CLINICAL SCIENCE
Comparison of a Supraglottic Gel Device and an
Endotracheal Tube in Keratoplasty Performed Under
General Anesthesia: A Randomized Clinical Trial
Gilles Guerrier, MD,* David Boutboul, MD,* Sylvie Rondet, MD,* Dalila Hallal, MD,*
Jacques Levy, MD,* Lea Sjögren, MD,* Jean-Marc Legeais, MD, PhD,† Romain Nicolau, MD,†
Chadi Mehanna, MD,† Jean-Louis Bourges, MD, PhD,† and Charles Marc Samama, MD, PhD*
Purpose: To assess the safety of a laryngeal mask (i-gel) in
keratoplasty performed under general anesthesia.
Methods: Patients with indications for keratoplasty (n = 110) were
enrolled in a prospective study and randomly assigned to the i-gel (n =
55; 30 lamellar keratoplasty and 25 penetrating keratoplasty) or
tracheal tube group (n = 55; 29 lamellar keratoplasty and 26 penetrating
keratoplasty). Perioperative complications and the recovery time were
compared between the 2 groups using the t test or x2 test. Contra-
indications to elective use of the laryngeal mask airway (esophageal
reflux, extreme obesity, oropharyngeal pathology, or expected difficult
intubation) and expected difficult intubation were excluded.
Results: No surgical perioperative complications were reported in
either group. There was a significantly greater incidence of coughing
at extubation and/or after extubation in the tracheal group (40/55;
73%) than in the laryngeal mask group (3/55; 5%) (P , 0.001). The
recovery time was shorter in the i-gel group (80 minutes; 95%
confidence interval, 75–86) compared with that in the tracheal tube
group (88 minutes; 95% confidence interval, 82–95) (P = 0.03).
There were no significant differences in the incidence of sore throat
and hoarseness between both devices.
Conclusions: The use of i-gel for keratoplasty under general
anesthesia appears to be safe, reduces the risk of potential ocular
hypertension during recovery, and saves recovery time.
Key Words: keratoplasty, general anesthesia, laryngeal mask
(Cornea 2016;35:37–40)
P enetrating keratoplasty remains one of the most performed
procedures throughout the world, although lamellar
techniques are developing increasingly.1 General anesthesia
Received for publication July 4, 2015; revision received July 15, 2015;
accepted July 15, 2015. Published online ahead of print October 28, 2015.
From the *Service d’Anesthésie-Réanimation, Hôpital Cochin, Assistance
Publique-Hôpitaux de Paris, Université Paris Descartes, Paris, France; and
†Service d’Ophtalmologie, Hôpital Cochin, Assistance Publique-Hôpitaux
de Paris, Université Paris Descartes, Paris, France.
The authors have no funding or conflicts of interest to disclose.
Reprints: Gilles Guerrier, MD, Hôpital Cochin, 27 rue du Faubourg
Saint-Jacques, 75014 Paris, France (e-mail: guerriergilles@gmail.com).
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Cornea  Volume 35, Number 1, January 2016
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
(GA) is used less frequently in penetrating keratoplasty than
local anesthesia because of its higher cost, greater complexity,
and preoperative fasting requirement. However, it is usually
hard to obtain full cooperation from patients, especially with
extended operation times. Therefore, surgeons have to
perform penetrating keratoplasty quickly, limiting the time
available to treat pathological changes in the anterior
chamber, and therefore decreasing the success rate of the
procedure.2 When penetrating keratoplasty is performed
under GA, tracheal intubation is the standard means of
performing GA to secure the airway. However, insertion of
a tracheal tube often requires the use of neuromuscular
blockade, risks intubation difficulties, potentially increases
intraocular pressure (IOP),3 and is a potent stimulus to
coughing during recovery. Coughing after anesthesia for
penetrating keratoplasty may cause a rise in intraocular
tension and increase the possibility of suturing damage.4
New airway devices have been introduced for airway
management, trying to offer simple and effective alternatives to
tracheal intubation, namely laryngeal mask airway (LMA) and
supraglottic gel device (i-gel).5 The LMA has been success-
fully used for more than a decade. It has an airway tube that
connects to a mask with a cuff that is inflated. It has been used
in managing difficult airways and in different kinds of surgeries
with minimal changes in hemodynamic responses.6 However,
incorrect positioning of the LMA leads to aspiration, pneumo-
nia, air leakage, and partial airway obstruction with an
overinflated or malpositioned cuff.7,8 The i-gel airway (Inter-
surgical Ltd, Workingham, Berkshire, United Kingdom) is
a second-generation supraglottic device introduced in 2007,
with an anatomically designed mask made of a soft, gel-like,
and transparent thermoplastic elastomer. It has no cuff to
inflate, providing easy use and stability after insertion, and also
reducing the possible stimulatory effects of tissue compres-
sion.9 Despite its increasing use,10,11 studies of its safety during
keratoplasty are still lacking. Likewise, some practitioners are
reluctant to use the technique because of potential intra-
operative device dislodgement with a lack of access to the
airway during ophthalmic surgery.
The purpose of the study was to compare perioperative
anesthetic and surgical complications, time of recovery, and
the incidence of coughing in patients undergoing a kerato-
plasty procedure under GA using either an i-gel or an
endotracheal tube.
www.corneajrnl.com | 37
 Guerrier et al Cornea  Volume 35, Number 1, January 2016

METHODS the tracheal tube group and in the i-gel group, respectively, at
Patients with indications for keratoplasty (n = 110) the 0.05 significance level with 0.8 power. STATA (Stat
were enrolled in a prospective study and randomly assigned Corp, College Station, TX) was used for analysis. To avoid
to i-gel (n = 55; 30 lamellar keratoplasty and 25 penetrating the effects of crossover and dropout, which may break
keratoplasty) or tracheal tube group (n = 55; 29 lamellar the random assignment to the treatment groups, we used the
keratoplasty and 26 penetrating keratoplasty), from Novem- intention-to-treat analysis, providing unbiased comparisons
ber 15, 2014 to February 15, 2015. Approval from the local among the treatment groups. Quantitative and qualitative
institution review board (Approval CPP: SC 3239, November variables were compared using paired Student t tests and x2
10, 2014) and signed written consent from all patients were tests, respectively. When the frequency of events was #5 or
obtained. values did not follow normal distributions, Fisher exact and
Contraindications to elective use of the LMA (esoph- Mann–Whitney tests were used.
ageal reflux, extreme obesity, oropharyngeal pathology, or
expected difficult intubation) were exclusion criteria. Pre-
medication (hydroxyzine 1 mg/kg) was given approximately RESULTS
1 hour before induction of anesthesia to all patients. Demographic data are summarized in Table 1. There
Anesthesia was induced with intravenous sufentanil (0.25 were no significant differences between the study groups with
mg/kg) and propofol (3 mg/kg). After induction, either respect to age and smoking habits/preoperative respiratory
a tracheal tube or laryngeal mask was inserted according to disease. LMAs were inserted at the first attempt in 50 patients
randomization. Neuromuscular blockade was provided by and at the second attempt in 5 patients. A satisfactory airway
atracurium (0.5 mg/kg) when necessary. Size 7.0 tracheal control was not achieved with the laryngeal mask in 3
tubes and size 3 LMAs were used for female patients, patients, who were subsequently intubated. All 55 patients
whereas size 8.0 tracheal tubes and size 4 LMAs were used in the tracheal tube group were intubated at the first attempt.
for male patients. The number of attempts for insertion was No surgical perioperative complications were reported
recorded, including in patients with unexpected difficult in either group. There was a significantly greater incidence of
intubation. If insertion of the i-gel was poor or impossible, coughing at extubation and/or after extubation in the tracheal
the protocol dictated that the trachea should be intubated. group (40/55; 73%) than in the laryngeal mask group (3/55;
Satisfactory ventilation parameters included good chest 5%) (P , 0.001). In the tracheal tube group, 4/55 patients
expansion, no audible leaks, a capnograph square box demonstrated stridor in the recovery period compared with
waveform, and expired volume $7 mL/kg. Patients’ lungs 0/55 in the LMA group (P = 0.9). There were no significant
were ventilated with 50% air in oxygen and sevoflurane differences in the incidence of sore throat and hoarseness
adjusted to achieve a bispectral index between 40 and 60. between both devices (Table 2). The recovery time was
The minute volume was adjusted to obtain an EtCO2 of 35 to shorter in the i-gel group (80 minutes; 95% confidence
38 cmH2O. At the end of the surgical procedure, residual interval, 75–86) than in the tracheal tube group (88 minutes;
neuromuscular blockade was reversed with neostigmine 95% confidence interval, 82–95) (P = 0.03) (Table 3).
40 mg/kg and atropine 15 mg/kg, when appropriate. The
patients were extubated once spontaneous respiration had DISCUSSION
returned. The term “extubation” will refer to the removal of
The main finding of this study was that i-gel is
both LMA and tracheal tube.
apparently safe when used appropriately in patients undergo-
Demographic characteristics, preoperative comorbid-
ing keratoplasty. No perioperative anesthetic complications
ities, and the American Society of Anesthesiologists physical
occurred in our patients when using the i-gel, providing
status (ASA score) were recorded. The incidence of coughing,
stridor, and sore throat was noted both at extubation and for
a 60-minute period in the recovery room. The patients were
questioned about the presence of sore throat and hoarseness TABLE 1. Clinical Characteristics According to the Device
before leaving the postanesthesia care unit. An adverse Used for Ventilation
airway event was defined as oxygen desaturation of 90% or I-gel, n = 55 Tube, n = 55
less; significant airway trauma; or other major adverse events. (95% CI) or (95% CI) or
Characteristics n (%) n (%) P
Perioperative time intervals were recorded, delays from
induction to admission to recovery, time of surgical pro- Age, yr 55 (49–60) 47 (41–53) 0.98
cedure, and also the time of device insertion and time spent in Sex
recovery. On removal of the supraglottic airway, note was Female 30 (54) 29 (53) 0.85
made if any blood was visible on the device, indicative of Male 25 (46) 26 (47)
trauma in the upper airway. Smoking or respiratory disease 20 (36) 22 (40) 0.75
ASA
1–2 51 (93) 49 (89) 0.47
Statistical Analysis 3–4 4 (7) 6 (11)
A sample size of 55 subjects per device was needed to Weight 73 (69–76) 91 (65–116) 0.08
detect a 50% reduction of the incidence of coughing, CI, confidence interval.
assuming the incidence of the event was 60% and 30% in

38 | www.corneajrnl.com Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
 Cornea  Volume 35, Number 1, January 2016 Supraglottic Gel Device and the Endotracheal Tube

this context, the difference between both groups must be

TABLE 2. Operative Characteristics and Outcome According
ascribed to the airway adjunct used. Because the supraglottic
to the Device Used for Ventilation
devices tend to provoke swallowing rather than cough reflex
I-gel, n = 55 Tube, n = 55 in patients recovering from GA, a greater incidence of
(95% CI) or (95% CI) or
Variable n (%) n (%) P coughing and stridor may therefore be expected. The i-gel
lies in a wholly supraglottic position, whereas the tracheal
Type of keratoplasty
tube passes through the vocal cords into the trachea and may
Penetrating 39 (53) 30 (55) 0.85
therefore be presumed to provide a greater stimulus to sensory
Lamellar 26 (47) 25 (45)
receptors involved in airway protection reflex arcs.
Anesthetic procedure
It is acknowledged that a significant postoperative
Myorelaxation 0 49 (89) ,0.001
increase in IOP after keratoplasty should be avoided. The
Reversal 0 30 (54) ,0.001
mechanism of the increase in IOP after cough has not been
Outcome in recovery
extensively explored, partly because of practical issues. With
No cough 52 (95) 15 (27) ,0.001
recent improvements in suture material, it may be less
Cough before extubation 0 22 (40) ,0.001
important to prevent patients from coughing postoperatively
Cough at extubation 3 (5) 18 (33) ,0.001
after keratoplasty.12 However, coughing, laryngospasm, and
Sore throat 5 (10) 5 (10) 1
breath holding all have the capacity to cause hypoxemia, and
Stridor 0 4 (7) 0.9
a technique that reduces their incidence must therefore be
Bronchospasm 0 1 (2) 1
considered a potential advantage to the patient.
Blood on device 2 (4) 0 0.9
Importantly, the use of the LMA is not universally
Dysphagia 4 (7) 1 (2) 0.7
applicable to patients undergoing this type of surgery. Any
CI, confidence interval. patient at risk of regurgitation should be intubated with
a cuffed tracheal tube. This group of patients includes those
with a clear history of esophageal reflux, with proven hiatal
hernias and gastric outflow obstruction. In addition, it may
contraindications were respected and initial ventilation quality not be satisfactorily possible to inflate the lungs of patients
was checked. No axial rotation and malpositioning during with low compliance or high airway resistance without
surgery were reported. Four of 55 intubated patients suffered substantial leak around a LMA. These patients may be better
stridor in the recovery period, all of them having smoking suited to intubation with a tracheal tube. However, in practice,
habits with or without chronic obstructive pulmonary disease. such contraindications are infrequent, and the use of the LMA
Although this was not significantly different from the is applicable to the majority of patients having keratoplasty
incidence in the i-gel group, this device may be considered under GA.
in this population. Although not significant, blood staining of Another major finding of our study is the shortening of
the device at extubation and dysphagia after extubation was delays when using the i-gel, particularly time spent in
lower for the tube group than in the i-gel group. These recovery. Although not significant, time between the end of
complications may be caused by compressing the microvas- surgery and admission to recovery was shorter in the i-gel
cular structures and terminal nerve endings in these tissues.9 group. The efficient use of operating theater time is important
However, those local complications were minor and for several reasons. First, operating sessions are plainly an
infrequent. expensive and limited resource. In addition, more efficient use
Another important finding of this study was that cough of operating theaters may help in clearing waiting lists.
occurred less often in patients whose airway was controlled Finally, operating time frame is often used as an index of
using the i-gel than in those whose airway was secured with surgeons’ performance. In the literature, identified reasons for
a tracheal tube. The anesthetic protocol ensured that both delay during elective operating lists mainly refer to patients
groups were extubated at an equivalent depth of anesthesia. In transfer to and from the operating theater,13 but we found no
studies focusing on the anesthetic technique as a potential
solution to save time. Finally, because the administration of
a neuromuscular blocker is not necessary to insert an i-gel
TABLE 3. Surgical Delays According to the Device Used for device, several benefits may be expected, including saving
Ventilation time at induction (although not significant in our study), less
I-gel, n = 55 Tube, n = 55 allergic reactions, and optimized cost-effectiveness.
Theater Timings (95% CI) (95% CI) P Our study has from several limitations. First, insertions
Induction to incision 16 (14–18) 16 (14–18) 0.58 were performed in patients who had a normal airway; our
End of surgery to recovery room 9 (8–10) 11 (9–13) 0.06 results may not apply to patients with a difficult airway.
End of surgery to extubation 22 (18–26) 22 (19–25) 0.55 Second, our results are specific to the anesthetic administered
Recovery length of stay 80 (75–86) 88 (82–95) 0.03 and might not apply to other anesthesia regimens. Finally, the
Induction to exit from recovery 156 (147–165) 171 (162–179) 0.009 study was conducted on patients who were not at risk of
room aspiration. Hence, the results from these data cannot be
CI, confidence interval. extrapolated to all other groups of patients. Regardless of
limitations, our study applied rigorous methodology to

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. www.corneajrnl.com | 39

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
 Guerrier et al
compare the i-gel and tracheal tube and is the first report
involving patients undergoing keratoplasty. In addition, we
are confident in the homogeneity between the evaluated
patients, including differences in the use of neuromuscular
blocking drugs and reversal, the depth of anesthesia, and the
expertise of the anesthesiologist inserting the device.
In summary, the use of i-gel for keratoplasty under GA
appears to be safe, reduces the incidence of immediate
postoperative upper airway complications after GA for
keratoplasty, and saves recovery time. Its use in this situation
may therefore offer an advantage to many patients. Future
research should assess the cost-effectiveness of using supra-
glottic devices in ophthalmic surgery.
REFERENCES
1. Agence de la biomedicine. Le rapport medical et scientifique du prélèvement
et de la greffe en France. Activité de prélèvement, de greffe de cornée et
d’inscription en attente de greffe. Available at: www.agence-biomedecine.fr/
annexes/bilan2013/donnees/prelevement/02-cornee/synthese.htm. Accessed
May 3, 2015.
2. Wang X, Dang GF, Li YM, et al. General anesthesia versus local
anesthesia for penetrating keratoplasty: a prospective study. Int J
Ophthalmol. 2014;7:278–282.
3. Ismail SA, Bisher NA, Kandil HW, et al. Intraocular pressure and
haemodynamic responses to insertion of the i-gel, laryngeal mask airway
or endotracheal tube. Eur J Anaesthesiol. 2011;28:443–448.
40 | www.corneajrnl.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Cornea  Volume 35, Number 1, January 2016
4. Holden R, Morsman CD, Butler J, et al. Intra-ocular pressure changes
using the laryngeal mask airway and tracheal tube. Anaesthesia. 1991;46:
922–924.
5. Bein B, Scholz J. Supraglottic airway devices. Best Pract Res Clin
Anaesthesiol. 2005;19:581–593.
6. Ziyaeifard M, Azarfarin R, Massoumi G. A comparison of intra- ocular
pressure and hemodynamic responses to insertion of laryngeal mask
airway or endotracheal tube using anesthesia with propofol and
remifentanil in cataract surgery. J Res Med Sci. 2012;17:503–507.
7. Weiler N, Latorre F, Eberle B, et al. Respiratory mechanics, gastric
insufflation pressure, and air leakage of the laryngeal mask airway.
Anesth Analg. 1997;84:1025–1028.
8. Barker P, Langton JA, Murphy PJ, et al. Regurgitation of gastric contents
during general anaesthesia using the laryngeal mask airway. Br J
Anaesth. 1992;69:314–315.
9. Levitan RM, Kinkle WC. Initial anatomic investigations of the I-gel
airway: a novel supraglottic airway without inflatable cuff. Anaesthesia.
2005;60:1022–1026.
10. Uppal V, Gangaiah S, Fletcher G, et al. Randomized crossover
comparison between the i-gel and the LMA-Unique in anaesthetized,
paralysed adults. Br J Anaesth. 2009;103:882–885.
11. Shin WJ, Cheong YS, Yang HS, et al. The supraglottic airway I-gel in
comparison with ProSeal laryngeal mask airway and classic laryngeal
mask airway in anaesthetized patients. Eur J Anaesthesiol. 2010;27:
598–601.
12. Foulds WS. The changing pattern of eye surgery. Br J Anaesth. 1980;52:
643–647.
13. Marjamaa R, Vakkuri A, Kirvelä O. Operating room management: why,
how and by whom? Acta Anaesthesiol Scand. 2008;52:596–600.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.