Secondary Logo

Journal Logo

Airway management

Sore throat following three adult supraglottic airway devices

A randomised controlled trial

L’Hermite, Joël; Dubout, Elisabeth; Bouvet, Sophie; Bracoud, Laure-Hélène; Cuvillon, Philippe; Coussaye, Jean-Emmanuel de La; Ripart, Jacques

Author Information
European Journal of Anaesthesiology: July 2017 - Volume 34 - Issue 7 - p 417-424
doi: 10.1097/EJA.0000000000000539
  • Free


This article is accompanied by the following Invited Commentary:

Hinkelbein J, Greif R, Diemunsch P, Kranke P. Publication and innovation in airway management: quality not quantity! Eur J Anaesthesiol 2017; 34:408–410.


Sore throat is a common complaint after surgery.1,2 It affects patient satisfaction and can affect activity after discharge. Postoperative sore throat could be the result of direct trauma by rigid materials inserted into the upper airway, physical tensile stresses imposed by laryngoscopy or distending forces generated by the pressure of the cuffs or cushions.3,4 During general anaesthesia, the airway is usually managed with tracheal intubation. However, airway injuries and complications of airway management with tracheal intubation are frequent.5 Supraglottic airway devices (SADs) offer an alternative to traditional tracheal intubation with potential benefit for ease of fit, disturbance of the airway6 and in particular sore throat.7

The laryngeal mask airway (LMA) was historically the first SAD (cuffed, orally inserted hypopharyngeal airway) and was introduced into clinical practice in the 1980s.8,9 The incidence of sore throat after the LMA ranges from 9 to 17.5%.10,11 Two new SADs were developed in 2007: the LMA Supreme (LMA-S; Intavent Orthofix, Maidenhead, UK) and the I-gel (Intersurgical Ltd., Wokingham, UK).12,13 The LMA-S is a single-use, latex-free inflatable LMA with gastric access.12 The I-gel is a single-use supraglottic device featuring an additional tube for fitting a gastric suction catheter, but no inflatable cuff as its constituent thermoplastic elastomer provides an airway seal.13 To date, no randomised clinical trial comparing their sore throat incidence with the LMA Unique (LMA-U) has been reported.

The aim of this single-centre randomised controlled trial was to compare the incidence of sore throat with the LMA-U, LMA-S and the I-gel. Secondary endpoints were clinical performance and adverse events.

Materials and methods

Trial design and participants

The study took place at the Nimes University Hospital, Caremeau Centre in Nîmes, France, between April 2009 and September 2012. This was a prospective randomised single-blind controlled parallel three-group trial. Ethical approval was provided by our Institutional Ethics Committee (Comité de Protection des Personnes Sud Méditerranée III, Faculté de Médecine Nîmes, France on 24 July 2008 (Ref: 2008/07/08). According to French law (code de Santé Publique: Art L.1121–1–2° and R.1121-3) and our Institutional Ethics Committee norms, signed informed consent was not required for a ‘routine practice’ study. However, during preoperative evaluation, written information was provided and patients were advised that they could refuse to participate.

Eligible participants were adult patients more than 18 years scheduled to undergo elective surgery lasting less than 2 h under general anaesthesia. We excluded those who refused to participate, were pregnant or nursing, were undergoing emergency surgery, had had a sore throat within the 30 days preceding surgery, had an increased risk of aspiration of gastric contents (diabetes mellitus, gastro-oesophageal reflux, BMI > 35) and those who had expected airway difficulties requiring tracheal intubation.

Patient randomisation

All patients were screened during preoperative evaluation (≥48 h before surgery) and written information was provided. The day before surgery (or the same day in case of outpatient surgery) informed consent was obtained. On the day of surgery, patients were allocated into three parallel groups to receive LMA-U (LMA-U group), LMA-S (LMA-S group) or I-gel (I-gel group) using an online randomisation application designed for clinical research projects. Randomisation was performed in balanced blocks with a 1 : 1 : 1 allocation ratio. The randomisation list was computer generated using SAS Software version 9.4 (SAS Institute Inc. Cary, North Carolina, USA) by the methodologist from Biostatistics Department of Nimes University Hospital appointed to the study and not involved in treatment allocation.

Anaesthesia and supraglottic procedures

After arrival in the operating theatre, an intravenous cannula was inserted and standard monitoring (noninvasive assessment of blood pressure, pulse oximetry, electrocardiography) was attached. After preoxygenation (3 min by bag and mask with 100% oxygen), patients were induced with intravenous sufentanil (0.3 μg kg−1), lidocaine (30 mg) and propofol (2.5 to 3 mg kg−1), in the supine position with the head on a standard pillow, 7 cm in height. The lungs were manually inflated via a facemask using 100% oxygen. A Guedel airway (Intersurgical Ltd, Wokingham, UK) was used as advised by the attending anaesthesiologist. If required, additional boluses of 0.5 mg kg−1 intravenous propofol were given until an adequate level of anaesthesia was achieved for placement. The head position was left to the discretion of the attending anaesthesiologist. The SADs were lubricated with a water-soluble agent (KY jelly, Johnson & Johnson Ltd, New Brunswick, New Jersey, USA), and inserted at least 1 min after completion of induction, but only when the jaw was relaxed and there was no lash reflex. According to the manufacturer's instructions, the LMA-U and LMA-S were inserted with the cuff fully deflated. The manufacturer's weight-based recommendations were used for size selection (LMA-U and LMA-S: size 3, >30 to 50 kg; size 4, >50 to 70 kg; size 5, >70 kg; I-gel: size 3, 30 to 60 kg; size 4, 50 to 90 kg; size 5, >90 kg). After the SAD was inserted into the pharynx, the cuff (LMA-U and LMA-S) was inflated with air until the maximum recommended inflation volume was reached. The intra-cuff pressure was immediately set at 60 cm H2O using a handheld pressure gauge (Endotest; Rüsch, Kernen, Germany). Devices were fixed according to the manufacturers’ instructions. Two attempts were allowed before insertion was considered a failure. An insertion attempt was defined as placement of the SAD in the mouth. A failed attempt was defined as removal of the SAD from the mouth. An effective airway was defined as three square-wave capnograph traces during manual ventilation. If an effective airway could not be achieved, a different size was used. Two attempts were allowed with the new size before insertion was considered a failure. In the case of a second failure despite the change of size, tracheal intubation was performed. The time between picking up the SAD and obtaining an effective airway (defined as three square-wave capnograph traces during manual ventilation) was recorded. Oropharyngeal leak pressure was determined automatically by the respiratory system (Taema, Air Liquide Medical System, Antony, France). Pressure-controlled ventilation was initiated with a mixture of 50% oxygen and 50% air, respiratory rate 12 cycles min−1, zero end-expiratory pressure and an inspiratory to expiratory ratio (I : E) 1 : 2 and adjusted to maintain the end-tidal CO2 at around 35 mmHg. An appropriately sized orogastric tube (14-F for LMA-S size 3 and 4, 16-F for LMA-S size 5; 12-F for I-gel size 3 and 4, 14-F for I-gel size 5) was inserted through the supraglottic device. Orogastric tube insertion was subjectively scored as very easy, easy, difficult, very difficult or impossible. Correct placement of the orogastric tube was confirmed by an anaesthesiologist upon either aspiration of gastric fluid or detection of injected air by auscultation over the epigastrium.

Anaesthesia was maintained with either a target-controlled infusion of propofol or 1.5 to 2% sevoflurane in a mixture of 50% oxygen and 50% air. Analgesia was achieved with 5 to 10-μg sufentanil boluses, as required. Also, approximately 30 to 60 min before the end of surgery, in the absence of any contraindication, 100-mg ketoprofen and 1-g paracetamol were administered intravenously for postoperative analgesia. At the end of surgery, anaesthesia was discontinued, and the SAD was removed when the patient was able to open his or her mouth to command. The SAD cuff was deflated as the SAD was removed.

Measured variables and time of measurement

Collected data included patient characteristics (weight, height, calculated BMI, age, sex, American Society of Anesthesiologists score), airway assessment (Mallampati class, mouth opening, dentition) and anaesthesia details (time, maintenance). Effective airway time (time between picking up the EAD and obtaining an effective airway, defined as three square-wave capnograph traces during manual ventilation) was noted. Ease of insertion and removal of the SAD were also noted. Ease of use of the SAD was graded using a subjectively five-point scoring system: very easy, easy, difficult, very difficult and impossible (insertion only). The numbers of attempts, the need to change size, the need for tracheal intubation were noted. For each patient, the following complications occurring during insertion, maintenance and removal were noted: aspiration or regurgitation; hypoxia (peripheral oxygen saturation measured by pulse oximetry <90%); bronchospasm; laryngospasm; airway obstruction; coughing or retching; hiccup; blood staining of the SAD; and tongue, lip or dental trauma. In the event of intraoperative failure of the SAD, the need to intubate was recorded.

In the postoperative period, patients underwent a structured interview 2 h (H+2) and 24 h (H+24) after removal of the device before leaving the recovery room. H+24 postoperative evaluations were held at the bedside (in-patient) or by phone call (day case surgery). Patients were asked about sore throat (constant pain, independent of swallowing), dysphonia (difficulty or pain on speaking), neck and/or jaw pain, nausea and/or vomiting. Dysphagia with liquids and/or solids was evaluated at H+2 and H+24 after removal of the device. Symptoms were graded by the patient as null, mild, moderate or severe.

Primary endpoint, data collection and blinded observers

The primary endpoint of the study was the rate of sore throat at 24 h (H+24). Sore throat with the LMA-U has been estimated at about 17%.8 When we performed the study in 2008, limited information was available concerning the rate of sore throat with LMA-S and the I-gel. In 2008, a pilot study evaluated the LMA-S in 22 patients but the authors did not report postoperative sore throat (H+2).12 The expected postoperative sore throat incidence for the LMA-S was less than 5%. A pilot observational study evaluated the I-gel in 71 women. Only one case of mild sore throat occurred (1.40%).13 The expected postoperative sore throat incidence for the I-gel was less than 5%. Sample size was based to test the hypothesis that there is a difference of at least 12% of postoperative sore throat between the LMA-U (17%) and both the I-gel (<5%) and the LMA-S (<5%). Therefore, to detect a 12% difference in the primary endpoint with a global bilateral risk α of 5 and 90% power, 546 patients (182 patients in each of three groups) were required to allow the two comparisons. Patients were unaware of the SAD used. Unblinded trained observers collected the data during anaesthesia and in the recovery room and a blinded trained observer collected the data after removal of the SAD.

Statistical analysis

When complete data were available for all randomised patients, an intention-to-treat analysis was performed. As some loss of data for evaluation of the primary outcome was likely, an incomplete intention-to-treat analysis was then conducted. No imputation for missing data was undertaken. Comparisons between groups were performed using χ2 tests or Fisher exact tests when appropriate and one-way ANOVA or Kruskal–Wallis analysis. All comparisons were two sided and a P value of less than 0.05 was required to exclude the null hypothesis. The Bonferroni correction for multiple comparisons was used for sore throat intensity analysis and effective airway time insertion. Studied data were expressed as mean ± SD, or median with interquartile or in frequencies and percentages when appropriate. Statistical analysis was conducted using SAS (release 9.4; SAS institute, Cary, North Carolina, USA).


Between April 2009 and September 2012, 546 patients were enrolled and assessed for eligibility. There were nine exclusions (Fig. 1) leaving 537 patients for intention-to-treat analysis. These were allocated as follows: 177 patients included in the LMA-U group, 180 in the LMA-S group and 180 in the I-gel group (Fig. 1). Patient characteristics were similar between groups (Table 1).

Fig. 1
Fig. 1:
CONSORT (Consolidated Standards of Reporting Trial) trial flow diagram.
Table 1
Table 1:
Baseline characteristics

Postoperative (H+24) sore throat:

  1. Primary endpoint: Sore throat at H+24 was assessed in 436 out of 537 patients because of missing data in 101 (18.8%), and was reported by 104 (23.9%). The incidence of postoperative (H+24) sore throat was not significantly different between groups (P = 0.34; Table 2).
  2. Postoperative (H+24) sore throat intensity (Table 2): Sore throat intensity (none, mild, moderate or severe) at H+24 was statistically different (P = 0.03) between groups. Pairwise comparisons were made regarding H+24 sore throat intensity. By applying the Bonferroni correction (significance P < 0.0167), the difference was significant between the LMA-S and I-gel group (P = 0.0111). Among patients with H+24 sore throat, intensity was higher in the LMA-S group than the I-gel group (25.8% moderate intensity vs. 5.4% and 6.5% severe intensity vs. 0).
Table 2
Table 2:
Postoperative sore throat: primary endpoint (24-h postoperative sore throat) and secondary endpoints (category of 24-h postoperative sore throat intensity)a

Secondary endpoints included:

  1. Morbidity: H+24 dysphagia with liquids was higher in the LMA-S group (P = 0.0065; Table 3). Otherwise, postoperative morbidities were not statistically different at the various times (Table 3).
  2. Clinical performance: Data regarding airway leak pressure, dynamic airway compliance, complications during maintenance and alternative airway management (tracheal intubation) in case of failure of the SAD, are summarised in Table 4. Airway leak pressure was significantly lower in the LMA-U group than in the LMA-S and I-gel groups (P < 0.0001). Complications occurring during insertion, maintenance and removal were not statistically different (Tables 4 and 5).
  3. Ease of use of the devices: Device insertion time, success on the first attempt, ease of insertion and removal, ease of insertion of orogastric tube (LMA-S and I-gel groups), complications during insertion and removal are summarised in Table 5. Effective airway time was significantly different between groups (P = 0.0376; Table 5). Pairwise comparisons were made for effective airway time insertion (LMA-U vs. LMA-S, P = 0.82; LMA-S vs. I-gel, P = 0.0187; LMA-U vs. I-gel, P = 0.0392) and by applying the Bonferroni correction (significance P < 0.0167), the difference was between I-gel group and the other two groups. Effective airway time was a little faster in I-gel group than in LMA-U and LMA-S groups. There were no significant differences in ease of insertion and removal, ease of insertion of orogastric tube (LMA-S and I-gel groups), complications during insertion and removal between groups (Table 5).
Table 3
Table 3:
Postoperative morbidity data
Table 4
Table 4:
Clinical performance of supraglottic airway device
Table 5
Table 5:
Airway management and complications during insertion and removal of the supraglottic airway device


In this prospective randomised controlled trial, we found that incidences of postoperative sore throat at H+24 were similar between the LMA-U, LMA-S and I-gel. However, in patients with sore throat, risk of moderate or severe sore throat intensity was higher in the LMA-S group compared with the LMA-U and I-gel groups. H+24 postoperative dysphagia with liquids was higher in LMA-S group. Other postoperative signs of poor tolerance of the SADs (neck and/or jaw pain, dysphonia, nausea and/or vomiting, dysphagia with solids) were similar between groups. Effectiveness of the SADs assessed by airway leak pressure was lower with the LMA-U compared with the LMA-S and I-gel. Finally, the ease of use assessed by device insertion time seemed better with I-gel.

Our study provides additional information to assist the practitioner in choosing the best SAD. Theoretically, the best SAD should cause little postoperative sore throat, and then with low intensity, and low morbidity; it should be easy to use, rapidly introduced at the first attempt, and should provide high airway leak pressure. We selected postoperative sore throat at H+24 as our primary endpoint because this was the outcome that would have the greatest impact on the patient. In the light of our results, even though the incidence of postoperative sore throat at H+24 was similar between groups, practitioners should be discouraged from using the LMA-S because of a greater intensity of sore throat and a greater incidence of dysphagia with liquids observed 24 h after mask removal. Finally, if we base the selection criteria on device placement time and airway leak pressure, because of its better airway leak pressure, the I-gel should be the LMA of choice among the three SADs studied.

Despite our working assumption, we failed to find any significant difference regarding the incidence of postoperative sore throat after using the LMA-U, I-gel or LMA-S. With the LMA, the overall incidence of postoperative sore throat can be as high as 42%,14,15 but the use of manometry can reduce pharyngolaryngeal complications by 70% compared with routine care without manometry.16 In a randomised controlled trial, Seet et al.16 showed that inflation of the LMA cuff with a pressure of less than 60 cm H2O was associated with a decreased incidence of sore throat at 24 h postoperatively (3.1 vs. 13.6%, P < 0.008) compared with routine care. In our study, although intra-cuff pressure was routinely checked (60 cm H2O), the incidence of sore throat at 24 h in LMA group was 22.2% (Table 2), similar to previous reports,11–17 but seven times higher than reported by Seet et al.16 This difference may be explained by a greater use of Guedel-type airway in our trial. Indeed, in our study a Guedel-type airway was used whenever advised by the attending anaesthesiologist, whereas Seet et al.16 used it only when bag-mask ventilation was difficult. Furthermore, time measurement of intra-cuff pressure was performed just after insertion of the device, whereas in the study by Seet et al.,16 intra-cuff pressure was measured once regular spontaneous breathing had been achieved. Our results might have been different if we had used spontaneous instead of pressure-controlled ventilation.

If we speculate that the incidence of sore throat is mainly related to the inflatable cuff, the expected result would have been less pain in the I-gel group. In a recent systematic review and meta-analysis of the I-gel vs. LMA in adults, the relative risk (95% confidence interval) of sore throat was in fact lower with the I-gel, 0.59 (0.38 to 0.90; P = 0.02).18 These results are derived from studies where monitoring of intra-cuff pressure was not common practice. In our study, we assume that the similarity in the incidence of postoperative sore throat between the I-gel group and others (LMA-U and LMA-S) is related to systematic monitoring of intra-cuff pressure set at 60 cm H2O using a handheld pressure gauge immediately after insertion of devices.

Regarding the intensity of sore throat, our results are in accordance with a recent meta-analysis in which the LMA-S was associated with a greater intensity of sore throat (RR 2.56, 95% confidence interval 1.60 to 4.12) than the I-gel group.19

The performance of the I-gel vs. the LMA-S and the I-gel vs. the LMA were recently subject to two meta-analyses.18,19 The faster insertion of the I-gel observed in our clinical trial agrees with the meta-analysis published by Montblanc et al.18 who reviewed randomised clinical trials of the I-gel vs. LMA. However, Montblanc et al.18 emphasised that the difference in placement time was not clinically relevant, as is the case for our clinical trial. Moreover, Chen et al.19 did not observe any significant differences in device placement time between the I-gel and the LMA-S in their meta-analysis. Our finding of better airway leak pressure of the I-gel (vs. the LMA-S and LMA) are in accordance with these two recent meta-analyses.18,19

Our study has several limitations. First, data on the primary endpoint were collected in 436 patients, with a high rate of patients lost to follow-up: 101/537 missing data (18.8%). The lack of statistical difference between groups could be the consequence of insufficient power. Second, we systematically tested ease of insertion of an orogastric tube in the I-gel and LMA-S groups. This could have potentially increased the incidence and/or severity of sore throat in these groups. In our clinical practice, when an I-gel or a LMA-S is used, the insertion of an orogastric tube is performed in particular to verify an empty stomach. This practice, although not mandatory, is used in other institutions. Therefore, to ensure that our results could be generalised to clinical practice, it seemed logical to do so. Finally, because we studied only adult patients, further studies might be useful to confirm our results in children.

We conclude that, in anaesthetised nonparalysed adults who underwent pressure-controlled ventilation with passage of an orogastric tube (through the LMA-S and I-gel) and when monitoring of intra-cuff pressure is routinely performed (LMA-U and LMA-S), no significant difference between LMA-U, LMA-S and I-gel was found in term of postoperative (H+24) sore throat. Our study emphasises that selection criteria of SADs should be based on clinical performance criteria (such as device insertion time and mainly airway leak pressure) and postoperative tolerance criteria (such as dysphagia with liquids).

Acknowledgements relating to this article

Assistance with the study: we would like to thank Dr Mariella Lomma for proofreading the manuscript.

Financial support and sponsorship: this work was supported by the Division of Anaesthesia Intensive Care, Pain and Emergency, University Hospital of Nîmes, Nîmes, France.

Conflicts of interest: none.

Presentation: Preliminary data for this study were presented as a poster presentation at the French Society of Anaesthesia and Intensive Care Annual Congress, 18 to 21 September 2013, Paris.


1. Riding JE. Minor complications of general anaesthesia. Br J Anaesth 1975; 47:91–101.
2. Loeser EA, Stanley TH, Jordan W, et al. Postoperative sore throat: influence of lubrication versus cuff design. Can Anaesth Soc J 1980; 27:156–158.
3. McHardy FE, Chung F. Postoperative sore throat: cause, prevention and treatment. Anaesthesia 1999; 54:444–453.
4. Chandler M. Tracheal intubation and sore throat: a mechanical explanation. Anaesthesia 2002; 57:155–161.
5. Domino KB, Posner KL, Caplan RA, et al. Airway injury during anesthesia: a closed claim analysis. Anesthesiology 1999; 91:1703–1711.
6. Nicholson A, Cook TM, Smith AF, et al. Supraglottic airway devices versus tracheal intubation for airway management during general anaesthesia in obese patients. Cochrane Database of Syst Rev 2013; 9:CD010105.
7. Tanaka A, Isono S, Ishikawa T, et al. Laryngeal resistance before and after minor surgery. Anesthesiology 2003; 99:252–258.
8. Brain AI. The laryngeal mask: a new concept in airway management. Br J Anaesth 1983; 55:801–805.
9. Brimacombe J. A proposed classification system for extraglottic airway devices. Anesthesiology 2004; 101:559.
10. Brimacombe J, Keller C. Bleeding, dysphagia, dysphonia, dysarthria, severe sore throat, and possible recurrent laryngeal, hypoglossal, and lingual nerve injury associated with routine laryngeal mask airway management: where is the vigilance? Anesthesiology 2004; 101:1242–1244.
11. Higgins PP, Chung F, Mezei G. Postoperative sore throat after ambulatory surgery. Br J Anaesth 2002; 88:582–584.
12. Van Zundert A, Brimacombe J. The LMA Supreme: a pilot study. Anaesthesia 2008; 63:209–210.
13. Richez B, Saltel L, Banchereau F, et al. A new single use supraglottic airway device with a noninflatable cuff and an esophageal vent: an observational study of the i-gel. Anesth Analg 2008; 106:1137–1139.
14. Bick E, Bailes I, Patel A, et al. Fewer sore throats and a better seal: why routine manometry for laryngeal mask airways must become the standard of care. Anaesthesia 2014; 69:1304–1308.
15. Brimacombe J, Holyoake L, Kellar C, et al. Pharyngolaryngeal, neck, and jaw discomfort after anesthesia with the face mask and laryngeal mask airway at high and low cuff volumes in males and females. Anesthesiology 2000; 93:26–31.
16. Seet E, Yousaf F, Gupta S, et al. Postoperative pharyngolaryngeal adverse events: a prospective randomized trial. Anesthesiology 2010; 112:652–657.
17. Grady DM, McHardy F, Wong J, et al. Pharyngolaryngeal morbidity with the laryngeal mask airway in spontaneously breathing patients. Anesthesiology 2001; 94:760–766.
18. De Montblanc J, Ruscio L, Mazoit JX, et al. A systematic review and meta-analysis of the i-gel vs laryngeal mask airway in adults. Anaesthesia 2014; 69:1151–1162.
19. Chen X, Jiao J, Cong X, et al. A comparison of the performance of the I-gel vs. the LMA-S during anesthesia: a meta-analysis of randomized controlled trials. PLoS ONE 2013; 8:e71910.
© 2017 European Society of Anaesthesiology