This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Cortínez, L. I. Articles by Muñoz, H. R. Search for Related Content PubMed PubMed Citation Articles by Cortínez, L. I. Articles by Muñoz, H. R.

---

TECHNOLOGY, COMPUTING, AND SIMULATION

Performance of the Cerebral State Index During Increasing Levels of Propofol Anesthesia: A Comparison with the Bispectral Index

From the Departamento de Anestesiología, Facultad de Medicina, Pontificia Universidad Católica de Chile. Santiago, Chile.

Address correspondence and reprint requests to Luis I. Cortínez, MD, Departmento de Anestesiología, Hospital Clínico U.C., Marcoleta 367, Santiago, Chile. Address e-mail to licorti{at}med.puc.cl.

	Abstract

Top Abstract Introduction METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
BACKGROUND: The cerebral state monitor is a new device to measure depth of anesthesia. In this study we compared the cerebral state monitor with the bispectral index (BIS) monitor during propofol anesthesia.

METHODS: Fifteen healthy patients received a continuous infusion of propofol (300 mL/h). The cerebral state index (CSI) and the BIS values were recorded until burst suppression ratio 60%. Baseline variability, prediction probability, and agreement analysis between indices were evaluated. Clinical markers of loss of consciousness were also assessed.

RESULTS: Mean awake BIS and CSI values were 95.6 and 91.6, respectively (P = 0.01). BIS and CSI prediction probability values (mean ± sd) were estimated to be 0.87 ± 0.08 and 0.86 ± 0.08, respectively (NS). The CSI tended to stabilize at values of 60–40 when estimated propofol concentrations at the effect site increased from 5 to 8 µg/mL. The BIS stabilized at values of 40–20 when the propofol concentrations at the effect site increased from 7 to 10 µg/mL. The mean BIS-CSI difference was –7.4 with 95% limits of agreement of 22.2 and –36.9. The BIS and CSI correlation with the burst suppression ratio was –0.60 and –0.97, respectively (P < 0.01). Predicted BIS and CSI values for loss of eyelash reflex in 50% and 95% of the patients were different (P < 0.05).

CONCLUSION: The overall performance of both monitors during propofol induction was similar. However, the different dynamic profiles of these monitors indicate that BIS may be a more useful index for evaluating intermediate anesthetic levels, whereas CSI may be better for evaluating deeper anesthetic levels.

	Introduction

Top Abstract Introduction METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Electroencephalographic (EEG) monitors are increasingly being used to measure the hypnotic effect of anesthetic drugs. The most extensively validated device used for this purpose is the bispectral index (BIS) monitor (Aspect Medical Systems, Newton, MA). This monitor calculates the "bispectral index" using multiple subparameters of the EEG as a measure of the hypnotic effect (1,2). The cerebral state monitor (CSM) (Danmeter A/S, Odense, Denmark) has recently been introduced as a new device to measure the hypnotic component of anesthetics drugs. Both the BIS and the CSM use proprietary algorithms to process the EEG signal and estimate dimensionless indices that are expected to approach 100 for the awake patient and progressively decrease as hypnosis increases. Both manufacturers state that index values between 40 and 60 indicate an adequate depth of hypnosis.

The cerebral state index (CSI), calculated by the CSM, uses four subparameters derived from the time domain analysis (burst-ratio) and frequency domain analysis (-ratio, ß-ratio, ß-ratio–-ratio) of the EEG (3). These subparameters are then used as inputs to a fuzzy logic inference system that calculates the CSI. Fuzzy logic is a problem-solving control system methodology that incorporates a simple, rule-based "IF X AND Y THEN Z" approach to a solving control problem rather than attempting to model a system mathematically. The potential advantage of this approach is that because the relationship between the EEG subparameters and the clinical state cannot be easily modeled by a mathematical function, the neuro-fuzzy method might offer a better alternative to establish this relationship (4).

The aim of this study was to compare the change in EEG indices by both the CSI and the BIS over a wide range of hypnotic effects induced by propofol.

	METHODS

Top Abstract Introduction METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
After institutional ethics committee approval (Facultad de Medicina, Pontificia Universidad Católica, Santiago, Chile), written informed consent was obtained from 15 ASA I patients, aged 20–40 yr, scheduled to undergo surgery under general anesthesia. Exclusion criteria included a body mass index >30%, any known cerebrovascular disease, long- or short-term (within the previous 48 h) intake of any drug acting on the central nervous system, and any known adverse effect to the study drugs.

After arrival in the operating room, standard monitoring (heart rate, noninvasive arterial blood pressure and arterial oxygen saturation) was initiated and a peripheral IV line was installed. At this time, the sensors of the Aspect A-2000 BIS® monitor (version XP) and the CSM were attached according to the manufacturer’s recommendations. The smoothing time period of the BIS monitor was set at 15 s. The data of the CSM and BIS were automatically recorded in intervals of 1 and 5 s, respectively. BIS values were recorded using Microsoft® Hyper Terminal Version 5.1(Microsoft Corporation) and transferred to a computer hard disk for off-line analysis. The values of CSI were recorded using the Danmeter A/S CSM capture V2.02 onto the computer hard disk.

All patients were breathing spontaneously through a facemask delivering 100% oxygen at 10 L/min throughout the study period. No patient received preoperative medication or any other drug during the study period. Before propofol administration a baseline period of 2 min was recorded. During this period, patients were kept undisturbed with the operating room in silence. Then, a continuous infusion of 1% propofol was started at 300 mL/h using a Pilot 2 syringe infusion pump (Fresenius Vial Infusion Systems, Brézins, France) and maintained until either burst suppression levels 60% were reached in both monitors or until mean arterial blood pressure <50 mm Hg. Three clinical end points of loss of consciousness were assessed every 5 s from the start of propofol infusion and recorded when they occurred. These clinical end points were: loss of response to verbal command, loss of eyelash reflex, and drop of a weighted syringe from the patient’s hand.

The performance of both monitors was compared during baseline by computing the mean and the coefficient of variation (CV%) of the BIS and CSI before propofol administration. During anesthesia, the predictive capacity of both indices for detecting the level of anesthesia was evaluated with the prediction probability (P_k) statistic. To determine the P_k statistic, estimated propofol concentrations at the effect site (CePROP) were calculated and used as an objective measure of the level of anesthesia. Pharmacokinetic parameters of Schnider et al. (5) and a k_e0 value predicting a peak effect at 1.6 min after a bolus dose (5,6) were used to estimate CePROP. The P_k between CePROP and both EEG indices was then calculated for each individual patient using the PKMACRO, developed by Smith et al. (7). The P_k can range from 0.5 to 1. A P_k value of 0.5 means no predictive ability (50% chance) and a P_k of 1 means that the monitor always correctly predicts increments or decrements in the level of anesthesia. In addition, agreement between CSI and BIS throughout the study period was assessed with a Bland–Altman analysis.

The burst suppression ratios (BSR) recorded by each of these monitors were used to evaluate the influence of burst suppression activity on BIS and CSI. Agreement between the BSR calculated by both monitors was also assessed.

EEG indices values were compared with paired Student’s t-test. The probability of being unconscious at different BIS or CSI levels was calculated using logistic regression analysis, and multiple comparisons between these values were performed with ANOVA. A P value 0.05 was considered statistically significant. Statistical analyses were performed using R (language and environment for statistical computing, freely available from http://www.r-project.org/).

	RESULTS

Top Abstract Introduction METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
All patients, seven women and eight men, age 30.5 ± 6.5 yr, weight 69.0 ± 12.5 kg and height 168.5 ± 8.8 cm (mean ± sd) completed the study. All patients were able to maintain spontaneous ventilation throughout the study period. The lowest mean arterial blood pressure values recorded in each patient were 76.3 ± 10.5 mm Hg (mean ± sd). In all patients, BIS and CSI values progressively decreased during propofol infusion and the end point of a BSR 60% was reached.

Before propofol administration, BIS values were higher with less fluctuation in time than CSI values. The mean (CV%) was 95.6 (3.9) and 91.6 (7.8), respectively (P = 0.01).

The predictive ability of both indices measured with the P_k statistic (mean ± sd) were 0.87 ± 0.08 for BIS and 0.86 ± 0.08 for CSI (NS). The individual raw BIS and CSI values versus CePROP are shown in Figures 1a and b.

View larger version (19K): [in this window] [in a new window]   Figure 1. (a) Plots of the individual measured bispectral index (BIS) values and cerebral state index (CSI) values. (b) Their relationship with estimated propofol concentrations at the effect site (CePROP). BIS and CSI values progressively decreased during propofol infusion in all patients.

The degree of agreement between BIS and CSI, and between the BSR calculated by each monitor, is shown in Figures 2a and b. At very deep levels of anesthesia, the BIS was relatively insensitive to burst suppression until a BSR of 40% or more (Fig. 3a). This was not the case with CSI, which was very sensitive to burst suppression changes (Fig. 3b). The Pearson correlation coefficient (r) between BIS and BSR (r = –0.60) was lower than the one obtained between CSI and BSR (r = –0.97), (P < 0.01).

View larger version (15K): [in this window] [in a new window]   Figure 2. Bland–Altman plots. (a) The differences between bispectral index (BIS) and cerebral state index (CSI). (b) Difference between burst suppression ratio (BSR)BIS and BSRCSM are plotted against their averages. The solid line is the mean difference (bias) and the dotted lines are the mean difference plus and minus 1.96 times the standard deviation of the differences (95% limits of agreement). Figure 2a shows that the agreement between BIS and CSI vary with the level of anesthesia. The mean BIS-CSI difference was –7.4 with 95% limits of agreement of 22.2 and –36.9. Figure 2b shows that there is a good degree of agreement with very little bias between the BSR calculated by each monitor. The mean BSR difference was –3.4 with 95% limits of agreement of 30.3 and –37.1.

View larger version (12K): [in this window] [in a new window]   Figure 3. (a) Plot of the individual measured bispectral index (BIS) values and cerebral state index (CSI) values. (b) Their relationship with burst suppression ratio (BSR).

The CePROP values at three clinical end points of unconsciousness for 50% and 95% of patients are shown in Table 1. The predicted BIS and CSI values at these three clinical end points for 50% and 95% of patients are shown in Tables 2 and 3. The three clinical end points occurred at different BIS and CSI values (P < 0.01). Figures 4a–c show the probability curves of these three clinical end points with both monitors. Finally, the predicted CePROP that have a 95% probability of producing BIS or CSI values 60, 50, 40, 30 and 20 are shown in Figure 5.

View larger version (9K): [in this window] [in a new window]   Figure 4. (a) Probability curves of loss of eyelash reflex, (b) loss of verbal contact, (c) and drop a weighted syringe, (d) at different cerebral state index (CSI) values. The horizontal lines represent the 95% confidence intervals of the probability 50% and the probability 95%.

View larger version (14K): [in this window] [in a new window]   Figure 5. Estimated propofol concentrations at the effect site (CePROP) that have a 95% probability of producing relevant cerebral state index (CSI) values. This figure shows that if propofol administration is titrated to obtain an index value between 40 and 60, the CSI may result in deeper anesthesia than bispectral index (BIS).

	DISCUSSION

Top Abstract Introduction METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
In general, during propofol induction, both monitors showed similar predictive capacities for detecting the level of anesthesia. However, neither the BIS nor the CSI linearly decreased at increasing levels of CePROP. Although the CSI tended to stabilize at values of 60–40 at intermediate levels of hypnosis (when CePROP increased from 5 to 8 µg/mL), the BIS stabilized at values of 40–20 at deeper anesthetic levels (when CePROP increased from 7 to 10 µg/mL). The different dynamic profiles showed by both monitors might indicate a superiority of BIS with respect to CSI to adequately discriminate intermediate anesthetic levels, and a superiority of CSI with respect to BIS at deeper anesthetic levels.

The predictive capacity of CSM and BIS for detecting the level of anesthesia evaluated with the P_k statistic showed a good performance for both monitors (mean BIS-CePROP P_k value of 0.87 and mean CSI-CePROP P_k values of 0.86). This is consistent with a recent study by Zhong et al. (8) comparing the CSI with the BIS in patients undergoing general anesthesia with propofol. In this study, the authors performed stepwise increments in propofol concentration and found that both indices showed similar correlations between CePROP-BIS (r² = 0.787) and CePROP-CSI (r² = 0.792). Jensen et al. (4) compared the performance of the CSI with that of the BIS and the A-Line ARX index. For this, they used raw EEG records from two earlier studies (9,10) and performed an off-line calculation of the CSI. In agreement with our results, the authors showed that during propofol infusion at 300 mL/h (protocol 1) the behavior of the CSI and the BIS was comparable. However, they found that the correlation between propofol concentration and CSI (–0.94) was higher than that of the BIS (–0.82) and the A-Line ARX index (–0.89). It is possible that because the CSI algorithm was calculated by these same authors using a similar data set (15 of 50 patients received propofol until BSR >80%) (4), this might have favored the CSI’s performance. In addition, possible differences in the way the authors calculated the CSI (off-line calculation from filtered EEG signals acquired with an A-Line monitor) might also have influenced their results.

Recently, Anderson and Jakobsson (11) compared the CSM with the BIS monitor in a population of day-surgery patients using several anesthetics. Interestingly, the authors found that when BIS values decreased below 40, the CSI tended to stabilize at higher values. This tendency was also observed in a study by Jensen et al. (4) and in the present study (Fig. 2a). The different relationships observed in our study between indices and CePROP (Figs. 1a and 1b) suggest a better dynamic profile of BIS to discriminate intermediate levels of anesthesia, but a better profile of CSI at deeper anesthetic levels. This is also in agreement with the second protocol in the study by Jensen et al. (4). In that section, the authors assessed the level of consciousness through the observer’s assessment of alertness and sedation (OAA/S) scale at different steady-state CePROP and showed that BIS, but not the CSI, adequately discriminated between OAA/S levels 3 and 2, whereas the opposite occurred at deeper anesthesia (OAA/S levels 0 and 1).

During deep anesthesia, an EEG pattern known as "burst suppression" may appear. This pattern is characterized by alternating periods of normal to high voltage activity changing to low voltage or even isoelectricity (2). Although both monitors use their own algorithms to quantify this phenomenon (2,12), a good agreement was observed between the BSR calculated by each monitor (Fig. 2b). With the initiation of burst suppression activity, we found that the dynamic profile of the CSI improved and was highly correlated to BSR indicating a virtually total dependency on BSR at all BSR values (Fig. 3a). On the contrary, BIS was not significantly affected by this phenomenon until a BSR 40%. This is probably the reason for the superiority of CSI over BIS to adequately detect deeper anesthetic levels.

Electromyographic (EMG) activity can contaminate EEG signals, and could alter the value of EEG-derived indices. In our protocol, we did not include administration of a muscle relaxant to test this possibility. Although, it has been shown that the neuromuscular block level does not alter BIS during propofol anesthesia (13), this possibility has not been explored with the CSI. It might be hypothesized that part of the stabilization of the CSI at values of 60–40 at intermediate anesthetic levels might have been due to EMG artifact in the CSI. However, we did not observe significant EMG activity (recorded with the CSM electrodes) during these periods. A different protocol design with and without muscle relaxant is probably necessary to exclude this hypothesis.

The three clinical end points of loss of consciousness used in this study were reached at different indices values either with the BIS or with the CSM. Predicted BIS and CSI values for loss of eyelash reflex in 50% and 95% of patients were different (Tables 2 and 3). For the other two clinical end points used in this study (loss of verbal contact and drop of a weighted syringe), no differences between monitors were observed. The discrepancies observed between monitors for loss of eyelash reflex and not for the other two clinical end points were probably because loss of eyelash reflex occurred at hypnotic levels that occur within the less dynamic range of the CSI (Table 1). We also observed a wide interindividual variability of BIS and CSI values at these three clinical end points. There have been many studies analyzing the relationship between clinical end points and EEG-derived indices during propofol anesthesia (14–16). Most of these studies have found that these indices are useful tools for predicting the anesthetic level. However, there are wide variations in the values (14–16). Different definitions of clinical end points and different propofol administration schemes are mainly responsible for these disparities. To accurately estimate the value of the BIS or CSI associated with a clinical end point under dynamic conditions, it is important to adequately predict the estimated effect site concentrations of the drug, and to know the exact time delay of the index calculation. The method of administration of propofol has been shown to affect the accuracy of the pharmacokinetic model (17), resulting in a reduction of the ability of prediction as the rate of administration of propofol increases.

One limitation of this study is that we infused propofol at a relatively rapid rate (300 mL/h). We used this infusion rate because it was used in two studies comparing EEG monitors (4,18), which allowed as to compare our results with earlier findings. Another limitation of the present study is that we have not corrected for possible time delays in index calculations. This was because we wanted to show the real values that the clinicians would see using either monitor during propofol infusion. These factors might explain the higher EC₅₀/EC₉₅ values for loss of verbal contact (4.7/6.6 µg/mL) found in our study, compared with those estimated by Struys et al. (19) at steady-state (2.9/3.8 µg/mL). A recent study (12) showed that BIS and CSI have different time lags for reacting to a simulated change in the level of anesthesia. These delays ranged between 14 and 155 s and varied depending on the speed of change of the anesthetic level (12). This makes time delay corrections of their indices very difficult to perform. As a consequence, it is not surprising to find wide differences between pharmacodynamic studies using these monitors, and their results should always be interpreted cautiously and with clinical judgment.

	CONCLUSION

Top Abstract Introduction METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
The overall performance of both monitors during propofol induction was similar. However, the different dynamic profiles showed by both monitors indicate that BIS may be a more useful index for evaluating intermediate anesthetic levels, whereas CSI may be better for evaluating deeper anesthetic levels. Further study is needed to determine if a significantly different dose of propofol would result when using either the CSI or BIS to guide drug administration.

	Footnotes

 
Accepted for publication November 6, 2006.

	REFERENCES

Top Abstract Introduction METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 

1. Sigl JC, Chamoun NG. An introduction to bispectral analysis for the electroencephalogram. J Clin Monit 1994;10:392–404.[Web of Science][Medline]
2. Rampil IJ. A primer for EEG signal processing in anesthesia. Anesthesiology 1998;89:980–1002.[Web of Science][Medline]
3. Rodriguez BE, Jensen EW, Martinez P. Definition of a new index for depth of anesthesia using EEG sup-parameters combined by fuzzy logic. Abstract presented at Seventh EuroSIVA Meeting, Lisbon, 2004.
4. Jensen EW, Litvan H, Revuelta M, et al. Cerebral state index during propofol anesthesia: a comparison with the bispectral index and the A-line ARX index. Anesthesiology 2006;105:28–36.[Web of Science][Medline]
5. Schnider TW, Minto CF, Shafer SL, et al. The influence of age on propofol pharmacodynamics. Anesthesiology 1999;90:1502–16.[Web of Science][Medline]
6. Struys MM, De Smet T, Depoorter B, et al. Comparison of plasma compartment versus two methods for effect compartment– controlled target-controlled infusion for propofol. Anesthesiology 2000;92:399–406.[Web of Science][Medline]
7. Smith WD, Dutton RC, Smith NT. Measuring the performance of anesthetic depth indicators. Anesthesiology 1996;84:38–51.[Web of Science][Medline]
8. Zhong T, Guo QL, Pang YD, et al. Comparative evaluation of the cerebral state index and the bispectral index during target-controlled infusion of propofol. Br J Anaesth 2005;95:798–802.[Abstract/Free Full Text]
9. Vereecke HE, Struys MM, Mortier EP. A comparison of bispectral index and ARX-derived auditory evoked potential index in measuring the clinical interaction between ketamine and propofol anaesthesia. Anaesthesia 2003;58:957–61.[Web of Science][Medline]
10. Struys MM, Jensen EW, Smith W, et al. Performance of the ARX-derived auditory evoked potential index as an indicator of anesthetic depth: a comparison with bispectral index and hemodynamic measures during propofol administration. Anesthesiology 2002;96:803–16.[Web of Science][Medline]
11. Anderson RE, Jakobsson JG. Cerebral state monitor, a new small handheld EEG monitor for determining depth of anaesthesia: a clinical comparison with the bispectral index during day-surgery. Eur J Anaesthesiol 2006;23:208–12.[Web of Science][Medline]
12. Pilge S, Zanner R, Schneider G, et al. Time delay of index calculation: analysis of cerebral state, bispectral, and narcotrend indices. Anesthesiology 2006;104:488–94.[Web of Science][Medline]
13. Greif R, Greenwald S, Schweitzer E, et al. Muscle relaxation does not alter hypnotic level during propofol anesthesia. Anesth Analg 2002;94:604–8.[Abstract/Free Full Text]
14. Irwin MG, Hui TW, Milne SE, Kenny GN. Propofol effective concentration 50 and its relationship to bispectral index. Anaesthesia 2002;57:242–8.[Web of Science][Medline]
15. Iannuzzi M, Iannuzzi E, Rossi F, et al. Relationship between bispectral index, electroencephalographic state entropy and effect-site EC50 for propofol at different clinical endpoints. Br J Anaesth 2005;94:613–6.[Abstract]
16. Milne SE, Troy A, Irwin MG, Kenny GN. Relationship between bispectral index, auditory evoked potential index and effect-site EC50 for propofol at two clinical end-points. Br J Anaesth 2003;90:127–31.[Abstract/Free Full Text]
17. Schnider TW, Minto CF, Gambus PL, et al. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998;88:1170–82.[Web of Science][Medline]
18. Vereecke HE, Vasquez PM, Jensen EW, et al. New composite index based on midlatency auditory evoked potential and electroencephalographic parameters to optimize correlation with propofol effect site concentration: comparison with bispectral index and solitary used fast extracting auditory evoked potential index. Anesthesiology 2005;103:500–7.[Web of Science][Medline]
19. Struys MM, Vereecke H, Moerman A, et al. Ability of the bispectral index, autoregressive modelling with exogenous input-derived auditory evoked potentials, and predicted propofol concentrations to measure patient responsiveness during anesthesia with propofol and remifentanil. Anesthesiology 2003;99:802–12.[Web of Science][Medline]

This article has been cited by other articles:

				A. E. Delfino, L. I. Cortinez, C. V. Fierro, and H. R. Munoz Propofol consumption and recovery times after bispectral index or cerebral state index guidance of anaesthesia Br. J. Anaesth., August 1, 2009; 103(2): 255 - 259. [Abstract] [Full Text] [PDF]

				R. Fuentes, L. I. Cortinez, M. M. R. F. Struys, A. Delfino, and H. Munoz The Dynamic Relationship Between End-Tidal Sevoflurane Concentrations, Bispectral Index, and Cerebral State Index in Children Anesth. Analg., November 1, 2008; 107(5): 1573 - 1578. [Abstract] [Full Text] [PDF]

				C. N. Sessler and K. Varney Patient-Focused Sedation and Analgesia in the ICU* Chest, February 1, 2008; 133(2): 552 - 565. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Cortínez, L. I. Articles by Muñoz, H. R. Search for Related Content PubMed PubMed Citation Articles by Cortínez, L. I. Articles by Muñoz, H. R.