Comparative Efficacy of the Air-Q Intubating Laryngeal Airway during General Anesthesia in Pediatric Patients: A Systematic Review and Meta-Analysis

Air-Q® (air-Q) is a supraglottic airway device which can be used as a guidance of intubation in pediatric as well as in adult patients. We evaluated the efficacy and safety of air-Q compared to other airway devices during general anesthesia in pediatric patients by conducting a systematic review and meta-analysis. A total of 10 studies including 789 patients were included in the final analysis. Compared with other supraglottic airway devices, air-Q showed no evidence for a difference in leakage pressure and insertion time. The ease of insertion was significantly lower than other supraglottic airway devices. The success rate of intubation was significantly lower than other airway devices. However, fiberoptic view was better through the air-Q than other supraglottic airway devices. Therefore, air-Q could be a safe substitute for other airway devices and may provide better fiberoptic bronchoscopic view.


Introduction
There are many supraglottic airway devices (SADs) which are used for the management of a difficult airway and as a conduit for tracheal intubation [1]. The endotracheal intubation assisted by SADs has many advantages including easy insertion, better alignment of the glottis opening, and continuous patient oxygenation and ventilation. Moreover, the haemodynamic stress response to intubation by SAD is less than that of direct laryngoscope [2]. Also, these devices can be a good alternative for patients with previous history of difficult intubation, restricted neck mobility, and stability of cervical spine [3]. Moreover, SAD provides the ability to overcome upper airway obstruction and provision of a hands-free airway with a relatively straightforward path to the larynx [4].
Among these SADs, Air-Q Intubating Laryngeal Airway (air-Q, Mercury Medical, Clearwater, FL, USA) is a SAD intended for allowing for airway maintenance under general anesthesia as well as an aid for tracheal intubation with a cuffed tracheal tube in both adults and pediatric patients. The design of air-Q includes a large airway tube inner diameter (ID), a short airway tube length, and a tethered, removable standard 15 mm circuit adapter [5]. Also, specially designed ridges inside mask cavity provide lateral stability to help prevent tip from bending backward and improve mask seal. These features enable direct insertion of a curved shaft, the lack of a grill in the ventilating orifice, and an easily removable airway adapter [5].
Several studies reported the efficacy of air-Q in pediatric patients comparing other airway devices such as laryngeal mask airway (LMA), i-gel, Aura-I, Cobra Perilaryngeal Airway (CobraPLA), and fiberoptic guided intubation. However, the findings are variable and the reported outcomes from several studies are conflicting. To date, no systematic review nor meta-analysis regarding air-Q has been performed. Therefore, we aimed to evaluate the efficacy of air-Q compared with other airway devices in pediatric patients.

Methods
This systematic review was conducted following the guidelines of the PRISMA statement [6].

Systematic Search.
We conducted a systematic review and meta-analysis for RCTs which compared the air-Q with the other airway device during general anesthesia in pediatric patients. Literature searches were conducted in MEDLINE, EMBASE, Cochrane Library, KoreaMed, KMBASE, and Google Scholar inclusive at June 1, 2014, and updated at September 2015. The search strategy combining free text and related search is attached in the Appendix.

Selection of Included Studies.
The study's inclusion and exclusion criteria were determined before systematic search. Two review authors (Eun Jin Ahn and Si Ra Bang) independently scanned the titles and abstracts identified by the variety of search strategies described above. If the report was determined not eligible from the title or abstract, the full paper was retrieved. Potentially relevant studies, chosen by at least 1 author, were retrieved and evaluated in fulltext versions. The articles that met the inclusion criteria were assessed separately by 2 authors (Eun Jin Ahn and Si Ra Bang), and any discrepancies were resolved through discussion. If agreement would not be reached the dispute was resolved with the help of third investigator (Hyun Kang).

Inclusion and Exclusion Criteria.
We included randomized controlled trials which compared the air-Q and the other airway device during general anesthesia in pediatric anesthesia. The group used air-Q as an airway device considered as an experimental group. Otherwise, group used other laryngeal mask airway devices and fiberoptic bronchoscope as an airway device considered as a control group. We excluded data from abstracts, posters, case reports, comments or letters to the editor, reviews, and animal studies. There was no limitation in language to select studies.

Study
Outcomes. The outcome data were divided into two series as if the air-Q is used as a SAD or used as a conduit of tracheal intubation. In series of air-Q used as SAD, the outcomes included success rate of insertion of airway device, oropharyngeal leakage pressure, insertion time, and ease of insertion airway device. In series of air-Q used as a conduit of intubation, the outcomes included success rate of intubation, the number of attempts, intubation time, and fiberoptic glottis view. Also, complications including desaturation, sore throat, blood staining of device, and laryngospasm were evaluated.
We compared the number of attempts through the rate of successful intubation at first attempt. Ease of insertion of airway device or intubation through airway device was analyzed by comparing the number of easiest level of difficulty. The extracted outcome data of fiberoptic glottis view grade were divided into 4 grades [13][14][15][16] (visible vocal cords/visible vocal cord with posterior epiglottis/visible vocal cord with anterior epiglottis/no visible vocal cord) or 5 grades [8][9][10][11][12] (visible vocal cords/visible vocal cord with posterior epiglottis/vocal cord anterior visible <50% obstruction/visible vocal cord with anterior epiglottis, >50% obstruction/no visible vocal cord). The scenarios were divided into best and worst which include the incidence of fiberoptic glottis view grade with visible vocal cord and no visible vocal cord.
We performed subgroup analyses for comparing groups, SADs, and fiberoptic guided intubation. SADs were also divided into Aura-I, CobraPLA, i-gel, and LMA series which include LMA-Unique and LMA-Flexible. We also performed subgroup analyses based on the generations of SADs. Firstgeneration devices included LMA-Unique, LMA-Flexible, and CobraPLA. The second-generation devices included Aura-I, i-gel, and LMA-Fastrach. The sensitivity analysis was performed to rule out the excessive effect of single study on heterogeneity.

Validity
Scoring. The quality of eligible studies was assessed independently by two authors (Chong Wha Baek and Yong Hun Jung) of our review group using the tool of "risk of bias" according to Review Manager software (version 5.3, the Cochrane Collaboration, Oxford, UK). The quality was evaluated based on the following seven potential sources of bias: random sequence generation; allocation concealment; blinding of the participants and their parents; blinding of outcome assessment; incomplete outcome data; selective reporting. The methodology of each trial was graded as "high," "low," or "unclear," to reflect a high risk of bias, low risk of bias, and uncertainty of bias, respectively (Table 3).

Data Extraction.
All interrelated data in each included study were independently extracted onto a spreadsheet by 2 authors (Eun Jin Ahn and Geun Joo Choi) and cross-checked. The spreadsheet included the following indexes: (1) name of first author, (2) year, (3) name of journal, (4) study design, (5) registration of clinical trial, (6) risk of bias, (7) number of patients in study, (8) sex of patients, (9) age, (10) weight, (11) height, (12) ASA physical status, (13) type of surgery, (14) number and experience of device user, (15) number of allowance of insertion trials, (16) device size, (17) induction method, (18) maintenance agent, (19) use of muscle relaxant, and (20) intervention/control. If there were missing data, we made attempts to contact the authors to chase the data of included studies. And if the author could not be contacted, the data were acquired from estimated value in the figure or graph.

Statistical Analysis.
We conducted this meta-analysis by Review Manager (version 5.3, the Cochrane Collaboration, Oxford, UK) and Comprehensive Meta-Analysis software (version 2.0, Biostat, Englewood, NJ, USA). Three authors (Eun Jin Ahn, Geun Joo Choi, and Young Cheol Woo) independently input all data to the software. For dichotomous data, we calculated pooled risk ratio (RR), odds ratio (OR), and 95% confidence intervals (CIs). If the 95% CI included a value of 1, we considered the difference not to be statistically significant. We calculated the mean difference for continuous data, also reported with 95% CI. We used the Chi-squared test and the -squared test for heterogeneity. A level of 10% significance ( < 0.1) for the Chi-squared statistic or 2 greater than 50% was considered to indicate considerable heterogeneity. The Mantel-Haenszel random-effect model was used for these studies. The Mantel-Haenszel fixed model was used for studies that did not demonstrate significant heterogeneity [17,18]. For data expressed with median and interquartile range, we changed mean and standard deviation via data extraction method from Cochrane handbook for systematic reviews of intervention [17].
We estimated publication bias using Begg's funnel plot and Egger's linear regression test. If the funnel plot was visually asymmetrical or the value was found to be <0.1 using Egger's linear regression test, the presence of a possible publication bias was identified [19].

Risk of Bias.
In all the included studies, the random sequence generation method was performed and seven studies used allocation concealment [8,10,[12][13][14][15][16]. Three RCTs were registered in clinical trial [8,9,12,16] and no incomplete data was reported. All included studies reported no sponsorship. The overall risks of bias are shown in Table 4.

Success Rate of Device Insertion.
Success rate of device insertion was compared in seven studies [8-10, 12, 14-16]. The combined results showed no evidence for a difference, RR 1.00 (0.98 to 1.03), = 0.90, 2 = 0%. In most studies, the success rate of device insertion was 100%. Only in one study [16], the success rate of device insertion of air-Q was 97.5% (39/40).

Total Success Rate of Intubation.
The total success rate of intubation was compared to other airway devices in five studies [7,9,11,13,16]. Among these studies, four studies performed the assisted fiberoptic intubation through the airway devices. However, in the single study of Kleine-Brueggeney et al. [16], both groups performed blind intubation through the airway devices rather than assisted fiberoptic intubation.     [7,9,11,13]. The combined results showed no evidence of difference in the rate of successful intubation at first attempt in air-Q compared to other airway devices. However, the result showed a tendency of the rate of successful intubation at first attempt in air-Q which was higher than other airway devices, RR 1.08 (95% CI 0.99 to 1.17), = 0.29, 2 = 20%.

Time to Intubate.
The time to intubation was compared to other airway devices in five studies [7,9,11,13,16]. The combined results showed no evidence of a difference in the time to intubation, MD 1.341 (95% CI −4.44 to 7.13), Senior anesthesiologists from the pediatric anaesthesia division Elective ophthalmic surgery 3 Air-Q/Aura-I = 0.04, 2 = 60%. Subgroup analyses were performed based on airway devices, SADs, fiberoptic guided. The combined results of four studies [9,11,13,16] which compared air-Q and other SADs showed no evidence of difference, MD −0.18 (95% CI −6.20 to 5.84), = 0.07, 2 = 58%. Also, in a single study of Sohn et al. [7], there was no difference in time to intubation between free-handed and air-Q assisted tracheal intubation ( = 0.13).
The worst scenario of fiberoptic view showed no evidence of difference in the air-Q compared to other SADs, RR 1.10 (95% CI 0.59 to 2.07), = 0.54, 2 = 0%. Performing subgroup analysis based on the generations of airway devices, the worst scenario of fiberoptic view in air-Q showed no evidence of difference compared to first generation of SADs, RR 0.78 (95% CI 0.30 to 2.03), = 0.57, 2 = 0%. Also, there was no evidence of difference in the worst scenario of fiberoptic view in air-Q compared to second generation of SADs, RR 1.44 (95% CI 0.62 to 3.34), = 0.25, 2 = 27%.
There was no evidence of difference for the incidence of blood staining on device, RR 0.57 (95% CI 0.22 to 1.50), = 0.53, 2 = 0%. However, in the single study of Girgis et al. [13], the incidence of blood staining on device was significantly higher in patients with CobraPLA than Air-Q (30% versus 6.7%).

Publication Bias.
There was no evidence of publication bias detected by Egger's linear regression test and funnel plot. There was no value of Egger's regression test that was showed to be <0.1 which is indicative of publication bias.

Discussion
The major finding of our meta-analysis was that air-Q is more difficult to insert than other airway devices. Also, the total   success rate of intubation and the rate of successful intubation at first attempt showed a tendency lower than other airway devices. However, the fiberoptic view in best scenario was better in air-Q than other SADs. Also, in a safety analysis, the incidence of sore throat was lower in air-Q than other airway devices. The combined analysis of OLP, success rate of device insertion, insertion time, time to intubation, and the worst scenario of fiberoptic view could not reveal the difference between air-Q and other airway devices. A tendency of higher success rate of intubation at first attempt in air-Q than other airway devices might be caused by better view of fiberoptic videoscope of air-Q. Using fiberoptic bronchoscope to guide tracheal intubation through a SAD is an established technique for securing the airway in children when conventional laryngoscopy is failed [11]. Also, SAD provides the ability to overcome upper airway obstruction and provision of a hands-free airway with a relatively straightforward path to the larynx [4,7,11]. The total success rate of blind intubation in Kleine-Brueggeney et al. 's study was only 10% (6/60), while the total success rate of air-Q involving four studies which used air-Q as a fiberoptic guidance for intubation [7,9,11,13] was 82.5% (147/178). Because of high incidence of worst scenario in pediatric patients and potential injury to the epiglottis or the glottis, blind intubations through SADs are not recommended in several studies [16,28,29]. The infants have large and floppy epiglottis preventing visualization of glottis [29]. In a pilot study of Sinha, the incidence of best scenario of fiberoptic bronchoscopy was higher in infants compared to other studies aimed at children [28][29][30]. In conclusion, it is recommended that SGA be used as a fiberoptic bronchoscope guidance of tracheal intubation not as a guide of blind intubation in children. Also, further studies should be needed to apply the result of this study according to patients' age.
Performing the subgroup analysis, OLP was higher in air-Q than in Aura-I. The possible reasons for better OLP in air-Q are due to the unique features which include (i) curved and rigid airway tube which approximates the upper oropharyngeal airway with the glottis, (ii) mask ridges which improve the transverse stability of the bowl and support the lateral cuff seal, and (iii) raised mask heel [8,10,25]. Even though i-gel has shown higher airway leak pressure compared with other airway devices in children [31], the subgroup analysis showed no evidence of difference. These unique features of air-Q, especially raised mask heel, would have caused better OLP; otherwise they caused the following issues. The insertion of air-Q was more difficult and needed longer time compared to other SADs.
There are SADs which are simply "airway device" which may or may not protect against aspiration in the event of regurgitation. These SADs are called first generation including classic LMA, LMA-Flexible, laryngeal tube, and Cobra Perilaryngeal Airway. Second-generation SADs including LMA ProSeal, LMA Supreme, and i-gel provide a higher leak pressure and offer a drain tube to separate the respiratory and gastrointestinal tracts and minimize the risk of aspiration [32,33]. Air-Q, classified in second-generation SAD [34], showed no evidence of difference in OLP compared to first-generation SADs in this meta-analysis. This result suggests that air-Q has little advantage in OLP compared to first-generation SADs and no further studies would be needed.
Through safety analysis, the incidence of sore throat was lower in patients in air-Q than other airway devices including CobraPLA and LMA. This result can be associated with the lower incidence of blood staining on device in air-Q than CobraPLA. Comparing CobraPLA and air-Q, "Cobra head" of the CobraPLA is more stiff in comparison with the cuff of other SADs including the air-Q which lead to more mucosal injury [13]. Therefore, we suggest that air-Q might be less traumatic SAD compared to other airway devices.
In two studies, self-pressurized air-Q was used rather than air-Q with balloon [10,12]. Self-pressurized air-Q has an inner aperture at the junction of the airway tube and the mask cuff, creating an open airspace between the two and allowing the pressure to be self-regulated, which might provide easier device insertion and reduced risk for prolonged overinflation of the cuff and pressure-related injuries to the pharyngeal mucosa without the need for cuff pressure monitoring [10]. To rule out the effect of difference between air-Q which requires cuff inflation and self-pressurized air-Q, subgroup and sensitive analysis was performed. However, no result was changed through subgroup analysis and sensitive analysis.
There are some limitations in our study. First, because of air-Q being a newly developed SAD and a large variety of study designs, the number of studies involved in each subgroup was small. Second, large difference in the operator's experience (from trainee to senior anesthesiologists) could be a confounding factor.
In summary, we found that the air-Q could be a safe substitute for other intubating laryngeal mask airway devices and might provide better fiberoptic bronchoscopic view and shorter time to guide intubation.