The bispectral index (BIS)
The BIS monitor displays a real-time electroencephalography (EEG) trace, acquired from a frontotemporal montage. The monitor generates a dimensionless number on a continuous scale of 0-100, with 100 representing normal cortical electrical activity and 0 indicating cortical electrical silence. The influence of pre-existing neuropathology on BIS values is unknown. As with any EEG signal, BIS is subject to interference and artefact, particularly from electromyographic (EMG) activity, which can artificially elevate the recorded BIS. The display also shows a signal quality index and an indicator of EMG interference.
Because there is no ‘gold standard’ monitor against which to compare BIS, studies have used predictive probability outcome measures - that is, the likelihood of various clinically relevant end-points occurring (loss of consciousness, recovery of consciousness, postoperative recall, suppression of learning) at different BIS values. From these various studies, broad guidelines have emerged to aid the interpretation of BIS values. The probability of postoperative recall is very low when BIS is kept <60 intraoperatively. Studies comparing BIS-titrated anaesthesia with clinical judgement of anaesthetic requirements show reduced anaesthetic consumption and slightly quicker awakening using BIS, supporting the suggestion that, in trying to ensure lack of awareness, we tend towards excessively deep anaesthesia. However, BIS values tend to display considerable variability within study populations, so that, for example, one patient may be unresponsive to command at a BIS of 75, but another may still be responsive at a BIS of 70. This makes it difficult to identify sensitive and specific threshold values that are widely applicable. No BIS value predicts an individual’s threshold for loss or recovery of consciousness. BIS values discriminate between awake and asleep states but with considerable overlap of values and no clear-cut transition between awake and asleep values at the end of surgery.
How does bispectral analysis work?
The phase relationships between the component waves of different frequencies that make up the composite EEG are not considered with traditional power spectral analysis. Bispectral analysis combines traditional power spectral analysis with interrogation of these phase relationships. A number of other EEG subparameters to produce a proprietary combination of derived EEG descriptors are incorporated. BIS was developed by recording EEG data from healthy adults, who underwent repeated transitions between consciousness and unconsciousness, using several different anaesthetic regimens. The raw EEG data were time stamped at various clinical end-points. A multivariate logistic regression was used in offline analysis and identified those features of the EEG recordings that best correlated with clinical depth of sedation/anaesthesia, and these were then fitted to a model. The resulting algorithm generates the BIS. In calculating BIS from raw EEG, the relative weighting of the various subparameters changes, as some descriptors correlate better with clinical measures of light sedation while others correlate better with deeper levels of reduced consciousness. The BIS algorithm was initially validated prospectively on a second cohort of healthy volunteers and, subsequently, on various patient groups. It has been refined on several occasions.
BIS and various anaesthetic agents
Many published BIS studies have controlled for surgical stimulus by excluding it, so the applicability of BIS values taken from these studies to patients undergoing surgery remains uncertain.
BIS is therefore best described as a monitor of the depth of the hypnotic component of anaesthesia or sedation. It appears to track the effect-site concentration of hypnotic drugs and their effect on the cortical EEG. This is in keeping with BIS reflecting gradually decreasing effect-site concentrations at the end of anaesthesia.
BIS demonstrates a dose-response relationship with inhalational and hypnotic intravenous agents, such as propofol and midazolam, which is independent of the agent(s) being used and correlates with clinical assessments of the level of consciousness. Bispectral analysis is the first processed EEG technique to be correlated well with behavioural assessments of level of consciousness. Ketamine, however, causes EEG activation, complicating BIS interpretation.
In the absence of surgical stimulation (conditions applying during most early studies), the use of opioids produces clinical changes in depth of sedation or anaesthesia that are not reflected by decreases in BIS. This is a major drawback of using BIS to assess depth of balanced anaesthesia, as it does not fully reflect the synergistic effect of opioids with hypnotic agents. Nevertheless, when opioids are used during surgery, BIS values do decrease, perhaps illustrating the counteraction of arousal by pain. Not surprisingly, BIS is not able to predict movement in response to surgical stimulation; the anatomical site of generation of such reflexes is likely to be at the level of the spinal cord and therefore unlikely to register on the cortical EEG.
Baseline BIS values are not reduced by nitrous oxide, at inspired concentrations of up to 50%. Furthermore, the addition of nitrous oxide to established anaesthesia has little or no effect on BIS in the absence of surgical stimulation. However, during surgery, the antinociceptive effects of nitrous oxide may be responsible for the observed decrease in BIS.
BIS and paediatrics
Because healthy adult EEG data were used to authenticate the BIS algorithm, it cannot automatically be extrapolated to young children, as the paediatric EEG only approaches the adult pattern by about 5 years of age. However, early investigations suggest that BIS may be valid in children older than 1 year of age.
BIS and intensive care
BIS can be used as a continuous monitor of sedation in adult intensive care, and investigations have concluded that it is a useful reflector of the great interindividual variations in pharmacokinetics and pharmacodynamics of sedatives in critically ill patients. However, comparative studies using either the Ramsey sedation score (itself not validated) or the Sedation-Agitation Score have generated unimpressive correlation statistics. There is methodological dissatisfaction with comparing serial, cross-sectional, subjective observations of patient comfort with a quasi-continuous monitor of cortical activity, and it is unsurprising that there is limited correlation between the two. There are ongoing investigations into the role of BIS in paediatric intensive care. Early studies in convenience samples comparing BIS with the COMFORT paediatric sedation scoring system again suggest limited correlation.
[i] Applications of bispectral index monitoring in the pediatric intensive care unit. J Intensive Care Med 2004; 19: 111-6
[ii] Reduction in the incidence of awareness using BIS monitoring. Acta Anaesthesiol Scand 2004; 48: 20-6
[iii] Clinical comparison of three different anaesthetic depth monitors during cardiopulmonary bypass. Anaesthesia 2005; 60: 189-93
[iv] Impact of age on both BIS values and EEG bispectrum during anaesthesia with sevoflurane in children. Br J Anaesth 94: 810-20
[v] Monitoring depth of anaesthesia by EEG Whyte and Booker Contin Educ Anaesth Crit Care Pain 2003; 3: 106-10.