Secondary Logo

Journal Logo

Airway management

Best position and depth of anaesthesia for laryngeal mask airway removal in children

A randomised controlled trial

Thomas-Kattappurathu, George; Kasisomayajula, Ananth; Short, Judith

Author Information
European Journal of Anaesthesiology (EJA): September 2015 - Volume 32 - Issue 9 - p 624-630
doi: 10.1097/EJA.0000000000000286
  • Free



The laryngeal mask airway (LMA) has been ubiquitous in anaesthetic clinical practice since it was invented more than 30 years ago1 and was introduced into paediatric practice around 1990.2,3 A number of studies have demonstrated the safety and efficacy of the LMA in paediatric anaesthesia,4–8 but the manufacturer's recommended technique for removal of the LMA has proved controversial in this group of patients. In adult anaesthesia, routine recommended practice is to leave the patient undisturbed at emergence from general anaesthesia and remove the LMA when the patient awakens and is able to obey commands.9 In children, the technique of LMA removal is highly variable, as children are not comparable with adults in this regard.10 Advantages of removal of the LMA in the deeply anaesthetised child may include the avoidance of biting on the stem of the LMA, accumulation of secretions within the pharynx and dislodgement of the LMA with the risk of airway stimulation.

There have been a number of studies comparing removal of the LMA whilst deeply anaesthetised with fully awake, but the results are inconsistent.11–18

Given the controversy in the literature, we undertook a two-centre observational audit of airway complications after LMA removal, noting the depth of anaesthesia and patient position at the time of removal, which suggested that there may be some advantages to the lateral position. The results of that audit were presented as a poster at the Association of Paediatric Anaesthetists’ annual scientific meeting in May 2008. The present study was designed to test under formal clinical trial conditions whether there are any differences in airway complications associated with removal of the LMA in the deeply anaesthetised or awake planes of anaesthesia, with the patient in the supine or lateral position.

Materials and methods

This randomised, single-blind, noncrossover study was approved by the Sheffield Research Ethics Committee, Sheffield, UK (Chair: Dr C.A. Moore) on 6 October 2009 (REC reference number 09/H1308/109) and was registered by the Sheffield Children's NHS Foundation Trust ‘Children's Clinical Research Facility’ (Study number CA08014), who retain a copy of the full protocol.

Parents of all the children who were scheduled to have day surgery were sent patient and parent information leaflets along with their appointment letter. On the day of surgery, suitable children were identified at the routine preanaesthetic visit by the anaesthetist by determining that anaesthesia using an LMA was appropriate for that patient undergoing that procedure and that randomisation into any of the potential study groups would be appropriate. Other inclusion criteria were American Society of Anesthesiologists’ (ASA) physical status 1 to 2, age 1 to 16 years and informed consent from parents (with assent from patients as appropriate). Patients were excluded if they were ASA 3 or above, had clinically significant congenital heart disease or gastro-oesophageal reflux or an anticipated difficult airway. Other exclusion criteria included infants under 1 year of age, patients scheduled to undergo airway or dental surgery (which may be associated with a blood-soiled airway postoperatively) and patients for magnetic resonance scanning (due to the remote location). Recruitment began on 7 December 2009 and was completed on 8 October 2010. The trial observations were completed for each patient on the day of surgery; no further follow-up was required.

All patients were anaesthetised by a consultant paediatric anaesthetist or a senior trainee with at least 3 months’ experience of paediatric anaesthesia. We wished our study to reflect routine practice as far as possible but attempted to minimise bias by standardising a number of important aspects of the anaesthetic technique. After induction of anaesthesia using either propofol or sevoflurane, a classic single-use LMA (Solus, Intersurgical Ltd., Wokingham, UK) was inserted using the anaesthetist's usual technique. A manometer was used to ensure that the cuff pressure remained below the manufacturer's recommended maximum of 60 cmH2O. All patients received a volatile-based anaesthetic using sevoflurane. The use of nitrous oxide, local anaesthetic techniques and supplementary analgesia (including opioids), pressure support ventilation strategies and continuous positive airway pressure (CPAP) were all allowed as appropriate for the type of surgery and documented accordingly.

Approximately 5 min prior to the end of the surgery, each patient was allocated randomly to one of four groups for removal of the LMA. Randomisation was achieved by the use of random numbers (generated in advance by computer, using SPSS Version 17.0), which determined the order of allocation of successive recruits to each group. Sealed envelopes detailing the allocation for each recruit and the relevant data forms were stored securely in the theatre department. Neither block randomisation nor strata were applied. The four groups were:

  1. lateral deep, removal deeply anaesthetised in the lateral position;
  2. lateral awake, removal awake in the lateral position;
  3. supine deep, removal deeply anaesthetised in the supine position; and
  4. supine awake, removal awake in the supine position.

At the completion of surgery, the end-tidal sevoflurane concentration was titrated to 2.2% to ensure adequate depth of anaesthesia for patient movement and airway manipulation. The patient was then transferred to the transport trolley and placed in the supine or lateral position as required. For a further minute, patients breathed sevoflurane in oxygen to maintain the end-tidal sevoflurane concentration of 2.2% and to allow washout of nitrous oxide or air.

In awake groups, the sevoflurane was then discontinued and the patient transported to the postanaesthetic care unit (PACU) with supplemental oxygen delivered via a ‘T-bag’ (Ultimate T-Bag; Intavent Orthofix Ltd., Maidenhead, UK) at a flow rate of 3 l min−1 (inspired oxygen concentration ∼50%). The LMA was removed by the PACU nurse when the patient was fully awake, defined by being able to open the eyes, showing purposeful movements to remove the LMA or obeying commands.

In the deeply anaesthetised groups, after breathing sevoflurane in oxygen for 1 min as described above, the LMA was removed by the anaesthetist. Any airway complication occurring immediately after removal were managed by the anaesthetist and recorded (e.g. airway manipulation, insertion of an oropharyngeal airway, change of position). In all four groups, supplemental oxygen was delivered via a Hudson mask at a flow rate of 6 l min−1 (inspired oxygen concentration ∼50%) after removal of the LMA.

In all four groups, any airway complications (defined as desaturation <90%, stridor or upper airway obstruction requiring manipulation of the airway/use of airway adjuncts, complete laryngospasm with paradoxical respiratory movements, retching/vomiting or excessive secretions requiring suction or biting on the stem of the LMA) occurring within 1, 5 and 15 min following removal of the LMA were recorded. During emergence from anaesthesia and the study observation period following removal of the LMA, the attending anaesthetists or PACU nurses were free to manage any complications or adverse events as appropriate.

The primary outcome measure was the number of patients experiencing one or more complication in each group.

Because not all airway complications recorded are associated with the same level of risk, we wished to determine the clinical impact of the different depths of anaesthesia and patient positions for LMA removal. Our secondary outcome was obtained by the creation of a novel ‘Clinical Importance’ score, as follows. At a meeting of the consultant paediatric anaesthetists working in our institution (n = 12), we asked the clinicians to agree a consensus score on a scale of 1 to 10 for each of the six airway complications we were to observe, according to the risk they attached to each. The percentage contribution of each score to the possible total was then calculated. This created a standardised scoring system with a total of 100 points, so that each patient could have a maximum score of 100 if they suffered all six complications, as summarised in Table 1.

Table 1
Table 1:
Consensus scoring of airway complications by consultant anaesthetists for creation of ‘Clinical Importance’ score


Due to the wide range of complication rates reported in previous studies of LMA removal in children, and because there were no directly comparable studies, we conducted a two-centre pilot observational audit prior to designing the present study. This indicated that the complication rates in children placed lateral for LMA removal ranged from 8.7% under deep anaesthesia to 12.8% awake; in the supine groups, complication rates ranged from 17.3% with deep anaesthesia deep to 30.7% awake (unpublished data). Estimating these to be true population proportions, and based on a study power of 80% and significance level of 5%, we calculated that a sample of 54 patients would be required in each group in order to detect a statistically significant difference of at least 15% among the groups with highest and lowest complication rates should one be present.

The primary outcome and the nonparametric aspects of the demographic data were analysed using the χ2 measure of association. Parametric demographic measures were analysed using analysis of variance (ANOVA), including the ‘Clinical Importance’ score calculated for each group. In all cases, a significance level of 5% was used and the data were analysed using SPSS Statistics for Windows, version 17 (IBM, Armonk, New York, USA).


A total of 216 patients were recruited as planned and data were analysed for 212, four patients having been excluded for protocol violations (Fig. 1). There were 146 boys and 66 girls, with ages ranging from 1 to 16 years. There were no differences in the demographic parameters and the size of LMA among the four groups (Table 2).

Fig. 1
Fig. 1:
CONSORT flow diagram.
Table 2
Table 2:
Demographic data for study population

Table 3 summarises the primary outcome measure, which is the number of patients with one or more airway complication recorded for each group and the overall number of complications. Where partial upper airway obstruction occurred, this was recognised immediately and managed by the attending anaesthetist or nurse. In most cases, simple airway manoeuvres or insertion of an orophayngeal airway was sufficient to restore a clear airway. Seven initially supine patients were also turned into the lateral position to aid in airway management during recovery, but no patients were turned from the lateral to supine position at any point during the study. No patients experienced laryngospasm on LMA removal and there were only two cases of desaturation below 90%. A patient in the supine deep group desaturated briefly to 66% when airway obstruction associated with stridor occurred immediately after removal of the LMA; manual support of the airway was required in addition to an oropharyngeal airway. A further patient in the supine awake group desaturated to 88% after biting on a partially removed LMA. It was noted that excessive secretions requiring suction clearance and biting on the stem of the LMA were complications recorded only in the awake removal groups.

Table 3
Table 3:
Number of occurrences of each airway complication recorded for patients in the four groups after removal of the laryngeal mask airway

Analysis of the overall complication rates showed highly significant differences between the four groups (P = 0.001). This is summarised in Table 4 alongside the results for the secondary outcome, the mean ‘Clinical Importance’ score ascribed to the complications that occurred in each group. Analysis of these scores showed that there were highly significant differences (P < 0.001) between the groups, with more complications of clinical significance occurring in the supine deep group. Although complications of several types occurred in all four groups, it was the high incidence of upper airway obstruction in the patients whose LMA was removed when they were deeply anaesthetised in the supine position, which led to the significantly higher score for clinical importance in this group.

Table 4
Table 4:
Four group analysis of complication rates and the ‘Clinical importance’ score ascribed to airway complications in each group


This study was designed to address a relevant clinical question regarding the influence of patient positioning for LMA removal in children on the incidence of airway complications. The results of the study provide evidence for a reduction in complications when patients are turned into the lateral position for removal of the LMA while deeply anaesthetised. We also attempted to stratify the airway complications by their clinical impact. The lowest scores were obtained in the lateral position with the lowest incidence of airway obstruction when the children were awake and in the lateral position.

We wished to undertake a study that would be as clinically relevant as possible. Therefore, we aimed to keep patient exclusion criteria to a minimum and to allow patients undergoing a wide range of procedures to be included. We excluded infants under 1 year of age, as this group of paediatric patients has been shown previously to have a higher rate of airway complications associated with LMA use19–21 and they may also be more difficult to position laterally. We also excluded patients who would have a blood-soiled airway postoperatively, as these patients have been shown previously to benefit from LMA removal once completely awake.17 It is generally agreed that removal of the LMA should not take place during emergence from anaesthesia at a time before airway reflexes have fully returned due to an increased risk of airway reactivity and also that patient movement should not be undertaken during emergence for similar reasons. For this reason, we defined the depth of anaesthesia at which the patient should be transferred to the transport trolley and positioned and for LMA removal in the ‘deep’ groups corresponding to an end-tidal sevoflurane concentration of 2.2%, which has been shown previously to be sufficient for uncomplicated airway manipulation.18,22 Unlike a number of previous studies investigating airway events on LMA removal, we did not include coughing as a complication. This was partly due to the difficulty of clearly defining a level of coughing that would be considered undesirable and the fact that a small amount of coughing may actually be helpful in clearing secretions after removal of the LMA. In a study that included coughing as a complication, it was noted that in most cases, coughing was minimal and presented no hazard to the patient.23

There have been many previous studies to investigate the complications associated with LMA removal in the awake or deep planes of anaesthesia in children, but no previous studies have specifically discussed the influence of patient position at the point of LMA removal. However, many studies have mentioned patient positioning as part of the study methodology. For example, all patients turned into the lateral position for LMA removal (both deeply anaesthetised and awake)10,12,16–17 or all patients turned lateral after LMA removal.13,18 In one study, the position of the deeply anaesthetised patients was not clear, but all patients having awake LMA removal were turned lateral.15 Although the reasoning for lateral positioning of patients in the studies cited is not discussed, its use implies some advantage over the supine position for LMA removal in children, even though this is not the standard practice in adults. Our study supports this view, showing that approximately half of anaesthetised children required manipulation of the airway (head tilt, chin lift, jaw thrust) or an airway adjunct (most commonly an oropharyngeal airway) in order to avoid upper airway obstruction while supine compared with a much lower proportion when lateral. It is the best practice for anaesthetists to minimise this risk, particularly at a point when they may be handling their patients over to other healthcare professionals for recovery. Even routine replacement of the LMA with an oropharyngeal airway may not preclude the need for additional supportive airway manoeuvres in the supine patient.

We have also shown that, although the risk of airway obstruction is lower when the LMA is removed with the patient awake rather than deeply anaesthetised, the overall complication rates in both awake groups were increased by the incidences of excessive secretions and biting of the stem of the LMA seen in these patients. The effect of excessive secretions is not unimportant, particularly if they pool above the LMA cuff in the pharynx during emergence. This risk can be reduced in the lateral position, by allowing secretions to drain away. As mentioned above, lateral awake LMA removal is preferred for patients after dental or ENT surgery to avoid pooling of blood-soiled secretions and the need for suction of the pharynx, which could disturb haemostasis. In the lateral position, secretions isolated from the pharynx by the LMA are expelled horizontally from the mouth as it is removed. Even though our study was done in patients without soiled airways, the results support this practice as the overall complication rate was considerably lower in the awake patients who were turned lateral than those who remained supine.

Although unvalidated, our locally created ‘Clinical Importance’ score also concurs with the view that the greatest clinical risk occurs when the LMA is removed in the deeply anaesthetised supine patient and review of the distribution of complications clearly indicates that airway obstruction is the prominent reason for the higher score in this group. The mean scores were generally low, as the highest scoring potential complications (laryngospasm and desaturation) were rare. This may have been due to the strict control of depth of anaesthesia for airway manipulation or patient movement and appropriate management of any airway obstruction before desaturation could occur. The lowest score was seen in the lateral awake group, reflecting the very low incidence of airway obstruction in these patients.

We acknowledge that there are limitations in our study design. The sample size was calculated from the findings of an observational audit, which acted as a pilot study in the absence of any directly relevant published data. However, our sample sizes were comparable to those used in many of the previous studies of a similar nature. The anaesthetic technique was not fully standardised, which could potentially result in bias, although we did not see a wide variation in practice in our groups. Preexisting respiratory conditions such as upper respiratory tract infection, runny nose or snoring were not considered as exclusion criteria because we wished to allow our study sample to reflect our routine patient population as closely as possible. We are aware that in awake removal of the LMA, any complications are likely to occur immediately, whereas after deep removal, airway complications may be delayed. We planned 15 min of observation following removal of the LMA assuming that this time period was enough for daycase patients to awaken from sevoflurane anaesthesia. In fact, all patients in the study had emerged from anaesthesia and regained airway reflexes by the end of the study observation period. A potential source of bias, however, is that any immediate complications occurring after LMA removal in anaesthetised patients were recorded by the anaesthetist, whereas all other observations were made by nurses in the PACU. Although we believe that it is important to distinguish between the nature of airway complications and their associated risk during studies of this type, we recognise the limitations of our unvalidated, locally developed ‘Clinical Importance’ score and acknowledge that further work is required in order to develop a more rigorous and validated scoring system.

Although our study was a single-centre randomised trial conducted at a tertiary children's hospital, we believe it has external validity, as we studied a sample of children who may be found in any population of children undergoing daycase surgery and chose a study design that allowed our findings to be transferable to other settings.

In conclusion, the present study provides some evidence of an advantage in turning anaesthetised children into the lateral position before removal of the LMA, either at a deep plane of anaesthesia or awake. If there are reasons to remove the LMA while the patient is supine and deeply anaesthetised, airway obstruction must be anticipated and managed as appropriate.

Acknowledgements relating to this article

Assistance with the study: we would like to thank Ann Hageman and Tracy Elliot for their assistance with data collection and analysis.

Financial support and sponsorship: the study was funded by the Children's Charity, Sheffield Children's NHS Foundation Trust: Study number CA08014.

Conflicts of interest: none.

Presentation: this study was presented in part at the Annual Scientific Meetings of the Association of Paediatric Anaesthetists of Great Britain and Ireland, 19 to 20 May 2011, Torquay, the British Association of Day Surgery, 16 to 17 June 2011, Leeds and the European Society for Paediatric Anaesthesiology, 22 to 24 September 2011, Palma de Mallorca.


1. van Zundert TCRV, Brimacombe JR, Ferson DZ, et al. Archie Brain: celebrating 30 years of development in laryngeal mask airways. Anaesthesia 2012; 67:1375–1385.
2. Mason DG, Bingham RM. The laryngeal mask airway in children. Anaesthesia 1990; 45:760–763.
3. Johnston DF, Wrigley SR, Robb PJ, Jones HE. The laryngeal mask airway in paediatric anaesthesia. Anaesthesia 1990; 45:924–927.
4. McGinn G, Haynes SR, Morton NS. An evaluation of the laryngeal mask airway during routine paediatric anaesthesia. Pediatr Anesth 1993; 3:23–28.
5. Lopez-Gil M, Brimacombe J, Alvarez M. Safety and efficacy of the laryngeal mask airway. A prospective survey of 1400 children. Anaesthesia 1996; 51:969–972.
6. Parry M, Glaisyer HR, Bailey PM. Removal of LMA in children. Br J Anaesth 1997; 78:3373–3444.
7. White MC, Cook TM, Stoddart PA. A critique of elective pediatric supraglottic airway devices. Pediatr Anesth 2009; 19 (Suppl 1):55–65.
8. Patel B, Bingham R. Laryngeal mask airway and other supraglottic airway devices in paediatric practice. CEACCP 2009; 9:6–9.
9. Brimacombe JR, Brain AIJ, Berry AM. The Iaryngeal mask airway instruction manual. 4th ed.1999; Maidenhead, UK: Intavent-Orthofix, pp. 25–26.
10. Baird MB, Mayor AH, Goodwin APL. Removal of the laryngeal mask airway: factors affecting the incidence of postoperative adverse respiratory events in 300 patients. Eur J Anaesthesiol 1999; 16:251–256.
11. Laffon M, Plaud B, Dubousset AM, et al. Removal of laryngeal mask airway: airway complications in children, anaesthetized versus awake. Pediatr Anesth 1994; 4:35–37.
12. Gataure PS, Latto IP, Rust S. Complications associated with removal of the laryngeal mask airway: a comparison of removal in deeply anaesthetised versus awake patients. Can J Anaesth 1995; 42:1113–1116.
13. Kitching AJ, Walpole AR, Blogg CE. Removal of the laryngeal mask airway in children, anaesthetized compared with awake. Br J Anaesth 1996; 76:874–876.
14. Splinter WM, Reid CW. Removal of the laryngeal mask airway in children: deep anaesthesia versus awake. J Clin Anesth 1997; 9:4–7.
15. Samarkandi AH. Awake removal of the laryngeal mask airway is safe in paediatric patients. Can J Anaesth 1998; 45:150–152.
16. Pappas AL, Sukhani R, Lurie J, et al. Severity of airway hyperreactivity associated with laryngeal mask airway removal: correlation with volatile anaesthetic choice and depth of anaesthesia. J Clin Anesth 2001; 13:498–503.
17. Dolling S, Anders NRK, Rolfe SE. A comparison of deep vs. awake removal of the laryngeal mask airway in paediatric dental daycase surgery. A randomised controlled trial. Anaesthesia 2003; 58:1220–1234.
18. Sinha A, Sood J. Safe removal of LMA in children – at what BIS? Pediatr Anesth 2006; 16:1144–1147.
19. Mizushima A, Wardall GJ, Simpson DL. The laryngeal mask airway in infants. Anaesthesia 1992; 47:849–851.
20. Harnett M, Kinirons B, Heffernan A, et al. Airway complications in infants: comparison of laryngeal mask airway and the facemask-oral airway. Can J Anaesth 2000; 47:315–318.
21. López-Gil M, Mantilla I, Blanco T, et al. The Size 1 Proseal laryngeal mask airway in infants: a randomized, noncrossover study with the Classic laryngeal mask airway. Pediatr Anesth 2012; 22:365–370.
22. Lee JR, Kim SD, Kim CS, et al. Minimum alveolar concentration of sevoflurane for laryngeal mask airway removal in anaesthetized children. Anesth Analg 2007; 104:528–531.
23. O’Neill B, Templeton JJ, Caramico L, Schreiner MS. The laryngeal mask airway in paediatric practice: factors affecting ease of use during insertion and emergence. Anesth Analg 1994; 98:659–662.
© 2015 European Society of Anaesthesiology