Comparison of ocular microtremor and bispectral index during sevoflurane anaesthesia{dagger}

L. G. Kevin*,1, A. J. Cunningham1 and C. Bolger2

1 Department of Anaesthesia, Beaumont Hospital, Dublin 9, Ireland. 2 Department of Neurosurgery, Beaumont Hospital, Dublin 9, Ireland*Corresponding author: Leo G. Kevin, Department of Anesthesiology, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI 53226, USA

{dagger}No financial support was received for this work.

Accepted for publication: May 22, 2002


    Abstract
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Background. A practical and reliable monitor of depth of anaesthesia would be a major advance on current clinical practice. None of the present monitors is both simple to use and accurate. Ocular microtremor (OMT) is a physiological tremor that is suppressed by propofol in a dose-dependent manner. We studied OMT during propofol induction and nitrous oxide– oxygen–sevoflurane maintenance of anaesthesia in 30 patients, and compared OMT with the bispectral index (BIS) as a predictor of response to verbal command.

Methods. OMT was measured using the closed-eye piezoelectric strain-gauge technique. OMT and BIS were measured at specific times during the anaesthetic, including at loss of consciousness, at end-tidal sevoflurane 1 and 2%, and at emergence.

Results. OMT decreased significantly after induction, did not decrease as end-tidal sevoflurane was increased from 1 to 2%, and increased at emergence in all patients. By logistic regression, OMT was more sensitive and specific than BIS in distinguishing the awake from the anaesthetized state (OMT, 84.9 and 93.1% respectively; BIS, 75.7 and 69.0%).

Conclusions. OMT is suppressed by sevoflurane and accurately predicts response to verbal command. OMT may be a useful monitor of depth of hypnosis.

Br J Anaesth 2002; 89; 551–5

Keywords: anaesthesia, depth; anaesthetics volatile, sevoflurane; eye, ocular microtremor; monitoring, bispectral index


    Introduction
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Current estimation of intraoperative depth of anaesthesia is an inexact science. Autonomic activity (heart rate, sweating, lacrimation) and movement to incision are commonly used as indicators of inadequate anaesthesia. However these clinical signs do not predict depth of sedation.1 Whilst the electroencephalogram (EEG) may be the most accurate monitor of deep sedation, subtle changes may be difficult to detect and may be non-specific, affected by factors such as temperature, arterial carbon dioxide (PaCO2) and electrolyte fluctuations. Additional EEG-based technologies, including spectral edge frequency and the bispectral index (BIS), have been developed but have not demonstrated the accuracy and ease of use required to gain widespread acceptance. At least part of the difficulty likely is that monitors may measure only one of several components of anaesthesia, such as level of sedation or responsiveness to noxious stimuli. The BIS has been consistently reported to be a good general measure of sedation,2 3 and whilst the manufacturer of the BIS monitor makes no claim about BIS as an overall measure of anaesthetic depth, it is used for this purpose by many practitioners. However, there are concerns that BIS recordings may be affected by specific drugs in contradictory and unpredictable ways,4 5 so that individual patients may risk awareness at BIS levels lower than any defined threshold of unconsciousness.6 These and other concerns have stimulated the continued search for alternative monitors.

    High-frequency eye tremor, or ocular microtremor (OMT), is a physiological tremor of the eye present in all subjects and is related to tonic activity in brainstem oculomotor neurones.7 First described by Alder and Fliegelman in 1934,8 this phenomenon attracted scant attention for several decades. More recently it has been shown that the frequency of OMT in patients with head injury correlates with the level of consciousness.9 Coakley and colleagues studied the effects of anaesthesia on OMT, which was suppressed by thiopentone.10 In a recent study11 propofol caused dose-dependent suppression of OMT, suggesting that OMT may indicate the cerebral effect of propofol.

We set out to study the effect of sevoflurane anaesthesia on OMT and to compare OMT with BIS as a predictor of response to verbal command.


    Methods
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
After institutional ethics committee approval and with informed, written consent, we studied 30 patients (mean age 38.0 yr, range 17–59 yr) of ASA status I and II undergoing extracranial surgery under general anaesthesia. Exclusion criteria were age <18 and >60 yr, intracranial or ocular disease and previous ocular surgery.

Baseline BIS and OMT measurements were made before induction of anaesthesia. The A-2000® BIS monitoring system (software 4.0) (Aspect Medical Systems, Natick, MA, USA) was used, and Zipprep® electrodes (Aspect Medical Systems) were applied to the skin in the frontotemporal montage as recommended by the manufacturer.12 OMT was measured using the piezoelectric strain gauge technique (EyeTect, Belleville, IL, USA), described previously.13 Briefly, the sensor consists of a probe composed of a surface-mounted amplifier and a piezoelectric transducer which is coated in silicone rubber, and can detect eye movements corresponding to displacement of the sclera ranging from 12 to more than 3000 nm. The probe is applied gently to the closed eyelid for approximately 1 min. The signal from the sensor was amplified, low-pass filtered at 150 Hz and displayed on an oscilloscope (Fluke 123 Industrial Scopemeter; Fluke Industrial, Almelo, The Netherlands). Measurements were made in real time at specific times, using an averaged 3 s segment of OMT waves. We kept a hard copy of each measurement for later examination to ensure readings were free from possible interference from signals such as facial muscle activity and electrocautery.

Patients received a standardized anaesthetic that included premedication with temazepam 20 mg 1 h before surgery. Anaesthesia was induced with i.v. fentanyl 1 µg kg–1 and propofol 2–2.5 mg kg–1. End-tidal anaesthetic agent concentration was measured routinely by anaesthetic machine monitors (Dräger Cato, Lübeck, Germany). Muscle relaxation was assessed visually by train-of-four ratio (TOFR) using a peripheral nerve stimulator (Ministim IIIA; Life-Tech, Stafford, TX, USA). OMT and BIS readings were recorded manually at the following times, taking a single figure read off each monitor: (1) before induction; (2) at loss of consciousness (first failure to respond to verbal command); (3) immediately before incision; (4) at end-tidal sevoflurane 1% after 10 min of equilibration; (5) at end-tidal sevoflurane 2% after 10 min of equilibration; and (6) at emergence (first response to verbal command). Values at time-points 1, 2, 3 and 6 were made without reference to end-tidal sevoflurane. Every 20 s patients were asked to squeeze an independent observer’s hand until failure to respond (at induction) or first unequivocal response to command (at emergence). The study plan did not specify the choice of analgesic agent but most patients received supplementary fentanyl in response to changes in heart rate and arterial pressure. OMT and BIS data were not used to guide patient management. Patients were interviewed 1 day after surgery to identify cases of explicit recall.

Data analysis
Data are expressed as medians and interquartile values. Comparison of the variables at each time was with the Wilcoxon matched-pairs signed-ranks test. The abilities of OMT and BIS to distinguish the awake state (defined by response to verbal command) from the asleep state (loss of response after induction) and the steady-state asleep state (end-tidal 2% sevoflurane) from the awake state (first response at emergence) were compared using logistic regression models; sensitivity, specificity, positive predictive values and negative predictive values of the indices for the data set were determined, and receiver operating characteristic curves were derived. Analyses were performed using Stata 3.1 (College Station, TX, USA). After Bonferroni correction, P<0.05 was considered statistically significant.


    Results
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
We studied 30 patients (18 males, mean age 38 yr, range 17–59 yr). OMT frequency decreased significantly at loss of consciousness (first failure to respond to verbal command): [85 (82–88) (awake) vs 48 (39–52) Hz, P<0.001], as did BIS [92 (90–96) vs 74 (65–80), P<0.001]. There were no significant differences for OMT between the time-points of loss of consciousness, pre-incision and end-tidal sevoflurane 1 and 2% (Fig. 1). BIS decreased significantly between loss of consciousness and pre-incision but did not differ significantly between end-tidal sevoflurane 1 and 2% (Fig. 2). OMT gave better logistic regression models than BIS to discriminate between the awake state and loss of consciousness at the beginning of anaesthesia, and to discriminate between the anaesthetized patient (end-tidal sevoflurane 2%) and the patient after recovery. Performance values for OMT and BIS models at return of consciousness are shown in Table 1 and the receiver operating characteristic curves are shown in Figures 3 and 4.



View larger version (24K):
[in this window]
[in a new window]
 
Fig 1 Ocular microtremor (OMT) at each time-point in temporal sequence (n=30). The boxes represent the median (point) and interquartile (25, 75) values. The extended bars represent the entire range. A=awake; L=first loss of response to verbal command; I=before surgical incision; 1%=end-tidal sevoflurane 1%; 2%=end-tidal sevoflurane 2%; E=first response to verbal command. *P<0.001.

 


View larger version (29K):
[in this window]
[in a new window]
 
Fig 2 Bispectral index (BIS) at each time-point in temporal sequence (n=30). The boxes represent the median (point) and interquartile (25, 75) values. The extended bars represent the entire range. A=awake; L=first loss of response to verbal command; I=before surgical incision; 1%=end-tidal sevoflurane 1%; 2%=end-tidal sevoflurane 2%; E=first response to verbal command. *P<0.001.

 

View this table:
[in this window]
[in a new window]
 
Table 1 Performance values (%) for logistic regression models of ocular microtremor (OMT) and bispectral index (BIS) to distinguish the asleep (end-tidal sevoflurane 2%) from the awake state at emergence (first response to verbal command)
 


View larger version (12K):
[in this window]
[in a new window]
 
Fig 3 Receiver operating characteristic (ROC) curve for OMT to discriminate the steady-state asleep state (at end-tidal 2% sevoflurane) from the awake state (first response at emergence). The area under the curve provides a measure of the discriminatory performance of the model. An area of 1 indicates 100% accuracy and an area of 0.5 indicates performance no better than chance; the central diagonal line indicates an area of 0.5.

 


View larger version (13K):
[in this window]
[in a new window]
 
Fig 4 Receiver operating characteristic (ROC) curve for BIS to discriminate steady-state anaesthesia (at end-tidal 2% sevoflurane) from the awake state (first response at emergence).

 
Neuromuscular block was used in 10 patients included in the study. OMT signals were not obliterated by these agents (even at TOFR=0), and although OMT amplitude decreased significantly (by 40–60%), amplitude remained sufficient for accurate measurement of OMT frequency, which was unaffected by neuromuscular block. Four patients among 20 who did not receive neuromuscular blocking agents moved at the time of skin incision. OMT frequency was significantly higher before incision in patients who moved compared with those who did not move (P<0.05), whereas BIS did not differ significantly between these patients. No patient reported recall of intra-operative events despite specific questioning on the day after surgery.


    Discussion
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
The i.v. anaesthetic agents thiopentone10 and propofol11 suppress OMT, suggesting that OMT may be useful as a marker of unconsciousness. The effects of volatile anaesthetic agents on OMT have not been studied previously. The primary objective of this study was to assess the effect of volatile anaesthetic on OMT and the secondary objective was to compare OMT with BIS, a device that is used currently to monitor depth of anaesthesia. We found that OMT is suppressed by sevoflurane. Compared with BIS, there was less overlap between conscious and unconscious values at induction of anaesthesia, and less of a hysteresis effect at emergence, allowing more accurate identification of the patient who has regained consciousness. Importantly, OMT was still measurable even during profound neuromuscular block (TOFR=0). OMT may therefore allow the identification of the conscious but paralysed patient.

The piezoelectric strain-gauge technique to measure OMT was first described in 1968.14 Measurements are relatively easy to obtain and to date records have been obtained from more than 250 normal subjects and patients. In previous reports, OMT readings were obtained by placing a piezoelectric probe on the anaesthetized sclera.911 14 In our study the probe was placed over the closed eyelid. This is comfortable for the patient and reproducible results can be obtained. No cases of scleral or other eye damage have been noted during use of the open-eye technique and we expect that the closed-eye technique will give even greater safety.

We chose i.v. induction with propofol and the opioid fentanyl to assess the monitoring techniques under conditions reflecting routine clinical practice. It is difficult to quantify the effects of propofol and fentanyl (and other administered opioids) on OMT, although the duration of surgery in most of our patients and the reported effect on OMT of stopping propofol11 suggest that sevoflurane was predominantly responsible for the continued suppression of OMT during surgery. An increase in end-tidal sevoflurane from 1 to 2% did not further decrease OMT. The lack of effect may reflect the number or variety of patients studied or other confounding factors, such as choice of analgesic agent and intensity of surgical stimulus. It may be a real effect, indicating either that the increase in end-tidal concentration was insufficient to cause a change in OMT or that sevoflurane, when administered in sufficient concentrations to achieve unconsciousness, produces an all-or-none type response in OMT.

We excluded patients less than 18 and greater than 60 yr old. In adults, normal OMT frequency is known to be 84 (SD 6) Hz,15 and in subjects more than 70 yr of age the mean peak count frequency is approximately 7 Hz lower than in younger subjects.15 There are no published OMT data in children. Thresholds to define unconsciousness may require adjustment for age which cannot be defined by the present data. Parkinson’s disease,16 multiple sclerosis17 and head injury9 alter OMT. As suggested previously,11 a more useful threshold to define unconsciousness may be a percentage reduction from the patient’s baseline OMT.

We chose to compare OMT with the BIS, an EEG-derived measure of the hypnotic component of anaesthesia. In bispectral analysis, EEG waves are initially subjected to a Fourier transform and the phase correlation of the resulting sine waves is then analysed.18 Other EEG features are then combined using an algorithm that gives a single measurement, the BIS. The BIS is on a scale of dimensionless units in which 100 represents the awake state, 70 deep sedation and 60 general anaesthesia, continuing to 0, which indicates isoelectric EEG. In several studies (as in our study), BIS predicted the ability of patients undergoing anaesthesia to respond to command.19 20 Furthermore, BIS correlates with end-tidal sevoflurane concentration20 and measured blood propofol concentration.21 However, the BIS value required to ensure unresponsiveness and suppression of memory formation varies greatly between studies,22 and a recent report highlights the possibility of explicit awareness at BIS values generally taken to indicate unconsciousness reliably.6 A further difficulty in the determination of specific BIS thresholds is the marked hysteresis effect.2 23 In other words, BIS values associated with unconsciousness at induction of anaesthesia commonly occur with consciousness during emergence. This effect leads to risk of awareness if anaesthetic drug doses are titrated using a specific value. We noted this effect, which also occurred to some extent with OMT, as demonstrated by logistic regression analysis. As reported elsewhere,2 BIS tended to rise gradually as anaesthesia was lightened at the end of surgery. In contrast, OMT tended to remain depressed until just before the first response to verbal command. These results suggest that BIS can provide a graded measure of impending emergence. However, OMT may exclude the aware patient more accurately. In the only other controlled study of anaesthetic effects on OMT,11 OMT readings at emergence are not reported.

We did not observe a statistically significant effect on BIS of increasing end-tidal sevoflurane from 1 to 2%, although there was a trend towards decreased BIS. We may not have studied a sufficient number of patients to find this effect as it is likely to be small. Katoh and co-workers report that BIS does not show further decreases when end-tidal sevoflurane concentration exceeds 1.4%.20

In summary, both OMT and BIS decreased at induction and remained depressed during sevoflurane anaesthesia. When using response to verbal command to indicate consciousness, OMT showed less overlap between the conscious and unconscious states and more accurately identified patients regaining consciousness after anaesthesia. OMT is depressed by i.v. and inhalational anaesthetic agents and recordings can be obtained from a probe placed over the closed eyelid. Further studies are needed of this technique at extremes of age, in patients with pre-existing central nervous system pathology and during different surgical procedures.


    Acknowledgement
 
Eye probes were provided free of charge by EyeTect, LLC (Belleville, IL, USA).


    References
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
1 Domino KB, Posner KL, Caplan RA, Cheney FW. Awareness during anesthesia: a closed claims analysis. Anesthesiology 1999; 90: 1053–61[ISI][Medline]

2 Gajraj RJ, Doi M, Mantzaridis H, Kenny GN. Comparison of bispectral EEG analysis and auditory evoked potentials for monitoring depth of anaesthesia during propofol anaesthesia. Br J Anaesth 1999; 82: 672–8[Abstract/Free Full Text]

3 Iselin-Chaves IA, Flaishon R, Sebel PS, et al. The effect of the interaction of propofol and alfentanyl on recall, loss of consciousness, and the bispectral index. Anesth Analg 1998; 87: 949–55[Abstract]

4 Vernon JM, Lang E, Sebel PS, Manberg P. Prediction of movement using bispectral electroencephalographic analysis during propofol/alfentanyl or isoflurane/alfentanyl anesthesia. Anesth Analg 1995; 80: 780–5[Abstract]

5 Detsch O, Schneider G, Kochs E, Hapfelmeier G, Werner C. Increasing isoflurane concentration may cause paradoxical increases in the EEG bispectral index in surgical patients. Br J Anaesth 2000; 84: 33–7[Abstract]

6 Mychaskiw G, Horowitz M, Sachdev V, Heath BJ. Explicit intraoperative recall at a bispectral index of 47. Anesth Analg 2001; 92: 808–9[Free Full Text]

7 Coakley, D. Minute Eye Movement and Brain Stem Function. Florida: CRC Press, 1983; 49–56.

8 Alder FH, Fliegelman F. The influence of fixation on the visual acuity. Arch Ophthalmol 1934; 12: 475–83

9 Coakley D, Thomas JG. The ocular microtremor record and the prognosis of the unconscious patient. Lancet 1977; 1: 512–5[Medline]

10 Coakley D, Thomas JG, Lunn JN. The effect of anaesthesia on ocular microtremor. Br J Anaesth 1976; 48: 1121–2

11 Bojanic S, Simpson T, Bolger C. Ocular microtremor: a tool for measuring depth of anaesthesia? Br J Anaesth 2001; 86: 519–22[Abstract/Free Full Text]

12 Aspect Medical Systems. BIS Clinical Reference Manual. Natick, USA: Aspect Medical Systems, 1997

13 Sheahan NF, Coakley D, Hegarty F, Bolger C, Malone J. Ocular microtremor measurement system: design and performance. Med Biol Eng Comput 1993; 31: 205–12[ISI][Medline]

14 Benji H, Thomas JG. Three electronic methods for recording ocular tremor. Med Biol Eng 1968; 6: 171–8[ISI][Medline]

15 Bolger C, Bojanic S, Sheahan NF, Coakley D, Malone JF. Dominant frequency content of ocular microtremor from normal subjects. Vision Res 1999; 39: 1911–5[ISI][Medline]

16 Bolger C, Bojanic S, Sheahan NF, Coakley D, Malone JF. Ocular microtremor in patients with idiopathic Parkinson’s disease. J Neurol Neurosurg Psychiatry 1999; 66: 528–31[Abstract/Free Full Text]

17 Bolger C, Bojanic S, Sheahan NF, Malone J, Hutchinson M, Coakley D. Ocular microtremor (OMT): a new neurophysio logical approach to multiple sclerosis. J Neurol Neurosurg Psychiatry 2000; 68: 639–42[Abstract/Free Full Text]

18 Sigl J, Camoun NG. An introduction to bispectral analysis for the electroencephalogram. J Clin Monit 1994; 10: 392–404[ISI][Medline]

19 Liu J, Singh H, White PF. Electroencephalographic bispectral index correlates with intraoperative recall and depth of propofol-induced sedation. Anesth Analg 1997; 84: 185–9[Abstract]

20 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[ISI][Medline]

21 Glass PS, Bloom M, Kearse L, Rosow C, Sebel P, Manberg P. Bispectral analysis measures sedation and memory effects of propofol, midazolam, isoflurane, and alfentanyl in healthy volunteers. Anesthesiology 1997; 86: 836–47[ISI][Medline]

22 Drummond JC. Monitoring depth of anaesthesia. Anesthesiology 2000; 93: 876–82[ISI][Medline]

23 Sleigh JW, Donovan J. Comparison of bispectral index, 95% spectral edge frequency and approximate entropy of the EEG, with changes in heart rate variability during induction of general anaesthesia. Br J Anaesth 1999; 82: 666–71[Abstract/Free Full Text]