ANESTHESIA AND THE ELECTROENCEPHALOGRAM

With the advent of the then high-tech EEG in 1950, Courtin and colleagues^[22] sought to monitor the brain and devise a servocontrolled system that could adjust the anesthetic concentration administered on the basis of the electroencephalographic pattern. It was an ingenious idea, but despite available descriptions of the EEG during anesthesia, knowledge and technology were not adequate at that time. Even using more modern devices with much higher degrees of computer power, servocontrolled administration of anesthetics remain imperfect at best.^{[23] [24]} Because all anesthetic drugs do not produce exactly the same changes in electroencephalographic pattern as anesthesia deepens, generic correlation of the EEG with depth of anesthesia across all anesthetic techniques remains an elusive goal. One of the major reasons that the EEG has been difficult to use for assessing anesthetic depth is that most modern anesthetics use many different classes of drugs, all of which have significant electroencephalographic effects, for premedication, induction, and

1517

Much research has been directed at developing a processed electroencephalographic parameter that is indicative of depth of anesthesia. These processed parameters fall into two broad categories: parameters derived from power analysis of the raw electroencephalographic data and parameters derived from bispectral analysis of the EEG. Power analysis processes data from the raw EEG related to frequency and amplitude of the waveforms over time but does not consider phase relationships between the component waves. Bispectral analysis is a somewhat more complex process that includes phase relationship data. Several processed electroencephalographic parameters from Fourier analysis have been investigated as indicators of anesthetic depth. These parameters include spectral edge frequency 95% (SEF95), which is the frequency below which 95% of the electroencephalographic activity falls; median frequency (i.e., frequency at the midpoint of the power spectrum); peak power frequency (i.e., frequency with the highest electroencephalographic power), and relative delta power (i.e., percent of the EEG power in the delta band). Some investigators have had success in using one or more of these parameters to predict the depth of anesthesia.^[14] Others have found that although these parameters change during anesthesia, they were not consistently predictive of depth of anesthesia as assessed by response to stimuli, movement, or return of consciousness during emergence from anesthesia.^{[25] [26]} In particular, these parameters depend on the type of anesthetic agent or combination of agents used. Whereas one parameter may correlate well with a primarily volatile anesthetic technique, it may perform less consistently during narcotic-based anesthesia.

Encouraging results have been obtained using the BIS to monitor anesthetic depth (see Chapter 31 ). The BIS is a processed parameter derived from multiple features generated by bispectral analysis of the EEG. Through clinical trials, features of the bispectral analysis that were predictive of response to stimuli under the effects of a variety of anesthetic agents were identified. These features were combined to a multivariate index using discriminant analysis.^{[18] [27]} After the BIS was developed, it was further tested and empirically refined to improve predictive ability.^[27] The BIS is displayed as a numeric value from 0 to 100 and can be trended over time. Many clinical trials have investigated the ability of the BIS to monitor anesthetic depth and predict response to stimuli. Trials have been directed at determining BIS values predictive of loss of consciousness, loss of recall, and prevention of movement in response to surgical stimulation. The BIS value indicative of a certain level of consciousness varies somewhat between different anesthetic techniques^[27] ; however, the ability of the BIS to predict loss of consciousness and lack of recall during sedation has been consistently demonstrated by a number of investigators using a variety of drugs and drug combinations.^{[14] [27]} However, the ability of BIS to predict hemodynamic response to surgical stimulation or movement in response to surgical stimulation has been less consistently demonstrated and may depend more on the anesthetic technique.^{[28] [29]} These results have suggested to some that the anesthetic state involves at least two different components. One is a reflection of hypnosis and consists of loss of consciousness and recall; the BIS value is indicative of this state. However, the obtundation of hemodynamic and movement responses to noxious stimuli is less well predicted by the BIS and probably is mainly mediated at the spinal cord level.^[27] BIS monitoring of the level of sedation during an awake cranitomy with cortical mapping is shown in Figure 38-5 .

Anesthetic drugs affect the frequency and amplitude of electroencephalographic waveforms. Although each drug class and each specific drug has some explicit, dose-related electroencephalographic effects ( Table 38-3 ), some basic anesthesia-related electroencephalographic patterns may be described. Subanesthetic doses of intravenous and inhaled anesthetics usually produce an increase in frontal β activity and abolish the α activity normally seen in the occipital leads in the awake, relaxed patient with the eyes closed. As the patient goes to sleep with general anesthesia, the brain waves become larger in amplitude and slower in frequency. In the frontal areas, small β activity seen in the awake patient slows to the α range and increases in size. In combination with the loss of the occipital α activity, this phenomenon produces the appearance of a shift of the α activity from the posterior cortex to the anterior cortex. Further increases in the dose of the inhalation or intravenous agent produce further slowing of the EEG, and some agents have the capability to totally suppress electroencephalographic activity (see Table 38-3 ). Other agents (e.g., opioids, benzodiazepines) never produce burst suppression or an isoelectric EEG despite increasing dose because they are incapable of completely suppressing the EEG or because cardiovascular toxicity of the drug (e.g., halothane) prevents administration of a large enough dose.

Intravenous Anesthetics

Barbiturates, Propofol, and Etomidate

Despite widely varying potencies and durations of action, all of the intravenous anesthetics produce similar electroencephalographic patterns. Figure 38-6 shows the electroencephalographic effects of thiopental (see Chapter 10 ). These drugs all follow the basic anesthesia-related electroencephalographic pattern described previously, with

1518

Figure 38-5 Bispectral index (BIS) tracings from an awake craniotomy with intraoperative cortical mapping. Propofol infusion was titrated against the BIS value to provide adequate sedation during the operative procedure. Termination of the propofol infusion during the cortical mapping is followed by a prompt increase in the BIS value and recovery of the patient to a fully alert state.

Ketamine

Ketamine does not follow the basic anesthesia-related electroencephalographic pattern. Anesthesia with ketamine is characterized by frontally dominant rhythmic, high-amplitude theta activity. Increasing doses produce intermittent polymorphic δ activity of very large amplitude interspersed with low-amplitude β activity. ^[30] Electrocortical silence cannot be produced with ketamine. Electroencephalographic activity may be very disorganized and variable at all doses. Recovery of normal electroencephalographic activity, even after a single bolus dose of ketamine, is relatively slow compared with barbiturates. There is no information available about the relationship between emergence reactions after ketamine and the EEG. Ketamine has also been associated with increased epileptiform activity.^[30]

Benzodiazepines

Despite different potencies and durations of actions, benzodiazepines also follow the basic anesthesia-related electroencephalographic pattern. As a class, however, these drugs are incapable of producing burst suppression or an isoelectric EEG.

Opioids

As a class, opioids do not follow the basic anesthesia-related electroencephalographic pattern. In general, opioids produce a dose-related decrease in frequency and increase in amplitude of the EEG. If no further doses of opiates are given, α and β activity will return as drug redistribution occurs. The rapidity of return depends on the initial dose and on the drug. Remifentanil is associated with the most rapid return to normal.^[31] Complete suppression of the EEG cannot be obtained with the opioids. Epileptiform activity occurs in humans and in animals receiving large to supraximal clinical doses of opioids. For example, sharp wave activity is relatively common after induction of anesthesia with fentanyl, with 20% of patients showing this phenomenon after 30 µg/kg, 60% at 50 µg/kg, 58% at 60 µg/kg, and 80% at 70 µg/kg. This epileptiform activity is mainly observed in the frontotemporal region.^[32]