if this model is also adequate to describe the relationship between propofol effect-site concentrations and Spectral Entropy. Redmond. SPECTRAL ENTROPY AND BIS MEASURING PROPOFOL’S ELECTROENCEPHALOGRAPHIC EFFECT 1457 among the three EEG variables (SE. Aspect Medical Systems) and the Entropy electrodes were positioned. The EEG was recorded continuously using an Aspect A-2000 BISTM monitor (version XP) and the Entropy ModuleTM (Datex-Ohmeda). The infusion was stopped hereafter until BIS recovered to values of 60. The first curve (Equation 2a) reaches from E0. the presumed maximum drug effect. (12). AND SIMULATION ELLERKMANN ET AL. (15) have shown that this bi-sigmoidal model adequately fits the correlation between effectsite concentrations and BIS. to Eplateau. COMPUTING. an effect site was introduced into the model: dCeff ⁄ dt ϭ (Cpl Ϫ Ceff) · ke0 (1) Cpl is the plasma concentration. BIS values were recorded and transferred to computer hard disk using the software program HyperTerminal (Microsoft.5 k⍀. and ␭BS is the steepness of that relationship. Study data were automatically recorded in intervals of 5 s and transferred to a computer hard disk in real time for further offline analysis. The propofol infusion was continued until substantial BS occurred (50% and more) or mean arterial blood pressure decreased to Ͻ60 mm Hg. as published by Kreuer et al.456 minϪ1 (16) was chosen to calculate propofol effect-site concentrations obtained by Equation 1. RE. the EEG value in the absence of propofol. and BIS). Subsequently. Impedance was automatically considered adequately low by the BIS monitor and the Entropy Module if kept to less than 10 k⍀ and 7. The skin of the forehead was prepared with 70% isopropanol. which extends from Eplateau to Emax. Propofol effect-site concentrations were obtained by simultaneous pharmacokinetic and pharmacodynamic modeling (13). The relationship between estimated effect-site concentrations and EEG variable values was modeled by a variation of the classical Emax model (14) with two linked sigmoidal curves describing the EEG effect of propofol with (BS) and without BS (no BS). RE. Kreuer et al. Propofol plasma concentrations were calculated within Excel as described by Bruhn et al. Exclusion criteria were a history of any disabling central nervous or cerebrovascular disease or patients who had received central nervous system active drugs. Spectral Entropy also revealed a pronounced plateau leading to a biphasic dose-response curve. All patients were spontaneously breathing throughout the experiment through a face mask delivering 100% O2. Ceff is the apparent effect-site concentration.0) onto the computer hard disk. Anesthesia was induced by a continuous 2% propofol infusion of 2000 mg/h (451 Ϯ 77 ␮g ⅐ minϪ1 ⅐ kgϪ1). where a transition occurs to the second curve (Equation 2b). Measurements were then stopped. and ke0 is the first-order rate constant determining the efflux from the effect site. and BIS) and estimated effect-site concentrations of propofol was investigated by simultaneous pharmacokinetic and pharmacodynamic modeling. written informed consent was obtained from 20 patients aged 43 Ϯ 12 yr (range. onto the temporal-frontal area of the patient’s forehead. To eliminate the hysteresis between plasma concentrations of propofol and the EEG variable values (SE. (15): For Ceff Յ Cplateau: E ϭ E0 ϩ (Eplateau Ϫ E0) * ͓Ceff ⁄ C50 no BS͔␭ no BS ͓1 ϩ (Ceff ⁄ C50 no BS)␭ no BS͔ (2a) For Ceff Ͼ Cplateau: E ϭ Eplateau ϩ ͑Emax Ϫ Eplateau͒* ͓͑Ceff Ϫ Cplateau͒/͑C50 BS Ϫ Cplateau)]␭ BS (1 ϩ ͓͑ Ceff Ϫ Cplateau)/(C50 BS Ϫ Cplateau)]␭BS) (2b) Both sigmoidal curves have their own variables. and the patient’s trachea was intubated for surgery. Measurements were started after the electrode impedance check of each monitor was completed. we visually investigated the dose-response curve without any underlying model.102:1456–62 TECHNOLOGY. depth of anesthesia was again increased by a propofol infusion at 2000 mg/h and then decreased by stopping the propofol infusion in a way that targeted BIS values ranged from 40 to 60.

In a second step. and Ptx are the respective probabilities that two data points drawn at random. the PK value describes the probability that the EEG variable correctly predicts which of the data points is the one with the larger (or smaller) anesthetic drug concentration. the PK value is independent of scale units and does not require knowledge of underlying distributions or efforts to where Pc. maximizing R2 is equivalent to minimizing SSE. In the remaining 10 patients.. RE. the actual PK value we measure is 1-PK. and BIS was investigated with the model-independent prediction probability (PK) (18). and with replacement from the population are a concordance. a discordance.. PK can be computed for any degree of coarseness or fineness of the scales. The effect-site concentration was obtained modelindependently by incorporating a ke0 value of 0. Because propofol concentration increases as BIS. we estimated a standard deviation of 0. increasing propofol concentrations led to substantial BS (Ͼ50%). Results In 10 patients. A PK value of 1 means that the values of the predicting variable (SE. All tests were two tailed. it is equivalent to nonlinear regression with ordinary least squares. The PK values were calculated on a spreadsheet using the Excel 2000 software program and the PKMACRO written by Warren Smith (18). RE. (A) Correlation between state entropy (SE) and propofol plasma concentration for a patient in whom an increasing propofol plasma concentration leads to burst suppression. or a x-only tie. The relationship between SE and changing propofol concentrations in patients with the best and worst fit (R2 value) is displayed in Figure 3 together with the corresponding fit for the same patients for BIS. Pd. A power analysis was performed using PK as the target variable. Furthermore. independently. i. Based on a previous study (20).05 (19).1458 TECHNOLOGY.5 means that the values of the indicator predict no better than by chance only. We chose the coefficient of determination (17) (R2) as an objective function: R2 ϭ 1 Ϫ SSE SST ˆ) ͚ (y Ϫ y i i n 2 ϭ1Ϫ iϭ1 n (3) i i 2 ៮) ͚ (y Ϫ y iϭ1 SSE is the sum of squared errors and represents the sum of squares of the differences between observed measurements y ˆ i for a given time and the corresponding model prediction. To detect a difference in PK between BIS and Entropy in a paired study design with a significance level of 5% (␣ ϭ 0. The sample size calculation was aimed to show a clinically relevant difference in PK of 0. SPECTRAL ENTROPY AND BIS MEASURING PROPOFOL’S ELECTROENCEPHALOGRAPHIC EFFECT ANESTH ANALG 2006.05 for PK.e. and SE decrease. (B) Correlation between SE and propofol effect-site concentration for the same patient.05. RE. or BIS) and the predicted drug effect. A PK value of 0. as shown exemplary for one patient in Figure 2. AND SIMULATION ELLERKMANN ET AL.18). Thus. As a nonparametric measure. Because SST is independent of the model variables. The computations were performed on a spreadsheet using the Excel software program. estimated propofol effect-site concentration).g. propofol infusion was stopped before substantial BS occurred and after mean arterial blood pressure had decreased to less than 60 mm Hg. with statistical significance defined as P Ͻ 0. Calculated effect-site concentrations were adequately fitted using a bi-sigmoidal Emax model (Equation 2a and 2b). PK has been defined as: PK ϭ Figure 1. Our aim was to maximize the correlation between the measured drug effect (SE.05) and a power of 82% (␤ ϭ 0.102:1456–62 linearize or to otherwise transform scales. A single sigmoidal curve (Equation 2a) was sufficient to fit the data for these patients. data are presented as mean Ϯ sd.5 * Ptx) (Pc ϩ Pd ϩ Ptx) model was therefore used to fit the data of BIS and Spectral Entropy. Statistical calculations were performed by Student’s t-test or Wilcoxon test where appropriate. the correlation between effect-site concentrations and SE. The relationship between SE and changing propofol concentrations in patients with the best and worst fit (R2 value) is displayed in Figure 4 . Given two randomly selected data points with distinct anesthetic drug concentrations. or BIS value) always correctly predict the value of the variable to be predicted (e. SST is the total sum of squares and represents the sum of squares of the differences between each actual measurement and the average of all the measurements (y ៮ i). Variables were optimized with the Solver tool within Excel using nonlinear regression with ordinary least squares. PK fully uses the available data without imposing additional arbitrary constraints.456 minϪ1 into Equation 1. COMPUTING. RE. (Pc ϩ 0. at least 14 patients had to be included.

Each symbol represents an EEG variable of a 5-s epoch. 2) and (B) the worst fit for SE. (21).07).84) for the bi-sigmoidal model (P ϭ 0. A single sigmoidal curve (Equation 2a) was sufficient to fit the data. (C and D) The relationship between bispectral index (BIS) and propofol plasma concentrations (Cpl) and propofol effect-site concentrations (Ceff) for the same two different patients as in (A and B). Table 1). and RE (0. together with the corresponding fit for the same patients for BIS.10) (P ϭ 0 0.18 Ϯ 1. COMPUTING. Table 1). No significant differences between SE and RE were detected for the fitting variables. (C and D) The relationship between bispectral index (BIS) and propofol plasma concentrations (Cpl) and propofol effect-site concentrations (Ceff) for the same two different patients as in (A) and (B).88 Ϯ 0. We therefore validated our R2 values by refitting our data using the pharmacokinetic parameter set published by Schnider et al.102:1456–62 TECHNOLOGY. SE (0.89 Ϯ 0. Figure 3. One might criticize that R2 values were calculated on the basis of the estimated propofol plasma concentration using the Marsh et al. R2 ϭ 0.09) and RE (0. and PK values. bold line) is superimposed over the data points of each figure. (A and B) The relationship between state entropy (SE) and propofol plasma concentrations (Cpl) and propofol effect-site concentrations (Ceff) for two different patients in whom burst suppression was not reached.06).72 Ϯ 1. The bi-sigmoidal fit (Equation 2a and 2b. and changing propofol effect-site concentrations (Ceff) of a patient in whom BS took place. A plateau was reached at smaller drug concentrations with Cplateau 3. as judged by the PK values. BIS showed a significant advantage in discriminating between changing propofol effect-site concentrations in comparison with SE and RE. (A) Typical example of the time courses of the measured electroencephalographic (EEG) variables state entropy (SE).01 for BIS versus SE). (A and B) The relationship between state entropy (SE) and propofol plasma concentrations (Cpl) and propofol effect-site concentrations (Ceff) for two different patients in whom burst suppression was reached. The correlation of SE and RE could be best described by a linear function (RE ϭ 1. (12) pharmacokinetic parameter set.26 ␮g/mL for BIS (P ϭ 0.42) compared to BIS (2. Discussion In this study. This significant difference was not mirrored by a significant difference in R2 values of BIS versus SE and RE. calculated R2 values remained unchanged. Each symbol represents the electroencephalographic (EEG) variable of a 5-s epoch. Each symbol represents the electroencephalographic (EEG) variable of a 5-s epoch. Calculated R2 values describing the correlation of the EEG variables and the calculated propofol effect-site concentrations were comparable between BIS (0.ANESTH ANALG 2006. Figure 4. burst suppression ratio (BS).84 Ϯ 0.65 Ϯ 2. (A) represents the best fit (same patient as in Fig. This parameter set is slightly imprecise during the first few minutes in estimating the propofol plasma concentration. The fitted curve is superimposed over the data points of each figure (bold line). bispectral index (BIS).025.92 Ϯ 0. (A) represents the best fit and (B) the worst fit for SE. However.985). SPECTRAL ENTROPY AND BIS MEASURING PROPOFOL’S ELECTROENCEPHALOGRAPHIC EFFECT 1459 Figure 2.03.08).77 Ϯ 0. All individual fits and their corresponding data points of all 20 patients are shown in Figure 5 for SE and BIS. .08 ϫ SE.76 Ϯ 0. AND SIMULATION ELLERKMANN ET AL. Calculated PK values proved to be significantly better for BIS (0. We therefore showed the performance of SE compared with BIS in the figures and added all results of RE into Table 1. response entropy (RE).52 ␮g/mL for SE versus 5.06) compared with SE (0. Pharmacokinetic variables revealed a significantly steeper slope factor ␭no BS for SE (4.68 Ϯ 0. R2 values.

Fitting the data with a bi-sigmoidal curve adequately described this biphasic response.31 Ϯ 1.60 Ϯ 1. Equation 2a) 91.34 Ϯ 0.00 0.06 86. (24) investigated the correlation of propofol with SE and BIS and observed higher PK values for BIS and a significantly higher Spearman rank correlation for BIS compared with SE.84) for the bisigmoidal model and a steeper ␭no BS for SE (4. BIS ϭ bispectral index.61 Ϯ 2. . and BIS) in the absence of the drug. ␭ BS ϭ the steepness of the concentration-response relationship between Eplateau and Emax.41 Ϯ 15.80 4. as evident for SE and RE) followed by a pronounced plateau (expressed by lower Cplateau values for SE and RE).73 2. however.76 1.26 3.18 Ϯ 1.50 23. revealing a significantly steeper slope factor ␭no BS for SE (4.09 Ϯ 3. Previously. RE.70 Ϯ 0.33 Ϯ 16.42) compared to BIS (2. SPECTRAL ENTROPY AND BIS MEASURING PROPOFOL’S ELECTROENCEPHALOGRAPHIC EFFECT ANESTH ANALG 2006.75 Ϯ 0.001 The bi-sigmoidal model was applied to fit the data when burst suppression (BS) occurred within a measurement (n ϭ 10).13 1.80) for the monosigmoidal model. If increasing propofol concentrations lead to an initial steep decrease of the EEG index value (higher slope factor ␭no BS. leading to a less graded response of the EEG variable and thereby loosing discrimination compared with BIS. and BIS (20). be explained by the fitting algorithm for R2.09 0.81 1. E0 ϭ measured encephalographic variable value (SE.10 3.78 6.15 0.11 Ϯ 5.28 5.102:1456–62 Table 1. Pharmacokinetic and Pharmacodynamic Variables Variable E0 Emax C50 no BS (␮g/mL) ␭no BS Eplateau Cplateau (␮g/mL) C50 BS (␮g/mL) ␭BS ke0 (minϪ1) E0 Eplateau C50 no BS (␮g/mL) ␭no BS ke0 (minϪ1) BIS Ϯ SD SE Ϯ SD RE Ϯ SD t-test BIS-SE Bi-sigmoidal model (n ϭ 10.73 Ϯ 1. we could show a close correlation of SE and RE with changing sevoflurane effect-site concentrations and equally high PK values among SE. Figure 5.16 Ϯ 0.025 0.17 87.07 0.17 25.38 Ϯ 0.47 97.00 6.48 Ϯ 0.47 Ϯ 1.61 34. Vanluchene et al. SE ϭ state entropy. This plateau is broader for the Entropy variables compared with BIS.51 95. RE.56 0.96 Ϯ 15. ke0 ϭ first-order rate constant determining the efflux from the effect site.46 6.1460 TECHNOLOGY. In the same study. investigated only increasing propofol concentrations. Equation 2a and 2b) 96.61 Ϯ 2. Cplateau ϭ propofol concentration at Eplateau. C50 BS ϭ propofol concentration associated with 50% decrease from Eplateau to Emax. sevoflurane (22). the same authors describe a less smooth decrease of SE and RE after increasing propofol concentrations compared with BIS.65 Ϯ 0.89 Ϯ 1.71 31. the authors report a steeper slope factor ␭ for SE and RE.72 Ϯ 1.64 Ϯ 1.79 Ϯ 8. Our results confirmed their findings. in part. The mono-sigmoidal model was applied to fit the data when BS was not reached within a measurement (n ϭ 10).68 Ϯ 0. RE ϭ response entropy. The relationship of (A) state entropy (SE) and (B) bispectral index (BIS) and the calculated propofol effect-site concentrations for all patients (n ϭ 20). Eplateau ϭ index value where a transition occurs to the second curve. Emax ϭ encephalographic variable value corresponding to maximum drug effect.33 Ϯ 3. Corresponding individual fits for each patient for (C) SE and (D) BIS are presented.30 Ϯ 7.37 0.28 Ϯ 0.68 Ϯ 0.04 Ϯ 2.52 3.54 Ϯ 1.70 14. In a more recent study (25). The discrepancy between R2 and PK values can.59 1.74 0.06 Ϯ 2.66 6.72 2.65 Ϯ 2.45 Ϯ 1. AND SIMULATION ELLERKMANN ET AL. then calculated R2 values can be high because the R2 values only describe the degree of concordance of the measured data with the modelpredicted data.86 Ϯ 1.86 Ϯ 3. and it was our intention to evaluate the monitors for increasing and decreasing propofol effect-site concentrations.00 1.00 0.32 Ϯ 0.04 Ϯ 12. the calculated PK will be low because the monitor does not distinguish between increasing drug concentrations where a plateau is evident. ␭no BS ϭ steepness of the concentration-response relationship between E0 and Eplateau.84 4.38 Ϯ 0. COMPUTING.42 5.56 0.03 Ͻ0.65 Ϯ 2.11 Monosigmoidal model (n ϭ 10. as indicated by the early onset of the plateau described by lower Cplateau values (Table 1). These studies.94 7. However.00 19. We showed that increasing propofol concentrations leading to BS exert a biphasic response in the BIS and Entropy Index.42 Ϯ 0.80 4.04 Ϯ 0. This phenomenon has been described for isoflurane (15).80) compared to BIS (3. C50 no BS ϭ propofol concentration associated with 50% decrease from E0 to Eplateau.54 Ϯ 1.14 0.17 Ϯ 3. and desflurane (23) between BIS and Narcotrend Index versus drug effect.78 5.

In: Linear statistical models (an applied approach).99: 34–41. Dose-response relationship between sevoflurane concentrations and Narcotrend and bispectral index. Anesthesiology 1999.363:1757–63. Minto CF. 16. A simple Excel spreadsheet approach to predict i. Application of bispectral index and narcotrend index to the measurement of the electroencephalographic effects of isoflurane with and without burst suppression. Clin Pharmacol Ther 1979. 8. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Dutton RC. Ellerkmann RK. 3.ANESTH ANALG 2006. The possible effect of aggregation of the molecules of hemoglobin on its dissociation curves. However. and RE responds faster to changing EEG and EMG signals because of shorter time windows for signal interpretation. White M. 13. J Clin Monit Comput 2000. et al. as measured by the PK. As a time constant. Anesthesiology 2004.43:545–9. The influence of age on propofol pharmacodynamics. Narcotrend monitoring allows faster emergence and a reduction of drug consumption in propofol-remifentanil anesthesia. Spectral entropy and bispectral index as measures of the electroencephalographic effects of sevoflurane. and optimized time-frequency windows ranging from 1. et al. Larsen R. 18. McNeil J. . Yli-Hankala A. 4.2 and 1. Bouillon TW. Anesthesiology 1996. 19. 2 seconds for the canonical univariate variable (29). revealing information about the efflux of propofol from the effect site. Simultaneous modeling of pharmacokinetics and pharmacodynamics: application to d-tubocurarine. Ekman A. However.88: 1170–82. Vakkuri A. Marsh B. 22. et al. Smith NT. 7. Smith WD. 2. A primer for EEG signal processing in anesthesia. Sheiner LB. et al. Acta Anaesthesiol Scand 2004.102:1456–62 TECHNOLOGY.48:154–61. 11. Ana ¨ sthesiologie und Intensivmedizin 2002. in part.101:847–54. Shafer SL. Shafer SL. Hill AV. Annila P. Leslie K. Lennmarken C. 12. Hoeft A. Minto CF. Maja V. The different ke0 values estimated for SE and BIS can. but also about how long the monitor requires to detect the drug effect. Description of the Entropy algorithm as applied in the Datex-Ohmeda S/5 Entropy Module. Schnider TW. 1990:174–83. J Clin Anesth 2000.43:708–10. Pharmacokinetic model driven infusion of propofol in children. this difference was not mirrored by significantly different R2 values between BIS versus SE and RE.67:41–8.90:1502–16. 15 seconds for the BIS (Methods).10:392–404. Lancet 2004. Sarkela M. SPECTRAL ENTROPY AND BIS MEASURING PROPOFOL’S ELECTROENCEPHALOGRAPHIC EFFECT 1461 Schmidt et al. References 1. et al. ke0 is assumed to remain constant over the entire observed propofol-concentration range. Simple coefficients of determination and correlation. 10. Larsen R. Vozeh S. Anesthesiology 2003. et al. Comparative evaluation of the Datex-Ohmeda S/5 Entropy Module and the Bispectral Index monitor during propofol-remifentanil anesthesia.21 minϪ1. Bruhn J. Myles PS. 5. Further clinical trials are required to show that the improved discrimination of the BIS is clinically relevant. Alves TM. Viertio-Oja H. Anesthesiology 1998.101: 1275–82. Acta Anaesthesiol Scand 1999.103:A823. This is perhaps only an adequate approximation to reality. J Physiol 1910. Reduction in the incidence of awareness using BIS monitoring. Sebel PS.89:980–1002.48:20–6. (26) reported that RE indicated emergence from anesthesia 11 seconds earlier than SE and 12. Schmidt et al. 20 seconds for the Narcotrend Index (28). Rampil IJ. (19) and Vakkuri et al. 6. Kenny GN. et al. Bispectral index (BIS) and burst suppression: revealing a part of the BIS algorithm. Bruhn J. We report higher ke0 values for SE compared with BIS independent of the underlying model used to fit the data (Table 1). Morton N. Buchinger et al. Sigl JC. Vakkuri et al. Clinical impact of hypnotictitration guidelines based on EEG bispectral index (BIS) monitoring during routine anesthetic care. The ke0 value not only contains information about the time delay the anesthetic drug requires to reach the effect site. Sandin R. be explained by different time windows required for calculating the index. Bouillon T.25:358–71. et al.4 seconds earlier than BIS. Our data investigating increasing and decreasing propofol plasma concentrations enabled us to estimate the time constant ke0. 40: iv–vii. The manufacturer claims that RE is able to detect upcoming arousal by detecting increasing EMG activity. Improved artifact algorithms may influence ke0 values by increasing the amount of interpretable EEG data within a certain time window and therefore possibly permit a better resolution of the time course of drug effect. J Clin Monit 1994. 1964. Bowerman B. Shannon C. Anesthesiology 2004. 21. Korttila K.92 to 60 seconds for SE and RE (10). 17. Kreuer S. Kreuer S. Standl T.v. Liermann VM. 9. Lindholm ML. Bruhn J. Johansen JW. The mathematical theory of communication. In summary. Bruhn J. COMPUTING. Kreuer S. lead to higher ke0 values. Bischoff P. ke0 values for propofol range between 0. our clinical investigations did not include emergence from anesthesia or surgical stimuli. Sigl JC. EEG bispectral index monitoring in sevoflurane or propofol anaesthesia: analysis of direct costs and immediate recovery.16:593–6. Schnider TW. Stanski DR.84:38–51. Shorter time windows for calculating the respective EEG index. Weaver W. Urbana. depending on the EEG variable or monitor used to measure the EEG effect. (26) have provided information concerning the clinical benefit of the Entropy ModuleTM.101:1283–90. Chamoun NG. Br J Anaesth 1991. We could not detect any significant differences between SE and RE. Biedler A. Anesthesiology 1998. IL: University of Illinois. An introduction to bispectral analysis for the electroencephalogram. O’Connell R. Anesthesiology 2005. 15. Gambus PL. anesthetic drug concentrations [German]. Measuring the performance of anesthetic depth indicators. 14. Boston: PWS-KENT. BIS seems to show slight advantages in predicting propofol effect-site concentrations compared with SE and RE. Schmidt GN. Ro ¨pcke H.12: 433–43. Acta Anaesthesiol Scand 2004. et al. (19) showed that the SE index correlated better with sedation levels compared with BIS. 20. AND SIMULATION ELLERKMANN ET AL. as well as the pharmacokinetic parameter set used to calculate the propofol plasma concentration and the model used to estimate the effect-site concentration of propofol (27–32). Bispectral index monitoring to prevent awareness during anaesthesia: the B-Aware randomised controlled trial. Anesthesiology 2004.

Cortinez LI.90:92–7. Heyse BE. Vereecke H.1462 TECHNOLOGY. Gambus PL. Narcotrend index versus bispectral index as electroencephalogram measures of anesthetic drug effect during propofol anesthesia. Acta Anaesthesiol Scand 2004. Kreuer S. Time-frequency balanced spectral entropy as a measure of anesthetic drug effect in central nervous system during sevoflurane.48:145–53. delta power. et al. Vanluchene AL. 31. Mortier EP. Shafer SL. Thas O. Bruhn J. De Smet T.101:1269–74. et al. Comparison of plasma compartment versus two methods for effect compartment– controlled target-controlled infusion for propofol. COMPUTING. Talja P. Anesthesiology 2005. Anesthesiology 2004.103:A820. Anesthesiology 1999.90:1502–16. .92:399–406. Billard V. The influence of age on propofol pharmacodynamics. Bialas P. et al. 24. et al. Kreuer S. 27. Anesthesiology 2004. Depoorter B. Schnider TW. et al. 26. Spectral entropy measurement of patient responsiveness during propofol and remifentanil: a comparison with the bispectral index. Anesth Analg 2004.101:34–42. Br J Anaesth 2004. Vanluchene AL. Anesthesiology 2000. Wakeling HG.102:1456–62 23. Ibacache ME. 28. and thiopental anesthesia. Zimmerman JB. Estimation of the plasma effect site equilibration rate constant (ke0) of propofol in children using the time to peak effect: comparison with adults. AND SIMULATION ELLERKMANN ET AL. Vakkuri A. Minto CF. propofol. Application of bispectral and narcotrend indices to the measurement of the electroencephalographic effects of desflurane with and without burst suppression. Yli-Hankala A. Clin Pharmacol Ther 1997. Targeting effect compartment or central compartment concentration of propofol: what predicts loss of consciousness? Anesthesiology 1999. 93:645–54. Munoz HR. 30. Howell S. propofol. 32. and bispectral index as EEG measures of alfentanil.61:45–58. Struys MM. et al. and midazolam drug effect. Altermatt FR. 29. Struys MM. Wilhelm W.98:692–7. SPECTRAL ENTROPY AND BIS MEASURING PROPOFOL’S ELECTROENCEPHALOGRAPHIC EFFECT ANESTH ANALG 2006. Chamoun N. Glass PS. Grundmann U. A comparison of spectral edge. 25. Spectral entropy as an electroencephalographic measure of anesthetic drug effect: a comparison with bispectral index and processed midlatency auditory evoked response. et al.