Onset of propofol-induced burst suppression may be correctly detected as deepening of anaesthesia by approximate entropy but not by bispectral index

J. Bruhn*,1, T. W. Bouillon1 and S. L. Shafer1,2

1Department of Anesthesiology, Stanford University School of Medicine, Stanford, CA, USA. 2Palo Alto VA Health Care Center, Palo Alto, CA, USA*Corresponding author: Department of Anaesthesiology and Intensive Care Medicine, University of Bonn, Sigmund Freud Str. 25, D-53105 Bonn, Germany

Accepted for publication: April 5, 2001


    Abstract
 Top
 Abstract
 Introduction
 Patient 1
 Patient 2
 Discussion
 References
 
The bispectral index (BIS) is a complex EEG variable that combines several disparate descriptors of the EEG into a single value. Approximate entropy is a novel EEG measure that quantifies the regularity of a data time series such as EEG. We report two patients in which the EEG effect of propofol was quantified very similarly by BIS and approximate entropy. However, at the beginning of burst suppression of the EEG, BIS did not indicate an increased anaesthetic drug effect, while approximate entropy did.

Br J Anaesth 2001; 87: 505–7

Keywords: monitoring, depth of anaesthesia


    Introduction
 Top
 Abstract
 Introduction
 Patient 1
 Patient 2
 Discussion
 References
 
The bispectral index (BIS, Aspect Medical Systems, Framingham, MA, USA) is a complex EEG variable that combines several disparate descriptors of the EEG into a single value, the so-called BIS value. This value has been correlated with behavioural assessments of sedation and hypnosis.1 The BIS has been used clinically to titrate depth of anaesthesia2 and to quantify the pharmacodynamic action of anaesthetic drugs in the laboratory setting.3

Approximate entropy is a novel EEG measure which quantifies the regularity of a data time series such as the EEG.4 Approximate entropy quantifies the predictability of subsequent amplitude values of the EEG, based on the knowledge of the previous amplitude values.

We report two patients in which we measured the BIS and approximate entropy at the time of onset of burst suppression of the EEG.


    Patient 1
 Top
 Abstract
 Introduction
 Patient 1
 Patient 2
 Discussion
 References
 
After local Institutional Ethics Committee approval and informed consent a 28-yr-old healthy female volunteer participated in a study of the action of propofol on the EEG and respiration. Assessment including drug testing, an ECG, medical history, and physical examination showed no abnormalities. Spontaneous ventilation continued throughout, and there was no type of stimulation during the study. The EEG was recorded continuously with a bipolar montage Fp1-A1,Fp2-A2 (international 10–20 system of electrode placement), using Zipprep electrodes (Aspect). Electrode impedance was kept below 2 k{Omega}. EEG recordings were performed with an A-1000 EEG monitor (Version 3.22, Aspect). The raw EEG signal was filtered with a 30 Hz low-pass filter and a 0.5 Hz high-pass filter. In addition to the processed EEG variables, the raw EEG signal was sampled at 128 Hz and stored on a computer hard disk for further off line analysis (Datalogger, Aspect). Approximate entropy of the EEG was calculated off-line with the settings N=1024 (length of the epoch, i.e. 1024 EEG amplitude values per epoch), m=2 (number of previous values used for the prediction of the subsequent value) and r=0.2 * SD (filtering level) as previously described.4 Propofol concentrations were ramped up/down to predetermined concentrations using a computer controlled infusion pump (STANPUMP). All concentrations refer to effect compartment concentrations predicted by STANPUMP based on Schnider’s parameter set.5 After having established a propofol target concentration of 3 µg ml–1 at the time 09:17, the propofol target concentration was increased to 6 µg ml–1 at the time 09:32 (see Fig. 1). The BIS decreased to 35, the approximate entropy decreased to 0.5. At 09:47 the propofol target concentration was increased to 9 µg ml–1. The BIS remained unchanged at 35, the approximate entropy further decreased to 0.35. The visual inspection of the raw EEG and the suppression ratio calculated by the Aspect monitor showed new appearance of burst suppression pattern (Fig. 1).



View larger version (21K):
[in this window]
[in a new window]
 
Fig 1 (Top) Propofol concentrations predicted by STANPUMP vs time. (Bottom) BIS (bold line), approximate entropy (ApEn, thin line), and suppression ratio (SR, dashed line) at increasing propofol target concentrations in Patient 1.

 

    Patient 2
 Top
 Abstract
 Introduction
 Patient 1
 Patient 2
 Discussion
 References
 
A 31-yr-old healthy female volunteer participated in the same study. The EEG montage, the EEG monitor settings and the approximate entropy calculation were the same as in Patient 1.

After having established a propofol target concentration of 2 µg ml–1 at the time 10:17, the propofol target concentration was increased to 4 µg ml–1 at the time 10:37 (see Fig. 2). The BIS decreased to 40, the approximate entropy decreased to 0.7. At 10:57, the propofol target concentration was increased to 8 µg ml–1. The BIS remained unchanged at 40, the approximate entropy further decreased to 0.4. The suppression ratio calculated by the Aspect monitor showed beginning appearance of burst suppression pattern (Fig. 2).



View larger version (20K):
[in this window]
[in a new window]
 
Fig 2 (Top) Propofol concentrations predicted by STANPUMP vs time. (Bottom) BIS (bold line), approximate entropy (ApEn, thin line) and suppression ratio (SR, dashed line) at increasing propofol target concentrations in Patient 2. From 10:17 to 10:57 BIS nearly completely parallels (i.e. overlaps) approximate entropy. The abrupt and brief change of EEG parameters at 10:39 was caused by artifacts.

 

    Discussion
 Top
 Abstract
 Introduction
 Patient 1
 Patient 2
 Discussion
 References
 
Between propofol target concentrations from 2 to 6 µg ml–1 the EEG effect of propofol was quantified very similarly by BIS and approximate entropy. This is remarkable as approximate entropy is a relatively simple single value derived from the time domain, while BIS is a complex index of several variables, requiring third order statistics.1 As this concentration range covers the clinically relevant range (with co-administration of opioid), the BIS adequately reflects EEG changes in the therapeutic range of propofol.

The higher variations of the approximate entropy values compared with the BIS values at low propofol target concentrations in Patient 1 are because approximate entropy is calculated without artifact rejection, while an artifact rejection mode is incorporated in the BIS algorithm.

In the first patient, burst suppression appeared after increasing propofol target concentration from 6 to 9 µg ml–1. Although, burst suppression ratio is a part of the BIS algorithm,1 6 this did not change the BIS. As the suppression part of the burst suppression pattern is classified as highly regular, the approximate entropy algorithm correctly classifies increasing burst suppression ratio as increasing anaesthetic drug effect.7 The decrease of approximate entropy indicated the increased pharmacodynamic action.

In the second patient, the BIS remained unchanged despite an increase of the propofol target concentration from 4 to 8 µg ml–1. The suppression ratio calculated by the Aspect monitor (Fig. 2) failed to indicate additional CNS suppression. The change in approximate entropy value indicated the pharmacodynamic situation. At this time, quantification of the phase coupling by the BIS appeared less sensitive to drug effect than quantification of regularity by approximate entropy.

With an older BIS version (version 3.12) and a combination of sufentanil, nitrous oxide, and isoflurane Detsch and colleagues recently reported paradoxical increases of BIS when isoflurane concentrations was increased.8 This was noted in the same range of BIS values between 30 and 40, where we found that the more recent BIS version (version 3.22) gave a value of BIS that failed to indicate increased propofol effect.

BIS values between 40 and 60 have been suggested as clinically adequate during surgery. The range of uncertainty between phase coupling and burst suppression ratio is at BIS values between 30 and 40. Targeting BIS between 40 and 60, a BIS value between 30 and 40 is always a ‘too deep’ state of anaesthesia. The cases we report demonstrate that, based on the BIS alone, it may be difficult to determine how much ‘too deep’ the state of anaesthesia is, which is clinically relevant information. Choosing the suppression ratio as a second trend line on the Aspect monitor, or visual inspection of the raw EEG on the screen for burst suppression pattern would have been helpful to provide a more exact ‘depth of anaesthesia’ at a BIS value between 30 and 40, particularly in the first patient.

We only noted this problem with the BIS with these specific degrees of propofol drug effect at the onset of burst suppression.

When BIS values are between 30 and 40, an increased drug pharmacodynamic effect may not be adequately reflected by the BIS value. The calculation of the burst suppression ratio may be helpful, as new evidence of burst suppression or an increasing burst suppresion ratio with unchanged BIS can highlight this problem.

Approximate entropy, a measure sensitive to signal regularity, indicated EEG changes after the occurrence of burst suppression more readily than the BIS.


    References
 Top
 Abstract
 Introduction
 Patient 1
 Patient 2
 Discussion
 References
 
1 Rampil IJ. A primer for EEG signal processing in anesthesia. Anesthesiology 1998; 89: 980–1002[ISI][Medline]

2 Song D, Joshi GP, White PF. Titration of volatile anesthetics using bispectral index facilitates recovery after ambulatory anesthesia. Anesthesiology 1997; 87: 842–84[ISI][Medline]

3 Billard V, Gambus PL, Chamoun N, Stanski DR, Shafer SL. A comparison of spectral edge, delta power, and bispectral index as EEG measures of alfentanil, propofol, and midazolam drug effect. Clin Pharmacol Ther 1997; 61: 45–58[ISI][Medline]

4 Bruhn J, Röpcke H, Hoeft A. Approximate entropy as an EEG measure of anesthetic drug effect during desflurane anesthesia. Anesthesiology 2000; 92: 715–26[ISI][Medline]

5 Schnider TW, Minto CF, Gambus PL, Andresesn C, Goodale DB, Shafer SL, Youngs EJ. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998; 88: 1170–82[ISI][Medline]

6 Glass PS, Bloom M, Kearse L, Rosow C, Sebel P, Manberg P. Bispectral analysis measures sedation and memory effects of propofol, midazolam, isoflurane, and alfentanil in healthy volunteers. Anesthesiology 1997; 86: 836–47[ISI][Medline]

7 Bruhn J, Röpcke H, Rehberg B, Bouillon T, Hoeft A. EEG approximate entropy correctly classifies the occurrence of burst suppression pattern as increasing anesthetic drug effect. Anesthesiology 2000; 93: 981–5[ISI][Medline]

8 Detsch O, Schneider G, Kochs E, Hapfelmeier G, Werner C. Increasing isoflurane concentration may cause paradoxical increases in the EEG bispectral index in surgical patients. Br J Anaesth 2000; 84: 33–7[Abstract]