Providing adequate sedation and analgesia during gastrointestinal (GI) endoscopy is increasingly a standard practice in many institutions [1–3]. However, the depth of sedation is a continuum, and progression from moderate to deep sedation and general anaesthesia can be fast and subtle. Therefore, precise monitoring of the level of consciousness can avoid over-sedation. There is no objective measurement of sedation that is universally accepted. Current methods of determining whether the patient has reached an optimal sedative level rely upon subjective clinical assessments, which may be unreliable and interrupt the procedure.
The application of a brain monitor in sedation for GI endoscopy was first conducted with a bispectral index (BIS) . We decided to use another method, the A-Line® auditory evoked potential index (AAI) [5,6], which may be more sensitive than BIS [7–10]. The AAI is mapped into an index value ranging from 100 (fully awake) to 0 (no brain activity) and suggests that a target range of 15–25 is adequate for ensuring an appropriate anaesthetic level. However, data are lacking with regard to numeric AAI range during sedation for GI endoscopy. In addition, AAI version 1.6 is a newly developed composite index . Thus, we designed this prospective, observational study to investigate the use of AAI version 1.6 in patients receiving sedation induced by midazolam and alfentanil in GI endoscopy and to further define the AAI range corresponding with optimal sedation.
Materials and methods
Following approval of the institutional Ethics Committee and patient informed consent, 30 adults, of either gender, below 65 yr of age, American Society of Anesthesiologists (ASA) Grade I or II, scheduled consecutively for combined oesophagogastroduodenoscopy (OGD) and colonoscopy under sedation were studied. Exclusion criteria included neurological disorders, hearing impairment and history of habitual sedative medication or alcoholism.
OGD and colonoscopy were, respectively, performed sequentially by a gastroenterologist and a colorectal surgeon. A standardized total intravenous sedation technique was performed in all patients by the same anaesthesiologist (Table 1). Spontaneous ventilation was maintained with oxygen supplementation at 3 L min−1 via a nasal prong. Head-tilt, chin-lift, jaw thrust manoeuvres or manual ventilation via a facemask were performed if the SPO2 was <90%. The sedation score was evaluated with the Observer’s Assessment of Alertness/Sedation (OAA/S) scale  by one operating anaesthesiologist. After the loading dose, the procedure commenced when the patient lost response to normal command (OAA/S < 4). The OAA/S scale was then checked every 2 min and incremental boluses were titrated to meet the target end-points (OAA/S < 4).
According to the ASA guideline for sedation and analgesia , minimal sedation defined as anxiolysis and normal response to verbal stimulation, corresponds with OAA/S score 4. Moderate sedation/analgesia, synonymous with ‘conscious sedation’ defined as the presence of purposeful response to verbal or light tactile stimulation, is approximately equivalent to OAA/S score 3. Deep sedation, defined as purposeful response to repeated or painful stimulation, corresponds with OAA/S score 1. By the same token, OAA/S score 2 was a state between moderate to deep sedation.
AAI (version 1.6) was monitored with an AEP Monitor/2 (Danmeter A/S, Odense, Denmark) . The skin of the forehead, where the electrodes were attached, was rubbed with fine abrasive paper and alcohol. Three silver/silver chloride electrodes (Danmeter, Odense, Denmark) were positioned at the mid-forehead (+), right forehead (reference) and right mastoid (−). Electrode placement and adhesion to skin were adjusted until the electrode impedance was less than 1000 Ω. Auditory stimuli in the form of intermittent clicks were applied through bilateral earphones (9 Hz; 2-ms duration, automatic 45–75 dB sound pressure level). The mid-latency auditory evoked potential (MLAEP) analysis window was 20–80 ms, and latency and amplitude changes in the AEP were equally weighted. The data were transferred to a personal computer for subsequent analysis with an AAI GraphTM software package (Danmeter A/S, version 1.6, Odense, Denmark).
The AEP Monitor/2 was operated by an independent research fellow and the screen was out of sight of the endoscopists and the operating anaesthesiologist. AAI, OAA/S scores, as well as non-invasive mean arterial pressure, heart rate (HR) and SPO2 were recorded by the research fellow before the loading dose was given and afterwards every 2 min until the procedure was finished. To minimize the influence on AAI when the sedation was under scoring (OAA/S) with verbal or physical stimulation, all AAIs were recorded beforehand by the research fellow.
After the procedure, patients were sent to the ward for continuous follow-up of vital signs, ability to walk, presence of nausea/vomiting, pain and bleeding every 30 min for 2 h. After the patients could walk steadily without dizziness, they were requested to answer a modified questionnaire including amnesia (amnesia score: 1. remember everything about the endoscopy, 2. forget a few things about it, 3. forget most of the event, 4. do not remember anything)  and attitude towards present and future acceptance of endoscopic examination (Yes or No).
Data are presented as mean ± standard deviation (SD) or median (range). Spearman’s rank correlation was performed to evaluate the relationship between OAA/S scores with AAI, HR, mean arterial pressure and SPO2. In addition, the relationship between AAI and electromyographic (EMG) value was also analysed. After trend test for AAI and OAA/S by linear regression, U-tests were carried out to estimate significant differences between subsequent OAA/S levels. Receiver operator characteristic (ROC) curves were created to estimate the best ‘cut-off values’ of different levels of sedation. The relative risk of oxygen saturation lower than 95% was calculated for transition from OAA/S = 3 to 2. P < 0.05 was considered statistically significant. All statistical analyses were conducted with the statistical package for social sciences 13.0 (SPSS Inc., Chicago, IL, USA).
Around 203 recordings were obtained from 30 patients (14 males and 16 females) aged 30–61 yr, body weight 38.4–76.6 kg. Patient characteristics data showed that the mean (SD) dose requirement of midazolam and alfentanil were 3.0 mg (0.8) and 803.6 μg (272.8), respectively. Median (range) duration of procedure was 8.5 (5–16) min.
The mean (SD) AAIs in various OAA/S scores in numerical order were 81.2 (15.0) at score 5, 63.2 (11.6) at score 4, 48.8 (8.8) at score 3, 36.5 (3.7) at score 2 and 29.0 (3.0) at score 1. There was a significantly positive correlation between AAI and OAA/S scores (ρ = 0.886, P < 0.001). Linear regression revealed AAI = 14.01 × (OAA/S score) + 8.23 (P < 0.001) and the R2 was 0.695. Statistically, the AAIs differed significantly at various OAA/S scores. A box plot indicating the distribution of various AAIs for related OAA/S scores is shown in Figure 1. However, no correlations were shown between OAA/S with BP and HR. In addition, the correlation between AAI with EMG was also noted (ρ = 0.413, P < 0.001).
To determine the optimal AAI thresholds of different sedative levels, ROC curves were made between subsequent OAA/S scores (Fig. 2a–c). The optimal AAI cut-off values for minimal, moderate and moderate to deep sedation were considered to be 54, 42 and 34, respectively, given the sensitivity value 90%, 93% and 94% at the least compromise of its specificity 88%, 92% and 100%.
A scatter plot showing the distribution of SPO2 for each OAA/S score is presented in Figure 3. The correlation between the SPO2 and OAA/S score (ρ = 0.704, P < 0.001) was significant. More episodes of SPO2 below 95% were noted when OAA/S < 3. The relative risk of SPO2 < 95% for OAA/S = 2 compared with 3 was 15.98 (95% CI: 3.94–64.81).
The postanaesthetic period was uneventful in all patients. No complaints of nausea/vomiting or pain were noted. All patients could walk steadily without evidence of dizziness (near full recovery) within 2 h after completing the procedure. Seven patients (23%) had partial recall and the other 23 (77%) had complete amnesia. All 30 patients expressed their satisfaction at the present sedation and willingness for future GI endoscopy.
In the results of Bower and colleagues, the correlation between BIS and OAA/S scores was moderate (ρ = 0.59) . We found that AAI was closely correlated with sedative level assessed by OAA/S score (ρ = 0.886) under sedation with midazolam and alfentanil for combined OGD and colonoscopy. BIS is an objective, electroencephalogram (EEG)-based monitor for depth of sedation. Several studies have compared AAI with BIS and revealed that AAI was more sensitive in monitoring depth of anaesthesia and in predicting recovery to consciousness [7,8,14]. The Spearman’s rank correlation coefficient and linear regression with OAA/S levels for AAI were all higher than those for BIS during propofol sedation . Some authors also doubted the usefulness of BIS in titrating boluses of propofol to attain an adequate level of sedation during colonoscopy, because there was a substantial time-lag between BIS and OAA/S score . Furthermore, Bower and colleagues have demonstrated that BIS levels varied increasingly with deepening levels of sedation . As mean BIS levels decreased, the data revealed increasing SDs of BIS. That is to say, inter-individual variations increased as sedation deepened. This may indicate that it is more difficult to accurately assess sedation by BIS as patients attain deeper levels. Although we did not monitor BIS, we found that the variation of AAI decreased (demonstrated by the gradually reducing standard deviation) when sedation deepened. This finding is very similar to the results of Litvan and colleagues, who suggested that the reliability of AAI increases as sedative level deepens .
Ge and colleagues have evaluated the performance of AAI as a monitor of sedation induced by propofol or midazolam in patients receiving low thoracic epidural anaesthesia for gynaecologic surgery . The Spearman’s rank correlation coefficient (ρ = 0.958 vs. 0.977, propofol vs. midazolam, respectively) and linear regression (slope = 17.6 vs. 17.91, propofol vs. midazolam, respectively) were both higher than those of our study (ρ = 0.886 and slope = 14.01, respectively). These differences may be explained by the analgesic effect provided by the epidural anaesthesia. AEP index was considered a measure of the overall balance between noxious stimulation, analgesia and hypnosis . In our study, plasma levels of the anaesthetics might not be constant. Instead, bolus injections of midazolam and alfentanil caused considerable fluctuations of drug concentration and the uncomfortable sensation and stress from the endoscopic examination may not be sufficiently suppressed. Consequently, noxious stimulation may transiently raise AAI as a result of increased electrocortical activity, but this cortical recognition may not be clinically apparent, as assessed by the OAA/S. In Ge’s study, epidural anaesthesia could totally overcome the surgical painful stimuli; therefore, AAI only showed the component of cortical hypnosis.
We used AAI version 1.6 which incorporates MLAEPs, EEG and burst suppression, but found no burst suppression in our study. This suggests that the sedative level in our procedure was not deep enough to cause burst suppression. In case of low AEP signal quality, AAI is calculated from the spontaneous EEG activity. This prevents interpreting low AAI values falsely as a result of disconnection of the headphones . Besides, EMG activity can disturb the MLAEP waveform at this level of sedation and cause excessive artefacts and noise contamination, leading to rejection from further analysis . Weber and colleagues reported a highly significant positive correlation between AAI1.4 and EMG value (ρ = 0.72) , which suggested possible interference from EMG activity. However, the results of our study with the use of AEP-Monitor/2 showed a relatively low correlation coefficient between AAI1.6 and corresponding EMG values (ρ = 0.413, P < 0.001), suggesting less interference from EMG activity.
We found that cardiovascular function could be usually maintained during sedation . The major concern during sedative procedures is the risk of airway compromise and deoxygenation due to apnoea. The pulse oximeter is an invaluable monitor in outpatient settings. However, desaturation following sedation is a late event and hypoventilation may occur earlier, especially in the presence of oxygen supplementation. In case of significant oxygen desaturation (SPO2 < 90%), resuscitation may be difficult, particularly when the endoscope has been inserted orally. Therefore, we felt it reasonable to increase the alarm threshold of SPO2 up to 95% during sedation under an FiO2 of about 30% in our cases. We measured SPO2 at different sedative levels and found that there was a moderate to high correlation between SPO2 and OAA/S score (ρ = 0.704, P < 0.001). The relative risk of SPO2 < 95% for OAA/S = 2 compared with 3 was 15.98. Consequently, we recommended that OAA/S score 3, corresponding to moderate sedation, is the optimal level for combined OGD and colonoscopy procedure with minimal risk of ventilatory compromise. According to the manufacturer of the AEP Monitor/2, AAI values of 60 or more are indicative of the ‘awake’ state, and the range of 15–25 is suggestive of surgical anaesthesia. By reckoning the validity of distinct cut-off values, we estimated that AAIs of about 42–54 could offer a sedative level that corresponded to OAA/S score 3 with high sensitivity and specificity.
It has been emphasized that the clinical end-points for sedation and analgesia during endoscopy need to be reappraised and we should aim to achieve anxiolysis and amnesia rather than ptosis and hypnosis . For this reason, the dosage of sedation regimen in this study was designed based on previous empirical practices in conscious sedation for endoscopy . The dose of midazolam as recommended by Froehlich and colleagues was 0.035 mg kg−1 with or without opioids. We modified the regimen by adding alfentanil (8 μg kg−1) for prevention of the gag reflex in ODG and painful stimulation in colonoscopy [20–22]. Incremental bolus rather than target-controlled infusion was chosen because the procedures were relatively short.
The limitation of our study is that the obtainment of AAI data at OAA/S score 1 (n = 8) was insufficient and the cut-off values of moderate to deep sedation may be questionable due to limited sample size. However, it is unnecessary to achieve deep sedation for pure endoscopic examination.
In conclusion, the AAIs correlated well with clinical changes of the hypnotic effect of the combined use of midazolam and alfentanil. AAI1.6 can provide the endoscopist with an effective, safe and objective real-time index for depth of sedation without disrupting the procedure. The measured EMG interference was less with this AEP Monitor/2. Precise monitoring of sedative level may hopefully prevent inadequate depth or overshooting of sedation.
The authors wish to thank Professor K. C. Wong and Professor Jhi-Joung Wang for revising the manuscript. We also thank Ms. Pui-Ching Lee, consultant of medical statistics, for her statistical assistance.
1. American Society of Anesthesiologists Task Force. Practice guidelines for sedation and analgesia by non-anesthesiologists. A report by the American Society of Anesthesiologists Task Force on sedation and analgesia by non-anesthesiologists. Anesthesiology
2002; 96: 1004–1017.
2. Keeffe EB. Sedation and analgesia for endoscopy. Gastroenterology
1995; 106: 932–934.
3. Waring JP, Baron TH, Hirota WK et al.
Guidelines for conscious sedation and monitoring during gastrointestinal endoscopy
. Gastrointest Endosc
2003; 58: 317–322.
4. Bower AL, Ripepi A, Dilger J, Boparai N, Brody FJ, Ponsky JL. Bispectral index monitoring of sedation during endoscopy. Gastointest Endosc
2000; 52: 192–196.
5. Kurita T, Doi M, Katoh T et al.
Auditory evoked potential index predicts the depth of sedation and movement in response to skin incision during sevoflurane anaesthesia. Anesthesiology
2001; 95: 364–370.
6. Struys MM, Jensen EW, Smith W et al.
Performance of the ARX-derived auditory evoked potential index as an indicator of anaesthetic depth: a comparison with bispectral index and hemodynamic measures during propofol administration. Anesthesiology
2002; 96: 803–816.
7. Nishiyama T, Hanaoka K. The A-line ARX index may be a more sensitive detector of arousal than the bispectral index during propofol–fentanyl–nitrous oxide anaesthesia: a preliminary investigation. Can J Anaesth
2004; 51: 539–544.
8. Nishiyama T, Matsukawa T, Hanaoka K. Is the ARX index a more sensitive indicator of anaesthetic depth than the bispectral index during sevoflurane/nitrous oxide anaesthesia? Acta Anaesthesiol Scand
2004; 48: 1028–1032.
9. Ge SJ, Zhuang XL, Wang YT, Wang ZD, Li HT. Changes in the rapidly extracted auditory evoked potentials
index and the bispectral index during sedation induced by propofol or midazolam
under epidural block. Br J Anaesth
2002; 89: 260–264.
10. Nishiyama T, Matsukawa T, Hanaoka K. A comparison of the clinical usefulness of three different electroencephalogram monitors: bispectral index, processed electroencephalogram, and Alaris auditory evoked potentials
. Anesth Analg
2004; 98: 1341–1345.
11. Vereecke HE, Vasquez PM, Jensen EW et al.
New composite index based on midlatency auditory evoked potential and electroencephalographic parameters to optimize correlation with propofol effect site concentration: comparison with bispectral index and solitary used fast extracting auditory evoked potential index. Anesthesiology
2005; 103: 500–507.
12. 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: 244–251.
13. Brouillette DE, Leventhal R, Kumar S et al. Midazolam
versus diazepam for combined esophogastroduodenoscopy and colonoscopy. Dig Dis Sci
1989; 34: 1265–1271.
14. Gajraj RJ, Doi M, Mantzaridis H, Kenny GNC. Comparison of bispectral EEG analysis and auditory evoked potentials
for monitoring depth of anaesthesia during propofol anaesthesia. Br J Anaesth
1999; 82: 672–678.
15. Chen SC, Rex DK. An initial investigation of bispectral monitoring as an adjunct to nurse-administered propofol sedation for colonoscopy. Am J Gastroenterol
2004; 99: 1081–1086.
16. 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: 351–358.
17. Weber F, Zimmermann M, Bein T. The impact of acoustic stimulation on the AEP monitor/2 derived composite auditory evoked potential index under awake and anesthetized conditions. Anesth Analg
2005; 101: 435–439.
18. Weber F, Bein T, Hobbhahn J, Taeger K. Evaluation of the Alaris auditory evoked potential index as indicator of anaesthetic depth in preschool children during induction of anaesthesia with sevoflurane and remifentanil. Anesthesiology
2004; 101: 294–298.
19. Froehlich F, Schwizer W, Thorens J et al.
Conscious sedation for gastroscopy: patient tolerance and cardiorespiratory parameters. Gastroenterology
1995; 108: 697–704.
20. Holloway AM, Logan DA. Pain relief for outpatient colonoscopy: a comparison of alfentanil
with fentanyl. Anaesth Intens Care
1990; 18: 210–213.
21. Chokhavatia S, Nguyen L, Williams R, Kao J, Heavner JE. Sedation and analgesia for gastrointestinal endoscopy
. Am J Gastroenterol
1993; 88: 393–396.
22. Diab FH, King PD, Barthel JS, Marshall JB. Efficacy and safety of combined meperidine and midazolam
for EDG sedation compared with midazolam
alone. Am J Gastroenterol
1996; 91: 1120–1125.