| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Department of Anesthesiology, University Hospital Regensburg, Germany
Address correspondence and reprint requests to Frank Weber, MD, Department of Anesthesiology, University Hospital Regensburg, Franz-Josef-Strauß-Allee 11, D-93053 Regensburg, Germany. Address e-mail to frank.weber{at}klinik.uni-regensburg.de.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The AEP Monitor/2 (Danmeter A/S, Odense, Denmark), a recently commercialized system for depth of anesthesia monitoring, extracts the MLAEP from the EEG-signal by using an autoregressive model with an exogenous input adaptive method (4). A monitoring variable indicating the patient’s hypnotic state, the so-called composite AAITM (AAI), is then calculated from the MLAEP and the EEG (5).
Several recent studies with the A-LineTM Monitor (Danmeter A/S, Odense, Denmark), an earlier version of the AEP Monitor/2, suggested that the AAI might be helpful in distinguishing between the awake and unconscious state and in the detection of intraoperative awareness with recall (6–10).
Schmidt et al. (11) reported on the inability of the A-Line monitor to detect unnoticed disconnection of the headphones (HP) from the patients’ ears during anesthesia, leading to the danger of falsely interpreting low AAI values attributed to HP disconnection as a deep stage of anesthesia. The new AEP Monitor/2 no longer exclusively relies on the auditory stimulus for AAI calculation. In case of low AEP signal quality, the AAI is calculated from the spontaneous EEG activity, whereas disconnection of the HP core from the monitor results in an immediate disconnection alert on the screen; furthermore, index calculation is interrupted.
The purpose of this study was to evaluate the impact of acoustic stimulation on AAI calculation under awake and anesthetized conditions and at the moment of transition from consciousness to unconsciousness. We hypothesized that HP disconnection from the patients’ ears results in immediate significant alterations of AAI values under both awake and anesthetized conditions. The primary end-point of our study was the alteration of the AAI after HP disconnection from the patients’ ears under awake and anesthetized conditions.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
A standardized total IV anesthesia (TIVA) technique was performed in all patients. A slow bolus injection of remifentanil 0.5 µg/kg was given over 30 s, followed by a constant infusion at a rate of 0.2 µg · kg–1 · min–1. Two minutes later anesthesia was induced with a target-controlled infusion of 3.0 µg/mL propofol (Orchestra® Base Primea; Fresenius-MCM GmbH, Alzenau, Germany). After loss of consciousness (LOC) and establishment of manual ventilation via face mask, a laryngeal mask airway was inserted, and patients were mechanically ventilated to maintain end-tidal carbon dioxide concentrations of 35–40 mm Hg. Patients did not receive neuromuscular blocking drugs. Both remifentanil infusion and propofol target-controlled infusion were left unchanged, and patients remained undisturbed until the end of the study period.
Vital signs were monitored using a three-lead electrocardiogram, pulse oximetry, noninvasive automatic arterial blood pressure and measurements of inspired and end-tidal concentrations of oxygen and carbon dioxide.
AAI monitoring was started with the awake patients over 5 min with HP correctly positioned (cHP) and then over 5 min with HP disconnected from the patients’ ears (dHP). The cord of the HP was kept connected to the monitor throughout the study. For induction of anesthesia, patients were randomly allocated by a computer-generated list to one of two study groups: induction was performed either under cHP conditions (group C; n = 8) or dHP conditions (group D; n = 8), and AAI values at the moment of the start of the induction (AWAKE) and at the moment of LOC, defined as the moment of loss of responsiveness to verbal command, were specifically recorded.
Ten minutes after laryngeal mask airway insertion AAI data were recorded with cHP over 5 min, followed by another 5-min period with dHP, both under steady-state anesthetic conditions.
The Narcotrend® (MT MonitorTechnik, Bad Bramstedt, Germany; software version 4.0), an EEG monitor designed to measure the effects of anesthetics on the brain in terms of "depth of anesthesia," was simultaneously used. Several studies have indicated the reliability of the Narcotrend for monitoring depth of hypnosis (12–14).
MLAEP were recorded using the AEP Monitor/2 (Danmeter A/S, Odense, Denmark; software version 1.6). After the skin was prepared with alcohol and abraded with gauze, three silver-silver chloride electrodes (Medicotest A/S, Olstykke, Denmark) were positioned at the mid forehead (+), left forehead (reference), and left mastoid (-). Electrode placement and skin preparation were performed until the electrodes’ impedance was <1000 Ohms. The MLAEP were elicited with a binaural click stimulus of 2 ms duration and repetition rate of 9 Hz. The MLAEP analysis window was 20–80 ms. A dimensionless index, the so-called composite AAI, ranging from 99 (patient fully awake) to 0 (very deep hypnosis) is then calculated from the MLAEP or the EEG, depending on the signal-to-noise (SNR) ratio. In case of an SNR ratio <1.45, the AAI is no longer calculated from the AEP but is based on an analysis of the ß-ratio and the extent of burst suppression of the spontaneous EEG.
AEP Monitor/2 data were transferred to a personal computer for subsequent analysis with the AAI GraphTM software package (Danmeter A/S; version 2.0).
After skin preparation with alcohol and gauze, two silver-silver chloride electrodes (Medicotest A/S, Olstykke, Denmark) were positioned on the left and right lateral parts of each patient’s forehead with the maximum achievable distance and a third electrode was positioned on the mid-forehead, serving as a referential electrode. Electrode placement and skin preparation were performed until the electrodes’ impedance was <6000 Ohms. The EEG of each patient was recorded continuously with 128 samples per second, a 0.5-Hz high-pass filter, a 45 Hz low-pass filter, and a 50 Hz notch filter, using the Narcotrend® EEG monitor (MT, MonitorTechnik). Narcotrend EEG processing leads to a variable called the Narcotrend Index (NI), a dimensionless scale from 0 (very deep hypnosis) to 100 (wakefulness). Detailed information about the development of the Narcotrend algorithm has been given by Schultz et al. (15,16).
Sample size calculation indicated that 14 patients would be required to detect a 20% difference in awake AAI values with cHP versus dHP with a power of 90% at an 
-level of 0.05. To compensate for nonevaluable patients, 16 patients were enrolled in the study.
Data were tested for normality using the Kolmogorov-Smirnov method. AAI and NI values at different study conditions were compared using one-way repeated-measures analysis of variance or the Friedman test as appropriate. In case of significant differences Dunnett’s post hoc test was performed to account for multiple testing. Data were analyzed using SigmaStat, version 3.0 (Systat Software GmbH, Erkrath, Germany). Results were considered significant when P < 0.05.
The ability and accuracy of predicting the clinical end-points "Awake" and "LOC" were evaluated using the prediction probability (PK), which compares the performance of indicators having different units of measurements, as described by Smith et al. (17). PK was calculated using a custom spreadsheet macro, the PK-MACRO, described and provided by Smith et al. (17). We used the jackknife method to compute the standard error of the estimate. A PK value of 1.0 means that the parameter (e.g., the AAI or NI) predicts the states (e.g., Awake versus LOC) correctly 100% of the time. A PK value of 0.5 means that the prediction is no better than chance alone. A PK value of <0.5 indicates an inverse relationship.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Compared with cHP AAI values (median [interquartile range (IQR)]) observed 1 min before disconnection of HP (78.5 [63.3–96.8], awake dHP AAI values were significantly lower, showing an immediate decline during the first minute under dHP conditions (66.4 [56.1–79.4]) and stable conditions during the next 4 min. Under anesthetized conditions a moderate decline in AAI values under dHP conditions was observed from the second to fifth minute. NI values were not affected by HP disconnection. Detailed information on the progression of AAI and NI values under cHP and dHP conditions are given in Fig. 1. AAI values (mean ± sd) under awake conditions with cHP (75 ± 15) and dHP (58 ± 16) declined to cHP (51 ± 9) and dHP (43 ± 7) at the moment of LOC. NI values were 96 ± 1 under awake conditions, declining to 48 ± 24 at the moment of LOC. Results of PK analysis for discrimination between Awake and LOC are displayed in Figure 2.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Schmidt et al. (11) recently reported the inability of the so called "click detection" of the Alaris AEPTM Monitor to detect inadvertent disconnection of the HP during anesthesia. The AAI was exclusively calculated from the AEP signal and therefore the "click detection" was implemented to account for inadvertent disconnection of the HP from the patients’ ears. In their study, Schmidt et al. (11) found significant decreases of AAI values as a consequence of HP disconnection from 76 ± 21 to 28 ± 19 (mean ± sd) under awake conditions, whereas under anesthetized conditions the AAI remained unaltered after HP disconnection. The authors concluded that, in anesthetized patients, low AAI values attributed to HP disconnection and missing auditory stimulation could be falsely interpreted as a deep stage of anesthesia, and in these particular cases an "AAI controlled" reduction of anesthetics would implicate the risk of decreasing anesthesia and awareness.
To make our results comparable to those reported by Schmidt et al. (11), we used an almost identical experimental setup with respect to AAI recordings and the TIVA regime. The AEP Monitor/2 is not only an AEP monitor but also an EEG monitor. In case of low AEP quality, the AEP Monitor/2 calculates the AAI exclusively from the spontaneous EEG and with increasing signal quality from the AEP again. The "signal quality bar" of the AEP Monitor/2, as an indicator of the quality of the extracted signal, provides the anesthesiologist with the necessary information about the actual source of AAI calculation. Formally, AAI calculation with the AEP Monitor/2 does not require acoustic stimulation (attachment of the HP to the patient). Under these circumstances the SNR is 1 and the AAI is calculated from the spontaneous EEG. As a consequence, the Alaris AEP monitor’s "click detection," which was based on the estimation of the SNR, is no longer integrated into the new AEP Monitor/2.
In this study, under awake conditions, as a consequence of artifact contamination of the AEP, we observed frequent fluctuation between AEP-derived and EEG-derived AAI values, sometimes more than 2 shifts within 10 seconds. Under awake conditions high electromyographic activity, as displayed on the screen of the monitor, is a common finding and is most likely to cause artifact rejection of the AEP signal. Unfortunately, the AAI GraphTM software package does not distinguish between AEP-derived or EEG-derived AAI values, and we were unable to mark all transitions manually with the event marker menu of the monitor. Although no formal calculation of the ratio of AEP-derived and EEG-derived AAI values could be performed, we are able to state that under both awake and anesthetized cHP conditions AAI values were mainly AEP derived, with an immediate turnover to 100% EEG-derived AAI values under dHP conditions.
We observed significant decreases in AAI values as a consequence of transition from AEP-derived to EEG-derived index calculation and increases with a turnover from EEG-derived to AEP-derived calculation both under awake and anesthetized conditions. These differences should be interpreted with great caution because, according to the manufacturer of the AEP Monitor/2, AAI values of 50 or more are indicative of the "awake" state, with floating changes to "light anesthesia" in the AAI range of 30, "surgical anesthesia" in the range of 15–25, and "deep anesthesia" indicated by AAI values <15. We observed median (IQR) AAI values of 73 (59–89) under awake cHP conditions, declining to 58 (50–66) under dHP conditions. This finding is in contrast to the data for the Alaris AEP reported by Schmidt et al. (11), where awake AAI values under dHP conditions (28 ± 19) were within a range indicative for light to surgical anesthesia. Under anesthetized cHP conditions we observed AAI values of 20 (16–28), declining to 15 (13–18) under dHP conditions. AAI values under cHP and dHP conditions remained within the typical range for the particular clinical condition (awake or anesthetized).
We conclude that, in contrast to the former Alaris AEPTM Monitor, the new AEP Monitor/2 no longer carries the risk of false assignment of a patient to a deep hypnotic state as a consequence of inadvertent HP disconnection. There are differences between the AEP-derived and the EEG-derived AAI, however, and anesthesiologists should be aware that, especially under anesthetized conditions, sudden changes of the AAI in the observed range of approximately 5 U are not necessarily indicative of a change of anesthetic depth, but may be a consequence of a turnover from AEP-derived to EEG-derived (or vice versa) AAI calculation.
Chan et al. (21) recently reported data on the performance of the AEP Monitor/2 under cHP and dHP conditions. They measured awake cHP-AAI (mean ± sd) of 89 ± 8, declining to 61 ± 4 after HP disconnection and found no alterations of AAI values under anesthetized conditions. The awake cHP AAI data measured by Chan et al. were remarkably higher, with a much smaller sd than reported by Schmidt et al. or in this study. We have no explanation for these contradicting findings. Awake dHP AAI data reported by Chan et al. (21) were comparable to those in our study.
Although not a primary end-point of the study, we also evaluated the performance of the AEP Monitor/2 with respect to its ability to detect the transition from consciousness to unconsciousness during induction of anesthesia under both cHP (PK 0.95) and dHP (PK 0.83) conditions. Interestingly, we observed a PK value of 0.88 for the Narcotrend under cHP conditions, whereas under dHP conditions PK was 1.0. We are unable to explain these findings. Further efforts should be made to address this issue.
The AEP Monitor/2 is currently the only commercially available depth of anesthesia monitor based on the analysis of MLAEP. The results of this study have, therefore, no implications for other types of depth of anesthesia monitors.
In conclusion, the combination of AEP-derived and EEG-derived AAI calculation appears to be an important and clinically relevant feature of the new AEP Monitor/2. However, the observed different AAI values under AEP-derived or EEG-based calculation show that further improvement of the system with a better harmonization of the two ways of index generation would be desirable to prevent anesthesiologists from misinterpreting the evolution of the patient’s hypnotic state.
|
Footnotes |
|---|
Accepted for publication January 19, 2005.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
