BJA Advance Access originally published online on August 23, 2006
British Journal of Anaesthesia 2006 97(5):666-675; doi:10.1093/bja/ael223
Correlation of the A-LineTM ARX index with acoustically evoked potential amplitude
1 Department of Anaesthesiology and Intensive Care, University of Bonn Bonn, Germany
2 Department of Anaesthesiology, Charité Campus Mitte Berlin, Germany
*Corresponding author. E-mail: wenningman{at}uni-bonn.de
Accepted for publication May 24, 2006.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionConclusionReferences |
|---|
Background. Automated indices derived from mid-latency auditory evoked potentials (MLAEP) have been proposed for monitoring the state of anaesthesia. The A-LineTM ARX index (AAI) has been implemented in the A-LineTM monitor (Danmeter, V1.4). Several studies have reported variable and, in awake patients, sometimes surprisingly low AAI values. The purpose of this study was to reproduce these findings under steady-state conditions and to investigate their causes.
Methods. Ten awake unmedicated volunteers were studied under steady-state conditions. For each subject, the raw EEG and the AAI were recorded with an A-LineTM monitor (V1.4) during three separate sessions of 45.0 (1.6) min duration each. MATLABTM (Mathworks) routines were used to derive MLAEP responses from EEG data and to calculate maximal MLAEP amplitudes.
Results. The AAI values ranged from 15 to 99, while 11.4% fell below levels which, according to the manufacturer, indicate an anaesthetic depth suitable for surgery. Inter-individual and intra-individual variation was observed despite stable recording conditions. The amplitudes of the MLAEP varied from 0.8 to 42.0 µV. The MLAEP amplitude exceeded 2 µV in 75.3% of readings. The Spearman's rank correlation coefficient between the MLAEP amplitude and the AAI value was r=0.89 (P<0.0001).
Conclusions. The version of the A-LineTM monitor used in this study does not exclude contaminated MLAEP signals. Previous publications involving this version of the A-LineTM monitor (as opposed to the newer A-Line/2TM monitor series) should be reassessed in the light of these findings. Before exclusively MLAEP-based monitors can be evaluated as suitable monitors of depth of anaesthesia, it is essential to ensure that inbuilt validity tests eliminate contaminated MLAEP signals.
Keywords: index, A-Line ARX; monitoring, evoked potentials; mid-latency auditory, evoked potentials; response, post-auricular
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionReferences |
|---|
Mid-latency auditory evoked potentials (MLAEP) have been proposed as an alternative to the spontaneous EEG for monitoring the depth of anaesthesia.1 MLAEP are the electroencephalographic responses after predefined auditory stimuli and involve commonly the 10–80 ms portions of the acoustically evoked potential (AEP). Many studies suggest that certain features of the MLAEP are changed in a concentration-dependent manner by volatile and i.v. anaesthetics, e.g. amplitudes and latencies of the EEG response.2 However, the quantitative assessment of these variables critically depends on visual inspection of the waveform and manual labelling of peaks by an investigator.2 3 This process requires time, experience and is open to investigator bias.
In order to overcome these problems and to make it feasible to routinely use MLAEP in clinical practice, automated indices have been proposed.4–9 One of these, the A-LineTM ARX index (AAI), has been implemented in commercially available anaesthesia monitors, first, as the only EEG parameter (A-LineTM monitor series, Danmeter, Odense, Denmark) and, in a successor model, together with other variables based on the spontaneous EEG (A-Line/2TM monitor series). The AAI had been proposed10 11 and subsequently been tested in more than 50 publications as a single variable suitable for monitoring depth of anaesthesia (A-LineTM), suggesting a clinical performance that looks comparable with that of anaesthesia monitors based on the spontaneous EEG. Somewhat confusing has been the variability of AAI signals during different phases of anaesthesia and before induction; in awake patients very low levels10 12–17 have been recorded that according to the manufacturer suggested ‘a state of surgical anaesthesia’.18
In order to address the aspect of variability and abnormally low AAI values we decided to investigate the performance of the A-LineTM monitor under steady-state conditions. In analogy to our (unpublished) observations on awake patients before induction we repeated similar measurements under controlled conditions on awake volunteers. Contrary to previous clinical studies was our ability to simultaneously record the raw EEG signals with a specifically modified A-LineTM monitor. By correlating AAI values with the underlying raw EEG signals we are able to obtain information which should be considered when interpreting results from clinical studies conducted with the A-LineTM instead of the newer A-Line/2TM monitor series. While the A-Line/2TM no longer exclusively relies on MLAEP data, our results may provide clues as to how the performance of (completely or partly) MLAEP-based monitors can be improved.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionReferences |
|---|
Subjects
After approval from the local Institutional Ethics Committee, informed consent was obtained from 10 healthy volunteers (5 females and 5 males, American Society of Anaesthesiologists physical status I, mean age 28 yr, ranging from 22 to 51 yr). They were not under the influence of any centrally acting drugs. No subject had a history or showed signs of hearing abnormalities and responded immediately on being addressed in a quiet voice during OAA/S testing (see below), indicating that auditory input processing was fully functional.
Measurements were performed during the daytime in a quiet room with dim lighting in the presence of an investigator. The subjects were asked to lie relaxed on a couch but stay awake, although they were requested to close their eyes in order to avoid sensory distractions and minimize EEG artifacts because of eye movements. Silence was strictly maintained and no communication took place throughout the experiment except when testing for wakefulness (see below).
Wakefulness was confirmed using the Observer's Assessment of Alertness/Sedation Scale.19 A typical clinical evaluation of whether or not a patient is awake involves a verbal command, with a correctly responding patient being considered awake.20 Thus, clinically, a patient with an OAA/S score of 5 is considered awake. Subjects were asked in a quiet voice 20 min after the beginning of each recording to move the right hand, left foot, etc. This was repeated at 5 min intervals. Three recordings were obtained from each volunteer, spaced over the period of a month, the middle recordings being separated from the others by 1 or 3 weeks, respectively.
Auditory evoked potentials: recording and analysis
Auditory evoked EEG responses were recorded using the A-LineTM AEP monitor (AAI version 4.0, software version 1.4, Danmeter, Odense, Denmark). They were elicited with a binaural click stimulus of 65 dB intensity, 2 ms duration and a frequency of 9 Hz. The (bipolar) EEG was recorded between the left mastoid and a frontopolar electrode with the left forehead as reference, using silver–silver chloride gel-filled electrodes (Danmeter, Odense, Denmark) after cleaning the skin with saline and an abrasion band (3M Deutschland GmbH, Neuss, Germany). Electrode impedance was kept <5 k
and was monitored continuously.
Online monitoring of AAI
The proprietary algorithm used by the A-LineTM monitor to calculate the AAI value has been described in detail elsewhere.21 In brief, EEG sweeps after the auditory stimulus are preprocessed by the A-LineTM monitor in the window from 20 to 80 ms by applying artifact rejection and 16–150 Hz IIR Butterworth digital band-pass filtering. An autoregressive model with exogenous input (ARX model) is applied in order to minimize the number of sweeps that need to be averaged before the AAI value can be calculated. A post-smoothing of the index results in a total update delay of 6 s. This index is displayed online, but the monitor also generates a text file containing the values of the AAI and electrode impedances at a rate of 1 s–1.
Offline signal processing of raw EEG signals
The A-LineTM monitor used in this study was, in contrast to the standard commercial version, equipped with a port enabling the extraction of raw EEG data. The EEG was continuously recorded, digitized at a rate of 900 Hz and stored by the monitor. The EEG data were transferred after the measurement via HyperterminalTM (Hilgraeve, Monroe, MI, USA) to a PC for further analysis. The entire offline analysis was performed using custom written programs in MatlabTM 7.0 (Mathworks, Natick, MA, USA). The EEG output data were calibrated by using inputs of known amplitudes from a signal generator.
Data were rejected for further processing if EEG amplitudes were outside 97.5% of the maximum possible range of the amplifier. The following 440 data points were also rejected to allow the amplifier to recover from saturation. High-pass filtering was performed with a digital 6th-order Butterworth filter with cut-off frequency of 16 Hz (MatlabTM Signal Processing Toolbox 6.2, Mathworks, Natick, MA, USA). We filtered each complete (45 min) EEG record. The EEG analysis including artifact rejection was fully automated and was observer independent.
For each subject, individual EEG responses to acoustic stimuli were averaged in blocks of 512 successive sweeps containing the first 72 data points (80 ms) per sweep. No other form of data smoothing was used. The averaged data for each block, which is known as the MLAEP, represents a period of approximately 57 s each: 40–51 successive MLAEP were obtained for each subject in each session.
When the MLAEP were quantitated, the concept of peaks and troughs was not used. Instead, the peak-to-peak amplitude of any given averaged MLAEP was determined as the difference of the maximum and the minimum value within the investigated time window. The monitor writes AAI values together with the corresponding time value into a data file every second and, into a separate file, the continuous EEG data with a sampling rate of 900 Hz. Pressing the stop key on the monitor terminated the writing process and served as the time point to synchronize both data files. AAI values were averaged over the same time intervals as the corresponding MLAEP.
Statistical analyses
MatlabTM routines (Statistics Toolbox 5.0 and Curve Fitting Toolbox 1.1, Mathworks, Natick, MA, USA) were used for all statistic analyses, curve fitting and creation of figures. Values are reported as mean (SD). When AAI values were correlated with MLAEP peak-to-peak amplitudes, the Spearman's rank correlation coefficient and its P-value were calculated.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionReferences |
|---|
A-LineTM AAI
The rapidly extracted auditory evoked potential indices (AAI) were recorded from all 10 volunteers and all 30 sessions are shown in Figure 1 in their entirety (45 min each). Quite clearly, the inter-individual variability can be substantial. However, there are also examples of considerable intra-individual variability, although the recording conditions were the same. In particular, irrespective of possible differences in levels of relaxation or vigilance, each subject responded immediately on being addressed in a quiet voice during OAA/S testing. Thus, each subject always had OAA/S scores of 5 whenever tested, i.e. 20, 25, 30, 35 and 40 min after the start of the experiment.
|
Figure 2 shows the AAI values as a function of time, averaged over all 30 recording sessions. It can be seen that, at least in the beginning, the average AAI value declines with time. Note also that the first few minutes of high AAI values is the period where normally a recording would be obtained from the awake patient just before anaesthesia and surgery.
|
When all subjects and recording sessions were pooled (1350 min total), AAI values ranged from 15 to 99; artifact rejection was below 2.7% (Fig. 3). The distribution of AAI values in the histogram of Figure 3 is broad, tapering off for AAI values below 30. The AAI value of 99 is the highest value permitted by this version of the A-LineTM monitor and occurs with an artificially high frequency, which will be described in the legend of Figure 8.
|
153 min (out of 1350 min), representing 11.4% of the overall recording time for all 10 subjects. The height of bin 99 is excessively large because the algorithm of the A-LineTM monitor truncates an intermediary index leading to the final AAI index that cannot exceed a value of 99 (see also There were 11.4% of all the AAI values that were 30 or below (anaesthetic depth suitable for surgery according to the manufacturer).
Such low AAI values were not evenly distributed between subjects and recording sessions but inter-individual and intra-individual variation was observed (Fig. 4A). The percentage of a recording session with AAI values at or below 30 varied between 0 and 50%, depending on the subject and individual session. Thus, in three subjects the AAI value was never below 30 while, at the other extreme, the AAI value of Subject 5 remained at 30 or below for more than half (23 min) of the second recording session. Seven subjects showed AAI values of 30 or below during at least one session. In six subjects the AAI values decreased below 30 in each one of the three individual sessions (Fig. 4A).
|
30 and at least 6 s duration. The longest episode lasted 264 s. The average episode length was 27.6 s. The total time of these episodes was 9212 s (Not all episodes with AAI values at or below 30 were of short duration (Fig. 4B). As each AAI value is calculated from the MLAEP signals recorded over a period of 6 s (though updated every second) the shortest episode is 6 s. A histogram of all episodes (n=338) with an AAI value not exceeding 30 for at least six consecutive seconds shows that these periods can be substantially longer than 6 s, lasting up to 264 s. The average episode length with AAI values at or below 30 was 27.6 s (Fig. 4B).
MLAEP responses
In order to understand the reasons for the inter-individual and intra-individual variability of the AAI values, we inspected the raw EEG data from which the AAI values had been calculated. Examples for four different sessions are shown in Figure 5A–D. The insets show the respective AAI values as a function of time during the recording session. There were sessions where the MLAEP responses of a subject were consistently large (Fig. 5A), consistently small (Fig. 5B), where they alternated between two modes (Fig. 5C) or where they displayed the entire spectrum between large and small MLAEP responses (Fig. 5D).
|
The EEG data from which the MLAEP responses in Figure 5A–D were derived have been re-averaged for each session in a different manner: individual sweeps were sorted into one of four bands (see figure legend), depending on their simultaneously recorded AAI values. After sorting, the sweeps for each band were then averaged separately and are shown in Figure 5E–H, respectively. Clearly, the peak-to-peak amplitude of the average MLAEP response for the band with AAI values above 94 was larger than the peak-to-peak amplitude of the average response when AAI values ranged between 61 and 95 and so on. The smallest average MLAEP response was obtained for the band with AAI values below 31. This can be seen best in Figure 5H where averages of MLAEP responses from all four bands of AAI values are represented, but also in the other three examples where one or more bands of AAI values were not observed and where, therefore, the corresponding average MLAEP response is missing. It is clear that large MLAEP responses correlate with large AAI values, whereas small MLAEP responses correlate with small AAI values. That this finding is generally true for all 30 sessions can be seen in Figure 8.
A peak-to-peak amplitude for each of the 1418 MLAEP responses (representing averages of 512 consecutive sweeps of acoustically evoked EEG responses) was calculated for two time windows (1–19 and 20–80 ms) separately by subtracting the minimum value from the maximum value within a respective window. Figure 6 shows the histogram of peak-to-peak amplitudes for the time window 20–80 ms (which is the time window used by the monitor for calculating AAI values). While this distribution is skewed, it does not include an obvious outlier as did the histogram in Figure 3. The peak-to-peak amplitudes in the 1–19 ms window are considerably larger than those in the 20–80 ms window (Table 1). Indeed, some AEP amplitudes were so large that they could be detected in the EEG trace even without any averaging (Fig. 7).
|
|
|
In Figure 8 the AAI values are plotted against the peak-to-peak amplitudes in the 20–80 ms window of the MLAEP responses for all subjects and for all sessions, resulting in a Spearman's correlation coefficient, r=0.89 (P<0.0001).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionReferences |
|---|
We have demonstrated low monitor index values and variability in the AAI in awake patients using the A-LineTM anaesthetic depth monitor. This may compromise its function: a large signal variability leads to a blurring of the separation between values recorded during different states of anaesthesia. Similarly, if AAI values can be very low in the awake state, then there is perhaps not sufficient range of the index remaining to discriminate between different levels of anaesthetic depth. Such a failure to discriminate has been reported for sevoflurane 1.0–2.0 vol%10 and desflurane 3.0–6.0 vol%.18 Therefore, knowing the conditions under which large variability and low AAI values are observed, and understanding their reasons, is essential in order to optimize the sensitivity and the specificity of the monitor.
The issue whether or not purely MLAEP-based monitors can be used to monitor depth of anaesthesia remains unresolved. For example, the successor model to the A-LineTM monitor, the A-Line/2TM monitor, no longer relies entirely on MLAEP signals but switches to the analysis of the spontaneous EEG when the MLAEP signal becomes small. However, before MLAEP-based monitors can be evaluated as monitors of depth of anaesthesia it is essential to ensure that they are indeed measuring genuine MLAEP signals. This latter aspect is the main focus of this study.
Low AAI values in awake subjects
In fully alert subjects we recorded episodes of AAI values below 30, sufficiently low to suggest, according to the manufacturer, a state of surgical anaesthesia. These episodes were not observed in all individuals but only 3 out of 10 subjects did not show any episode where AAI values reached 30 or less. The number and lengths of these episodes varied within and between subjects. The longest uninterrupted episode lasted more than 4 min (264 s) in an individual in whom more than one-third of the recording session showed AAI values of 30 or less. During this episode testing for alertness occurred but the AAI was not affected and remained below 30. Low AAI scores in fully awake, unpremedicated subjects in the absence of any anaesthetic have been reported before.14 During propofol anaesthesia, Ge and colleagues15 found that the 50% predicted probability of responding to a verbal command (using the isolated forearm technique) occurred at an AAI value of 29. Other incidents of low AAI values in fully alert subjects though not being the main focus of the respective study have been seen before.10 12 22
One reason why episodes of low AAI values in awake subjects before induction of anaesthesia may not have been noticed or mentioned could be that these studies concentrated on observing the effects of anaesthetics on the AAI. In most studies where MLAEP during anaesthesia have been studied, the awake state has been recorded as baseline only for relatively short periods (if at all). Had the recording in this study been limited to the first 2 min, only one episode of a low AAI value (31) would have been observed in 1 out of 30 sessions (Fig. 2). The recording time of 45 min per session was chosen to be considerably longer than the time typically available to record from a patient before induction of anaesthesia. Extending the recording time well beyond the first 2 min was necessary to reveal incidences of low AAI in so many of our subjects, all of whom had shown ‘normal’ awake AAI values in the beginning. The initial decrease of AAI values with time (Fig. 2) may reflect a combination of reduction in muscle tension, alertness and focused attention and an increase in drowsiness. The opposite effect may be expected after testing for alertness. In agreement, averaging the AAI values 30 s before and 30 s after the verbal command (time of testing indicated in brackets), we found the AAI average to be slightly higher by 7.4 (20 min), 3.7 (25 min), 3.2 (30 min), 1.8 (35 min) and 1.2 (40 min).
Variability of the AAI
As the variability of the AAI was a central point of the study, maintenance of a steady-state without disruptions was considered important. The surroundings of the subjects did not change, neither did the subject move nor was there any outside influence on the subject except for the responsiveness tests. Testing for alertness (OAA/S) requiring a movement response from the subject presented a possible disruption and, therefore, the recording period was divided into two equal 20 min periods. During the first period, there was no testing for the OAA/S score except at the end of that period, and during the second period, subjects would be tested every 5 min. As subjects always responded immediately when addressed in a soft voice, the OAA/S score was constant throughout with a value of 5. There is thus no indication that it would have assumed any other value had the testing been conducted more frequently or at any other time point. Low AAI values were observed both during the first and the second 20 min period (Fig. 2). A short, transient increase in the AAI after the first OAA/S testing after 20 min (Fig. 2) may reflect a startle response or otherwise increased muscle tension, or increased levels of alertness which the OAA/S scale is not specific enough to resolve.
The AAI was variable (ranging from 15 to 99) not only between subjects but also for an individual subject between different recording sessions and even within a single session (Fig. 1A and B, Fig. 4A, and insets Fig. 5A–D). This variability has also been observed in unsedated subjects by Dullenkopf and colleagues.14 They observed AAI values ranging from 25 to 99. In subjects before the induction of anaesthesia the following ranges have been reported: 38–133,10 47–99,12 61–99,13 58–100,15 55–9916 and 51–99.17
Variability of the auditory evoked EEG potentials
In order to obtain an understanding why the index showed this behaviour we examined the raw EEG data from which the AAI was derived. This was possible because the A-LineTM monitors used in this study had been modified from the commercial version, thus permitting the extraction of the raw EEG data. Except for high-pass filtering and artifact rejection we did not subject the raw data to any of the other numerical algorithms used by the A-LineTM monitor such as autoregression analysis, etc. The only additional algorithm used was that of averaging blocks of 512 sweeps (ca. 57 s). The resulting MLAEP were also very variable in shape as can be seen from Figure 5.
The peak-to-peak amplitude of an averaged MLAEP signal (512 sweeps) was quantified by determining the difference between its maximum and minimum values within a certain time window. Clearly, in the 20–80 ms time window (used by the A-LineTM monitor) the sizes of the recorded MLAEP responses (peak-to-peak amplitudes) varied over a very large range from 0.8 to 42.0 µV (Fig. 6 and Table 1). Therefore, the MLAEP did not only vary in shape (Fig. 5) but they also showed large variations in their amplitudes.
The large variations observed in the AAI (Fig. 5) and the large variations observed in the MLAEP amplitudes (Fig. 6) appear to be related. Clearly, the sizes of the recorded MLAEP responses correlated well with the corresponding AAI values (Figs 5 and 8).
Large-amplitude contaminations
The average size of 5.5 µV in the 20–80 ms window (Table 1) was much greater than would be expected for the amplitude of a typical MLAEP (1 µV or less).23–26 Of all recorded MLAEP 75% had amplitudes exceeding 2 µV and thus were larger than typical MLAEP amplitudes. This implies that a great number of the recorded MLAEP responses are not pure MLAEP signals of cortical origin but that other responses have been superimposed.
When AAI values did not exceed 30 (135 occurrences), the peak-to-peak amplitudes were not greater than 2.5 µV with three exceptions. Outside this range, peak-to-peak amplitudes could become very large (Table 1), going above 40 µV in the 20–80 ms window and exceeding 100 µV in the 1–19 ms window. The peak-to-peak amplitudes could thus become much larger than is characteristic for genuine MLAEP.
It is well known that myogenic responses may seriously distort MLAEP signals (Fig. 7). Picton and colleagues27 described the post-auricular muscle reflex as a large positive peak (today's sign convention) at 11.8 ms and negative valley at 16.4 ms, pointing out that it was variable between awake subjects and even within subjects. Tooley and colleagues28 showed post-auricular muscle reflex responses of almost 30 µV between peak and valley. O'Beirne and Patuzzi29 characterized the post-auricular muscle reflex systematically and provided examples of even larger amplitudes with similarly placed electrodes (active electrode in the mastoid position, reference electrode on the forehead).
Strategies for recording ‘true’ MLAEP responses
There may be several reasons why many of the peak-to-peak amplitudes of the MLAEP recorded with the A-LineTM monitor in this study are large. Understanding these and taking counter measures should help to reduce this problem.
MLAEP are more subject to contamination by post-auricular muscle reflexes when recorded from electrodes placed at the mastoid position (as in the above examples) rather than at other positions.28 Therefore, alternatives to placing electrodes at the mastoid position should be explored in the future design of MLAEP-based monitors.
The amplitude of the post-auricular muscle reflex increases with click stimulus strength.29 30 Our subjects described the click sounds (65 dB) as very loud. The click duration of 2 ms of the A-LineTM monitor was considerably longer than commonly used in other studies (0.1–0.5 ms). As reducing the volume of the click sound (implemented in A-LineTM monitor versions later than 1.4) appears to be beneficial, click volume and click profiles should be optimized.
Because of the post-auricular muscle reflex the designers of the A-LineTM monitor did not use the 1–19 ms window to calculate the AAI10 but used the 20–80 ms window instead (as we did). Indeed, the peak-to-peak amplitudes in the 1–19 ms window are even greater (Table 1). The post-auricular muscle reflex may still influence later parts of the MLAEP for two reasons. First, repolarization of a high-amplitude post-auricular muscle reflex appears to continue beyond 20 ms. Figure 5 shows that a large peak in the 16–18 ms region is also followed by a large trough which extends well beyond 20 ms. Second, high-pass filtering may distort the post-auricular muscle reflex to such an extent that oscillations resulting from the filter algorithms may be present at time points as late as 50 ms (Fig. 5). These phenomena have been described.29 Thus, the post-auricular muscle reflex may well distort the MLAEP responses beyond 20 ms.
Therefore, it seems essential to adopt a strategy that results in the reduction of the contaminating signal as ignoring the first 20 ms of the signal and filtering of the signal does not appear to result in a MLAEP signal of purely neurogenic origin.
The last point also illustrates the importance of verifying that the signal from which an automated index is obtained is indeed an appropriate MLAEP. Steps in this direction have already been suggested and undertaken. In order to verify that the signal reflects the response to an auditory stimulus the signal-to-noise ratio of the mid-latency components of the AEP has been implemented in later versions of the A-LineTM monitor.31 Alternatively, the AEP brainstem component can be used.32 However, these approaches still do not consider the shape, the amplitude of the signal or both as a criterion for whether or not a MLAEP signal of neurogenic origin is present. The problem here is that there is poor agreement between experts regarding the signal quality of perioperatively recorded AEPs. As a consequence, results obtained by one expert may not easily be reproduced by a different expert.3
|
Conclusion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionReferences |
|---|
In conclusion, the version of the A-LineTM monitor used in this study does not exclude MLAEP signals that are contaminated by signals that are much larger than is characteristic for genuine MLAEP. The variability of the AAI correlated with the variability of the peak-to-peak amplitudes of these large signals. Previous publications involving an A-LineTM monitor of similar versions should be reassessed in the light of these findings. Before exclusively MLAEP-based monitors can be evaluated as suitable monitors of depth of anaesthesia, it is essential to ensure that they perform validity checks and thus be able to distinguish between genuine and contaminated MLAEP signals. In addition, conditions for recording neurogenic MLAEP should be improved.
|
Acknowledgments |
|---|
We thank Dr Eric W. Jensen for detailed discussions and providing us with the modified version of the A-LineTM monitor. We thank Dipl. Stat. Claudia Nicolay for help with the statistical analysis. Support was provided solely from institutional and departmental sources.
|
Footnotes |
|---|
Presented in part at the annual meeting of the European Society of Anaesthesiologists, Lisbon, Portugal, June 5–7, 2004. 
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionReferences |
|---|
1 Sebel PS, Heneghan CP, Ingram DA. Evoked responses—a neurophysiological indicator of depth of Anaesthesia? Br J Anaesth 1985; 57:841–2
2 Thornton C and Sharpe RM. Evoked responses in Anaesthesia. Br J Anaesth 1998; 81:771–81
3 Schneider G, Nahm W, Kochs EF, et al. Quality of perioperative AEP-variability of expert ratings. Br J Anaesth 2003; 91:905–8
4 Jensen EW, Nygaard M, Henneberg SW. On-line analysis of middle latency auditory evoked potentials (MLAEP) for monitoring depth of anaesthesia in laboratory rats. Med Eng Phys 1998; 20:722–8[CrossRef][ISI][Medline]
5 Kochs E, Stockmanns G, Thornton C, Nahm W, Kalkman CJ. Wavelet analysis of middle latency auditory evoked responses: calculation of an index for detection of awareness during propofol administration. Anesthesiology 2001; 95:1141–50[ISI][Medline]
6 Mantzaridis H and Kenny GN. Auditory evoked potential index: a quantitative measure of changes in auditory evoked potentials during general anaesthesia. Anaesthesia 1997; 52:1030–6[CrossRef][ISI][Medline]
7 Munglani R, Andrade J, Sapsford DJ, Baddeley A, Jones JG. A measure of consciousness and memory during isoflurane administration: the coherent frequency. Br J Anaesth 1993; 71:633–41
8 Thornton C and Jones JG. Evaluating depth of anesthesia: review of methods. Int Anesthesiol Clin 1993; 31:67–88[ISI][Medline]
9 Thornton C. Evoked potentials in anaesthesia. Eur J Anaesthesiol 1991; 8:89–107[ISI][Medline]
10 Alpiger S, Helbo-Hansen HS, Jensen EW. Effect of sevoflurane on the mid-latency auditory evoked potentials measured by a new fast extracting monitor. Acta Anaesthesiol Scand 2002; 46:252–6[CrossRef][ISI][Medline]
11 Litvan H, Jensen EW, Galan J, et al. Comparison of conventional averaged and rapid averaged, autoregressive-based extracted auditory evoked potentials for monitoring the hypnotic level during propofol induction. Anesthesiology 2002; 97:351–8[CrossRef][ISI][Medline]
12 Anderson RE, Barr G, Jakobsson JG. Correlation between AAI-index and the BIS-index during propofol hypnosis: a clinical study. J Clin Monit Comput 2002; 17:325–9[CrossRef][Medline]
13 Barr G, Anderson RE, Jakobsson JG. A study of bispectral analysis and auditory evoked potential indices during propofol-induced hypnosis in volunteers: the effect of an episode of wakefulness on explicit and implicit memory. Anaesthesia 2001; 56:888–93[CrossRef][ISI][Medline]
14 Dullenkopf A, Schmitz A, Lamesic G, Weiss M, Lang A. The influence of acupressure on the monitoring of acoustic evoked potentials in unsedated adult volunteers. Anesth Analg 2004; 99:1147–51
15 Ge SJ, Zhuang XL, Wang YT, Wang ZD, Chen SL, Li HT. Performance of the rapidly extracted auditory evoked potentials index to detect the recovery and loss of wakefulness in anesthetized and paralyzed patients. Acta Anaesthesiol Scand 2003; 47:466–71[CrossRef][ISI][Medline]
16 Kreuer S, Bruhn J, Larsen R, Hoepstein M, Wilhelm W. Comparison of Alaris AEP index and bispectral index during propofol-remifentanil anaesthesia. Br J Anaesth 2003; 91:336–40
17 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[CrossRef][ISI][Medline]
18 Kreuer S, Bruhn J, Larsen R, Bauer C, Wilhelm W. Comparison of BIS and AAI as measures of anaesthetic drug effect during desflurane-remifentanil anaesthesia. Acta Anaesthesiol Scand 2004; 48:1168–73[CrossRef][ISI][Medline]
19 Chernik DA, Gillings D, Laine H, et al. Validity and reliability of the Observer's Assessment of Alertness/Sedation Scale: study with intravenous midazolam. J Clin Psychopharmacol 1990; 10:244–51[ISI][Medline]
20 Stoelting RK and Longnecker DE. Eger EI. Minimum alveolar concentrations in man on awakening from methoxyflurane, halothane, ether and fluroxene anesthesia: MAC awake. Anesthesiology 1970; 33:5–9[CrossRef][ISI][Medline]
21 Jensen EW, Lindholm P, Henneberg SW. Autoregressive modeling with exogenous input of middle-latency auditory-evoked potentials to measure rapid changes in depth of anesthesia. Methods Inf Med 1996; 35:256–60[ISI][Medline]
22 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[CrossRef][ISI][Medline]
23 Bell SL, Smith DC, Allen R, Lutman ME. Recording the middle latency response of the auditory evoked potential as a measure of depth of anaesthesia. A technical note. Br J Anaesth 2004; 92:442–5
24 Scherg M and Volk SA. Frequency specificity of simultaneously recorded early and middle latency auditory evoked potentials. Electroencephalogr Clin Neurophysiol 1983; 56:443–52[CrossRef][ISI][Medline]
25 Newton DE, Thornton C, Konieczko KM, et al. Auditory evoked response and awareness: a study in volunteers at sub-MAC concentrations of isoflurane. Br J Anaesth 1992; 69:122–9
26 Picton TW and Hillyard SA. Human auditory evoked potentials. II. Effects of attention. Electroencephalogr Clin Neurophysiol 1974; 36:191–9[CrossRef][ISI][Medline]
27 Picton TW, Hillyard SA, Krausz HI, Galambos R. Human auditory evoked potentials. I. Evaluation of components. Electroencephalogr Clin Neurophysiol 1974; 36:179–90[CrossRef][ISI][Medline]
28 Tooley MA, Stapleton CL, Greenslade GL, Prys-Roberts C. Mid-latency auditory evoked response during propofol and alfentanil anaesthesia. Br J Anaesth 2004; 92:25–32
29 O'Beirne GA and Patuzzi RB. Basic properties of the sound-evoked post-auricular muscle response (PAMR). Hear Res 1999; 138:115–32[CrossRef][ISI][Medline]
30 Fox JE, Peyton MB, Ragi E. Lability of the postauricular and inion microreflexes, studied in the normal human subject. Electroencephalogr Clin Neurophysiol 1989; 72:48–58[CrossRef][ISI][Medline]
31 Martinez P, Trillo L, Jensen E, Rodriguez B, Litvan H. Validation of an algorithm for auditory evoked potential detection quality. Eur J Anaesthesiol 2002; 19:A-94, 24
32 Schneider G, Kochs EF, Arenbeck H, Gallinat M, Stockmanns G. Signal verification of middle latency auditory evoked potentials by automated detection of the brainstem response. Anesthesiology 2004; 101:321–6[CrossRef][ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||