Airway Complications in Intubated Versus Laryngeal Mask Airway–Managed Dentistry: A Meta-Analysis

Abstract

Objective:

Serious airway complications can occur with inadequate airway management during general anesthesia (GA). This meta-analysis investigated randomized controlled trials that compared perioperative technique failures and airway complications, including hypoxia, during GA for dentistry using endotracheal intubation or a laryngeal mask airway (LMA) for airway management.

Methods:

A systematic search of electronic databases and gray literature was completed. Independent reviewers assessed eligibility, performed data extraction, completed risk of bias assessment, and judged the quality of results through Grading of Recommendations, Assessment, Development, and Evaluation. Risk ratios (RRs) for airway complications, with 95% CIs, were calculated. Heterogeneity was quantified using the I² statistic. Sensitivity and age-subgroup analyses were explored.

Results:

Six trials were deemed eligible from a total of 9076 identified reports. The airway management intervention for these trials was LMA. Technique failures or effect differences in airway complications were not detected except for postoperative hypoxia, where LMA use had a decreased risk (RR, 0.22; 95% CI, 0.06-0.77; I² = 0%; moderate quality). A similar effect was seen in the pediatric analysis (RR, 0.10; 95% CI, 0.01-0.84; I² = 0%; moderate quality). Additionally, LMA use reduced pediatric sore throat risk (RR, 0.08; 95% CI, 0.04-0.15; I² = 0%; moderate quality).

Conclusion:

Use of an LMA in dentistry may have the potential to reduce the risk of postoperative hypoxia, particularly in pediatric patients, although further study is required.

General anesthesia (GA) has a profound importance in dentistry, particularly for patients with anxiety, dental phobias, inadequate pain control, or special needs (such as mental or physical impairments, poor cooperation, or behavior issues), or those who require full mouth rehabilitation.^1,2 Five million Americans undergo dental extraction surgery each year.³ It is estimated that 46% of these surgeries require sedation or GA.⁴ In Canada, early childhood caries in patients aged 12 to 59 months is the most common indication for GA, comprising 31% of all hospital day surgeries within this age group.⁵

The gold standard for airway management is endotracheal intubation,⁶ a vital component of GA since the late 1800s that provides a secure patent airway, eases ventilation, and protects against laryngospasm and aspiration.⁷ However, intubation may be associated with complications including failure of placement, trauma to the oral or nasal anatomy, bronchospasm, obstruction, aspiration, epistaxis, postoperative sore throat, and damage to the laryngeal tissues.^8,9 Therefore, anesthesia providers may utilize less-invasive nonintubation techniques to avoid many of these potential complications. One method of nonintubated airway management is the use of a laryngeal mask airway (LMA) device that is placed above the vocal cords in the hypopharynx.^10,11

Unlike most medical surgeries, dental procedures typically impinge upon the airway. In North America, approximately 60% of dentist anesthesiologists do not routinely intubate,¹² which clearly identifies variance in practice. Airway management is a topic of interest because the loss of airway reflexes associated with GA can lead to significant airway complications.¹³ To our knowledge, there is no systematic review comparing airway management techniques for GA for dentistry. Therefore, a systematic search of the literature is required to assess if nonintubated airway management using an LMA is safer compared to intubation, in terms of airway complications. Objectives for this study were the following:

1. To compare airway management success for intubated and nonintubated (LMA) patients under GA for dental procedures in randomized controlled trials (RCTs).

2. To compare the relative risk of anesthetic airway complications between intubated and nonintubated (LMA) patients under GA for dental procedures in RCTs.

METHODS

In this review, we followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) criteria.¹⁴

Inclusion and Exclusion Criteria

We included RCTs that would match the following Population, Intervention, Comparison, Outcomes, and Study design (PICOS) framework. Any study not meeting our PICOS framework below was excluded.

• Population: patients undergoing GA for dentistry, defined as a procedure performed by a dental health care professional. No age restriction was applied.

• Intervention: airway managed by use of an LMA.

• Comparison: airway managed by endotracheal intubation.

• Outcome: The primary outcome was successful airway management, defined as the absence of death, case cancellation because of an airway concern, or an intraoperative change from the planned airway management technique. Secondary outcome measures were perioperative airway complications or potential complications, including hypoxia (defined as a reduced arterial oxygen saturation recorded after airway device removal and prior to discharge home, as well as the oxygen saturation criteria of each individual article); bronchospasm (defined as respiratory wheezing recorded prior to discharge home); airway obstruction (such as complete/partial obstructions or laryngospasms recorded after airway device removal and prior to discharge home); laryngeal soiling (visualized intraoperatively or aspiration recorded prior to recovery); sore throat (utilized the maximum recorded incidence); and perioperative nausea and vomiting (PONV). For the latter, if the incidence of PONV was measured at different time points, the value assessed nearest to the beginning of the recovery phase was utilized. Furthermore, data were excluded if recorded after discharge home or if recorded following a prolonged monitored recovery or recovery indicated as greater than 1 hour.

• Study design: RCTs.

Search Strategy

A wide nonspecific strategy (Supplemental File S1) was utilized, adhering to the listed PICOS framework. We searched CENTRAL, Ovid MEDLINE, Ovid EMBASE, and Scopus from inception to week 4 of January 2019. The search was limited to English articles. Unpublished literature and gray literature were searched by screening the first 100 results in Google and a 2-year hand search of relevant journals (Journal of Oral and Maxillofacial Surgery, International Journal of Oral and Maxillofacial Surgery, and Anesthesia Progress) and relevant conference proceedings (American College of Oral Maxillofacial Surgeons Conference, Canadian Association of Oral Maxillofacial Surgeons Conference, International Conference on Oral Maxillofacial Surgery, and International Association for Dental Research). Completed and ongoing trials were searched using ClinicalTrials.gov, World Health Organization International Clinical Trials Registry Platform (who.int/ictrp), and ProQuest (www.proquest.com), along with a hand search of bibliographies of eligible RCTs.

Screening and Data Collection

Two review authors (J.P. and C.G.) independently reviewed and screened trials from the search. Disagreements were resolved through discussion and consensus, or by consulting other review authors (A.A. and M.W.), with the primary reasons being procedures not provided by dentists, the study not being an RCT, or procedures not limited to GA. For eligible trials, the following information was extracted: participants, interventions, controls, outcomes, and study design, along with general study characteristics such as author, year, country, clinical setting, participant age, dental interventions, and American Society of Anesthesiology classification (Table 1). An attempt to contact the authors was made in case of missing information from any of the studies.

Table 1.

Demographics of Included Studies*

Source	Control (Intubation route)	Country	Setting	Population
Goodwin et al,15 1993	Nasal	United Kingdom	Hospital	60 ASA 1-2 patients for third molar dental extractions under GA
Quinn et al,16 1996	Nasal	United Kingdom	Educational facility	100 ASA 1-2 patients for oral surgery (predominantly wisdom tooth extraction) under GA
Hung et al,17 2005	Unknown route	Taiwan	Hospital	146 ASA 1-2 patients with mild-moderate developmental disabilities for nonspecific dental treatment under GA.
Kundra et al,18 2009	Oral	India	Educational facility	66 ASA 1-2 pediatric patients for cleft palate surgery under GA
Zhao et al,19 2014	Nasal	China	Hospital	171 ASA 1-2 pediatric patients for nonspecific dental surgery under GA.
Keles and Kocaturk,20 2018	Nasal	Turkey	Educational facility	70 ASA 1-2 pediatric patients for a wide variety of dental treatment (majority restorations) under GA

^*RCT indicates randomized controlled trial; ASA, American Society of Anesthesiology; and GA, general anesthesia.

Assessment of Risk of Bias in Included Studies

Three independent reviewers (J.P., C.G., and M.Z.) evaluated each included study using the Cochrane risk of bias assessment tool.²¹ We evaluated 7 domains (random sequence generalization, allocation concealment, blinding of participants and personnel, blinding outcome assessment, incomplete outcome data, selective reporting, and other bias). A final judgment for a domain was given based on the following: low risk if judged to be of low risk; moderate risk if judged unclear risk; high risk if judged to be of high risk. Disagreements were resolved through discussion and consensus or by consulting a fourth reviewer (A.A.).

Quality of Evidence

To assess the quality and confidence in the main results of pooled data, 2 review authors (J.P. and M.Z.) used the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) assessment tool (GRADEpro GDT, McMaster University, 2015; developed by Evidence Prime, Inc; available from gradepro.org). Five domains of study limitations—risk of bias, inconsistency, indirectness, imprecision, publication bias, and overall certainty of bias—were evaluated. Each domain's risk level was judged as very severe, severe, or no concerns. This either downgraded or upgraded the level of evidence to reach an overall judgment of either high, moderate, low, or very low.²² Consensus was reached by consulting a third reviewer (A.A.; Table 2).

Table 2.

GRADE Assessment for Meta-Analyses*

^*GRADE indicates Grading of Recommendations, Assessment, Development, and Evaluation.

Quantitative Data Synthesis

In the meta-analysis, we analyzed summary effects by considering risk ratios (RRs) together with the 95% CIs calculated for dichotomous outcome using Review Manager 5.3 software (Cochrane Collaboration). Random-effects models, with the Mantel-Haenszel method, were used to estimate pooled probabilities. We inspected the graphic display of the trials' estimated treatment effects (along with their 95% CIs) to assess heterogeneity in the generated forest plots. In order to identify and quantify heterogeneity, we used a χ² test (with a significance level of 0.1 as the cutoff value) and I² statistic values, respectively.²³ I² values were used based on suggestions from the Cochrane Handbook for Systematic Reviews of Interventions.²⁴

Subgroup Analysis and Investigation of Heterogeneity

Potential heterogeneity among studies was explored in a number of a priori subgroup and sensitivity analyses. We stratified data into subgroups based on age of participants, categorized as pediatric (population under the age of 16) and adult (population aged 16 or greater). We also performed sensitivity analysis to explore the risk of bias on outcomes, in which studies that were judged to have high risk of bias in the domain of random sequence generation were excluded.

RESULTS

Search Results

Figure 1 presents the study flow diagram. From the initial search of 9076 records, after 4239 duplicates were removed, 4837 articles were screened for title and abstract. Of these, a total of 12 articles were screened in full text, and 6 met our inclusion criteria.^15-20 The 6 other studies were excluded because of procedures not provided by dentists, not being RCTs, or procedures not limited to GA.

Figure 1.

Study flow diagram.

Study Characteristics

Table 1 explains the characteristics of these 6 included RCTs, 3 of which were conducted in educational facilities^16,18,20 and 3 in hospital settings.^15,17,19 Two RCTs were conducted in the United Kingdom^15,16 and 1 study each was conducted in China¹⁹ Taiwan,¹⁷ Turkey,²⁰ and India.¹⁸

Participants

Three RCTs were conducted on pediatric populations (n = 307),^18-20 2 were conducted with adults (n = 160),^15-16 and 1 was conducted on a mixed population aged 4 to 36 years (n = 146).¹⁷ Three RCTs were conducted with primarily surgical dental cases,^15,16,18 1 was conducted with a variety of dental procedures (the majority of which were restorations),²⁰ and 2 studies did not specify the type of dental procedures performed.^17,19

Classification of the Interventions vs Controls

The intervention in the 6 eligible RCTs was the use of an LMA device.^15–20 The control in all included trials was endotracheal intubation delivered through nasal cavity^15,16,19,20 or oral cavity¹⁸ or not specified by invesitgators.¹⁷

Risk of Bias

Risk of bias assessment and justification can be viewed in Figure 2 and Supplemental File S2. Two studies were assigned low risk of bias for all domains.^18,20 Four studies were assigned unclear risk of bias for at least 1 domain.^15-17,19 One study was assigned high risk of bias for at least 1 domain.¹⁹

Figure 2.

(A) Risk of bias summary: judgments from authors on risk of bias domains among individual included studies. (B) Risk of bias summary: judgements from authors on risk of bias domains across all included studies

Effects of Intervention on Airway Management Success and Failure

There were no recorded airway management failures identified in the included trials.

Effect of Intervention on Risk of Airway Complications

This meta-analysis showed that GA cases managed with an LMA resulted in a significant reduction of hypoxia (RR, 0.22; 95% CI, 0.06-0.77; P = .63; I² = 0%; moderate quality). It should be noted that the complication of hypoxia occurred only postoperatively, in recovery before discharge home. The meta-analysis failed to demonstrate any differences for airway obstruction (RR, 0.79; 95% CI, 0.16-3.86; P = .17; I² = 41%; very low quality), laryngeal soiling or aspiration (RR, 0.52; 95% CI, 0.08-3.32; P = .26; I² = 21%; low quality), sore throat (RR, 0.20; 95% CI, 0.02-2.08; P < .00001; I² = 97%; very low quality), and perioperative nausea or vomiting (RR, 0.13; 95% CI, 0.01-1.91; P = .01; I² = 84%; very low quality; Table 2 and Figure 3).

Figure 3.

Forest plot for airway complications from all included studies.

Sensitivity Analysis

Our sensitivity analysis (Figure 4) focused on studies without high risk of bias and with stronger methodological quality. This was accomplished by excluding the study by Zhao et al,¹⁹ which was assessed to have a high risk of bias in the domain of randomization. This was determined through our risk of bias assessment as a result of inadequate randomization of intervention and control in the methodology as presented in the above study. The results were similar to those of the primary analysis.

Figure 4.

Sensitivity analysis forest plot for airway complications from all included studies.

Subgroup Analysis

Our subgroup analyses focusing on pediatric populations showed that LMA-managed GA cases resulted in a significant reduction of hypoxia (RR, 0.10; 95% CI, 0.01-0.84; P = .77; I² = 0%; moderate quality) and sore throat (RR, 0.08; 95% CI, 0.04-0.15; P = .81; I² = 0%; moderate quality) but did not affect other airway complications (low to very low quality; Table 2 and Figure 5). Our subgroup analyses focusing on adult populations showed no significant effect from LMA use on any airway complications (low to very low quality; Table 2 and Figure 5).

Figure 5.

Age subgroup meta-analyses of airway complications presented in forest plot.

DISCUSSION

This meta-analysis identified 6 RCTs that would answer the a priori defined PICOS question. Although the comprehensive search strategy for nonintubated interventions included around 40 different key words/MeSH terms and variations (Supplemental File S1), only dental studies with LMAs were eligible for inclusion. This is striking and illustrates the need for further research and high-quality studies investigating other nonintubated airway management techniques. Common alternative methods for airway management during GA for dentistry that avoid placement of endotracheal tubes and LMAs include the use of nasopharyngeal airways and nasal cannulas.²⁵ The avoidance of intubation also is referenced in the literature by many terms, such as nonintubated,^14,15 open airway,²⁷ laryngeal mask airway,¹⁸ or supraglottic airway,²⁸ or by the level of anesthesia.^29,30

This study included dental procedures performed under GA because the surgical field often impinges upon anatomic spaces typically used during anesthetic management. It is believed that “sharing the airway” during these types of procedures results in a higher propensity for airway contamination and instigates airway complications. Nondental intraoral surgeries were excluded to respect the authors' professional interest in dentistry.

No airway management failures were stated in the eligible articles. This may mean that both airway techniques were equally safe or that any intraoperative concerns were successfully managed. Training and case selection may also have contributed to the lack of airway management failures. Alternatively, this finding could have resulted from a lack of direct investigation of this specific question, or relevant studies may have been inadequately powered to detect potentially rare events. Furthermore, there is a strong possibility of the inability to detect potential events. Therefore, future research is required to address the important question of success.

There were justifications for pooling all airway complication data regardless of age. Reasons included that both LMAs and endotracheal tubes are utilized in all ages, that these airway devices are size adjusted based on patient age or weight,^31,32 and that airway complications cause morbidity and mortality for both age segments.^33–35 Further analysis was performed on age-specific subgroups, because physiological differences exist (eg, children have a predisposition for laryngospasms³⁶) and anesthesia practitioners may find this information clinically relevant. The study by Hung et al,¹⁷ which incorporated subjects 4 to 36 years of age, was included in our primary analysis, but was not included in subgroup analyses because subjects were not segmented into adult or pediatric populations.

Our primary pooled results suggest that the use of an LMA for airway management significantly reduces the risk of postoperative hypoxia when compared to intubation. This result was potentially limited by high risk of bias within the pooled studies, although sensitivity analysis still resulted in a significant effect of a reduced risk of postoperative hypoxia with LMA intervention. Performing a meta-analysis can require judgment decisions from the authors, and sensitivity analysis provides the reader with an alternative decision-making tree with which to assess the robustness of the results. Our sensitivity analysis removed studies judged to have a high risk of bias from the meta-analysis, which could have inaccurately influenced results, as any effects could have resulted from this methodology. Therefore, the similarities between the sensitivity analysis and the primary analysis give credence to our postoperative hypoxia result. The reduction of postoperative hypoxia occurrences was 78% (1 − RR 0.22) for all ages and 90% (1 − RR 0.10) among the pediatric population. This has significant clinical relevance because hypoxia can lead to patient mortality and serious morbidities including cardiac arrest and brain damage.¹³ Thus, this result may impact the anesthesia provider's choice of airway management. Alternatively, the presence of clinically relevant hypoxia is difficult to determine, not only because the definition of hypoxia varied (only Zhao et al¹⁹ defined hypoxia as an arterial oxygen saturation of 90% or less, whereas Kundra et al¹⁸ and Quinn et al¹⁶ defined an event as less than 95%) but also because the duration of each event was not clearly specified.

Furthermore, clinical interpretation may be limited by the wide confidence intervals (0.06-0.93 and 0.06-0.77) and the moderate level of evidence (Table 2). GRADE analysis (Table 2) was utilized to allow readers to obtain the quick clinical relevance of a comparison, specifically the certainty of a finding, as a factor of the risk of bias, indirections or relatability of the studies, publication bias, inconsistencies such as heterogeneity within the meta-analysis, and imprecision such as the confidence interval. The moderate level of certainty implies from the GRADE working group that “further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate” and that “true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.”²² As a comparison, high quality would suggest that “future research is very unlikely to change our confidence in the estimate of effect.”²²

A reason for the reduction in postoperative hypoxia may be related to the greater invasiveness of intubation.¹² This suggests a possible explanation that removal of the endotracheal tube may instigate increased airway reactivity and lead to postoperative hypoxia. Similarly, as all studies included in the hypoxia analysis performed extubation either “awake”^16,18 or “after regaining protective airway reflexes,”¹⁹ there exists the potential that an alternative deep-extubation technique may have produced different results. Although all the hypoxia-analyzed studies did not specify causation for hypoxia,^16,18,19 Kundra et al¹⁸ did note congruent airway issues, including breath holding, persistent coughing, laryngospasm, and a category noted as “other.” The lack of specifics surrounding the hypoxia events is a limitation in this meta-analysis. Perhaps greater relevance to the reader may be found in the study by Quinn et al,¹⁶ which identified frequent epistaxis in 38% of nasally intubated patients compared with 0% in LMA patients. It is conceivable that this difference in bleeding within the airway could correlate to postoperative hypoxia and should be further investigated.

We also noted a significant reduction of 92% (1 − RR 0.08) in the occurrence of sore throats among the pediatric population. This finding may relate to the growth changes in pediatric airways from a spherical shape to the elliptical shape seen in adults.³⁷ It is possible that an LMA exerts less pressure on vulnerable pharyngeal tissues of younger patients. The medical literature has evidence to support a reduced sore throat in pediatrics when an LMA is used,³⁸ though no statistical effect has been shown.³⁹ Reducing postsurgical pain, even in the oropharynx, is a consideration in postoperative care after a dental procedure under GA. Follow-up in these instances may present challenges, as a child may have difficulty distinguishing between dental pain and a sore throat, which may affect choice of analgesic management.

Excluding postoperative hypoxia, the primary meta-analysis (Figure 3) did not find a statistically significant effect from the LMA intervention on the following complications: airway obstruction, laryngeal soiling or aspiration, sore throat, and PONV. However, these findings may have clinical relevance because they suggest a statistical comparability between the gold standard of intubation and LMA intervention. One possible explanation for this could be that pharmacological agents or physiology of anesthesia has greater weight associated with airway complications than does the mode of airway management. Alternatively, the inability to detect a significant difference between the interventions may actually be a result of the limited number of articles and their associated sample sizes or power. A certainty level of low to very low was determined through the GRADE assessment relating to the above comparisons (Table 2). A low level would imply, “Further research is very likely to have an important impact on our confidence in the estimate of the effect and is likely to change the estimate.”²³ Therefore, extreme caution should be exercised when considering the relevance of our findings. However, the use of an LMA instead of intubation has been shown to be cost-effective for dental surgeries.⁴⁰ The readers should be aware that pooled analysis could not be performed for bronchospasm because only 1 set of data was available. Readers should also be aware of the considerable heterogeneity in the pooled results for airway obstruction, sore throat, and PONV, which may imply that the pooling of these individual studies for meta-analyses may be inappropriate.

Excluding postoperative hypoxia and pediatric sore throat, subgroup analysis (Figure 5) did not find a statistically significant effect difference between LMA and intubation for pediatric airway obstruction, pediatric airway soiling, or adult airway obstruction. All other airway complications could not be assessed because of lack of eligible studies. The lack of significant statistical differences mirrors the primary analysis for airway obstruction and airway soiling. As a result, this may further support the above-mentioned 3 possible explanations: that anesthetic management outside of the realm of airway management largely contributes to the risk of airway complications; that the risk of certain airway complications is truly similar between both interventions; and that as a result of the small sample sizes and poor GRADE assessment, it is impossible to detect an actual difference. Because of the scarcity of eligible papers, only airway obstruction can be compared between pediatric and adult subgroups. Analysis displayed no difference between interventions for these 2 subgroups. This may imply that any anatomical or physiological differences between age groups may not be a leading cause of airway risk. However, readers should appreciate that although the airway obstruction pediatric subgroup effect had a relatively small confidence interval with very low heterogeneity, the adult airway obstruction had a very wide confidence interval with significant heterogeneity. This may imply that pooling of data for this meta-analysis may not be appropriate. Readers should take caution with the statements that adult and pediatric airway obstruction risk differences are similar and that no effect difference is true. The most likely reason for this lack of confidence in the adult population is that the number of pooled subjects is near half of what was pooled in the comparative pediatric analysis.

To our knowledge, there is no previous review comparing airway management modalities within dentistry that may guide clinical practice. Although this review searched all types of nonintubated airway managements, only studies using an LMA intervention met our selection protocol.^17-19 This is an interesting finding considering use of a “nasopharyngeal or oropharyngeal airway” contributes to more than 60% of airway management within pediatric dental cases in the United States.⁴¹ Therefore, there is a significant gap in the current literature especially relating to “guarded-airway” techniques with or without the use of throat packs or adjunct airways. Readers should be aware that there is literature regarding the “guarded airway” and “nasopharyngeal or oropharyngeal airways” within dentistry, but this study did not identify those studies because the search strategy screened for RCTs.

Our findings are similar to those of the review by van Esch et al.⁴² However, they compared LMA use to intubation for a variety of medical surgeries that did not involve the airway. They stated that the proprietary LMA-Supreme could be associated with reduced airway complications and that LMAs overall have either similar complications or complications that cannot be statistically differentiated from intubation.⁴²

This meta-analysis recognizes that literature exists where other airway strategies are examined. Examples include the articles by Strauss et al²⁶ and Wang et al.⁴⁷ These articles underwent a full read for eligibility but were rejected because they were not RCTs. The clinical trial of 40 adults by Strauss et al²⁶ compared an intubation group to a “guarded” airway group (ie, patients that did not receive any advanced airway) and noted that although neither of the groups experienced bronchospasm or obstruction, PONV and sore throat were more common in the intubated group (20% vs 10% and 30% vs 0%, respectively). Wang et al⁴⁷ studied a prospective cohort of 200 subjects comprising adult and pediatric patients divided into 2 groups: intubation and nonintubation with nasal cannula. The study noted that the following airway complications were more common in the nonintubated group compared to the intubated group: hypoxia (9% vs. 0%), perioperative nausea or vomiting (7% vs. 0%), aspiration (4% vs. 0%), and obstruction (4% vs. 0%). The conflicting nature of these 2 studies, along with the absence of high-quality RCTs on the topic of non-LMA, nonintubated airway management, further illustrates the need for additional research.

Our review has several strengths. We applied a comprehensive search strategy and strict inclusion criteria to meet the assumption of similarity and to reduce selection bias. To reduce the potentials of human error, 2 reviewers independently screened articles, assessed for full-text eligibility, and performed data extraction, and 3 independent reviewers assessed risk of bias. In both cases, disagreements were settled with discussion or another author was approached. Our sensitivity and subgroup analyses were informative from the standpoint of potential impact of covariates that might have affected the outcome.

This meta-analysis has some limitations. Four non-English articles were excluded; this may have influenced the results, especially because conclusions reached only very low to moderate levels of certainty (Table 2). Three studies^15-17 had unclear risk of bias in the random sequence allocation, which may have potentially affected our outcome.^15-17 Furthermore, the majority of articles were judged to have an unclear risk of bias in the domain of allocation concealment.^15-17,19 The authors of this review made the decision that imperfect blinding or lack of blinding to study personnel and outcome assessment would likely not result in an increased risk of bias and that blinding may be unreasonable and potentially dangerous to participants. The authors also did not consider lack of blinding of participants to pose an increase in risk of bias. Additionally, small sample sizes utilized in eligible articles are a major concern to overall review quality. Given the relatively small number of studies included, the imprecision and uncertainty in our results is expected, and may be improved in future updates on this topic with the availability of new trials. Heterogeneity is a limitation to our results; however, heterogeneity among articles may be difficult to reduce within the field of anesthesia because of the vast array of practice motifs and medications available. Lastly, confounding factors such as differences in pharmacologic management may have the potential to exert influence over airway complication outcomes, which warrants further investigation.

CONCLUSION

In summary, this review may suggest that LMAs are an acceptable alternative to intubation for airway management during dental procedures. Moderate-quality evidence suggests that LMAs may potentially reduce the risk of postoperative hypoxia, particularly in the pediatric dental population, although further study is required. However, specifically regarding postoperative hypoxia, clinicians should take professional judgment and training into consideration regarding whether to implement LMAs into practice. This review's GRADE certainty levels were overall very low to moderate, suggesting further research is very likely to influence conclusions. There is a need for further well-designed RCTs comparing intubation with other nonintubated airway management strategies (ie, LMAs, nasopharyngeal airways, and other airway adjuncts).