The auditory evoked response as an awareness monitor during anaesthesia

E. Loveman1, J. C. Van Hooff2 and D. C. Smith3

1Wessex Institute for Health Research and Development, University of Southampton, Bassett Crescent East, Southampton SO16 7PX, UK. 2Department of Psychology, University of Portsmouth, King Henry I Street, Portsmouth PO1 2DY, UK. 3Department of Anaesthesia, Southampton General Hospital, Tremona Road, Southampton SO16 6YD, UK*Corresponding author

Accepted for publication: November 14, 2000


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
We investigated the relationship between the latency of the Nb wave of the auditory evoked response (AER) and periods of awareness during propofol anaesthesia. In the anaesthetic room before cardiac surgery the AER was recorded continuously in 14 patients. Awareness was measured by the ability of the patient to respond to command using the isolated forearm technique (IFT). The Nb latencies were shorter when the patients were able to respond than at loss of response (P<0.001). In six patients who repeated this transition from response to loss of response, there was a high and significant correlation between Nb latencies. None of the patients had any recollection of events after the initial induction of anaesthesia as measured by explicit and implicit memory tests. These results suggest that the Nb latency of the AER may represent an indication of awareness in individual patients, but wide inter-patient variability limits its practical usefulness. In addition, because no evidence of memory was demonstrated, even when patients were known to be awake, the relationship between AER and memory processing remains unclear.

Br J Anaesth 2001; 86: 513–8

Keywords: brain, auditory evoked response; anaesthetics i.v., propofol; measurement techniques, isolated forearm technique; anaesthesia, depth; memory, implicit


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Changes in the middle latency auditory evoked response (AER) reflect changes in anaesthetic depth,13 and to some extent changes as a result of surgical stimulation (Fig. 1).4 5 Some studies have attempted to use the AER to distinguish between consciousness and unconsciousness during target-controlled infusions of propofol.6 7 The AER technique, however, has rarely been correlated with objective evidence of awareness such as the isolated forearm technique (IFT),8 or with implicit memory.9 10



View larger version (12K):
[in this window]
[in a new window]
 
Fig 1 Schematic representation of the AER.

 
We examined the correlation between the latency of wave Nb of the AER and periods of awareness, determined by the ability to respond to command with an isolated forearm. The response to verbal command can be distinguished from the response to surgical stimulation, as it must involve some degree of cognitive processing, rather than a simple reflex type movement. To establish the extent of this cognitive processing we also presented lists of words before the induction of anaesthesia and during periods of awareness for use in post-operative tests of implicit memory. Previous evidence has suggested that autonomic reflexes are not reliable measures of anaesthetic depth,11 but these are still used in everyday anaesthetic practice. We therefore also observed the relationship between episodes of awareness and arterial pressure.


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
After approval for the study from the Southampton Joint Research Ethics Committee and written informed consent, we studied 14 patients (aged 54–75 yr, ASA II or III) having cardiac surgery. Patients were introduced to the study by means of an information sheet explaining the use of the AER and IFT, and they were informed that a number of word lists would be played during the anaesthetic room procedure. Patients with poor left ventricular function, more than moderate valvular stenosis, hearing difficulties or any past medical history of epilepsy or mental illness were excluded from the study, as were those with a high pre-operative anxiety score as measured by the Hospital Anxiety and Depression Scale.12

Pilot studies in spontaneously breathing patients showed that it was impossible to record a meaningful AER when the patients were awake or lightly sedated because of interference from muscle artefact. We therefore used response to command with the IFT as our indicator of awareness. To eliminate the confounding effect of surgical stimulation, the entire procedure was undertaken in the anaesthetic room before surgery. After pre-medication with oral diazepam 10 mg, i.m. morphine 10 mg and prochlorperazine 12.5 mg patients were taken to the anaesthetic room approximately 1 h before the scheduled time for surgery.

In the anaesthetic room full invasive monitoring with continuous recording, including automated ST segment analysis, was started according to standard practice in our unit. The patient’s free left upper arm was wrapped with soft surgical padding (Soffban; Smith and Nephew) and a padded orthopaedic tourniquet was placed over this. The AER was monitored using the Northwick Park system, as described previously.13 Briefly, the EEG was recorded with a prototype amplifier, digital signal processor system (Loughborough Sound Instruments, Loughborough, UK) and bespoke software running on a 386SX personal computer. Disposable silver/silver chloride electrodes (MSB) were placed on Fpz and the two mastoids as in the International 10:20 system.14 Closely fitting binaural ear pieces were used for delivery of both words and the auditory stimuli (clicks) for evoking the AER. The AER was generated by averaging 512 EEG epochs of 125 ms duration beginning at the click stimuli.

Sixteen words with comparable occurrence frequencies were divided into four lists. The words were selected from a previous implicit memory study in non-anaesthetized volunteers, in which they were found to have similar priming rates (Loveman, van Hooff and Gale, submitted for publication). Each list of four words was recorded onto a compact disc and copied five times. For the post-operative implicit memory tests the same female voice was used for the recording. A compact-disc player presented the first list of words (five times) before induction of anaesthesia with fentanyl 100 µg and propofol 1 mg kg–1, followed by an infusion of propofol at 2 mg kg–1 h–1. As soon as consciousness was lost (absence of eyelash reflex) the tourniquet was inflated to 300 mm Hg and vecuronium 0.1 mg kg–1 was administered. After tracheal intubation, mechanical ventilation of the lungs was commenced and adjusted to an end-tidal carbon dioxide tension of 4.5–5.0 kPa. When the AER was considered stable, the propofol infusion was turned off. At 1 min intervals the patient was called by their name and asked to squeeze the experimenter’s hand. The request for response to command was made individually to each patient by the same experimenter. If patients had not responded after 20 min the tourniquet was deflated for 10 min and then re-inflated for a further 20-min period. Thus, the maximum inflation time for any patient was 40 min and all patients responded within this period.

When the first unequivocal response was made, the patient was reassured and informed that they were now going to hear another word list. This word list was again played five times. To enable clear hearing, the commands to squeeze the experimenter’s arm were stopped during presentation of the word list. After presentation of the word lists, the propofol infusion was started again at a rate of 6 mg kg–1 h–1 for 10 min, reducing to 4 mg kg–1 h–1 for a further 10 min and then 3 mg kg–1 h–1 until the response to command was lost. All patients remained responsive to command after presentation of the word list. Testing for response continued each minute and when the response to command had been lost for 3 consecutive min the patient was then played the third word list (five times). All patients were unresponsive to command after the presentation of the third word list. A fourth and final word list was not presented to serve as a baseline measure and the order of the four word lists was rotated between patients. In six patients there was sufficient time to stop the propofol infusion again and repeat the above procedure without the word lists. The patients were then prepared for surgery.

Within 36 h of surgery, after having been transferred to the main ward, patients were assessed for both explicit and implicit memory. To establish any explicit recall of intra-anaesthetic events patients were asked the following four questions:15

What is the last thing that you remember before you were put to sleep for your operation?

What is the first thing that you remember on waking from your operation?

Do you remember hearing anything during the operation?

Did you dream of anything during the operation?

To assess implicit memory (defined as memory without conscious recollection16) patients were presented with 18 pre-recorded spoken word stems, corresponding to the first syllables of the 16 stimulus words of the four lists (e.g. PIC from PICNIC or MEL from MELLOW), plus two practice stems. They were asked to verbally complete each stem with the first word that sprang to mind (e.g. PIC can become PICNIC but also the more typical words such as PICTURE or PICKLE). The word stem completion test has been used successfully in previous anaesthetic studies17 and is a commonly used method for assessing implicit memory in psychological investigations (for a review see Schacter16). Furthermore, the current word stem completion test, including the specific stimuli, the auditory presentation mode and the verbal response format, was previously validated in a study using healthy volunteers (Loveman, van Hooff and Gale, submitted for publication). In addition to this, a forced choice recognition task was undertaken. Patients were played a list of 28 words (the 16 words from the four lists plus 12 fillers) and asked to respond if they felt they remembered hearing any, or if any seemed familiar. This technique has also been used in anaesthetic studies.17 18 Finally, explicit memory was reassessed with the question, do you remember squeezing the experimenter’s hand at any time during the anaesthetic period?

Measurement of the latency of the Nb wave was undertaken off line by each author independently. Repeatability between us was significant, (r=0.93, P<0.05) and analysis was therefore continued with the scoring of one rater.


    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
IFT
Despite the potential problems,11 we found the IFT easy to apply, and had no difficulty using it or interpreting responses. The average time of first response to command was 15 min, with a range of 5–30 min. No patients were responsive at the first time of asking. In all patients an unequivocal response was given, with all patients being able to squeeze the experimenter’s hand. No reported problems with the isolated arm were documented post-operatively. Patients stopped responding to command after commencement of the propofol infusion at between 4 and 10 min (mean time 6 min).

AER
The mean latency of Nb was 48.4 (SD 4.3) ms at first response to command and 54.6 (6.7) ms at loss of response to command. A paired samples t-test revealed that this difference in Nb latency was statistically significant (t=–5.11, df 13, P<0.001). The individual Nb latencies for each of the patients are shown in Figure 2. In all but two patients (patients 9 and 10) the Nb latencies were longer at loss of response than at response. However, the range of Nb latencies were quite large which illustrates the problem of specifying a general cut-off point. For example, a cut-off point of 60 ms, as suggested by Schwender and colleagues,19 would in our study lead to a sensitivity of 100% (percentage of patients for which the presence of a response to command could be correctly predicted) but also to a very low specificity of 29% (percentage of patients for which the absence of a response to command could be correctly predicted). Similarly, a cut-off point of 51 ms20 would result in a sensitivity of 79% and a specificity of 71%, and a cut-off point of 44.5 ms8 would results in a sensitivity of 21% and a specificity of 93%.



View larger version (28K):
[in this window]
[in a new window]
 
Fig 2 Individual changes in Nb wave latency with and without response to command of the isolated forearm.

 
In six patients we had time to make a second measurement of Nb latency with and without response to command (Table 1). Nb latencies for the first response (i.e. first measurement) were not different from those for the second response (i.e. second measurement) (t=0.425, df 5, P=0.68). Furthermore, there was a high correlation between Nb latencies for first and second response (r=0.80, P<0.05). Similarly, Nb latencies for the first loss of response were not different from those for the second loss of response (t=–1.81, df 5, P=0.13), and the values correlated highly with each other (r=0.85, P<0.05).


View this table:
[in this window]
[in a new window]
 
Table 1 Individual ranges of Nb latency in six patients with response to command and at loss of response
 
Arterial pressure measurements
Arterial pressure measurements are shown in Table 2. A repeated measures ANOVA showed a significant main effect of recording period (F(2,22)=9.66, P<0.001). Planned comparisons showed significant differences between baseline and the first response to command of the isolated arm (t=–2.19, df=11, P<0.05), baseline and when responses were no longer present (t=2.43, df=11, P<0.05) and between the first response and first loss of response (t=4.06, df=11, P<0.01).


View this table:
[in this window]
[in a new window]
 
Table 2 Arterial pressure measurements (mm Hg) for 14 patients obtained at commencement of the study (baseline), on the first movement and when the movement had ceased
 
Explicit and implicit memory measures
Results of the five questions designed to establish evidence of explicit memory suggested that patients had no recollection of events during the study period, despite the fact that each patient made several responses to command during the study. In eight patients, the last thing remembered before induction of anaesthesia was before arriving in the anaesthetic room. The remaining patients had some recollection of i.v. cannulation procedures and only two patients could remember that a word list was played before induction of anaesthesia. These two patients were nevertheless unable to correctly recall the individual words. All patients, however, were responsive and co-operative when they arrived in the anaesthetic room. There was no evidence of implicit memory for any of the presentation conditions, including those for which it was known that patients were awake (i.e. before induction of anaesthesia and when responsive to command).


    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
A significant difference was demonstrated between the mean latency of Nb when movement occurred in response to command on the isolated forearm and when it ceased. In an early study Thornton and colleagues8 compared the latency of the Nb wave with responses on the IFT in seven patients. Using light anaesthesia with N2O, they suggested a threshold of 44.5 ms as a cut-off for awareness as demonstrated by the IFT. In two other studies, using averaged group data, an Nb threshold of 60 ms was predictive of cessation of spontaneous movement during both propofol and isoflurane anaesthesia without neuromuscular block19 and in the region of 50 ms with no response to command or post-operative recall during isoflurane anaesthesia.20 In our study the range of Nb latencies at which the transition occurred from the anaesthetized state to first response and from first response to no response of the isolated forearm was quite large. Some patients had no responses at latencies where others remained responsive. This finding hinders the definition of a general cut-off point in the latency of the Nb wave as a method of measuring awareness during anaesthesia.2 With regard to our results, in order to reduce the possibility that any patients were deemed to be asleep but were actually awake, a cut-off point of 58 ms would be required (sensitivity 100%). However, this would mean that 71% of patients (specificity of 29%) would have to be anaesthetized to a point much deeper than just at loss of response.

High correlations were found between Nb latencies for the first and second measurements in all six patients who repeated the transition from response to no-response. This means that there is minimal hysteresis in the AER, with Nb latency at both periods of response to command and at both periods of loss of response to command being largely similar. This finding, although based on a small number of patients, adds further weight to the possibility of providing individual rather than general cut-off points for awareness during anaesthesia. However, the practical implications need to be the focus of further investigation.

Measurements of arterial pressure at response to command were significantly different from baseline (pre-anaesthesia), and at loss of response to command. It should be noted, however, that arterial pressure was not monitored throughout the experimental procedure, and that changes were measured at two single points in time, for example, response and no response of the isolated forearm. Thus, we cannot comment on what the arterial pressure was at other times.

Measures of memory
We found no evidence of implicit or explicit memory during anaesthesia. In the word stem completion task, manifestations of implicit memory would be a tendency to complete the word stems with the words recently encountered (i.e. before and during anaesthesia), but this was not found. Similarly, in the forced choice recognition task, implicit memory would be if patients recognized at least some of the presented words based upon feelings of familiarity, but this was not found. These results suggest that there is no memory, explicit or implicit, during anaesthesia. However, this does not account for the lack of priming effects for the word list presented in the pre-medicated state before anaesthesia induction, or for the word list presented during the anaesthesia period while patients were responsive to command, using the IFT. This finding is comparable to Russell’s investigations with the IFT.21

Some studies using the word stem completion test17 have demonstrated memory effects during general anaesthesia, and others have provided no evidence for implicit memory.22 23 Similarly, a number of studies using sub-anaesthetic doses of anaesthesia in volunteers have found an inconsistent relationship between responsiveness during sedation and later recall, as measured by implicit or explicit measures of memory.24 One difficulty with all these studies is that the depth of anaesthesia has not been assessed at the time of presentation. This means that it cannot be ruled out whether the memory effects demonstrated were because of period of inadequate or light anaesthesia.

Perhaps the implicit memory test may not be a reliable enough measure over such a period of time. Roediger and colleagues,25 however, reported that implicit memory, as demonstrated by a word stem completion test, could persist for up to 1 week in normal populations. The lack of implicit memory in our study may reflect an effect on memory from prolonged anaesthesia and sedation (as patients are all sedated up to and during their stay on the intensive care unit).

In summary, we found a correlation between changes in wave Nb of the AER with changes in awareness as measured by the IFT. With the exception of two patients, individual latencies changed in the same direction with increasing depth of anaesthesia, although the variation between individuals was wide. This means that some patients were responding where others had stopped responding and as such reduces the application of the AER for anaesthetic monitoring. No evidence of memory for inter-anaesthetic events was demonstrated.


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
1 Thornton C, Heneghan CPH, James MFM, Jones JG. The effects of halothane and enflurane with controlled ventilation on the auditory evoked potentials. Br J Anaesth 1984; 56: 315–23[Abstract]

2 Thornton C, Newton DEF. The auditory evoked response: a measure of depth of anaesthesia. In: Jones JG, ed. Depth of Anaesthesia. Clinical Anaesthesiology. London: Bailliere Tindall, 1989; 559–85

3 Schwender D, Klasing S, Madler C, Poppel E, Peter K. Midlatency auditory evoked potentials and purposeful movements after thiopentone bolus injection. Anaesthesia 1994; 49: 99–104[ISI][Medline]

4 Thornton C, Koniecko KM, Jones JG. Effect of surgical stimulation on the auditory evoked response. Br J Anaesth 1988; 60: 372–78[Abstract]

5 Schwender D, Golling W, Klasing S, Faber-Zullig E, Poppel E, Peter K. Effects of surgical stimulation on midlatency auditory evoked potentials during general anaesthesia with propofol/fentanyl, isoflurane/fentanyl and flunitrazepam/fentanyl. Anaesthesia 1994; 49: 572–78[ISI][Medline]

6 Davies FW, Mantzaridis H, Kenny GN, Fisher AC. Middle latency auditory evoked potentials during repeated transitions from consciousness to unconsciousness. Anaesthesia 1996; 51: 107–13[ISI][Medline]

7 Doi M, Gajraj RJ, Mantzaridis H, Kenny GN. Relationship between calculated blood concentration of propofol and electrophysiological variables during emergence from anaesthesia: comparison of bispectral index, spectral edge frequency, median frequency and auditory evoked potential index. Br J Anaesth 1997; 78: 180–4[Abstract/Free Full Text]

8 Thornton C. Barrowcliffe MP, Koniecko KM, Ventham P, Dore CJ, Newton DEF, Jones JG. The auditory evoked response as an indicator of awareness. Br J Anaesth 1989; 63: 113–15[Abstract]

9 Schwender D, Kaiser A, Klasing S, Peter K, Poppel E. Midlatency auditory evoked potentials and explicit and implicit memory in patients undergoing cardiac surgery. Anesthesiology 1994; 80: 493–501[ISI][Medline]

10 Villemure C, Plourde G, Lussier I, Normandin N. Auditory processing during general anaesthesia: A study with an implicit memory task and auditory evoked potentials. In: Sebel PS, Bonke B, Winograd E, eds. Memory and Awareness in Anesthesia. Englewood Cliffs, Prentice Hall, 1993; 99–106

11 Russell IF. Conscious awareness during general anaesthesia: relevance of autonomic signs and isolated arm movements as guides to depth of anaesthesia. In: Jones JG, ed. Depth of Anaesthesia. Clinical Anaesthesiology. London: Bailliere Tindall, 1989; 511–32

12 Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psych Scand 1983; 67: 361–70[ISI][Medline]

13 Hett DA, Smith DC, Pilkington SN, Abbott TR. Effect of temperature and cardiopulmonary bypass on the auditory evoked response. Br J Anaesth 1995; 75: 293–6[Abstract/Free Full Text]

14 Cooper R, Osselton JW, Shaw JC. EEG Technology. London: Butterworths & Co Ltd, 1980

15 Ghoneim MM, Block RI. Learning and consciousness during general anesthesia. Anesthesiology 1992; 76: 279–305[ISI][Medline]

16 Schacter DL. Implicit memory: history and current status. J Exp Psychol: Learning, Memory Cognition 1987; 13: 501–18[ISI]

17 Bonebakker AE, Bonke B, Klein J, et al. Information processing during balanced anaesthesia: evidence for unconscious memory. Memory Cognition 1996; 24: 766–76[ISI][Medline]

18 Millar K, Watkinson N. Recognition of words presented during general anaesthesia. Ergonomics 1983; 26: 585–94[ISI][Medline]

19 Schwender D, Daunderer M, Mulzer S, Klasing S, Finsterer U, Peter K. Midlatency auditory evoked potentials predict movements during anesthesia with isoflurane or propofol. Anesth Anal 1997; 85: 164–73[Abstract]

20 Newton DE, Thornton C, Konieczko KM, et al. AER and awareness: a study in volunteers. Br J Anaesth 1992; 69: 122–9[Abstract]

21 Russell IF. Midazolam-alfentanil: an anaesthetic? An investigation using the isolated forarm technique. Br J Anaesth 1993; 70: 42–6[Abstract]

22 Jelicic M, de Roode A, Bovill JG, Bonke B. Unconscious learning during anaesthesia. Anaesthesia 1992; 47: 835–7[ISI][Medline]

23 Andrade J, Jeevaratnam D, Sapsford D. Explicit and implicit memory for names presented during propofol sedation. In: Bonke B, Bovill JG, Moerman N, eds. Memory and Awareness in Anaesthesia III. Assen: Van Gorcum, 1996; 10–16

24 Andrade J. Investigations of hypesthesia: using anesthetics to explore relationships between consciousness, learning and memory. Consc Cognition 1996; 5: 562–80

25 Roediger HL, Weldon MS, Stadler ML, Riegler GL. Direct comparison of two implicit memory tests: word fragment and word stem completion. J Exp Psychol: Learning, Memory Cognition 1992; 18: 1251–69[ISI][Medline]