The BIS® monitor (Aspect Medical Systems, Natick, MA) uses a form of processed cortical electroencephalogram (EEG) to quantify the hypnotic effects of anesthetic drugs. This device uses both Fourier transformation and bispectral analysis to compute a single number, the bispectral index (BIS) which ranges from 100 (awake) to 0 (isoelectric EEG) (1,2 3,4 5–9 10–12
Validation of BIS poses significant problems in infants and young children. We cannot rely on a simple response to voice command as a “gold standard” for sedation or sleep, and other clinical end points can sometimes be ambiguous. In babies, it can be difficult to distinguish purposeful actions from nonspecific startle responses. Because it is not possible to ask normal children to volunteer for a study of general anesthesia, we must use surgical patients and thereby introduce the variable of surgical stimulation. Despite these shortcomings, we believe it is feasible to describe clinical markers of sedation and hypnosis in pediatric surgical patients and to measure average BIS values under defined conditions. We specifically propose the following two hypotheses:
1. BIS values obtained before anesthesia, during maintenance, and at emergence will be similar to those reported for adults.
2. Increasing concentrations of a general anesthetic will produce a dose-related decrease in BIS over the range used for clinical anesthesia.
We tested the first hypothesis by observing BIS during routine general anesthetics involving infants and children. BIS was recorded at various clinical endpoints, but the information was never shared with the person giving anesthesia. The data were compared with historical adult “controls,” because we had previously measured the average values for BIS at similar case milestones in adult patients (6
Methods
After institutional approval and informed consent, 77 pediatric patients were studied under the two protocols. Fifty-five patients were initially observed with BIS measurements made throughout an unrestricted general anesthetic. In the second protocol, 22 infants and children were anesthetized with various end-tidal concentrations of sevoflurane in nitrous oxide (50%–60%) and oxygen, and BIS measurements were made at each concentration. Patients were considered eligible for enrollment in either protocol if they were between 1 mo and 12 yr old, ASA physical status I or II, and scheduled for elective surgery requiring general anesthesia. Patients having cardiovascular, major vascular or neurosurgical procedures were excluded, as were those with large anticipated fluid shifts or body temperature changes. Patients were also excluded if they had significant cardiovascular, respiratory, or neurologic disease or if they were taking chronic medications known to affect the EEG.
BIS was measured on an Aspect Medical Systems Model A1050 EEG monitor by using commercially available, disposable BisSensor® strips (Aspect Medical Systems) designed for use in adults. The strip uses three gel electrodes, two active, one ground. Placement on the forehead approximated the recommended placement in adults, i.e., the proximal lead was above the nasion, and the distal lead was midway between the tragus of the ear and the outer canthus of the eye. No additional skin preparation was used, but the impedance of each electrode was <2 kΩ (ground, <10 kΩ). BIS version 3.3 was recorded continuously. The smoothing window was set at 15 s and the update rate at 2 s. The algorithm within the BIS® monitor sets limits for electrode impedance and signal quality, and no BIS value is displayed if the signal has too much noise or artifact. We accepted the limits set in the monitor: if the BIS was displayed, it was recorded.
An initial group of 55 patients had BIS recorded during routine general anesthetics. Sedative-hypnotic drugs used included methohexital, propofol, sevoflurane, isoflurane, desflurane, nitrous oxide, and diazapem. The choice of drugs, the method of airway management, and the administration of opioids, local anesthetics, and neuromuscular blockers were at the discretion of the primary anesthesia caregiver. During the observational protocol, the BIS was recorded, but the anesthetist was not aware of the BIS value. This was done to ensure that the BIS value was not used to guide the conduct of the anesthetic. Study personnel recorded BIS either manually (36 patients) or by using a commercially available software program (Datalogger, Aspect Medical Systems) (19 patients). BIS values were analyzed for three prospectively-defined case milestones: before the induction of anesthesia, during maintenance, and at emergence from anesthesia. In addition, a “nadir BIS,” the lowest BIS value occurring within the 10-min period after the induction of anesthesia, was recorded. Maintenance was defined as the period starting 5 min after the nadir BIS until the hypnotic was turned off. Emergence was defined as the time when any of the following first occurred: 1) eyes open spontaneously, 2) crying or phonating, or 3) purposeful movements. Comparison values for BIS in adults at the same clinical end points were obtained from data recorded previously in 26 adults (6
Pediatric subjects were stratified into two groups based on age, infants (0 to 2 yr) and children (>2 to 12 yr), to assess if there were significant differences in BIS between infants and children as well as between pediatric and adult patients. Mean BIS and standard deviation were assessed at each milestone and compared by using Student’s two-tailed t -test for significance. During the maintenance phase, the mean BIS was calculated as a weighted average from readings taken every 5 min during this period of the study. To minimize bias, clinical end points were judged by study personnel who were not aware of the BIS value.
Power to detect a difference of 10 units in BIS at each milestone was evaluated retrospectively by using a two-sided test with significance level of 0.05.
Twenty-two patients (11 infants and 11 children) were studied while receiving sevoflurane as the primary anesthetic, and BIS measurements were recorded at five steady-state end-tidal concentrations of sevoflurane. No premedication was given. Anesthesia was induced with sevoflurane by mask in 60% N2 O/O2 or with a single bolus dose of propofol IV. BIS measurements were delayed for a minimum of 20 min after propofol. Anesthesia was maintained with sevoflurane in nitrous oxide and oxygen (50%–60% N2 O). However, no additional sedative or hypnotic drugs were given. Tracheal intubation was performed in those cases in which clinically indicated by using either mivacurium or cisatracurium to provide muscle relaxation. Ventilation was controlled or assisted to maintain end-tidal CO2 levels between 35 and 45 mm Hg. Body temperature was maintained above 35.5°C in all cases. During the maintenance phase, sevoflurane was adjusted initially to a measured end-tidal concentration of 4% and maintained at this level (±0.1%) for at least 5 min. End-tidal sevoflurane was measured (±0.1%) by using one of two calibrated monitors (Datex Ultima Capnomac or Ohmeda RGM5250; Datex-Ohmeda, Tewksbury, MA).
Because intense surgical stimuli can increase BIS, we attempted to make our measurements during periods of relatively low surgical stimulus. Four readings of BIS at each end-tidal sevoflurane concentration were recorded at 30-s intervals; BIS readings were rejected as “unstable” if the BIS varied by more than 10 units during any 30-s interval. The process was then repeated with target sevoflurane concentrations of 3.0%, 2.0%, 1.5%, 1.0%, and 0.5% in sequence. The BIS was not available to the anesthetist or used to guide the administration of anesthesia. After the completion of data collection, the conduct of the anesthetic was at the discretion of the anesthetist. Again, subjects were stratified into two groups based on age: 0 to 2 yr, and 2 to 12 yr. A prospective power analysis was done before initiation of the trial and 11 subjects per group were deemed necessary to show a 10-point BIS difference between adults and children or infants (α = 0.05 and β = 0.8). Mean BIS and standard deviation were determined at each concentration of sevoflurane. A semi-logarithmic plot of sevoflurane concentration versus BIS was generated, and data were fit to an inhibitory sigmoid Emax model:MATH by using a commercial nonlinear regression program (SlideWrite Plus© 3.0; Advanced Graphics Software, Sunnyvale, CA). E0 was specified to be 100, and the scale factor (gamma) was set at 1. Although we could not use much larger doses, it is fairly certain that BIS will become 0 at a sufficiently high concentration of sevoflurane, so Emax was constrained to 100. The EC50 is the sevoflurane concentration corresponding to half-maximal effect (a BIS of 50), and this value was estimated, with 95% confidence intervals, for each group.
Results
All enrolled patients completed the study. All data collected from the studied patients are presented here; however not all endpoints could be assessed for each patient. Patients ranged in age from 1 mo to 12 yr, and the demographics are listed in Table 1 .
Table 1: Demographics
BIS responses during general anesthesia were observed in infants and children and compared with values measured in adults. We were able to measure awake BIS in 26 unpremedicated pediatric patients. In both pediatric and adult populations, awake BIS values were between 90 and 100, then decreased precipitously after induction. BIS values usually increased from this nadir during maintenance, then increased rapidly during emergence. BIS values for case milestones are reported as mean ± SD and are presented in Table 2 .
Table 2: BIS (Bispectral Index) Values at Case Milestones in Unpremedicated Children and Infants Versus Adults
No difference between unpremedicated children and adults is seen in BIS values before the induction, during maintenance, or at emergence. BIS nadir is lower in children than adults, and this difference is statistically significant. This sample size had power >90% to detect a true difference in BIS of 10 units at all 4 comparison points. When the data are stratified by age, infants and children do not have significantly different BIS values at case milestones (Figure 1 ); however, statistical power is reduced to <70% by the smaller sample sizes.
Figure 1: Bispectral index (BIS) values at case milestones in adults, infants, and children. The nadir BIS for the combined pediatric groups was significantly lower than that for adults. Error bars represent SD.
Four children received rectal methohexital before the induction. Compared with unpremedicated children, those premedicated with the barbiturate had significantly smaller BIS values before the induction and at emergence. These data are presented in Figure 2 .
Figure 2: Bispectral index (BIS) values in children receiving methohexital (n = 4) versus unpremedicated (n = 73) children. Error bars represent SD. Pre-induct. = preinduction, Maint. = maintenance, Premed = premedication.
All 22 subjects were able to complete the concentration-response protocol, although not all data points were obtained for every patient. Figure 3 shows the individual data relating BIS to measured end-tidal concentration of sevoflurane. Each point is the average of four readings over 2 min. BIS decreased monotonically in both infants and children as the concentration of sevoflurane increased (Figure 3 ). Full axis curves are displayed in this figure, but care must be taken not to extrapolate beyond the available data points. As stated in Methods, it is quite possible that the EEG would actually be isoelectric (BIS = 0) at a sevoflurane concentration of 10%.
Figure 3: Bispectral index (BIS) values versus end-tidal sevoflurane concentration. The difference between infants and children was statistically significant (see text). The solid circles represent the children (•) and the open circles represent the infants (○).
When infants and children are compared, the EC50 of sevoflurane was slightly greater in infants than in children. This difference was statistically significant (P < 0.05), and the data are shown in Table 3 . There was no difference in the amount of opioids or muscle relaxants administered.
Table 3: EC50 of Sevoflurane in Infants and Children
Discussion
We found that BIS values in awake and anesthetized children and infants were comparable to values in adults, whether comparing with adults receiving an IV anesthetic (3,5,6 7
In these typical, uncontrolled general anesthetics, the awake values were in the 90 to 100 range, and the maintenance values were usually 40–50, just as we see in adults. As expected, methohexital premedication produced substantial sedative effects, and this was reflected in lower BIS values. In our study, the youngest subject was five weeks old; yet even this infant had an apparently “adult” response.
One important difference between our pediatric and adult data is the use of inhaled vs IV induction of anesthesia. In the observational arm of this study, the majority of children had an inhaled induction, usually with 4%–8% sevoflurane. The adult patients all received 1–2 mg/kg of propofol for the induction of anesthesia. The BIS reached a considerably lower nadir in the children, and this may reflect a relatively greater hypnotic effect or the occurrence of EEG suppression with the large concentrations of sevoflurane. The inverse relationship of BIS to the end-tidal sevoflurane concentration was monotonic over the range tested. This relationship was similar to the sevoflurane dose-response reported in adults (7
Although it is not readily apparent in Figure 3 , in four of the subjects, BIS was increased slightly when the end-tidal sevoflurane levels were 4% as opposed to 3%. We have no explanation for this, but this “near suppression” phenomenon has been reported previously in patients receiving large concentrations of volatile anesthetics (13 14,15
We also demonstrated that increased end-tidal levels of sevoflurane were required to achieve a given BIS in infants than in older children. The same age-related difference has been demonstrated previously for halothane minimum alveolar concentration (16,17
The practice of anesthesia in children is different from that in adults. Traditionally, pediatric anesthetists have relied more on volatile anesthetics than on a nitrous-opioid technique. It is possible that pediatric patients could benefit from techniques that have gained wide acceptance in adult anesthesia: the use of much smaller concentrations of volatile anesthetics for hypnosis and moderate doses of opioids for analgesia. With a monitor such as the BIS, there is some basis for titrating to hypnotic end points in children, and techniques such as these may be evaluated objectively. In this preliminary investigation, BIS appears to provide accurate clinical information in children and shows promise for use in pediatric anesthesia practice in similar ways to its use in adults.
We gratefully acknowledge the statistical support and efforts of Jeffrey C. Sigl, PhD and Paul Manberg, PhD at Aspect Medical Systems, Natick, MA and statistical advice from YuChiao Chang, PhD.
References
1. Rampil IJ. A primer for EEG signal processing in anesthesia. Anesthesiology 1998; 89:980–1002.
2. Sigl JC, Chamoun NG. An introduction to bispectral analysis for the electroencephalogram. J Clin Monit 1994; 10:392–404.
3. Glass PS, Bloom M, Kearse L, et al. Bispectral analysis measures sedation and memory effects of propofol, midazolam, isoflurane, and alfentanil in healthy volunteers. Anesthesiology 1997; 86:836–47.
4. Kearse LA, Rosow C, Zaslavsky A, et al. Bispectral analysis of the electroencephalogram predicts conscious processing of information during propofol sedation and hypnosis. Anesthesiology 1998; 88:25–34.
5. Flaishon R, Windsor A, Sigl J, Sebel PS. Recovery of consciousness after thiopental or propofol. Anesthesiology 1997; 86:613–9.
6. Gan TJ, Glass PS, Windsor A, et al. Bispectral index monitoring allows faster emergence and improved recovery from propofol, alfentanil, and nitrous oxide anesthesia. Anesthesiology 1997; 87:808–15.
7. Katoh T, Suzuki A, Ikeda K. Electroencephalographic derivatives as a tool for predicting the depth of sedation and anesthesia induced by sevoflurane. Anesthesiology 1998; 88:642–50.
8. Song D, Joshi G, White PF. Titration of volatile anesthetics using bispectral index facilitates recovery after ambulatory anesthesia. Anesthesiology 1997; 87:842–8.
9. Liu J, Singh H, White PF. Electroencephalographic bispectral analysis predicts the depth of midazolam induced sedation. Anesthesiology 1996; 84:64–9.
10. Torres F, Anderson C. The normal EEG of the human newborn. J Clin Neurophysiol 1985; 2:89–103.
11. Holmes GL, Lombroso CT. Prognostic value of background patterns in the neonatal EEG. J Clin Neurophysiol 1993; 10:323–52.
12. Scher MS, Sun M, Hatzilabrou GM, et al. Computer analysis of eeg-sleep in the neonate: methodological considerations. J Clin Neurophysiol 1990; 7:417–41.
13. Detsch O, Schneider G, Hapfelmeier G, et al. Increasing isoflurane may induce paradoxical increases in the EEG bispectral index (BIS) [abstract]. Anesthesiology 1998; 89:A363.
14. Doi M, Gajraj RJ, Mantzaridis H, Kenny GNC. Effects of cardiopulmonary bypass and hypothermia on electroencephalographic variables. Anaesthesia 1997; 52:1048–55.
15. Hayashida M, Chinzei M, Uchida K, Hanaoka K. Bispectral Index as an indicator of cerebral function during profound hypothermia [abstract]. Anesthesiology 1999; 93:A611.
16. Gregory GA, Eger EI II, Munson ES. The relationship between age and halothane requirement in man. Anesthesiology 1969; 30:488–91.
17. Nicodemus HF, Nassiri-Rahimi C, Bachman L, Smith TC. Median effective doses (ED50) of halothane in adults and children. Anesthesiology 1969; 31:344–8.
