Secondary Logo

Journal Logo

Original Article

The asleep–awake technique using propofolremifentanil anaesthesia for awake craniotomy for cerebral tumours

Olsen, K. S.*

Author Information
European Journal of Anaesthesiology: August 2008 - Volume 25 - Issue 8 - p 662-669
doi: 10.1017/S0265021508003633
  • Free



Awake craniotomy, the commonly used term for a craniotomy, where the patient is awake during whole or part of the operation, has for many years been used for epilepsy surgery. The technique has also been used in recent years for patients undergoing surgery for resection of cerebral tumours located near certain cortical areas. The advantage of an awake craniotomy is that this technique allows for intraoperative mapping of important areas such as motor cortex or areas involved in language functions, making it possible for the surgeon to avoid these areas and thus to prevent significant postoperative functional disabilities. Furthermore, it allows the surgeon to operate on patients otherwise considered inoperable due to the localization of the tumour close to important areas.

Anaesthesia for awake craniotomy may, however, be a challenging procedure for the anaesthesiologist. The success of the technique relies first of all upon the skilled performance of local anaesthetic infiltration and nerve blocks to produce effective analgesia at the operative site. The patient must be under general anaesthesia (or be sufficiently deeply sedated) to be able to tolerate the injection of a substantial volume of local analgesics as well as the first phase of the operation with the placement of the head frame, the skin incision, the craniotomy and the dural opening. At the same time the patient should shortly thereafter during the mapping procedure be awake, pain free and fully co-operative with sufficient spontaneous respiration. Furthermore, the procedure exposes the patient to potential anaesthetic hazards such as respiratory depression, airway obstruction, haemodynamic instability, disinhibition and nausea.

Many different anaesthetic methods have been tried to overcome these obstacles. The purpose of the present study was to present our initial experience with and validate an alternative to the often-used ‘asleep–awake–asleep' technique, namely the ‘asleep–awake' technique for tumour resection, which allows for close monitoring (patient awake) and additional mapping, if necessary, throughout the whole phase of surgery.


From May 2004 to February 2006, 25 awake craniotomies were planned. Following each operation, selected parameters (e.g. time of anaesthesia, time of the awake period, complications, treatment of complications, doses of remifentanil administered during the awake period and type of mapping) were entered into a database. After 25 craniotomies the database was reviewed as were all the anaesthetic cards to collect further information. Thus, this study was a retrospective study with some prospectively recorded data. The team involved consisted of two neurosurgeons, two neuroanaesthesiologists, a neuropsychologist, three nurse anaesthetists and four operating-theatre nurses. One or more from each group participated in each operation.

Patients were selected for the operation by the neurosurgeon and subsequently assessed by the neuropsychologist and the anaesthesiologist regarding the suitability of this particular patient for this technique. The patient was carefully informed about the operation, the awake period and the potential problems and complications. Premedication (oral midazolam) was only used as an exception. Propofol and remifentanil were used for induction and maintenance of anaesthesia. Remifentanil infusion was started at 0.7 μg kg−1 min−1. When the patient could feel an effect of this infusion, propofol was injected as a bolus dose until loss of eyelid reflexes. Then an infusion of propofol was started (approximately 0.17 mg kg−1 min−1). A laryngeal mask was placed and the patient was ventilated with an oxygen–air mixture at an FiO2 of 0.6 and end-tidal CO2 (etCO2) 4.0–4.5 kPa. After a few minutes both infusions were reduced (typically by about 40–50%). As soon as the patient was asleep, an arterial catheter was placed in a radial artery and a urinary catheter (with a temperature sensor) was placed in the bladder. If the mean arterial pressure (MAP) decreased below 65 mmHg phenylephrine 0.1 mg intravenously (i.v.) was administered as necessary. Before the pins of the head frame were screwed into the cranium, the infusion of remifentanil was increased by 25% for a few minutes. Lidocaine 20 mg mL−1 plus epinephrine 5 μg mL−1 was used as local anaesthetic as the pins of the head frame were placed and bupivacaine 5 mg mL−1 plus epinephrine 5 μg mL−1 (15–20 mL) was used at the incision site and for nerve blocks of relevant peripheral nerves of the scalp (performed by the surgeon).

Before the surgery, paracetamol i.v. (1–2 g) was infused. Dehydrobenzperidol (0.6 mg), ondansetron (4 mg) and dexamethasone (8 mg) were administered as antiemetics. Mannitol was given preoperatively on request of the surgeon. The patient was covered by a blanket, which was connected to a forced-air temperature management unit (BairHugger®; Arizant Inc., Eden Prairie, MN, USA). Transparent drapes were used to separate the face of the patient from the operation field.

After the craniotomy and the dural incision, the surgeon estimated the time remaining to the beginning of the mapping, and the propofol infusion was reduced accordingly for some minutes, after which both infusions were stopped. The laryngeal mask was removed when the patient was awake. During the anaesthesia phase and in the awake phase of the operation, the following parameters were monitored: electrocardiogram (ECG), oxygen saturation, temperature, blood pressure (BP) (measured invasively), etCO2 and respiratory rate. To monitor the respiratory rate in the awake phase, the etCO2 was measured via a binasal cannula with a port for etCO2 monitoring (MAC-SAFE Nasal Cannula; Unomedical Inc., McAllen, TX, USA). Oxygen, 3–4 L min−1, was administered via the same catheter. An infusion of remifentanil in a small dose (typically 0.03 μg kg−1 min−1) was started if the patient felt any pain.

The mapping (electrocorticography and/or localization of motor, sensory and language areas by cortical stimulation) was performed by the neurosurgeon and the neuropsychologist (the latter communicating with the patient during the awake phase). If electrocorticography was carried out, a neurophysiologist participated as well.

We aimed to keep the patient awake after the mapping, allowing for additional mapping if necessary in order to be able to discover developing symptoms caused by the resection. However, if an infusion of remifentanil in a relatively low dose (i.e. a dose not depressing the respiratory rate too much) could not alleviate the pain, then the patient would be re-anaesthetized. Accordingly, the patients were told that if they felt uncomfortable with the procedure they could be re-anaesthetized.

The patients stayed in the neurointensive care department for 24 h. In the first postoperative hours analgesia consisted of paracetamol and, if necessary, an infusion of a small dose of remifentanil, which was reduced successively.


Patient data and data regarding the tumours are shown in Table 1. Perioperative data are shown in Table 2. The mean anaesthesia time was 145 min. The mapping was carried out as planned in 23 cases. One patient (no. 8) had a tight brain, which increasingly protruded through the dural opening during the awakening phase. Accordingly, the laryngeal mask was not removed, the anaesthesia was deepened, hyperventilation was applied and the operation carried out without mapping. In another patient (no. 10) it was impossible to obtain a tightly fitting laryngeal mask and a relative high FiO2 was necessary to keep the patient sufficiently oxygenated. The patient was eventually intubated (preoperatively) with an endotracheal tube and during light anaesthesia only motor mapping was carried out. Postoperatively, the patient was mechanically ventilated due to atelectasis. In a third patient (no. 13) the laryngeal mask was not tight either and this patient had to be intubated. The patient was awakened and extubated without complications before the mapping, which was carried out as planned.

Table 1
Table 1:
Patient characteristics and perioperative data.
Table 2
Table 2:
Perioperative data.

One patient (no. 2) had a grand mal seizure (related to cortical stimulation), which was easily controlled by thiopental 50 mg. A second attack a few minutes later was controlled by thiopental 50 mg and midazolam 0.5 mg. No problem with airway control or respiration was encountered after these injections. The patient received a loading dose of phenytoin i.v. All the succeeding patients (except no. 14, who was allergic to phenytoin) received a loading dose of phenytoin preoperatively. Three patients had minor, localized convulsions, which stopped either spontaneously or after flushing the brain with cold saline (no. 11 and no. 23) or after administering a small dose of propofol (no. 21). In one of these patients (no. 23) the resection in one area was stopped as several small seizures with an accompanying few seconds of aphasia developed. Two patients (nos. 12 and 14) shivered immediately after they were awakened. This might have been due to a low central temperature although both patients confirmed that they did not feel cold. In one of these patients the rectal temperature was 36.3°C, but in the other patient no temperature was recorded at the time of awakening. In both patients, the shivering stopped after a small dose of midazolam (2 and 0.5 mg, respectively). One patient experienced nausea and vomited once and therefore received an additional dose of antiemetics with good effect. In one patient the MAP increased (from 80 to 120 mmHg) after awakening and this was effectively treated with labetalol. No patient was hypertensive (MAP > 90 mmHg) during anaesthesia.

The median dose of propofol administered for induction was 200 mg (2.4 mg kg−1) with a range of 150–350 mg. The doses of propofol and remifentanil administered during the anaesthesia phase were adjusted according to the BP and pulse rate and hence it was not possible to generate useable data regarding mean values. The median time (n = 21) from stopping the infusions to laryngeal mask removal (and the patient being awake) was 15 min (range 5–39 min). Typically after a further 2–3 min the patient was ready to undergo mapping. Except for the previously noted problems the removal of the laryngeal mask was uneventful. The patients either regained spontaneous breathing shortly before they woke up or started breathing at the same moment they woke up. A few patients were very tired during the awake phase (Table 2), but in all patients awakened mapping could be performed. All the patients awakened stayed awake during the rest of the surgery. The mean awake time was 165 min. The longest awake time was 245 min.

In five patients, resection of the tumour near areas identified as important during the mapping was modified due to the development of symptoms (Table 2).

If a patient experienced pain, this was easily managed with an infusion of remifentanil in a small dose (19 patients received remifentanil for an average time of 87 min, the mean value of the average dose in each patient was 0.036 μg kg−1 min−1; the mean value of the maximal dose given to each patient was 0.054 μg kg−1 min−1) under close monitoring of the sedation level and of the respiratory rate (Table 2). No patient had a respiratory rate below 10 breaths min−1 during this phase. The highest etCO2 recorded was 6.1 kPa and the mean was 5.35 kPa (n = 20) and the highest arterial carbon dioxide tension measured was 6.1 kPa. There was no episode of hypoxia during either the anaesthesia period or the awake period of the operation. Seven, 14 and 18 patients received remifentanil during the mapping, resection of tumour and closure, respectively. Two patients needed a supplement of local analgesic during the last part of the operation (suturing muscle and skin after 2 h and almost 5 h of surgery, respectively).

In the first postoperative hours all patients were comfortable with no or minimal pain. No postoperative complications seemed to be related to the postoperative infusion of remifentanil. All patients were satisfied with the awake part of the operation. However, a few of the patients had some degree of amnesia related to this period although they had had intact cognitive functions and had been awake.


Many different anaesthetic techniques for awake craniotomy have been described in the literature [1-7]. Methods ranging from local analgesia supplemented with either light or more deep sedation (monitored conscious sedation) [5] to general anaesthesia with the patient intubated have been applied [7]. Several of the methods have been adapted from those used for epilepsy surgery, where mapping has been used for many years. The general aim, however, is that the patient is alert and fully cooperative during the mapping phase.

The advantage of a monitored conscious sedation technique is that the patient, in principle, is awake and has spontaneous respiration during the whole procedure. Potential complications related to the wake-up procedure are thus omitted. On the other hand, it may be difficult to titrate the level of the sedation with a consequent risk of respiratory depression and/or airway problems. In a prospective study comprising 21 patients undergoing monitored conscious sedation (using midazolam, fentanyl or sufentanil with or without propofol), the subjective experiences of the patients were addressed. It was concluded that the method was useful and safe. However, a significant number of patients experienced minor or more significant problems with this method [8]. A variation of the monitored conscious sedation technique is neurolept analgesia [9], typically based on dehydrobenzperidol and fentanyl. Differences in the half-lives of these drugs make this kind of sedation difficult to titrate and it has been associated with excessive sedation, higher incidence of pain and convulsions [8]. Other drugs with sedative and analgesic effects such as clonidine or dexmedetomidine that have less respiratory depressant action have been used with some success [3,10,11].

Keifer and colleagues [12] used a propofolremifentanil technique with spontaneous ventilation. However, they frequently observed respiratory depression. In 69 of 96 patients at least one 30-s epoch of apnoea was recorded. Several techniques have been applied to minimize problems with the airway and with respiratory depression during the first part of the operation. Yamamoto and colleagues [13] used non-invasive positive pressure ventilation. Controlled ventilation has been applied via an endotracheal tube [7], a laryngeal mask [2] or a cuffed oropharyngeal airway [14], all of which typically have been removed during the transition period between anaesthesia and arousal. The laryngeal mask may be preferable compared to the endotracheal tube (easier to remove without the patient coughing and easier to replace intraoperatively). On the other hand, it may be a problem to use a laryngeal mask or an airway in patients with risk of aspiration, low lung compliance, high airway resistance, limited opening of the mouth, etc. Furthermore, it may be difficult or impossible in some patients to obtain a tight laryngeal mask fit as we experienced with two patients in the present series.

Most centres re-anaesthetize the patients (or increase the depth of sedation) after the mapping phase and replace the laryngeal mask [1,4], the airway [14] or the endotracheal tube [7]. The patients in the present study, however, stayed awake during the rest of the operation, allowing for additional mapping or for stopping the resection of the tumour if symptoms should occur. This was the case in five patients (Table 2) where the resection of part of the tumour was omitted or modified. The symptoms of these patients subsided in due course. One might suppose that if the patients had not been awake in these situations, the resections would have been more extensive with more severe symptoms as a result postoperatively. This suggests that the asleep–awake method could be advantageous compared with the asleep–awake–asleep method and that studies concerning this could be relevant.

Recently, some series with propofolremifentanil anaesthesia without [1-3,6] or with [1,4] a laryngeal mask have been reported. Only a few studies, however, comparing different methods of anaesthesia have been published. Sareng and colleagues [1] retrospectively compared three different methods: (1) monitored conscious sedation (fentanyl, dehydrobenzperidol or midazolam and propofol); (2) the same regime as (1) but using a laryngeal mask; and (3) total i.v. anaesthesia using propofol and remifentanil (for an asleep–awake–asleep technique, where the laryngeal mask was replaced after the mapping). The data were inadequate, but the impression was that the third regime was superior. Manninen and colleagues [15], in a randomized study, compared remifentanil and fentanyl in conjunction with propofol and concluded that remifentanil was a good alternative to fentanyl for monitored conscious sedation for awake craniotomy and that both techniques were well accepted by the patient. In the present study, propofolremifentanil anaesthesia provided a stable and sufficient anaesthesia. After both infusions were stopped simultaneously the patients generally first obtained spontaneous ventilation and shortly thereafter woke up. The patients could, as they were instructed about immediately before the induction of the anaesthesia, take a deep breath and open the mouth on command allowing the laryngeal mask to be removed.

Several anaesthetic challenges are related to an awake craniotomy. The most critical phases are the transition from controlled ventilation to spontaneous respiration (increase in PaCO2) and the removal of the laryngeal mask (insufficient respiration, coughing). Several factors can increase the tension of the brain and thereby induce protrusion of the brain through the dural incision. Hypoventilation increases PaCO2 with subsequent cerebral vasodilatation and increased cerebral blood volume. Coughing increases the intrathoracic pressure and reduces the outflow of cerebral venous blood with increased cerebral blood volume as a result. Lightening of anaesthesia during the wake-up phase increases the cerebral metabolic rate, the cerebral blood flow and the cerebral volume. Furthermore, handling of these problems requires free airways and sufficient ventilation. Manual ventilation via a facemask may be difficult (due to the draping, the fixation of the head to the frame and the position of the head). Replacement of a laryngeal mask generally seems to be without problems, whereas intubation using a laryngoscope could be difficult under these conditions. A fibrescope-assisted intubation could be an alternative but may be time consuming, which under these circumstances could be fatal. These considerations were relevant in one patient (no. 8) in whom an increasing protrusion of the brain was seen during the wake-up phase. The patient, probably reacting to the laryngeal mask, had a regular but shallow respiration. In this situation we had to decide whether or not to remove the laryngeal mask. On the one hand removal of the laryngeal mask could have improved respiration but on the other hand it could have violated the free airway. We chose to re-anaesthetize and hyperventilate the patient, and the operation was carried out without mapping.

A difficult airway preoperatively may be considered as a relative contraindication for awake craniotomy [16], and the balance between the possible benefits and the risks must be considered in each patient. It was impossible to obtain a tight laryngeal mask fit in two patients. One of these (no. 10) weighed 116 kg. Initially, the laryngeal mask was tight, but after placement of the head frame and positioning of the head in a position suitable for the surgeon, it was not possible to obtain adequate ventilation. Accordingly, the patient was intubated and was not awakened during the operation. The second patient (no. 13) was also intubated, but it was decided that it would be acceptable to extubate this patient to be able to carry out the mapping. The extubation was uncomplicated and the rest of the operation was carried out with the patient awake and with no complications.

In retrospective studies of different anaesthetic techniques, these problems and complications are found with varying frequencies [5,8,9,17]: Too heavy sedation 3%, nausea and vomiting 1–8%, agitation 7%, pain 8–10%, seizures 9–16%, tight brain 1.4% and induction of general anaesthesia during the operation 0.9–2%. Given the limited numbers of patients in the present study, the complications recorded seem to be in accordance with these numbers. No patient in our study, however, could be described as agitated. Amnesia for the awake period, as experienced by some patients in the present study, has been described before and it has been suggested that between 8% and 37% of patients will have no or little recollection of the procedure [18].

Due to the pain related to the surgery or due to the discomfort from the protracted placement on the operating table, several patients had an infusion of a relatively small dose of remifentanil during the awake part of the operation. No patients had complications related to this infusion. The highest dose administered during this phase was 0.12 μg kg−1 min−1. Herrick and colleagues [19], in a study on the effect of sedative doses of remifentanil on intraoperatively recorded inter-ictal spike activity, found that the mean decrease in respiratory rate during a remifentanil infusion of 0.1 μg kg−1 min−1 was 8 breaths min−1 and that one patient transiently developed a respiratory rate of 4 breaths min−1. We did not see a respiratory rate below 10 breaths min−1 in any of our patients and the carbon dioxide level did not seem to be inappropriately high. In no case did the surgeon complain of a tight brain. The relatively high frequency of patients receiving remifentanil during the awake period thus does not seem to cause a problem for the asleep–awake technique. Remifentanil was also used for postoperative analgesia. Remifentanil is often used in our neurointensive care department and the nurses are familiar with this drug, which makes it possible to titrate the analgesia very precisely and hence to reduce the postoperative analgesia fast. Anaesthesia-related postoperative problems after an awake craniotomy are not common and do not seem to be related to the awake course. On the contrary, a shorter postoperative stay in intensive care and an earlier discharge have been described [17]. Some awake craniotomies are now even scheduled as ambulatory surgery [20].

In conclusion, we find the ‘asleep–awake' technique using propofolremifentanil anaesthesia in combination with a laryngeal mask to be a rational and safe method of anaesthesia for awake craniotomies, provided the laryngeal mask is a tight fit. The method is well tolerated by patients and allows for additional mapping during the operation and for stopping of resection should symptoms occur. The complications we experienced are comparable with the results of other studies of awake craniotomy. All of our patients were satisfied with the method. The use of remifentanil in the awake phase of the operation and postoperatively was effective and no complications were encountered.


1. Sareng A, Dinsmore J. Anaesthesia for awake craniotomy – evolution of a technique that facilitates awake neurological testing. Br J Anaesth 2003; 90: 161–165.
2. Hans P, Bonhomme V, Born J et al. Target-controlled infusion of propofol and remifentanil combined with bispectral index monitoring for awake craniotomy. Anesthesia 2000; 55: 255–259.
3. Berkenstadt HPA, Hadani M, Unofrievich I, Ram Z. Monitored anesthesia care using remifentanil and propofol for awake craniotomy. J Neurosurg Anesthesiol 2001; 13: 246–249.
4. Tongier WK, Joshi GP, Landers DF, Mickey B. Use of laryngeal mask airway during awake craniotomy for tumor resection. J Clin Anesth 2000; 12: 592–594.
5. Andrews Danks R, Aglio LS, Gugino LD, Black PM. Craniotomy under local anesthesia and monitored conscious sedation for the resection of tumors involving eloquent cortex. J Neurooncol 2000; 49: 131–139.
6. Johnson K, Egan T. Remifentanil and propofol combination for awake craniotomy: case report with pharmacokinetic simulations. J Neurosurg Anesthesiol 1998; 10: 25–29.
7. Huncke K, Van de Wiele B, Fried I, Rubinstein EH. The asleep-awake-asleep anesthetic technique for intraoperative language mapping. J Neurosurg 1998; 42: 1312–1316.
8. Danks RA, Rogers M, Aglio L, Gugino L, Black PM. Patient tolerance of craniotomy performed with the patient under local anesthesia and monitored conscious sedation. Neurosurgery 1998; 42: 28–36.
9. Archer DP, McKenna JM, Morin L et al. Conscious-sedation analgesia during craniotomy for intractable epilepsy: a review of 354 consecutive cases. Can J Anaesth 1988; 35: 338–344.
10. Mack PF, Perrine KP, Schwartz TH, Lien CA. Dexmedetomidine and neurocognitive testing in awake craniotomy. J Neurosurg Anesthesiol 2004; 16: 20–25.
11. Ard JL, Bekker AY, Doyle WK. Dexmedetomidine in awake craniotomy: a technical note. Surg Neurol 2005; 63: 114–116.
12. Keifer JC, Dentchev D, Little K, Warner DS, Friedman AH, Borel CO. A retrospective analysis of a remifentanil/propofol general anesthetic for craniotomy before awake functional brain mapping. Anesth Analg 2005; 101: 502–508.
13. Yamamoto F, Kato R, Sato J, Nishino T. Anaesthesia for awake craniotomy with non-invasive positive pressure ventilation. Br J Anaesth 2003; 90: 382–385.
14. Audu PB, Loomba N. Use of oropharyngeal airway (COPA) for awake intracranial surgery. J Neurosurg Anesthesiol 2004; 16: 144–146.
15. Manninen PH, Balki M, Lukitto K, Bernstein M. Patient satisfaction with awake craniotomy for tumor surgery: a comparison of remifentanil and fentanyl in conjunction with propofol. Anesth Analg 2006; 102: 237–242.
16. Sartorius CJ, Wright G. Intraoperative brain mapping in a community setting-technical considerations. Surg Neurol 1997; 47: 380–388.
17. Blanshard HJ, Chung F, Manninen PH, Taylor MD, Bernstein M. Awake craniotomy for removal of intracraniel tumor: considerations for early discharge. Anesth Analg 2001; 92: 89–94.
18. Whittle IR, Midgley S, Georges H, Pringle AM, Taylor R. Patient perception of “awake” brain tumour surgery. Acta Neurochir 2005; 147: 275–277.
19. Herrick IA, Craen RA, Blume WT, Novick T, Gelb AW. Sedative doses of remifentanil have minimal effect on ECoG spike activity during awake epilepsy surgery. J Neurosurg Anesthesiol 2002; 14: 55–58.
20. Bernstein B. Outpatient craniotomy for brain tumor: a pilot feasibility study of 46 patients. Can J Neurol Sci 2001; 28: 120–124.


© 2008 European Society of Anaesthesiology