A comparison of the i-gel™ with the LMA-Unique™ in non-paralysed anaesthetised adult patients


Correspondence to: Dr B. Bein
E-mail: bein@anaesthesie.uni-kiel.de


This study assessed two disposable devices; the newly developed supraglottic airway device i-gel™ and the LMA-Unique™ in routine clinical practice. Eighty patients (ASA 1–3) undergoing minor routine gynaecologic surgery were randomly allocated to have an i-gel (n = 40) or LMA-Unique (n = 40) inserted. Oxygen saturation, end-tidal carbon dioxide, tidal volume and peak airway pressure were recorded, as well as time of insertion, airway leak pressure, postoperative sore-throat, dysphonia and dysphagia for each device. Time of insertion was comparable with the i-gel and LMA-Unique. There was no failure in the i-gel group and one failure in the LMA-Unique group. Ventilation and oxygenation were similar between devices. Mean airway pressure was comparable with both devices, whereas airway leak pressure was significantly higher (p < 0.0001) in the i-gel group (mean 29 cmH2O, range 24–40) compared with the LMA-Unique group (mean 18 cmH2O, range 6–30). Fibreoptic score of the position of the devices was significantly better in the i-gel group. Post-operative sore-throat and dysphagia were comparable with both devices. Both devices appeared to be simple alternatives to secure the airway. Significantly higher airway leak pressure suggests that the i-gel may be advantageous in this respect.

Reusable supraglottic airway devices are established in clinical anaesthesia and have been shown to be safe and efficient [1–3]. However, especially for emergency airway management, there has been increasing demand for disposable devices. The LMA-Unique™ (LMA-U, Intavent Orthofix, Maidenhead, Berkshire, UK) was introduced in 1997 as a disposable device at approximately one-third of the cost of an LMA-Classic™ (cLMA, Intavent Orthofix, Maidenhead, Berkshire, UK). The cuff, back plate, airway tube and pilot balloon of the LMA-U are manufactured from clear medical grade polyvinylchloride and more curved than the softer silicone airway tube of the cLMA. Brimacombe et al. in a study of 60 adult patients concluded that disposable and reusable laryngeal mask airways perform similarly in paralysed patients [4]. Recently, a new single-use supraglottic airway device was introduced into clinical practice. The i-gel™ (Intersurgical LTD, Wokingham, Berkshire, UK) is made up of a thermoplastic elastomer [styrene ethylene butadiene styrene (SEBS)]. One main component of this new device is a non inflatable gel-like cuff that is claimed to achieve a perfect fit to the pharyngeal and laryngeal structure and to enable rapid, easy, safe and reliable application [5]. The device incorporates a gastric channel that allows gastric tube drainage [5]. In the present randomised study we tested the hypothesis that ease of insertion, airway leak pressure and fibreoptic position differ between the i-gel and the LMA-U in non paralysed, anaesthetised patients. For the i-gel, we also assessed ease of gastric tube placement.


After approval of the institutional review board and written informed consent, eighty patients with American Society of Anesthesiologists physical status 1–3 undergoing general anaesthesia for routine minor gynaecological surgery were enrolled and randomly allocated (by opening of a sealed envelope) to the i-gel (n = 40) or LMA-U (n = 40) group, respectively. Patients were excluded from the study if they were less than 18 years of age, had a known or predicted difficult airway, mouth opening less than 2.5 cm, a body mass index more than 35 kg.m−2, or if they were at risk of aspiration. All patients were premedicated with midazolam 0.1 mg.kg−1 orally 30 min before induction of anaesthesia. Routine monitoring was established including non-invasive blood pressure (NIBP), peripheral oxygen saturation (Spo2) and heart rate (HR; S/5, Datex-Ohmeda, Helsinki, Finland). Pre-oxygenation with 100% oxygen via face mask for 3 min was followed by standardised induction of anaesthesia with 2 mg.kg−1 propofol and 1 μg.kg−1 remifentanil, and anaesthesia was subsequently maintained with propofol (4–8 mg.kg−1.h−1) and remifentanil (0.2–0.5 μg.kg−1.h−1). Depth of anaesthesia was controlled with bispectral index (BIS XP™; Aspect Medical Systems, Natick, MA, USA). Successful ventilation was assured with capnography and bilateral chest auscultation. Following device placement, patients were ventilated with a tidal volume (VT) of 7 ml.kg−1, at a respiratory rate of 10 breaths min−1. Respiratory variables were recorded directly after device insertion and 10 min thereafter, respectively. Haemodynamic and BIS values were recorded before and after device insertion, followed by measurements every 5 min throughout the study period. Adverse events during anaesthesia were recorded and defined as aspiration/regurgitation, hypoxia (Spo2 < 90%), bronchospasm, airway obstruction and dental trauma.

After loss of eyelash reflex, the patient’s head was positioned in a neutral position and the airway device inserted according to the manufacturer’s instructions. A size 4 (i-gel and LMA-U) was used for all patients, since it has been shown previously, that a gender based approach to laryngeal mask size selection is appropriate [6, 7]. Also, weight of the patients enrolled was well within the range recommended by both manufacturers for the use of size 4 devices. All LMA-U devices were tested for leaks before insertion. Both devices were lubricated with water-soluble lubricant. The cuff of the LMA-U was initially inflated with 20 ml of air. Resulting cuff pressures were recorded in all patients with an LMA-U in situ (high fidelity pressure gauge; Smiths Medical, Kirchseeon, Germany; pressure range ± 500 mmHg, accurate to ± 0.5%). Correct insertion was assessed by proper chest expansion, the presence of end-tidal CO2, absence of audible leak and lack of gastric inflation. Time to successful insertion was measured from ‘touching’ the device after loss of eyelash reflex until the first expiratory VT > 200 ml. All devices were inserted by a single anaesthesiologist (H.F.) with experience using each of the study devices (> 50 uses with the i-gel, and > 1500 insertions with the LMA-U).

After successful placement of the i-gel, a 12 FG gastric catheter was advanced in the oesophagus via the gastric channel, followed by aspiration of gastric fluid. A pH≤2.5 was judged as indicative of gastric fluid (pH test paper; Bayer Vital, Leverkusen, Germany).

After taking baseline values, cuff volume in the LMA-U device was increased twice by another 10 ml (total cuff volume: LMA-U 40 ml). After recording of the resulting cuff and airway leak pressures, cuff volume was finally reduced to baseline values.

Airway leak pressure was determined by adjusting the expiratory valve of the breathing circle to 40 cmH2O (fixed fresh gas flow 3 l.min−1) and recording the pressure when equilibrium was reached [8]. By listening with a stethoscope over the epigastrium, air entrainment in the stomach was detected during measurement of oropharyngeal leak pressure.

Fibreoptic position of the mask bowl (i-gel/LMA-U) was determined by passing a fibreoptic endoscope through the airway tube to a position 1 cm proximal to the end of the tube. The resulting view on the glottic aperture (i-gel/LMA-U) was scored using an established scoring system: score 4, only vocal cords; score 3, vocal cords plus posterior epiglottis; score 2, vocal cords plus anterior epiglottis; score 1, vocal cords not seen, but function adequate [9]. If ventilation was impossible after insertion of the device, the position of the device was adjusted by gently pushing or pulling it, and adequacy of ventilation was reassessed. If more than three attempts were necessary for placement of the device or more than 3 min were needed, the attempt was terminated and attributed as failed, and the airway managed as clinically indicated. Upon completion of the study protocol, the anaesthesiologist gave a subjective assessment of the handling of either device, which was rated as excellent, good, fair or poor. Furthermore, 18–24 h after surgery, all patients underwent a structured interview by a medical student blinded for the airway device used [10]. The interview comprised the degree of sore throat, hoarseness, and dysphagia. Symptoms were graded as nil, moderate and severe. Patients were unaware of the airway device used in their individual case.

Variables studied were overall success rate for ventilation, times for airway insertion, cuff pressures and resulting leak pressure, expiratory tidal volumes, fibre optic view, subjective assessment of the handling of either device, and the frequency of postoperative sore throat, dysphagia and hoarseness.

Statistical analysis and sample size calculation

Sample size calculation was based on two previous studies on the LMA-U and i-gel [11, 12]; accordingly, we calculated the sample size to detect at least the difference between both devices which was described previously for the primary endpoint (airway leak pressure) with an α-error of 0.05 and a power of 0.9. For a difference of 6 cmH2O and a standard deviation of 8 cmH2O we calculated 40 patients per group. Non-parametric data between groups were analysed with the Mann–Whitney U-Test, while parametric data were compared with unpaired student’s t-test. Proportions and anaesthesiologists’ experiences were compared with Fisher’s exact test or χ2-test, as appropriate. Changes of parameters within each group over time were analysed with one way repeated analysis of variance with Bonferroni correction for multiple comparisons and Friedman’s test with Dunn’s post test correction, as appropriate. Parametric data are given as mean (SD), and nonparametric data are presented as median, interquartile range, and range. p < 0.05 was considered significant.


There were no significant differences between groups with regard to demographic data and number of predictors of a difficult airway (Table 1).

Table 1.  Demographic data and predictors of a difficult airway. Data are given as mean (SD) or as absolute numbers.
Groupi-gel (n = 40)LMA-U (n = 40)p-value
Height; cm167 (6)166 (7)0.25
Weight; kg 69 (12) 68 (10)0.12
Age; years 57 (12) 55 (10)0.42
ASA I/II/III 12/27/1 10/29/10.25
Duration of surgery; min 79 (15) 75 (16)0.41
Mouth opening; mm 49 (8) 50 (9)0.29
Thyromental distance; mm 69 (15) 70 (14)0.3
Mallampati Class I/II/III 14/19/7  8/20/120.22
Mobility of atlantooccipital junction
 <15°  2  31.0
 >15° 38 371.0

There were no significant differences in cardio respiratory variables, tidal volume (Vtex) and BIS values which all remained stable throughout the study period (Table 2), except for a higher airway pressure (Paw) in the i-gel group (Table 2).

Table 2.  Bispectral index (BIS), heart rate (HR), mean arterial pressure (MAP), oxygen saturation (Spo2), expiratory tidal volume (Vtex) and airway pressure (Paw) before and after insertion and 10 min after placement, respectively. Data are mean (SD).
BeforeAfter10 minBeforeAfter10 min
  1. *p < 0.05 vs i-gel before, †p < 0.05 vs i-gel after, ‡p < 0.05 vs LMA-U before, §p < 0.05 vs LMA-U after.

BIS; units96 (3)33 (10)*44 (10)*†96 (3)35 (9)‡39 (9)‡§
HR; beats.min−172 (13)58 (9)*55 (8)*†71 (14)61 (11)‡55 (9)‡§
MAP; mmHg102 (14)68 (13)*†70 (13)*†105 (13)70 (13)‡81 (10)‡§
Spo2; %97 (2)98 (1)99 (1)97 (2)99 (1)99 (1)
V tex; ml 452 (60)459 (65) 442 (52)445 (57)
P aw; cmH2O 14 (3)14 (3) 13 (3)12 (2)

In the i-gel group, a successful primary airway was established in 36 patients (90%) on the first attempt and in four patients (10%) on the second attempt. In the LMA-U group, device insertion was successful in 34 patients (85%) on the first attempt and in five patients (13%) on the second attempt. In one patient (2%) in the LMA-U group, device insertion failed even after three attempts and the patient was managed by tracheal intubation. Time required for the first adequate ventilation was comparable between groups (i-gel: median 15 s; range 10–60 s; LMA-U: 17 s; 11–180 s) (p = 0.45). Subjective assessment of handling was comparable with the i-gel and the LMA-U (Table 3). In no patients did arterial saturation decline to less than 90% during insertion. The airway leak pressure was significantly higher in i-gel group (mean 29 (5) cmH2O) compared with the LMA-U group [18 (5) cmH2O (cuff 20 ml); 20 (5) cmH2O (cuff 30 ml) and 22 (5) cmH2O (cuff 40 ml)] (p < 0.0001) (Table 4).

Table 3.  Subjective assessment regarding the overall performance. Data is given as absolute numbers.
 i-gel (n = 40)LMA-U (n = 40)
Table 4.  In vivo cuff pressures and airway leak pressure. Data are given as median [25/75th percentile] and (range).
DeviceCuff pressure;
Airway leak pressure; cmH2O
  1. *p < 0.0001 vs 30 and 40 ml, respectively, †p < 0.0001 vs i-gel, ‡p < 0.0001 vs 40 ml.

I-Gel 29 [26–30] (20–40)
LMA-U 20 ml 70 [50–100] (30–120)*18 [15–22] (6–29)†‡
LMA-U 30 ml120 [120–130] (100–120) 20 [17–25] (8–31)†
LMA-U 40 ml140 [140–140] (120–140)22 [18–26] (10–33)†

Fibreoptic control of the position of the devices was significantly better in the i-gel group (Table 5) (Figs 1 and 2). No gastric inflation occurred with the i-gel device, whereas gastric inflation was observed in the LMA-U group in three patients (7.5%). Gastric catheter placement was successful in all patients in the i-gel group. Insertion of a gastric tube was very easy (first attempt) in 36 cases (90%) and easy in four cases (second attempt) (10%). In all patients, fluid aspirated showed a pH≤2.5 and therefore the gastric tube was assumed to be in the correct place.

Table 5.  Fibreoptic score of the position of both devices.
Fibreoptic scorei-gel (n = 40)LMA-U (n = 40)
4 = only vocal cords visible2013
3 = vocal cords plus posterior epiglottis visible159
2 = vocal cords plus anterior epiglottis visible313
1 = vocal cords not visible but function adequate24
0 = vocal cords not visible and function inadequate00
Figure 1.

 Fibreoptic score of the position of the i-gel (left) and the LMA-U (right). 4 = only vocal cords visible, 3 = vocal cords plus posterior epiglottis visible, 2 = vocal cords plus anterior epiglottis visible, 1 = vocal cords not visible but function adequate, 0 = vocal cords not visible and function inadequate. *p < 0.005 vs LMA-U.

Figure 2.

 Fibreoptic views: typical positions with the i-gel and with the LMA-U. (a) i-gel score 4; (b) LMA-U score 4; (c) i-gel score 3; (d) LMA-U score 3; (e) i-gel score 2; (f) LMA-U score 2; (g) i-gel score 1; (h) LMA-U score 1.

Blood staining after removal of the devices occurred rarely (5%) and was comparable between groups. There were no differences regarding postoperative airway morbidity (Table 6). No major adverse events occurred during the intra- and immediate postoperative period in any patient in both groups.

Table 6.  Subjective assessment of sore throat and dysphagia. Data are given as absolute numbers. No difference between groups.
Devicei-gel (n = 40)LMA-U (n = 40)
Sore throatDysphagiaDysphoniaSore throatDysphagiaDysphonia


The main findings of our prospective, randomised study are as follows: insertion success, time required for adequate ventilation and subjective rating of the handling of the I-Gel and the LMA-U were comparable. However, airway leak pressure was significantly higher in the i-gel group and the fibreoptic score was also significantly better in the i-gel group.

The practicability and handling of a new airway device is of importance, because especially in difficult situations such as the ‘can't ventilate, can't intubate’ scenario, an easy-to-use supraglottic airway device may help to improve the patient’s outcome considerably. For the inexperienced and untrained user, the complex act of inserting a supraglottic airway device is a significant disadvantage, especially in emergency situations [13–15].

Regarding the time of insertion for the i-gel, the results of our study are in agreement with the data of Wharton et al. In their study, time to successful insertion and success rate (first attempt 83%, overall 95%) were comparable to our results [12]. Gatward and Richez also showed comparable data for placement of the i-gel [16]. The time needed for the correct placement of either the i-gel and LMA-U is acceptable and comparable with previous reports [11] and recommended ventilation aids such as the face mask [17] or the intubating laryngeal mask [18].

The high (in vivo) cuff pressures in the LMA-U group found in the present investigation are in agreement with the mucosal pressures found in the cadaver study of Keller et al., which were obtained with a sophisticated, high fidelity recording technology. Filling the cuff of the LMA-U with a predefined cuff volume (20, 30 and 40 ml, respectively) resulted in cuff pressures largely above recommended values. This may be of clinical significance, since in daily clinical practice the recommended intracuff pressure is often ignored and cuff over inflation unnoticed, especially during short procedures with spontaneously breathing patients [19]. Consequently, the high cuff pressures in the LMA-U group may be harmful for the oropharyngeal mucosa when applied for a longer period of time. Nandwani et al. have shown that increasing the cuff volume of the laryngeal mask airway displaces the larynx anteriorly. Higher cuff volumes may be observed especially during nitrous oxide administration [20]. The wedge-shaped tip is then displaced from the wedge-shaped hypopharynx causing proximal displacement of the cuff, movement of the epiglottis into the bowl and exposition of the oesophageal inlet. A malposition of a supraglottic airway device increases the risk of leakage. If the leakage is sufficiently large, a ballooning of the stomach may occur. This can lead to a deterioration of respiratory mechanics and increase the likelihood of regurgitation and thus the risk of aspiration. In contrast, the non-inflatable cuff of the i-gel is semi-rigid and cannot be folded over, or over inflated, thus diminishing the risk of both airway obstruction and mucosal damage.

The fibreoptic view depends on the hypopharyngeal position of the supraglottic airway device and whether the epiglottis is folded down during insertion or not. It has been shown previously that the best fibreoptic view is accomplished if the LMA cuff is inflated to 20 ml [21]. Fibreoptic examination via the i-gel yielded an acceptable view of the vocal cords (view 3 and 4) in 87% of the patients. This is comparable with the data from Levitan et al. obtained in fresh cadavers [5]. In contrast, an acceptable fibreoptic view of the vocal cords in the LMA-U group was recorded only in 55 % of the patients. Most studies, however, have shown little or no correlation between fibreoptic position and function of the supraglottic airway device [1, 9], but the fibreoptic score confirms that the i-gel occupies an excellent anatomic location to ensure either unimpeded ventilation or passage of a tracheal tube [5, 22].

The effective airway leak pressure is important, especially in patients with increased respiratory resistance, in patients with chronic obstructive pulmonary disease, and during cardiopulmonary resuscitation, since a higher airway leak pressure increases the likelihood that a preset tidal volume can be applied. Further, studies on the most common alternative supraglottic airway devices suggest that a mean Paw of more than 20 cm H2O increases the risk of leakage with resultant insufficient ventilation and increased risk of aspiration [18, 23–25].

Our research on airway leak pressure using the i-gel yielded a mean airway leak pressure of 29 ± 5 cmH2O without gastric inflation [26]. This is higher than values reported for the cLMA [1] or LMA-U [11] and similar to the LMA ProSeal™ (PLMA, Intavent Orthofix, Maidenhead, Berkshire, UK) [27]. Consequently, in our study the i-gel provided a better seal than the LMA-U. In contrast, Schmidbauer et al. recently concluded that both the PLMA and the cLMA provided a better seal of the oesophagus in contrast to the i-gel [28]. However, this study was performed on fresh cadavers.

If a supraglottic airway device lacks the facility for active (suction) and passive (drainage) emptying of the stomach, this may put the patient at an increased risk of regurgitation, thereby precluding application of such a device in patients with full stomach [3, 29]. Gibbison et al. described, that in two cases (from 280) the i-gel completely protected the airway from risk of aspiration [30]. The drain tube is designed to reduce gastric inflation and the risk of regurgitation, but its efficiency is unproven. In our study, placement of a gastric tube was very easy or easy in all patients in the i-gel group.

Post-anaesthesia airway morbidity has gained widespread attention, especially in a health environment where cost containment is essential and patient satisfaction is of high priority [31]. The causes of postoperative adverse events such as sore throat, dysphonia and dysphagia after general anaesthesia using supraglottic airway devices are dependent on the depth of anaesthesia, the method of insertion, the number of insertion attempts [32], the mode of ventilation used [33], the time of anaesthesia [32] and on the type of postoperative analgesia provided [10]. In our study, we intended to control some of the above mentioned variables by limiting the attempts of insertion to three, using the same intracuff pressure for the LMA-U and measuring the depth of anaesthesia. The perioperative adverse events rate was infrequent for both devices. The low morbidity rate in both groups in our study is noticeable and could have been due to the high rate of first attempt success. The incidence of blood on the surface of both the i-gel and the LMA-U after removal is in agreement with reports from previous studies [34].

Some limitations of our study should be noted. The present study was performed in patients with normal airways during controlled positive pressure ventilation by investigators trained in the use of each device. Therefore, we cannot comment on results obtained in spontaneously breathing patients, during difficult airway management or with naïve users, though we speculate that the results found would be similar in these scenarios.

The better seal with the i-gel device suggests an alternative to the LMA-U for positive pressure ventilation, especially in patients where higher airway pressures are necessary to achieve a sufficient tidal volume.


We are grateful to LMA Company, Bonn, Germany, and to the Intersurgical Company, St Augustin, Germany for supplying the devices used in this investigation.