Figures
Abstract
We conducted a large retrospective study to investigate the confounding factors that predict Ce ROC under propofol-based TIVA with TCI. We recorded sex, age, height, weight, Ce LOC, Ce ROC, total propofol and fentanyl consumption dose, and anesthetic time. Simple linear regression models were used to identify potential predictors of Ce ROC, and multiple linear regression models were used to identify the confounding predictors of Ce ROC. We found that Ce ROC correlated with age, sex, Ce LOC, and both total fentanyl and propofol consumption dose. The prediction formula was: Ce ROC = 0.87 - 0.06 × age + 0.18 × Ce LOC + 0.04 (if fentanyl consumption > 150 μg; if not, ignore this value) + 0.07 × (1 or 2, according to the total propofol consumption dose, 1 for a propofol amount 1000-2000 mg and 2 for a propofol amount > 2000 mg). We simplified the formula further as Ce ROC = 0.87 - 0.06 × age + 0.18 × Ce LOC. In conclusion, Ce ROC can be predicted under TCI with propofol- and fentanyl-based TIVA. The confounding factors that predicted propofol Ce ROC are age, sex, Ce LOC, and total consumption dose of propofol and fentanyl.
Citation: Chan S-M, Lee M-S, Lu C-H, Cherng C-H, Huang Y-S, Yeh C-C, et al. (2015) Confounding Factors to Predict the Awakening Effect-Site Concentration of Propofol in Target-Controlled Infusion Based on Propofol and Fentanyl Anesthesia. PLoS ONE 10(5): e0124343. https://doi.org/10.1371/journal.pone.0124343
Academic Editor: Francesco Staffieri, University of Bari, ITALY
Received: November 26, 2014; Accepted: February 28, 2015; Published: May 4, 2015
Copyright: © 2015 Chan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: All relevant data are within the paper.
Funding: The authors have no support or funding to report.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The target-controlled infusion (TCI) machine provides a function to estimate the effect-site concentration (Ce) and the elimination time of propofol. Iwakiri and Nishihara (2005) suggested that patients can be expected to awaken quickly upon completion of the procedure and discontinuation of drug administration if we know the propofol effect-site concentration at loss of consciousness (CeLOC) [1]. Nunes and Ferreira (2005) reported that the propofol effect-site concentration at return of consciousness (CeROC) was related to CeLOC and the patients age, and they concluded that CeROC can be estimated by combining information about CeLOC and the patients age [2]. Shafer and Doze (1998) found that CeROC was related to the patients age, sex, weight, and type of surgery [3]. By knowing the associated factors that predict CeROC, the anesthesiologist should be able to estimate the emergence time and to provide a fast emergence to shorten the anesthesia-controlled time. However, the three reports mentioned above included small sample sizes. The purpose of this large retrospective study was to estimate the confounding factors that predicted propofol CeROC under TCI.
Materials and Methods
This retrospective study retrieved information from the electronic database and anesthetic records of the Tri-Service General Hospital (TSGH; Taipei, Taiwan, Republic of China). The Ethics Committee of TSGH approved the study (TSGHIRB No: 100-05-168). IRB allow waiving the requirement for obtaining informed consent and patient records was anonymized and de-identified prior to analysis. Data collected by six different anesthesia providers with average twelve years of experience. At least two anesthesiologists independently reviewed the case information in the database and the medical records for that case. The study included 794 patients (464 women and 330 men) classified as American Society of Anesthesiologists (ASA) physical status I to III, who were aged 18–88 years and were scheduled for one of the following types of surgery: elective surgery of the extremities; spine surgery; exploratory laparotomy; laparoscopy; ear, nose, or throat surgery; or ophthalmological, genitourinary, chest, gynecological or breast cancer surgery. The exclusion criteria were known neurological disorders, pregnancy, medication affecting the central nervous system, uncontrolled hypertension, recent use of psychotropic drugs, chronic alcohol consumption, severe obesity (body mass index > 35 kg m2−1), combined inhalation anesthesia with propofol, or incomplete data.
Anesthetic techniques used in our routine practice
No medication was administered before induction of anesthesia, as in our clinical practice. Regular monitoring included electrocardiography (lead II), pulse oximetry, and noninvasive measurement of blood pressure, respiratory rate, and end-tidal carbon dioxide pressure. Three electrodes [A-Line Auditory Evoked Potential electrodes; Danmeter, Odense, Denmark] were positioned at the mid-forehead (+), left forehead (reference), and left mastoid (–). Anesthesia was induced using a total intravenous anesthesia (TIVA) technique with intravenous (i.v.) fentanyl (2–3 μg kg−1) and 2% lidocaine (1.5 mg kg−1). Continuous infusion of propofol (Fresfol 1%) was delivered using a TCI system (Base Primea, Fresenius Kabi AG, Bad Homburg, Germany) with Schnider’s kinetic model of a Ce of 3.0–5.0 μg ml−1. Propofol was adjusted to keep the Auditory Evoked Potential Index (AAI) between 15 and 25 during maintenance of anesthesia.
Rocuronium (10 mg, i.v.) was given as required by the return of neuromuscular function. We maintained the hemodynamic parameters in the 20% range of preoperative values. Hypertension and tachycardia were treated with 1 μg kg−1 of i.v. fentanyl if the AAI was in the set range. After two unsuccessful treatments, 5 mg of i.v. labetalol was given. Hypotension was treated with fluid, and 5 mg of i.v. ephedrine was given if the AAI was in the set range. Atropine (0.5 mg) was given if the HR was < 50 bpm and accompanied by hypotension [4]. TCI of propofol was turned off at the last surgical suture. At the end of the surgical procedure, 2 mg neostigmine and 1 mg atropine were given intravenously. The conduct of anesthesia, including fluid management, was determined by the attending anesthesiologist. When the patient regained consciousness with smooth respiration, the tracheal tube was removed and the patient was sent to the post-anesthesia recovery room for further care.
In our anesthesia chart and computer system, we had recorded age, sex, height, weight, CeLOC, CeROC, anesthesia time, surgical time, and both propofol and fentanyl consumption dose. Anesthesia time was defined as the time from anesthesia induction to extubation of the tracheal tube, and surgical time was defined as the time from the skin incision to covering with the dressing. Emergence time was defined as the time from the end of surgery to extubation. The total dose of propofol and fentanyl were defined as the total dose of propofol and fentanyl administered from the induction of general anesthesia to cessation of drug administration. CeLOC was defined as the propofol effect-site concentration at the time of loss of consciousness without eyelash reflex, and CeROC was defined as the propofol effect-site concentration at the time when the patient opened his/her eyes in response to his/her name being called loudly at 30-s intervals, which corresponds with Observer’s Assessment of Alertness/Sedation Score 4–5.
Statistical analysis
The patients’ characteristics are presented as minimum, maximum, mean, and standard deviation. Before multivariable analysis, data are needed thorough univariate analyses between independent variables and dependent variable, such as correlation analysis.Simple linear regression models were used to identify potential predictors of CeROC. Multiple linear regression models were used to assess the independent predictors for CeROC. Statistical analyses were performed using SPSS Version 16.0 software (SPSS Inc, Chicago, III). Statistical significance was defined as p< 0.05.
Results
The patients’ characteristics are shown in Table 1. According to the records, anesthesia induction and maintenance were smooth in all cases. No patient reported memory of the operation either spontaneously or when questioned about it on the day after the operation.
The regression coefficient (β) and coefficient of determination (R2) against CeROC are shown in Table 2. In the univariate analyses, positive correlations were observed between CeROC and CeLOC (R2 = 0.286, p < 0.01, Fig 1), between CeROC and fentanyl consumption (R2 = 0.04, p < 0.01, Fig 2), and between CeROC and propofol consumption (R2 = 0.029, p < 0.01, Fig 3). A negative correlation was observed between age and CeROC (R2 = 0.288, p < 0.01, Fig 4). The formula from a multiple linear regression analysis was obtained as:
Statistically significant correlation (R2 = 0.286; p < 0.01) and positive slope.
Statistically significant correlation (R2 = 0.04; p < 0.01) and positive slope.
Statistically significant correlation (R2 = 0.029; p < 0.01) and positive slope.
Statistically significant correlation (R2 = 0.288; p < 0.01) and negative slope.
This equation could be simplified as the following prediction formula (round off to the 2nd decimal place):
Both CeLOC (3.01 ± 0.48 vs. 2.88 ± 0.48 μg ml−1 respectively, p < 0.001) and CeROC (1.21 ± 0.29 vs. 1.13 ± 0.23 μg ml−1 respectively, p < 0.001) were higher in men than in women. Consumption of both propofol (0.111 ± 0.024 vs. 0.127 ± 0.024 mg kg−1 min−1 respectively, p < 0.001) and fentanyl (0.021 ± 0.010 vs. 0.025 ± 0.012 μg kg−1 respectively, p< 0.001) was lower in men than in women.
Discussion
The major finding of our study is that CeROC correlated with age, sex, CeLOC, and consumption dose of both total fentanyl and propofol. These results are consistent with those of previous studies [1, 3, 5]. The two key factors that predicted CeROC were age and CeLOC. We simplified the formula further as CeROC = 0.87 − 0.06 × age × 0.18 × CeLOC. (Because variables in fentanyl and propofol are very small to affect outcome).This result is consistent with the study by Nunes, which reported the equation CeROC = 0.8215 + 3.0589 × CeLOC/age [5]. If we apply our CeLOC and age data to the formula of Nunes, we obtain a CeROC of 1.01 and about 10% variation to produce a CeROC of 1.17 with our formula. Thus, these two formulas may be able to predict CeROC. This study provides a new strategy to address problem about predicting the awakening effect-site concentration.
We found that the patients’ age correlated inversely with CeROC in patients undergoing TIVA under TCI with propofol. That is, the older the patient, the lower the propofol CeROC. This finding corroborates the results of Nunes, who found a significant correlation between CeROC and patients’ age when using a different induction technique in 31 patients [5]. They reported that older patients had a lower CeROC with the same CeLOC. This may be explained by the fact that the weight of the human brain decreases by about 10% with age; the gray matter decreases more than the white matter [6]. Kreuer and Schreiber (2005) also found that the average normalised propofol consumption decreases with increasing patients age [7]. Schnider and Minto (1998) reported an increased sensitivity to propofol’s effect in elderly patients [8]. These data suggest that propofol dose should be reduced in elderly patients for pharmacokinetic and pharmacodynamic reasons [8, 9]. Schnider created a model of the age-related pharmacodynamic relationship between plasma propofol concentration and loss of consciousness, which we used in the present study.
We found that CeLOC correlated positively with CeROC in patients undergoing TIVA under TCI with fentanyl. This finding is also consistent with the results of previous studies [1, 5]. Iwakiri concluded that propofol CeROC can be estimated using CeLOC with the Marsh model and that CeLOC can be used as a guide to anaesthetic management to maintain a constant concentration [1]. They also theorized that such a clinical management strategy could be better than simply targeting the population-based average effect-site concentration. Nunes also reported that Ce ROC is positively related to CeLOC [5].
We also found that sex was significantly related to Ce ROC in patients undergoing TIVA under TCI with fentanyl. These results are consistent with previous studies showing that females emerge faster from TIVA than males [3, 10, 11]. The reason may reflect a possible hormonal influence on the effect of hypnotic drugs. In our study, both propofol and fentanyl consumption was higher in women than in men, possibly because females have a larger proportion of body fat and smaller water content compared with males [12]. This would affect the volume of distribution and therefore the initial concentration of many drugs used in anesthesia [12]. For lipid-soluble drugs, such as opioids and benzodiazepines, the volume of distribution is generally larger in females [12]. Conversely, for water-soluble drugs, such as neuromuscular blocking agents, the volume of distribution is generally smaller in females. The pharmacokinetic analysis produced a larger volume of redistribution and higher clearance in women, which may explain why females consume more propofol than males [13]. However, sex was not linearly related in the multiple linear regression analysis; possibly because of the inclusion of sex as nominal data.
We found that fentanyl and propofol consumption correlated positively with Ce ROC in patients undergoing TIVA under TCI. Analgesic and hypnotic drugs may interact with each other to achieve an adequate depth of anesthesia during surgery [12]. The analgesic concentration of opioid can affect the predictive performance of TCI [14–16]. We found that patients with more propofol consumption presented higher Ce ROC and the result might conflict with our knowledge, as we know the context-sensitive half-time increases after longer infusion times [17]. However, patients sometimes underwent prolonged anesthesia. Although the propofol consumption increased, the TCI machine provided a fixed Ce and then decreased the infusion rate with time. Moreover, the maintained propofol was guided by AAI, and we observed that the maintained Ce decreased with time. In addition, we used fentanyl mainly in the first 2 h of the operation, after which its administration was dependent on each patient’s hemodynamics. Therefore, the fentanyl dose must have been < 1.5 ng ml−1 in prolonged surgery (> 3 h). At a lower level of analgesia, the patient returns to consciousness at a higher hypnotic level, which may explain why the longer infusion in our study had a higher Ce ROC [18]. Thus, our formula for predicting Ce ROC produced a higher propofol and fentanyl consumption and Ce ROC. Although the results were statistically significant, the clinical effect was minimal.
Operating room (OR) time is expensive, and it is estimated that the cost of one OR to the health consumer or insurance carrier is about US $10–30 per minute [19]. Surgeons, anesthesiologists, and hospital administrators continue to try to find ways to increase OR efficiency and profitability in the face of decreasing insurance coverage and increasing costs of health care. In clinical practice, the ability to predict revenue and clinical productivity is important for strategic planning for anesthesiology groups to reduce total procedure time, induction and emergence from anesthesia. Moreover, if the total reduction in non-operative time is large enough to allow more patients to be treated in the OR during regular business hours, the overall financial impact must be considered. Recently, we included data from 1405 patients, with 595 patients receiving TIVA and 810 receiving desflurane anesthesia in ophthalmic surgery. The extubation time was faster (1.85 min) and the PACU stay time was shorter (3.62 min) in the TIVA group than in the desflurane anesthesia group [20]. In our previous studies, TIVA using a TCI system provided faster emergence than volatile anesthesia in different surgeries, which increased the OR turnover rate [18, 21–22]. These results will have an economic impact on increasing OR productivity and reducing labor costs because our ORs are consistently used for more than 8 hours daily. Chiang and Wu (2013) demonstrated that TCI systems may facilitate the clinical management of TIVA in endoscope examination [23]. The anesthesiologist sets only the desired effect-site concentration as the target and the TCI pump adjusts the rate of delivery of the anesthetic agent according to a pharmacokinetic model. It is also important for evaluating the profitability of providing care to the general TIVA under TCI procedure. Previous studies have reported that these TCI anesthetic techniques are associated with a high level of patient and surgeon satisfaction, faster recovery time, and better hemodynamic and respiratory stability and safety [23–25].
Previous studies have reported conflicting results. Vuyk reported that the propofol concentrations at which 50% and 90% of the patients showed loss of consciousness were 3.4 μg ml−1 and 4.34 μg ml−1, respectively [26]. We found a lower CeLOC than that reported by Vuyk, possibly because their study did not include opioid administration [26]. Shafer reported that the predicted blood propofol concentrations at which 50% and 95% of patients were awake after surgery were 1.07 μg ml−1 and 0.52 μg ml−1, respectively, with intermittent bolus propofol [3]. In our study, we calculated the Ce of propofol and delivered propofol with the TCI machine. Schuttler and Kloos (1988) reported a Ce ROC of 1.59 μg ml−1 after 131 min of TCI propofol infusion, and Jung and Yang (2011) reported a Ce ROC of 1.4 μg ml−1 and CeLOC of 3.4 μg ml−1 with the Schnider model [27–28]. The values in these reports are all higher than those in our study. The differences between our study and the previous studies may relate to differences in the ethnicity of the patients, shorter propofol infusion times compared with ours (176 min), and the use of different opioids (sufentanil vs. fentanyl).
There are some limitations to this study. First, it was a retrospective study. (For example: without analgesic protocol) Second, our data reflect the experiences of a single academic medical center and therefore may be biased toward the hospital’s subspecialization. Third, the study lacks systematically collected data about patient satisfaction and potential adverse sequelae of the technique involved. Fourth, propofol causes hypotension, particularly in volume depleted patients [29]. In this study, transient hypotension was resolved by fluid challenge and ephedrine treatment from anesthetic records. This may be due to this study excluded emergent surgery and less volume depleted patients. Fifth, the retrospective feature of the study does not allow to analyze important data such as impact of propofol on cerebral hemodynamics and the effect of propofol on postoperative recall [30–31]. Finally, because the R2 was only 0.432 for the formula presented here, it is likely that more confounding factors are involved, and these should be investigated further.
In conclusion, we confirmed that age, sex, CeLOC, and both total fentanyl and propofol consumption amounts were independent factors that predicted Ce ROC during propofol infusion under TCI with the Schnider model. Our results are applicable to clinical practice. Patients can be expected to awaken quickly upon completion of the procedure and discontinuation of drug administration. However, further study is needed to investigate the impact on OR turnover rate.
Author Contributions
Conceived and designed the experiments: CHC ZFW. Performed the experiments: SMC CHL YSH CCY CYK ZFW. Analyzed the data: MSL. Contributed reagents/materials/analysis tools: MSL. Wrote the paper: SMC ZFW.
References
- 1. Iwakiri H, Nishihara N, Nagata O, Matsukawa T, Ozaki M, Sessler DI. Individual effect-site concentrations of propofol are similar at loss of consciousness and at awakening. Anesth Analg. 2005; 100: 107–110. pmid:15616062
- 2. Nunes CS, Ferreira DA, Antunes L, Amorim P. Clinical variables related to propofol effect-site concentrations at recovery of consciousness after neurosurgical procedures. J Neurosurg Anesthesiol 2005; 17: 110–114. pmid:15840999
- 3. Shafer A, Doze VA, Shafer SL, White PF. Pharmacokinetics and pharmacodynamics of propofol infusions during general anesthesia. Anesthesiology 1988; 69: 348–356. pmid:3261954
- 4. Lin BF, Ju DT, Cherng CH, Hung NK, Yeh CC, Chan SM, et al. Comparison between intraoperative fentanyl and tramadol to improve quality of emergence. J Neurosurg Anesthesiol 2012; 24: 127–132. pmid:22089326
- 5. Nunes CS, Ferreira DA, Antunes L, Lobo F, Santos IA, Amorim P. Individual effect-site concentrations of propofol at return of consciousness are related to the concentrations at loss of consciousness and age in neurosurgical patients. J Clin Anesth 2009; 21: 3–8. pmid:19232933
- 6. Brody H. The aging brain. Acta Neurol Scand Suppl 1992; 37: 40–44.
- 7. Kreuer S, Schreiber JU, Bruhn J, Wilhelm W. Impact of patient age on propofol consumption during propofol-remifentanil anesthesia. Eur J Anaesthesiol 2005; 22: 123–128. pmid:15816591
- 8. Schnider TW, Minto CF, Gambus PL, Andresen C, Goodale DB, Shafer SL, et al. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998; 88: 1170–1182. pmid:9605675
- 9. Schnider TW, Minto CF, Shafer SL, Gambus PL, Andresen C, Goodale DB, et al. The influence of age on propofol pharmacodynamics. Anesthesiology 1999; 90: 1502–1516. pmid:10360845
- 10. Gan TJ, Glass PS, Sigl J, Sebel P, Payne F, Rosow C, et al. Women emerge from general anesthesia with propofol/alfentanil/nitrous oxide faster than men. Anesthesiology 1999; 90: 1283–1287. pmid:10319774
- 11. Hoymork SC, Raeder J. Why do women wake up faster than men from propofol anesthesia? Br J Anaesth 2005; 95: 627–633. pmid:16169889
- 12. Vuyk J. Drug interactions in anesthesia. Minerva Anestesiol 1999; 65: 215–218. pmid:10389395
- 13. Vuyk J, Oostwouder CJ, Vletter AA, Burm AG, Bovill JG. Gender differences in the pharmacokinetics of propofol in elderly patients during and after continuous infusion. Br J Anaesth 2001; 86: 183–188. pmid:11573657
- 14. Lysakowski C, Dumont L, Pellegrini M, Clergue F, Tassonyi E. Effects of fentanyl, alfentanil, remifentanil and sufentanil on loss of consciousness and bispectral index during propofol induction of anesthesia. Br J Anaesth 2001; 86:523–527. pmid:11573626
- 15. Lentschener C, Ghimouz A, Bonnichon P, Pepion C, Gomola A, Ozier Y. Remifentanil-propofol vs. sufentanil-propofol: optimal combinations in clinical anesthesia. Acta Anaesthesiol Scand 2003; 47: 84–89. pmid:12492803
- 16. Hentgen E, Houfani M, Billard V, Capron F, Ropars JM, Travagli JP. Propofol-sufentanil anesthesia for thyroid surgery: optimal concentrations for hemodynamic and electroencephalogram stability, and recovery features. Anesth Analg 2002; 95: 597–605. pmid:12198044
- 17.
Glass PS, Shafer SL, Reves JG. Intravenous drug delivery systems. In: Miller RD: Miller’s Anesthesia. 7th ed. Philadelphia: Elsevier Churchill Livingstone; 2009. pp. 997–999.
- 18. Chen JL, Kuo CP, Chen YF, Chen YW, Yu JC, Lu CH, et al. Do anesthetic techniques affect operating room efficiency? Comparison of target-controlled infusion of propofol and desflurane anesthesia in breast cancer surgery. J Med Sci 2013; 33: 205–210.
- 19. Guidelines to the ethical practice of anesthesiology. ASA newsl 1979: 43: 3–4. pmid:10241538
- 20. Lu CH, Yeh CC, Huang YS, Lee MS, Hsieh CB, Cherng CH, et al. Hemodynamic and biochemical changes in liver transplantation: A retrospective comparison of desflurane and total intravenous anesthesia by target-controlled infusion under auditory evoked potential guide. Acta Anaesthesiol Taiwan 2014; 52: 6–12. pmid:24999212
- 21. Wu ZF, Jian GS, Lee MS, Lin C, Chen YF, Chen YW, et al. Analysis of anesthesia-controlled operating room time after propofol-based total intravenous anesthesia compared with desflurane anesthesia in ophthalmic surgery: A retrospective study. Anesth Analg 2014; 119: 1393–1406. pmid:25211391
- 22. Chan SM, Horng HC, Huang ST, Ma HI, Wong CS, Cherng CH, et al. Drug Cost Analysis of Three Anesthetic Regimens in Prolonged Lumbar Spinal Surgery. J Med Sci 2009; 29: 75–80.
- 23. Chiang MH, Wu SC, You CH, Wu KL, Chiu YC, Ma CW, et al. Target-controlled infusion vs. manually controlled infusion of propofol with alfentanil for bidirectional endoscopy: a randomized controlled trial. Endoscopy 2013; 45: 907–914. pmid:24165817
- 24. Yeganeh N, Roshani B, Yari M, Almasi A. Target-controlled infusion anesthesia with propofol and remifentanil compared with manually controlled infusion anesthesia in mastoidectomy surgeries. Middle East J Anesthesiol 2010; 20: 785–793. pmid:21526662
- 25. Horng HC, Kuo CP, Ho CC, Wong CS, Yu MH, Cherng CH, et al. Cost analysis of three anesthetic regimens under auditory evoked potentials monitoring in gynecologic laparoscopic surgery. Acta Anaesthesiol Taiwan 2007; 45: 205–210. pmid:18251240
- 26. Vuyk J, Engbers FH, Lemmens HJ, Burm AG, Vletter A, Gladines MP, et al. Pharmacodynamics of propofol in female patients. Anesthesiology 1992; 77: 3–9. pmid:1610007
- 27. Schuttler J, Kloos S, Schwilden H, Stoeckel H. Total intravenous anesthesia with propofol and alfentanil by computer-assisted infusion. Anesthesia 1988; 43 Suppl: 2–7.
- 28. Jung SM, Yang CW, Oh JY, Cho CK, Kang PS, Lim YS, et al. Predicted effect-site concentration of propofol and sufentanil for gynecological laparoscopic surgery. Acta Anaesthesiol Scand 2011; 55: 110–117. pmid:21039351
- 29. De Comso G, Congedo E, Clemente A, Aceto P. Sedation in PACU: the role of propofol. Curr Drug Targets 2005; 6: 741–744. pmid:16305451
- 30. De Comso G, Cancelli I, Adduci A, Merlino G, Aceto P, Valente M. Changes in hemodynamics during isoflurane and propofol anesthesia: a comparison study. Neurol Res 2005; 27: 433–435. pmid:15949243
- 31. Aceto P, Congedo E, Lai C, Valente A, Gualtieri E, De Comso G. Dreams recall and auditory evoked potentials during propofol anaesthesia. Neuroreport 2007; 18: 823–826. pmid:17471074